From ca71598b3031aa90e44543abcd319674a7b4dabf Mon Sep 17 00:00:00 2001
From: Pat Alt <55311242+pat-alt@users.noreply.github.com>
Date: Wed, 13 Sep 2023 13:29:00 +0200
Subject: [PATCH] slowly slowly
---
bib.bib | 2837 ------------------------------------------
paper/aaai/paper.pdf | Bin 1944190 -> 1945527 bytes
paper/appendix.tex | 2 +-
paper/bib.bib | 636 +++++++---
paper/body.tex | 8 +-
5 files changed, 481 insertions(+), 3002 deletions(-)
delete mode 100644 bib.bib
diff --git a/bib.bib b/bib.bib
deleted file mode 100644
index daed3cba..00000000
--- a/bib.bib
+++ /dev/null
@@ -1,2837 +0,0 @@
-@TechReport{kingma2017adam,
- author = {Kingma, Diederik P. and Ba, Jimmy},
- date = {2017-01},
- institution = {arXiv},
- title = {Adam: {A} {Method} for {Stochastic} {Optimization}},
- doi = {10.48550/arXiv.1412.6980},
- note = {arXiv:1412.6980 [cs] type: article},
- url = {http://arxiv.org/abs/1412.6980},
- urldate = {2023-05-17},
- abstract = {We introduce Adam, an algorithm for first-order gradient-based optimization of stochastic objective functions, based on adaptive estimates of lower-order moments. The method is straightforward to implement, is computationally efficient, has little memory requirements, is invariant to diagonal rescaling of the gradients, and is well suited for problems that are large in terms of data and/or parameters. The method is also appropriate for non-stationary objectives and problems with very noisy and/or sparse gradients. The hyper-parameters have intuitive interpretations and typically require little tuning. Some connections to related algorithms, on which Adam was inspired, are discussed. We also analyze the theoretical convergence properties of the algorithm and provide a regret bound on the convergence rate that is comparable to the best known results under the online convex optimization framework. Empirical results demonstrate that Adam works well in practice and compares favorably to other stochastic optimization methods. Finally, we discuss AdaMax, a variant of Adam based on the infinity norm.},
- annotation = {Comment: Published as a conference paper at the 3rd International Conference for Learning Representations, San Diego, 2015},
- file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/1412.6980.pdf:application/pdf},
- keywords = {Computer Science - Machine Learning},
- shorttitle = {Adam},
-}
-
-@TechReport{xiao2017fashion,
- author = {Xiao, Han and Rasul, Kashif and Vollgraf, Roland},
- date = {2017-09},
- institution = {arXiv},
- title = {Fashion-{MNIST}: a {Novel} {Image} {Dataset} for {Benchmarking} {Machine} {Learning} {Algorithms}},
- doi = {10.48550/arXiv.1708.07747},
- note = {arXiv:1708.07747 [cs, stat] type: article},
- url = {http://arxiv.org/abs/1708.07747},
- urldate = {2023-05-10},
- abstract = {We present Fashion-MNIST, a new dataset comprising of 28x28 grayscale images of 70,000 fashion products from 10 categories, with 7,000 images per category. The training set has 60,000 images and the test set has 10,000 images. Fashion-MNIST is intended to serve as a direct drop-in replacement for the original MNIST dataset for benchmarking machine learning algorithms, as it shares the same image size, data format and the structure of training and testing splits. The dataset is freely available at https://github.com/zalandoresearch/fashion-mnist},
- annotation = {Comment: Dataset is freely available at https://github.com/zalandoresearch/fashion-mnist Benchmark is available at http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/},
- file = {:xiao2017fashion - Fashion MNIST_ a Novel Image Dataset for Benchmarking Machine Learning Algorithms.pdf:PDF},
- keywords = {Computer Science - Machine Learning, Computer Science - Computer Vision and Pattern Recognition, Statistics - Machine Learning},
- shorttitle = {Fashion-{MNIST}},
-}
-
-@Online{mw2023fidelity,
- author = {Merriam-Webster},
- title = {"Fidelity"},
- url = {https://www.merriam-webster.com/dictionary/fidelity},
- language = {en},
- organization = {Merriam-Webster},
- urldate = {2023-03-23},
- abstract = {the quality or state of being faithful; accuracy in details : exactness; the degree to which an electronic device (such as a record player, radio, or television) accurately reproduces its effect (such as sound or picture)… See the full definition},
-}
-
-@InProceedings{altmeyer2023endogenous,
- author = {Altmeyer, Patrick and Angela, Giovan and Buszydlik, Aleksander and Dobiczek, Karol and van Deursen, Arie and Liem, Cynthia},
- booktitle = {First {IEEE} {Conference} on {Secure} and {Trustworthy} {Machine} {Learning}},
- title = {Endogenous {Macrodynamics} in {Algorithmic} {Recourse}},
- file = {:altmeyerendogenous - Endogenous Macrodynamics in Algorithmic Recourse.pdf:PDF},
- year = {2023},
-}
-
-%% This BibTeX bibliography file was created using BibDesk.
-%% https://bibdesk.sourceforge.io/
-
-%% Created for Anonymous Author at 2022-12-13 12:58:22 +0100
-
-
-%% Saved with string encoding Unicode (UTF-8)
-
-
-
-@Article{abadie2002instrumental,
- author = {Abadie, Alberto and Angrist, Joshua and Imbens, Guido},
- title = {Instrumental Variables Estimates of the Effect of Subsidized Training on the Quantiles of Trainee Earnings},
- number = {1},
- pages = {91--117},
- volume = {70},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Econometrica : journal of the Econometric Society},
- shortjournal = {Econometrica},
- year = {2002},
-}
-
-@Article{abadie2003economic,
- author = {Abadie, Alberto and Gardeazabal, Javier},
- title = {The Economic Costs of Conflict: {{A}} Case Study of the {{Basque Country}}},
- number = {1},
- pages = {113--132},
- volume = {93},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {American economic review},
- year = {2003},
-}
-
-@InProceedings{ackerman2021machine,
- author = {Ackerman, Samuel and Dube, Parijat and Farchi, Eitan and Raz, Orna and Zalmanovici, Marcel},
- booktitle = {2021 {{IEEE}}/{{ACM Third International Workshop}} on {{Deep Learning}} for {{Testing}} and {{Testing}} for {{Deep Learning}} ({{DeepTest}})},
- title = {Machine {{Learning Model Drift Detection Via Weak Data Slices}}},
- pages = {1--8},
- publisher = {{IEEE}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2021},
-}
-
-@Article{allen2017referencedependent,
- author = {Allen, Eric J and Dechow, Patricia M and Pope, Devin G and Wu, George},
- title = {Reference-Dependent Preferences: {{Evidence}} from Marathon Runners},
- number = {6},
- pages = {1657--1672},
- volume = {63},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Management Science},
- year = {2017},
-}
-
-@Article{altmeyer2018option,
- author = {Altmeyer, Patrick and Grapendal, Jacob Daniel and Pravosud, Makar and Quintana, Gand Derry},
- title = {Option Pricing in the {{Heston}} Stochastic Volatility Model: An Empirical Evaluation},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2018},
-}
-
-@Article{altmeyer2021deep,
- author = {Altmeyer, Patrick and Agusti, Marc and Vidal-Quadras Costa, Ignacio},
- title = {Deep {{Vector Autoregression}} for {{Macroeconomic Data}}},
- url = {https://thevoice.bse.eu/wp-content/uploads/2021/07/ds21-project-agusti-et-al.pdf},
- bdsk-url-1 = {https://thevoice.bse.eu/wp-content/uploads/2021/07/ds21-project-agusti-et-al.pdf},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2021},
-}
-
-@Book{altmeyer2021deepvars,
- author = {Altmeyer, Patrick},
- title = {Deepvars: {{Deep Vector Autoregession}}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2021},
-}
-
-@Misc{altmeyer2022counterfactualexplanations,
- author = {Altmeyer, Patrick},
- title = {{{CounterfactualExplanations}}.Jl - a {{Julia}} Package for {{Counterfactual Explanations}} and {{Algorithmic Recourse}}},
- url = {https://github.com/pat-alt/CounterfactualExplanations.jl},
- bdsk-url-1 = {https://github.com/pat-alt/CounterfactualExplanations.jl},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2022},
-}
-
-@Software{altmeyerCounterfactualExplanationsJlJulia2022,
- author = {Altmeyer, Patrick},
- title = {{{CounterfactualExplanations}}.Jl - a {{Julia}} Package for {{Counterfactual Explanations}} and {{Algorithmic Recourse}}},
- url = {https://github.com/pat-alt/CounterfactualExplanations.jl},
- version = {0.1.2},
- bdsk-url-1 = {https://github.com/pat-alt/CounterfactualExplanations.jl},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2022},
-}
-
-@Unpublished{angelopoulos2021gentle,
- author = {Angelopoulos, Anastasios N. and Bates, Stephen},
- title = {A Gentle Introduction to Conformal Prediction and Distribution-Free Uncertainty Quantification},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2107.07511},
- eprinttype = {arxiv},
- file = {:/Users/FA31DU/Zotero/storage/RKSUMYZG/Angelopoulos and Bates - 2021 - A gentle introduction to conformal prediction and .pdf:;:/Users/FA31DU/Zotero/storage/PRUEKRR3/2107.html:},
- year = {2021},
-}
-
-@Misc{angelopoulos2022uncertainty,
- author = {Angelopoulos, Anastasios and Bates, Stephen and Malik, Jitendra and Jordan, Michael I.},
- title = {Uncertainty {{Sets}} for {{Image Classifiers}} Using {{Conformal Prediction}}},
- eprint = {2009.14193},
- eprinttype = {arxiv},
- abstract = {Convolutional image classifiers can achieve high predictive accuracy, but quantifying their uncertainty remains an unresolved challenge, hindering their deployment in consequential settings. Existing uncertainty quantification techniques, such as Platt scaling, attempt to calibrate the network's probability estimates, but they do not have formal guarantees. We present an algorithm that modifies any classifier to output a predictive set containing the true label with a user-specified probability, such as 90\%. The algorithm is simple and fast like Platt scaling, but provides a formal finite-sample coverage guarantee for every model and dataset. Our method modifies an existing conformal prediction algorithm to give more stable predictive sets by regularizing the small scores of unlikely classes after Platt scaling. In experiments on both Imagenet and Imagenet-V2 with ResNet-152 and other classifiers, our scheme outperforms existing approaches, achieving coverage with sets that are often factors of 5 to 10 smaller than a stand-alone Platt scaling baseline.},
- archiveprefix = {arXiv},
- bdsk-url-1 = {http://arxiv.org/abs/2009.14193},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- file = {:/Users/FA31DU/Zotero/storage/5BYIRBR2/Angelopoulos et al. - 2022 - Uncertainty Sets for Image Classifiers using Confo.pdf:;:/Users/FA31DU/Zotero/storage/2QJAKFKV/2009.html:},
- keywords = {Computer Science - Computer Vision and Pattern Recognition, Mathematics - Statistics Theory, Statistics - Machine Learning},
- month = sep,
- number = {arXiv:2009.14193},
- primaryclass = {cs, math, stat},
- publisher = {{arXiv}},
- year = {2022},
-}
-
-@Article{angelucci2009indirect,
- author = {Angelucci, Manuela and De Giorgi, Giacomo},
- title = {Indirect Effects of an Aid Program: How Do Cash Transfers Affect Ineligibles' Consumption?},
- number = {1},
- pages = {486--508},
- volume = {99},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {American economic review},
- year = {2009},
-}
-
-@Article{angrist1990lifetime,
- author = {Angrist, Joshua D},
- title = {Lifetime Earnings and the {{Vietnam}} Era Draft Lottery: Evidence from Social Security Administrative Records},
- pages = {313--336},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The American Economic Review},
- year = {1990},
-}
-
-@Unpublished{antoran2020getting,
- author = {Antor{\'a}n, Javier and Bhatt, Umang and Adel, Tameem and Weller, Adrian and Hern{\'a}ndez-Lobato, Jos{\'e} Miguel},
- title = {Getting a Clue: {{A}} Method for Explaining Uncertainty Estimates},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2006.06848},
- eprinttype = {arxiv},
- year = {2020},
-}
-
-@Article{arcones1992bootstrap,
- author = {Arcones, Miguel A and Gine, Evarist},
- title = {On the Bootstrap of {{U}} and {{V}} Statistics},
- pages = {655--674},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The Annals of Statistics},
- year = {1992},
-}
-
-@Article{ariely2003coherent,
- author = {Ariely, Dan and Loewenstein, George and Prelec, Drazen},
- title = {``{{Coherent}} Arbitrariness'': {{Stable}} Demand Curves without Stable Preferences},
- number = {1},
- pages = {73--106},
- volume = {118},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The Quarterly journal of economics},
- year = {2003},
-}
-
-@Article{ariely2006tom,
- author = {Ariely, Dan and Loewenstein, George and Prelec, Drazen},
- title = {Tom {{Sawyer}} and the Construction of Value},
- number = {1},
- pages = {1--10},
- volume = {60},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Economic Behavior \& Organization},
- year = {2006},
-}
-
-@Article{arrieta2020explainable,
- author = {Arrieta, Alejandro Barredo and Diaz-Rodriguez, Natalia and Del Ser, Javier and Bennetot, Adrien and Tabik, Siham and Barbado, Alberto and Garcia, Salvador and Gil-Lopez, Sergio and Molina, Daniel and Benjamins, Richard and others},
- title = {Explainable {{Artificial Intelligence}} ({{XAI}}): {{Concepts}}, Taxonomies, Opportunities and Challenges toward Responsible {{AI}}},
- pages = {82--115},
- volume = {58},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Information Fusion},
- year = {2020},
-}
-
-@Article{auer2002finitetime,
- author = {Auer, Peter and Cesa-Bianchi, Nicolo and Fischer, Paul},
- title = {Finite-Time Analysis of the Multiarmed Bandit Problem},
- number = {2},
- pages = {235--256},
- volume = {47},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Machine learning},
- year = {2002},
-}
-
-@Article{barabasi2016network,
- author = {Barab{\'a}si, Albert-L{\'a}szl{\'o}},
- title = {Network {{Science}}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Network Science},
- year = {2016},
-}
-
-@Unpublished{bastounis2021mathematics,
- author = {Bastounis, Alexander and Hansen, Anders C and Vla{\v c}i{\'c}, Verner},
- title = {The Mathematics of Adversarial Attacks in {{AI}}--{{Why}} Deep Learning Is Unstable despite the Existence of Stable Neural Networks},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2109.06098},
- eprinttype = {arxiv},
- year = {2021},
-}
-
-@Article{bechara1997deciding,
- author = {Bechara, Antoine and Damasio, Hanna and Tranel, Daniel and Damasio, Antonio R},
- title = {Deciding Advantageously before Knowing the Advantageous Strategy},
- number = {5304},
- pages = {1293--1295},
- volume = {275},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Science (New York, N.Y.)},
- shortjournal = {Science},
- year = {1997},
-}
-
-@Book{berlinet2011reproducing,
- author = {Berlinet, Alain and Thomas-Agnan, Christine},
- title = {Reproducing Kernel {{Hilbert}} Spaces in Probability and Statistics},
- publisher = {{Springer Science \& Business Media}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2011},
-}
-
-@Misc{bernanke1990federal,
- author = {Bernanke, Ben S},
- title = {The Federal Funds Rate and the Channels of Monetary Transnission},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- publisher = {{National Bureau of Economic Research Cambridge, Mass., USA}},
- year = {1990},
-}
-
-@Article{besbes2014stochastic,
- author = {Besbes, Omar and Gur, Yonatan and Zeevi, Assaf},
- title = {Stochastic Multi-Armed-Bandit Problem with Non-Stationary Rewards},
- pages = {199--207},
- volume = {27},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Advances in neural information processing systems},
- year = {2014},
-}
-
-@Article{bholat2020impact,
- author = {Bholat, D and Gharbawi, M and Thew, O},
- title = {The {{Impact}} of {{Covid}} on {{Machine Learning}} and {{Data Science}} in {{UK Banking}}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Bank of England Quarterly Bulletin, Q4},
- year = {2020},
-}
-
-@Book{bishop2006pattern,
- author = {Bishop, Christopher M},
- title = {Pattern Recognition and Machine Learning},
- publisher = {{springer}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2006},
-}
-
-@Article{blaom2020mlj,
- author = {Blaom, Anthony D. and Kiraly, Franz and Lienart, Thibaut and Simillides, Yiannis and Arenas, Diego and Vollmer, Sebastian J.},
- title = {{{MLJ}}: {{A Julia}} Package for Composable Machine Learning},
- doi = {10.21105/joss.02704},
- issn = {2475-9066},
- number = {55},
- pages = {2704},
- urldate = {2022-10-27},
- volume = {5},
- abstract = {Blaom et al., (2020). MLJ: A Julia package for composable machine learning. Journal of Open Source Software, 5(55), 2704, https://doi.org/10.21105/joss.02704},
- bdsk-url-1 = {https://joss.theoj.org/papers/10.21105/joss.02704},
- bdsk-url-2 = {https://doi.org/10.21105/joss.02704},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- file = {:/Users/FA31DU/Zotero/storage/7AY87FGP/Blaom et al. - 2020 - MLJ A Julia package for composable machine learni.pdf:;:/Users/FA31DU/Zotero/storage/D69YSMVF/joss.html:},
- journal = {Journal of Open Source Software},
- langid = {english},
- month = nov,
- shorttitle = {{{MLJ}}},
- year = {2020},
-}
-
-@InProceedings{blundell2015weight,
- author = {Blundell, Charles and Cornebise, Julien and Kavukcuoglu, Koray and Wierstra, Daan},
- booktitle = {International Conference on Machine Learning},
- title = {Weight Uncertainty in Neural Network},
- pages = {1613--1622},
- publisher = {{PMLR}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2015},
-}
-
-@Article{borch2022machine,
- author = {Borch, Christian},
- title = {Machine Learning, Knowledge Risk, and Principal-Agent Problems in Automated Trading},
- pages = {101852},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Technology in Society},
- year = {2022},
-}
-
-@Unpublished{borisov2021deep,
- author = {Borisov, Vadim and Leemann, Tobias and Se{\ss}ler, Kathrin and Haug, Johannes and Pawelczyk, Martin and Kasneci, Gjergji},
- title = {Deep Neural Networks and Tabular Data: {{A}} Survey},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2110.01889},
- eprinttype = {arxiv},
- year = {2021},
-}
-
-@Article{bramoulle2009identification,
- author = {Bramoull{\'e}, Yann and Djebbari, Habiba and Fortin, Bernard},
- title = {Identification of Peer Effects through Social Networks},
- number = {1},
- pages = {41--55},
- volume = {150},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of econometrics},
- year = {2009},
-}
-
-@Article{bramoulle2020peer,
- author = {Bramoull{\'e}, Yann and Djebbari, Habiba and Fortin, Bernard},
- title = {Peer Effects in Networks: {{A}} Survey},
- pages = {603--629},
- volume = {12},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Annual Review of Economics},
- year = {2020},
-}
-
-@Unpublished{branco2015survey,
- author = {Branco, Paula and Torgo, Luis and Ribeiro, Rita},
- title = {A Survey of Predictive Modelling under Imbalanced Distributions},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1505.01658},
- eprinttype = {arxiv},
- year = {2015},
-}
-
-@Book{brock1991nonlinear,
- author = {Brock, William Allen and Brock, William A and Hsieh, David Arthur and LeBaron, Blake Dean and Brock, William E},
- title = {Nonlinear Dynamics, Chaos, and Instability: Statistical Theory and Economic Evidence},
- publisher = {{MIT press}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {1991},
-}
-
-@InProceedings{buolamwini2018gender,
- author = {Buolamwini, Joy and Gebru, Timnit},
- booktitle = {Conference on Fairness, Accountability and Transparency},
- title = {Gender Shades: {{Intersectional}} Accuracy Disparities in Commercial Gender Classification},
- pages = {77--91},
- publisher = {{PMLR}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2018},
-}
-
-@Unpublished{bussmann2020neural,
- author = {Bussmann, Bart and Nys, Jannes and Latr{\'e}, Steven},
- title = {Neural {{Additive Vector Autoregression Models}} for {{Causal Discovery}} in {{Time Series Data}}},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2010.09429},
- eprinttype = {arxiv},
- year = {2020},
-}
-
-@Report{card1993minimum,
- author = {Card, David and Krueger, Alan B},
- title = {Minimum Wages and Employment: {{A}} Case Study of the Fast Food Industry in {{New Jersey}} and {{Pennsylvania}}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- school = {{National Bureau of Economic Research}},
- year = {1993},
-}
-
-@InProceedings{carlini2017evaluating,
- author = {Carlini, Nicholas and Wagner, David},
- booktitle = {2017 Ieee Symposium on Security and Privacy (Sp)},
- title = {Towards Evaluating the Robustness of Neural Networks},
- pages = {39--57},
- publisher = {{IEEE}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2017},
-}
-
-@Article{carlisle2019racist,
- author = {Carlisle, M.},
- title = {Racist Data Destruction? - a {{Boston}} Housing Dataset Controversy},
- url = {https://medium.com/@docintangible/racist-data-destruction-113e3eff54a8},
- bdsk-url-1 = {https://medium.com/@docintangible/racist-data-destruction-113e3eff54a8},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2019},
-}
-
-@Article{carrell2009does,
- author = {Carrell, Scott E and Fullerton, Richard L and West, James E},
- title = {Does Your Cohort Matter? {{Measuring}} Peer Effects in College Achievement},
- number = {3},
- pages = {439--464},
- volume = {27},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Labor Economics},
- year = {2009},
-}
-
-@Article{carrell2013natural,
- author = {Carrell, Scott E and Sacerdote, Bruce I and West, James E},
- title = {From Natural Variation to Optimal Policy? {{The}} Importance of Endogenous Peer Group Formation},
- number = {3},
- pages = {855--882},
- volume = {81},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Econometrica : journal of the Econometric Society},
- shortjournal = {Econometrica},
- year = {2013},
-}
-
-@Article{carrizosa2021generating,
- author = {Carrizosa, Emilio and Ramırez-Ayerbe, Jasone and Romero, Dolores},
- title = {Generating {{Collective Counterfactual Explanations}} in {{Score-Based Classification}} via {{Mathematical Optimization}}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2021},
-}
-
-@Article{cascarino2022explainable,
- author = {Cascarino, Giuseppe and Moscatelli, Mirko and Parlapiano, Fabio},
- title = {Explainable {{Artificial Intelligence}}: Interpreting Default Forecasting Models Based on {{Machine Learning}}},
- number = {674},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Bank of Italy Occasional Paper},
- year = {2022},
-}
-
-@Article{chandola2009anomaly,
- author = {Chandola, Varun and Banerjee, Arindam and Kumar, Vipin},
- title = {Anomaly Detection: {{A}} Survey},
- number = {3},
- pages = {1--58},
- volume = {41},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {ACM computing surveys (CSUR)},
- year = {2009},
-}
-
-@Article{chapelle2011empirical,
- author = {Chapelle, Olivier and Li, Lihong},
- title = {An Empirical Evaluation of Thompson Sampling},
- pages = {2249--2257},
- volume = {24},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Advances in neural information processing systems},
- year = {2011},
-}
-
-@Article{chetty2011adjustment,
- author = {Chetty, Raj and Friedman, John N and Olsen, Tore and Pistaferri, Luigi},
- title = {Adjustment Costs, Firm Responses, and Micro vs. Macro Labor Supply Elasticities: {{Evidence}} from {{Danish}} Tax Records},
- number = {2},
- pages = {749--804},
- volume = {126},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The quarterly journal of economics},
- year = {2011},
-}
-
-@Article{cortes1995supportvector,
- author = {Cortes, Corinna and Vapnik, Vladimir},
- title = {Support-Vector Networks},
- number = {3},
- pages = {273--297},
- volume = {20},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Machine learning},
- year = {1995},
-}
-
-@Article{crawford2019variable,
- author = {Crawford, Lorin and Flaxman, Seth R and Runcie, Daniel E and West, Mike},
- title = {Variable Prioritization in Nonlinear Black Box Methods: {{A}} Genetic Association Case Study},
- number = {2},
- pages = {958},
- volume = {13},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The annals of applied statistics},
- year = {2019},
-}
-
-@InProceedings{dai2022counterfactual,
- author = {Dai, Xinyue and Keane, Mark T and Shalloo, Laurence and Ruelle, Elodie and Byrne, Ruth MJ},
- title = {Counterfactual Explanations for Prediction and Diagnosis in Xai},
- eventtitle = {Proceedings of the 2022 {{AAAI}}/{{ACM Conference}} on {{AI}}, {{Ethics}}, and {{Society}}},
- pages = {215--226},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2022},
-}
-
-@Article{danielsson2021artificial,
- author = {Danielsson, Jon and Macrae, Robert and Uthemann, Andreas},
- title = {Artificial Intelligence and Systemic Risk},
- pages = {106290},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Banking \& Finance},
- year = {2021},
-}
-
-@Article{daxberger2021laplace,
- author = {Daxberger, Erik and Kristiadi, Agustinus and Immer, Alexander and Eschenhagen, Runa and Bauer, Matthias and Hennig, Philipp},
- title = {Laplace {{Redux-Effortless Bayesian Deep Learning}}},
- volume = {34},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Advances in Neural Information Processing Systems},
- year = {2021},
-}
-
-@Article{dehejia1999causal,
- author = {Dehejia, Rajeev H and Wahba, Sadek},
- title = {Causal Effects in Nonexperimental Studies: {{Reevaluating}} the Evaluation of Training Programs},
- number = {448},
- pages = {1053--1062},
- volume = {94},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of the American statistical Association},
- year = {1999},
-}
-
-@Article{dell2010persistent,
- author = {Dell, Melissa},
- title = {The Persistent Effects of {{Peru}}'s Mining Mita},
- number = {6},
- pages = {1863--1903},
- volume = {78},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Econometrica : journal of the Econometric Society},
- shortjournal = {Econometrica},
- year = {2010},
-}
-
-@Article{denhengst2020reinforcement,
- author = {den Hengst, Floris and Grua, Eoin Martino and el Hassouni, Ali and Hoogendoorn, Mark},
- title = {Reinforcement Learning for Personalization: {{A}} Systematic Literature Review},
- issue = {Preprint},
- pages = {1--41},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Data Science},
- options = {useprefix=true},
- year = {2020},
-}
-
-@Article{deoliveira2021framework,
- author = {de Oliveira, Raphael Mazzine Barbosa and Martens, David},
- title = {A Framework and Benchmarking Study for Counterfactual Generating Methods on Tabular Data},
- number = {16},
- pages = {7274},
- volume = {11},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Applied Sciences},
- options = {useprefix=true},
- year = {2021},
-}
-
-@Article{dhurandhar2018explanations,
- author = {Dhurandhar, Amit and Chen, Pin-Yu and Luss, Ronny and Tu, Chun-Chen and Ting, Paishun and Shanmugam, Karthikeyan and Das, Payel},
- title = {Explanations Based on the Missing: {{Towards}} Contrastive Explanations with Pertinent Negatives},
- volume = {31},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Advances in neural information processing systems},
- year = {2018},
-}
-
-@InProceedings{dombrowski2021diffeomorphic,
- author = {Dombrowski, Ann-Kathrin and Gerken, Jan E and Kessel, Pan},
- booktitle = {{{ICML Workshop}} on {{Invertible Neural Networks}}, {{Normalizing Flows}}, and {{Explicit Likelihood Models}}},
- title = {Diffeomorphic Explanations with Normalizing Flows},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2021},
-}
-
-@InProceedings{dorffner1996neural,
- author = {Dorffner, Georg},
- booktitle = {Neural Network World},
- title = {Neural Networks for Time Series Processing},
- publisher = {{Citeseer}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {1996},
-}
-
-@Article{epstein1979stability,
- author = {Epstein, Seymour},
- title = {The Stability of Behavior: {{I}}. {{On}} Predicting Most of the People Much of the Time.},
- number = {7},
- pages = {1097},
- volume = {37},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of personality and social psychology},
- year = {1979},
-}
-
-@Online{barocas2022fairness,
- author = {Solon Barocas and Moritz Hardt and Arvind Narayanan},
- title = {Fairness and Machine Learning},
- url = {https://fairmlbook.org/index.html},
- urldate = {2022-11-08},
- bdsk-url-1 = {https://fairmlbook.org/index.html},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- month = dec,
- year = {2022},
-}
-
-@Article{falk2006clean,
- author = {Falk, Armin and Ichino, Andrea},
- title = {Clean Evidence on Peer Effects},
- number = {1},
- pages = {39--57},
- volume = {24},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of labor economics},
- year = {2006},
-}
-
-@Unpublished{fan2020interpretability,
- author = {Fan, Fenglei and Xiong, Jinjun and Wang, Ge},
- title = {On Interpretability of Artificial Neural Networks},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2001.02522},
- eprinttype = {arxiv},
- year = {2020},
-}
-
-@Article{fang2011dynamic,
- author = {Fang, Hanming and Gavazza, Alessandro},
- title = {Dynamic Inefficiencies in an Employment-Based Health Insurance System: {{Theory}} and Evidence},
- number = {7},
- pages = {3047--77},
- volume = {101},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {American Economic Review},
- year = {2011},
-}
-
-@Article{fehr2000cooperation,
- author = {Fehr, Ernst and Gachter, Simon},
- title = {Cooperation and Punishment in Public Goods Experiments},
- number = {4},
- pages = {980--994},
- volume = {90},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {American Economic Review},
- year = {2000},
-}
-
-@Article{fix1951important,
- author = {Fix, E and Hodges, J},
- title = {An Important Contribution to Nonparametric Discriminant Analysis and Density Estimation},
- number = {57},
- pages = {233--238},
- volume = {3},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {International Statistical Review},
- year = {1951},
-}
-
-@Book{friedman2008monetary,
- author = {Friedman, Milton and Schwartz, Anna Jacobson},
- title = {A Monetary History of the {{United States}}, 1867-1960},
- publisher = {{Princeton University Press}},
- volume = {14},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2008},
-}
-
-@InProceedings{gal2016dropout,
- author = {Gal, Yarin and Ghahramani, Zoubin},
- booktitle = {International Conference on Machine Learning},
- title = {Dropout as a Bayesian Approximation: {{Representing}} Model Uncertainty in Deep Learning},
- pages = {1050--1059},
- publisher = {{PMLR}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2016},
-}
-
-@InProceedings{gal2017deep,
- author = {Gal, Yarin and Islam, Riashat and Ghahramani, Zoubin},
- booktitle = {International {{Conference}} on {{Machine Learning}}},
- title = {Deep Bayesian Active Learning with Image Data},
- pages = {1183--1192},
- publisher = {{PMLR}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2017},
-}
-
-@Article{galizzi2019external,
- author = {Galizzi, Matteo M and Navarro-Martinez, Daniel},
- title = {On the External Validity of Social Preference Games: A Systematic Lab-Field Study},
- number = {3},
- pages = {976--1002},
- volume = {65},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Management Science},
- year = {2019},
-}
-
-@Article{gama2014survey,
- author = {Gama, Jo{\~a}o and {\v Z}liobait{\.e}, Indr{\.e} and Bifet, Albert and Pechenizkiy, Mykola and Bouchachia, Abdelhamid},
- title = {A Survey on Concept Drift Adaptation},
- number = {4},
- pages = {1--37},
- volume = {46},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {ACM computing surveys (CSUR)},
- year = {2014},
-}
-
-@Unpublished{garivier2008upperconfidence,
- author = {Garivier, Aur{\'e}lien and Moulines, Eric},
- title = {On Upper-Confidence Bound Policies for Non-Stationary Bandit Problems},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {0805.3415},
- eprinttype = {arxiv},
- year = {2008},
-}
-
-@Book{gelman2013bayesian,
- author = {Gelman, Andrew and Carlin, John B and Stern, Hal S and Dunson, David B and Vehtari, Aki and Rubin, Donald B},
- title = {Bayesian Data Analysis},
- publisher = {{CRC press}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2013},
-}
-
-@Article{gilbert1998immune,
- author = {Gilbert, Daniel T and Pinel, Elizabeth C and Wilson, Timothy D and Blumberg, Stephen J and Wheatley, Thalia P},
- title = {Immune Neglect: A Source of Durability Bias in Affective Forecasting.},
- number = {3},
- pages = {617},
- volume = {75},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of personality and social psychology},
- year = {1998},
-}
-
-@Article{gneezy2006uncertainty,
- author = {Gneezy, Uri and List, John A and Wu, George},
- title = {The Uncertainty Effect: {{When}} a Risky Prospect Is Valued Less than Its Worst Possible Outcome},
- number = {4},
- pages = {1283--1309},
- volume = {121},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The Quarterly Journal of Economics},
- year = {2006},
-}
-
-@InCollection{goan2020bayesian,
- author = {Goan, Ethan and Fookes, Clinton},
- booktitle = {Case {{Studies}} in {{Applied Bayesian Data Science}}},
- title = {Bayesian {{Neural Networks}}: {{An Introduction}} and {{Survey}}},
- pages = {45--87},
- publisher = {{Springer}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2020},
-}
-
-@Article{goldsmith-pinkham2013social,
- author = {Goldsmith-Pinkham, Paul and Imbens, Guido W},
- title = {Social Networks and the Identification of Peer Effects},
- number = {3},
- pages = {253--264},
- volume = {31},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Business \& Economic Statistics},
- year = {2013},
-}
-
-@Unpublished{goodfellow2014explaining,
- author = {Goodfellow, Ian J and Shlens, Jonathon and Szegedy, Christian},
- title = {Explaining and Harnessing Adversarial Examples},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1412.6572},
- eprinttype = {arxiv},
- year = {2014},
-}
-
-@Book{goodfellow2016deep,
- author = {Goodfellow, Ian and Bengio, Yoshua and Courville, Aaron},
- title = {Deep {{Learning}}},
- publisher = {{MIT Press}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2016},
-}
-
-@Article{goodfriend2005incredible,
- author = {Goodfriend, Marvin and King, Robert G},
- title = {The Incredible {{Volcker}} Disinflation},
- number = {5},
- pages = {981--1015},
- volume = {52},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Monetary Economics},
- year = {2005},
-}
-
-@Article{graham2017econometric,
- author = {Graham, Bryan S},
- title = {An Econometric Model of Network Formation with Degree Heterogeneity},
- number = {4},
- pages = {1033--1063},
- volume = {85},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Econometrica : journal of the Econometric Society},
- shortjournal = {Econometrica},
- year = {2017},
-}
-
-@Article{greene2012econometric,
- author = {Greene, William H},
- title = {Econometric Analysis, 71e},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Stern School of Business, New York University},
- year = {2012},
-}
-
-@Article{grether1979economic,
- author = {Grether, David M and Plott, Charles R},
- title = {Economic Theory of Choice and the Preference Reversal Phenomenon},
- number = {4},
- pages = {623--638},
- volume = {69},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The American Economic Review},
- year = {1979},
-}
-
-@Article{gretton2012kernel,
- author = {Gretton, Arthur and Borgwardt, Karsten M and Rasch, Malte J and Sch{\"o}lkopf, Bernhard and Smola, Alexander},
- title = {A Kernel Two-Sample Test},
- number = {1},
- pages = {723--773},
- volume = {13},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The Journal of Machine Learning Research},
- year = {2012},
-}
-
-@Unpublished{griffith2020name,
- author = {Griffith, Alan},
- title = {Name {{Your Friends}}, but {{Only Five}}? {{The Importance}} of {{Censoring}} in {{Peer Effects Estimates}} Using {{Social Network Data}}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2020},
-}
-
-@Unpublished{grinsztajn2022why,
- author = {Grinsztajn, L{\'e}o and Oyallon, Edouard and Varoquaux, Ga{\"e}l},
- title = {Why Do Tree-Based Models Still Outperform Deep Learning on Tabular Data?},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2207.08815},
- eprinttype = {arxiv},
- year = {2022},
-}
-
-@Misc{group2020detailed,
- author = {Group, Open COVID-19 Data Working},
- title = {Detailed {{Epidemiological Data}} from the {{COVID-19 Outbreak}}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2020},
-}
-
-@InProceedings{gupta2011thompson,
- author = {Gupta, Neha and Granmo, Ole-Christoffer and Agrawala, Ashok},
- booktitle = {2011 10th {{International Conference}} on {{Machine Learning}} and {{Applications}} and {{Workshops}}},
- title = {Thompson Sampling for Dynamic Multi-Armed Bandits},
- pages = {484--489},
- publisher = {{IEEE}},
- volume = {1},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2011},
-}
-
-@Book{hamilton2020time,
- author = {Hamilton, James Douglas},
- title = {Time Series Analysis},
- publisher = {{Princeton university press}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2020},
-}
-
-@Article{hamon2020robustness,
- author = {Hamon, Ronan and Junklewitz, Henrik and Sanchez, Ignacio},
- title = {Robustness and Explainability of Artificial Intelligence},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Publications Office of the European Union},
- year = {2020},
-}
-
-@Article{hamzacebi2008improving,
- author = {Hamza{\c c}ebi, Co{\c s}kun},
- title = {Improving Artificial Neural Networks' Performance in Seasonal Time Series Forecasting},
- number = {23},
- pages = {4550--4559},
- volume = {178},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Information Sciences},
- year = {2008},
-}
-
-@InProceedings{hanneke2007bound,
- author = {Hanneke, Steve},
- booktitle = {Proceedings of the 24th International Conference on {{Machine}} Learning},
- title = {A Bound on the Label Complexity of Agnostic Active Learning},
- pages = {353--360},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2007},
-}
-
-@Article{hansen2020virtue,
- author = {Hansen, Kristian Bondo},
- title = {The Virtue of Simplicity: {{On}} Machine Learning Models in Algorithmic Trading},
- number = {1},
- pages = {2053951720926558},
- volume = {7},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Big Data \& Society},
- year = {2020},
-}
-
-@Article{hartland2006multiarmed,
- author = {Hartland, C{\'e}dric and Gelly, Sylvain and Baskiotis, Nicolas and Teytaud, Olivier and Sebag, Michele},
- title = {Multi-Armed Bandit, Dynamic Environments and Meta-Bandits},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2006},
-}
-
-@Article{heckman1985alternative,
- author = {Heckman, James J and Robb Jr, Richard},
- title = {Alternative Methods for Evaluating the Impact of Interventions: {{An}} Overview},
- number = {1-2},
- pages = {239--267},
- volume = {30},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of econometrics},
- year = {1985},
-}
-
-@Article{hershfield2011increasing,
- author = {Hershfield, Hal E and Goldstein, Daniel G and Sharpe, William F and Fox, Jesse and Yeykelis, Leo and Carstensen, Laura L and Bailenson, Jeremy N},
- title = {Increasing Saving Behavior through Age-Progressed Renderings of the Future Self},
- issue = {SPL},
- pages = {S23--S37},
- volume = {48},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Marketing Research},
- year = {2011},
-}
-
-@InProceedings{ho1995random,
- author = {Ho, Tin Kam},
- booktitle = {Proceedings of 3rd International Conference on Document Analysis and Recognition},
- title = {Random Decision Forests},
- pages = {278--282},
- publisher = {{IEEE}},
- volume = {1},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {1995},
-}
-
-@Article{hochreiter1997long,
- author = {Hochreiter, Sepp and Schmidhuber, J{\"u}rgen},
- title = {Long Short-Term Memory},
- number = {8},
- pages = {1735--1780},
- volume = {9},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Neural computation},
- year = {1997},
-}
-
-@Unpublished{hoff2021bayesoptimal,
- author = {Hoff, Peter},
- title = {Bayes-Optimal Prediction with Frequentist Coverage Control},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2105.14045},
- eprinttype = {arxiv},
- file = {:/Users/FA31DU/Zotero/storage/IQK27WVA/Hoff - 2021 - Bayes-optimal prediction with frequentist coverage.pdf:;:/Users/FA31DU/Zotero/storage/K8EAZA25/2105.html:},
- year = {2021},
-}
-
-@Misc{hoffman1994german,
- author = {Hoffman, Hans},
- title = {German {{Credit Data}}},
- url = {https://archive.ics.uci.edu/ml/datasets/statlog+(german+credit+data)},
- bdsk-url-1 = {https://archive.ics.uci.edu/ml/datasets/statlog+(german+credit+data)},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {1994},
-}
-
-@Online{hoffmanGermanCreditData1994,
- author = {Hoffman, Hans},
- title = {German {{Credit Data}}},
- url = {https://archive.ics.uci.edu/ml/datasets/statlog+(german+credit+data)},
- bdsk-url-1 = {https://archive.ics.uci.edu/ml/datasets/statlog+(german+credit+data)},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {1994},
-}
-
-@Unpublished{houlsby2011bayesian,
- author = {Houlsby, Neil and Husz{\'a}r, Ferenc and Ghahramani, Zoubin and Lengyel, M{\'a}t{\'e}},
- title = {Bayesian Active Learning for Classification and Preference Learning},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1112.5745},
- eprinttype = {arxiv},
- year = {2011},
-}
-
-@Article{hsee1996evaluability,
- author = {Hsee, Christopher K},
- title = {The Evaluability Hypothesis: {{An}} Explanation for Preference Reversals between Joint and Separate Evaluations of Alternatives},
- number = {3},
- pages = {247--257},
- volume = {67},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Organizational behavior and human decision processes},
- year = {1996},
-}
-
-@Article{hsee2004music,
- author = {Hsee, Christopher K and Rottenstreich, Yuval},
- title = {Music, Pandas, and Muggers: On the Affective Psychology of Value.},
- number = {1},
- pages = {23},
- volume = {133},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Experimental Psychology: General},
- year = {2004},
-}
-
-@Article{hsieh2016social,
- author = {Hsieh, Chih-Sheng and Lee, Lung Fei},
- title = {A Social Interactions Model with Endogenous Friendship Formation and Selectivity},
- number = {2},
- pages = {301--319},
- volume = {31},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Applied Econometrics},
- year = {2016},
-}
-
-@Unpublished{immer2020improving,
- author = {Immer, Alexander and Korzepa, Maciej and Bauer, Matthias},
- title = {Improving Predictions of Bayesian Neural Networks via Local Linearization},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2008.08400},
- eprinttype = {arxiv},
- year = {2020},
-}
-
-@Unpublished{innes2018fashionable,
- author = {Innes, Michael and Saba, Elliot and Fischer, Keno and Gandhi, Dhairya and Rudilosso, Marco Concetto and Joy, Neethu Mariya and Karmali, Tejan and Pal, Avik and Shah, Viral},
- title = {Fashionable Modelling with Flux},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1811.01457},
- eprinttype = {arxiv},
- year = {2018},
-}
-
-@Article{innes2018flux,
- author = {Innes, Mike},
- title = {Flux: {{Elegant}} Machine Learning with {{Julia}}},
- number = {25},
- pages = {602},
- volume = {3},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Open Source Software},
- year = {2018},
-}
-
-@Unpublished{ish-horowicz2019interpreting,
- author = {Ish-Horowicz, Jonathan and Udwin, Dana and Flaxman, Seth and Filippi, Sarah and Crawford, Lorin},
- title = {Interpreting Deep Neural Networks through Variable Importance},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1901.09839},
- eprinttype = {arxiv},
- year = {2019},
-}
-
-@InProceedings{jabbari2017fairness,
- author = {Jabbari, Shahin and Joseph, Matthew and Kearns, Michael and Morgenstern, Jamie and Roth, Aaron},
- booktitle = {International {{Conference}} on {{Machine Learning}}},
- title = {Fairness in Reinforcement Learning},
- pages = {1617--1626},
- publisher = {{PMLR}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2017},
-}
-
-@Article{jackson2007meeting,
- author = {Jackson, Matthew O and Rogers, Brian W},
- title = {Meeting Strangers and Friends of Friends: {{How}} Random Are Social Networks?},
- number = {3},
- pages = {890--915},
- volume = {97},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {American Economic Review},
- year = {2007},
-}
-
-@Unpublished{jeanneret2022diffusion,
- author = {Jeanneret, Guillaume and Simon, Lo{\"\i}c and Jurie, Fr{\'e}d{\'e}ric},
- title = {Diffusion {{Models}} for {{Counterfactual Explanations}}},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2203.15636},
- eprinttype = {arxiv},
- year = {2022},
-}
-
-@Article{johansson2005failure,
- author = {Johansson, Petter and Hall, Lars and Sikstr{\"o}m, Sverker and Olsson, Andreas},
- title = {Failure to Detect Mismatches between Intention and Outcome in a Simple Decision Task},
- number = {5745},
- pages = {116--119},
- volume = {310},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Science (New York, N.Y.)},
- shortjournal = {Science},
- year = {2005},
-}
-
-@Article{johnsson2021estimation,
- author = {Johnsson, Ida and Moon, Hyungsik Roger},
- title = {Estimation of Peer Effects in Endogenous Social Networks: {{Control}} Function Approach},
- number = {2},
- pages = {328--345},
- volume = {103},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Review of Economics and Statistics},
- year = {2021},
-}
-
-@Article{jolliffe2003modified,
- author = {Jolliffe, Ian T and Trendafilov, Nickolay T and Uddin, Mudassir},
- title = {A Modified Principal Component Technique Based on the {{LASSO}}},
- number = {3},
- pages = {531--547},
- volume = {12},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of computational and Graphical Statistics},
- year = {2003},
-}
-
-@Article{joseph2021forecasting,
- author = {Joseph, Andreas and Kalamara, Eleni and Kapetanios, George and Potjagailo, Galina},
- title = {Forecasting Uk Inflation Bottom Up},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2021},
-}
-
-@Unpublished{joshi2019realistic,
- author = {Joshi, Shalmali and Koyejo, Oluwasanmi and Vijitbenjaronk, Warut and Kim, Been and Ghosh, Joydeep},
- title = {Towards Realistic Individual Recourse and Actionable Explanations in Black-Box Decision Making Systems},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1907.09615},
- eprinttype = {arxiv},
- year = {2019},
-}
-
-@Unpublished{jospin2020handson,
- author = {Jospin, Laurent Valentin and Buntine, Wray and Boussaid, Farid and Laga, Hamid and Bennamoun, Mohammed},
- title = {Hands-on {{Bayesian Neural Networks}}--a {{Tutorial}} for {{Deep Learning Users}}},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2007.06823},
- eprinttype = {arxiv},
- year = {2020},
-}
-
-@Misc{kaggle2011give,
- author = {Kaggle},
- title = {Give Me Some Credit, {{Improve}} on the State of the Art in Credit Scoring by Predicting the Probability That Somebody Will Experience Financial Distress in the next Two Years.},
- url = {https://www.kaggle.com/c/GiveMeSomeCredit},
- bdsk-url-1 = {https://www.kaggle.com/c/GiveMeSomeCredit},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- publisher = {{Kaggle}},
- year = {2011},
-}
-
-@online{kagglecompetitionGiveMeCredit,
- author = {Kaggle Competition},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- title = {Give Me Some Credit, {{Improve}} on the State of the Art in Credit Scoring by Predicting the Probability That Somebody Will Experience Financial Distress in the next Two Years.},
- url = {https://www.kaggle.com/c/GiveMeSomeCredit},
- bdsk-url-1 = {https://www.kaggle.com/c/GiveMeSomeCredit}}
-
-@Article{kahneman1979prospect,
- author = {Kahneman, Daniel and Tversky, Amos},
- title = {Prospect {{Theory}}: {{An Analysis}} of {{Decision}} under {{Risk}}},
- pages = {263--291},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Econometrica: Journal of the Econometric Society},
- year = {1979},
-}
-
-@Article{kahneman1990experimental,
- author = {Kahneman, Daniel and Knetsch, Jack L and Thaler, Richard H},
- title = {Experimental Tests of the Endowment Effect and the {{Coase}} Theorem},
- number = {6},
- pages = {1325--1348},
- volume = {98},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of political Economy},
- year = {1990},
-}
-
-@Article{kahneman1992reference,
- author = {Kahneman, Daniel},
- title = {Reference Points, Anchors, Norms, and Mixed Feelings},
- number = {2},
- pages = {296--312},
- volume = {51},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Organizational behavior and human decision processes},
- year = {1992},
-}
-
-@Unpublished{karimi2020algorithmic,
- author = {Karimi, Amir-Hossein and Von K{\"u}gelgen, Julius and Sch{\"o}lkopf, Bernhard and Valera, Isabel},
- title = {Algorithmic Recourse under Imperfect Causal Knowledge: A Probabilistic Approach},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2006.06831},
- eprinttype = {arxiv},
- year = {2020},
-}
-
-@Unpublished{karimi2020survey,
- author = {Karimi, Amir-Hossein and Barthe, Gilles and Sch{\"o}lkopf, Bernhard and Valera, Isabel},
- title = {A Survey of Algorithmic Recourse: Definitions, Formulations, Solutions, and Prospects},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2010.04050},
- eprinttype = {arxiv},
- year = {2020},
-}
-
-@InProceedings{karimi2021algorithmic,
- author = {Karimi, Amir-Hossein and Sch{\"o}lkopf, Bernhard and Valera, Isabel},
- booktitle = {Proceedings of the 2021 {{ACM Conference}} on {{Fairness}}, {{Accountability}}, and {{Transparency}}},
- title = {Algorithmic Recourse: From Counterfactual Explanations to Interventions},
- pages = {353--362},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2021},
-}
-
-@InProceedings{kaur2020interpreting,
- author = {Kaur, Harmanpreet and Nori, Harsha and Jenkins, Samuel and Caruana, Rich and Wallach, Hanna and Wortman Vaughan, Jennifer},
- booktitle = {Proceedings of the 2020 {{CHI}} Conference on Human Factors in Computing Systems},
- title = {Interpreting Interpretability: Understanding Data Scientists' Use of Interpretability Tools for Machine Learning},
- pages = {1--14},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2020},
-}
-
-@Article{kehoe2021defence,
- author = {Kehoe, Aidan and Wittek, Peter and Xue, Yanbo and Pozas-Kerstjens, Alejandro},
- title = {Defence against Adversarial Attacks Using Classical and Quantum-Enhanced {{Boltzmann}} Machines},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Machine Learning: Science and Technology},
- year = {2021},
-}
-
-@Unpublished{kendall2017what,
- author = {Kendall, Alex and Gal, Yarin},
- title = {What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1703.04977},
- eprinttype = {arxiv},
- year = {2017},
-}
-
-@Article{kihoro2004seasonal,
- author = {Kihoro, J and Otieno, RO and Wafula, C},
- title = {Seasonal Time Series Forecasting: {{A}} Comparative Study of {{ARIMA}} and {{ANN}} Models},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2004},
-}
-
-@Book{kilian2017structural,
- author = {Kilian, Lutz and L{\"u}tkepohl, Helmut},
- title = {Structural Vector Autoregressive Analysis},
- publisher = {{Cambridge University Press}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2017},
-}
-
-@Unpublished{kingma2014adam,
- author = {Kingma, Diederik P and Ba, Jimmy},
- title = {Adam: {{A}} Method for Stochastic Optimization},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1412.6980},
- eprinttype = {arxiv},
- year = {2014},
-}
-
-@Article{kirsch2019batchbald,
- author = {Kirsch, Andreas and Van Amersfoort, Joost and Gal, Yarin},
- title = {Batchbald: {{Efficient}} and Diverse Batch Acquisition for Deep Bayesian Active Learning},
- pages = {7026--7037},
- volume = {32},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Advances in neural information processing systems},
- year = {2019},
-}
-
-@Unpublished{kuiper2021exploring,
- author = {Kuiper, Ouren and van den Berg, Martin and van den Burgt, Joost and Leijnen, Stefan},
- title = {Exploring {{Explainable AI}} in the {{Financial Sector}}: {{Perspectives}} of {{Banks}} and {{Supervisory Authorities}}},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2111.02244},
- eprinttype = {arxiv},
- year = {2021},
-}
-
-@Article{kydland1982time,
- author = {Kydland, Finn E and Prescott, Edward C},
- title = {Time to Build and Aggregate Fluctuations},
- pages = {1345--1370},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Econometrica: Journal of the Econometric Society},
- year = {1982},
-}
-
-@Unpublished{lachapelle2019gradientbased,
- author = {Lachapelle, S{\'e}bastien and Brouillard, Philippe and Deleu, Tristan and Lacoste-Julien, Simon},
- title = {Gradient-Based Neural Dag Learning},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1906.02226},
- eprinttype = {arxiv},
- year = {2019},
-}
-
-@InProceedings{lakkaraju2020how,
- author = {Lakkaraju, Himabindu and Bastani, Osbert},
- booktitle = {Proceedings of the {{AAAI}}/{{ACM Conference}} on {{AI}}, {{Ethics}}, and {{Society}}},
- title = {" {{How}} Do {{I}} Fool You?" {{Manipulating User Trust}} via {{Misleading Black Box Explanations}}},
- pages = {79--85},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2020},
-}
-
-@InProceedings{lakkaraju2020how,
- author = {Lakkaraju, Himabindu and Bastani, Osbert},
- booktitle = {Proceedings of the {{AAAI}}/{{ACM Conference}} on {{AI}}, {{Ethics}}, and {{Society}}},
- title = {" {{How Do I Fool You}}?" {{Manipulating User Trust}} via {{Misleading Black Box Explanations}}},
- pages = {79--85},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2020},
-}
-
-@Unpublished{lakshminarayanan2016simple,
- author = {Lakshminarayanan, Balaji and Pritzel, Alexander and Blundell, Charles},
- title = {Simple and Scalable Predictive Uncertainty Estimation Using Deep Ensembles},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1612.01474},
- eprinttype = {arxiv},
- year = {2016},
-}
-
-@Unpublished{laugel2017inverse,
- author = {Laugel, Thibault and Lesot, Marie-Jeanne and Marsala, Christophe and Renard, Xavier and Detyniecki, Marcin},
- title = {Inverse Classification for Comparison-Based Interpretability in Machine Learning},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1712.08443},
- eprinttype = {arxiv},
- shortjournal = {arXiv preprint arXiv:1712.08443},
- year = {2017},
-}
-
-@Thesis{lawrence2001variational,
- author = {Lawrence, Neil David},
- title = {Variational Inference in Probabilistic Models},
- type = {phdthesis},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- school = {{University of Cambridge}},
- year = {2001},
-}
-
-@Article{lecun1998mnist,
- author = {LeCun, Yann},
- title = {The {{MNIST}} Database of Handwritten Digits},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- shortjournal = {http://yann. lecun. com/exdb/mnist/},
- year = {1998},
-}
-
-@Article{lee2003best,
- author = {Lee, Lung-fei},
- title = {Best Spatial Two-Stage Least Squares Estimators for a Spatial Autoregressive Model with Autoregressive Disturbances},
- number = {4},
- pages = {307--335},
- volume = {22},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Econometric Reviews},
- year = {2003},
-}
-
-@Article{lerner2013financial,
- author = {Lerner, Jennifer S and Li, Ye and Weber, Elke U},
- title = {The Financial Costs of Sadness},
- number = {1},
- pages = {72--79},
- volume = {24},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Psychological science},
- year = {2013},
-}
-
-@Article{list2004neoclassical,
- author = {List, John A},
- title = {Neoclassical Theory versus Prospect Theory: {{Evidence}} from the Marketplace},
- number = {2},
- pages = {615--625},
- volume = {72},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Econometrica : journal of the Econometric Society},
- shortjournal = {Econometrica},
- year = {2004},
-}
-
-@Article{lucas1976econometric,
- author = {Lucas, JR},
- title = {Econometric Policy Evaluation: A Critique `, in {{K}}. {{Brunner}} and {{A Meltzer}}, {{The Phillips}} Curve and Labor Markets, {{North Holland}}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {1976},
-}
-
-@InProceedings{lundberg2017unified,
- author = {Lundberg, Scott M and Lee, Su-In},
- booktitle = {Proceedings of the 31st International Conference on Neural Information Processing Systems},
- title = {A Unified Approach to Interpreting Model Predictions},
- pages = {4768--4777},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2017},
-}
-
-@Book{lutkepohl2005new,
- author = {L{\"u}tkepohl, Helmut},
- title = {New Introduction to Multiple Time Series Analysis},
- publisher = {{Springer Science \& Business Media}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2005},
-}
-
-@Article{madrian2001power,
- author = {Madrian, Brigitte C and Shea, Dennis F},
- title = {The Power of Suggestion: {{Inertia}} in 401 (k) Participation and Savings Behavior},
- number = {4},
- pages = {1149--1187},
- volume = {116},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The Quarterly journal of economics},
- year = {2001},
-}
-
-@Book{manning2008introduction,
- author = {Manning, Christopher D and Sch{\"u}tze, Hinrich and Raghavan, Prabhakar},
- title = {Introduction to Information Retrieval},
- publisher = {{Cambridge university press}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2008},
-}
-
-@misc{manokhin2022awesome,
- author = {Manokhin, Valery},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- title = {Awesome Conformal Prediction}}
-
-@Article{manski1993identification,
- author = {Manski, Charles F},
- title = {Identification of Endogenous Social Effects: {{The}} Reflection Problem},
- number = {3},
- pages = {531--542},
- volume = {60},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The review of economic studies},
- year = {1993},
-}
-
-@Article{markle2018goals,
- author = {Markle, Alex and Wu, George and White, Rebecca and Sackett, Aaron},
- title = {Goals as Reference Points in Marathon Running: {{A}} Novel Test of Reference Dependence},
- number = {1},
- pages = {19--50},
- volume = {56},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Risk and Uncertainty},
- year = {2018},
-}
-
-@Article{masini2021machine,
- author = {Masini, Ricardo P and Medeiros, Marcelo C and Mendes, Eduardo F},
- title = {Machine Learning Advances for Time Series Forecasting},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Economic Surveys},
- year = {2021},
-}
-
-@Article{mccracken2016fredmd,
- author = {McCracken, Michael W and Ng, Serena},
- title = {{{FRED-MD}}: {{A}} Monthly Database for Macroeconomic Research},
- number = {4},
- pages = {574--589},
- volume = {34},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Business \& Economic Statistics},
- year = {2016},
-}
-
-@Article{mcculloch1990logical,
- author = {McCulloch, Warren S and Pitts, Walter},
- title = {A Logical Calculus of the Ideas Immanent in Nervous Activity},
- number = {1},
- pages = {99--115},
- volume = {52},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Bulletin of mathematical biology},
- year = {1990},
-}
-
-@Article{migut2015visualizing,
- author = {Migut, MA and Worring, Marcel and Veenman, Cor J},
- title = {Visualizing Multi-Dimensional Decision Boundaries in {{2D}}},
- number = {1},
- pages = {273--295},
- volume = {29},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Data Mining and Knowledge Discovery},
- year = {2015},
-}
-
-@Article{miller2019explanation,
- author = {Miller, Tim},
- title = {Explanation in Artificial Intelligence: {{Insights}} from the Social Sciences},
- pages = {1--38},
- volume = {267},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Artificial intelligence},
- year = {2019},
-}
-
-@InProceedings{miller2020strategic,
- author = {Miller, John and Milli, Smitha and Hardt, Moritz},
- booktitle = {Proceedings of the 37th {{International Conference}} on {{Machine Learning}}},
- title = {Strategic {{Classification}} Is {{Causal Modeling}} in {{Disguise}}},
- eventtitle = {International {{Conference}} on {{Machine Learning}}},
- pages = {6917--6926},
- publisher = {{PMLR}},
- url = {https://proceedings.mlr.press/v119/miller20b.html},
- urldate = {2022-11-03},
- abstract = {Consequential decision-making incentivizes individuals to strategically adapt their behavior to the specifics of the decision rule. While a long line of work has viewed strategic adaptation as gaming and attempted to mitigate its effects, recent work has instead sought to design classifiers that incentivize individuals to improve a desired quality. Key to both accounts is a cost function that dictates which adaptations are rational to undertake. In this work, we develop a causal framework for strategic adaptation. Our causal perspective clearly distinguishes between gaming and improvement and reveals an important obstacle to incentive design. We prove any procedure for designing classifiers that incentivize improvement must inevitably solve a non-trivial causal inference problem. We show a similar result holds for designing cost functions that satisfy the requirements of previous work. With the benefit of hindsight, our results show much of the prior work on strategic classification is causal modeling in disguise.},
- bdsk-url-1 = {https://proceedings.mlr.press/v119/miller20b.html},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- file = {:/Users/FA31DU/Zotero/storage/46I2QMPI/Miller et al. - 2020 - Strategic Classification is Causal Modeling in Dis.pdf:;:/Users/FA31DU/Zotero/storage/NWREET6B/Miller et al. - 2020 - Strategic Classification is Causal Modeling in Dis.pdf:},
- issn = {2640-3498},
- langid = {english},
- month = nov,
- year = {2020},
-}
-
-@Article{mischel1988nature,
- author = {Mischel, Walter and Shoda, Yuichi and Peake, Philip K},
- title = {The Nature of Adolescent Competencies Predicted by Preschool Delay of Gratification.},
- number = {4},
- pages = {687},
- volume = {54},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of personality and social psychology},
- year = {1988},
-}
-
-@InProceedings{mittelstadt2019explaining,
- author = {Mittelstadt, Brent and Russell, Chris and Wachter, Sandra},
- booktitle = {Proceedings of the Conference on Fairness, Accountability, and Transparency},
- title = {Explaining Explanations in {{AI}}},
- pages = {279--288},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2019},
-}
-
-@Book{molnar2020interpretable,
- author = {Molnar, Christoph},
- title = {Interpretable Machine Learning},
- publisher = {{Lulu. com}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2020},
-}
-
-@Book{morgan2015counterfactuals,
- author = {Morgan, Stephen L and Winship, Christopher},
- title = {Counterfactuals and Causal Inference},
- publisher = {{Cambridge University Press}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2015},
-}
-
-@Article{mosteller1951experimental,
- author = {Mosteller, Frederick and Nogee, Philip},
- title = {An Experimental Measurement of Utility},
- number = {5},
- pages = {371--404},
- volume = {59},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Political Economy},
- year = {1951},
-}
-
-@InProceedings{mothilal2020explaining,
- author = {Mothilal, Ramaravind K and Sharma, Amit and Tan, Chenhao},
- booktitle = {Proceedings of the 2020 {{Conference}} on {{Fairness}}, {{Accountability}}, and {{Transparency}}},
- title = {Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations},
- pages = {607--617},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2020},
-}
-
-@Book{murphy2012machine,
- author = {Murphy, Kevin P},
- title = {Machine Learning: A Probabilistic Perspective},
- publisher = {{MIT press}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2012},
-}
-
-@Book{murphy2012machine,
- author = {Murphy, Kevin P},
- title = {Machine Learning: {{A}} Probabilistic Perspective},
- publisher = {{MIT press}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2012},
-}
-
-@Book{murphy2022probabilistic,
- author = {Murphy, Kevin P},
- title = {Probabilistic {{Machine Learning}}: {{An}} Introduction},
- publisher = {{MIT Press}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2022},
-}
-
-@Article{nagel1995unraveling,
- author = {Nagel, Rosemarie},
- title = {Unraveling in Guessing Games: {{An}} Experimental Study},
- number = {5},
- pages = {1313--1326},
- volume = {85},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The American Economic Review},
- year = {1995},
-}
-
-@Unpublished{navarro-martinez2021bridging,
- author = {Navarro-Martinez, Daniel and Wang, Xinghua},
- title = {Bridging the Gap between the Lab and the Field: {{Dictator}} Games and Donations},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2021},
-}
-
-@InProceedings{nelson2015evaluating,
- author = {Nelson, Kevin and Corbin, George and Anania, Mark and Kovacs, Matthew and Tobias, Jeremy and Blowers, Misty},
- booktitle = {2015 {{IEEE Symposium}} on {{Computational Intelligence}} for {{Security}} and {{Defense Applications}} ({{CISDA}})},
- title = {Evaluating Model Drift in Machine Learning Algorithms},
- pages = {1--8},
- publisher = {{IEEE}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2015},
-}
-
-@Book{nocedal2006numerical,
- author = {Nocedal, Jorge and Wright, Stephen},
- title = {Numerical Optimization},
- publisher = {{Springer Science \& Business Media}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2006},
-}
-
-@Misc{oecd2021artificial,
- author = {{OECD}},
- title = {Artificial {{Intelligence}}, {{Machine Learning}} and {{Big Data}} in {{Finance}}: {{Opportunities}}, {{Challenges}} and {{Implications}} for {{Policy Makers}}},
- url = {https://www.oecd.org/finance/financial-markets/Artificial-intelligence-machine-learning-big-data-in-finance.pdf},
- bdsk-url-1 = {https://www.oecd.org/finance/financial-markets/Artificial-intelligence-machine-learning-big-data-in-finance.pdf},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2021},
-}
-
-@Online{oecdArtificialIntelligenceMachine2021,
- author = {{OECD}},
- title = {Artificial {{Intelligence}}, {{Machine Learning}} and {{Big Data}} in {{Finance}}: {{Opportunities}}, {{Challenges}} and {{Implications}} for {{Policy Makers}}},
- url = {https://www.oecd.org/finance/financial-markets/Artificial-intelligence-machine-learning-big-data-in-finance.pdf},
- bdsk-url-1 = {https://www.oecd.org/finance/financial-markets/Artificial-intelligence-machine-learning-big-data-in-finance.pdf},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- publisher = {{OECD}},
- year = {2021},
-}
-
-@Book{oneil2016weapons,
- author = {O'Neil, Cathy},
- title = {Weapons of Math Destruction: {{How}} Big Data Increases Inequality and Threatens Democracy},
- publisher = {{Crown}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2016},
-}
-
-@Article{pace1997sparse,
- author = {Pace, R Kelley and Barry, Ronald},
- title = {Sparse Spatial Autoregressions},
- number = {3},
- pages = {291--297},
- volume = {33},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Statistics \& Probability Letters},
- year = {1997},
-}
-
-@Unpublished{parr2018matrix,
- author = {Parr, Terence and Howard, Jeremy},
- title = {The Matrix Calculus You Need for Deep Learning},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1802.01528},
- eprinttype = {arxiv},
- year = {2018},
-}
-
-@Unpublished{pawelczyk2021carla,
- author = {Pawelczyk, Martin and Bielawski, Sascha and van den Heuvel, Johannes and Richter, Tobias and Kasneci, Gjergji},
- title = {Carla: A Python Library to Benchmark Algorithmic Recourse and Counterfactual Explanation Algorithms},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2108.00783},
- eprinttype = {arxiv},
- year = {2021},
-}
-
-@Book{pearl2018book,
- author = {Pearl, Judea and Mackenzie, Dana},
- title = {The Book of Why: The New Science of Cause and Effect},
- publisher = {{Basic books}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2018},
-}
-
-@Article{pearl2019seven,
- author = {Pearl, Judea},
- title = {The Seven Tools of Causal Inference, with Reflections on Machine Learning},
- number = {3},
- pages = {54--60},
- volume = {62},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Communications of the ACM},
- year = {2019},
-}
-
-@Article{pedregosa2011scikitlearn,
- author = {Pedregosa, Fabian and Varoquaux, Ga{\"e}l and Gramfort, Alexandre and Michel, Vincent and Thirion, Bertrand and Grisel, Olivier and Blondel, Mathieu and Prettenhofer, Peter and Weiss, Ron and Dubourg, Vincent and others},
- title = {Scikit-Learn: {{Machine}} Learning in {{Python}}},
- pages = {2825--2830},
- volume = {12},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {the Journal of machine Learning research},
- year = {2011},
-}
-
-@Book{perry2010economic,
- author = {Perry, George L and Tobin, James},
- title = {Economic {{Events}}, {{Ideas}}, and {{Policies}}: The 1960s and After},
- publisher = {{Brookings Institution Press}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2010},
-}
-
-@Article{pfaff2008var,
- author = {Pfaff, Bernhard and others},
- title = {{{VAR}}, {{SVAR}} and {{SVEC}} Models: {{Implementation}} within {{R}} Package Vars},
- number = {4},
- pages = {1--32},
- volume = {27},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Statistical Software},
- year = {2008},
-}
-
-@Book{pindyck2014microeconomics,
- author = {Pindyck, Robert S and Rubinfeld, Daniel L},
- title = {Microeconomics},
- publisher = {{Pearson Education}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2014},
-}
-
-@Article{pope2011numbers,
- author = {Pope, Devin and Simonsohn, Uri},
- title = {Round Numbers as Goals: {{Evidence}} from Baseball, {{SAT}} Takers, and the Lab},
- number = {1},
- pages = {71--79},
- volume = {22},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Psychological science},
- year = {2011},
-}
-
-@InProceedings{poyiadzi2020face,
- author = {Poyiadzi, Rafael and Sokol, Kacper and Santos-Rodriguez, Raul and De Bie, Tijl and Flach, Peter},
- booktitle = {Proceedings of the {{AAAI}}/{{ACM Conference}} on {{AI}}, {{Ethics}}, and {{Society}}},
- title = {{{FACE}}: {{Feasible}} and Actionable Counterfactual Explanations},
- pages = {344--350},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2020},
-}
-
-@Article{qu2015estimating,
- author = {Qu, Xi and Lee, Lung-fei},
- title = {Estimating a Spatial Autoregressive Model with an Endogenous Spatial Weight Matrix},
- number = {2},
- pages = {209--232},
- volume = {184},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of Econometrics},
- year = {2015},
-}
-
-@Article{rabanser2019failing,
- author = {Rabanser, Stephan and G{\"u}nnemann, Stephan and Lipton, Zachary},
- title = {Failing Loudly: {{An}} Empirical Study of Methods for Detecting Dataset Shift},
- volume = {32},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Advances in Neural Information Processing Systems},
- year = {2019},
-}
-
-@Unpublished{raghunathan2019adversarial,
- author = {Raghunathan, Aditi and Xie, Sang Michael and Yang, Fanny and Duchi, John C and Liang, Percy},
- title = {Adversarial Training Can Hurt Generalization},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1906.06032},
- eprinttype = {arxiv},
- year = {2019},
-}
-
-@Unpublished{raj2017taming,
- author = {Raj, Vishnu and Kalyani, Sheetal},
- title = {Taming Non-Stationary Bandits: {{A Bayesian}} Approach},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1707.09727},
- eprinttype = {arxiv},
- year = {2017},
-}
-
-@InProceedings{rasmussen2003gaussian,
- author = {Rasmussen, Carl Edward},
- booktitle = {Summer School on Machine Learning},
- title = {Gaussian Processes in Machine Learning},
- pages = {63--71},
- publisher = {{Springer}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2003},
-}
-
-@InProceedings{ribeiro2016why,
- author = {Ribeiro, Marco Tulio and Singh, Sameer and Guestrin, Carlos},
- booktitle = {Proceedings of the 22nd {{ACM SIGKDD}} International Conference on Knowledge Discovery and Data Mining},
- title = {"{{Why}} Should i Trust You?" {{Explaining}} the Predictions of Any Classifier},
- pages = {1135--1144},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2016},
-}
-
-@Article{romer1989does,
- author = {Romer, Christina D and Romer, David H},
- title = {Does Monetary Policy Matter? {{A}} New Test in the Spirit of {{Friedman}} and {{Schwartz}}},
- pages = {121--170},
- volume = {4},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {NBER macroeconomics annual},
- year = {1989},
-}
-
-@Article{rudin2019stop,
- author = {Rudin, Cynthia},
- title = {Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead},
- number = {5},
- pages = {206--215},
- volume = {1},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Nature Machine Intelligence},
- year = {2019},
-}
-
-@Article{sacerdote2001peer,
- author = {Sacerdote, Bruce},
- title = {Peer Effects with Random Assignment: {{Results}} for {{Dartmouth}} Roommates},
- number = {2},
- pages = {681--704},
- volume = {116},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The Quarterly journal of economics},
- year = {2001},
-}
-
-@Article{sadinle2019least,
- author = {Sadinle, Mauricio and Lei, Jing and Wasserman, Larry},
- title = {Least Ambiguous Set-Valued Classifiers with Bounded Error Levels},
- number = {525},
- pages = {223--234},
- volume = {114},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- file = {:/Users/FA31DU/Zotero/storage/YXQ8N76A/Sadinle et al. - 2019 - Least ambiguous set-valued classifiers with bounde.pdf:;:/Users/FA31DU/Zotero/storage/ZHB56F3V/01621459.2017.html:},
- journal = {Journal of the American Statistical Association},
- publisher = {{Taylor \& Francis}},
- year = {2019},
-}
-
-@InProceedings{satopaa2011finding,
- author = {Satopaa, Ville and Albrecht, Jeannie and Irwin, David and Raghavan, Barath},
- booktitle = {2011 31st International Conference on Distributed Computing Systems Workshops},
- title = {Finding a" Kneedle" in a Haystack: {{Detecting}} Knee Points in System Behavior},
- pages = {166--171},
- publisher = {{IEEE}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2011},
-}
-
-@InProceedings{schut2021generating,
- author = {Schut, Lisa and Key, Oscar and Mc Grath, Rory and Costabello, Luca and Sacaleanu, Bogdan and Gal, Yarin and others},
- booktitle = {International {{Conference}} on {{Artificial Intelligence}} and {{Statistics}}},
- title = {Generating {{Interpretable Counterfactual Explanations By Implicit Minimisation}} of {{Epistemic}} and {{Aleatoric Uncertainties}}},
- pages = {1756--1764},
- publisher = {{PMLR}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2021},
-}
-
-@Book{schutze2008introduction,
- author = {Sch{\"u}tze, Hinrich and Manning, Christopher D and Raghavan, Prabhakar},
- title = {Introduction to Information Retrieval},
- publisher = {{Cambridge University Press Cambridge}},
- volume = {39},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2008},
-}
-
-@Article{shafir1993reasonbased,
- author = {Shafir, Eldar and Simonson, Itamar and Tversky, Amos},
- title = {Reason-Based Choice},
- number = {1-2},
- pages = {11--36},
- volume = {49},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Cognition},
- year = {1993},
-}
-
-@Article{simonson1989choice,
- author = {Simonson, Itamar},
- title = {Choice Based on Reasons: {{The}} Case of Attraction and Compromise Effects},
- number = {2},
- pages = {158--174},
- volume = {16},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of consumer research},
- year = {1989},
-}
-
-@Article{sims1986are,
- author = {Sims, Christopher A and others},
- title = {Are Forecasting Models Usable for Policy Analysis?},
- issue = {Win},
- pages = {2--16},
- volume = {10},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Quarterly Review},
- year = {1986},
-}
-
-@InProceedings{slack2020fooling,
- author = {Slack, Dylan and Hilgard, Sophie and Jia, Emily and Singh, Sameer and Lakkaraju, Himabindu},
- booktitle = {Proceedings of the {{AAAI}}/{{ACM Conference}} on {{AI}}, {{Ethics}}, and {{Society}}},
- title = {Fooling Lime and Shap: {{Adversarial}} Attacks on Post Hoc Explanation Methods},
- pages = {180--186},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2020},
-}
-
-@Article{slack2021counterfactual,
- author = {Slack, Dylan and Hilgard, Anna and Lakkaraju, Himabindu and Singh, Sameer},
- title = {Counterfactual Explanations Can Be Manipulated},
- volume = {34},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Advances in Neural Information Processing Systems},
- year = {2021},
-}
-
-@Article{slovic1974who,
- author = {Slovic, Paul and Tversky, Amos},
- title = {Who Accepts {{Savage}}'s Axiom?},
- number = {6},
- pages = {368--373},
- volume = {19},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Behavioral science},
- year = {1974},
-}
-
-@Unpublished{spooner2021counterfactual,
- author = {Spooner, Thomas and Dervovic, Danial and Long, Jason and Shepard, Jon and Chen, Jiahao and Magazzeni, Daniele},
- title = {Counterfactual {{Explanations}} for {{Arbitrary Regression Models}}},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2106.15212},
- eprinttype = {arxiv},
- shortjournal = {arXiv preprint arXiv:2106.15212},
- year = {2021},
-}
-
-@Article{srivastava2014dropout,
- author = {Srivastava, Nitish and Hinton, Geoffrey and Krizhevsky, Alex and Sutskever, Ilya and Salakhutdinov, Ruslan},
- title = {Dropout: A Simple Way to Prevent Neural Networks from Overfitting},
- number = {1},
- pages = {1929--1958},
- volume = {15},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The journal of machine learning research},
- year = {2014},
-}
-
-@Unpublished{stanton2022bayesian,
- author = {Stanton, Samuel and Maddox, Wesley and Wilson, Andrew Gordon},
- title = {Bayesian {{Optimization}} with {{Conformal Coverage Guarantees}}},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2210.12496},
- eprinttype = {arxiv},
- file = {:/Users/FA31DU/Zotero/storage/XFGZAB9J/Stanton et al. - 2022 - Bayesian Optimization with Conformal Coverage Guar.pdf:;:/Users/FA31DU/Zotero/storage/RPWYDPVW/2210.html:},
- year = {2022},
-}
-
-@Article{sturm2014simple,
- author = {Sturm, Bob L},
- title = {A Simple Method to Determine If a Music Information Retrieval System Is a ``Horse''},
- number = {6},
- pages = {1636--1644},
- volume = {16},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {IEEE Transactions on Multimedia},
- year = {2014},
-}
-
-@Article{sunstein2003libertarian,
- author = {Sunstein, Cass R and Thaler, Richard H},
- title = {Libertarian Paternalism Is Not an Oxymoron},
- pages = {1159--1202},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {The University of Chicago Law Review},
- year = {2003},
-}
-
-@Book{sutton2018reinforcement,
- author = {Sutton, Richard S and Barto, Andrew G},
- title = {Reinforcement Learning: {{An}} Introduction},
- publisher = {{MIT press}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2018},
-}
-
-@Unpublished{szegedy2013intriguing,
- author = {Szegedy, Christian and Zaremba, Wojciech and Sutskever, Ilya and Bruna, Joan and Erhan, Dumitru and Goodfellow, Ian and Fergus, Rob},
- title = {Intriguing Properties of Neural Networks},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1312.6199},
- eprinttype = {arxiv},
- year = {2013},
-}
-
-@Article{thaler1981empirical,
- author = {Thaler, Richard},
- title = {Some Empirical Evidence on Dynamic Inconsistency},
- number = {3},
- pages = {201--207},
- volume = {8},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Economics letters},
- year = {1981},
-}
-
-@Article{thaler2004more,
- author = {Thaler, Richard H and Benartzi, Shlomo},
- title = {Save More Tomorrow{\texttrademark}: {{Using}} Behavioral Economics to Increase Employee Saving},
- number = {S1},
- pages = {S164--S187},
- volume = {112},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of political Economy},
- year = {2004},
-}
-
-@Article{tversky1981framing,
- author = {Tversky, Amos and Kahneman, Daniel},
- title = {The Framing of Decisions and the Psychology of Choice},
- number = {4481},
- pages = {453--458},
- volume = {211},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Science (New York, N.Y.)},
- shortjournal = {science},
- year = {1981},
-}
-
-@Article{ungemach2011how,
- author = {Ungemach, Christoph and Stewart, Neil and Reimers, Stian},
- title = {How Incidental Values from the Environment Affect Decisions about Money, Risk, and Delay},
- number = {2},
- pages = {253--260},
- volume = {22},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Psychological Science},
- year = {2011},
-}
-
-@Unpublished{upadhyay2021robust,
- author = {Upadhyay, Sohini and Joshi, Shalmali and Lakkaraju, Himabindu},
- title = {Towards {{Robust}} and {{Reliable Algorithmic Recourse}}},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2102.13620},
- eprinttype = {arxiv},
- year = {2021},
-}
-
-@InProceedings{ustun2019actionable,
- author = {Ustun, Berk and Spangher, Alexander and Liu, Yang},
- booktitle = {Proceedings of the {{Conference}} on {{Fairness}}, {{Accountability}}, and {{Transparency}}},
- title = {Actionable Recourse in Linear Classification},
- pages = {10--19},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2019},
-}
-
-@Article{vanboven2000egocentric,
- author = {Van Boven, Leaf and Dunning, David and Loewenstein, George},
- title = {Egocentric Empathy Gaps between Owners and Buyers: Misperceptions of the Endowment Effect.},
- number = {1},
- pages = {66},
- volume = {79},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of personality and social psychology},
- year = {2000},
-}
-
-@Book{varshney2022trustworthy,
- author = {Varshney, Kush R.},
- title = {Trustworthy {{Machine Learning}}},
- publisher = {{Independently Published}},
- address = {{Chappaqua, NY, USA}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2022},
-}
-
-@Unpublished{verma2020counterfactual,
- author = {Verma, Sahil and Dickerson, John and Hines, Keegan},
- title = {Counterfactual Explanations for Machine Learning: {{A}} Review},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2010.10596},
- eprinttype = {arxiv},
- year = {2020},
-}
-
-@Article{verstyuk2020modeling,
- author = {Verstyuk, Sergiy},
- title = {Modeling Multivariate Time Series in Economics: {{From}} Auto-Regressions to Recurrent Neural Networks},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Available at SSRN 3589337},
- year = {2020},
-}
-
-@Article{wachter2017counterfactual,
- author = {Wachter, Sandra and Mittelstadt, Brent and Russell, Chris},
- title = {Counterfactual Explanations without Opening the Black Box: {{Automated}} Decisions and the {{GDPR}}},
- pages = {841},
- volume = {31},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Harv. JL \& Tech.},
- year = {2017},
-}
-
-@Article{wang2018optimal,
- author = {Wang, HaiYing and Zhu, Rong and Ma, Ping},
- title = {Optimal Subsampling for Large Sample Logistic Regression},
- number = {522},
- pages = {829--844},
- volume = {113},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Journal of the American Statistical Association},
- year = {2018},
-}
-
-@Book{wasserman2006all,
- author = {Wasserman, Larry},
- title = {All of Nonparametric Statistics},
- publisher = {{Springer Science \& Business Media}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2006},
-}
-
-@Book{wasserman2013all,
- author = {Wasserman, Larry},
- title = {All of Statistics: A Concise Course in Statistical Inference},
- publisher = {{Springer Science \& Business Media}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2013},
-}
-
-@Article{widmer1996learning,
- author = {Widmer, Gerhard and Kubat, Miroslav},
- title = {Learning in the Presence of Concept Drift and Hidden Contexts},
- number = {1},
- pages = {69--101},
- volume = {23},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Machine learning},
- year = {1996},
-}
-
-@Unpublished{wilson2020case,
- author = {Wilson, Andrew Gordon},
- title = {The Case for {{Bayesian}} Deep Learning},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2001.10995},
- eprinttype = {arxiv},
- year = {2020},
-}
-
-@Article{witten2009penalized,
- author = {Witten, Daniela M and Tibshirani, Robert and Hastie, Trevor},
- title = {A Penalized Matrix Decomposition, with Applications to Sparse Principal Components and Canonical Correlation Analysis},
- number = {3},
- pages = {515--534},
- volume = {10},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Biostatistics (Oxford, England)},
- shortjournal = {Biostatistics},
- year = {2009},
-}
-
-@Article{xu2020epidemiological,
- author = {Xu, Bo and Gutierrez, Bernardo and Mekaru, Sumiko and Sewalk, Kara and Goodwin, Lauren and Loskill, Alyssa and Cohn, Emily and Hswen, Yulin and Hill, Sarah C. and Cobo, Maria M and Zarebski, Alexander and Li, Sabrina and Wu, Chieh-Hsi and Hulland, Erin and Morgan, Julia and Wang, Lin and O'Brien, Katelynn and Scarpino, Samuel V. and Brownstein, John S. and Pybus, Oliver G. and Pigott, David M. and Kraemer, Moritz U. G.},
- title = {Epidemiological Data from the {{COVID-19}} Outbreak, Real-Time Case Information},
- doi = {doi.org/10.1038/s41597-020-0448-0},
- number = {106},
- volume = {7},
- bdsk-url-1 = {https://doi.org/10.1038/s41597-020-0448-0},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Scientific Data},
- year = {2020},
-}
-
-@Article{yeh2009comparisons,
- author = {Yeh, I-Cheng and Lien, Che-hui},
- title = {The Comparisons of Data Mining Techniques for the Predictive Accuracy of Probability of Default of Credit Card Clients},
- number = {2},
- pages = {2473--2480},
- volume = {36},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Expert systems with applications},
- year = {2009},
-}
-
-@Article{zhang1998forecasting,
- author = {Zhang, Guoqiang and Patuwo, B Eddy and Hu, Michael Y},
- title = {Forecasting with Artificial Neural Networks:: {{The}} State of the Art},
- number = {1},
- pages = {35--62},
- volume = {14},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {International journal of forecasting},
- year = {1998},
-}
-
-@Article{zhang2003time,
- author = {Zhang, G Peter},
- title = {Time Series Forecasting Using a Hybrid {{ARIMA}} and Neural Network Model},
- pages = {159--175},
- volume = {50},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {Neurocomputing},
- year = {2003},
-}
-
-@Unpublished{zheng2018dags,
- author = {Zheng, Xun and Aragam, Bryon and Ravikumar, Pradeep and Xing, Eric P},
- title = {Dags with No Tears: {{Continuous}} Optimization for Structure Learning},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1803.01422},
- eprinttype = {arxiv},
- year = {2018},
-}
-
-@Article{zhu2015optimal,
- author = {Zhu, Rong and Ma, Ping and Mahoney, Michael W and Yu, Bin},
- title = {Optimal Subsampling Approaches for Large Sample Linear Regression},
- pages = {arXiv--1509},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- journal = {arXiv},
- year = {2015},
-}
-
-@Article{barber2021predictive,
- author = {Barber, Rina Foygel and Candès, Emmanuel J. and Ramdas, Aaditya and Tibshirani, Ryan J.},
- title = {Predictive inference with the jackknife+},
- doi = {10.1214/20-AOS1965},
- issn = {0090-5364, 2168-8966},
- number = {1},
- pages = {486--507},
- urldate = {2022-12-13},
- volume = {49},
- abstract = {This paper introduces the jackknife+, which is a novel method for constructing predictive confidence intervals. Whereas the jackknife outputs an interval centered at the predicted response of a test point, with the width of the interval determined by the quantiles of leave-one-out residuals, the jackknife+ also uses the leave-one-out predictions at the test point to account for the variability in the fitted regression function. Assuming exchangeable training samples, we prove that this crucial modification permits rigorous coverage guarantees regardless of the distribution of the data points, for any algorithm that treats the training points symmetrically. Such guarantees are not possible for the original jackknife and we demonstrate examples where the coverage rate may actually vanish. Our theoretical and empirical analysis reveals that the jackknife and the jackknife+ intervals achieve nearly exact coverage and have similar lengths whenever the fitting algorithm obeys some form of stability. Further, we extend the jackknife+ to \$K\$-fold cross validation and similarly establish rigorous coverage properties. Our methods are related to cross-conformal prediction proposed by Vovk (Ann. Math. Artif. Intell. 74 (2015) 9–28) and we discuss connections.},
- file = {:Barber2021 - Predictive Inference with the Jackknife+.pdf:PDF},
- journal = {The Annals of Statistics},
- keywords = {62F40, 62G08, 62G09, conformal inference, cross-validation, distribution-free, jackknife, leave-one-out, stability},
- month = feb,
- publisher = {Institute of Mathematical Statistics},
- year = {2021},
-}
-
-@TechReport{chouldechova2018frontiers,
- author = {Chouldechova, Alexandra and Roth, Aaron},
- title = {The {Frontiers} of {Fairness} in {Machine} {Learning}},
- doi = {10.48550/arXiv.1810.08810},
- eprint = {1810.08810},
- note = {arXiv:1810.08810 [cs, stat] type: article},
- abstract = {The last few years have seen an explosion of academic and popular interest in algorithmic fairness. Despite this interest and the volume and velocity of work that has been produced recently, the fundamental science of fairness in machine learning is still in a nascent state. In March 2018, we convened a group of experts as part of a CCC visioning workshop to assess the state of the field, and distill the most promising research directions going forward. This report summarizes the findings of that workshop. Along the way, it surveys recent theoretical work in the field and points towards promising directions for research.},
- archiveprefix = {arxiv},
- file = {:chouldechova2018frontiers - The Frontiers of Fairness in Machine Learning.pdf:PDF},
- keywords = {Computer Science - Machine Learning, Computer Science - Data Structures and Algorithms, Computer Science - Computer Science and Game Theory, Statistics - Machine Learning},
- month = oct,
- school = {arXiv},
- year = {2018},
-}
-
-@Article{pawelczyk2022probabilistically,
- author = {Pawelczyk, Martin and Datta, Teresa and van-den-Heuvel, Johannes and Kasneci, Gjergji and Lakkaraju, Himabindu},
- title = {Probabilistically {Robust} {Recourse}: {Navigating} the {Trade}-offs between {Costs} and {Robustness} in {Algorithmic} {Recourse}},
- file = {:pawelczyk2022probabilistically - Probabilistically Robust Recourse_ Navigating the Trade Offs between Costs and Robustness in Algorithmic Recourse.pdf:PDF},
- journal = {arXiv preprint arXiv:2203.06768},
- shorttitle = {Probabilistically {Robust} {Recourse}},
- year = {2022},
-}
-
-@InProceedings{stutz2022learning,
- author = {Stutz, David and Dvijotham, Krishnamurthy Dj and Cemgil, Ali Taylan and Doucet, Arnaud},
- title = {Learning {Optimal} {Conformal} {Classifiers}},
- language = {en},
- url = {https://openreview.net/forum?id=t8O-4LKFVx},
- urldate = {2023-02-13},
- abstract = {Modern deep learning based classifiers show very high accuracy on test data but this does not provide sufficient guarantees for safe deployment, especially in high-stake AI applications such as medical diagnosis. Usually, predictions are obtained without a reliable uncertainty estimate or a formal guarantee. Conformal prediction (CP) addresses these issues by using the classifier's predictions, e.g., its probability estimates, to predict confidence sets containing the true class with a user-specified probability. However, using CP as a separate processing step after training prevents the underlying model from adapting to the prediction of confidence sets. Thus, this paper explores strategies to differentiate through CP during training with the goal of training model with the conformal wrapper end-to-end. In our approach, conformal training (ConfTr), we specifically "simulate" conformalization on mini-batches during training. Compared to standard training, ConfTr reduces the average confidence set size (inefficiency) of state-of-the-art CP methods applied after training. Moreover, it allows to "shape" the confidence sets predicted at test time, which is difficult for standard CP. On experiments with several datasets, we show ConfTr can influence how inefficiency is distributed across classes, or guide the composition of confidence sets in terms of the included classes, while retaining the guarantees offered by CP.},
- file = {:stutz2022learning - Learning Optimal Conformal Classifiers.pdf:PDF},
- month = may,
- year = {2022},
-}
-
-@InProceedings{grathwohl2020your,
- author = {Grathwohl, Will and Wang, Kuan-Chieh and Jacobsen, Joern-Henrik and Duvenaud, David and Norouzi, Mohammad and Swersky, Kevin},
- title = {Your classifier is secretly an energy based model and you should treat it like one},
- language = {en},
- url = {https://openreview.net/forum?id=Hkxzx0NtDB},
- urldate = {2023-02-13},
- abstract = {We propose to reinterpret a standard discriminative classifier of p(y{\textbar}x) as an energy based model for the joint distribution p(x, y). In this setting, the standard class probabilities can be easily computed as well as unnormalized values of p(x) and p(x{\textbar}y). Within this framework, standard discriminative architectures may be used and the model can also be trained on unlabeled data. We demonstrate that energy based training of the joint distribution improves calibration, robustness, and out-of-distribution detection while also enabling our models to generate samples rivaling the quality of recent GAN approaches. We improve upon recently proposed techniques for scaling up the training of energy based models and present an approach which adds little overhead compared to standard classification training. Our approach is the first to achieve performance rivaling the state-of-the-art in both generative and discriminative learning within one hybrid model.},
- file = {:grathwohl2020your - Your Classifier Is Secretly an Energy Based Model and You Should Treat It like One.pdf:PDF},
- month = mar,
- year = {2020},
-}
-
-@Book{murphy2023probabilistic,
- author = {Murphy, Kevin P.},
- date = {2023},
- title = {Probabilistic machine learning: {Advanced} topics},
- publisher = {MIT Press},
- shorttitle = {Probabilistic machine learning},
-}
-
-@TechReport{artelt2021evaluating,
- author = {Artelt, André and Vaquet, Valerie and Velioglu, Riza and Hinder, Fabian and Brinkrolf, Johannes and Schilling, Malte and Hammer, Barbara},
- date = {2021-07},
- institution = {arXiv},
- title = {Evaluating {Robustness} of {Counterfactual} {Explanations}},
- note = {arXiv:2103.02354 [cs] type: article},
- url = {http://arxiv.org/abs/2103.02354},
- urldate = {2023-03-24},
- abstract = {Transparency is a fundamental requirement for decision making systems when these should be deployed in the real world. It is usually achieved by providing explanations of the system's behavior. A prominent and intuitive type of explanations are counterfactual explanations. Counterfactual explanations explain a behavior to the user by proposing actions -- as changes to the input -- that would cause a different (specified) behavior of the system. However, such explanation methods can be unstable with respect to small changes to the input -- i.e. even a small change in the input can lead to huge or arbitrary changes in the output and of the explanation. This could be problematic for counterfactual explanations, as two similar individuals might get very different explanations. Even worse, if the recommended actions differ considerably in their complexity, one would consider such unstable (counterfactual) explanations as individually unfair. In this work, we formally and empirically study the robustness of counterfactual explanations in general, as well as under different models and different kinds of perturbations. Furthermore, we propose that plausible counterfactual explanations can be used instead of closest counterfactual explanations to improve the robustness and consequently the individual fairness of counterfactual explanations.},
- annotation = {Comment: Rewrite paper to make things more clear; Remove one theorem \& corollary due to buggy proof},
- file = {:artelt2021evaluating - Evaluating Robustness of Counterfactual Explanations.pdf:PDF},
- keywords = {Computer Science - Machine Learning, Computer Science - Artificial Intelligence},
-}
-
-@Article{guidotti2022counterfactual,
- author = {Guidotti, Riccardo},
- date = {2022-04},
- journaltitle = {Data Mining and Knowledge Discovery},
- title = {Counterfactual explanations and how to find them: literature review and benchmarking},
- doi = {10.1007/s10618-022-00831-6},
- issn = {1573-756X},
- language = {en},
- url = {https://doi.org/10.1007/s10618-022-00831-6},
- urldate = {2023-03-24},
- abstract = {Interpretable machine learning aims at unveiling the reasons behind predictions returned by uninterpretable classifiers. One of the most valuable types of explanation consists of counterfactuals. A counterfactual explanation reveals what should have been different in an instance to observe a diverse outcome. For instance, a bank customer asks for a loan that is rejected. The counterfactual explanation consists of what should have been different for the customer in order to have the loan accepted. Recently, there has been an explosion of proposals for counterfactual explainers. The aim of this work is to survey the most recent explainers returning counterfactual explanations. We categorize explainers based on the approach adopted to return the counterfactuals, and we label them according to characteristics of the method and properties of the counterfactuals returned. In addition, we visually compare the explanations, and we report quantitative benchmarking assessing minimality, actionability, stability, diversity, discriminative power, and running time. The results make evident that the current state of the art does not provide a counterfactual explainer able to guarantee all these properties simultaneously.},
- file = {Full Text PDF:https\://link.springer.com/content/pdf/10.1007%2Fs10618-022-00831-6.pdf:application/pdf},
- keywords = {Explainable AI, Counterfactual explanations, Contrastive explanations, Interpretable machine learning},
- shorttitle = {Counterfactual explanations and how to find them},
-}
-
-@TechReport{mahajan2020preserving,
- author = {Mahajan, Divyat and Tan, Chenhao and Sharma, Amit},
- date = {2020-06},
- institution = {arXiv},
- title = {Preserving {Causal} {Constraints} in {Counterfactual} {Explanations} for {Machine} {Learning} {Classifiers}},
- doi = {10.48550/arXiv.1912.03277},
- note = {arXiv:1912.03277 [cs, stat] type: article},
- url = {http://arxiv.org/abs/1912.03277},
- urldate = {2023-03-24},
- abstract = {To construct interpretable explanations that are consistent with the original ML model, counterfactual examples---showing how the model's output changes with small perturbations to the input---have been proposed. This paper extends the work in counterfactual explanations by addressing the challenge of feasibility of such examples. For explanations of ML models in critical domains such as healthcare and finance, counterfactual examples are useful for an end-user only to the extent that perturbation of feature inputs is feasible in the real world. We formulate the problem of feasibility as preserving causal relationships among input features and present a method that uses (partial) structural causal models to generate actionable counterfactuals. When feasibility constraints cannot be easily expressed, we consider an alternative mechanism where people can label generated CF examples on feasibility: whether it is feasible to intervene and realize the candidate CF example from the original input. To learn from this labelled feasibility data, we propose a modified variational auto encoder loss for generating CF examples that optimizes for feasibility as people interact with its output. Our experiments on Bayesian networks and the widely used ''Adult-Income'' dataset show that our proposed methods can generate counterfactual explanations that better satisfy feasibility constraints than existing methods.. Code repository can be accessed here: {\textbackslash}textit\{https://github.com/divyat09/cf-feasibility\}},
- annotation = {Comment: 2019 NeurIPS Workshop on Do the right thing: Machine learning and Causal Inference for improved decision making},
- file = {:mahajan2020preserving - Preserving Causal Constraints in Counterfactual Explanations for Machine Learning Classifiers.pdf:PDF},
- keywords = {Computer Science - Machine Learning, Computer Science - Artificial Intelligence, Statistics - Machine Learning},
-}
-
-@TechReport{antoran2023sampling,
- author = {Antorán, Javier and Padhy, Shreyas and Barbano, Riccardo and Nalisnick, Eric and Janz, David and Hernández-Lobato, José Miguel},
- date = {2023-03},
- institution = {arXiv},
- title = {Sampling-based inference for large linear models, with application to linearised {Laplace}},
- note = {arXiv:2210.04994 [cs, stat] type: article},
- url = {http://arxiv.org/abs/2210.04994},
- urldate = {2023-03-25},
- abstract = {Large-scale linear models are ubiquitous throughout machine learning, with contemporary application as surrogate models for neural network uncertainty quantification; that is, the linearised Laplace method. Alas, the computational cost associated with Bayesian linear models constrains this method's application to small networks, small output spaces and small datasets. We address this limitation by introducing a scalable sample-based Bayesian inference method for conjugate Gaussian multi-output linear models, together with a matching method for hyperparameter (regularisation) selection. Furthermore, we use a classic feature normalisation method (the g-prior) to resolve a previously highlighted pathology of the linearised Laplace method. Together, these contributions allow us to perform linearised neural network inference with ResNet-18 on CIFAR100 (11M parameters, 100 outputs x 50k datapoints), with ResNet-50 on Imagenet (50M parameters, 1000 outputs x 1.2M datapoints) and with a U-Net on a high-resolution tomographic reconstruction task (2M parameters, 251k output{\textasciitilde}dimensions).},
- annotation = {Comment: Published at ICLR 2023. This latest Arxiv version is extended with a demonstration of the proposed methods on the Imagenet dataset},
- file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/2210.04994.pdf:application/pdf},
- keywords = {Statistics - Machine Learning, Computer Science - Artificial Intelligence, Computer Science - Machine Learning},
-}
-
-@Misc{altmeyer2022conformal,
- author = {Altmeyer, Patrick},
- date = {2022-10},
- title = {{Conformal} {Prediction} in {Julia}},
- language = {en},
- url = {https://www.paltmeyer.com/blog/posts/conformal-prediction/},
- urldate = {2023-03-27},
- abstract = {A (very) gentle introduction to Conformal Prediction in Julia using my new package ConformalPrediction.jl.},
-}
-
-@InProceedings{welling2011bayesian,
- author = {Welling, M. and Teh, Y.},
- date = {2011-06},
- title = {Bayesian {Learning} via {Stochastic} {Gradient} {Langevin} {Dynamics}},
- url = {https://www.semanticscholar.org/paper/Bayesian-Learning-via-Stochastic-Gradient-Langevin-Welling-Teh/aeed631d6a84100b5e9a021ec1914095c66de415},
- urldate = {2023-05-15},
- abstract = {In this paper we propose a new framework for learning from large scale datasets based on iterative learning from small mini-batches. By adding the right amount of noise to a standard stochastic gradient optimization algorithm we show that the iterates will converge to samples from the true posterior distribution as we anneal the stepsize. This seamless transition between optimization and Bayesian posterior sampling provides an inbuilt protection against overfitting. We also propose a practical method for Monte Carlo estimates of posterior statistics which monitors a "sampling threshold" and collects samples after it has been surpassed. We apply the method to three models: a mixture of Gaussians, logistic regression and ICA with natural gradients.},
- annotation = {[TLDR] This paper proposes a new framework for learning from large scale datasets based on iterative learning from small mini-batches by adding the right amount of noise to a standard stochastic gradient optimization algorithm and shows that the iterates will converge to samples from the true posterior distribution as the authors anneal the stepsize.},
- file = {:welling_bayesian_2011 - Bayesian Learning Via Stochastic Gradient Langevin Dynamics.html:URL;:welling2011bayesian - Bayesian Learning Via Stochastic Gradient Langevin Dynamics.pdf:PDF},
-}
-
-@Article{gill2010circular,
- author = {Gill, Jeff and Hangartner, Dominik},
- date = {2010},
- journaltitle = {Political Analysis},
- title = {Circular {Data} in {Political} {Science} and {How} to {Handle} {It}},
- doi = {10.1093/pan/mpq009},
- issn = {1047-1987, 1476-4989},
- language = {en},
- number = {3},
- pages = {316--336},
- url = {https://www.cambridge.org/core/journals/political-analysis/article/circular-data-in-political-science-and-how-to-handle-it/6DF2D9DA60C455E6A48FFB0FF011F747},
- urldate = {2023-05-15},
- volume = {18},
- abstract = {There has been no attention to circular (purely cyclical) data in political science research. We show that such data exist and are mishandled by models that do not take into account the inherently recycling nature of some phenomenon. Clock and calendar effects are the obvious cases, but directional data are observed as well. We describe a standard maximum likelihood regression modeling framework based on the von Mises distribution, then develop a general Bayesian regression procedure for the first time, providing an easy-to-use Metropolis-Hastings sampler for this approach. Applications include a chronographic analysis of U.S. domestic terrorism and directional party preferences in a two-dimensional ideological space for German Bundestag elections. The results demonstrate the importance of circular models to handle periodic and directional data in political science.},
- file = {Full Text PDF:https\://www.cambridge.org/core/services/aop-cambridge-core/content/view/6DF2D9DA60C455E6A48FFB0FF011F747/S1047198700012493a.pdf/div-class-title-circular-data-in-political-science-and-how-to-handle-it-div.pdf:application/pdf},
- publisher = {Cambridge University Press},
-}
-
-@InProceedings{liu2023goggle,
- author = {Liu, Tennison and Qian, Zhaozhi and Berrevoets, Jeroen and Schaar, Mihaela van der},
- date = {2023-02},
- title = {{GOGGLE}: {Generative} {Modelling} for {Tabular} {Data} by {Learning} {Relational} {Structure}},
- language = {en},
- url = {https://openreview.net/forum?id=fPVRcJqspu},
- urldate = {2023-05-15},
- abstract = {Deep generative models learn highly complex and non-linear representations to generate realistic synthetic data. While they have achieved notable success in computer vision and natural language processing, similar advances have been less demonstrable in the tabular domain. This is partially because generative modelling of tabular data entails a particular set of challenges, including heterogeneous relationships, limited number of samples, and difficulties in incorporating prior knowledge. Additionally, unlike their counterparts in image and sequence domain, deep generative models for tabular data almost exclusively employ fully-connected layers, which encode weak inductive biases about relationships between inputs. Real-world data generating processes can often be represented using relational structures, which encode sparse, heterogeneous relationships between variables. In this work, we learn and exploit relational structure underlying tabular data to better model variable dependence, and as a natural means to introduce regularization on relationships and include prior knowledge. Specifically, we introduce GOGGLE, an end-to-end message passing scheme that jointly learns the relational structure and corresponding functional relationships as the basis of generating synthetic samples. Using real-world datasets, we provide empirical evidence that the proposed method is effective in generating realistic synthetic data and exploiting domain knowledge for downstream tasks.},
- file = {Full Text PDF:https\://openreview.net/pdf?id=fPVRcJqspu:application/pdf},
- shorttitle = {{GOGGLE}},
-}
-
-@TechReport{du2020implicit,
- author = {Du, Yilun and Mordatch, Igor},
- date = {2020-06},
- institution = {arXiv},
- title = {Implicit {Generation} and {Generalization} in {Energy}-{Based} {Models}},
- doi = {10.48550/arXiv.1903.08689},
- note = {arXiv:1903.08689 [cs, stat] type: article},
- url = {http://arxiv.org/abs/1903.08689},
- urldate = {2023-05-16},
- abstract = {Energy based models (EBMs) are appealing due to their generality and simplicity in likelihood modeling, but have been traditionally difficult to train. We present techniques to scale MCMC based EBM training on continuous neural networks, and we show its success on the high-dimensional data domains of ImageNet32x32, ImageNet128x128, CIFAR-10, and robotic hand trajectories, achieving better samples than other likelihood models and nearing the performance of contemporary GAN approaches, while covering all modes of the data. We highlight some unique capabilities of implicit generation such as compositionality and corrupt image reconstruction and inpainting. Finally, we show that EBMs are useful models across a wide variety of tasks, achieving state-of-the-art out-of-distribution classification, adversarially robust classification, state-of-the-art continual online class learning, and coherent long term predicted trajectory rollouts.},
- file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/1903.08689.pdf:application/pdf},
- keywords = {Computer Science - Machine Learning, Computer Science - Computer Vision and Pattern Recognition, Statistics - Machine Learning},
-}
-
-@Comment{jabref-meta: databaseType:biblatex;}
diff --git a/paper/aaai/paper.pdf b/paper/aaai/paper.pdf
index af2560f2fc09202c587302bed15d8f4197df921b..e3685c61834b3c4116322632349c43aa11a7f475 100644
GIT binary patch
delta 118186
zcmV(wK<U5!)ONSuc90_jG&MGt;e-(=f2})fkK;IU-}hH^J_K<iZG|F5O4=Y;%ueQT
z6Xb9~c7WR;8z5?%ZVMw@K0R)x@7JeZ)WfEGcMkdFVv*vbSZ|hp{Pf*F{m$BlG)qk8
z^uwpagHCPY^8CRSCP`EG@M-_>-<uEb-@kAB$E~(%bLbw^V$+_uH5@$yKQ`6wf3tw?
z*Y6+y=hMGEY<1#H`LKm4%Pi%oe)p*DrmBbIVQhH#hp(5WI#)y8p8GH<uQ9$Hc^+$%
zBF$tX+0xm=AhmgU-M)tphpwv6e$U?n;(*Jv-KlB<YCI3VJ5;-2tePID<&&pLX)T?+
zD3YwmVbf)zbIq%~JGbZ8Q#;DWfBNzHRQEmXbef<{N~axuzb#c#mX3B?YN6b_XWZIu
zknzPs0pnd3#^<o~7N}o1%j3g7@o2nPSXrkV|Crj%noxjk>6izg1U0KO{;ExV9>#j8
zvF|VVoe$l628dcO$KlX+k1%{oL{_jN*qh;fe4!0=_(86u^@@#jGqn6We{HEg^QdQJ
zf$ekqmHSWa-ZwpW@zVKv$B1uU+76iERPUCX5bAe+*N$C}+joZ7bsA2$=cCnoyvl@o
zKRp&|BO|K)zVm%ApAm-w_Uws20EOo!05(7N^>ZU5_(U;s%!Ge>+?vwK(7GS$^9y5$
zMs4?*TgR$r;L_EPK#QuYf16jvbiS!^CkO*0(QiL?UHkF~k}>$m@)IcwAwT>!zUs~l
z9$Tcx7wM<-7D%VdI52Ks4HZMR%(92Acl9$~|2X)4!q`v8(C_7*cH79kELn?ARTOZl
zs%hFC>>m$a<h&2;8LIv>LvZ=@OIv5c!9e^H3Ay9W8&1C^2i>13e_U9@$pK9PTkcRb
ztvZQ%4Aw*Mn}hVXt8sR92?_V1YVk3K`19z`1DxI24RA!33f&QS@t~Suiaj@PB^R+L
zlqLjv?@tUZn33jdBa}amr;3DMr5m|CZU}QLVDolD?LY{G@er~1{SNNjbBp-|K6v)W
z3a9!@4Z_BKcu`@@f9=_qa&C%1UTw+^t1V=B((g>4g!}%40_ypb18L^Em*&812ZLbq
zs9Xt4o4~<5>2QqQ1FUjpQMVTD2H6a0cm_ZlcI|%L$;{UA%+OQG$f|DIj*SN@OX6*`
zQq5R}kL>2;hnOsM$hSImqT*7N`1aD$(NbOlaOW1eKqN-Se=k%YAr4)2@+`<fli&=j
zte{``3gspkpV^qKsq!IWZonuNFpT=ib{1$mBU5RUXt35Y^nGM1mt=Vs1}u#4?Oa)!
zXp{ZgTxOn&IU_QYabPUxUA2*^Qn0@^4@0i)C<>l2n_rYw?9Kp`vU`!56nveviaS;)
zMkN@N4zIE_f3P_cBt-T6yEGp#AfTD^00@!+eHTyzo9+bqCt<{=iC4?F4A0=t{mPIV
zIFLVGYT!Hc4Y1h{<G%J}nV|JW2JQWcuVxSg9|@u0&LP}gsdy7^Vprq2j0*}909G;x
zu?9vU2<Ie-h$UG-Kqyo<^$>IcqCcLcyXrvik1@Gwf2!^=Q>a|EE$C3%=9_*tb+!nX
zWD5}z7%*8%{0O$i&`VbVh0X(ysdlIsW#(FDK&2($f%NwEHmdlDl$K3Ln!!xR+OKa1
z%*X&F-?oP>&}|C`<e?W}M&#}DPKM#`+HsHxiNr9BNPN%Y>3r=d07ld=aus$5F9_j%
z%LDrMe~A&K-Vums8XQQDNiXZP4Vw1bdm(k`3es{OZqN7J-uqva(rneSEpAt_vKB6%
zWWp7bF{V+k{M#(Z*;OyGTVTh*qT{fmJ2h};)t*&0khkYv^h&YE&&uw(>WP5^pENR6
zAq)%n9jTyuUNMIgCEO;Mi<&G7iL;&EPv~Sbe+t0oGjnU>_b>j52a2tO-ta5xBHJJV
zyMEwtOagG^JL#>@F;9$UGqBx#4dh^JWzsS*j9OX^YA*$rhwdUF2aa4gE>X>d&J41h
zLceh&^`k`nC=510vg-@8DY+7)eE>r31HwiM2%0;LS`g@cH;EBW(DQ^9H}URb+kiVH
zf6>u$WCTX0ngK5g;s9_qcgKXgf)j%9GqU?yc=e<twMYw0u@DkOU;MdekiDju^77}A
z<RX(@>C8+!@tyRr!vH-O3=nIi=hO#Y;b#^lXbf#jDlL;3D8s!EaAgs4>^USE>;5RQ
zbT{^c90Am!-Moe{K7{b&83n&$GY0LTe^4SJvdfdA07I#>B$qfmecYxBuIVFqSroJf
zPb=8xMSz^c=NX3aCQ<s0cvqQNysMB>>_^pm&xlI!EN3v!*gP=`gvm&d#wSx5QCvty
zzCsZc%5AuuXUm}1O3<5$cMS;f1^{&x2WRUz*e>GWY~i!#=wJ5jYCoko`M$ZJe+a^c
zLlj2Qgwggo_~l`f-^&fpLr<PU1;kO{MIP-<xV>CB#M>w=wuCF~2Vd>EK~sRHME)1G
zS;|@fJahJk!1n7CLlMOiSu_?OnT610U=_#n9y12mAz>Esf<I9Nfdg=Nh!lSZW_&^a
zsCngXuv9CRzFZ%4#WRbVv2!eAe>r@nuq5N#2ORAK>^n*BdiZq|1t@eiBY7eX?4JWh
z{L~*Jl9GY1KKZ)nrzSk($MaSaYiW_>#$BJJzK+awS#J7^uTYsR6f~kkn98l&Tfniz
z(PbIT3}C=MGJU@eF=9TyawrWUzr^<-mY`C%H3i)n#?5_L0Qxv1VCNY8f6z7!3-k0n
z><HBP_uTm3KaO2&;T;-elq*m{UENdcR8IE;$Q?LN9I{i781SAu#gR;>P!vY2C5?MS
zA8LWSo7k?ktwg}3+Wd)EDPZmcWANB1t7O2!*W-@DQ0}Tb#)BU}9QigUtL$NS_t@D>
zw(?n2eX9Qchku~4Y$#Uae>vr2N<N$^^Ud&j39>`z)1~nzARXtda=zno?eQVLV)Rh4
zyX6ONgO4jryIfsV@2)+0=Kri*ib$3iHA%=TkI)8!#WFVqhmBgf2a~u|5~u&jgN#~C
zQzk{4KWrDHH0r^0^Is8}(<)K03AJ#zQDgVGLx9XRL)`AVfP)oNf3W*Cx*<KCO1AQ|
z94=dF2-xuy6#t7X<rX7|hB}P+Pnp9l2$UyjX_o~0*T|UxvSvk~kR#fQo>3|R19Bte
zyk5TCkxAy9Ky0m4!<|lBu7*YWk$lHIIIRGDIo7))gRB|yAj`3ZQ<!+tmVb!QqD-8#
zaE2r+ZP2M&M+grCe{KG-825eT67!@?r!2$Ik6Y)=CjYy~txfYNV`actQ|6^U*Fl8?
zlRKxcq_<|lqm;p{qQL#H7`d-Q0-eKI3O5E6qwiz1$uzOSuh7Q*7TRQqRV&)e=4~Yf
zCMoi3Ay`f1K+f`|af{M8%O(MEkxA6GbS$?39;=zJ=v&msf9m2KbZ0B+klc+pt+Pw9
zO-pcXUAex@b@E9AdvGo(rR4aTAqfih+#;0?zbD|6Q~g)=OVXUvi20^zd$CxD@w^)#
zisJNSFk!eC?QCKnvK+{4@=42HDmDutor8cd7{7|RIWl+1Qv7z{oIJZMK3ghy89edd
zwKaOB9Y-{re@e2#ESNw=xOc@#zarP!ukvp3up0OO%DeZs{m+qAvLw^RZ{tq4;?AqB
zOx)?N_p@T-)qWBV1b(VH4+UU6{I@ZWaG5b8w$Ba7X}-R4PDGFrEAo(o3(qWwT82l3
zWMZCpjHKP`J$K!1t>=!a*~S7niP+?ERQ&Pj0aW2*e}O!w;!Bl)=)}MdFhNOe5<yL#
zE<_X4vjOycf!N*0KdSo}_lt)_0;V13hjpL?0DF#9IxlS!HTV_thhXg87iPLZB>c5L
zvHLDlMm8{~sl+AZmk-4+AL|#}5cxBBCw56qp4izn`QeASQ{l`yc#$kSf4c?sthA^m
z$^@5sf792)IPn>@h1Ll2ts%6eI<{*WT=7rm|2zWop|ASH7-XP`^BIioHP@?n98@k}
zt&J);C50XdZ4n^4k4slewyVVS=_tuKXbsUY{{!a?IPqc%wRzVKH6C$S({6H%zz57y
zLa4TrF?cUOzxzm{F8&ea&oKK4OwciSX)j|me>%XdIykch@3;cP({sD2KO-qTxBmU(
zA3iX2j}KIll&buu7*-bMN`PYX)9>H?$D3=|EXhmrO-nrSgh74)ga|f6B>)25Hlhod
z8$oyByV06FtQL$bt`qHUnP`W`kCeH0#b%K~JTpf4)fH=gbrNC0FC~!>Kt^)FE#o12
ze_PB9oZ|*10~%>x3s<2w;}mZS3&Un5!8;{dwW4wmb6Y#Y$YmaQPV}#}7l#6pV3xd+
zy3TYUit|gHtJR`M5{5rb+u=+#C|JL%#$F~njm;1bWG|GpGcG&^ymE^ZoV$Tr5%p+t
zPAOw|Mfx6qhVoblaVjS$VF3Dy0z8w?e-XxgHA;_0dgJ?x-vw9QlD0ZCw8vrw^4btB
z4^Vn@^g=TJ!r;pntUJ7sm#Y6t0k@hHh(=`#ImMp6_ZP;W{n)V%vjB&L?`O`+XH!50
zy;Ub4&vAallyasEPl-$z3adg!VXFF)!jd-(226@&WZ#zTVhL3?Nlhm9qL}T)f1;`<
zTP38ZhfEy&Mx2Ky{zSfoh#qb$Oh>61hY4Q6FbTQA3p{Sq&3$XKVnV0H*LQ(f00^g1
zU)}pKt~s^z=mRjNx~fIp1mNEBk6gQCiyVsoZtUafY`9+&a%up80>VrmzllJbYm<0H
z=MQbbKYJfOZTS}nbRO!%YZy*te+_0ScTv`z=}?5ah{9eKJQQw9H}}Co<n)-ZsHVyt
zqD<jd7B-f5<d;1PN6!3PNmTO+ABC<1|AJ<NlXc9y4y&)oTsU#j;CBaci<iO@Xx_`@
z0*9I``(_h`OM(Je#E&|~Jd0f)s%Q7|zHmVlrPa4DZ%H8FhAMfOe`^hAf09yvtDU!@
z;tzmT3BX$`fk8GV`<pa{(@6%Q)2(NzrrdR^rr70)JHCjow(lzB<}ci$oA3C8Y8a6J
z7jB&ReOynLu3uB{s|H~t1uS2&lZG&RZd&E*XhI1WDrRA|VAt`)2u@C_Xn>T&#AB+x
zD(n=Luw32NZ;$}?tC-EMe<KyCozs-tXHN@I*L#U-29DseOhi1h6lQ@*=WG^qh|NTZ
z^+6U&xd0Uc6aYzyVJ8(6lJMiS&s0(oK8krYAoU6J;UDRt#$M}I%dZ?$dUB4Qc+&`k
zTIlJuqUZESMrjRTNHBU*@cfI`_Hk>Knr{JJzPXSkSCjcE*Q@~Te|jpQl{WZ#6pSo<
zL&Qg`*5EtvF*<qdE0W9=R~afUgjIJUZO5|ulV85*uB|)B>|Ap7+rmLHd!!!#n9JoH
z$E8k9PYYpL*tLe0(yIg;z51yb6K5*e6NFwCuVT{K@zbGVR>ohB<`C+~df=#x@I=8r
z?w`L!9AgtV=U%@le^iFhC)GJUl-t7H<D;z1r{YFxlEy&nRk<OeR@Rq>N(3m~^@$2r
z5gaCVxuW;7mL$`K)ZiC4HTc-~eyZh@CZ|k8;IX^&cqMPz`E{Kh;X-IrTt_d|yLyL5
zM9|X!u49%vP7_M>{p-wh;?nZBU#DE91CwM^xyt<BRbp0Qe^uNt=OS(^0AcL10#Jn@
z6uq@wFltlvpVW)4?z31GITpqI2alq>p{-sPWfPr>`DGRQu3x^PSA0GIuoO6c)m0$*
zt%D%76##|6O082Y!Hal9GnG2Af1~uFTW2Y)xlrR%WgP@>9QWGW0x#?oe*+-?4HX&I
z%mrT9SQU8Te=sgsn<vOPmHI~hinV-Jh9aYKC85-@s`f%m)McO^DdIwqY6P-Uin!w8
zIbXz7SE?6vr3iBjj@7AiSr;}d>96<)9d|C_!j{xJr&9@+rV7<i!u6JC1BAON=+Z=d
zt`U`6Lts%~jc6|D!m*2lu@rFG@MjTwUBG3ha-#1(34HqQ{{UfyHK><?9U2n>IG2%I
z9TNjGH#V2ygb^rzy<Ayy<VJFS-(NvLsTpRMQ3tM3C=?@Uw3a-xJK<rhY_Ha_U^mFd
zPB+j3&^w}jz4IzuUCkLv3cs+c5`{t?S(#rR#jkGv@NeJ7*;TIc#3Wa@_gAsIN=>FB
zn_k^+ul~9^w#BBe`l_kRw!2<iovymFd|%ZMSV>lW)9i(R-ZxCTEBn4|nYOLE9@?I&
zZYPb}Q`djE*3qhI``vZCdc6MI?e|yF)mld?k8}R&ecK$a*Cviur@AfoRdKg3xqfK2
z<-Sw+5{x{)(i+~14GulcRA!T_wXsU)F?VFHq2DN4eN}AUKeX36TQ#S8D{oKdz4mWw
zr%WW1|6x~u;dh;jRvUN^W_6DvhZQ*$d#0DyI$r&J+!u9$i_{@iuYQc;=<8STSIgT)
z&4ps$HC(zYx#>e$uR}{vwnblVp-mdC?wghiMSEBEZP7k*wWY5=v}M=9x0%KZ#AO`D
zuu|h(b&{w!$>59hr7r`^;Z&X(-vH-v!#(nJc~hBxB76Cs=CIbOR`9LaOm+n_h{=GL
zszcFM`$w+cH=9$3Q^X1mAAdLd0w(ig6qyYzC{lBHvjGk`T<aBVJyBklJrCPPlVKa3
zt0;;6!2g?@7&o1~R!Nrl7GLnNjg3?qYZ-SoOgob0GwGxUV3c_1;Pb#IoN{JZT}@oT
zjgqT>MCGZ?d@u2pRf!f>{g1#lz;F_mUacq1wq@P)co_0*)$gc~!s?alUDL12dIQ_P
z<)_W5Kc0FnZ<|9=)!g_)1qbhZS~>@=aDLdK(X5Thl{LwHduhi?+!vGQtJ_^!bLr4@
z@NpZ3&knW#)<-UHP_SSfPWmI4Xy4(@hql;%R^j7OW9LZ>>(On73W=a3SvAMLI#j=u
z(xke-Hu37dY=?%xV79n6(dwW0r#X>q1A22W9qmf)>Au<T8+`i%a(rMQ!O4N3>dK<s
z?6}<A{i6gairxGZ4^a7nSJ&l})rEdds*+e_&;Q64<rAJsRTODpFRdLniUQmh!`Zff
zSL;cuFJKgD2Eqf5Y!YRw18@TGZZzrKBub|_2tpMt$B>ac&{n#SK<uhKxAQSf2WnK3
zqZAlh8Q)ht=?l+ds;8a3=2m*jn9!}xuQW(?>-@>oQGh>Vm1#J`puJ@LkNc+U0!w5f
z<uW@FZ=qvuRhACN;Z5KK=PF6?EGU<M592Z>C=X&7(X8Fazjz<q+B`lV^^A+KkD<-4
z<{_g}+`maHlwuP}HXU}VQW(Nq9?r%zlZTJegPn{i0hOCf$Mmm(F(n`a-PxFcwT&6t
z&*WJGjI77yGVw*0iTqLS`jE!Jj-JCTzhUd#$F|P=)~(a$TQ?pUI9NI)_(2VS!4TF4
zcGBr_5<|D}2v#XKs6n-gT*7Vs9V9MnP-ZTkL<f>$x=DUvW?Fr=Fd(^gc-l!i1uTJS
zBvZ}LdbH!=u_A>6LzG9Z(3q)of2uc#Ms7lUW^?5t`~1!SRddWfVUEMMDj9wudj$KJ
z7`=enGPP?2L9h@7A_f3hhvL|OF-^cY(0G~b1&t03y1wY8Np~zZrM%Yo&dVU!FN^+!
zjG}YnU&CSnK%EK~&SeJ6n*2X4Ypl{Nn=UJks{|m0h7Ju@h7TxVFXKO!b+PXsxzbQO
z>)UMrR+Jk?HtRf7HlJfp*^xLWM1sHETSXij=YFH0RL~iN0)uIK*zj+EbvQPiJl%kC
z0;WS%_d^eDb9&f?c0}xo(;m?JF9gF8b?3?I)lVlz<pHB20_5+8qx@yassJ6HM}w*W
zvjy3u<E|NTLE*0Gz~ZZ=RvZEQ7n_}!7dlpMVP4>5oy{040>E2i2V15Etu~;0aE4{*
zA6H4<xO)uD%@{pB;-RU30Wfw0Gcvvsv%=-`@iOX#fmoNKWq05b8A<R98`L)2(}oNq
zs0Rn60jy<{#r~KDr?{`%>kPmsm%ZUyZK|FuG91H(7Hk-rEV|Vy1ARSTbv8Y~-aG+U
zayO4mA&$-UnQ43z(0SOm85i!d)|*&BQ4k#1$j$7eF9Om4bQ5rY{z6hOpbB79-{6-3
z=!;d9n;Gca2gObVn=xth5a^0`Xxi25z&^5_IB7I0E?DcL4rnA7gP0$<^>t87q~kQ#
zo>^n1v&A-Vt3z3%bHST6McWpSV%6O9??nrUPB*z~=+joffo#w%9e8b~2VU#pz~Au0
z54&o!lcKc$Q0@1B+_nP&N4yKvw!CMjlMzRh!6QXePs|j{_@PW@`NK6z@D1|o4qaF0
z3uXbQWCOmqL4D>#iqKYAlI%nj4S?m~%0N@pYpd#7!=N{G_^>gV=kO~^)}daMt&y+K
z4<M;b8>L=l+9+X_i8Rs{RY$@V=vCh_9sO}dy^%L*p&~DTdx6lj)k8&ljkU?RRY*;I
zha=YLh9ID!LCL^YhA@(w`AmKTM2deG7#PhF5=_B!5>A3Q0EiVFwVuTq)Fwa)8WrA)
zI2SM$q?KLTUrU#c9#QC~Y1`5tK<!kj7uupbHuYBU25uMq?gU|6LQb&V<^EoA<qgOP
zjeb_DZOc}F{1EV`nEE96xp(KuC!70_I<!{BI)6(0tbLn>Cy-q({5hLM6XlCFkOAmv
z;Y!d&BY+5>KZQjS;5VH<iAAV1IndZ;S&*zrqTwU3)Ccd`%v72#CbTWP>Y-+bn`YPr
zzr}WoE?KU0$8y8|%z!XLJQIy-8oPrH!u2WoBBa!RokG`1<aID%K={bHpy_1Pajvp_
zftV2X;ftvtMfg3GrBFx=Axvm-B(p}iS!GMd1ZrK``kln;UDNNl^zu9IVGJG2J$RXS
zaN`Q0PaK7ovH0mN^JBH88Sqo_-EGJlj?L~K<;ng5<W;{rR2wF2VWm-5tq;Zf3*Nf!
zvFpo!134>G%_;~BAS!z<ky3+S*$UhSs#o=9f7+6)huR&o<Vjwsga~Tc_jagwv?(z~
zU>q!QotgU=Umy)xrf55qJ4sBO>Ci51e#TxOxq8X559V~NKMD)=!m*D%79`t>yJ{~V
z(U(4wfGsp4woeTn;5WQ07saVEX1*fKOFVCXP<R*T2y(^c&;LAh2uSqQRl^J~e<F#L
z%_h<#Xv)A>oC|!VZJ0T_z9t#)#CPHptcv`(-{CG<(?ekclPdPIu%KC#P8KW68tkcX
zkcv(IET&Q1;Wu-~I+B8r2k0|jKSr`xMp76F>u?~t(sPH37fy=0EVm>N(uEE|%3Vo+
zAqSU;#nB^T9rYT?D`?i8yu3nfNXHXRo>~g!s1^4o4!s#N$M~0@wBcYxbLu?gJ#EEA
zzqs5kpUUtrzYOoYfi;hk`I^(}BmffPBm|`av4IHUm>FU@RAEuJWP5m02^hq<N$bHL
zW5En&zn{LJdT^Kx+6k85J1)+lGzd(8WR5sleT(aiZV#;R&J+N$JU*PQ2}TXwQyiG8
z*iO!LWN6#sWblI|i+R$s$78jD`AC&xuUOib#a0w#nyV-^=N<)PGjyM{4Gx_7X3Nz1
z0ar|X2f}$@pl)QDjkLvpfI^W_EXYvM-+;^;3@g>O*Lf0KH#(8iE|^rfQzssOvSMt$
zCWnt6RtAVNkiE2-AJHH~_o9Dej4>mE>BZEe&-({^w(ED*@oA3N7Nn9%pCz-^=zGv^
z!R}wfN6Vj1z`c|8jz;g~aSx`|eC@?r4Z1GH%_*A`jaw4r??ZzkE4FelurN4{u>u>n
zDO*egz#2vFu>i}+)?$k+gOrzln_b01=K@xl^G^)&4G(UNvg)VP1jO%2_~LWq68r@w
zIJG2Nu{ARSJ*N=^Nu+VP{j`WCsWC*EVs1m3g2+^2`GAG?BqdQ!dUdLsy8$PP!rFp}
z2r``vic?EsjhJV{Dl}#=^35OLit<6q59rJ)xx6h&CQwZjJ+Qzi^kPGQo;H#NfiKfB
ziX(5x#*?fDc6tbP1Z&-k0H0&8vC>+CDc|s*Di+BO-Zv^}k0PR(GJo`TJh1rM*XJ8D
zF>Plj4HE{e9o>HCCI?tX(V)x2Srj@|Mh`~pe}N?kyz`Hc_f9K2e(iE(58ZOf<;b44
zAcHk~?~<AKZstT-WNG++QxbG0MA|$jQ2KLV6%r6QgUkxee&%cV)svU7=A_lQx3i5)
zl+$h=y+W>(Tm>r70mKprp)K(p3u{MypSBXSMPduFc<|y?M~YcAA_(i;#&cVZ4oE5@
zj|xS54KnNjS&}WsloLS$>y)}FJ3YNn@ob?M^B)!}9=zK4T}e5ANe7qp(x+>b9i0I`
z2VBg&)#one+4o}=$9DF8wu@6r<)Q@uy+jBJR))>Ll)Qo|gA<R@9WLwwa(MMTai{W0
zzY%#P4-p85l1H=w%P-A;%_L_`kI5ruHpw3&8*%=~6M_IBdp70({H|LZ@Ga>Wgjnw8
z($>c|Tx<vRhBY^Tg+KT(px$oy$p?7>#deDh0VLvzGr0$q(p&@vKb?xYukOXz;Dh3_
z#YuE&=p;G@+t}Qn#F_0ebcUx+XE41!sJH3-0h{Pc9#n%K{`^w~yZJTyxeP`x_R|}W
z6a8&s!}*a&?D;GM2`q*c!We1calV$LU?!W4RN2*uapEw4B)1H9&rx4I1ib0oi3=Ey
z8=;%|^X0}dLN#fu9PmP7CrPt^MkL|@gHy!yVb_CTB1BQNTYkW`N3Ku=6O&~dsS^yr
zlK<`H$*C*3DJ)FeJY1unusml;k=i~L)M`l~wUKzYk<u64dlkSC8NiX)ezH5@X3_u5
z(xXS+>EWS&q+CGl*Ap97b=QL(#$@(0U#+uDIqMd(k$^KNoa!tLPn~AZrPGoXiG;yD
zj#PG!2f$7a?3w8CP=rEo@t9M3!+n2{Za0OG^+raSWp3m!N2d`VPh1x*?_`7u0w01c
zMKB)A<>9W}ZpA8C4mSpbmuF8#7yxEQv+TAfT;<MxDx0}a_moRcGfoG0a>)JBZV{it
zfXWbCj_`nP$qmjSDQ`AJW(BDJT0?(|=|hYf%3!Cr^{n~0Q26SwP&b!+tQ`2D{0WXU
z_*4mYJSDXLNb$zUG#-L<IzwFSPMpt~i#?FjyU@j7X#7Q66(Z;ZgL!%C8_v#^ASLb#
zylCTp)yr4jssZv_@i<hJ#%r~oIXs%3d7{YRut+`~3X<;kLF;OWLzLSa@l5Z^s%7un
z7>2<IL(T=&(1TKLL&1v#AZa?V;(xyS%XjZyg<_`wD$;?0Skyzmk5#$fj%O!|gVrx;
z?Ibx%CQwLDSAWI~oQ|8!J%jjklG*{!JND6k@T1}wTqT}(88q(zx22i=h?i<qujj|z
zBtFwD(G>M&lFyV%48s6;Sa7u_o3;8ynAK2BtdSTQwv*t(+GWZvy;xndAHAgZsc>1M
zRT@1bTK0L!mrZvro?II@`ZGz7;i$Igt0fnWb1P61)xrCS;o&6La%wT<4iwbT)&GWn
zEX4@m=+Gi$A*O0%+!tXSY^N5mGj5gsODN#%R5Y&4U!zaVkA334z~6$4H|b)MbU91e
z$R<`vZpVZ#u1dm=z2E@D7V_d&e27q#L@>VCf`qf)h9re3l%8O;WA^?W%N89&H%ukh
z2>D~kfp^YOpq8WTBOEn`x&Zl}4)Vc&_cb38kXjF=eD|)QXr1=y6+nZ<Ir$7c9Yw^y
z%PDwY=hW{xXoM*^^k8&Am13`L$saz?fJyMW66plSatJ(2xlp6iiS#8fBgdV=W5-SA
zo<XGWNU+3^kzCAA+AEtqNXESc!pN8F9dr~4z+Zx)SQ@=ySQ~*GzexcPEc0!D9pgQM
zW_7nOHt*MW&CekY@yZK~WuLDQ5G!!J?32+xkbHJHJl)PqM{E6}P^sZx4s%=Yyv4#I
zJOY%n+ZqyhF(6~z{H%SD3|BT1(Rs{4V3VKa*HKQbu+HGDHi%mJk`5HlEevY<wyEm=
zf{&Oc<~_UC#j8_|VVagN%vth(1wT|UJaz{oVG_}3cf61AUR4h{hY3a4@s|)s*jMXe
z2&Naaoq-2Cvt%VB@N%n1pHl6CS2mb&8E120d%5(JG5gGtu|0(3hKDMAQ4X|v=dX%c
zxxoX!$F|&7^KB&h3ElHO=D|KPLD*Sw?Uq52waJJcLd;wCyJp+$n};EP0+A*@S&fRN
z{8F+SScyJC4T<bHO2P?hn6i<N+UXsqrJedvkWhmqHp@zbP6xf;aUB)^5FUmXlBJlt
zkV6BO*d)yVW!TWjN8CVZ897}?gq7!;T$I1Ne`Ndu&%Q4<&0SU0>#sOR7hAjogagRf
zLv?8NrZ|c#QY#xT<kri7z6YT~p_-Hs!uRjqyp)?D7_+PSi3Eev(Fnw+9%|XtoA?;9
zX`8NFV=S~eqM(ytoxicL+<J}}rw)!f{|<THxXZ-eKm^sMxFks?Kso`>aQxeWl0C3~
zrbCpH7q@9a4Qwt2Dd@WnE$qXwYJFhSMLrjtOT@icoTw3~J^GA)L%`e}8=jYPhL+j+
zWZ0v3OfoTJ(f(Lqffl$TIn~XV?oP!Q?oKVldt~f1{Z-C&w;0L9HAeM)apy-i^K9kt
z<@lo^i#Xu5Kg;5?^N@wVR*Cm4&{9a}s$y3?>^MycfEl76fLWVTF==k~Aw_(CBiw*u
zB`Hy4<1OqtH%*~`_>_R20|{NQb0i*(EyWLR#RC;%4rIYJ^uUxtrGS<PQ%ESfu+e!s
z9L33*XfOO{R?vVDi0&twQ1s6)NyiUq+Q9bJ-Nn#lNa*t|@`=z#f|wd*_$IkE%4K+y
z9N={#o=Yn?$$!`lPK9XoEsW0R8_52|9s@w(RgI~3FSzo5k_K$ib&Xykt8yo-@uq<h
zYcdt1RnbqftzWStEqGrYWFw`ICvZY!j4$af1twZ`n^VVWm6#5SsxkMT1S0Sq1HRHk
zq#;QZD-u_(<o!TSPe?xT)%MhIB;A3De}6*dha_fr!K;`+_sDcMQg|dU_z~!@lfO!z
zG8)C-QH@1^pQgW3m5f(*F6l4rh9-QEM(0BqhmIoWpZe9H%-@d5T{GkY3q?$eT!ar5
zg(H1me47I}E0_N^aCW64NybwEfVJo)!#qjG22){&b@Kcxe&b!kv=98h3$Qp{d~5>y
zrz7uz4(mA3+5qb*Z_iyw2yo5=Y$y#WuS_nv;FG3*C_f4bldQ^M7=FC^LSgs230C!n
zn1t&0O(Vg1{z%t*-~HmYH^uFTe<#5izI1zm?rz2L>fAiji-;0J^QpRa)~5>KBL)&!
zqJQ#WqlF8c+LZjea8Y4DXrYTR3+9{Pe~GZ1pI&kq1lNNU$N!e#Z!OHPCPo0_#z^>c
zFEKBF6mf2*dYDvz6Ag?763^cDu0xpzMvZR3H-Gx>-EA;&Jjzw7vqhdYa7<T<<#0JM
zG|iru#`wZtkAxq+dV@Ea)~}04g6TkiTeh3>2s~MHgYm<QByL#FXuK-^`fW(oK+(i=
zFFKh)3`3#!CAY)FJo35cv5NCW?)ec%N1kwh6g?7P$?*=?sVxLfCo#Fzp~0UT{ymfh
z2Gu&QEfVtKbxx+FE{pdR-S<`LpgLS~oK8nSNgNb;_8FcvNjxmJ-YbeubICV6OPqx_
zIjEG5_g6LrAowAX(O*&f1QdXJY;O7ifn&DgKO0d@zV&*NC0@&weR<z==~&d{xyY@5
zk6SHL+rO<I(Bug4DE{-2kLjc|PnZ4-y2tC7QNfY~Q-A~-kT_t>^le+<Y(LceNTgoo
zGem$S=SY(4oKr4cgl*(jRz{A5R!lei$2~JqLoMj;%BII;yGN!>u(^-J;j5w!k@oi*
zq}^wi?F_pAL~bX%5ru+Iae2rlMI&8*oHuUuFxlgpNwdu3p%*T4UK$jLBX78~4?)bl
zNHdNyzFvEN`dy^5c-XM#-y{q4sN(q3%zvkh9puh%BWV$&QEp7Fy&R*#H1qLdNyAf~
zgxDkcCo@b&Jm5VFF-Tpzyc@dA|8}ugyqLr&$=YJ!#B(&`k|B=)9{s&5CmZ~Knc@bw
z{!I{pVB6oZ&6HdcO~nI?JIfFc9h3fC)l9D?ahz+T|FdMmoTbd5f#unX&KXujL=1h{
zRg2O8iz$cBTox6WC6qlQD5JQ)uZyFAXX7-x)Et87BNsT&#CR_qmFL%Pgm&On^Ur0w
zV;I!(KJ#V>N6LY>W3s>8!`wH2EJdtO8=W?uAY`QDzj_+4p=dkCnWNwh<kas57D3Ly
zr#h~1JR=<&&KNCEN7xMI@8q8&QTjW-)t2&amhe_1{_83KPnW{unGf7J4#^@mohx>H
z1d5#=O?0@IP6jV02t_zB{;01;^v)C%lYaf(AhAR&-c_|vqL@h;kVKO+e<=0|LNHd=
z&HkSP#{ZgnFOfh=qUM9`JV=7lcn?avuHM|hg8x@)e|UBKhyMf6WNT5A)F>tdH8?Uk
zm*IpFD1YT#TW{nxmVWPF;eDDi#-&KSN$Ow_OgiqMgCv+tH#Q4Q23S><Y%5Vo6-gDh
z=hyGKkdmlk+3rqfcYuWfiy|pr4i68%b73VHG5jSLX&Pnv!sKxz<Md*?fAQ;!DAu_V
zR6a{mkrZ)EZ<&2U?q_fI$;DUgiyz=Gv8}GT*?;w{-7nw0`0TaLF7gOkG#Bq~F0{HZ
zsfbJ}FW&7g{;_&p-5$I0a;>CVB`^8;sy#MC+1*?w@v7Jk$D;1J;HK>`p;yDPYpUj!
za|?dt{P-!>F_e|?OgN`%hIUp{uhPqZzWWA-yiVdM*EtUazw66Op;y0MCVEvi*Z>Yl
z<$tTTt8S~NsPQdZRr}($<O2MtE^h{``#EO~cV)w=o37pS7g)uXw)}(!?}qUMZ#)@~
zXB#4dj`ixcY|5?}${oKRdn5!j@Y!ovw29D(G*OZoXqKA$@iG_C4o>Y3n&l6?m65P2
zS?cOv1)5AFEwrofmfJcGfu=+n30JRZc7LJK<&hez{QY`O6j%!(qd3m|h564nZ?8jr
z#E}9fvQ0#`0nS`Y8Ku$^(h*C=d4Q;#Um=6V*{Z1bZ9mw&W>@tC5cgR1cg3bIIoqyk
ze7bOUytBrm%e4Tqzk`a9NLIz+(6z<(j<Yupc3OH)m>5Wqjm!2$VCJyPrV1sl!hici
zxgC0LwcxJq%er1~i@Ms_>0{Zx-IX;{D#0PA!XNr@SHR?jP^)dxaKdpLza6@AS8a!C
z+9sTTY_?@L6tL4CIHLs0?cuXN;`g@|)ZpZB2h<MG4c9;b>6#YDkR(K3%Gnlud9}!!
z?E=gXEi6{6+xC~9tJKvmSZcqg`F}xfv0H}*!;|a!qN}hesf49;u|L#h#5u1?CP2#%
zmqx>mp-dVfCPG8P{MpFy)p3EhJr<++`}LVkAa$gra=hRsYtq&JNL*$NtT~ET%sbQt
z2n%9le1*ksT&m={Eu=b|caP(4BYyjmw>fMq?(lTsGP$a7-7=7}b>DK1?SC8cm<rW!
zcXO<{Y_N5>bEvW@?}|$iuYRlAW5;Ff4HscXK!#*#B$M3D{=4trynPpzYcd^W&9*Lj
zCv89onLP2uZeRTr$D-`MfSzRzTbFi(ig}ihV<`(cb=ljzZCw@}XKu<ZP`TuMo*U%$
z$F6H{FZGId5)?t5o&Dl>F@J!Gq#1=a?7yzv9k(TnOX?V86>iqHd5N!f^I-m*1tH(c
zkh*9JtZhGTTD#=jg*TqMZElHEpx&-Ll+BLQ+|srNKAS{D{lJB7!%wCPBsO)i{bjvr
z-}5U=CrVaBU)DD))N!&B1y&@KJ$it3q>NYhcNMM$CXrL{$y0<x@PGKk0v7O?n)4)%
zQOX%9s)mKs+I8=VN$4cW<&tHST+u+VXw0x^0rQaeX8eU&uPc~yg;Epig48AHhb4Dq
zec;?<P+Zdf0Wh;KMzMy#i1}5T)z8P>ZMiR-;fhOUj0tR5*PNgoM&FfOW`dLp^lxYG
z4uKp_#v`?wBP@!Aoqrby#)vcD-5vX@lX7~ey#-GcNHT`y2jVBIBTOF_Nf@fXtza%(
zY*CuPc^rvlLj&BG1F*kkxt5<mf8MlnHk|W8GZqtN;Q8u08joq<!CY%$C9g!V3y38X
zDfg&gz%h*Pul(ir<-JX`G)Ax4*EIZDr>A>+?fOVml!Bb1K7U@p{sf-s2>l9xEIf@k
zh4XD+4Q>Nt33UZekJ?bNu;sgQ4|NB3ib77+Ol<}z3a39&9=P2KzP=>+a>Z{TKkD{&
zEMyMNS6`Jlf~Il{wM0BdM?#J=?28cOq0|ITso8rgNk^e)2VwC|s-q;&{RuQQe!D77
zU{%7bP%Q=5wts~kLXu)wGRw=r43Ug*ZpMa3;n8EQGO<!o5}34s>=EFirtuoP91l_a
zFrX=s-V?%=8W`2b%ZoE7kK(DnZ%Yg8Wjt>MQ~)`u0RP1`pvVB0M7)jQG>P%Bh*YlK
z$b|?rrGOym#jM`XTuJ3~kRPSf>1NMKn8ehbi3E^UMt>%;t^C<mQl2Ul8I<9(YJaiS
z()oT9GLBN6ESNqpNez?|XC~?OjFZeRym~qg%oBu2I>Na30cx2jRwE()I7%5PRvw`y
zcJg=yb<2?tb|vg3+I|k8OGM<LKOu_~6o(Adrn(B``sOkL{4+p`_aBtO{yCjeq5z*|
z3#xdd$bTZKX5eQrf&nlY98nhSfiM6u<FD;4ot(6XGd`U|YXdB;KZddkf-NbFVoZY=
zL&z=DRT<z&nM6h;lL&A;cA}N%{!;YZw3S!c>Hw5L(E&#ybOv+oIiJN=KIw&v{u98?
zX);pyyTDd}B{KTVR`7;-vfrJ`4)`*T7lQg78GlhE3=K^_fi=khxi47*s4!m?+%Pk8
z$H7IBz+m%5!7~fH-4IgFkc;WKr4`z7Sz$*uN@A|al9;9N6<4|EV|1;PC`sh<EGl;|
zLmOGaXwr|FnkO*eyjh66pJExqa;I5PZi8eu!pb!DOu?-l7f0D2uxX+YPBJpev&27S
z-+$oID$>=iWji6DNHjcLaV}y9eC}~OWUE{IS~Xk~ZB~@g91F5rV48LlCV(fl<ze8e
zEV+Veg4MSxMqW}%lKxMip&VgL_9PgZ<AQ`6M?q3@(rQ6s#5kDADw16NH}P~oAe?5&
z2+$xgi(~H|nA?*g`D#}T#h*9{%0#<&Vt;1dAd9r|+j~|==B+p_&%~fY1T_m(rkm1g
zl*Oh!4h#521DT<6_%=p-Oh`*4Q6??7Ag*Fa{sS!JP~AK*Y@Rpv9ggQuxW6sh`<j*G
z3gzT)PPKykzY(jnhq4<U0;>}HIQ`va2Cxt)p2$(_z;swIO_U`XR_V$f?Roq%*nc*j
zU>gCdiA<i)HdoxL6Ho%ECena}f@(QQJOxA~lTh)%c{DJ1{%R@9RhaW2`v7V-Xens0
z$5#&xU`0j)8$6(oK?~e8QU6HF0D+GC;|Xx4o(ay>zYjPw&jM#=3C?_@5DZvPNFQ`&
zJ@I3#*#ZkGwJMw473ZNu#OJOn5PvG%^Bc(ysMnQSt9=8C*s<!}u$c*Rs>U93X{6Lb
zj^Jh@5&N*-6g}JF4GV#qB{MXDgunjsJDkoHS17okKW=vIz96|Ald$E_i|t+2lxKwp
zH7wc2iK8}#ObnPq!kfhlzS>;}6>i<OpP9C?25~Sv?=JP~VbfLiLzqF3ZGYXgWG-j?
z<nL8TPJWc<iTH?H#8VBQ0O*Eld#nN1shBo%PQt$AVmoa}RX6;+FWr!0@ji&P5rvrr
zk9uQ*StgUUmAAjOb5WlQv~5V6^lYoQ1htJCZ8v2Y+NG1O{~^%9SDu4-8~P@*fpH!V
zOiZGQER{yiHUU;3k5333>wicI^-NLp|3^<;EQ5|mMrKPlo;E1N7bfKnKv%nGTQTr0
zo^(ZB{cCh?db7w(q>?2e7So88@kk#BhnGX~tWv?2Y-!!NLM!WFrOdZOy9LlPuvt;9
zzV3=$HR32xFqPK+tiCOpTWd74Mx?JEnqpsVE!e<ye*5*eUonhh6Mv;OMYVO^lDVqC
z3Qz!D7b%$tQ9<*_U`}*}^(k8Shz0AIk30L%NyzXcBdMi-SgH_*Cct2FxiHTf(SnZ^
z>>RJVI1X(>(5zrZ(+Tta6enq5iVC__bG8<Nv4Jc{_RnVHdU5^k#jj(CLn369D4>G?
z#x^;IIQ-+E@x=}b;eRBEBp~;Ds=B`b7AI?3eew3i4;%>L8>K+e1K0{PNF$+?Hz<!3
zX_PAggUqC@bRFD~la1m7WXDIWj3WKy8S39Nqztr`s3)b&td`v-LX7o2W`G&!nYm<S
z1ara&>nA2~GRrR5RGMVL&6PgZ)aO<w`8hCU0f+2JxWBNvo_~a<3TMxxCSr?#NnXBu
zNh!1XCLrq+Y++cK*vyU?;lwuTO)xP5=uLn&nAxGSP8nF=6y*ANNS~uk{v<PHkyvhX
zG*YNouD%aZ2Wn0r<vDyFk{KW}CxhAN!NmtP-&J@MJg%kEo(3L@(un2Q3!%@%I%{MC
zy5{U8p@Op8bAJh)K#RKN!NRH>WJO~9Bc`wZ6|x;-D2!4cyBQW}Gy(d|Dp9yh4)FTn
zP;D7D;p?HTADVVwO*y-ATRafxhnF*OIMFcDiRnXdSdo3lhr`$+1=i3&kbD8r3C9Sq
zy8%oY%k)v{sq$yAm3~TmnolP_pR}Io=je<2Ir<`=Wq<0_f}A=lNjmxFAvtMu{66Am
zGTEfo&p-dzR}GZF&@bN(kBWiJtcRSFb+K_aX^%EWz~;+8Mwwn}GfHG0bzC?>X4NHl
zq9F!d7HHnU_rea2_a-D7aszT?hp#%~6x!R$V#7287M9BhzYa}3NIp`84g0C-NsMin
zbY!3$_<!9J2$Sasg#J`BX^)>!gNnFgXG*|y;Xo?CT5CPJU$aPMG87qyb|H+tGOHeg
zsbKC`{7RY`JSXJK?>6QOBNO&}VlGG@x?&5`)0U7U14QNau5DdB<cUFvqwuUeJtH->
zfDgp;rKUbFHPLO9p6V8PIv)w-6`jrASw*3+et&7m3j4)Cv`>2H6Nm4I46rW`1x5qZ
zHf&?m<b_%IXMVh^>L4yak6iXlkZF7UCV1kYM4DCN0L=ts3mq-7Slu3Btj$oeNh#s2
z1Nn2q;!tu~(T5<E(vhCl`giN^=`;0r{gnQ0p3>j5|2_RZor9_BChRo_fxh%^O*>~X
zy?-nj5Ay#Ax$5WYV;mfij89{Wmuf&7k9gvnz%X*yKlwC_Im-59j0Gsh(u+HB0gI5>
zRHUm*IfbC?_?_a^F@eFjQmp#D4Y*<hTvH8YkRc(R=cmGFA(rKISldWibazMD()M7_
zrn<VdL7&R7lU#Y?Q<-zfEXyOK&GQ3uEPpb?7XNy{pup;VpjCi-ajYI~H9?LV_sOAa
zaU#D}JAx(9da)aw2`r-k6vRt|QiYMa7y#$wsnj51hoY)`ezB{hFzgK4f3T5oMgNPf
z(3Q1~jlz5zZ+BwQ4uzgM^Y$`=alH26oJ?dYjxnmV9;!oK@*5qDY+cWt{Rxw3DSzVB
zaCfxD4^>(3Y`SQCXQYTS2Q$n4q3RH<W8jeTrR3;oN#^XBgHX#r-e_g<(Iy~{KsJ(}
zc~&jtUCscJkRZ)6F%zVq-Z%wJ$DRz6=}E|iU^PIpyv;JS+vT}@6yoK`6lk|j<^k?V
z01Cx_&oSzpb8pM!82+|Kb5E9<z<=T*FQnwaghV!TSwgOJSxOh2C(*NfaCzvxj5wE%
zQOf1$6^9LE=Eax=Sd0)4X!_S*es|@fxG9X%`a1A=?2pc-ZB{z~=v;Akw9$i*J9I_W
zSO{h;1T!ce+C$sBK=C#-fUZ0oJ6j2lU+z8jE;+#4*cqg&=BDgQz6e0C{C{k_9U(${
zh~sk+;TiPw6fWp^c12+k!gmr^yo_`{>qM4tLI~$X7S>>MP3~ZhIdy*Izj;o6GT^##
zJtM>urvyj-6!iiOm>P8Nz`~Vu?Fn8t<#1opt)f)3!v%A9Gmo&XNd7VW-WwF=*po7R
zS7jdWC^JEl^MRArlQgPCf`9!cnBXvfOi@@dCi-e$p-+lF_>d^wp09>f_R`6wr2qoD
zhth}cknlS4^&HOO>k8d3{Kgwo;El1ZjkJex+?5=&9CX}O^@FdzPDBI+aBx>Vu!Jr;
zq!jfDtpZP5ZMonFbjZACazyx@9V9l{0#b-L{l;}euDsxoCjVZ*YJY-GMD_SJAvwj4
zEuyvI9J5>8o^Ip}!@F{OpLZz2+yzq4;>$6&=#QYkS^$DZY#ma#K&Ihg!*Mhy*f`|e
zlXQW*J5YgdT9I(1%T18gfSfIRD@dENy1m=jyCKW#b0;yob4{9Jcu|v%IQ`c#B)qbz
z*^srp`=Jw%c@$?1QhyCY)301Gus{6ZQc{=)hQHYUb9XleQuBK`@Oo@m9L|GgV5rCa
z!Nn`*ZW+)(0He9X(b$@fLV+<%^a2o-+B+6D-Scxklo2hurGi9XHoZ-@_R}W=Q*<P<
z4+v<AxY;+2^A*DP%a~ku)J@x-&M!SWcln}&QSNG!o*5PF*?$KRU{@R%-kAX78)E$6
zV#02qEqNhjgz*l*{CJgaM4chMCW}Oxe2j<YH1coe4+)vKL4yJ<bnZ1dXXDNmf!rPj
zJB4f?0>qZ)RWw@9aG3SufZ9cY*l0IU2Ay1~9ddH&ds6~nA&^^UpuK_)_=xXKxFEWj
z^Ub%yy0%h~TYqW#Om+LylWCzGFUj<C0G8t&6r)ke6?uMsTx*=5pl~>upXVqA^BdVx
z1lrM{Zstn@B(tf(n&CAVx~cNW{Nt3Y-X=Yd*V8D4r9h=cSt_9v4s9j{bJ8DUug_2H
z^`Cw-y$|kySy;vyQ;R^eZH9ftDx0PzLoT{B()uYE9e*%L_EB-f3)f4kd`r>jAV_lK
zuEOcyA&wPB6V1YeiMp`)lp9N&XRM5zp0u$hYa$Jf$KD?|pmC3pOCU=`PTF(g*p)PM
zk>-5rrS4?LgjJyMqAEwGay}iMUd1{7Nb|GT4WLgLZu-ZV@3G@AANV;#gk5p#jOObk
z&4H#1TYm~0V7M-aNKFO{JUvH)1r}9}J9{gevo9L7nsBJv6oe(+7~o%9?Yt9P!YSMT
z;<Ae25sQ92gN5`m!JIrkV?vMP$f)1HJ7CLZ+9UFas&We(b9bXXk3D#m<38Df?mJOF
z!JEo?;F>3f4LuICG{&VoX;ahQb+R(LbsDR^#ea4j>Z<J6jK>k3Fu{f8>iXA7nOh7(
z=8<^(9$x}_wH~Y7&OBy4-@k=Lw-<ORzz}i@eoU&ROhT$9-|e-U3f382Q3fWyAnLJM
zD!vpYo34ID*A^M!GQ1l=SC=ur=g{W^myme}v)K&@IV?e}{f@D*r2CY+YC@1RYus<u
z?tf@ig4C)i&d&uIzN%uJ(+`W-W~oVZ;X8KhZtD3k0XPD$>V35R?h~S%g)lbwvlBCv
z!?%!}mZ8n;db*WJ=rmub0RHtdU|uVNUNSI;V+&Z(L*}zZ4~3RwQ6kj;Ev88PC>1=@
zp0j-F4_>|x#quH&N%jPmmw;drm>$N^=6{dI(V<5uS;WzK;Z?3Ffp-#Zmrc-_m#2P2
z(sgj(0L04iI#lY&@F=_n4N%Ar!`Gm9%sED7h>TOWGMi9u358y-XVuO?ag=n3#m(bb
zJbw-rHwd&dryV_p#m&E~V*Alp9EOGmyY|E7R2Yk+i-3Qa0>ee<#|p^qZY_qnwSTMU
zB)<1927u9laUz^@?G*{kb|C=P8F6wA2Jh>Q(>TNLSzMh{cbE;dtz+9xW81c!G<F)>
zPTttIZQHilG<IX#>G{sy=jvR{G1hOGcvi$Vb{Tf^nHM+O<c@YrTmhMn@p`jCqs;kz
zboUEKbC-5Dd@$P;ByuN8@DUST0QsC+oEQdUg(e@o@9>MJ1M#2w*LL`-yD}Wp1OY5*
z+pi$J0HLpu#2Dw)u3bPWV|N+h`w}dHQ)8)l1`rUVeQq%isGQwOsCv58gtYw$$DeL=
z7}b1p<+K_z$cR__{?%5^d9>UQ>+*HJQqAf5K^1x6dEKJH+HL?2vbSu>@=-?j{LF=#
z^F~0ozN<YwF2{nAUqbLT`DO(Wgz-H<9oNSEfvrWThwF*Dod>#&1-hU@n$K&gaM^Yd
zKGk2PiZ8<wX&i8TS*gXCq&2T`6;&B|SXEt)p@=XZJ+2ztV)rSpPs2@?9%~*-LXl2a
zO_lZ66E;YDM&mjr9#FLljkq%q`l(-AnNXK8a(jbKm}u9#_<i`r|3QN<Tm-?H*_*n!
zI-42W{ol(WJ%JDp0+b_tKMs|-`9WCW9F+ZkEB?&ftp96}U+G;W|7~;nSJ!XSZ}w-W
z5Ef-mj4_=oS1C>}=^%yih(G6N2iq2*B!RQMvupeN#WGXE59nqe9}QI*b!(ew5SSpN
z;>CDFuH$Ks{I7Gm>|G3%#Pi4F?crDsuFSb&?n(QdD~E44HlnI$_w15&bDWE^eG{HI
zv+cAO1-EfT;H8~+Cr7I0Z+4wsUWeD^U+;>JhI;$dZT@HVBm0!YBDua!hpBDk%}v1P
zB|Q7py!xP`0aEFx?*l~NwU5XBqiWwo07?B};7E+ev9NNgul8AwhvUer!njJW6xKIj
ztc~EJE|rR?iwGW4VHj|ms~2qlsB*|Vf4+0Cw(D63(tqKqNR{Y;<*Jw5r||td<BZ=I
z(8UKgeht3Rs8aGJl=S373QPTOzQ6z5quhQVQtjaj(CBXLTXBCNLlPw$`Z<4P-=mJ7
z@kuyqbWygKc>YzmL%0_$-x#QwjUyQVd5%W}BZMai_NU>@y9oW!rptIwb^W??xYBys
z-1k5O@UB;$!LUCnSpA@F(YlG$Zc%3&O@vt+(Ci(oA5b5^7h7c~^gs9cAvf~Xm-b!W
zM0~N`N#paG#Aoko_*NW`Ifo1F{gLj1`<Ez(uMNCkH-1+g3o|rq?)O=He=6g5nldkI
z149WsD{j*O|KYxi{{+U4*d}=#TK2ZB726~M@Eyu`en1%;ML*g23v0Bok0ntR)+4Gk
z`Od@4RL!sX$@>e+Jn<;I37S$ikRh)1qiYRcuY4elOV=}eHcf3}KUE^Gc=5EcQ$;vp
z1>@7}oSK0?!S%XyTk8)#Jf~=u70}?&p^3)SCj}m9hfyBJCjfD(7|h!wgbEm_N{%N2
zljy4#^NnsU^F4|(DHh9M{&a+@H-}Nw3TKD%##_=l&;}L(Ki_7HcBdVDVd|)rA58wh
zW9h`)P1Y0_BdpF|K-R0h5@)z$nK46p9o2)ouqJ~3ak3#LBj3qck{<t)zsv;gB3t}Z
ze-4&g=c2M-h#?gWeE#<m=dB6;kt|ahpbY6XRTA3o|KW`T&i0d1HtLbX_DPez$<Hcr
zrJUhDm1Z*?i`#`z5Mq^)GxJ3yDiwuZ-!usWUIoo!y7k3ExHR5Sj;`r=y|qPAPH3yj
zQ|-?oiFh%;>c%%R!x^yJ=Vvw5R6?t(h|8W6SAz<EBv?|9QGo+K&E@e?9w0gih$t^j
z{p~b>1HGaF1#f3K1Nq89QHFb^2%DCg26?g`lL->_9pwh%Qu~Hb3g!<Yw|60aN@3OR
zyv7_)q%rkQ1wI2T(BHhwCj@Oh?d9rddiKS>jVvw?L#$es1>*@EP@^#j*W7gGv|36S
zpw$>Ukt&S8{iV{%XbW6wd=lP(0n5D5XpgZfA?JrBb4f9Q0USIHgD-v^*vG=wK8>o)
zFRyTS-Z}WKAmpdUGOMAOtas=XXInRiR2rl4jnZ0OXa;@EGhbD!e!t1RsD-Zf?0%^e
z2>6ev(2GvoPGgVLb3$+RKN0^zdh06GAh8Tp^_s3suCQ(4FG_~wGmmY6r{?!Kde*zg
zI`6&A#$eb`2Mojh3&&2U=a1$sFhD&5j*<oM2EHhcISF*t0wmlt;Am}$HWA-KUU5ek
zGEbI6d@Einj;O+&@SRF4j~j_@#0k{!=-2Cf00JlRpr7@iABjMgR3;p2{^FVWQU&r!
zSuK=x>)-JM{nt9<%fV*g7p*XSB5wCjhHhisY)P4_G-9o%cG_M3BAd)ig_i+M15#>!
z4ZD?evd%JXFbAG?VcNVYd;WAp4<Dey+v@;wh4|!$pFiXCw1KIRuKdl$;9P*&)|`ea
zoVXTBaeWy2E%fz&70{WGI`W1B$C|Oe9b;TBk|?Gj&RZ<6B}WOM{G|$j_djjEW*CNk
zs2c~}?M#mk?@CW6p&#fbizn-qB(OB-VCHSj(+Z7>P^b1w(Nc%?+<dFX<%4Am(c!gm
z<9p6cu3;`rVzQI)=hqA<^;og`B(Q<v?Db{yT#`v+P${e;oD+yu*8&yq2pOSM&5vyq
z7#UkLn1(rGcI^f>o|L$5LNiyFM2w(KeHV=UTI)5LQQPwO>Ac0yjBxj=N$tN(_AVN&
z6Ae9s6Z?uUURJK*@U2n@UrvbUr9A>g1`Zy=U}8Xg)5JAupT`Tu<D7`YQ*e>wgVL(w
z9xIWbjh~VZg14xT7C)j(;UW(YD=R&o)=;EQZu|Q9gouC!RS0YlzBhPbM5%<N@pr0?
z3RIzwcsrCTFgxBn-3$dW8xN0e^%0E;YOV9?vbqw4Y6=fJMiA`Wcq$#q&cKnot`a1n
zQO-dn@MI&jBfKX(|3C3hME&%d`%GbR(@bGu);JerzGIH#$Z)ig?Qx}#oybMuzgAhM
zpf=%4ch3PcPAnAEH)1TT%LF<0Ke+x8V+WRAQ6Y3fMvE%w_!dFicML&T`;uz2Ojpz6
zQ8&*kNSyKfUH?$KQyAn@_%WyHA|`{30z(DTz@m9+-GiKqRJ|Xhe$se+nw4%*D>Z;o
zpdwk95M8E91i0&jXdl*`ooi4jau%=&!xU`oq_P5tmZ`lMHet(+k^LQs<3tYo!8RHq
z$^vdw;^sD7=QX<N_uv?YWzq5J#L^McDds_I#RmS9*+la7O@-Ptptp^}M!eE=DNpL%
z<yN9>D5!0YXBxOVJD-X5BT%wIaGUpyR{=I2BnPOBb!rs4;Be#us-wl4&K|%uk75Gp
z?=Jv^QnowUpL&rGA=wDhbjAK0hHMBgsboR;cJE+P@du1Sc1W9`ff?Vg#dQ-@&3^lF
zm#c>fa^+h_6_C0c6@K?GWLEl7UegXSM-Q+Y!izxia?(QuY;U$apEj_@J7rMRh`672
zd|Y;?6UY^*9n=Y0BU=ll^~Av#r^450YGpumd7&{Hr(SRRaUZWb^(uZP@)Nj}K?5uN
zJ$LtWW3yK_j$9xyW8(O)rP!Y@9fYH%q33nB=N|^osXmSsV-dvOY?DF44OXiG(nYb_
zPvUG*2!-nhw0VDp?X+T-Vm(9xXZlan>t9UFjW*9Nd6yWGd5OA@dpz_MPzm;=jMjnV
z!;P|@dg*9UZPU>nvd0#(_bLM37!<)tEi&eMaagZn%M-iWGd~m#v4>bKxc~uP>1JPs
z6!Oz#8?30#OR0)q6)<1$2fr*Wx;j{(y&^1-Yd_+20~CU@60oBSX;`NUQYOR;t&UM_
z&5;J()3CKp>H6CFQYz}I`o1fZd`5tB#|vd4qPojgKk0w^YaW_Gi8$hHaK1Hp?yV|+
zkCs-}`L{?c=~O(5(A^7f1}NuMdl%e0qyLfQCh@<_tnkBUS9|q+=kVG4-CQGK3K=;2
zal7Z@6TOPS3IF~9Sx#N1V2*4wU-$1zO;{U~m$lMwPI;D$58S|1U~_~7L0T3lARniA
zL*nW19nJ)K1YwQy45JOBDYuy@^!wzGC=;htbZDcDJ0O4KA(>G^^4AY1HjE(&0U4FR
zIN%5%g_NM+^rjbrSa2*Ry)9v}$GP|oE);3kRN)BmwK+hCN;8;Xk5Nxny3T>afG|Pz
zQ;M=Xs>R`udLCY!jVwL-4nr7l#<fmQKLe1!QdNu*oJCMKS6K%+Rj$#on#CNr^~XWe
zsoj#O@~7y|4iqNnzDLdnF3?CTSI_V?kLv$cfha<-P4yZdyvn9bH=DJ>Hi7!`+!a0*
z>_;L~xZX<$f0ExS=P}}-PG&<f--aCW{yw=G!0LsU*Vvq(e0FY&u7?6TQ_f_Lt>h&m
zgYRHHUGS88oa};6pPsq?%PuWpvB>zLAYFNfM_LPrC=8MI2KjSG<ah%m1vSXdCC`U`
zta`(bHfXT9fB)M&N!_$Syw8A{PIn9j-8}3&pnG=^$bdIxGy(4_reuUL);N9lcscUh
zPPmCgtT4Zm<+KjastN$G=lDM{q8LXL=cu@z`b_k)%MO2R;b27Q?5~@0pexc&h!+tA
z;4@vhRW#!V==Y^zvCK70+%jqX-bR&I#}yfXvvePmAF$iQLEvUhW{SjAEI<6sQ_=Oi
zQ!9qf7OZkX`lgL^m^$sZz>ExdasoKTe>^4@9UbGX^Teb}7=s6(H-E@pk%A%}GjBR{
zmZF`o%UsFcX=q1Y`rULu`)ov~KRa!1H_*x+J8?~aJwzhdKpY_Rzr(&{%bBzy!B0VC
zYk+FD#WX_aQfBjvUcEbm_$J;*)nZs*wR)j2k@zjD?Y7a>3A6f(YoYdcAbNXL(r*-o
zVGYT)y5pus2bdgy#vbK5CQdgYPtT+Q)IPP2NkUNlax1072plgv!~q70n$S9iYe1K<
zsy_!8B8$2?3%w8)!hism$+a@<l1AhDI>aQxfkyMJkbKXz1|L)mhEe(nGe@0F3fc<7
zfEI0Oz=+InT`+vek$ja?25!wfc#8sm8|6UjNhoS~o);V-=Gkp7-hav&L-)nQ?zCs7
z!#a6b-}J~Q=^x`dr&N(Jby%XOYV7Cpe@>rg_O?5F)+@q?#liSFZF{K$JUyxGe%5_l
z67S8&b{>he?^GBXm_AHJt1WCxB81kTkF&iMQ?8%RzrBC-ZLP*wXN8FJs5fJCNAzl;
z#|;^OR1yHlOX**>zc@eMK_~BS5ilm}r{PY*+`ok9{OJd&I?yKqM$MB#GaL9G^S`ug
z8Qu=er<Z4bh0j1OhW)nvKJ4Ss1cuGPj5;O^AwrXw#Ju=9?kB`{2|L;2m!8Z%I&i2C
z?$ZUhUqw9kE3GN9wS$wnV<_zGs7Yg*51D@5I2i(mbH*_13X<>dgOoOuJoujbFa7_H
zJy45b<gJ6y(_rTWp(p~n|MsT&zKLlrgNfeR@rAoIelwhBO!V?h7HOHmH)g`9hC(W-
zi*Mb>r4Xj|j2NEDaf75-!ka{&C{y7CA(NF8=}`tas0>I#KVR~!rLnOMwx)FV;|Np6
zZyf_<FA~>h4~CA0Ewg<Vmx2-ahY49Gk-muLrerYUHyEC(d~HKP;mnz^#<Bz^Rha_&
zLp(6m;i3@6>`8%2JNq>|sF?be4z~G?ay5Q#*TyEjTUE@_z6mEXb&jOn@VHm?f4NY~
zeVxF3%nbn;BW8_c!$;K~;rT!tMb3e7gfYNokjVh1XIe7=`-ta!|F?r!T*d=g)^W@k
z5(F#0)@$%)yOwb`#_KOEfAYgyA|024rcD(hS7TC}L4T!m_bs-;PjcuF4j&rr-i+oX
z+003M*F@F9gyBZZmcFSxQ3d6^g}<uT0<E__@Nu>blm(|h3?+Qa4hi$mq_LME9q55l
z_~!zK{H9P?^+d+C13B!PWLH?0l}i#(O?v(8Nt?IfO^0C`GhPF8w*ZM#l!Fpw0@A`m
ziqO4YQz~o(fkX8WFpvLeQyGaa^wAoFr2o<H;z~lrH8$O9g0Jw)RUvj{p9sS5F$`%B
z9nxjNvMeYP&#M=C_OiIuh;#AB26+HH6{K#M(*S71N0&vPWEydo7d`I+OTx52K9p+B
z`UC2FX(PBgmTN_EBkrQjY#iaz@=)Y(hp*}U4y<!UESUr%LqW}uw`DbWoI<xhss9GM
zS-=~iZ-UvU3`|>08y$qsgKuR~{v92;CP#f_@R05X7sUW2mlpeHtJqbrxP1l?$IRVn
z(_2DR@qL?J|8$`xXK6UmwugtHer)|}ozi@IUbT5xz*-X#d5bhC-wj!m&)!f{K=wZK
z(@myAYSM{{h`BF1jOJe(J;!XoLC1S|p3UZR{f_<hQJb>s|8~{d!-9X-a=Xf!>rC0u
zabpRo7FF&H>fRkqoWb9u?Tres9II(IsL<`o>P7@X`Z07YHhyXUVFQo%lQ3|!ANt#>
zQf64h$bBu*S5)_rq%qZu*{0Yy8l!lDitW=Kf<bM-GNl>=2NyaRJvB?q6`3o_Y)1X=
z>ScWmm#W6^nDQyF7vr@+$EaBE?9?<d1j)OcBT~BMBy<1DF9k3ZqCf+4Y11ePa<pJx
zR}ZXt7#Id^L^)+}37*~*){N=KQIdS`f9kgXL%AF|l014dQY;+^uGGs#g%tLm{^y{N
z*Jt#w86)Gk$&SofWUk+%D$8fz`Drhn43Nbi_$8y~cEdE&P_v{caFHQJp|YXli|#af
zC??zNGyjv_R3*aH(HjO>+pS%9_QllW5b%DK=TEtHt=QWBJ5%OT53sqtjg>VlTvath
z<`oNf{LFFs)I&)xD<dBxE!er~Ab!{PtYscg`$Nm75*DuSo0X%~QB`rvC(9vCfyw2j
z!M2u1$bHwUr9s1Ueo!NFMULb&rM>vXtWn2E7iG?vh$Gd7%v%LiW5$X;*F?D{>P3iL
zM2dX$B+;1Dc=IAdrz6lAdQm80ZV->$C40I-Jt*7s$Ia~jM_GeRk799#0CovwC>90I
z=ra@V#knPZbjzc{5TbgVEMRs$=G;U;*GM@I(Apo7Pn~h|aVGMPw$b>pAK>3?!H^_T
z2Z41l?+xHW@yY;Y2*F3B$BB_E-l)l@1|uW<#nZwu^SoJN;IO4@M$VIY`DG(XBc+wX
zgU}HLaDOgRyg@d&47>^t*#3Ga+GbPH6EHg^gEG+iprjRyE!=%%J?uDN*&bk`rYOjG
z2wY@axt>zVIw_Kxzjly)zfxK_JE%jM9FuxkJclQ-?_&Tz=<#@KSj{M$YDg8ZD&@2Q
z!@GGR9Leit@gbJAporhw>UxQ&Rl+)mCh8DrM)5w1W+KWnd^RMx)RqbKJm?O->P)Gw
zw_jld6#ITP1<)>?NZJIY&t|dqWE@gsS#=-Adm_+l-4E-&wwdlE8fBR&A66ukZPT+G
zCaE(-#oPcIM5w$ioo$xVyfEACfd8JmqE%4dYvn|EQuE)rdIT~mAykuzFe0-j1Br5C
z7ORJ>q>mym2nA-MIMBnP4C?rNA;-cb>CMXNK0Nn<kdn<L&#;t+AsO)jP6g=*<t){^
zgD%{J<zzM1BX7yiu=-oX+~hOJ(^0)}ZW@&z%zuF`+!g9SQMEY)@j}uf7fICSI`r~q
zbz1wo>9@pYf#UnRif=_0LH+%Udv9C_lBF&|OcQ*9qXSn2pHn#*zj|v_<9tbwTQTxR
z$^zK?5wL93#ivf)Y<oK*YN=c^zDo1$Z@ojebPygSGxcHX^ABH7lma6+sTm*T(9;j|
z4e9~9b4w?aRQ7VGtYy8$9M~@~S<I-#@jbKI?n91d!+mw!G91rdt-E3jXch=tDVmJ6
zfGo7~5D{w=eA^s1*&^d^^NUcgI5#7js~#31i0x=sD1wuO4l|J7$33;Q_E|L!H34xB
z&@qzS&0djlYvi=Y!S=`^=Yv&@SZ=aLI$uCsu%adn_r>zddQcy#(k<VlZ;wmHk1--m
z-!FrJHrN4#J(M3(7I<y@$dlOn-%P`A1XgJui{GIl%)_3AyI7NVW<NklQzHo8Ulu04
zw64etgD>kgn<KfGC*QEyLN0m5Ed9m=p8$)a^!@pHHo;#^;g|C>2e11$(5_VPstJJG
zG1`sKSTrY`pf2#AU7CB{&d9V8gmC(9jpPQI@gx4RMt&1S-GUP3@z9R>uy`VE97N=O
zB9rial0@;W&1FELMtd&p7P*<$ZeV)8F2@}6uw8lCLD(t|s>Qc%P~y)qz8-g1#+Wx>
zaDGCD()Ur;K%r&*uJ0?TIkN;`syM*cD^@+VF|juf^R_#~vijR|Q90^ei&bHeAfraj
zy2O+$CST4sGUlYfN<QzD)g3V{J^S%6rP3%V$d;-uLrW3dunY+^^ybbVU0!FJ?I~;5
z$}3fIHgBa87>&R5EDrH!7V|efx!1=QLMr-3{^Xvh7hx#Jl?WW7xsJ?NVFdhwX`-mn
zlhgf0K_KE+nj?4I2_7Kx#6Hw&jxv^S(=A+RT*%o3k)Sz<OJ~-SM=o5b&4Rk;`jARz
zVy5#l+ws!w5Zt%Xf;hhhK#5NDB?%@CH5pTC4?V9KP!rgmz}g|>n{wi}(yyw)=Br%-
zNi&=-9W%ht?=OlkcJku)4S~1f-(U$~Ij^A^+AGrzPmmb|zt3wzcOb1&=7p$-UO^H3
z4HM<V_Sf}z0(=BCJ|b@XS7TGbN2i=X!4&&_28BsG8s4Tl_8;ud+%^^!;@vYqwWkJ6
znLoPxOG-p4yM9G4722Fg_nBsXVf%ax@cw`KfC+>(92?93!~@}vBB1{X>_373Cy4(9
z`JbTv6LkBd2#m}(80-IFf&QcBUbt+wrTxq6AKKN(i|rWsuS#!UzT9O+m3u6BK65?S
zwv8)W#IPe>BC5LT^zU5&IRJuYB5iX=#vBIh2NCqi=N!1y_oJ-d=&h0-#XGH<cFNd+
zbuaMrP=B`qbq;hy71Phx5ml^y(?)XeduX=Te}A86BiDZ)c9VBb74rkBXDi`+cw$-7
zR-T)BUp=qh9=k2-{o962W@t<!PxZw90ZKD}lclRL#!8R{ZRM_$W$WW7ingd%Pp)QX
z70_{HDy2_0p1PPWni$^8M)T@-{5_wnn$#+b__-c(X)C&X%SFY<VDhf@{kWeUJk@z$
zG%3hXPwdt6JHuysh^zpNiy609Qs$7j7WTT~VnUHY53e$W_#cOlvJ#cj9n1aD<uajt
zI^Ml~9R^YCx;^c=Y7XLs#oU^ja-ig`y+C9s0z!2$`CMI(zx(g*3p3Jf_mXENbVdF|
zd!hc3MzHINnFBfhPl60_S?c3MDppu!lE|2ZWQyoLZ;lw(t{s3aIEp)H^doo*7%iX`
zE>9K#5LGCrRO7uaYrv$d2=3G`(S;&Sc_TL415w%NXC*%!d)cQpQv&wUUko-yoE^et
z#Z~_%Gx(!t&_z10-P3HJqsWaV#?6{ew1Z+ykkNz+{T*(zoklrYnwSVI7ZpvAvdYRy
z2K`Dw_(Y>}f&t)3lPO%N7fuR5pil$RA$a}_J;{IvkJQ~w51<7^MM?w68S;A$l_q=1
z^x)?M$856#|GY2f3yp{>wdUIMl*&eHeU`_D_wjHjSJ-QzV)7JdcOW`I3V4*DWc#YC
z`)J))9Ly}O8WKGNR(HYV<UrufjI5G>$rM41jD@cnngE2*uZjT^NL`nV+=LB_ra9b5
z`N;(C)N=>pt`ZjOs{fGMa-!70E{IkfFt}6bm(~ZsZ=%8RA^nEu-CSxuntHpj?uxWt
z#!beb^bH?N8Hl3Xa6?~3OKFp!7KxE}NS?IkaN<6E+R*3?VF0lyaUCjV*oe06j)>q9
z6}$n04nSJw`V|1hV-mSe844dbe4xqAlO9Avgg+CRQ-uvbmf=6EYeF|pbG+|Fjq)1E
zfJmU3KN8Ov{mqI30vqu$EGc1RS78WQqafK>+%Rm#mdTnPlI28CH6etG{euzXw))c9
ziU&dya^rd)>a3G}CPGlCI9=I)j8EFj6z;CX5r~9{VEj8e#9?NGYjol19l-X<LNY6m
zi6n`NKD&+iF<f`4r5vxXzUy-L{6XA*gtqi?RsL&HQ-M&K28;k2L%+MJz7LM}g<o9|
z1#(dM1O#dXLncMW!dL(qs(p_kBtaM~eo%Xg+R1;GmV6(K9RwtB^$Gyn`ZZ9IFe7kq
z1Z=S-ll;O_m3(#*C07hRB_)^pZ}(^>o%XL%I!!Ok^c;*_?%twM{u`7!+ggxWmfTVY
zc_t5YA=_@*Rpc&37={gdo}D`H9gS|yeS>3_RhvIWzhB(81Iyg$pJgkpoV`}cK}ZLv
zOzKe@-Qxq+s?s>`-|-V88AQsLtJz(!LqPh6Ise`b;m3^&_wPM53+$|G?>bD%<x&i&
z+jkb%T)1qA_%;3`u1I=l5M}INkBB%W)$R>;&$DGMSW5kN*@s_$*7>`uJg`&|AJ_PX
zPZ!d%tz#|m_<3Qesg?5v{}TTt@#<{kH=B=a+=AXlPo*Rizd%`#Pw`%uQ);wLw+CiR
z9~Gpn<Uu}-b+1J4$Uq#8C39C}9((@W9pwK&qSf6h-q*GlC7~U-WC=k(mX1k^H7yh#
z1MUySjMn#jDA$GW6wpnn93Ste*U!3b68APJ`G$GvF3oxw!^adtZGWY@X>NqoBZs49
z35K8MfED3<)i@(%P@|6`d}u&Wjsfe4+7tQ^?|&ORW~k)ypC>VVgwJwF$_*Ol69A(w
z?m3I)QIcYJQ5v}N5Gb{~fhN0!1mXCN@Q&ReT6(N8E^<+L=z<BE;)^Y~&3KWA0n;J2
zzPwn+o{)rx0<4B`r3oP%ECnXGXvGt>W?@XremQz!w?s&4#(UzWfL@n506=L*F{%+Q
zAf@IchQWps=8cPCo%i~3IuBlGrheMr2=MrqSdb=i(#OE5uKhB%&Sa-m(AmskEJ1D}
zp>ri3Q8t0dT9hA7o~hJY;)NcE?Va(6w&L+8{|emHf{D-FqAx|&`-O66YVeK4iBsRx
zUqcFM3JN6;_JU^mYyd3oClGx$K#DZja`h(vlj)0$`OEjzR<?(~6QuC*MwI%CTr?LI
zr4&Sc_68=pHtK2ID&Vq?{{5E^8XRBerYtM7hmjmp*JhT%<RI`Pn5bt;>(AUdKXw_o
zyN+)1(_mer_`~6YvF1+>&5jfFFsQZfy{?JrVvt2<nDB4%2S>0U2mmp|pzUg;^PuzI
z8Mi^tHgQ#(PIoGR=;L!)#mKzOim5`&7DZmc?xAAe($r+?w7Y<!TogOlysJT&-i1d0
zVAE)68(m#KJo0SUxxZA`wfn4R8)huaxW&PD(c#1oS+(ID2_Jf;1kpkl{b%vtGU^t8
z%Z;cd%CdRP!;wGj9U#|&a%WSTG`f4W`U!<c=?dnR^7i5+BY$k6^0e7~SjX6Hj}d;>
zY9(4W_IOFz))i@z8hcKf9v!B6-@xF_n7Ybp;Cn4&wdW)9%x8>Sln#;01Mb?z8iu~0
z4HuVorRd<%V$NN07)SSRr{Xf{#Ou=e^}b+W!(B1!6nhfR0aTCgc^UooUf*Kar%#u#
zS;7Cn`3np;-8tj!bl2x`Q{wVcX5!pr+#<^7F0g>M#H9Oq%I{)y7a45s<!FDpJs}^c
z?J+-N&gJSNO-Yg~<?2FNyJno1trK5X+FmvW<O!%~k|R(eqQYAbk9g0<ZXO51<MV~_
znb6l;rc;Yi0%XSb$P+ppU%{d3HZc{R#)l+`=tudC8E5o~eDS@VWsY4K%-3t`)o^-G
z;|~!WJSsD;TGpwgs_fiIX+ntMRe|mixLVx~Tc|SR;#Pd%hl|Y~Wa@4}&DgvT1`cUo
z_70iA;ftvlZ!>;SP(MlCS=+3Wh3={Xq<jm_+f`f-ph&F@GaSb*FPI!{7dSLMGuE;R
z`fK$lC84_$suF(Q*bRD!88%d55j&44;LK!%D~Xgw1d4u*BtskX_D#8Mh3LnKRp4fm
z5D}OZw$flef|hs*w;_hUR7x?cs^X}5Bbby=H`7MEo|4T7*o@R5{()T0fc^sd3ha|*
zw%D>Z5SW>0H)9Su6715|FWFMtoww_SMZGr7O272#j}P<5hEYH{)~R^o`O9flGh81x
z>)+1{G6(Cy@Oq9M5x4M5FU_`)>9ylm&}nhvb7+f|R6lH|ukNgT80-O@>4FN4i$kQ1
zeVdJdHvEMK`jJfKiE(OXg`;K~F<#YQNFezn0D|rlA-2}RMpcErg3^XGPnETj1QWwj
zB>hxV=9hd#^&z>Xpz)XE0AYXfr}7!-GIj;T{Wek-l_K_NmB^csu{Sy|XUo<`-L8N?
zYE;zGUO7_Nt9$IQZM+{fC45>cuAT^gusBb!5``l{M|RoQn$vXSP6|Uib7sNyLfS_a
zK-Zb3g--jTu76d%OTRx*DDyMIo^(FU=%k{mVP`<^qUlNm<9(c|gX-|?`i@t@s50<J
z%OBJ44(X@Of8HMWaD{ZZrH8dIHbh(YcovebWpC16>&^4Rh_r0m!{YTvMnvG}&Xo+^
z;1ziOBxV^(Zj;!0kTGvra(^&{`uSk}fkwQmlYf%nReQydZG6oNhjz)&B<fR*^Z;>Q
z=x%ClD|QX3>x8du?#ohYca=G^nKD4SB-<H^3kOQwP!>0Wt;3{&s0vlxHI8Qq6*P}X
zpNUtyYtr6w*|zCeEQWFIw94tj2ujK@e<2jD7Wvai^b%Fh@Bv>r9e$Qj=&G+0K$Jj*
z4?ncltlh`y$a2Fok!Yop_EB-`g&o8V_|sQ;r<+N)#ed$XrwavQX|--VmAW#%aBPKH
zV|XL{@^b>=VhAsAp>l1TOXMU#l^NA3HI3-#NZQ<lrt01mtOxV6f&3U7g$aWjO`+7d
z*Ov&?Jcr>>$|$0!G*%844x5D-Q0Pv@4!=c^MUOpw6qpd#EHsPDuk#p3>cGuZgqvT?
zKW?-V@;89a@kBS<h?!KysKpi=ai^tdBt8=+d*fg#ZHV@yPPDU)ZS^!OIZaz^_nb79
zaXIgYXa`67NpG$3qmZr{W58Pew!3no!r)NC$0(6BDwZb}kOxO=pnkvr)SYq2Un;<S
zJcDR6Gq=Vcj_D8Q3uPpnSj9zmxGd)(T@gFde&H>mXHOV(!48x*XD>W(Q<?nO!J9P-
zW>ByM4{kKU7K7zqsr##QIglEHsdYqdJzQ_u_UohU{Jl+w#FZe|sk+7qe)i9!CjuP%
zblwCuepABDCR!p6Qkps)z+Pb1^6LRt!E*`gHz*H^HF2lzehjufzq_7-22u|)7UOTb
z@wm50H4ZMd6QrdA94NBYjhA5EW<Fl>;pqi5>E^cJ0I@zNBv2?VMYkxo%DypV6Psic
z<2nY@8R*BXY#K$0FKAkbqH?(5x8M<9%sw$rXA|**ijtLv1!$KIfa5iYtf=STboVB*
zaiSs^>g*&O)ewqK{`0B_XfXPc9J7;eNxD~G#q+2@<!25qV!kIYa)Hr%)3IXqdETXk
z?PZ)P%il03R-tL`naY$1kgpI)k7X0m9!fQ}bf2GXXWi%Dr+NX}qAJH!xdv#mC4v29
z@L=KQEn_R}7f}QfK*i-B%$-u+6isdN<#h%=d<<&GN)toKJP_>e!?8YJAr{ld<6gfV
zP9NPak>vK1d(THT-9LL210V*w{r$bLBD4!43J+Ws!HQ>CQmm*<CKX)#M#i{uLAt4D
zna6ZE=L?D86e8yMr$W=Vm2E`GX>B9)%sW-&WkFSSLiV(xz<>FfIcV4gKW&nBqE4H1
z>T-)WsJkofEyGolo-|=h)fle+wG^GDUy=SwnoZ1KA{F)E(WW1RVec-F!B<p=E5CWb
zqO6GU6~5})p9w3#m5~YOhl|RvL~ok~ujK-yHsU}?=G27InG775;69}h!tEyfgNCLg
z?Q5<gbMf$B0AsFcKdX=0Dm-@7G?D9gPg|P<RbLNA=LHuFT^x~yDq}ri&!8kN6AeKG
zp*{A2nK*n6h0P7U(mw~Cv0#IFc4*+PzoaTC8xaWTXf&1esk%fbw7VdFgupLF6r@KI
zQYSV3s*V1Nz9#E}pQ1&r`9%c<r$0&^cGSh+<{n5w1(Y!Q?7Bfw_({&S-Y*~S{KJvW
zK~qjS)0;xS;1k&oFDr`RnHbS!f<(yP*_56|mm6r1<R2PBcGwMH{c&hJw#I!bea(*R
z3X(CwBQq63?IS-#4n5aw%=q+8bXh)HXG1whL14C>`mr)+?vbnWb8VV_Dkun#oIxMg
z>4@tO6EKVQPZ&>;^|WEz0vTJPbZ|&5OPTuTQRXP0#Cg(BB;44wAr^3YP><F;#+Iu;
zw^^)=F~P?g>(8m`nl864`#UDT)Ixa?+zZLiT^BiH*7Sv-BIZSizPYJZnWne>?7*iH
zm0pYn$7DVtL|r|g1PG*gUctncM)k>O@@MR^s{sEpqLe|o+%^V}w1!)ISQ}d$Lipu1
zu0IJX0$VKNpRD3RhzkzA{P-MZ&3SgJp}0u1co+n)TgCcuPsnTzfxRFxlVDpKw%9bN
z2qRGDeZMT#;__izY)Q4|88E~=XYGD@u8(OIZc~z*G$Y#6QQ4yQZx4~Eppf>wvB*p4
zL<5`|(-QE-96s3Jdw-XWs$>Z=S+pN*i@a~cS=q+q%KLpTs|9o$rzE2twQDXNCB{=Z
z0#a9c&=!o1TG(a91vNQGELOsX50-Fzwv-)3*nV9iD0jTPHXYNZe_6#Wi)>#OcU7kE
zpoYQ3LF37mC980jDeyxzvzte%T2Rw4p8&e@OGM^AgRdy?Xi!mc4AAQp?pl0-RT*Q2
zQ-UL4dtW7yoe_|V=+}?zkdhg;O}9*UJA?FFH*DolE@-jKP1MHpwiX<n7;qUsGNrs~
z8zu7?F?vuao068*>gUPXCi&mkIPrp~$&T^WW%^JG!6D_;<5s)Gnx5zlkiKtR7=f@M
z49}~Nfx;LK$DdFAuE~YNBL#Zo&VFzSfxQRykxM5z&Q?V;o;(gCbGw8rAQ7dA+RfYl
z_2$xR;nU!JHpj6IM*A-1KDu@|F}YJYzmc8|(z4eXwD>v>2&A;r@okRc6S4MB5|!!*
zXgtar?ov~0JHYKs`m4`T%J>s2i-Az$Pal<{IZu)3biqu;vn-8Hc`14GJ)4dEDI|=7
zZPIO~E`$?ZrLMgvU4-u3-kw-=p$|s{E~@6w??)?mN$K~==&J)#j1JzIh`_94N0Dn)
zIPMur(5DS!;)SU$5pLUXJYw1ce|D?NMt-fMVw7?VMsZ`ZXj)7R?UC~3EufOzXVw7B
zd9!Ps@TK(uDdYiZ<Zl?yQ9;+V{6OGdcR^Zv@V*@*w^+{afH>st3g7NC^Q+fVeA39Z
zWjv7I2e=~)Qdupnf<w5RPi2-glt#=bvb>%X=+br(aQP6E${Bg^KOqj|m4(KL)sRYw
zK5lz`8}0GiVj&F)WRl*{@d1VdM#U3juQXWDUAyWJ2D9KsO~xE>Br08Zffvx1kC;js
zV=}9Cc=$h$!AZYhtS-HN5!xkA|NR_rS-Ipy#fHAnP&EPLAxOCk#3WS2k{7L};Uu-g
z!Ott5h5KP+;LOW(bSrIhKjo<iOO2h4xCj#Mxvb_!%;!%*ER1v)Vgn53M>NAy;l$35
zM$V>2?{CTV{YS3=ucFZbo7dNNZCm?Ifi3xB@h*ia+)9y|_bC$%A0+60o0fl#H@ouy
zMKwh1VE)Z+QO0}M>ItW0>dX=~<LE`W%<+%@#YDQ>Nr=QopL)kCnFM)!`#rAz1px}}
z)<c!4t6)yT{k~Y~lMDbsn_(7+z1RxMUEjPw^;n6c;FJyhL85ytJJ0qP30(XTemY^#
zCwgJGSjDn(<iSSa^T0j8#<(C|aPWevj(@NLCx*pCDbZP3Ox2V!ZQzz;jD*WzfjuDx
zTIVT9DE;t@Bof>r&6PhHF^@7@07_5>VQ+LE*JaT`q8(@c1P(CXfL7y>*%A=?S?fOf
ze(1d7d#pK!6V<T_?flu8l^=;14RsgaEHasR{0Mi%n0s4Qds+3d@xJS7hs~|lFgQ*(
zOcg=-a2Kwt4nH7SS2OfeV(mvV3ZZ|z*7RLnz*d44LW0^v)13v7A^GQilqhVvdKd0H
zk@%_0Djw&N+#QHs3>$C=@?snlx9(Z<C55CVauKJoHxGVg&HCr<Xo}YL;^DuO#|=@f
z_f9m-`X?k4RItfD6)#DU`T!+{lufRH);#TppU{??NNaV8GFneJ3Vqsk>rlaEE|l~G
z8<$H&aOF=WDcwNcvsW7wpI?c{lQEn<uv?@SapC#<mhu4g@S>yr-pEJ&?Yx}RqP~b&
z`#-xZ0UwCpq4}o684sE*b}A1PF;T52)g#^rc%f%WREZJGAm@YS^As)_YbUJHcCUqq
zy%(TBEp1RIt5@N4T5KJg%bRNOcFcMC7^@Z{7w~YOMI}SY%38^Dr)v=e@qXTkyW{`x
zJ%iEk<O+OF*0k!%5*f=cLh41t$cn9#QVvR6a3EfrT%D?CjFjOE8wc|nFC*@yE+NXZ
z#&I-G?B#kfX#4VFxfyN?Bs%VykynP+iw<fL^CCF058VUZP-|_CZOSot7j`l4PCY-H
zM+p3P?*E0uWF!3lhPgEyd%C<PECx6$$Nxjjr0V_$$!}}*f2q%CZ>jo?0DJ8(vyRTg
zp?J)jsWOV}R3)02fQQSuPPs-z*Z-Y|xP>RM#4injEQ|s%cb{9@|6kuCUZ`X=o0Xa^
zwYL-SO{R2kR);Xo>F;#d`LuayyCM46*J|bZAZPR_7vU-+QrO-1YpmYrTi#mnxMFJu
zbo4pWKF0H8#;rRdqqdXJs{SCdjepJd-vGYoed|)j%4HP&ckGW`DqSCl|IUuh>z5o~
zJkdkDr((j#T{f+3GIiXR{H&{+IgraZ4j&ZKuEDafg=oV%i+>=f<$ZhyM8_U%GZ|NO
zjkzjy%u%loaygH~knQaYeJ50o)?s)7%BHwADdt|gFS$o5)xFWh!lTO-Wq5SeQh9B7
z?ER<4r`2KZAQff5Ol)|)>F705MKVe+2;UHh{hQ+-5;UY@kKXTgo_&vE@%mEvK51XJ
zjx;5l(8NP^Z{j^}1a>hnMNrf1XKF3d=p*i}lDgX+cAxudWrA$8j@D1xoZy83q06J?
zX>RHz->)<Gnyy8d`H;oywPne$sm2k77P2}cQYVNw^f*h7-9KNQ_)i1#zYjaD8HAXL
zs^-Q+ZY7OswXVF^x&pJ6x|YQ_o`VrHZo!)@nR+$ba60OxlOkQdpFMVo#MDY3h?aec
z(x9snFB)05VSpb$db|dFcrKX%Kzxu*V5`E>gXk`DrXO-mq$V=9F1>L17}6I2m|35V
z@{oMiejfczcBk3RtwKG1z9tvYF=acfCJi5T{j;XxLFU04{%iV*3*n$y(qx-h=~lyY
z|3)5Vy@?-1Y5T$4M=<x@nEk=p*c12qU8+a2`<tJ+uo{Ycb+kj5C_oel{W4FfHhZ_&
z63ciSFx^VQ`j$i)ckmk7)~&Y&Uk{)OSj5H;&_zof=z1DFJyLbTQJk%1;M-z-dl>v3
zZkRG968BwwvG}6#dysZg+pB_n^9wFKJ}arB1z$MDVCy+74%5=SrOR*R>|)2Y)vD!)
z#o0iHl1?1YIrgP`HOmVi9$vc<vCZ#{w9rKyd(Ktr6gS2pA1yHu-1>Uh9_P#VpA6E%
z#D*L5b3>pLcy>2ksv!_}Zz_mrdMfHYK*KIQ{b~6M*(dIn(J#8w9NKf}n%r;@n`pCa
z<HbPZ*`aax>2Ce|Q+pvO`R`4n5u$O=cOJvP0?Y9I>yvC!Q}bY;QrE<@;2Bv1Ie8+)
zQs}mKD`^pV1I}vUc`J!N%+{IFLn72GA>~a@!{lbH){Uhyuw*L=2Jc@92<igi%W$E)
z(w6UOrbzU*4ESH1*fOTDy!@^Wg%{W4p|pwN!$Kh;Li=7YaUa5?Ij|abS%d~CFC0e{
zYc_&sxr{K6-TrXkzhy@Fl(W+4i#x&lHS7}s8X}CafU{EnpTbcSPzYCV8%{U|eIj96
zPGWpb1l0^ED^wtM5t%@jLZnC*^e66%0wuFd5}gCYa=z5Esl%)zI8?EHE4BztL)s#m
z$oBnjY7`l;aW~#G>xOH;l~ZV&x1RS4S~BDV$BhW|3En+GZ+f>%6tm<GE9|U|;!uyC
zfsR^iR=goUi-wd@n2lAEJN%I<0T?EOpxAP}!Q9kiE_-GkDYO=`8#k6b=R;E4SNRch
zT#*@KYuTF69~=ul@Y-!F-o<a13Qe1TJD&gj_b>uy&Ns>MPaw0+sDsG||5P>Bt>0@b
z<a2X_G8z<6rHmi2uyWN&G8wD3mRiPgYFIEH_`>{x+KXr#eY?DqOZP0A;Q#y3$T}Jf
zn>jSVKgYb=<)l?6mT1Gmo2qC4*b#T?`nZl8B*o_{;mV4KD2awOs-*SIq(llSBwt$-
zS&KxmEFSEM^LDYn*SU=ei^I^oq_>~3G{13$w_*lDeu^1J=Y^372`EzAH-so1$Y;=@
zD`ynDXUt|HLzt4#W|sn^;`s(*T`&@(v}|IztOaR*v%xFoEpS-vqF#DW;)H0FuvC-D
zhL@TGX9ns>BcwtmUnf9x`1nmWo)M~=@yxa`-{Pm3Bbqd+1K20b1r{yu#8PB~3HY;G
z)!=}k;Oam`(K59uLShti?$F-&EgMN(`}yI8<WgK%CGa!-Rm+@&{rai@h&c+dArGf!
zx+qYhY?GLL@b{RMXVZptLG+X9AQ0yKkK~o`;Cx%;g%_AOnDb*$l?wBkCV*`~AJ2V<
zN>>cbo?v)-8veS2yh1r{GU#)jIav84-V4B*se)~^U%DHS?68)XX6WflPjImn=?CiD
z;wG(G2o<282p+PEz$lLvR?h{{%BLSQ?Rju#jMcMjQ&Bb}7+*lV<{&{rKgw6<Y=>zV
zo7xHMxUd&57er>D_7o|FAl6q&V9ZvfIS6A+>PAljMoRCZILjbg*{cLbytw{sia?VV
zSh;-x{X6+CZ~wO^Z+tRQz49Q;jfT>}2NzUk;U9NQE=!4EwE#Hc+DRC;#A1>CzrP(u
zZ-w&1;aZ6J7d`_Av2ftls-zwVxY^C@4Dt8uB9$YDA!;`eo;qtGO7?MG;A>mQxxvez
zdN}bd=65JWwB_+<I2UGvqmibi8vu?U{RnH?p#IaC$k6Sud%iFOU&C<X2dmsa-b6I7
zzE83VXli}u^PRTS<p%v^gBm8%_uJUBQlWNkugjt`W4ree57ZCw6F9qHJ;(3ZEVWeB
z2zON>5X-uxKr4HA)esJ_9tX@*3`6({1ap{~X*k@88aVinOD?tes@(J#2LYJfngf%;
z&}irb?R@UIaJN$YFbj3`QTozE6PhM@Hs|dws9Jutm*E24V{u=NpvXjBa3oOkx^$+_
zV{jVbJ!Cs|U_<D!zxkh#1iWe_n<=#;PHM0MH$I(lkOV(cxFFJ(iYS8Ab(gp6hgKmf
zD2o(523`17fDU-DbpKKH1n|v<T!U)w6uzq`NwZcG+WrmP2=4fBgQh5F$<N#~%yZ|0
zR!t0sqQNb+(&}DpxYbC3E~QLhwJ;WCNRzA6GT?fQib<nVa-Qh;fEyuqAbZiv65=6u
zRU?mEzmBxK2q)jrct;_6)+S1EJ|pGhRgDp!Iu(Ac{jyKy&Rirz0v2;u#>3OxxLlo2
zv3S3u%WUK2iem=ZJizn4c3}l^<pT%cJU<Ap6Wb=d8A7vBsHGB_(%_0Ic0j)#WqQ0X
zJM;upzvx@E=jceVot-Pm8}QPb^BT5DqQHiOSOex!V8|U=ZJQT#iKfb>KB*n_85Z$J
zsHaP@28<S9%j(X-f%=LZjh{_uwE7Tmcyb}p=4?4%H1mPE3|6GKc*J>Mi(u+lhLJ|#
zc?p={?B)Q{O#rcQA9F#x2S_cIp&?ALXCs4_T`(TO7vDxdJ^U2;_ItIWbj$=Ij7!R;
zC3M)ahN;Xujuh_LyR^{?_6d^e31gdXyuCrcCU_M^PU`|T5b(2(vU#O;fgeAE8vMmP
zco-1CkJBf&jDGvAp2PhG-qFV^dqMi1-|zvp+FB<pYu0Os26;x{^>&Yo9nCg|3rOZe
zC}jA`gP{oM05b)|BJd6KI;;d^8K2msKT3oC+pB{VQS)GSDJPgBasO45Wc;fp8-gL@
zqE**}g&k`IY)i20@5FQ1U+nXH|0;HLw8E&jlPl-RQogq^v6Ouf&@$%4LtLhtc16NV
z3D}07eXekn6ql?CQak!D9r{uf+$8&JcH<FLtv_S1rOu}d@fKqWo_7l#oUWaJ64fr`
zp{-|ch??C5Q$+8lVr`O;HTltCFF9MEUF7&tF~fKW1SR82Ff&`ps@o$7sdA#)PXBC>
zWoR#}H7t<}_zW(Y^@2_?8LRH(WG<YBP?sO1fu+;mF6%)^((G7wLaC0_P1i&BL5VV@
zkKc=2U&ow{UdrBv(q#&0S!+d$YCyy6zr9E}l|^kEd}WjV4(L>YY6;~&<wmq~`RNZT
zzgArefV(uUO6XRqI4MUH^0Je5R3&bzK?WpQ)@XwN-OE+_?-jz0P~Ve`B+-ad-{YO9
z>I)^WXOBDdPd@~9Cywwb9xRnC`rqRi@t1ixfu*?JLpZFhe{GJsB`?sEdHskgAVM~Z
zZ&#_NG2Dk(3)sE2=28t!IRvu5f4@)<;n+i_0vs7+%pa5(k%_D7)T}Y}Z_^nwx){bi
zm+PIgDF>iAS&&)28BeSx+E7HDn}5jRi=2m->@$<#@9Ej%f=Z+f?$3*lNFBUH8)&KY
z9ry(_FxS421fb=~@3enJ(u=uzVTq@F+NBs`EWR!^;>5#>{4={N5<i9CD9QnyCLmeo
z1}dv&_KHCQ5lUpI7*L8qZAaMVxduG<f1HLLYUjk#8<{bu%csaS(a-4hp-BH|%ID_G
zx)H>~&FpmKK-q3RK*DI0iP?hK(xLyYi#O;HKfE1(dT;c%eQTTHljzc0$?;zr-@3aT
zi70{WXc#GsKVD3U(62#$;d}&B{v*f3TVT}9iT1b0=Mn4gM+!O?DSn@wA#R;x-Csv6
zq1<C}Mwc&NheYg`eRyNjIygL=^iV>^;GZRoq)<E*@j6*)^=!A{g3lr*{m4NQeET+0
zNFXcnyEH0?S3kWHdZ4x|*(1!Iu#<6`Y#PY|JwU@)rI(kIH#;?;+GD|ow5#bS^Z+m}
zcr;j$Z%kw{p@*L@T&F(%Lw>9I)fPS!e>8)xT}s_>0*5DDut;wQbBOT%+D55@Z(Gw6
z_8fK3FFDMQ|KiB_XGPg8lxAnW#X33dwdNN1vK)ktY(>%skLEq_MULzFaDjUW4l)Bq
z<aQ&eU`bes!E;}Ji{}?oZa61u)&OuBf8*`(;6E#rz)-$O%jDISisO*|{At-%nS+yV
z7O0jn2`fv~l%FLt3cHnTRbRqgzlfFZRg^-pf-0Lya^%;PRI>?fXWVREsr(6jS-&gA
z?03?<8u4bxxtjgTS@G=nLMXP+(}k<%jqd2M0t=R}CGE{%Sl_wE?3PAR?+7%iH`a1w
zZ7nLaghmZhJ8Lio7kEy4DgAIyF>GX&Cgm;LSI8%b$`NB+pju{VJ>W2q5PlL?NgN5a
zZhR|1j;V|l7fQNP0XrflrQK1Z3#$2)O~HI^`Xz#jZK#b!p0-7XXDv%777dxTX+)qh
z+LBb_mvF(F1{w#MJ<o%jeGjC4?NdS~9*)opi}jy4CJIXV@dupaGP*iJ#FxrC2X}%w
zpb_l0q+~j)j04%Mu7XG19`TIox4F${1>|rRqlNlW!!e=K%~TY{;*%cy3G668#-qfZ
znD;8^gt2S9;ABR@_i*N8(~mkTSSbVrpVA%-DasI${g&GW_On{nRshV3tB1?q2L<kT
zlj0RLNO=br2$+A%v^<WVd|3v@5*lH-?M0+<g7FI=5zDu@b7OP=4*+&RiNBvEO0wCI
z7`?@F=Rzuim`WE1D_3ra@9Q7Ea3V<a=rvj;&z3QWSQTI~yKTi6#Ta_-zAVAuz{8<8
zwdQ<DK^uWob_qyPC4gMEabXT1fBD=5AbQGA=h7!v-oycPxII)-*yQG<B0OWK@+jGB
zaQOVTf)RQG@8oc45|&d5Rl^;Y3PvpVS*pL~vifi;5-%l_g0%4<6JETtSiO_7pPZeV
zes|!a8EH6W#Pudnpn^Osre91Tqls3rdXC#*>9FSOn(yLLusn7N1IANOe-T?;F-V34
z?qoYrU(s>8RjhtRG5=5l6O$B&pauXf%S}2aY``XoSxMLn*F(n|0130_7BJx_E)q}^
zX7UN-l=xICagqeBw}0oW{BT4`@(!>4<49^qCC;%)02|WFQLt+x^)gZrq!JISVye%v
zs+YCqN)PApW8>2I<^I+we^I=nv!hahPaOxSTHQPeQI0B<ynyG;7^+2f&7|l65H4C~
zyf+Ju0aYitl!iNno2mM$rvw`DIRR^zPlKDs^LQS-O36iQu5#?p2SaBaNlG5Gz%Kv8
zWKJ^Ufme-jnPnK#!WlofXGtkViwqTvnJ~MLM=|Zjhf1ZLbxj9@f9=+=RpL5Ni9-Q>
zibGL*jzirtNwz9Ur*G!F>j%QxgcVg1WloWiszN%|okkMRf_wRS0+No%eVvS>!ak0q
zMY-Dn?{5<7gZyYge!S4=8YG#D%HVDsG*>D^`y%p9Kb+udNNP*uLA#$pSlj!n##|vv
znYcof5)-g4(y)K1f2`}LrE^wRQC@N?&cxoOIpQQwNz&;1(z<FXBbfqc4B+-lg(ol*
zxKGss<>U+&W;2y#F;$oEpc#6;t__}Fk#o=pi*Pu>i!ESP@odN57yib&)a*qQjp5*U
zRTjNPbcn}N%&sjVT_KR}ExE;ZhSY0B7X^8gZ=7mIvt_wpf9M!y+5lHAg9cr*j>H6R
z!?ubmpYS1yyg<+8!DH{26YN|}D`$jdsMD_kvD$WJY$PCz*U?qs!uXNhoskKYM271f
z`ai~xsRe_AZu@RvYzK_oa_uQ*`e9*D==KLZ$koVEY3dr*LGJ>NW)a2iU|1)`$wXt%
z_nB|}JnbNee~WVXej<{&NE^1OYzZ63>UZ!!xEwKC|DI8v>0Piy`zvAQs+)BFjZT^`
zd_siPl{pHz=j-HTnl50<PmIv>6+$oM+x&2>7wGK4wpkFR>DLJk8P{l$K0@|e`o`WL
z>|q>?`^+O>#d?Q}{9Ka5ayF&AG(1srjn-l>u>#Avf5p(lzjt!rNKn#Lv(1PHjRW`-
z{OrPo_v<3(%Ovi187`*|4f%P4O66x<O#Ljq8y3d#bL6RUQ@L_%xEaio_w*1mPzei2
z$MIwaFxf4ay}+v5XKTpbK(8np#MXd%sL~GxX+eTWA0$4&8a9fq{qpWKz?o4MzvyCd
z9N|jIe<kwmycJC_g49$xXD}8I3pC{k#^G%JRnl|%ljTB16DCn=g7-i@cY|(?e4j!_
zV&e;4f{Cyx(E!+%8vJ>tam99Y73?k5wK%qx+f&0qjsH~}+*u0JL|?aE<J&kauUEaR
zpa@^%W`^_6fBMI%&^8N-ViP`mq|-wQtN3DjGO7x|CRz@ZNkTwz=zeRA4F;mrURy0P
z58;LO3&Ype7)Y({CT>QC_?x@?Z~h;_u)|=NEo})W135S{m*IpFD1YT!S$EqwmVVE#
z(4(A*Sj0}IALhhM5+`vg6T7Q>PFJ5fv_xAhQ>2Tel5&3id=~(z4RYj6-sZt32!L4b
z^4$f{@zt|mUaMr3Mycj{bagXQa%4mt37(FwW}|<Nua1&<oL!gs(S)bTn7wq*&w0Bn
zyVkwj)Qu~&4^=nMyMJQpUT0aCwRzY2Z`y2e-;a|W@p$~@h^cYjWVg8==s$j$xI$g|
zer=akv#go<0rMlS#$Sr8E9%M*G;QkEe?G03RhKt6M+{4zcFU}6BiC}Ax1Aq(*-h(3
zKVjD7NB_QhH;P9S7DuV_o1Pc9^AdkuP7`Y}E-LrDuFF~Eihr+KOjWdZG?JF+F)OB7
zm%GMpo^^f6SKStN?RLw4v@cYXl}q2FXx(#BH@O=W!}zX_ASnDEtN1zoF<;;KX};2?
z>$N^ZC%H~Uv~Cd5PnprO>=4<5tGQ|F#cBru8eN06Wnq`4VmswOtf)*uwT9-7=&NUE
zSI-zejYo_`OMfRa;xQjh7tj8D79}bbwry3i4gCi^k|`TaAdP8CE&uv=Jer{eYl%#1
zMqh2e#Ykb)SQ(Vd(dD!M8or-k#>v{0$2-!_SsKZNkF?=YDwW&W2ka_q$_KZz%ltlT
zoC2_zgF+XzB^%?Ic6?Ga(-P9;x4K$8OGKoM7-?;=R)4X*e2yLU%V*L=LW*Jet3ia&
zQIg94+aNMUPo{okkX?dl@I)sgtxUuc>F9M)&+?K86362=c?E^;NGzq6X8iZcX=BmN
zZ<b}1x0Y1Dyw+xIP?o5ONr`Fgc>m+NY|P7alnUMVI0ukq%eJ^K%A$J+jHr_6`@^Ze
z$3ctK41Y5A|9pv?BCfR6PYqR11>2S@nI?M`bXTrej4!Y>Em_35aC4rWoSf8t8)~g{
z%76&NM&l?J9)!4Nu7o-r*uPYf68(42p9h8(EMkJKhURf#=w#2(u$fO8Td>`+Mf}}m
ze|KaSTs2dNta`#3PtaTOz}ocDL|`78AkZu=#eYW=p&@0l-n15um4m6MeEH;4Z>5ah
z%Sg8@%)5^rzJpz(3DNYqwPbeT0h%+jHiMFZ0n*jR+1`!E2krRm_`+?O8Nq3C9as?a
z^I_`1a+q5V71NLi@nN+l(9G2^xCu~Gh~aAyYBr(0em?u{{POH*qO>u7fje9h!?ZH6
zs(+ZRG0`M}Wv|wy0zBqAq!wv{dCf}WWME*XzpoejRF1EPPF0W-2MrBJg<}icY*O`d
zI$y$z)Jd9pA>5&wS@BYY1bYw^9io|8ku<<+&U>{9YlU2Kga_!VgoguzttVu9r0^Si
zcXnaheA9%(>wvYeZEH>><S8+BA0f8ID}Q;Ck8VA&dt-gGOgFb46gD{wt^>_Nu5?7~
zV|&n6sszL`>M<Yea9gPimIs6VBwNWtVx$hRmAz%)-^1u6AWNEicCbhX79dy5)G8=S
zJ-t~#j)5mu(iptul&c>xJYlK?tvJ&F6{XiGDuRmnQ(Od-25GrFw$UK$ng<XPTYt#|
zGt!4*4Iu*)3U$DQA<TzjTE<8tBg2y)vX4>+<+T@=5c0%wuq7;LK;=sNgLGP<Y7+Qb
zu@&xxw4WVdA=Tr%3R}oQg{0a?>uqu3Ab>GzVj=e}IK<wSSf~)d?-eKZ17^_*iB6Np
zOrQ)y&tRpb+3U3KBb!nyLl7999DfGSlMo4nc~lf+zyx|Tck$uVt_d{zO`r-%#L7y^
zaG28XT<{=j4oZ+d;zdFU0$Wq@YUcr&0-o+Qfs7dSBk7#Y{Lxd=Jm{ce7zd+wlA4l>
zF@H%<zDjCRrA=y5p~<sy;?iqPIdeVNGMm(;&umh8C$kwq1DEGav`_0LYJWXNhwAM-
zCprvyPBibybE22ebJEDmZW5Z>CqD%gijtqUu#xVh3QBj<)2BPBxR&m8J^FMfy}5L!
zD+QB<+sRL9Hj<xGyW}UT7a%(u=kCSlKT&doLcjI=C)GEudX@h~<wpJ!jcde3>p<Kf
zefp#*#CR(yiW;93MaiXgQGfWMG$>%8PnrDr3_z;2Yg{Y-P$E<+((7;@bZyQ+9#p14
zN;Z<fA#M7+-pBLHD;Q!S;&lAx;_@Vv<y62(jqf<t&K}Jzsw6T}g|eKA0{i5!K?I`K
z0}_;yF$BkABS9$>k*+5wnPkbK3Ca`z|5#=XOV>X!`L{XT&q@AqN`EY<pPc-YfK{y@
z%l&Q1WFVY@+YSSdXk9@K`QzaXrZuaGo}~c?P?4*&yM0AQl&uFe%*63nF!-iB-k9J}
zhoJWlxiWkh#-xD+4!Y=&|F4uK+4>4}YQ`*~?8Vj?5|PYvpB&)ujHC|xlNcev2uUE9
z!!-iu;2wm8Bn(94SbtYw0X8YHy9rI!dm}Eid4HDZfD^)#RUUR@1u+=ckD&rSI>?_K
zmcNB}McBoIej`s|i#hRQ#{mM+G4gA84EvvkJaS3VwG$I`LCl>df&JNli3CXpk|^Be
zwd^v%L=M4$44ADnh;Ny!`-dAoAmE3MIWQp&G1;!1wgCW-l7F2wiWIQhZxq5F@ALpw
z1-Zu1)G?t*fEy~=ivU8wh2{J))IJ;n==_O^CqV!}W5<0J;q?C=2;eD&r#oZ7wC>3O
zY+^tn0wBc1DkXUc2ADu*nY8|J3@{>6JA~krP<eg`2uK}>5)hq8{R9vIA$Owzp?rqs
zF%Xb}NRfp}_<wy6NCF@bOB)Jqkw-*Ku7SXAcns~4=?VmrKoAK+>UaPF#&@@oeg_0x
z;-hNw1%)I}fCOd&iCt{U<}@87h(KhDP4Rx)1l+s18zCH$5Lb2Rpg_vU(wu@=o5-Jn
z2|P`(N%a&=fbH6qCLOkCdoV$zED-TyNn9S_6#$cVh<^^~<>qOK5bxp|QwJl0&?-uH
zrIEn?AV5e1$w2Z7)KdTfd{KcxxcWXIgmP~>{#jmufa|7rT{q+WxgHSW9qp2w5VZ#c
z1`8ht_Q%u!2)m~fl$6{BgssHT1_Gq}5I{hC<6Yp?4v_|@Km>NdlEEoR0v?zW&M71|
zg1|`#8Gpt8EDzU&rt0S)fQs!g-V+c&ut2g<&~fR(2te0ncD*Tb8`A?20Qji`2%yE2
z5Fq8dK7l>0J`e$5zl{upCrNq`0)(uKNXU*Ji~u|idNzW<?YnWo?$I2lbHAU307?eV
zs^GX;g|eEFsK5z~q*K3@OP~UN1;R%<!?y@v0)Gg>*KmB-B_Ig+QI|mI(4i3|&mJFJ
zK!Cq)LV%!#ZG#58HwTn-hi}AHh<Cg*iI(2Nh5>;T1WbE+_G&X<_Pwl@8d2s)Q67P`
z(%6g=Pf|n#p1dht<ijRjP4B9j@+x1;e0ICYtF8^5OTdFtmhRZU^dNLuFYI+pGp=vw
zsek)xo@TrN$MUA#Y`yxsdq2&~o6Z%#=I)ZG9^Yq8S7fEVZ-|DQWi@lxPR;oGA(Wy#
zKY4YIcA>}aiEoY=<v!WCarLi@s>`Qy_wqS!^Q@W9U$|GBQN=gsAK$o&-#MLzQ^_t9
z&wiQcHLc4syRL(&(#EEzqd||?2oLW`8h<+0fXC9`B>tS=KulRO&Z}u2Bz6CljZ&HW
z>Q&i+)=Z5b@Ue`?w2Q=yKOF&a$I;*DEg8S6Zu2tpO>g?z<7NBpVOAD*zTs8ha9Up%
z({FiS|GuwxDd}osuU^wm^JUZK6@6j2AIUIup1GP2MZWNjC(-&7UV02F96?e^p?_!f
zZC=&O)^)j{tp4mFNZl%16jPgtmhSp7p0CQ=x+%K(!rrh(#pirl)50i8Z^Zb#@`*gY
z4yqK|nIV}-%W)r{pPik#(n(zrZ@_QjjXzEIpKUCPP9tQZ#+UhY>H52OO9z?A@zs%1
z<7U}*U+bouKe$$GlQ=lHX>zLg0Dpn3it5(AQPTV}ySn&bGxmYRCH#Om9!?`)c<bkD
zH`XUv0&&(7_TPaL?J*k23sp{hg3*dheP1ujx&`BeV&E!0`S>Gomn!V8sOC89^1j!l
z?*%&2BpMrd<yyemj!uE76$;pdwPv=Q`coZc#+^O&rW2j&rfwEl>8kIWe1BF9-Lw}o
zG|Y;&Yl>?+Vp>x3<fh4U*R%xZ8fpk1Yh+lFKgIZSnN?l!dmQuWx(06ujrA)BnaP@e
z7GK=Q_m1uN4aW1wB-HqlGg$iAz>^ByJ0*jd3>G@_9Hh*rNVoXXba~mma2Iu)o?(vP
zth#4^F7v(#H9Q?}w&!J0-+z`%-{Z4wMB!}#L><5by<d~r>|2kqle?xaZ+zoBGz16;
z&!H$w$m|PurgY)HH=Z}Bd7CX3t7*`39JvNBDI^wW{izoSPvTGAn%eQR?fT6A_S|Gf
z5X}Y7{an4p$G`5&ta5kyF`41XJ~V(YAk`WgFgJ)p11eoUEbi;JSbr{hAaGx7k|S@5
z+@F&<DUE?5i~D6~m*4~$>WB4$YPYmv3oXMb+2!TQ`5t{|23A|~mA<>+kCA4>LEKjC
z1b(ul6Cb$O{gw1%z+_t!S-u=$N2IziJS&-@w-0%~$g0Y=r2sA?ho;N?clgfbz)h-q
z-#5O^mbbp{ovTa99DgZ)UYM8DZx44bLVR>G6ZbL`>+$=nt@3H{!d1VC+!IXUk#i1C
z^ZedbR<J;ptoW?T{ds3&e}1wW5r~pGX)3cOh(1}D(a&*YLa0TLcG<XdS!-m8S|L2p
zlzJG>gS_Ry*ADOP?yEMN_NS*ZaZ-WjN37enm;F20hCAPj)_-lK%ulef*p->EKg4)w
zyL=IH91NxJ)fT$=$_Z15y1ZDN){E<=CKdeJ-VxQMCSyLayruG+yt#9->Korh?s?Yt
zaFz~*_cRO?HZOfw?~>4q@H!u0z-e(qAKv8kqHgZzWXL6a?z1)Zr5ZW@TEMC6=I-UA
zH4eo$XEV|Nb${u%f%y{Ho}jMp_>C7INL&@-hLl_3nb-GJqb;|n1>->EzvOU9?ez$1
zTjwD4={fkqH~X~(?ashGEck2dRoR=f*zUTR7T%D2C>-s}VqVw2hdsJ&gFj_Rqhu9u
zd#3lbj_W>YcuE#QN;UrPLI3H^aT0}#x`A3)2bo^aBY*egF(a@^;nmP#saq#C+`1!V
zYn`RkgA8xEjIJ99<GU*5E><W-F0gZjZyzRffnerZ7k0BD^<yT96x%Y#^vQ-fb|i3J
zH@M(TG)rT|ZHZ3OkPh+4tZs_+I^-%hsQCtFkocHN2Y);jo}s9Dn}OdTZPQgNQ-}_y
zk!uG{V}BGC!Y}&eltn4qfpE!nvA!;z_jOrx#S}sp5UX}tcyl6paYAA4bi%z@*NQ69
zT>73x7bGr4iDs}jpU68HCb!?cf<b#|i;x^b;FL+%Z57?PO31_svTrY44HSjUmuC@7
z7Te|#{nl&P>KbOffvC+L+RrW2aP9Qh$}u-S34dT66KZ=OO-@{s9>2?`^|cFI>}xc6
z+kbVs^tW-c<t%8dPOWVd7JKUDH$qE#yKvQ>=TvivP`14rtcUOSK5|b0A#q&8AA$PO
zE&wLBWm;x!Yu(V?*AzkTr_HBL-r4XtH7<Nj&6)<ecdthDqt1o9?%{gRT;ZW!x&k0?
zS%3P<uF37sG*K-4&qGU(K|wdH$~QUGYXp?9{nllF*zgd8eQ<iA$G<)xuCM#*GgrN~
zD(}`dHC*!LUoYH8)(c1nFF;{A|Ck>UbwgYtDNh;fuz(%T^T}~fhXFfn`(+2Ub;Vzo
z-94O0ptkGgT@19W*;~BT!9dtwof~`%i+|Fm(BA(Hl4e=V>aHt11wQ+q;|^OxlpOGB
zEq}LsTK{|9o^mTyM{QS!NJEn4yrP8CGYithI0c70Lhq$bzo4n{*TM3BK=jG_@2-I|
z^F`L&fhs}hfF>I@#7U<P#qi-pVF@{?J9XNG<lY9puWFlsa$@JmS+rA|nSvdD2!GkW
zSi<GiTMn}IkCo~zq3^SSCGYBXPP`zQx+H0&!g~@OJKaP6N9|jG?OT64(!k9>iVmds
z1Cm<Z`5prX*)fWRuReBFY{x44eZQGw4K|dHt#c(1f_Y1;Az4<uJa0)Nnpx}m&@D^^
zzvM5`r*!|&@onb9Ak;NJ8RFjd%zy8ERCo6iQB6S3sFr-M$`!B6Y<f4juK#kcGq~%m
zwdeG<$nLDZP`Q0*!=X2v){yLQ871jcLTuP?gxIML{*o^oVlzcSdAg~T@3Y(6GPgHI
zDLcg|IF;}v-mG2Hbd4AO>GPc^ols`N#^88?Th#a<YxRx0u*!T!ls@k0o`2}Ru-5q8
zDHRa=yMn!cil>{+l`p4==`5}Ba{b_*pM0Xk)dSJNw>c$3^S-^l&Q@34wpVcF?yZ`{
z?(n*v*<n-2He44AnA8R@m7-niZ_xYeqROgiVZ{!ePbnuuk-KelOW(Z+>*2F<^5*$!
zrG2=$iVB!WxJH2CHfU+yXMatxuyKotDQ<Dr9OYwpT=U!pBB@J&Qtjn*Zbz}QVc<T}
z4MB&E+@<R2PH94iB&(*nv=*?wzwz@N_rp1d*$mkjijVQ|-OhWF6LdbZm6!8g0Odem
zxzy+7X4q-#Y`UD%s~c%f!-q8IOZR%3TWl|^!uqW1YZ|BCp8#9HS{n`hDVnUy<@)=8
z#}JwJKA(2&aB=s40nclmhL;?G2`2+JF*TRrgb^u!8ckQ@IPyNfg3~_dJusI1%^c<h
z=%#6)n<ezl&Tj8PID%+!>^xh63BP`;R8nFaXSV0C9B`@hUEfv8>HSZ~ALHyG_w$4&
z2ltN$adeRK%n!o!;C_DaZ|_PpYN-w<M?4I@sekmz|C!)fv6%el{_{ba_;D5;1bz_H
zjPE>u4ouHc;AbJnkL#+-=LrkEdi7`*JawyyYK{Q+{xM;Ud7_k!uwyQL`tpD?07DMv
zD4*&^>BSN4lg54;!$vId^Ejuima2fUaq8X4qFy(xJi%t-eQfF_b)41fO3UUEPb3Px
z7A1}N@W-kYmC&lLS{kWq`d;Gr2@gElH?TW@X0+qWgvFk$%se<}5(F>-2OfV+c$dPV
zpRg2Wn_0KbB#OPdkaDi7C)|)j^O3%KAql{{6EvNKfp>O8jkCIXgcF4+>|5BObwxz6
z+w_R?agxTKP)#M<7O5Kt-n1x?4&p(TN^d4{?AZ?D2mtqpZ18fTYFEN-x1p=`gomYn
zp&tUq-p#HQ#^HC)lQ7{v9nB3pVxi9xD;rmE{U!1|N+R#vr|+$i1{9Jgu|k6AQ#bCj
z9mfL}0my<V@}_gKJR$UyK1-<=wVnD{H$CQA%Cx#J7NXTCHF)NFr9tf9j1>4IOw2Q%
zhwh%D`KNlJc`G8%s!^4;bi~dR7BTLBCwUf(QH`fQFxahTiNtyAN2oqsB3+B;cCl2I
zXv79&v^qq=v}zQ_9HdF<H(6Trl+c!N*ne8CtGO)QfHR-IK)%EuBg;%JS0xA*&z*|q
z5oo?B#e+lNHXyl$RuhmtQUSt$1Bq<306J_4@d2p2Gz6t(zO_eVTY?x6;m)OhToLr4
zYUL6(Y<Fku(nOxEkq|rxz27E`=N=l2Ao5raT*4J{?;aHi2i$zQoZS;R7a;Ybl{6St
z4zUZk<||N9%ZfVY>PhK#Oa=xAUueBLIX(uyR6d=%l*QV?MZG+hXgmIxKO8SB)#~Fh
zQHTOROoyW2u>EVRHC#G23YtcLMn|ymcV(mk!GAg2scafa{DA}1Qnz-*XE(y|76Gv8
z0#rN-JUisZr|-)oi9HL?M`h&~_yZkFQ}5#I#l`gp&_SL-LsnqR2l4^<n;Hd?xSb_N
zFAe^FlEHZ%@TBWfSmZhIilB$h*xuJtG}Tse?qq4?3@tjgXS!K~fp6r0xcq=hJb)Et
znfG)^hne~IjK)0yf@B|nYbm#5jxOymP@fbHQJSE|!Om}$`W^&a&S}7OwQ6c%9p>QZ
z=v-KCoy)3`+FS!MhC&27B_r3uqh#E*uqg3&((}%>;7Tax_%*HtfsnP@iJ6(*JDoXI
zwWB6FpMlfipm9d0q3?@-T$87;V=pV<>sILju{~4U;yuSq&ZgCJ*^Vej$ox2B1Lp!n
z&(OPo+@)4s?6@A2Z^T0UZZsoeA-I>H2_Ru}=re~VM+3lYAsTW@49zKhf!#TPS%&sj
zwz7E{Ud2WPg;=*p2^5S1Sf?8_hb<|Is$9qWFa-N0(y<qc;Ao?NCfL0cqt>%ns0w)!
zW?+4`IsvVs6hLwm$_eviGiuPlefmL?58WEf9=J8YShtpgTMI&-renqo7?DRqw-zRT
z5adJdRP{5Q+_)u%F(6rSbXnjTIfhdiy7QcSud=O|lKM~-QNBu=x6=|cA%;+tN0*Od
zblTKwm;>U!Mh6go=E-33Jnb#sQ^7-;0FF_#c5I=vg_NeRXvemFkY$aA@Rh7-8ki%F
znx&h%thSbzfxkp&jb9?jd+sH<&-pg4wC2RHO_R6A$v2`vxgse+#MW|-$0Z?4b_@jO
z<sPSbK#Xl{1P9L{3y8wu4fH~lqBI6G_wIc9UXfAFcglr-jj6+0Q=4~m6I*9?HV<Hb
z_>Z~>95%tas`l6txGmi@jKYqO2q8*5+v)=-0}arNrd~fSsBzwLSdC#bPU`9Er~=wN
zB4oWh0_BnK$p`YlkHchtJ^)LZd;q?b4}?Z30zwU}vEhbYv3D+GfWcn=-&24r_A|x@
zRyqk%G|~xwb2i!y>kG`fSuG~9w?XYdErXc<3eCVS6Cfm9KlH~*nP)@)_h2qYp*>Ds
zc2xkEc82nV`ls`wEg5mO9vNQ*P}x#N*DF<wIjT7FV@P>-dJ`rPNL}8$6ah$Y$zVtJ
zsBmlOjQ?I`S-fqYkrF(jU<5O~Qw8yj>2IHv?VM46XM`aha*0OHWkuLea!sOK+VP*=
zVt1-oU{-7F%!oX1gG->f6$k*XL|e(ilm(!5VFGUEbd#Bo7qU~)l;VVjOzCsGflRC-
z_*y+QqS;UbqaggAf><p^6>BM)XBfkF*}qBg5#dsNYSAIlosW1{#CH^PY@_Y+<LSmI
zdzg)X6~SC$tg3}YZmEYImNiGj73l4A1x658pl3&4Y+V3)B)p)DMmc7M&yW%RjtK#e
zGVha5c<VZ`baC5s^gxv5hDO}g6lk~FwztKS5yE^CIQB(7fr<(WQ+Z;I5(4*=5Uhll
z%OkK?aMi~a@c}qJ{VGj5m7u+9bxR<fAAsF|DbOkrPW?G=T%PU@JZcRhKaAqPkfd`+
z03iT^h;lTV;Xcd4cUd|(K0YXA+}0>^8zts%hJ~=%o#8CRO|zOxV5RN2+IEOa*X_|=
zJ#W<0TD}b@o=ZZ2_8$_)3D*dnDB%xM?_&ur@Q~o02@NviK`@;ISN9RunW6<a!=%lB
zOoPxynV<n%+tsEf{t>n#Ga5GT1&4LDD1PS;!SaqLF+j?U|A|A4Y=JCt1u_7p)9K~0
ztBsL|vCH(%^)7e8QjhhF(@X2DK0w$~v_l^_s|zLdsLP9cZpT%<qR3Repwk{GcK)Oz
zDtGEZs;0&a1L%p#Hh#Be0~<5A_QDQ-m^mb6C;(inCFB(*bjIeHU4be(M_H1X2=iBR
zt7GZbqF$G{793g8hxFLk1Z@5WVgpoK|AGq<suOjYwqXMTojjEg$fyWH^J#^KZuA7*
zNDSi)ICjdzJ+;%!PlIs3Xn^}Rq5-l+qJhBzH{xr`fv$E6=-K7P)%iKWmsW~@kL>zG
zS>7eUAseZ>vd9P#_FYx=YbobXDEzkTT(yNU2rm-7U1*}HIRU1wF&4P0Jux|l1HKbA
z5cL>kjj<?lc8SLIWNlU~*4nrT2!@SI2jB`VwIt?H)P>lu)Xlw{f)E#t&<opn=jNg<
z7m(7F^>Edh04F0wwH}Qrs9ec^!LbR7M>1Eqvv{fLD@vcbZS5bz$Zb8z1oFI7H?d}k
zXU_5Ot<?7;0=Q+r<j*7Koo=Y-a=9v1p)_HUElbt9dIt2HvLUXpQdqVCx3te1L*fhW
zO9~ARtf1zuM5?g131`pum_#1?N#4)xx<#4fy<r2KdkJu}`<}GXM&}BD<*H=3YFR20
z-K}-p**^^%V)W4NsavMWP%P6pF=Ck}k*$K>-dx{xikfLMkkD0KS1uTZxTRCu3?Vu>
zmkm`gQCweCVaJ@i{_DDWvbfI<H(N+ZSms0Y*+Nzw+&-qUJccCF1UzqVvQmeHIyb3!
z;=~d~V=UOjq3_E$h&5$@99%(+NU5t8C{6J(PH4<JPnbFw*+bkC>w|j>IXN2^a&S&R
z^g)abeOQqAEar^wDdYrsKl0(ZALZ;I2|I6vGriVtCSQ+<DLe!7_o9BU^+GId$Cck3
z0OnLKpA;1dB2!~S<Sy;VVpkG?IrX}b+AcI5twOt)7vj!-QTQu=zUVkv@6>iovs(<<
zE}L<m$DEe^wi0vWL&Y{Z1-{v=3uLHn7Zk1MJ0w|o0J`xxwlO}(kyd<SxVI^G*wPS6
zP%|H`ANAt(u6EwZQrR~d3}yZ_On!H_R?`CtNCQ2{;v>&WK-jU>MeL&zO&;%Fg6Itb
zQM-qiJaz@rI0$=x5hHfp$?>wK%in+$SQthAUPFRT_5W_Qrm#<PM^UoRSfXLc^Swa=
zu4l}4X5&Y*LG3~SH@$bhzHLo81�>mQw&VUQS6}_6)#ZD5r3i`C%M&Eh{X*0<^2v
z)7uff^CJC6I8Bf1lB#ZD;A9nmb5)qU1mvFzoPe|0s1sy=z<dIhw6@n2W`ca1NEog9
z!SWb)G33tY0*0P|@Xg7U?XYpZtIinFb^>k6b4qrqsoe{QwB4_!hS;fanB3Cb8Z*-y
zeCI16h?#5KMe@H(PGP5Y+tQ1vNnd@!+9#hdkM~C#`AD=8zPmYMUO`vuwWcce*Ojzo
zN*7rpa0-ck+_MGiKTc@qbUhm1=z1x%G*7;mHx{bh{g%N#1!rz$$RN&d??#<Sl=*zC
zMxO-uUbM+H>GL%$_cv<ZWAR*4UCkd6<COW)?!EkwCWY+8iE!?xY!G&55!=sFdO19_
zl{F)$iIdaBL=7-zVf(kX2^+uSNi7UaMEjdE?wyK%jck>vn<YW#=>%IZO8l>%41g;H
z=Pc$*qFcq*HJDE;qxjGRo6a|tSYr5Oa&U;*`5ya~<vxq{=D)UlN@MyZUd*)uoB1HX
zGrd441e@B67_qY$M2v_-tA^iRi4+lK=apb6ZQp}9xIDYLhARQcMz|k^z}{hWm}Jb9
zgtN?3#)fA~%T~G^&|K3_9Pl3$S(VmIvCM=F_`TS9Rb!CSjGq0mub`J2fQ*qrpu=%!
z&CQPuaJswd54`yQNJ%LOn9qVdkAJYF^wWp?pZ*J2qo<;m;W`TwmpqRN6azRpG?(Fo
z5h#DvJZq2R#*yFmSLi+*pg~47`@z=)L9n~o4ZP>QT(Wb>0k#jIB+iI7MJgoa+3~MW
zRaZ61rg}yioOgqR;Z?J{+1>T%dUM{~zWU~UlrFL$i)FmHy;}$^1CcEf6$VOXi`(tu
z`{nIb8ZPs7Ra~uP6fV^b{l0tiks7<JX!(D5DB45b7W<+vTI$~HnsQTg^l;a3Z*$~E
zaV4VVXLzmL7j@qSS3ll<c=hh~)jwW|MF{_ig$$D*5o(cyQIJTn*z8|@|6{n=!r+HR
z7$jM;cra7<3z5Y^lqhIdi%+lq`sxh>mIPT63Iq(c3}B5oi-IT?1lXVQKJSYDY9)W7
zFkF86^LKxuqkZa&L&v>u@_w`9?()|HPx;Mz*lAo@XcdSkgT2k7KZG)R8kRT<gE&m)
zu;?s(2`oYdaR$JoVGw08?fYF>7kOJf@>vU@2GJ~vWV%%0)k>#fw2VRqEDS`5=Uk{X
zkm+KDJfLW)k4;l|^B2X*iz_&{;x~UKz4^9mH&tQZmH4hQ?~X5rf?k$Dj@aGyKfnES
zYhR5{x8*nC#n3(PpFe(jd-k|b&_hK*tmVsjD_RH<WU!|pNCH!kAfXB(8PTa;U&&~B
zI285PTm!>6&RttcDa~su5d%Fn-?kmlOz8x!OjR|w#zR@(Q&&DNXPysjgSUU^mvYPR
z<lAl8mrb2l^t>(lysUV3zsq}`*7E$e01kmIe!+k8nuq(wzF?-s%R>(L^eT#$0S&yK
zgB4A{QfU^1v9kB!kAL)l4h#%P=NWmu^(-TjARXtNUxnpKm6tUGS!2&5gOYbYcf@#F
zsUXo2U_Ar^vWx>n<bAQl%cFk`{P_E>2B<~4ynkH10YSl|ADeAaRfZ<`h^L2!UjA2G
z0809j@CjQNG~vJNP0<1w>i&^B{(1yVmp_D|+~i0s-gMVC$NL>UylrZbn0q86KH8<5
z`kFf0qB|7MUU2baLmwP|eOMV-Lk2P>APvzz`$z^tNuNb^I!u}H9rAw`Vqwa7fQ-J_
z?CNqukBlVjFZYM4K!MMY<KW1oFd<(z$DaC)s=x+pcW4a$amIZCWXXF_C^}qT(^*0O
zOrVLfAV~)XeE0V4+r|Tu5<wUav#$Ak^`^QK%j1^r9Q3rs9dLmRmn+LYIxu&3!aYSz
zicV+5J33`O2OSMyebs-#<Hj%$^lTcut~hB8Toq8~YdeMr>Uo^7Z0OtmQkt9kz^mTz
zFKv^~)Tv<{7zXWjJW!LcOl&(nAL<FKwebL?wXbPb-vZy_J5z4T1||TIjxDi(v@{lg
zLtq7>yj-rUeDm{a-F&98JzgFffwLMO2i#G|Ca-DInmg{ABdvcamn9Fe<nc1GOo2zE
zXW5vP2G0m+>GvgYEnT1Mvzw$cP;qwls>HFK(iVrN?JbI`igrNqVFyxS;3^3`>$24;
zO+mE-!jDzYc2!x^@4ji-JR>>>!`X4cFf(=o5KyWh1>MF93U&hxNn$Q^#9TNLbEhDN
ztg=}Deku(OC8&Sh2lK1px8<I?>8Xn?Na!+m1bRSxfFQhkq~1YZ;}C&fc0DmH3{UP6
zP3fAu{sGUO+pWmRgdvaMhT*B;hDoFj&7pCX=2&fMV9m>cgjReGK02JMHbUBS9gNbH
z8o1cpB_*Swiv$^T)J3tytA(A2Dg52#C|s<<z**a3BG!Ka>@7{qi4bY*D9I$yk+976
z;aHWqPt<{hIw;BmT*&;McfV_!J?+|US~PV7>L)7%4VdZdUEWb+T@*EqPsJNUwnl)A
z5TD?Hp@e^$a?2JVpf5^hq7(%noT==HPKhEb1RLAt2y|GjfrMM?X$e!KWK0x^rOpU2
zczP)8%)5WqSO)k&K=Wi)WW#oRlb}_S4rUb%9fn`biOP;zT@SbksQoR^BFT#9b?JzA
z*vB~Lh^d@0f=tD7S5Q}Jta_NTZF3xcsA=4s3%UtI&@sC;FjSR2Doc>kH$EFHOM^I0
z2XW?M9WayAL70KM-?ZhL54<kAp7uSoIT$0P15<y=_YnGz-Ea`{7NL&^rPbrEy_B+w
zz-#(@&xvJ>Jd>R}apJQINzf)bNTo}GU1RwtOZlMCFzN^!2Hq?IXpAI*9wWdqR*yWp
z%Dc`Obw%4z*G~=De1>%JKw#rLJ6>*EA~N@&eP%d-Nn}()!;2(b1?|~@+J8s?ocm`?
z1U-L~Q-fgx_N7?z0Xoq8G-5BghdT$yeJ};g6I;EgfWSOL`{$aThs$r?*%ekT<+jl^
zdv1X(2&&O|9Y%4LcL*gf$NFtu^bbbX@N(QBiOq2akno+IkoU9gXlC2955Ft#cNP5K
z50Kyy0}rPm9<m@#r4524U<(ooZa}0Z;qre2Vi{>nkE0r;=Z}Ey2GPBRMp5}ZK%q5o
z^B|c`prxU=nYFM<>H|0A&EjmL;dnE=G##~#W`W;z+vMAdCyd@CuWtlAtOPzK%)=N5
z?bXB8Gb$%B)`ic`(qLza=dK`F$m4{7)&h;v72_QYm3A3z&k#r<n0kV-VE7k$&8>e!
zY94yHHoq1;mwrqTYTcGwp0oa`nEPV51Tly(plPd+Dw@(c3xa5`w|UF!jn{kd*<eo9
zbRCa_+G=W+@ZY?b$rKx55{PJ^@^=Hm;vh>@ic?~cS%8i}XB<4(RCm0Rc0^km#f!<b
zAHpc~><C6)q9*|<=~)YC6iwzt0+oMOy8Hu8RiW#OizpfAhM{M@EYJhVtLENwDgn$c
z2W4~M8Sw-SQ4k8{;Ni)`M`7YegwAG&7(9YU4`^`!mY$vPlV=4PB%^}->^VvtggSBX
z=!x`1VF(sj>~cJwxd`q97Go_i(+W5UoLv+n9LPf!W9pz-V?p;&RF&J4(~o~wU+^-#
z&oi7VBOS!ySxyxWiC{|)TF|xw#Qke{H9gO(VRiwtX6KkyUdpWLh0Ln|udL=Q5|F%A
zDheb(w-UmPy2>WfgI$SF5!wZZs^EBxv6WBJlSM%+_5XA9aLmW(OMtsAjIxZZ_qnOS
zesv;jW{|>G5o=KB*fg~p{4#$C<N#7MRj`O^Hlb}i%%@>xmVsDO&uQ2+h(?dj*l4)g
zy6Ja(xM|>Kr4@MVB5}nJ#cO0dns7JtOrB|imq)f;Pn`pcoDxRBol~O|>JziDuSzB<
z66KQw7*LmL&Qntg6HkRN7Wp$akzss>O+bi@+&v%)C|*;Gu83#%I0Jve#4|3Pqe4n?
z2`c1QSC#s?gFDyjDOH*L7MPDtN$gsnffC6%`%}950;%>(nN#E4KaOO@ovs~=Ku#8a
zBETobz^4zTxpS~;G1}y(*o#CU#$?4iFLs2^Uc%f@5h=ddp1F)jD=jjsYh5md=0xDi
z*u$r>39|o7)xh_@duo5OewU)*N54%6tfFQi+Xv(EdMqojn?mth;VKCvu)$>BjgRAk
zNDYYkAE2Vt%jDk}7t<gj_;Ztd-QPS1gG&iJDbOLj-+%b7MF@xor*1N>E&5}Nsbr}*
zn~06dWU8rs7&2Qt3#kqkS=n2w7y3-HL)n}KbWDRSK4gRV)Z>3W`RIWNJo`9$0@X87
zG83Nx9pUf0vSN6KV-2xVJZuimtO)?%Lw7uI8UW`V>#4~=CYq3ej%RTxYe$$^X2yZS
z(56MO%P~E0Y&1P^*m%t85Cr+zu2WS34Sd^E7sWF(JhXj<JPgxU-jffG-A3AB_uF^Z
zyd_8fahaO;r2v0virz-ulQq@5K#+@X-s$1WbeMeggr97FV*qrB+cYOZp`4^|!66{S
zBxC_o+Q%_}UvmT(qq4AUv9&Q@{MBR7dEIOuIj(!>#G^H5#Z3}&7Hh8TE4*6HYn7ag
zz{s(&AlhNqNic%R2X>?>;^>FvFy~8L&x=snzB(2R3<ZDdrpE`o9=J7Bueia96xx{@
zC6h>e!cK!pnnkh4f^=vw37VKx6ouy}y4KYn7#ImBvRWy>sXIm%3=!NNo6Roo`f|eq
zMY|tTV>6_BOjnr(VKV5$)++!KkxgP*xb%Ay`NRDDkT-N{n+8Hh;e^oez&THL#gB(A
zP`7g(`7wV`*WGwT|0)Ab0<vfY8U+=mIoAJrKwE(Wj@l^<gi2j)Iny~hAwj4es{t(W
zaTeAa4M_N%@R_w>ltpSWS?7mP`+RS-o>1G2$NCbdpBY732={2=6IOQ&9e`l%-iBH2
zPz*`|K#-XPnDxANNKI<FY#?+}IpiQ?SSnRPXzPFNq!$E{AmFLe&svZtrZXHt;}!dg
z=+<B@q>Ix@=bZ)R?1csxtR*<h{5vniu>g}}GK;MSDTnzEg#m&{oC=(oZ8(MmTMrgc
zBBb6clRSro-*dYCCl4GF5T6X^`U8cxgb#oS798+~OUB9a=H|w<;^haQU6v@2vKoo2
zvJrnvVPZWnz+Vd12RNlG34W($26ziLSqb{X<YyGCeviYX!L^qcX9KQzv4VbQ`6?wG
zC3gyc0vaPeRaP^mNW~nN(u)|7q~fFsZO$-pdde@(ektsi%=QbPqTR5VPq;gmBJ{b(
zU?$3FK!DL1%4$P}8u!jrL>#N0u<2L`%GQ73Lc^4zV1tZnVXatCVXo=DwOiE3{kmw~
zcl)NdhAh~(L72@91um^kq_Iaj4{pFcR(inA2XE<)7}L50R$G$cBbRGqA&KSIR`g;g
zGWl4$e9E~Qx;63T^+I@G0$Q330H;0^x95yFOc*la*EB-5p|rJ8gDvJ-BQa<JmCb*h
zTlK624*fwCP+^lvEpT_UCp%KIkt>$bztb~Wwt{&gkwZ!q`^eRhlaLp9D_WU|yqhVe
zs*0F0XtCr}uubaHkl8xsN+|4TFh@*Hvui-(#BR`ex(J-V?~3Y<7wqVF0%1I&nXItb
zf&q>DHs2nN-DW1<w`^<CsBxI^c=Ug`&p(&@BSU4d1Y;t+M)Se+*-*L1bJ-r-!B?5(
z&O4t=5B3x0BtQXYm+9&3C3^aU*RT$wXv!sJH`vr|v2_c~irkjL7l+0+pNP+4!E}_I
zy;#z{ScbC%7(&{lh9|xndZ{W)%=ygkSU7->ro<>q!S6B@=T0hirZNW`jh}z4u?3%x
zUbBS4UV_rRKakAyBh1&-e%O_E(mwwg#jVMeWa;c3o2`eYldXqm&ej`<t#3N@`@!Mn
zTr(Gt0ITi5t+F=-fa7}veldADs6gu>5&(1R;@H}@Q=Ae|pD9;@UnVT9WeICG;Oocv
zM0<^p;A{RO!dwY9w_+<Fa6*4KhIZWupOis{7V1~!aWIQccsv%iECxkGM*x7Y-48c?
zNSj32tWuoPK4IofwJKp#tvF*ywO-S+<s7aT$f<M|2Fjir#gtL7{-L3M=H`$Jf0Ze4
z=!%xM-O|WJ7R^JZ5eSzgndub=PfpPE#;2fz^wJUrePLc$pUn$L7vz71)hy1RrccGH
zCkDp(0Ia?+siI#TuRjS^$6q8^J&Ofvk$IBhZZrlwMe)TkG;v7J2?2SL5Qu$3V3szF
zVm&*{7s?7a7Ch@TOcX?*+Z{3T7aJ?OJK9nX*GlkJFqvw%+E&J~vOezjU<v<f6}Bgi
zbnT;df&pi^u`<HXzoviJdIoWlo+hP!i<>&B;Fz6kTETIVs5w_HrN2l8%0)4TuO%wK
z7y$S^qYnsy->yjqjMli0_uhgVnPp!L5}#|Y!Eet{e&EK#mZ>M^A<`fYqx0!IAzoC$
zcL_40uYwE^TBB|^_7A3~=;v?c)+VeD1=x<}8#Q>qBYh{aoh^Usrrq0QTxv@wRcawT
zRv>=JUxd!71%uD~9BkSCxV7UPfJnJk#iaEdQwDb=Z%w6+`H5*XTbX2!=y5(Y2sD+d
zivv;<=R;>;E@rbN+)!N|okXiJgg?ajIVABo2=FbTr!8!R-{vl+c>8tu0uWy?pB+7!
z;2MJ)j_c(zY8HP@^A+}E<P)n{ja7cu!+S1NqTo8omqtps_6)9<7Oq40wy8ncwa6km
z<WJ_p`PU12%v)X)TbLk<J|P42&__nN*5d)cI;?nUCXcpQ#9hEkGPy@sa1Z*{x-*&y
zq#Z@93GLJ(-kmNs`T#MWM=yC{F1OUPa!aqbH&)EY$&i2BjfaaCC&OnH(9G$lJFbZn
zU>?OKY<?XYQ>b8^zLDx^thwt=HN9EGL{*HRSyGf2dMQ8y@}chsB*ge3@n)Fftucju
z=Yu|6<6Hnv$Ec#!%Oh~G*)67+xtbhS5wnP+`BF(rwcHhz`GN}DhS4M@XtzgmMrpIT
z&_F^R>9c=S<2FMVQ3l-Znx2XN5<l_5x_*A3!u4x<U>6UY-Htyg!#Nx%pr^8@-y-kI
zqMfge0kLj=;6vMtXQogsG>~>iVsM^&7J_<T-kZ-eeKq{x5T?&pIe~hN^qFt8W4P1t
z3zq4|Pi@9xWj9mm`-2gu-}qwUHjBm*5A&x3Jmo^3avD$_fN+;WAO`(NfBV2=*k@oR
z3@f2b;G}QQpP@M4;!i5FyQ1z2R{;Qp%l7$tZT#f9XnX#sn^Tbg1jq5@F_+;w3loPY
zq6vp5q6&v6q6>#7q6~*8q7Anvq7K>_12r);m*IpFD1YTUYj4~*lHcc7_&#kLuyv#;
z>XBeD$R?RV20OdCwGVfT%^slcmfhM>OFepw$M@^6idCd2HEk!s;vTp!-BgJpi>zY3
z`TN^%{`q^YFN&y0<MiV8?m{OQS)4~gWf!;mi@$C@w0Bj1Df3O$?y6Up*Rjer|1y>B
zSl>O??SDN1e(LsBGyIMoeyG~2FUR^)ip{V1QxW{58Kyf`{rGtKx7#0K;@q1!Rw~lz
zg`jEQjPy~B)8o|iIM&DNikc4$Du-b@a=Y#5|J2uAPtBhJ7VFJJQ|<5V7u#;);l~5R
zj<|x~U244<5VwnKM(0|JsL%x=b#GS&1H#I*S$~<(FdeTd)tiVww}*PbR<b!(WjhdX
zJd`7~9{9_BN1fV^UyS9J8g)B%_ABmO;gr9@xDxikY&`(Cm$BGPcZd=a%D}oi`e%Hr
zJv|QP@zhj|)?H^iPu%#igE^`Bz9pL9KT>C+@0CB}Xtqi42IaItbvukziPP!A%??kg
z*njN0c3+Qm*Om>vp0R<4yS_V8zhMX8><$sZ|LssWJb>9U03Ax4-)%oY`l_kT8lhL$
z5{xzt0Q>3=d9a+Ck>31@TU9nwH4qcVna6}G%OaWPmJ2@#nQ^bg>y<`Dp4dL$!mx$j
z!~vF;c@)Fp;*Mpz56mcGuuRSqDixS8kAI3nt|k<r!E}^nw%@D3Kyj4Ks}nMWa~$c!
zjZrp3N87fadgM8CHWFP#3IxUCaJZ9-Y?3CH0(#R<$F1Q^02<B&*w~GR4*+!Y!yn!e
z5P0V_5b(j!Y+!gb@{nC%&9O|pMgJIBbP;KpdAp~8iM0?R8I(sVjcwPurD@BiA%8@6
z<xma0JhLx|Tz%ixpM;2aAQPKMdU>et542=+=y@F8LW1^X-NI8uuf9ii?)VFwk7STW
zVie1L#w}Dr(79DYZidGe_Ij*$1mBk<pg9iogiL389qkcDD4r$!_XGs?(SpE>vx)$B
zR<RckDUiY~wLq90?I6Rnou`M9Mt@qU`y@cZBuyikJK5iImlO|LrgEc};K;!~L`k8k
z)B6CSVVe?YdA*R@ncT!WK<w)zhwUxm<@PqDE~;U?u6!w&Ir)hs%MvG=GEIyEpi^J&
zfDRTq(!b&5Zff?_Y0Iw%6557!>Z;|=M(9H?qZsg+^a!*%Qi7Qk-2ke}-G6~VhPnha
z<iQ3r!WH#+;AvOl6Ck@oH&na=I$=(%^PYfmLkc5zX0!{;$0?l6Q;8fzoS!jn&|uu4
z1%|)|_{k{EdOy**&ww?ic7%b=N#BhQPNyzoy#!N`+5;Rw_qXFfZ_Pfzf4MW`P)|4c
zFm}5`IRGu{St|ql_A&=jP=EVSH7EM4ulCcf3N74ekzIE@O=Icw7l23&_ItB+sj!m^
zi@1bw%jTM>-#4tE=F{nP8vL(_k}ggrsWsYQv5_LAPhqxMpTb+a`@&dE)FQ@!#x&Sr
zS+WTfD&gfL(X}9=XuCEo!qfx!P?bGvZ?XB;f%pWP$F8rq2jO`|Pk%<nYpdOHr|SC}
z)?}6o&-;4MJr8782r}|;AP_Vo8lraY`gj6Vc$_uDeKmjy0Ed?Zfkt6b2+*n^2?UyZ
zQ5Lhw`}<4S1rSzmyVe|0V0}9U*|62V-r?2<gbMCp8fleV>lp-)&iNQ)>&@v@0p)&k
zc+B5sag-KT2!8inNPjtlO{$zBV5B$<3Faqq4je894Be4iwC&j1mb;(td!QPMY0&2#
zWA^qV{gYIO1|PSkthtB%4~@F7K)g59K9J<32f}S_t>GKuK5X^X4&|Ovb$J@av0SoW
zxhGo<1mCb^L!^vHmD@|DH@(#}2%*F1)BiN}Cy;X<11yOKMt=f682X#7GS13G!Zag(
z%FK41lckVRCY+u3p&h5%cM=*Sg05&F+;1F+<-*Scfng$JFl0FkoeaI0>R1mIL5<}E
zz;Wt%nE*UgU{VZ5_2WD6B7iG-MEU^s8@8`UB|JXT3-&DF6`ZvWSi+$I9g$ksL=s-r
zQX<%K614#pS$`ivMB(R2yxB6iuTB*Tu7;n7`i{L7=rkT+zWW2cp@jkzg<?do4<r5W
zjM0R{Rs>+7g`aiw#`7-#3Y%q5rFpJVJh83M7!p$$6p=3XAQA@GpX?**$d>w<8eB{v
zH!R1Y#Usnjt^s!x09in$zw{kjgmYgSKB9-4nshldc@N~KsYicY&D8U={qXI3{I}8@
z0B`<G|25A&KA!3w3rxfS?S#Iqxc5Un9%ulA+8JITJGM=A1aTZ#usOs=we65e6#G~B
zZ}7%OQ;vW*I?fRR-q0VI1;Ce6Q}*->qyS4Awx+L3)Yl4BmN8}FosmJP(_k4!I6(Ub
zjEs?A&!-(9KYD*;z~@gwi<`F};ZG8vW?Ubtrnw&AEZQ**nUq!021}L-!R2qOT{#V`
znCj6f8T`O%8LOv`AbgxW0>&D9LUCBCBWtHYbM^pmY7{04rk7D3$29%TsoF8UnpuR1
zd~;)Dh$v7TCT3@0f`(+xKwSphJsTvj-Y~%Nq>%tq!q<N*87P+6Nw9YT13Zh;T$2T!
zEY0+#qbfnAdY!%v%_hJhbJqS{V3-7~ea|P>?5lw*<hklmo;fG{w*bazB*2mNF|I=%
zf26=U^g6u`%$0y&;oRoOFPbY^&6WCd8J<1gAY7Uw#(iKq+Mm#L%Ig(Jxpel#w=yse
zP(tL+HGY5n>eZ{)A)^MAqf!g9+UOM64a8v-0wd5kt73SA-uw9y3-GjO^1XADKF4ro
z+z@&?0&~JgkA&rn9h{(g+gq`R=Oc(zE~2c^UxpM69Y86M(zx(3qEmjYg^aXVdGxpD
zbjp<6=$_g}yu*oRy$2Ap<M?m3ftvKC_Jq6pp1Xg59)<Uh^Z?3%BnFC#)_D)`M27ru
z08m7I9ab#i7g&+&kC1cx$XDWzhwd6=z7^~UX+c4eK(8qdycWB|4$RBpRp?xZND3F8
zxe14Ql6(&Je6H-to{~L7P&m!ure|c6en~c^|LbH^_L;IN`8wHDJSCgr`KXz%m7x@z
zlq`R~xCYKB(&x@v5rH8C118pLIaGZp$`q$jHd7@;*(^$w7d|yM@=W2S{E8Cc@c;L~
zY<Z+pH`(jJ2n)LXC5%nxMWTb$NTW1y%$ofZvq!p&=Vr?G_;YakQZ<J>jZq2CRCQPN
zJsFV+TihC*b@P21`zR=Ex^0ESd#N%L#(;k>?!c*d!c*31q=e(7i28qMdBy_|fB{7E
zNJn~O+?qt6nX^~)y0o7ik2LGWwNmI;uN;}(p`$=un%l^Y^}N!8L=eBUD=<%Hf8hV`
z(=dkO?}>^sZ!?0&BYpWWrobCZE9zQFxe0NA1e43Rl|(Qv0D8V~Qtzh#TEGSAQXqeQ
z5^)@qep;;M#6Q?Pdr>U?H83BZg!m;fn11<ue+#7v{4C0q<sFmT^LV6dnVY~jn1%(1
zVKC)xJeI$4&uKhO{55bMXulIi_bA4ibF8zNKkl{z`nIUV_naw+!9<g470NERRTJ2%
z7@Q3Kbt+?df}{K?R#=2ra8KQIp|F3}+T53HEy2MRNxt-ozUSy5yF)tY*qT7($i7!;
zBab#q;l~~Cx`reE<O8k`P6o#d6czN1M8+t*4Z24%mj4&d&+DMc<Ef%gLaMMxT9Ame
zW6;Ecz0ldp#9EmwMSgZxSNYCaX}t<ytRt=TKfp-)A9Wy?+Z)dj#AlR!JokTzlTZpw
zM)7>M|094e7*yFj%r8O9Bo}M<bCuVI<>J<eC@zY5lM9PD<<=5&BEZorvUAI9k(CBh
zXEw^8L|UzoUf={w=ZLwl+K%FqD)CZ|3;xI=pXIfXVP^rU`n+xh!8@leKlkvQ<G*kj
zu75EVHYUHQ63))7sg(N@-voaxJ1JkTKFDHuj{VA+%@Vz^5EU6(N}HHy1@0l}I3rr&
z4H%A8e3NW2JR4ERHy$|*V5t9dUf%N6(p`r)Q~>~uy504a&B7~fiF<)nDqsiY>ws;y
zEw^=3n~=3mt&-NZ@CQI4?UBv#ft7g&+a=xEc^fp^bEUq*J6uE=&INxDaPt^Wc*{Hq
zdu<C60s}g%I5(0S62e-+Ofdygyhml<d%24}?}tg+1=b=3$ZGx7t67(xWNHyL*>@Ca
z&A31c=|#CNgIvF&<=9WQg)^9!PZ>c{fq%Ic$zPB6f(YdaIIr9tf+a?#dPe4}7|@eP
zuMmMi=$Uzyy<}d+0rP(fM4-x+*43Y8%Mt@+(iUbx>pjRM&z18rC=!tI>iNv2gr*av
zwh&G(GI{Xm$M*5r&yopFikK1<{1>L)4AX<xP35KAsL!?>YARWio8x5C;R4evKiiT-
z!O0aIuV+Bhf$WK>O#v_={k)IaHwjWxt;$A7LS^Y(oMCXxFc5#Ys(mm8BZPN|FD1qf
z<yQrdR!3f>3ny}4Vdi-ttT=OF>PiU0X-BLFDY1^FqRgW0e`o7aN?s_71i$9Q8)UG_
zi;CY^t&nM_6#)8NuGEw6Qi1-Di@=sEkx;$|D=jH44J}J$Tq8cC2c=9=<*u3b-03(q
zV{MjTK0!Ib15tmT2pV49Q^HHfR%z$i*K-YzDKMNTOIWkq*#+VJLuZHL0o)o$8~Uq3
zzS*scO%YD@Uy0*N?|mSc>Suqc4hCzqY#ofL*Q&S}?BQ@}0>udAdIe}zTw1yt>AB|0
zE(*$&F#2{bAlvfi!_ichV3PtZgR2t@1?F<FY_)#D3c-Jfn-g=pt{rNVF(+&*o~#^r
zF<mIMn=6wjc-~e+ohtq6PL;)*P_dvyrJ8tQ0L0+oP@|~9W?%g>)xBN1<sLb6bs{tj
zwz65Q4il`oDCs!ZsvA%FSyy(;DM(g>VwtBVSD1@zC>d<JEG#1+=P#~=XXN@C`5C1W
z@9H~O9zK88cEo^no&)t9g>P+~&6W=>Lny^&=V=9e^s*ZqNPz{IqAd#7RpYfQ>)jJ<
z-UR*vjsSf#P=}JoK45O5y*IV(*o{%Bxf>ZC_~Lkk5xF*k*Rw2F;VQTS&ghm3a;UZf
z0IoI$AIt{BKjoI8F_3yHlCzZ-2_<q0YaL^sMM!_NVEY;V@B4C(m4w%z`ph<%&Vf2D
zoYK{4!NbRiM`C$q3VFv_6eAYgqh_pVTMWi9Qb_X5OZIrmrPBjEx~lPc#@>(`{OQ>x
z2^t)QAC=~MG_JB8r<Q1CmUlk9){&f<^}C?6nauHtbwRq|!X{f9fT_`LaDgX|JN4y}
z>~4Rya8`9oFIG5NS4L8eHb()mGJVCC2l#|)bkxQcyDaGJjxE<o!uv((!KDPPkS2x`
zObJhAGm2Q<Gm{yf77sVB2$P<FVD_Y^ch#`thuBv8s3)s^l*@WzZP4nF0EpNTMq;}L
zK&)%RC7rp3Js5U<4SwtTLPri~tsH))UiN>og8_3jJb;N#2-<s9nDC<8{=Hf#<$?P3
zRm0VsIT!f?&?hp0XX7^_z@J*ihXq7veCkk#GBZh1M6-gzdl0!RH{OM$w88{f;U~a&
zsT|V51lr07E2;SO-ovOR&{$IOG!^elX8>YhELLp=A}OE_b@qZ8;D=JKcqURO)>eOU
zOSEjtl}7l_=q~h&TxZV#QXg$mr6VGK&HuHn-}O#@U*0no3(=NJdwoILhzGRx6^Ah*
zdY<c&+4Lf-SHzJr*&=Vl!GuRsL-M^KC4mR0@zCK(K&Vr5^jK1t;n;Pk`w>Jor?e3m
zd<~94*`O(a;C+QVVZu51&Xj)w_!)nB92ejaDg9FEswgu5YQ<s=+iuU(8~aw;%6EfZ
zd73NC;0=rMqn8ITD<2u*VZ9DYCJ3+~*KS)G^3XB!EZ0Gv&j|LztTy<o#ll}xbG#=M
z{MC`l)o11ezOt;~zbPk>$awSW+mMZqS1pqBYfds-sMSnf<dMkE0<NnJ3Z{Q5{eD^m
z1M_+I|1d`&^5VsawPzZAQN;S`{k;v;cw`1&6{Rvg!$&RlFpeD|WPmSVXVL<HwL2fd
zyKBHdS?+U&^Ea0vmL$sO2_M6Oe<d6(V&dz+JaNM>_Fcd@ezN{VsVLU*`4Lzbv!v3G
zwd05+C1b3mzCU*Lo}hDO0zQ8|5#NWk>N6@Wmmq-D*?P*7CNv>|6PJbKO`l_BYQeAz
z7xKsPY$B)}&Q!d*H9J<v(_Gq&wm!!9tsVPBP!k4{b9Km<B#pBIq;tFsYnp2qsKD7;
z;~M6Upj^u=XDw;9HuhpIiHe=I<P&a#%)mYp>N%WqpBQ9Knt{nfp!t7Ff`N&G>g#^$
z%WcC5oaZVK+H?aKJX?;qvg<0j(AwD7*kj8hG3~Gu`wkMUFRY_l85=-aG9|KQ@Byry
zh`{#h;<^1<$Y{>{|MbTXHzD>@F>IP+Kdeu_qH!O1+B?3pA~U-nXiapph=8Q<Gygys
z!xOIun7Vj;xw7coh&+GGuk~PjV{Bt6q#pS@GDp0g^UW;Uw;1GLzB1Czy>1hU&U3mp
zn)B`T+}?)4mhrH2zW<izd~+qw1NYykmsk2trzy|Rubkz%yIql6T3ud)2~Gg^hI0QB
zsr`oU4(9Xb7&Kc9?Xq>7^Su-vDR3h^uoHaGn|(!l70m{Jgr$G2XW5Ih(||DI``F`q
z9=QV{1$)IW8#e<R*ghKr79Ez!zp~Rjfvtas6R+2~t8}N3kGZd^6G}#jyG6IMVk>4)
zoO_-TzSwt1x@-k)=WpTz;nCAWc`5C^=tq8Sax>o0a#<N<AoHbFoZ8(prmslao12OQ
zi4Ayos2uKenIC^Yw0G>yB#bNOF!p87&|MGoxDIVcg)YenF*gwrWhtlaRaOA;JuS6h
zMCf4eOGfn8WRc;Yt_vn-u@@5wpiqnANRo}50f3wgGu_?l75Nx&q^{_Ny|1WN$BzOs
zY7w^Ks2D&w*52?eZB^~7;8h|z<nDTJz9@8L@BjPyezJdt#4_E6ay5S5+lr%nP8`~^
zHrPoy1hVi{b9>P~x7i2;p?YqryAc)a758E->3yNSFPOOeEyj58x|OXN%5AcUmb^UF
zxele{?U?KnE71FDojlc|ukQ~dPvQsC%1~e=ho&o?`0@Qup&CCKrG?X8Ci-MUc>dke
z^T9RNWw?K>{{wth_BIrFgSVxT$Y#O8FnKX2Qb}+J?yKSzhDd<%ytuadmanb8fBW`r
z_jy-S&4og~VU9%))_O1(6!{k7wy)8VR(yL9!)DXKUo7quuu$V`>e+lld7uF~Gn!G>
z75+<mxpHxV)l6cpmClpkjW&1b(cO=9@$Af~_og#42b8&nZa@Z(p1TqHL<rViG0FWZ
z>I{s@nPH};E<-F+fF+R6F=y!U&HLML{tE|5R?L@i(GDjAI507n;e-(=f9)FCa^$x0
zeZOKJsfq>yK!V4n64!cHyWTCY<5)`kVEch0IU0I7<d}!G^4F&ifX57FE2-oyr78{B
zY&05;zCpE-@TWE=*o0b(c#|Do{PIHhA;H}*0aYL+J^5NX=+*m!*}Q39{2l(3CA9@w
z+mYCpw=b^Vns}3lB-G*Ne|ASpNh4GUPx<Bt?`?TMw8h0%2ih|)@z?jhZHv6j`m(7R
z-W9zal)n{>@}o4eDt0~ms!RLw&@}!2$r)ACbqvF~SyA_GfwLKw);XhPP5q;k8kkD&
zD$7?^rCnDdI7=t496@HLE${dJL~qx$mN27yxc<n`eVdkbk@EwPfBNb61ITTwq)1G{
zB58&e!B=lXJrkTVB8pJTg!45JN!#iX=osnU6vwo^P{t#kY3@SkQ7n?gcXRW!PX#R;
zi#UWM)4tGg#4m3zV(Hnf1Lfc)|9(r<+u~64Ma!UU-;`O=@dNJ{2FN@O3Y_uzShVE<
zWZlt%FRpJ#MFB-He?cf<ORRKBqQaQ<ZW^-c*%?{{32-ZgpOA9>@RQu+@CF1Zq9ocp
z5Z+;9K))gorrcDUn^V=C);!luooJ!MWMh0I0zf8#d{guf7Ya7ke&UV2YO{UW7g<kw
zt5od!AT{53c(P@E?>0JPT*e{@kS_2p)hX{fNk-na2}BtCe<SY`kswTjQqQ50NTG}x
zNu2Lf0Y6Gb8>12tCYty82DF0AMrjEel^%{x2X4rQ2>ex<=S9unf7E5)t-!xdAB)z#
zcm-yBp^f(|{<FqIFRtsZINYr)aeZvEy_^1J+GqP2%9&_QNJ?Q;01`eGN&tHDlv12a
zB?SEfwKLKef03`BG&8o1Kt<Y9Qp$_1F`@EQc(Dyb(0Ko1>&qnYx`(pc6D;%$Ob>$q
z@nNQvhDK<BK_p_cwEmAxQ+IX&Skx`x>)8UaioCX{h7<Z`qOW7355{7GXt}`avdyZZ
zJKw`%iRbTOtHCjy?BSdDH@83}v<u)05~T!e@e=BJe@obHXga#Z67N3Vyml-2essQq
ze%1;g;U|>}TmbwA|B0(vIJZCWKxP>o0}UQPTIcf$HlY*mo2D;#g!Y5KuVD~KFe~s(
z;E>_}eNks?zF#Q~?jUl`f0}4$Vr(Z8)w1K?y6Jh)KOW00t<V7QF)zDp=sFBv_;?u5
zIBB9~f2-eHN)SeSjsTWYv$Kp}U-F<}w}5Pnlbhk*5u|Ccw@)og<>A=0eOmXdb8c4)
zodBGw<PeQvDWL$%zeArjhXSGyJo?_>>Vn~A-4CVnP{;~{$>FDu{LnW%sXG=~xwF9R
zDKQe&j%Sp1bqJXU$I#A#_HEJan<{6_W82)Ne|Ke7UMR0Gi|!?Zn~@|576(Vpjo=7Y
zGKY~&|Hzu4qnRdv-f4f#0*)3Vn5&<VYDB=ShH)TtO*p=Yqrl;~W&BMjlE{zAj|weA
zKm&tgq(3o|FXzq{UV|;nuzp@y3+z4|HTSb!D_N{1!M)9>$dz3ImBcL8qCD-~R6^m$
zf7&g_+M5+cPALwglAu61Qr^(Pk!$H~^8%AXPUp=0P|t<0f>@X-fBR7?v%*0vv`QRK
zH*oAj!fDS-vmO{!zyU0*-I1n3E|OLZC*1=?4!BjBUo!k*npfp+Sa)U4>D-KR2N4ez
zX4f8o@U&WL85)ESes`au5S-m0WX&paf5GqG*Zj*a0tR<wowkp5BBUPZ7b3-(B*ORr
zC=a=OnmqfiXxtc4wxHSZ;3K$_^K<cyF9NCmZxwT^SV#zpgk;p)zphYH$WZ7o<eKLZ
zNj#Exxz5bArFFvy;}M}8Rfz&tggOi9q*XKwmY1x)7)xWPy)jSTH)t6Tv+&cRe^Q}}
z$<V#U2jl1iI7@3DkZMUYRyX8{=?z&0{63Zi=^-GU+o&a1T54<yfC-NKb9q3wO>hd9
z)B6xKIfKYhc9cI8A3!=s^9hY(J>{R4Si85Fz{DPMR$#M*{CPEK++j!yeQ#f4@P+&c
z+(wc1O<QLCVgMV1&<nAWmJ<v`e|tt`sJY370La!!d3VXrtNX_}V75h@lIReR;w1dC
zyb)mvz$009Tkyz}DJMZx%2g1KM}vT41)@rau7pHgFo=rUhS99kf`3i9Di3?=LG10?
z=76IJT;1_lU~zZJZ_u0K@0b%H($nq<FzLwQ7qcRd(gA#;gNEPxHqA>cf0LOKtkyOu
zl$lUX()-_vwy`{sdqkY7vNVR42PYM`MZ?AhzMb%Q{KmE!PPcE*rfD%*@|E^(Ssu#w
z+%J$~GwB#(uy8T~c)rInk`VbB%8j%}B|%tQq&b60Ljf|`oO0m-pn17N@UCcKGnURu
z%tKTgv6$%H&qX%R=p4_Me_Hl5t<_eH%Nxlh#d9}|s1XzhwsCU(`t|EZ{9LU#FB6F<
z9wW#t?J!WlLQ5k^>e9An7$1&uhl+gW7X1!lB#&4>DSKsOx3)y@YB~iNrHoXiRux+Y
z7g^Iic71Ue7v!ylCykIK7+v~@p(@jrefiio8q<|ynte*|N#237e<=$x@L;)4TQ^vo
zPxZ*Y+K&T>v++{?`OHmM&5`UuLp8wdIxIoZEpgukZ5=9x(bf@dVe15f;{@9}fGc&h
zIPt;irWwFxZ11YlQEz$qvakw@^kH^ZbmZay<H62tYVt=gg7?VouVr31Cg=yKezFTT
zObKHOZw4#@raT^=e{#`nacsI0Y|=I&I2)ACVZx~ELG-p)x9|I7_wwp$qT39e2paHU
zqRUG3NmuO^>FXBDW+<ACzWVO^)tisk;*eWihq@G5kP!P>*|icz+aUVpVvBXXcRjir
z&Ofn7(htWKUBMj%!pQiv&FB`JX`d0k9L?p6FV>1$U-(fxe`?+&a?(DZM2o)+XNNo;
zBWfiZn(y4Ak2aW6chiz+i^^FQ*MLNi&C);bZW0tt`@D0jlR)Sxg`&>IiZMwM88z?G
zGR=pA?FKjkt#1xgivvp#|AALZk8^85p%2Nrzf>J=%tkS|k1DM*)_Qn64lU>uwQr-?
z+EbLUI1sXfe-vP7-Fs#NN1$k!c;u+6Vr{cNWS`2oeQ+meWw%=RdbwFzhupDas_GgZ
zO*=;(7dvNW-6kQpbX@#L7#RDCdhJ}wgt74blV;mpVI!GMY<yV$+wwE}8-BEf)WCs6
zy7Pb$l;YMIkeUF51EBOcs#KN&g?@ayw-aGI_jc&qfA2CIhFVcgHOUVvl>#Qr@l=z@
zwG1DkPB7xAC(`n9l|Qdg2t=SK3O}tGq7<<|Cgz-M;NJPCJB&hj2N_vcrdAYnQ+8GY
z=#ZP*#&|iaua@O=2+<)5rycwL3coPYqw4(y8>NZ2PECZ;{82$N?p*pSJEzgCDpXcI
ztOj%ae-Uda`%cQ|G|x$Ibzmcb;Y%SKkOGfYyW@vl<C3lMJxJ_8jV+CZ14dKo62Tiv
zq>(O%G~-==`H>-2*<)4*7A+X?iVPhF)R6*IzPJIv0Wyxtu5Rt#u{iR*Z@E=R_`v%1
zjvDoA+4Q(IUxDd%b{IQrNW52hdf-uW?4bzie>hW&XR5FW4=OMs@yB9&Ej%phO*|XH
z3;bTx@j=0+l;^3aqb6nV_Om27866|@UQCxB)nM;E5DOv<)KzxM0ZIXISVKF?XB3KU
z^DR1)NEadwIJ-m2b%jc-0|A*Qv5-l8ev?R1`6QV}(0I%+zK7#Hau-W-J&1GoLP+ov
ze}uNqUe+9lXLs{wRm89*54|JP*z2Ia==zBu-#em6F-|D~hxv?z0M0Y~!=sUbnmT8P
zL)w-U-h7$ZunmL#1*AhD!!g)1c)ZHQ2d=QBI}CiiR(Ak=DwX*8VUMy{=`rb0#B+tz
zYPH4TPenprKwaxN8X18)s<p-hPnh-de`x5&93Y0A9E@Wn2Z)t-(>K{Z?Ru*0LEi6d
z*~xF;6kRqxpz+<jdt-03@oehVwq_Iw9n898+N?CcjEI#6o}gMXi*R&ewLzF{W=NMS
z3mUqr{&$zt|Lf({V!5?=0dfKzo*`lO05;_W&OSS)d-aa)QUM4lds<;v2p_s0f2{Yu
z$IDT~!xOsxWw6&BD^?>FL^=KYW~1%()fm24{B0yu{QG3iVxeG;zk0$RFx8F!D>}0c
zl0+CE#Gk-%+FLW5I40N(+fhQVlgQrVgzhROngk8NNzm}X_JUIwHN=arKDC!Ik+*}2
z8cR$C9-JpEy6^y)Cyh*qN;Y_te;$uj$>7kk6nS^};mUa~?HkxzFbekji$udJ7!CI5
zx>mAk5L$G0n#<2fBFKVA*?)Ka&-XXi>-l*&C6AZP?j2UrL(e>8-))6uY+;X?)eSQ{
z^s6Ta$q_cz7q04IU%Cr_SK8X6+3wn+SgSbVM9BDAOXAJybQ9|6TWvAYe_eq{SIrwa
zxp;@+O7Ed_0r11k$5)o)y6dUCMKpG{Ab$7e^b;9MbABVk6)Oe@aS;l+Krj1CuQ|G|
zIkH4(GYVkre?SEUlH*S-uo}WXEmcO=6!3l8m0VR}>BN_(4xLg5z&)vSFI_>%zsoxy
zgp5`NV_qlnW*^Yt+>$=if5Kw&FxH!OF|=eofL<>?xxq@%esZ(5X*F+nnP+lF;%nFP
z)bp|t!OOA{A^mE(u1qL=89t+ejb$<~Vd?7`q>G+;)|g@-eMH85aVcK7#1vjqHjV?n
zCXZ~!qSu<bNXy=J@g8a((bWX*dZ=wm7~>a1CvvJ9d!KK=KclbHe^P#gRiJvNQ-G%f
zHK3_p2sC#s`8zr&UOW4F9jpbqT-Xom9_;sW6VKRdvLDusU&K-3Ir|un>zd^^cq@0C
z0p9D?*&Y(RPveva8(_}*w)Lq+x4HXn%^7CARoXVpK-$)rr<!A_6F^+4=jenC<|^T0
zAy9scLYe%C<*R$Fe__lCnN5)|om;9_KcsWHw!OxceD7h?@2VK{@6w6@D#>niUKZwl
zapM;TJ0Gktz`sgP67%oTpV%c%tR4AxnRb5i`Jd}CGb>cH*iR>kWn0s}TycNWu5WK8
zZ<<~Iz&-3W=jTTvNuot5k00~=3%HvBrV3?lWOH<KWnpa!m+U$V2!~(b4u@ah54T_7
z5I(2^I5U@lr5!1MR!wu;HW0o0SLjx9CPGP&pvYV@<5Z5FL{2AGd#O8Ph=e4xB%uOe
zROQ!qK>!jZGdlDlEf<ToZ$HV~<;~A`B%c(h$Z$4Ut|uf%X`W1GI7d-3n=GrzAHg%m
zB%G2c3Z4#1vnHHk5(U49aU2IsDlXc$R@Kn*%aax31SKec4*y(!oJ5l;Jf&%IiS{YP
z1OQrA{}AyPD)+p#GR7!Hk%0uxWXKsPgAam=)VjE*vI=7o?5PanJYeq5FwPYC<ul;c
zUwGRSou)a;vzdd6FX=8s<(WRxhB<}{THXrrw|)3kit8Z{Fdiz_YW^HWn8|Z2K#8+A
zXUQX%KwwjUGNpukMJS)X7m_(2t@iFtN?u&&cme&iW2|!67$sKtaR2=(y~nOuF$W;l
z$K+UC<<PE2xD0c+tEo@<2%?F}gXPzQ`i$KbT5eq)o+A?NJfwjFMeGA5nH4_WUM2ZC
z8Nm9m+||YHbxQXK;~huAbh(v`R=nNVi_35p1+3hEwxSlBgF~ISbXBwJDw_psg!C5@
z7x9PN<H*DjDxxAb{Xb3nIiF%sVb|*Oa2iJd^OdQt)@o#Gtm6X4t-o-`H=Sf76Nw9y
z#f6`)ckf1^0EZmsPi=+S1$z}vlMKVa!tnk}n_bPE##G$=m!gA~mg_K%0$S>h*3MuJ
z?b<tk$?(ug7NI;nx>-CvJS|7KB(SPT>a8lOsRi9{nHd@*>a7#HwqmkA^u;pIjh$#C
za!<A!=aQiW+&V8G=L>5+Zxw5xv?IK-7{%!^jmX-6k$l5j7z?|yvJ_>7*;!CX;(ClX
zbYn7S%1`~+`y9J%D%`*J<4|$B5v?CL+Ey2T=cSM^Z9CDP8g;Z9XHIvu(3&|YT1ugu
zg>(f+eHbR><UEFc{OKt#XYG|j)8gyxg;7x)3oqGDNNvg!k$~2wUQueB3KX`spkRj6
zH}ga5j%sdlWDFXnB`=jV+B2)%QZu{XOshIgDVkje(IRVjH3EVWN&asHN5Tr${s={X
zGYmsR5~wf^IvDR@WhADp)+fM)%<XoWCT6)r#d$qOY7KE>cwsO54gulix7HqWy<K;8
z3(Ni|dt$rEm^{Ex(@yafuem<B<0E<VaGqDASuxgHsp6{al<FG_9N9ggS*y6MoVimU
ztSxtUw5PWcl{@O+AH$;^HLIAAU=WgjU$lc9yu24m8_pzmoUJTbO|>1kJkY}Z+rh~V
zX-(VGJ*}AD+jxdHgS1+Oo4<j{>jEHqu=Ex$Us(m~34EDE97c#Xj4IgJj;Y#F_}0<~
z`vmzK3EAv4Ji|f7%$G{rXC5A@Q(Ja3$RLv?e9yq%u?KyM;eXAjb{>0ks;wk{Sws5@
z&CE)L+hDcw=;qD5<(qB)at3fu0&)i#Ft2H36*hDbp!G_?07%<ZpgP7q!$*L>I@A98
zj5`E#YB*^E7^cC!>HJb)=N}-)<)|7V^w!B{yn2Cu_iK>BVzr63-yQRR+<tkz8sa=d
zm|R^Az2q%?Isp03yceMVGZMTdv6>hAO>Jh|jhp!ZR2S^eiZ=-Fv=@G~?)&w9Ghg2P
z2XPoTIEOin5r;XA5{EgB6NfpC6o)yD6^A*E7Kb^F7l%2G7>7BH8HYKI8n-!(8<gw>
zH83_bm(h6~76UmkF_(dV8Yh3PxMOr>ZMP<zq+;9Hv2EL|*w&71vtrw}Z9A1zY*kQ6
z#dgwp-nUPm9_Q=tPxtw;$5{8|HRm<&Io8_b#7Zjk!lw4dW)k*xF7(U{OgsPupslg1
zv$DOdf<3n!y}X$z5bz~uhb1Q$buu$@0ovP%8M&D805r@@0b*t*02Y5105dl?H!L|o
z)ZW3<320&I0-#b;)}W@PrTb6GKPCWU&;QVTc{&3v>;M#B4{l~Q_71jYb}nBb{;wTX
z%*+5TmSzBRpp6+oR8dJwT0s&(C8?kWkTkP1b2734D7hNj08IdLKoc`NXESPmxxEv>
z=HCZ^iM^dE@SoD08NPpF2|ELf0L~6(CcrN@GY=CphksOb00%QCTcES^*E0a<46txA
zvUB<Bfr~u=XlG*MYWfd=FS)t>zW_Nn*?(Ese(An~DcL)_IGZ>D9b5okS(U^j{*BYc
z(#YkX+|Ix+HNf8d%hJ@|#Py$&{-yg0^QCn$0@^tPT+BRN{>gu3Yz8m|Iy=}Hd4A>o
z3g+Mh{1*&YXP}+Me<Gj*IGI@(Ihop+IXi!a`HKBdcmA_Z!2edgk%NPc=fAw||7G>R
zFaTYg&1}pWV3}FIa+<h&<+cFY!7~0cE7Er6_5fz4f6Gl>9sWb-X6E#-22uSpGt^%o
zj7;tAY&-#`X66pCj0*NHUnv1p|931i{FkA)920-vF7f~Ceg3bZC0uQ66pU=Y2H@X2
z2Jp3HjO+kkdj=o}_-E0$+WvoBjBJ56p8tPN|J7Q<?B8fb>}^c{t4`X*=&L2db{1c>
zFflOwTL^TP0D732Dgj+gECJ?5Hedbvmt4)x)Xd2SXlM4t>%UqGpl4=c`Y$?FOQ4Ch
z-9LW>u>V_SW@q|e;eS!}FYt`A;@S$|rD*>T%k^J|N?$|ZqU!1Jh3$WqB5!Z{zYhO|
z5fQQX0C>|gb8rFZS=pGs?%o$5ZVo2j|3k`u1!Dfsv%Hau6VL;o^Hnz!^S`S9pTGY+
z>HU{5;&vwXrvHqMii?q*>DQ$FFTp=*6IXvHr!Rv4b@{)l{*Uv&X5P%q!^{MBZQ0(0
zH^eGEE5oH8D{pjKOlPp4`P*oyLs^q*abZBMeV>2&k6rF6!?(InhUOWbj}FhJzYgyQ
zGBk(d{g^hCeW_*_1^B+C)B}YmTNGWKw1+{A`eg_iry9}Tmy2>c;MyEa>W6#N-<5y$
zOWz;~X1Z9Npf4Y(14!Kp1JKDH91#6X(i-D1x{;efQ0t5So>C+oKR!?nPmGO@^j$*?
z9N^LB#=(&D8e#dR{3T9y@icV0YclzOcyZ%&08yK8b<mbs6J<y+ZRyOEH?P^E@y}u$
zqOH+n*juWreYmxiPf-ujix#ncC>4ME?pk*|-yrD--ZnhEL;%thpGjN+@uC*r{}|Ka
z0v<kb^Z@&dlZl<N9zuq-y1tz%!lEB7;sk`A2)ta@gDrl-i=Tx(Wwto5!Pt*o;Z49e
z^DxS+=WgHxlt(0t`y5;=(`)rAa2&}Pa=k>ek`MZtC`P=wYsD4&C#8RQdM<yDTt4@@
z7uD7Fxbax}w|rDcp*!A$MCu96*qU8q9<1?_2a=+<?K~d(+F!@(RxqBhSTn%&W%P{0
z?NrS>i)#9<Pky}5m0%o(sb^69#gaKW+H;~02_#XoDwZIFd!|=$eu7GSgzpYhG~5;&
z>;N4uXw)QqN7**WU)(JNSVMnXn7N>PK#HQVwDkP*wDc3gMTP+k$wnHNS9vx?;4*4T
z{+|uiAkRQo{;VGS5yeOsyfVjM^V`jePS6C|k!MHN4Bfm!?LHj-LgaVN)s_LSXgz_@
zec2JJ=M4Cob2H+Y3A94<%0k{iDOUY+82zpIbemtFb9z#qZNv4!yzzfl3l9>eVz|69
z3{~pV28O!xj%q3QkS`Ad!5a)*g-|Z>InIB(%;8}M?fRw#))wuKUw0lr^ggs>?-E*k
zvg7M^%sD0D0;YDW7{!W51RJwXuA?kzLJdC5`^Fg^(yAwIGfVH9?!UKOZ(Xp^tKdFP
zc2d~CvYFKQ;4Op@uN!|t8aE-k)4FqO<PT!~!szrsO2}Ou3N?zSHY7+rFK2a6w8G;T
zq^Wm$T*oC7YVQ+KS!X<P&UGmB?<Kw$<@PjSeE?B`4%mfB1GTr?$_~<!stna(eb;$u
z{I;C*{<%Mh-KUSaQ*7#+KyP?E!>6a7#`Nto&C*3?rtHuju0nsHA8Iw=$^)V|>IC&Z
zyxe#VU+JKsrK`?Eat5R$TAqj0!lIQHgKogGsUsd7<PQYk0x{P);5W<Q7N5ML!@0>6
zg9-Lpz}!{jC$>L|wX>{p)00->0gCI#>Iw%*<eMKtc2tFkX=g)pz5FC}m(nmap-w>=
z`Xy*gH_kdH#zTMDl;>}L9Z#zaHtw8kJ&Ny+`TW$)=m86w!0Y}8yn>|bLOr#${Y3w^
z9i#Z}5_K?ZgJN$&;@G$Ay2~sK6hEu$lE-8JUpqaU1f>KQcmfg<p-xXSHom?Zt+WlU
zC@~W8g^>il9n*SPQ&)Sby0P2^GmAZPG@q%`==#_@(HDQRxmq)1fbVS22P?;*)!yjh
zx&F~>S-fjT<rqjm^)pd);#Tf?oQ?5wvo{8&UT3=~d4mgB6*iSAIF{O8?Y!F@h$k<}
zC7y4?I6?Ep8Q!avWZ%yz6InM0@z1X-YbK9W_mFg8clyx;90f2w-xI2mch`PVxl=~?
zu(&@#ggt+=g_(i2|1BwG_dNCY8!LnW0WM+CUH^6z?D2MOo4u^yFjrQcd7}}`KBYK<
zgJ^&_J@s;NTV<Z_=Mv=RWrw^Uro)l^F%RoFRu!p_cs1K3SC>LgnZgk$Bz9NpTiGpI
z$n;*#yn~$CW`f6ps!2sqxlV#F|E;rNJPkT}8;yT-6dbAjW}Xj@HW%S7FT8lOYie?L
zZqQprWSQ0yCenwbCqy9nD3zGlqUJZP2G#v@!x41EntwH&W05}llM5&(SJ-FPhf*|;
zE3yW&&Bvo6jI&RX^tn_p3-TS~bJAAFc;C1cWf$sa7I#iVooq#kgBsYqzTB+Kt(DIA
z+L?clkzHA9{f^&Ck18m=Lq!Rr&1BaLIVcuz<yq7*2Z#xCQ?*-auch0YW@a^3oaF0T
z7*2PU%SH@s<Ia>8=wyM4w6o9x9-sq@;o_VGKXOUF>6VyAq71Xb;=kL$;?D8MsS-9$
zY+*$Dza2-l;kxL^ws$=|hvsh&q+}j4c|d=Gs5j*2cibq=UgeYIS)>vFe34;{3n&b(
zZ!RZSgPJ`*3c3oJ=i=h~xKchMaau_g2QgI*FqQKMBTFS#Dupt$=G>`r*>$GhJj~v!
zNj3iXmW&*LmL1!&Lt2K#Okps>elnicC^*)*{%Oqd)15VJ-}uab;4wSjO7~o4_f&tR
znI~Q%*qpkr>vp9ZKS;w(3@+XDkHj6T+tYnA?|tv(^|K&BsP1`hNZfMXGZ#~6qXAXa
zam%bxfKEdg$^@M?=o;oC!3Il_K2sv-B%~^5T(=?#tikBi=j5S~>D0O1ggPAm{e{7M
z&;?Y}rL3NGO^c`-*GZ=?O5QM~tPOusc!bnvIrr_?s0t+Bk8vhGNg!4%j)5_hUhfo5
z51Of?&!>jtLDvB+cql}63AtnLZuOP_s%No^U+=qNq`GmEAw>=)oK9E3Yq}p6$Oy8(
zCeB6wTn_ZXs!*ifI~Xt;8oFvrD%q`JyEd*?>YY|zAs!b##iNkGrakXpf_s1J;n*(s
zL1N+Ggu6Mx2{aQ{P1kZ6jzdsiD9)AGso9JK<vqO*1W%q^Jy&7;F&Qn&0-n2)pq^Jc
z)IV2xW2JgIH$Vg*VtWbH7Ega(n21X!j;3~q_T(>Bq8Nh!>}zGpVp|xQdsS*ALcU$8
zAX@UI*3xN3ff6`UFz_W#4-$VQWBbwci>vDzONKQ&gk*i@60uWUOF{`>9mPlEVf3Z5
zMY^!%6rxB_``J^Ar9kU=X6Kkw=Hv}+;!!^EjoQaq%o)GyVHbt%7VW1kZ&4W3XocYq
zFeeY*-*UZA%%XIbzl=2h$)q-)%sWq6;Q$hu2Ip243ETolnw-q-I68k78n69|L&zw&
z1)H*oAU@-m=eKXYJKBR;6_o@=69p<?x5S7Xxoc;Gu=IY1#i4dg8phd<r*!qqTz0wO
z{^6fLf3JYZnkM-Z)FReg7kKmvd+0?5l>Y2s^rOnDck2dCF|x9MdrIU@gMWI#M6dOk
zP?r~wDltoxxOUsb^M!x5+zk>)X9Nvk=&Kby+ew@wh@0%k4?H>`_jqB9_Dqq$_77L|
z73FHh)}c}+qk4g#4DE%{uGRy+>0vC^=&MCSAhKS7%)oyytC$wpc>Z}{<#e-)FY}-y
zVk=8<V}iFZ59K=2xooZ{Bu?J4@Wg9bOty(vJ20*8a6)_tm>PdmNp=`B(Y_=pNdJEP
zOBDrVf9F~xZnAiCr%wj@ufiX}gR1VrHe@KX;qpeBhItfyr~DxrIWC)R^LVo0!F|Le
zeqxQZpvY9C!r^UuF%^@GWu~UKK%8kcpOa?Tn;%8gU|CT_&hUTv9?z*Tka4=tT%Zc2
z!sASwqIPl_98rIk#2!Y#h?A;0*AX$V_oGEjJRn{W-IAnES`6Th&u}wCit>X&ndN>M
zU3XMI335ML^v39v&Rvp~-5PPyi5@OOzluixw8l7Yq@t?mSoR8tRbP4gZFxZasL@3O
z`pqAw%WKK9_MU(H;KM<6uL^p4M2w)x2(`sN_v|)iduM+Q4O3>G>Q=m8ns5;iN8?9|
z@-u#OYllk9lECi=8)r?gc?s_yso_6Bwq>s(ixh)JKg;XGiKf{5C9;!T`o+<yyda_g
z4HP3Pp>aGO7B<a*s2W&_aJ_uZ<EidLCSq$)a&=;gJDU&(Rc4vihFaZ$RJM}>)d*Y;
zJ{&6>A~k=-MJ(8t^0dW+#Ws;C3%p%IJKr%M`PG+X{TYpM5p%+G?Vo~#WY_0eT5bdM
zAdcBqOB)7n@aj>?&=r*G-(b9Z4CLT|Z<8P5Nc6pJwWA4r7AqKRD<Suq8{2^ts3|Kl
zns6SFYvWQD(-omk1L9E8P}aQsKM=^uZ~EYq+46q~Q?6uYUS{d*8+>07(O9B?6^ax&
zd9{K~;1`AYGjP017Mc>L&-I0L-OJ`UViEVsIxDMm4LdIZyI7A4EZn!Ld1~FDr%eno
z`vb^|<K@lW;U+)twFikDh!9s%u4-vNwuDLT(ymr9*L2q5)^dlF81U=UMFS&88)gKN
z2^xPx@yW$gM`(mKR&F}%^w|9nHArZiGbu+<{7H4#UT4u<D<}yqQCN%8SS00sY?K%`
z6g3+?+_Cd-CIo*YB<E=&;V__DMSaE^a28ELGTS0XYsDy<ZMhZ?!AtdDcLcxbHg=EL
zMOF5v&jW>C2(dqi^!BPvIS}5~{Pl=*6o7xzLRHdXbzeJt5>J^)%z$Jk4P}U*0_#i?
zIE%1c;E9iR{xohJfT5(l7Ps#e<J$RjwjYDTfGwjlR_C##Y}Fgko$^ci+*>r*tMX{%
ze3BOc8`=^4)x!ZR0|e}An;U~bQK9D^qJabETzO*uIW-(8F=oYY)T)b@S(Q(4#uk5h
ze$O}-b$GPuJSskc30JyH{&p8-k#NPmDFD$WY`D4KG@9Ufj>(yXOwR~TZVdGWaueVz
zvL{L{=<vJAVv})U%-2t*g7ck44*sOw@S6*8rw3?;IxkM>Cr3*Q=-L>1_egW3y&m>4
zwka<~>#0h4GE9dkfib6aq=lQ>gB*X37l0df2%}gkb%KJa=oun;iD1Zj{tn2nc}_@5
zB4tlQe!pdGP)v=o5&q9%W-8b4&?77d6@^91*SqNxOJ5z8g0>B11xo<`tE*~~PWrg_
zqk7Ws*2f3=XE5TgFP|`eyR!Pa6An7~^J{N|tEVIQ-l{YZfDc-atM|gfWe0yglHGCS
za2P3FsSyL|ydpaj<Gw1rLunOcMGOk5r<Cp%N&*pP9=)C4^v9|7z`RM(W8ULZEP`Q&
zBvbFf+u!tD#X{=C0%wht7_akU?xdzn@lx+dew8q11of%z4wyD!h8M&thQ)8I9mG+F
z6ujNlJuX0%B!aF+K<=fpx&ePb0?qaDXK%aWgI?s;CQM2L7D#Sx@%v6)`M^M;MU5Cn
z;7YN;5R2R=f0ZAGpI!L8Py)eKrZ)$@PjCGsRl#BUvEyG<WQYyf&9um<gH=1N7ObSp
z-Kxq|BQTXR1{2AgzTMXV;d1AQV3uyeuuYYxm(WYq6X!76_o|oHZ*G6vd@gM)@RMs!
z<SBooofC(KZ48FtnvoIdUoL;8b4saJ@)UgUQQe+#dpTdZZKBco$lPw_V7(qBR28nL
zXSKbo`%#rbS>A!q1#pgZsK!jwSX&U0ft(GTW-N?jJ8d#AL4>e(|I}n}(W)#8l<EF9
z8!<`82@x?bSpOZ3t)+imc7sB@H-|bIx21BVV96%D@L&tTqy;z7pVgv?2A%-dL+GJy
zoLIhUuHfEUkK`+4P7Q_B7=5dSL-phzeob3&Zx}19ysitzmy|V2w?7Pza-*?UDL`lB
zPn_ZGKst*mhC=Gt^1;eM3Ok(XRm~=yt(|@Fv$q%zDp)bKpkaRiohH7It@O^0UH!Hc
zr0e|`CYwgfHKt+}Nr{(VVnnqp5(vC@Mao8!R;f%5Jtt@QOJ_;v#SF*QC?K!+uG|7D
zgp~6qQEG~ymS4#)COg=!#+nN?C9Q`)MsNikD85N;l)6c_C&}?Naw0kvAajULNnRlv
zUH4-zE)#emPC9>c&iU3MgQuvKGftYfa*_%;He=(C(!|$?zGtqc{JeQB?5*Jq$a%Z9
zQ3bDbDlE|CQv?Y4T)?T?7wmf?zp&3^+`j>~kL}&#A%%WUtzA@<cFe@s%qI(^p`n{u
zFz7BxJ%OIkv_Q8its;U!(Tj}c>+_dMEBeO>?3t9Cnf!k+6!O$bp2S*9WwwrHRHDez
zX!k<N#eyC5=U8!rVI$a4RMF+ThmBX&pFUWl3u0}WFYNaXDr@aQu-u{1!#_pIoc-2P
z{5MB}bjG3$ak1+j4Gpr94EnB^;E_VeYwxyiq7263vMBaYH2EACh<xL2?+<sM{U=(Y
ztPEAqpHzPftcE}d$~T$>?ZAz)0yUx$i-yX`z;-&CbYnGt6pw+DQJu}jtcPOh^Bwsw
zmMTZik}0Q>_DC}DctgC8gF&2|2Mpf*CMmg5C!20n!9B}mcY33qcE)jLHhZqZp4rL7
zlyaur<g1p18{)Cx-xMS5s@k4A66aP1S`}QSpKO0o9^CNv9k%TwT=Dfj2cd7wcb3$G
zZ)4`SPF!}9QyYOKtX5nl7!~7HTVXi)oS>H<WOM#KMx_7-T~5ofC|AB3h+bB_;74Q2
zok{|JnfnflEwhn8C|L#hQcYX?kLvzxc|{Nw4UEd5DR`7pTGSqlKhN(gLBZ9C{B%~k
zTKRt@vi>mPyA%Z=N+Gf>fLtgM<Qt|X&mrDq^|9Jp)eM1Gq;k8Vf-_%4O?Y4*(DlqR
zI}##yt%I~D6xS9;L|B7vk84^(+KiX8bE<a&TC6rzx9u0aca|9qXU#5@0?5D)lYM3)
z3F<N3RGh7lSQ*^Y@=m?+sXZF5?`XjbaNvJ5qf3VN+pLFdF|{4mL;OAyZ;RK#OWFEH
zU&6Z3z;J6OQV=_nk8@v+>aElg!8_@w7rq5cSVUuo9GlVn;faiAPV3{12rNHGTLM3t
z4wVD3cf-tApGho-N)mq4U>n?2WKWHqnT2o9^yw*^@+AAF(N6#&J72tCr@J92;7)&x
zQ$ZLP-n|+h3N`zj)N(frvub$bn}=TG<p6?jkTv@#d94f{*>*r<-h5=3)5I^Vprjix
za?xE7Xyr#4!&5u7xocg7${(v-Q2zRG-&zuDsJlqygC(l8;-z!A1pIiD<HVh*wO5cu
z@<TPsLZc_}y&dvn|6SABUE3hbHeY}BB9O5jk+B5coV5T)q~g^8^RNkmG?gEb;WLd-
zOZhB`6}K2)2?_EpXF5H@d91JEZCRRBvpgA?)m;;Zr`(IGSa(OO4o?%1iLQBi7#7rK
z;tn3sAY(SY4Qd=EqE*C#^QwzGK0em|TS|4Mr39-k)#DR3d{sxAS8iZ54@G~}^@C9c
z_7}3-pmantqe#@*s@i0`!@Zh?aXFg3hrR9Z_OZ!haQL9cG`62Dhx_ELZ3u)$e&pIR
zUZSaXbM<dyK0Vks-X*sNcN}*ex}ZepNeJL<=IMRE6O&tWDw)&gwSN9Q;ZW@g*D~!|
zbcnY;6MIW4&0c;|>S*6F$RK||7B({;>$W@-FKCW-zNc%tm~^%2Sswma>OeGBe57s!
zvpyBv{RZD_Xuw*xR9Df1=M8;~cvCR@{^@qE_g2G;G>y)17F4Xx{qeS*CFzJ=t*NgF
zQJvfsBP5Pp79Z-A8chaMIkaUuGoh(Q%yk1<I#p^RbwoS&+)(F^0)c;;huNb;KT_lb
zj`!;EkjE&8YOKa9Szs8ES-*aKa1f$m;gy*cLbnQPRN^x*LwY9^99<NWdulu*9T+M;
zwQ_S-J^nKW6mp)#hTn>ml*#aDm#&lR*~Oqo+zU6g=kc%Bpy>SXc$|XbJw~P|Ll}FJ
zVEE`RU9=O?qnyjnmb-tuM|bvs0b0@~by6MALhcC2pvShC)h)?Qu3QHifcl@m2_obq
zaM3ML3op10pgl)rC#@fQ`>9ayEYxW7Znh|rVWi_cBp8yN)N!9E)Ncim*t-`Kzi^{8
zq#OGkHI!_~OGo{$^_C*#P_CKHi`mM-)cXe+8famFr&&R!_~U;TW6MHMWr(p658!+V
zbeVG9<p#-eg-LOXE}&UOT93Dpw#lPl^QQ+j5vt?|udEoorN2wWc%XR|hUd8Dn2+ih
z>JHrC;JrLZp$UaFMZG8P8}t)JUZQO8pEcN4Zz<U>nHKqbCL%M8+WBS^L~jf7-wA%`
z%*iebkb_8`AVYr$iCRd<^NCF{afzK#y0HBIEcf7I_J+t74mLcHtLv>F9HG&KDQai3
zO%z=7S#xe_G#-+cs@qAJDe>)FUUuBC@c1L@Ix*Jj(5SS7(I|+mBOs^RLnq5EAxssi
zP}fBdt^UM402U2kT$rqM)#<=h&~J-d)By>FJfYiiOJ0BbHPz?lWL4OKrDL>NjZG8D
z6{MZU7RvejtIXVTPvm;1v|WjcAl)pMsS%nhI2abIyLQt+RNEZ$PX<DGhUDml`xwEK
z^(ix^29KlpR6|afHW<#X3`A#as3S+o?}nc>K}fSiD}mbV6(6!AcylALLpkF0wBNh1
zsXw;!hy8zXWAV1Gt5v#8;=XstEfn_oZ*L(8ofh*I2LTaLOpc<qWaEzdOb+0}#In*+
z&2O~{!qd9djf-RW(ZU1xgG9Ax!Xlnt3HbXSi>wbBcrj7*B1p#}%SJ|iKe1!e5G9(u
z>tx})FaYRPb=!uxSZS_h;D<40o=Z02#KVR%7)O5u3Fp90fYZPIVy-YUUIj2{93h@X
zHt|`6Juk!GO&m5QJgu}%Il6#Q;q<1Z>d-&MarzA7v#ipvil1)@FFtNeOu|uqN3#lU
zo7H7u2)@H1eh+&{&J6IHT1Jf1Fy>HxoQlrHD}6+w|IKPwfkm92nN3G<uvW_rn`lwQ
zNKSuTK<d7d#MVdFu(yjYv6Zb~#rroL-I~}138KiSzvwA*ekcRu0S9~(O*xQZswlmP
zN4Be>^1z=wAb*f&2RZc{!3~F2%N>y_ae7ew-E92wR>%^jW2VY!XN>VpI^m39$h+$l
z#|__OGHn1%K(oJ$Ow03<imB!q>o&gQGt;UJ+J{4bvfL8vR>77K4LYBwG*4YzUUG>j
z$yO%(tPe^W2|r+&yBfC-)W(fi-=HU`$J)J*;0$8`n?15;M=ykIq(g=7(%_chSwPEZ
z1G@%OyRY#Uzm!f~4mwRm*vP0LAKbQE>J~d@I-Px;n8$>H8OgBylW8wfWZ3uD3*o*p
ze!8Q7!tnPhy?id}@ol6h8qsA%iD>6Sg+ddbfJ>Nxd+?yzR;0Qp(PyC~$o}m+^38Ln
z?nEFet+t}0alLSdM9kthkejShgSW-mD;u3;JL1yiWik#grQl86EI8pm)*WuJstIU_
z%nXaZ6B4o7N==wbQ;Y|s6SZYx=ef{qw4X$O0sWMBFWVwQec5%a!qV~ACnL#i{AR-i
zwvsV}R+NZi4mb+!LEi{*Lc9yJ-e;6(Ys}sk;B8L~@*C|Cpyb3hSZJZ<;2L<cCBN0!
zqHB2t6xu-&Fd8N>>CxRLt%)8+H)T7WmxYvh5dWyzTrTfo{oxWN(36R!tYe7iQkpk^
zSqtaj96vbK&+<v+lki0R;h{m&QOcfpqwGm1EIoo?sQ?2KuhOxn=BWe1&@qv^m@y*i
z(`6UUMZ3U_*Lrg(Reh?K`#CVpx+~nqG$r`5sNe8Vh^bK&I#8{=DlW^d64o)G2GB|S
zlp-<pe^H_p0w!N|V!iE!GYdB(Bb3U28Pg#-IQ0XzT^AJJT!gcQ37&5itRjFB8aO63
zG-Lb3bcDqB(JGTzNg~KYH0-U>54c481mF6`rvQ4#P5}VB9+oa}Mu{N#LM5hK>%Vi^
zOugUg+dm9q7t`{8Jn-7-&mEVMX*96o-8<(r>3e7z!k#dZ<QwLM(41{4Jb6@qD~!7S
zdEXm_>~78hNvBynG=<^KxC&YEJCkxr!&yb*ivGs#N@1NH<D*oI`v($h6=J;qfo-z0
z%Ue=A{pK@aX4Fclf-SG9VX!nM2^J?H%fTb0y~hEwU2rJF_oG0o7~#;Wg&~j$*`(^a
zcz8{|S@JJxb7W{=ey`SkH6odRbs$+eTpz_icdtfTh}3`lb!2S0Ey+qep+1a21+%m|
zQsp^f7Eu~r@l07v;`Bt#>2oHHfna!M#{XPGd6Fb!AEhTet+T8V)(n}38$-6U$;7|+
zqiO3!-+szfvK~r!wv!s0W--2@UVU95yk%(Bp|&M_z1D<GW{V`L$h2C2Oag+gOzsR+
z(4xp&1p=bwviN=^SB|0(r}C};g-q9ix$yYVXq4A`2;Do);D)_t%YZC7`!u408*|*6
zCg+c37qteC&654~Dw@-!Ldm6+-c&V#a5JCXr<{NHw#4<?(H~yCW#gui3{RS6;Neok
z*kAZ`QN>zrF0G2V{-(cwJn*&O3Yh(k>{2ogpz%(mCDVW8xhZNB3WYbHaqx8*wlf~8
z?G&tDlKkmDps`b@6Yd@`d?oM{7TG&5kE=rFQXULqd<ju&=_veJZM^B6<Ho5%ve4Fw
zno5&1+^%<4#PuRaH>WyARDra=Sh!(3<mlg&);pkgV`H9CqFg<H`lc6c!6-y)(}9VO
z+U`K#wp%0ERiN`IbUBw5`aMuk4Cy8MFY=WsZf#cFN&#`0rrbqkr5vbqi@Njq-@P)|
zy4R8^%Mbodt+?h3I9wGLqUcXVaj-~NTsxJ(s!GUKp~1lKhljfCW*-#}O!|M3S;dI{
z#4y1NPb4(>8O9WU*=gOBHX$Zh)|=znS8Yr-xpF`GGzs)CTm52Hp1~++C;>OKlcjH;
z$fVR>qCFf97~_LkFm{<ZCnxn43bxjifsp4cqUY^Y)M1u(hjz~El&`hWc|L)>z_1R<
zm1`<&rB-#Rbzs#(aMn2ooBcxuuT8`J^Xl~5{#i$@X58t22|H@LJ%l8mUr-kVagrjW
znh$~!))L*wWX30S9?te@1$acwuLK#jMtzRgRF|q%t6-+=4|gINgUEFjVZ;4T^eiw5
z_S7B<5B~GH754Rs19MYrC2l2@rUKqMv7VR8UID>OT1<9TxWMyae$YHiH<OwoH${}L
z#*g1Di#Viz=bdPDkHR(HvmZpCA8|Ttnt0VDim{k{OxBI2o$MJW##cnjJGSrOCiF|N
zR&e@AJ>vjMdP_XlkQ<EAxjHij5gM?0PNoaj>lra;OdIeWHpM)i_|Lww894Rp?}>W|
zv-%~vsKnWkSK0o5?U!KPM+LqqssHTixZ8*@%ModR014%tTAb4-t9|AV8^~X`E(yeW
zE7X87oq!Ngc9JKu@ESTu^Ghji7Bs|}l;~`h<vz*9+KUPmcXWkijsl;N=Z+>UP4!Nd
zpT5x}j#twDX2?!AN`H4YXYfl%TOL_Cn88+suL8qg7#2yW>DNHS-y1x(9kk-MU{$e~
zM|$#q6U}#}k+;N*8Ow5dH5?5t^o2+OXDp2QhA!x$xwqUbEgSOama2ijPzLqMI`OXx
zup4{>CHQ;V;VC+$pPQfR9Xd%kaUq_~u6^kQ@p59nD;qAl1vU+^+biY~h*acA1L{^;
zC*-OP5}Z+g@gn{j$9!hp+KeE~U1xo$GGV2Ez|O|d=eKu9#@BkFIYn->vuQF{&)Zc@
zIndl&+#=v|v$lGj<$I3mn-I<96vmaXdhk<m`T+6SwEdM;<+s8JqP4I0abJEXHcY^T
zsEVNC#0%QRFeSkzPU<Lk(oX}q##gQ-NnXrgAJ^`=S1=GA{d}U8@UEs!tV_Jf5OhO-
za}|@r<(^)1QHe9IP=u_1k*HA(1w2Cc{K*5>`)x$6J|IS@LuLL<J>1MSxIFV5183CZ
zp^z<!cTMjj8hb~EuU`xD@ktN#?_z`>lu36FMG-&dcQt*fsHh5#`K9-bD7+nExKF=)
z=Yoi!s1fZeYQf)rz!=#N>`*AI7Tjuo_(qj_TD4tcv<Yb()wyz|6t(l^+qyDW`E|z5
zXgpouW`^)~<6e*S7^X5M6J~GOBAifDlW`@ld$i=S#=&5<iQZ>%f6r4^v7>;+pZ*Yw
zQ6s_lA*9@KZBEgp=aq8^0mJm)g5jFXqr<$yy)wyBZ1I1u4jNsE9qta#2wwJo#4imI
zThK(gjo(9iiw0l0g@MKt<Tn$a>mkG!C^Z{cgHU@)t6Ky4?f#Tp2e~)|rpfis35tu%
z8G)?gat<Fyy)n_;v=|n^`tR_BLFzqD-^EswaZ_Cct__8ns6;!O1ZU4V>$J|y6CMp3
z3j{&DQ$_QC3yK!&!sYOV)hXY9>kYT}FbUg*3nv#gpv-4@nO}%~U=T344!bGQ*0QSw
zuaY%(!E;&l@CWKU15c;;Ae+GhbEWz~uFIIS83KXNd}zk+_@$mFob<n@5lOaDrSaSK
zb1yXMAt{n}VoCV6FxneJE8`2pjVI>x(5;C)C1TK&#Eg~A52Lc`upL`}*9L?HD&(?K
zr@vcP)x>RARM1L*Ci>1KYQCI2cCu)(*3s~#Kl#pELRG4p5tO3%nhUTUmTlAXESO?k
zghHu-T$p~_R-j{aJp`BaoB*8l_@pIG6mDs*SsddSjFV=$lJN>XYxJnW;3am|g<-Se
zhff8UsyS9kG0^|kSg|~RNKHW&0>G4U8NGl|W5-L;#_aE-c(@*x4$tbyx><Ite0{Vn
zk=5P46^t=ccLqZ|L<dLpBe11tEa7>h%9$QVfW3E**hmR;H<DPFq(c=SHcf*>e;;5s
zycXUqh!6JZRXYxhG!+yS{Lw|?7q5zk8QMKF=z|x=T;$t#kJuD{^SwhoeMp?^ad^9e
z{SU+ILw$$Aj{?h*U}|MZ1_<A(RzoI<P1}asxnzh#Ey%*>SzZ$Btw45Zp#ux5;Y_sP
z$)dJwe}~*3N$*+JeR8;ZF?_c$@s4ml)b3D|Bi}Khk_moDNDNVpQC^*oiqdvhp$F-}
zKxhP8J@<OrQEG^PB%!ZqB~SY2YCT#5FQz&B6aN}E>gY2d2yzwu{gK`Era^CgFIgb3
zMdqNBJk3LOPh=jT2R?I7Q?qeD%3|j4>l$+iDLv!umE0*3mWfOP|NEQ<jmWzQ1*^1N
z%8FvPNRQrNzvGy)sB6Xhh57ud^cJ%m^~cO-q#R5>(0<>4CB{m{2h$`eVi$Zn%mr-Q
zoEIEcj*ZY<AcGfbZF|y-bbN?Iz9#lDSCGfPM3hD#(GDIhIEts9i+$nN(*wckw`B7}
z9Sx}o^T|yzB3|b~<OVFM$8m;HskoQlk2Pu0VtA24qZI8fdK)m}n}r1kg-nnZ6)teD
zwn~Q~7h^PkIhv<WeU<NOwKRTkK!;2GOx5}FGS$7uN^Cr%$I=_WP>>jr{sI<3XFo-V
z$1oi=#!Cr)saWL{6ks;hhkK6s^tnkzk`_>p9ItSKl;*kjo`Y?XMw^JtqWld3ow4Ic
ziNy7VA8(a8@b)t=;O&CudE0g~uP3pl$IP-$aF+Rhm(GxtsF?}kE=OP~*`&3x`)dVJ
z03eHK@oX0)IP>%`c4k{8yv%ED^t*EjzsA#O7%9CY(%a0@-ehBJ1-m5kGxQmEs^N^x
zvukw%caZ$<D3yQzN7s=ShkmeCe^!k_eA0Wm9-irj0@ToDx6k9x8NQ_$YQ4p<XYvp0
zAU+a*|CJU9DrzKT%R*mcy{X1xAun`($?4FF-y}FdsQHS7?a#kEE3JmP8@)jwbb5b$
z)_Q=3zYK-E$(}qKKCwzs50qH5k#2~qWxP8*els``oOeAN>~Y7~;dNiso_63Xy$e)b
zmK;F&H{RqA+MW!i=u;Z8IxIg(p%5&rIW8D~kM?Ka*m@pqW2Wh@w2P_{IWHN$9dv<d
zss4!(kW!EB^mK3atZvyR|E`eIGTpq_=Mao7Vx?8R%kWTzNq8WL;F8d8Hgb(D;Jb<=
zfc2q>`qPSK<IyO(5QiA<NP*|Oq>EDmAuU#4a_sNE4<`<S(lqK}*^trn^nE1Y>X2D~
zh>nx(oK!^+9Yvi%|6~R!vt$|uT!=HqSTxuE4iCFF_Ew+UVr11GFi{A<@aZfdtZ1hL
zg)-sQ&nm~X7kKLxwwrl{MNYW4t1x=q<Lu$FDWhJ+j9#+;oE8K8?V_4?_kGF5XA%xa
z_*Onp#(uQ(1NX#7Hv()~=i?7<{b<#HpCdG2g2}AjZ=7Qn4lv5Smrh2Yu*s^iZp0Zx
z9|sN3+!GIaC%C!6lMP*V9^cc18-tHVg(l65_UOypH~W&YB$aQ_Zbk3<b3WjU;&QRg
zKZ17KyoRdQ!gbY7?{cbRc0|q}SZk}VCW8f_Htc!h;<w%}423g0uM2>=WExq2G5{Cv
z^Ja@I>};-LyN0`cyVP>69V4&_nj!}45XlBP*PI9#(o`;h_Z2*5bV2!MlaBGQy;MQy
z0X-k)WiHDHJ7!B+8d09c0q8n+&Y%Q6a<KVKzazs@+Zuz%6Hp7FJxf|$V~hd4eh(o|
z0+O3n)IfAa|E-{toZhW`U*5xiS2GB5#3LJ}R7b@|y(SmoKrGSZP`J}Y@8%LEbU&RK
zc!q)H|0-ywBad!Y?--jpe*DK0{DKT}9LQ^l8Ycm9MBHi`quKrXx3JeoUy9(x2$IGo
zfBblh!z!lYd$2ty)@!oO+=MDq(|4B)h-qnW6DIX7_pKLfD#A*1Qt3Z`2?8&O?S~mq
zw)#^RemUEN{407791%q=v+_}4O}^nz*${@}`5bwVg2+AESd!%Rp)pAQgy@IUlMV$2
z%ECw)=WVot$PzN9o&n2n&90r6gKZD`FTUZwE&F!uzpY_ijd#JLBG}zN(L5F%LUswJ
zV;5`EGz(B}GDIu;KdoMW(C!{wH!i}JjSNotbJLLQKlq^Lg&yvLg<8$t!1@)3mBA6S
z4`B&fxx->y+@d7KF@n7kVcpzi3Gxbech(5h)XFQIY1s<LsfS;Pr<!``uSKN?%5@2w
z+BOUx?@z6mz~=w9fq+1D-(tgM3k8)N#E->MbT0hpKgB8~5@xi2Cr+Qbs%ma%%bkXS
z*1(y{PCeSIbSt#gs2%4R`qml7-nu0BL!4V2ZTxt>oku>Jm4PJIrxn}f5mit&@{*~N
z<(9v=$$@VGGDIZQm3A4Ei(5c^z98aRE`p-+&XtLmLGu!yMVXV>Ie1YVSk%;xyDBg#
zl6vm=6VSlV9&~qq7~;nBEe+e{`k-IRDv303OHO>x=5^^1+~Hn(=CcAbyi7O;2A|OL
z9hwp3Engl7twz;#DVN8SH5A(h-EPQ{;hVw0%C}q-1X%BmJtvL_j(V}b^$F)^8Fq5%
zY*K4Z>se(MY_-GTT@BR()w^@0B#EwS#S+ss3N3neK7ogSuc9?Rr<H_}=0e2k8*?1;
z9@a&}&f~J#9}k2g2Tmrw7BUyZxB}Aec4bBLF_PaK|5TOX_g*qRSzwY%>8H$8sl53J
zoi#3<f}xBGC-3I(y}i`(ZrvsgpkDRA0?PATbY<2OnBJ$|kMNcl<pw&h&J$uu-PIlq
zn1@QjZFO6J%iIsq+NpyRwI%yJG)$%5Qn{^RP2lIVN>y{>>RPx=S_R};gM?C6P6hop
z^o6k)QN({rMha7G{0w{Nx}J~G_GW|$=4gwkBR0j`kK=?00?;MyVc;tKLGy*Lt`)>~
z9i_B&H8eUF=@3+R(*Z;~8(fyU5w~wbGSzW7Q4mjm+@ZhFc+R^2Tq@N(vc7AX@4`CC
z2V=jg>%g40K6df=Ch(5Om+2$<X%wuhyM8!d0W$^}4wBj+=G||NxZW|x8B>}A&sY30
zr=;CskwmByVtAVO*{Up1kHn1ZY2jk4Q|XG0MzJQ`Dun!9#_(9e%9}Z-yY2|Lt96<w
z8E<!gL--|i7h{w4b4q(QGxqSP87{z}H97K!sI?d%%~b@P)5&R9F3G*mn`h4xl!pG~
z?OH(vaYqf7YIpEkxAy+1F5&3CG*glmVE<zq$JVLB2Xgm2Ixwc3RAe*&DpI1=0=^}<
z^u*5iHlw&l*Bn>!N15gi7EU9ZHBUP|lE9IFa{^DqZH3!A*|CQ)+iJ|2yj&P}pVr3h
zvd$aQwudspjVhNUD1V%`2RT?bbIp`#S`c_zi|FU7(!yP?{-q|sI*l-Wy8pmY?_9GA
zKWOh#%ndcQESd`9k5_+=HGcJSk6)y~)WJEeA2RCrN(Mss#%-~j_9K|Ao&f>*;W#mW
zm-P42slH=s2VA~0N59?ss|@q{;`%0Q-8Qh2A@4LG9Miq6wI1GV^_la<9FN6coGEA8
z?73!tTEs19|4B)IO`jpzEat~6!)L+}e4I#8W8D1S^Twr1vnE|edr6CV9x+WG<XBKT
zmOS6*vh`UvQJpYtJppOCF)Rg9h{Z2|+ELR^W3BS5PQt($Q${pSGdrmjr)}{(rO5|+
z<5#7$QjP{i%O_vhOV%|E|K5rTPivn|`S1bv(6I7&eXLqF<n=<^caLf~veWUsrnm?#
z6T2L*!0$tU*h&8|15u#s*ZE5G%$h7uK<A*7uW17WlxrsYI#)LqrWMUl6?SZYa7VHw
zVe4-@gDSc+vRW~30=%=6Gb?DypWC=Y9-m8}p@*P;M~eJBPIr!mc_o@Y>6qH|`ZdUy
z_rOBmJnS^`7b|Z{nwGCsu&sn2u{5<MjcOmAJW^?%FuuP5wO~gw5HE)ZQHtAqH`#Sk
zc0P7uGlTQ#-mLfu*3EeFhWVX;yFo5oF;iHi5&<8FwT%+wLnN{fPRJkQxvcA8O*A4o
zXnHUc$A1g1CFvsE!oF8c+IdN~#B~X@5}M?edHSa;L8yYaV-mZ_xr>#kOW2AsJTFyb
zvV(84hd_#S>z~91r9b}33Ev4nmwyIs)g?JupFmmGyDO*eJhWFS)<4jH6%4xhX4#xH
z_OY_|;^Dx@O8wzD#RA1cU2cl`;bN{(J9QCt$2&y0z`I!P&9!>q=%J5rJTZS$a+Bn@
zP@sG7yN<>FEV1piI=4MP4%5}mAX5hBJ*gmzbpO4ovt930F$`%H2t7s)m3@ht5SGFx
zu?+1$v^<=iQ`-TqsywKFx36pNCUKlD<brzLW88k4l8M~-K$uURM{RoC>}yoQE{>ZT
zYJ+7wJBE5=<tyiR?tc_1;g;I;R)$@J`pcM9^GDL*q*TC*P=G|w5prstQT8Q2UNAK~
z9p#h<wj7n{5Sh(#6oyFXp{_D&elsaFq<;J_uA7CCmuAHQ(vA{;nuJ}#fH|Rth2y!h
zlrT^}L9AnMx0gE7gWbYpBDk>~)XTMx_L09$0kMiJY|Hghw5SB2ATMF+q92brU8Nf;
z__^t;t5^H`@Qc%A9fp!7Kg0|%%qZA(2MB}Z_~)eNxtu-lFn<f5O7|o!sXKLpxHj`t
z$K{Yv3gSz4>*a%gcU>+yg)-`USV=uncrIoGBn~Ubp5`UFFQZU4JkbqKR3Qg*U0R#z
zC+E3GLV3w471~DoRvLDhuw0L5F!8{^gCT0%Pbj30y>vb20f3wD8$J+D=QNG1w>YNw
ztE9GvXj&2h(ctOi;9aqYrM~KR2fx9bx-Ca(3i)j!1ekt*6jm@?IYJQaAYP@?MKEM2
ztGGnE&Qkz}eNTaGEo5Sl(L@TAtOMdd3=UmP&^Ue7a<LLdtsAH>8!?OW!-nS!8;exB
z|F{a;BjE}(8e^nz6W(@jC{XzyKmkEnzLDN-6<va9#*x8BpLC{M3{B}7PL;nTTvVSW
ziC&fWI-xLsSags>-|Jaf`w~-E(@y?D2@CJyupJ7~I_dRqa>r-ylTc{bIg+S-9LIU4
z-oTDzX+QF=|J!ylRYYjN3@+d55M^17C$S-aBhc=g;3C2M?(YqQOZ>*tPCVAH-f;mP
zUXt!~#u@$bhIxu7CPqwZ&{eOkAEQzrD5eaulo*PCK0ZDE!^jidl#UkWVNyIOIHn7s
zxOlrqmH8^}yw$Xi5cI{2;iC8`JauTUXj_JcL6HoZ$P(rV4~ZSuQ$^%56gb9K#N!^f
zl3ZV9y@+D|Uf@AiP=xlvcC}_NN1Rj{^mCv)F4dCz6!)(#^DW1RyA3@}yoat>mroV=
zG(%{A>n0s*bO+S#AjjOk$|fBydLLuU=;ZEj_4f=lSwxzRjGO$Q3-AF#a{*(IhJ89x
zp>}A90el`iBz1o;C0pDGr6F7L!fy|D7;na431$(RXiC|wUB|!scLc0)QkHSR^dezH
z^%9`;7JBit_&cyvtgBv=JBT?~!bT}40Tb(gf08l9xwP}p??qYIv=XhlL<_Fj@=;Tr
zKAmV_5h=1%RH~O}q_l~*`A@VhCKT!;5_@?NU<TQM+Q)E)BZ!i+mp~$6Ijr~Pr!iTh
znIHG*f0D$*#UD#$y3ya1kSeT(g|}Rn>29`e2$><z(vwM<pJl#;5%mM!sZ&q?x)B+F
zt2q{u@GD8vjg%7YagPw3L3v^Nry+-Vw#h~RIPzRmf4!s%+J~_M{dNJ@^{dW!YC$?|
zEy^m<PnukXBBVH&ptwyQBRyKa$XX8+2fCkJs93lUwSG|ft$0a{6`&MANaofvXYu4b
zcd<6@>s~EBuNCJu)%(!oE(W?OW3O6&+ZYx0*de1nccO~<sO?R9U`j#T+3{;nTFSq3
zAAIr+`F=A39T&EseY~mOP3tWETW}Yu{fG5$Ze-bI+0qf80wN<4&`PfMnX-ExV7RFv
z?QZYJO|TUzr5U=zNxL4KOc=@ZpZ8!-K!JP?+NErmJ<=7LgV=6+a}_8g*lehOz{(H#
zB#m};*O0Sekj)+8Mg|hkoU5?W`nPKN<a3|FjU&wqfJ|L1dimmU!0?Tx?LFN{bMx9;
z%#WRe7gHh<9X^<j3|>UG^8{huiBT4mj01I_*PJl_cBqjsV1Q8~j06G=`L_5`+3t*&
zcPy-`7#Ql24btZML8syo*{9Zj$x1#0ck0f<8UlfzrKWZfOIJTC)|TegC;<5it5s2I
z!qRvr08H>qm1^#47Mqv+r?QAw55RD@ZDAh9tc&89bs?nP*hw(pdwKkD*n$Iu;sCfB
zx;^z@cvR1WFa|3v0+2LZh_*(NaX)|0es1{m`l*2#3u!)*tXAJs9xO0_1jfH!xVq}_
z0-|ldXNt1?I`-(z!X;*;7Nimv(VHaI$Byc$3rcO=XaMe@)PxwSH7lZ4%QSx}eVJoz
z@SyB_YK#aPhY%xS?UNK$z(fom(eS4lwOa$IHAA61MrGX#^b=nyRAzWX*veLd+pmoR
zaewGS7ZwiS-Qv0L2}V7Cmw?`bDxh<d9y39X@Q>mrlxAPmjE=Dzb}=)c?d&hvCjP90
z(jmG*NIugZw8!VHH*9#FZFgn}L|?-`KVs3+8m=-|-hb|<3Fg!9-CUT)H35UCK{8ak
ztlKO=LTIYPtqPX#U~Voc-ZS>DOm&F04+>sf=u7ZIMLsuV__<PlBt9z3Z;AJIiub)e
zm~F={DL=u|W7&-eSa@yTYHdHsTz;%0gqD~R**%Y(Xq1@>n{!I~xHG<c>i><&=7%pG
zOPsNk)D=%UiWe?->vzKJBbTArDl+U<c?!Z$B@u?M_$Aed*<9OJduh*tu|RSsVTodf
z&Sm0!#{;+p3&_NOllT*WgRUBM+==3|Ps<VQg0?k}t(HLJcJ8~iCFV#yGi}Fai`0c*
zbLNoRK4;&>dXv>LWN{WO^$UE+53bO9_Vw5BnQ`yKFb6J<MTC#B+@YlR5P?b8-^?n4
zk4F?JhsiK4xj&EklI<7Ld5@mSw1jLneT+>yI5C_;;{17k>+bgc@+R6Wk!X@BPRfc`
z-2)2^{ut57aa5f!9bT8pctS=*d6+U}yPyF-zFtCebqBp?0|X^CN$sbe$HNFIZDiA3
zF!aTLfo+~JxY%<C8(pS-GWsRG4R4KR?}UT+o<H+up?d6w353yBwMx9}!GMGZ{#X?7
zxPdJyFAQRTN-&Y9REuLVTfv);v+`e5hLPv^mZZq}E{tDj5t_L;`caDGVQ1W}c_1`#
zeBnLv`}k}Wx3?tmuNp-}LVXI%Dmigr@*&;!g09L>zD(0n$o6E_wb}fzczl@LjT}dV
z1rvCR7BJSxzD<7gzKC2crzyyi*E9Tqx9p|R+Tfyp-R-pL;%;WpFkgeJG&2=zsDdTl
zyw)kEmm{T!K>BzbeCBYi@R>0_H5R$|Sa{H;nI6a7Q!c`QOlWu-qv`yO5UPQYP&@*I
zc_5+>rY=$vDbRQ3ww!iV?^koLo}!~H+vyY4=O0j(;Jml9(aRi1e^NHGPX)G7yX83R
zMWzCO12w~^1HCm3k94m&e;ZSdq&EZ|H+(u*%PSS_OG~$lD6;!|+d#j`P1$7WE>T5o
zqvpSy#ZdJCYc&Uu+SkQG?;QRCtT!j64SkZIOW2&dYI5Va8F)2zx&<#>zng{L^uTBK
z<M{AbSUgTY>{J}?dx$TxY}jmWv$*~TUqF@ta+iS}8WWc<wj3n`G&L}np+_4O1u`%)
zIW?Doe;Oy3m$n=Yf3|Jg&K=vfZQDszY+DuERwb!aoK*bP^S*uh^f+JlpYHSH8spjv
zb78K%#@G}j%4!TE=8mR7Nk<1a1{OwUUVx&ty{Wsas-wN4Bab|TD$vp$@K??OM?oRx
z0yJ^6c61OoaRc%Kw1DOSaiAH1l@-9k!@~nd0T6R^@^-Pdf3$J~P;01a(a_P+|EJ_1
z6M(7re`x-Cx>{R004V?dcmVAjo$P@QZhs;EuN~EZK!BSS5MW_#2Ly;IDeK56N&%>)
z6g2=+KnI|Ui5)=M-PF$73?Of926S)*(f}+RT>y6fJ^;)d9n7u&Db1DfFP4Zazy#px
z1T?e$>jv~Pe*-%GqoM~m0bT5^U0whF1FT&EmM$g^Zhw2=<_NHMFtc+v|0jXJatp_Q
zCFJDd_}9YzukJ6HvZJe;tC@?nlN;b~RAq6=f9L6DW#aZvY**{QYJj7~UrTdGGxvW=
z`j_r6%wMgWiM4|(zzyi-_D?KRAi&(()ydAp`)};Oe_&27*8fVw-PPK`@;@`62e<$&
zO<c_FfUd58Vg6$O)1CjU6Y#%OZ{p-+=lw5l$A4M<&m64XT!D5LjBqThe`A`t{f%vD
z?EuI0&#cHeSU3V$nEx#|cX#>^od?k6Uk#%EXJ%;rCShXk=wRm!Fb7(|F)2E_{S67A
z{=Z|HfAPPM<o^bW|62(DZ=v`9jokmr(f@LZ|6lL(e}$HGx3g0;vHv>&|K2fxze~o%
z0q}Rv0OSGxEE+o#m;axSiM_R*_y5W3zglYn|DEpt0Zhit<ZoL<94!A*!_3J1Z=toT
zq_r2&T-n;q%nD#(V)wUa|B`Dsm;+txtQ~-Vf4TivTLBC#%*_8qr*36!X6x_|2^{}c
z0UgZ$YXyJl`d9u;(z-G#G8%OMn+5wXL*>7N;HK{F^f$NvQH_G5`Tsop10yQx=mqd)
zVBzEjFtBp7{JnpF6XIcI_y6C7{1*_*f1VXg++3`^0D6DxW@h<U_5b7h&y&G_fst@9
ze{(edXLQuuOdQPrPTT(!{G&E=cX9bk^1l|q->U!P{I9tO0=<A{a2u<RW_)2bS-IJ6
z4cLWa)8cwV11#WU;ZBv!>SZNCb&makSyB5u-;Ceu!x>v<c|SY7S1z1Bj%8_2CI+zV
zsQNR2zl#a{$!G>kP<JT1x#&(pm<%fsf3q*O;(V`{<@X?TIhi$24yIL94J+QE2xq(5
zTwt!BX@bZ+N`f%RpPY~a%rcu2F?K+qaWFD@xle?C7^jZBV@jrQL{4j$vv6(qt^
z@R?u-q+gJvxp^DAJT#kqLcV%%xqe^CYuBnnP1W%9!GBwICnkSP*hgo{zG5S?fBX^w
zJz^GKHBFTB{2dtqpTJs{B1ljCLu^i6*08Z>RpxS?bMse|Aa9m!?mAC`HT-Ey?5b22
zSsBOfC#oj3u=L1su5LuKxwIxTH%hjOj%KXD92a{do)CqEAX2PwL(pVR+JSI+ykWPS
zKM{;TA4siw)^KC~r82{qfHgQne>}^nXUu~$KeMo~j!gvZW!w|=t7<iSg3uv99$3as
z3zlfjjLrAOu?<LQp{&+~UetO|x!*?=Y_GVHsM0(3EyTQWGKCcPe)s8y?frVaeOVG5
zA&5&t#&c&wEpinHFy75$I?{btAXRd(f(eR60qp6zmmf&WruCyOE%bV^e?Rs2zv1U~
ziDu;X$8a8|@e$?%p03%k4k<l^G>S8RAVzq_*di?@j5eUjPJ%ufF<CO(oy*o~!!`l<
zx^-i5^emeG6vs%|gC&}-9|X$3&ul}By@6>4f~9@)+%D>LilwA*I)e4{@W@?BDk`kS
zBs=OQTKNcbA87SW2gTM|e<;^?YR-1rp7Vgp+sa6?Dl@g6Fsmcf%76exSDz8=RcV#Q
zul6eR@KU$keUi^pnQ{cmQ)tEU$@$Us*Gid_ebXKdlx`}Q&lj)dut%{Ej;JyU_iA+2
zM0xV?3U$I@c`Q+869G{MwDCM>qlhEli(Hjy{i!8Ihm2vf1~pg3e=!OxI8+ga0y8gC
z+zf!|sK0tjQDP{neGxzY&F;YFYUJQ$jm3TtC*pmVxMe9gXhR1p_)DeC1cmwQV0o}u
z!a=YW5+8I@(V*a5K3xAEzvD#wH1h7Sx^O*0cUiHMMhs~wXCTW^R-c;@tA@d0n2*Qt
z+?<7cTRU(Uc4^)6fBbq^W+-$={DZAQuIyswD!BP<>3JYuQ`a2rJM()&a9<QjH8_s4
zRxa4=r-^=oAE`-Yfa@bSqjnt@h*<gVd!-ao1zplUC34MUwU3K>LozLn&n(S%o4v9;
ze@Bj)Z>hDc8Z2C#d?h#AT_3`)iLahgm$|KkGoUp11FmeCe=J+u7yeE1IpAQE(KZD1
z^SF5Hv261V9CRFZsaXLVCR~6yx}!$(v9eN5_y>YQ>?jt~H`h89g*+`w4VRyTFKp;<
z3k_;ukG_3fC1;MmKZ`o@LRV0+<kVV%ds2{vSMo9G8Xqmi#Dj<t$Il4LcNNK}y#am}
z^KB`gegqV;e`o4ZcSZdx6GWrEkl|VO2nG`FA?LHg?NYvxuRp3NFK?xKw+P3eqjtxf
z2eTLAXXwjfB-Jwbs|=&Xq2-kcg2k-Vpq6jhzymMDa}fzzzVXdQLzJSvh`@%esh27j
zs;uWIJ`24D7V=H6Bi6vN?;pUTwMRQa8hcZ?hj7jJfAXb#&#-PR-o7ooiBJlwPVM)*
zvmnIh34R+r{unNAhhNK`<rsT^htvH50wu;bFbIGa%-E6r=?F$@{QikrxD3?iyMaHB
zlY&Q#I3ub5fq{J5&k|~{mAHAF?`&!3d3iXWgmv)Ed2K9sVn9>>E7x@`8$fipdv|LO
zG<Pw5e=!c((HDQ^NnPqCL>a=q&}`@97LD`p-)P5V9=*XoTr?qgF_)*le`H$6I5y*m
zciQzG>-wDG@|89d0EKSQU6i6-1=8XzA@J1BT3_z+);5m5Z&qZy=JE@gRgP><LLaqG
z`Ea2t>+7!Y$FDoX%`|)2Bo{VoPE)2&k-7_)f8TM`;g^oxs|VgmLIH`iBR`|^){_Nx
z3;7^G^Y+)s2cJ}NDSDOGF_jz@%#mMY;o-poq=j;NQKQo6_Z}i{OV~gD)Lv@P1q7hf
zXay{-=zzjqNJGQE5}PZXwEtlD*2xKQil5;t<+WunM^508p#WwW^7*;-#7W;@<QPl;
ze>BO&u;~%Ej4GYr3VF_)p-MPt7dG{z6vPH6w;eTkEi&FsT_nE}Z0uK$O6f5VvQtcs
zC~5MT|6Jx}CTT5~VBQ661zRmE;z~`Qy%J75r}*d)T4s%ww4#O&pUJg-_o98ei7o@k
znR(MhU{5>+a3{eHv6#ShdJ7}I-Z@QTe|e8$Eke<y<-{lV>c`>1-JS_!MAte5gLb0u
zw>~0{3T)ps)4q~~M4#Wb9v0~9(jr@ug=@2Okb6y`Rs=y=8KVE`tW$TT$AOoTxyj<`
z<Y;`Z7QbEY4J(J1(`E&mmi;v)mw*H7KoR@pr0yXOpJjd~c^e7NN@QTj6jy#YfA5Q{
z<^5E&TFHXyh(}o#_J)|vvRCu@$R~2*0CLaBqe@8t23g<jNc;Nr+i&oLr|AGDjF@LT
z{>2a3(`NGlgRI3*gHXD59zAPw!SX=|ae5fVMprCCCe73<6x*Adx2rD5mk+#kYq`^v
zq>zCSmoqDR$K_b^SNZ`LXkdT%e*=_K@SWe5QZjZQc}%=6kkoLygfUd$CB95EtBm@p
zzDhWJn2aagG(xu~B;wLwOrX|^r!|w?0aAMrwMSyr0m-<^E+ZVbqZx9gE1-v|)~^*r
zLPuSU*&+_l*K<45w5z@a7jYV|&Ju<^L=ohMY}D;w0Q-nj`B15Kxz-7<e`7-kc>~9?
zmxLzE^2Oquf0g7HIO{7!NkwuUJa#_eDyN9WixgH{64m;A&$|U_QN-3cN!H02g>^oB
zfA#LZJwQWFAw*jQ1`t=G;0`--!F)zEq6DfGx8a*`dCSVQ(uf6;9;T7Q2kCYwMpJih
zKx6K1ulzd(dGnwz8TSymf0dq{>)JAqt5qhPV6xkIK)&ziI=mx>=uCu1!Uy>#dz
z8oJ)h@jK@A(;}F4#e*at+7QM@fz>Vg=$Ib^1-C+iZ$UGm_u+`K8wqz7)#2+z<kn4$
zSouzmR;ITP@dBt=*}2bJ?tE{c5E>*ph)IPuo&qzDpz~+Yijyb|e>d6DuPu8s)wW)B
z5#C(eu5W3T_Gq=KU)e8Wf`VgR&31FaR&#Cga+q1J-qUD>c7E6<dj|d^4zL*~+96rZ
zG!0~s_A7;Q*X=1I$3jKvs{#+X4itY#I<s4C_H}Tv*eyMWv<hahTdM*wAs?)4OKO%N
z^jk2dfivf^DdUcSf9gJ2(e?mn-#n0u(qt*v;deYNuK=?g`oN7+#Dc_ELfs_9$+?wb
z++XZk2<_TkN2Q`0oTadsTPs+wq&0UUZUY#h4vjA46~TjRKh8c8p{ROwZ$ctQ3n+zu
z>=m$WS#aX|r|UwqVnpYhnP_5|tInI18GnpcG2C2y=B&q%e<vCFw{Hh!4s0~Mu;g68
z)%(;fSLvr_rb+x<IH~SkYE^J!g~RvundRoPZ)kebG-vIoMksst677T{GZykGIVriJ
z3Dyc>5WQ#bc=?0eAkW#qi`0nYhmTncWdTm04{O@h$UyG#l#!auJla}#25-WOB!SXL
zv#*3sX4f+wf0LNni#nt<EhizYi(X+^=7-)#90ILrI*iq_i=h<V*H$yvp@c{sXSp2p
zdKKUQmC~0(-nT4vz9mF>(Z+kO`0HA%j?hX%n{tX*dPH;Wyh^4zgx*tGP_GcI@C?){
zl&G9>U}!`Ni|WmY0lfE*i|zD*Yz!9NMmHt8i&|H#e>SejnNH;oZ|r4qg;F7jJ&UlV
zZBG2nsFtE4MnjF_>xOlemSdbA8zd4Im!eqrjK+`Dnp@O6&mgExb-UuNsFkfO*>hkV
zAxrJkuO%D=TyKK#qMZfxz^xz%?~1((%`YO<Tz<mgsDXQzGAXolG1qSv)1vV76L2X*
zub_m<e<LDYO^*uQ7m|!P4p7)l_bJ49r2;eb1cuW+!~Uc2L;1*W;Kyn%b)S6V>)OH`
z_x-vX;yM*Z#ZD@ARQN8mOfbFV_+D6bQC^y#k1i?L*^y1Q{SR}56*F`VDSJ#dM^FkT
zoo9Uu3xaBL*;ptg1D3^u(7$Zq;$I!@<SBS2f82bQn6)0O=R+%^kL2S@(p}8F?1;VQ
zZrFXxsF>BIfqmZXBPId;AVP7!h~#XZHKSi0?WbA%0<GQmB#3IuxhX=67sgZ&Dv;L=
zHO3AcmU1F?OTAVf2tfL-AitOj@knt`L*s%A(mjr^Nvyy|s?m2rpctL@;zQj<EJ{3K
zf7k-XI*M2z$|#G6(VZ{;Ku@f4FE{`(uc1<b_(OXWPQ{f+!YdGc$QS+Otv!(i?m>h&
zQJQ_E6jh9MNtG#Yiij3J5M<5Ep%e-7nEJw6SGD43A04mR;bfBW;MlFuKMbl*(_9>i
zcFP_BFxS+7Rxb)O#i%}=qYBTgC7O||fAi$ye3wQLrfZaN2ln?zrlU-7XnKFJ$*-oF
zk;XBF9!NS<v|sWU1k;;-)Q93YU%)*Ze#!@vDt=))HsRlvIFro-`Z^EF)+i}=9`=^q
zIwR742HaTVJ6F{xu-k)QYVJE5B$`OSi^`9<2nT|6e<8HwLpHRgaFe{x4S88te;HlX
z5zQDk;}axAp+siE{({OYg^`d%-$S<K)4<Y6Z0#To6*}4<_`v53d#-T2c|m75i^olD
z+mXl<1;p>5|9W6so=n1rtCx?fw9-P|ZOPA}e-(u-n%AxgOUhs|2h#QO|It3-afiEW
zLTrewHU91#l;W)83xp=}mRu(Qe_Q`~G03}TpepknUi5Jc?7-F}d*{f)Ui;fG1dQ-5
z+e(y(Ve??ptmKO5UIGQnb0Xc<HQk>OfFC2#xKSjeet45I>N5ywu#v9UV(2q5WjWv1
zmgIf~D|)`vcq?l4Vmx`3-*2#>aK7I!<)k$<@nw->NEVtKZMj6MnEFw3f8<-BXU4F@
z)6rG86)`JCR=P$P<#1LQPiHZ^KyikqfIv2h$*5FMd~xY_V9(llwPoYAmS3;!D8{ll
zqYD&IiP!7&hPj9+&ljS%@DxK>Lo6@-5}-|*Q+2Oa4UFSAEvh50Kq~OZ243{K&%Ebg
zj6eu2DOU2nGQ!M`fJfX&e@A9VS{HVp30!5uR=GC-()eCqNA0JMOHk0r(sDt)=loAw
zm4SZA{VZ6&{VWazm7e6UzOw>3e-OK0;g6#VfmvSoP=1KJL}DBkm!kiSuP2`@vp+20
zvNTV~z-PDLcBycYh0%hJ%VAu~e_8MYb|iCFOe3T#B1oR!@qmoae<s&CuZeP9pi-Op
zSn@wzAcRC*X9p3CE$g)h9TMoenJ+kcwy1J;XW?G)#>ofZV+tdEkI7ac^1S|-t&e${
zy<u<9GQFs&0qF&MwD1`$;s=eIFFam&A81(OkdK0=5$m)tB`1q-BGoF5<m2qiiP6|c
z1wQ;-nz-a3xzef$f0b2=>I`IB^X22`sgvuA$RRIt7WZgg%D8^=4(WnR%o&z@ZjwqT
zKp@8hy_b3p>kq?7Uvn&M`i>6mjiKo|o?YbObr~aAx?4C8!MtqO*xWXpQopXYM+iTA
zUZl4A`D0zDJGU18HlhO8Hu}9o+&i1PWz#d*CvIWb?nOm8e`)GTh>N!#6IH#^dP>_#
zo*v8$EY6$=?~t5Ke+zp(=aDB-fUsViLOA-gWlKs*mw1SJACs6|4$M#=LxdoTUD`of
zP`*XjQq?Q5I+b%HsIubpw*FUd<lPu%6f&>v5GMJzHE4%#6&Lt*$k|4}!(X_m`Hd@!
z_x1+VsyLWce;Q-wS!ni7TU^d~v==ZeYp-V4Lh|<<yu7!sIC=00QTaQ(ts5i_%sXw>
zjGx`A!!yua>OR^PO9a9v>a}C`pIQV*Tv8^2Nc`aD1tJfoNqd8_e?uLarB%Hey8LJe
z(HpLhy>QVy`}6JW{n3Z}DE_Gp(Mv!xN!P$LElD1sf76h*iUUL0bf2_Q=ha23UPL<y
z)rUSO+fq8Yk=+CF8Z$_)wqv<YL>2=j%Hb<sHLO82Jz~oO<(<fn=}`z53dvS%uVKAE
zenas?1@bqb4t{4t8{5vpo_?mQxU`{arl>_#I4|xAzW~M{C)vFHV!*!kkJokL<B@Hz
zIYv#yf5Dmn399=y2*#B!7jSs-Z>4mZ@X0PA#jAzPkPXEsFR-E)j(JTRms(}aAS1Ew
zw0Es8!+*Ganh26!XAb;588<?LT7*q3iOMz-l8D(ls``;LOV|Ao6oOX|^P`QF2Q`?c
zt{Gl<-H#d|?7S5Kr1=4Lpn+r{Y9m7rW?+a=fAHa|bE}ffI4uLL?z=_bMKgT``Z5Iq
z-bp8S`eh1f$+VxSed4kuKd?^Ev$j3>bh|2|!or5ckeNE~8tTHbWbf(9<r|Z!lzhXj
zjf`seS1vK-knvw>;d9Ofbg-nC`0U6I19NcpJJe%uA5^rv{IKyC(Cy|W&UdJe?*#f!
zf0}YEtTm~tyk%w(6)<O@tg(LYj7^Q*l#jcBF3)gKgw;7MbR5?*sCt#TEHY>N`GGIk
zwO%03E8O(VjhA>9{Yu9=0A;WKN^L%+^fTO5)C-H<ZMTh7B`=VOHKxbb|LGhMy$Lzn
zp}eRT$By)b!33}WnB<r}<lJ`Bd5W<rf9vPMMa}U`HoKdrvTkaG5DXP(M{^OKassFS
zVV8|Q59HiZ;|whd)erZm6EMVOaVOLG?cuV~RVbhP){}(#$FhpNJk;lLPQNhm#T{qD
zo}QA|8#<dww6bug+Tt(GNV{>SEmFn^iViNB3fZq26DAWYJWYfn{M7qcrlLele=QrD
z92k;g`pjvXM@S|8xk;`ACF-XK-LJizFkC2B^c<K5i*qLv%<ebV?k1-?ks9`UH)7%4
zill6_*fN?MnfyR}2GXNmI8G(m$xrivP+QnBH+ejk7AbgLaQQ9}Z=#cwH-&F%an#_{
zmP;>I$-%MlFvl5kkKgt;Sy0~Qe~)!4>YrYtzkzPT%@N8E6&-K>5FF#m>_~A$l9NpB
z*pe;2THTPLoE`?f%*44+V*Y(p?v;WWG+uH8^z8pwxhYY_c6oUF%^Y`KraX~p-Y*W<
z53PyQNl&Y1mYJYDE-!Y3z)AJ(th-N5IvN8@rF^%7l4(Urlrb<}ZII##e|!Q7w+u^n
zT!?l43TEO15dxak;CV$4jW4j5eD5fL*$QXg243;*rK*Il+pIsl|9Jam;!LmVg3Tn5
zC}z4pljEUWmz+3mbnD;?1rJV!#-yjcXMMRx8?8p!rOzAaCUDaOi~N!W{*owsuYF>D
zl<IP%DD+qxNfT#nNxHCNe?h}Qy(vmihfasUfkTk;u}TiKTw+F8f2T~D<;&Vvsf~vI
z6dDaoV3V(_f;3$c>7swk_>k$KfdJrGm0*G-<8WfeibtJ$jons|d=}h%B992zZZO;|
zTl6LAjjZ2>yMFSOIIUx5vxg3gqk*Bo5|L5}us<n4ih|I8es?Z2e}IEvGP)PPaAE(U
z6`*EJ&MoJTZeeklF?ywCIgC1HCl38{AI?5;v#H-!uK0pmX7)KdYhXZYo>E73PB~hm
z6M}vg%VG^Y<k-pN6ZOPPCx}_9Mo$N4J;W(?PtpA7%0mYZ+vH|L?CzU{n>i3JkSOZY
zm!~kn3Yhf~tCqW6f3&I>*!?-$4D+#qGmpZXd~U+K_H80@fKI~Rl<)UWD)F0#63eWS
zr>L4M1zG0p?CWDo;}wN0#?8W=`R>ogEzGYiw3XH1%-~|m&7o`tW74+|5K^^lid5Nm
zD9gz|3bl%!jSvNK{CBjm7qx4+Xz4StKEN&^yQTu78(DSef5p-}@q%&?hkX}$Jp&kR
z$<py*I1@=M%GO^}I|s5B)}oC_n19r!BQ{5?otpjqTba0JZk1MTLB#8|yu&v0>b
zx!hR7kq^U~)nVo{LWz$YzLUF($ol^bjD^$Q{s`3jIA=ggvxZwR$;6-zsxiabZ<qdn
zKe(V{lEe<8e<gA7hJ}BPe6^yJ<Cg_Xcc6LUmX9r0PUnmy@KDJs3W;T*LtE>b7Li*i
zb?3&rzWsC@d!UyeR@;~G>BT#Ity=d2(#qE_yj)^Yfh{}*Y3I?WxouwCIHKTe8Lz;A
zOG6E+dep|+qx{^VM_9mk?xDvf`Jy!S)KQFqhC}&%fB8UH)VJ38RB;C2GG&L|8YWdV
zIZM!l4u*xB6`g4A9y;y}CJxdUuqwl5Rh(YeL%@Tr#xED^uZrf75Ddf0>Y#A)<(db{
z?p9<NfQ_q*m$QtYY!}U!dVeyL^Fg8VQB$Y&?C6b>Ghf&8Hwi5A%Z~o03{PRIMAmdH
z5X+Kye>x`l7VZs?b>Xr!D6Xq$G4sR>@%JTccnY>?!_a&$W!2T(6oX$A7j7?SeZikV
z_zaTxsFRBstt)vx`Di=iD`<x`AXpLUQwm5?K(qFY$#X(|{?K~~r)$lQL!@?*QX^Ag
z{Mx&dCa@8jKDz7>Y0ksilFBEvLVhzGSfF>9e=>e%z&iDCUBGZJ&~S$==5)9i#9(v~
zI~zruuV<DFaXuD1$bwmi=X3~ryZ-*DhV}A*fKRx4Ctb?7mI*3ahr)&bVruD|b(>Gy
z_Q6aV!mxo3TkMeaoViw}rBZtK+_A%%cvI%=H=2CzaOFU^@Vug~dI0$#0n?6+cqv@f
zf0eCtS_uS0M8=Lfo*?rJgf!OktYWg7WGE(lR!|j>@w4hpFgh>OT=Md|9KR7}9+><n
z8c#p&7iMYtf}G<xY16H$5KCWOD?r||Ew7M9xQw(!>0_Kiu73II?mG}lc^5z6v+r*?
zyUmu8WT!q{aSHsJf#brbd+SZ0%V8Svf0okBhLhDTWG7pcVsdI{{fdTn1%Q5Vq1{0T
z+}#%DWqp=I$UDW3_z}s?7F4b1F3@MI^}Vsnwzp&=>_e-Dqh?=TW+Zu$w~$6N4dV2e
zue)s%M@t?(Qs5H~M-uZbPjD4#l2;2+w@z~qks>&RR9$_>%r#QH+5KA08W|a`e_}P}
z-T<?MGA@j(dH&VVUDY6X5yyD9%g)DM9M~DZ8?V<sbC<Q8^0@4n+S0Ql0o#>Zq-VCK
z7r{O+UAp($!%9mlWZG}MdM8%EpuF9GzwAW+p*8m60Y}x4OVWU0%plRZ16E{Fzk}IT
zByt@z+B;a+JGhXzMSso^^X|H%e?{$%&a(Y9#r{XoGp4q$@dL>KOhB{02(4Z!0ymgK
zf(lW?XJa0IagC&C`{UGzcJG@9S|b9#4OkFSrWE1sxA&9O#mf77LuE4rd)MRr7$?!-
zM@t*j9AXBkmFh+onEoT8<-mJtZm+pi-Gp2*R-<Rp&IWiy8hPvX0+EU>N2(dq?0*^1
zXKh|sBY8`s4wBWq{uvaAzQi$?r9`>?oyrQG3DWjz{#qQHRy8aC*mn0NWqSl8a^JJr
z02ionFyHXj>h<aua<GIo?{@PvVNX;NdQbT*a6%%2C#q}9EvFU8GR>AMsJFO}@n@VJ
z;@(byUg&E*QBmpeJ?2)`I*>1B1b@s66oRc}Y{Ka-b!9<7sPLH9K#)A8BknlLcF>yU
z{Uua_T(RW|Y~4IqJLQYM0uuad(g~$Xb~<bt3{xi_;*LdSFeYK=3qH+cmtEvE!REwe
z_>9W|_15@C*7ppjhn}<ckNSJ!tPsh>LH@zxIOop5g21HYNE_P7j`lnc5P#_y%DUYE
z*>8OHh9+GRO`PpC?H~l}@7jjmA$&MQku4)cGLU?`SZ!Vwp`pr5)Y1rtfkfpMh*z!C
z6Pe2%Fb9m7$3@mbVANelr^^s<#H~UYd>JmYbc#af;?g%&hf$@ta7+oa+H=Nn8}i2%
zgYSZy7lMoW8K2RWb96aU34imOuk-0~c|Jj>z`Frt2VA<W^$RIISk68~Nl=<)ve!0z
zLYav=onaU*?VVQr2$e;1??{u8-@B5OR+c(JfM%4hyw^!Od9h82J%Cm1Ta@!3D`Gek
zjR-KZXzW9$FWZz-u?OLELI|PI6z#RoiP|0-w!!($!5bI3ru{XhIDeDhk+L;qf8Y@q
zxInZ<uGfm@gX*CYeF%0_uBHR-7Lg2o!-h@NbfdVJ@KK`bzb5J-sfmqpgxlp7mtI%~
zJoRHogkF9hojy=)<7$08Vu4mSj93kZ&_&U5(E9_7f9p|Y5qKs20&ASh!s%?l26Y+f
z$PptmOg5b4bnbvctbdEqt}iP{9X)WD+Otr=z_M7rSr{t9HewNf;j)!GCm`l?kT>^;
zFL}p(3>$=(AUV<12Uo@@HfYHo2p%n_7^^P3F~<(%;rO$@&xODV@5vpnTQGrt|8yzS
z+E>We<Fidqwhd`(@Ra)|F);}~XU#*4+P=$Y*UMJef%nWl#eWBl7x)3S4p@mG1Pyyg
zx0j((P5{GE2XXL_9xwatBeXPrW{`*C=&?*=O5p^r2W4`JNKUEOS@hb0Z8?d2S$s-c
znu&9^{4iw6o5{cb(>UP{nZ1EaM#T^&bH_&(qj~Q&n!+mflO~tf>N8hjCN;F0e#8zN
zH!?>PBgY$VUVn-ZN=j&xTvgH35k?gMF+haj25H@kb&qJ&Y1v)fma#b{CR$-g;JDQA
zbWzP4<&Et+X5J9YR!e$dNudG80}WKmc*a^s#-CtxPwQi#zs&%=e7O@U+h0S#i(oW7
ziSO|K2km#?#uvR$KA-FFnKE9B&K8#6Y|ab%d9sRD?SBVTiO~rZuDyuQVwv<-s3jjw
zDa46>QpoT7RuVkC+nC!$a_AfRC5BMbeEYbr_o?&5;m3H@d-pvU@AciPY!o*tyY9W0
z#E3C&fKxGOMdZNbXS3rvBx+$Q?f_Q*AfkH7yA!6oB7_Pp>`ovsVz*B6ovEi$+Xr4*
zyY6f@-hWUi3-GcHpA-*v&ApWA=GTk(EmG0WrP}ojJ+yW^y{Yx7O192|Q6~FaDz1sj
zhc7y}O=gTJIx~eZ@a_b6s!t5(kf*5zSr`Cb$kgOS3e2Uf2zA|(bo<hosfV~Yx)a3m
zIoVphWcDt4fd0~Y_ouvpcgi2XW6*A5Dl6dq(tj*EAATXMoDLLmGgA6tQmR~ssqqiq
zHLGY`+YBdq{1`3mjeahmX5P_b41+T7Hnef1dg^zcB4JlUI@{Ftu2x|r`!zV4E?~T%
z2jnw_Qh>7u<G7O1tdzgD`<}#E_>3pm=<qs1NV?E3N2sTf#^Y6?gvrws%5M_ozAeKu
zEPoPX>S~}G^>Zm99mB|^a(H;hEUbLv(ngu!qp_Q(8}DhL+Xu{d!UNT~fesgQPI_ib
zE?_q{@t}+A<KF4;hpZ$0{AbpI<=1x5w+QwV;KO6eN?_VG7b{k*v`-6fP}ZEDL(~XN
zYf#6%(@=)vU<y+>MpTryZs6`Mqq<Kxr+@So5*M*D0^o2A$p_}te*t5_J&7G&SaKBR
zg=Xlyzw==!m3(HtFn9~b)I2nfLBLZtg`LK@v#z=h>lajQavzBU5;p3{HG(WlVleNj
zb%0u#5Ua=rSH;kd9~oeqD~!nF(aHK8JFXhl82bZ#`CS$8)6OtwZ`DFfdc3N(<A049
z!Fvg*fhVe^LD-v6!@m?y^23N;3YF3Te>-PmPZaFXNs%GkK=Pj9z>nB)aVTwA#gXWE
zkiTYF<2=3EbR)&R2XSLC9tbSS#Tl|uB1}h5s<lx)`1LCnLF#wKV1UxjpPIvhuG-6-
z=iTu1N}}@Lpn^fm`zh;<DTfARP=D1){(lZ2Iy}Ih-uF5ptN2Ankzr14O^DT56uVr6
z(kO?*SZ#}zH}r<|$mv}xSW0v0^y=M&dKE}=Hbfnoi1z~igoVucGb`W3V^;F+K!Y#H
zJQf#=^1)Kq(@(`i@PFFFtn*#mE)k|VecxZWdx8p+;+u4?Bgf}vvfG*3x_`?%(Xn4;
zG|lO!iF1?<^L<a^#D-g%)vq$g<iRt}2~g^2BLE8+9$ucnZeD*(GYUXoJw&!@PT|>Z
zFv>(eIhJbsgLx~7r{QZv3a-JNpo$?Toz_V60{MJr03ZINrWWH_UScx*^aZsACbywv
z)jewISlW9SJc%SA=o=^0Jbxo<*Y6*5rn-$v`p^7d)ck9rqkX-5pju#*ygn)5NP45c
z`u7X&igOjqh<2KFUGbk;`DTmUK&hXuZ>LlN2Z@5NeEDDpH?0)7Yg1Q2*uxs!vk^17
zO|uxy0k3QrTc1(uFE~75hyyozbTcE|A2eoF=8Q5|N@`B$gF<i?QGdjfcuG&Vb#7W{
zevp-F%M~J7pMqZh@wzjglVSaJUqFk;--os|k%D~uRe7#H7n!H!FdmBIXwd^Vk<JJ8
zk}H2({^~uK@-U2oL23Ar0=2WVY_RlbMb;m}`4%c=5W@<vxt{2k8dT0F2Xm8bkS6-0
z1a}EnKUbJf@y%^nM}Kl;8Md%g<GcawBTO+pKILNWJ;ZT1zi2fA0@d?-IbGiKdE1SY
zH+vCmk<(5n-}O0*=8a~(FVix|)@hLQye1!jkW-Py1X;zx=TP>+hRhu~7B)Ta8ROnM
zbEYouH=~RMrfZyOcga{SOVP3%QQ^}0&$%8>`Y}c5%PwH{mVZrwJ!U{)8|EvJ_s%~U
zj`9v6XQt7mtEC5Hk~xzx%38wcsC9UwHP-Ft&IN>BL6LS_Df4#(#(m1LB<^H@l44)-
z$?|91P`3s65?dB-7l|<cNCmpmydSjvMA2lLlkWspXDCIGu*74iTllpSU8AY46}l}M
z2j3Rb#(kIij(@-y^+gguPJxTQSSri<^4@Sg0qyX%aSjwo3M3Ng86h<`QxqCKcd)dG
z?rS4~1^<2vaSxQF*`#fwY%CP{dxog}pBk$+VWavC%7maYsLpFcL*j5|)l3MOgHB6~
z@y6vpP_K4{>efijPA|e;hdk=8?L*H<Eqh4mekU2~I)4(Qmqo`PwrRQQHp}_=jV8ZX
z?57<>mG;Y?$pl_1dKJ%Xjrm~iWjyN|dvQu{Sob=m=RLp6=o^GTs`C*fHkM?K<A?w;
zgppc&BM#q$B0ak(2jktw)J(o=;A~DjV`)=B0X)%;-#vJ*WVWx&eZ9YlFe1V7r7Du@
zyxXh(3V+i(Mqb$A8R+ENYe0Zpmio?NPN>l}KoVr_!oAGRjfH~t#_(fSp_Hc>(PI5X
z(SMV-E<bQO@VH2r(%ortx-q15066PBCR-|LkEbg)=!`65yyo*`xb>I-i;a${#R%6`
z5L9-WHC99jwKO9@$VTTY3%mz_a$3Sv@2&JGet-1-6i~k3>O^Rbs!R8252+DUzYma3
zV+2ihiWPx|rd0zQITAH`qU*yn+zcH_`I$r-iiJfBs27MU&b&lvN_wn6)F$Rj#8P!S
zM{x1)JXW_vR9X8II(X}VH+3s_qxl5`u?qsey+IPK1#Q2Wo7%r_E&+N@Rx<h;(~v#N
zxqmE;#HR@|u~F!duSgjT=)_?LO*B>1B0Zf6AJzwJNvpwM_J$+)Fw?nwJ+<CP=Uf-$
z={Uk8H`yjgf~}-;7$EYgYAAhTEj(=%c|;oFvl@+dNp>{g1tq1e?&rOQ@T+Nufk78O
z@2#p6dk&c#zrXlq#lwT6O=KWUx1JICi+_m4kJ~kNwVHUZfMCcAL$`=10C@QG%nEZj
z`ZASV?3-V>p1UO-KM2iKd-rvmq(g#pQV{pj03ESmNNQhh2eF+IHz#qc8CQIqe)}Jn
zd@);K2fV<b9{g@Gb9d89;&*8o(m*o&>nqOQA0eGn=4o1bXG4MFodw;N=-r`ECx5M#
z-nZ(?=&T*X*x4<o&T;TyWPMzDHA0}JkwGWjqrPM~g<Lhn^j>0>c&Fj7u3sya*6ph*
zh@~4??)@IiIvtUcQ(L+R-+3sq(cY693vE{*!C#QZ#L@eo%ksL*i=Nf#gD)m@c^}mD
z6tgT>)2W_5YfqWu>lKFjd|+`aP=CHhE-yFGHWN#6;b<n4gNjq|8p40(hCo4D+V6y`
zOdew{?k9(hhhPH*L5cNuQ6~t(2@jtiEX#=2%8(tUz5rs0H$P3w^q3SBbR8_Z4s0tu
z6yn1}_7*|76-RwuX)0K7$61=+<GrentY1ifn5*|-BH!3{a)162f_({R;eX)IN$5Y@
z%FDsHB)dh4B0F>Qs#cJ>elJfkmJ-+_tJ19k2e)2e_dGYO$?XA8fMjl}<%!7!kuC#C
z$(aryE~5I9ISF<x<i@~{m3|9J0JS=f*fP+<$Yl+hs-$Cw!er&R>qh3IqMMue9ieMR
z+fsr0W1Skdozhg)pkvvTRDTIZVMwgabuvB*9jdD=AmOLL`h-g<1GeEL{~s58pP3^H
z{bCyGS6{+UrbR*y5h;aD<%nAMVMDHIv+qYcD@)s~rG9$zi1S#`jK7Prs%*<h%9Wi*
zvASQo(RQ`<XaY1UU{z_8MuSB6JVR&dIipV8&0Xy(Jk_u^=z>L4rhkT-?wuPo@?$T_
zjb1C$Rh4fN5_RqOy?s2~XL?;i+HQYh+yzPVY7D71R4MwnPtFP$IW&&0sq%b~6ez6f
z@%G&Tcyz3yVa>SgOr=w5@v3@MZH~;sb5Q#St_0Z|<bwe@?O@l3wi$j{(?3MyjSASO
zq+E@1O_Rj^?~(IVH-D_LAGldi_RdhxqhK6|vNbd(PiS13vbXyas${gyx5L!TOV{=H
z3_^O3`!-*gY9JOG_^%3iO(qU{0*;Is9(w`L5J32_E#nZED?+1qpz!gb%58ntg@=FV
zJNKH*Ep2TExet$mJN+%jVDCD++!pCiE!m;`Jh-p~mz`9<;eTV<;tSZ$(Wxe@%V1`w
zcLA!es}t!1^Hqj3@nsM#o2eREdxeY&jk*l%(aY}`Oj4w>3kp)S(%}Tuobwo6iWzY!
z&jDEHVftUG8#fGUz&#dX7tqWsh=?7L#7>IlNZ%D~&|5d04ywYO+>X|vsCB7t>FR&_
z`6$)px#%OukbiY5bTuvDw2o5Qfl~d`{T)Ga58gV2cROw^{TU=>r435tf{Kt(S-tmI
zm~Z}g^73O1c>`ZRNQ-^PSmZ$_DYWG%X6)7!mb)2e1!!Voxcn9CDK*9ry<aj?iXCDO
zsSHE6D6|WI6NHcPzu9e2)FY%LtRapGrZf~xt$HjTU4M=;@a%(yhQU<AHint9&Lg0M
z)4ASow(%A~u0Ql9zGz$3n4iq;qjWaIaEQWHCkR(&QQ92x7KaMZss<Wdg6ZG#FAOO^
za9@KA)jAuRt7|XJ`@g{TB@`I1_NSqRE>CH7ZH9I&%!18&yNmJOefU^%bcftLzK6@w
z%?`xv;eUXM>(X7ns2pQ;OUj(64i<2+duDjC+nT)<EV;+Ap#H|b!S<SegO(+d2yIn%
zpe=&M)Zi=YaLbq#gr9)RMey{Q0YmA9(Bcw9MX5O$9j9dkaeD8ho`nW8hK*JD6VXRc
z^5skUCuqQyZ((jK(XxXO3Q>t3fsrFdH^jAq?0<2Ow4RdJOdtWA(i#niqj?tOXv!F2
z=V<9Pjy(Q}{YDeVwkg0ih@+;v<<8r%tF&gV`B8{8d|mQMaOr0m2Unik9IDP@bHJ&y
zb^0*VB!MWQ^c5OBm^nR!mcH=NA9KfQKjCPyjB!@DsFKuCC>?c*`VMn`)~oEWz4}QI
zfqyn;$KEh1ro}fl#U<W6ev9&N*iHN<%eRWLP&HHN-Id(YnLM0y5`3Q{VnJ}JZpmX6
zfp$g+EcoN%V&rm}6RNiD_gi^jr|_+o+lT0M-Tr^RSnAjhhg9Gnx?87rjtYnzOmYFD
zPvAu-LDj?+4LJ1r)QM^nl87QY7<nBG6@OI3`skvh(|Q29B`YE%eCg(EcmV@IDOf8(
zR{*DbZVwUV62ZpsMQPvfjO>>||A9YN5bYqQgHEsYX^SzE8lCNcU!#UYtkb<bGksq6
zolXw~^|AUEbni+SS6k=KsFIwCr`K$!13ql4__bA-ddun`6eI_drc0TPRVi1z27kKh
zUiGRNw9kZ3xg?YfeK2UHZDu~S+vIGlez;#Ol^cW<=Ipsp6+4#%0_k)FdA`P01~R%@
zb*D+UISZJPoo>DZmn%0=?ys0AAPq_7{Yc+@67@rmW6kSF>2c_`myCs6`aZ=?pJ=Kq
zO=0BoLCIvoOLi&^2`PxIyNpByqJLq(i#u&l+C}?uOP1S7_%ajxTvu|*(%ibT8=%f^
zc}(Vb%-g!2QYT9S=?!Odr<kp8OvL|rWz!Li(Vh8|uA>%Fdo~o>`Osdw&<yT#!{J9`
zMnt6t-EUEyg<<V2`<_muU_Xx$G?Bf@;chZt5txM;qq|#}Jr##`6C=T$t$%nHC%Okq
zgq5hkZg0+zAlEeI{GWw0IJufoaUU%&8?Z))sf+&RZDFDDzf)zEXq{Y28!7vYp7j2F
z8^U>BIAZWEG;dHU<Y>U9YhdJ(!8+q;GN0=$1|MZj!t+E)?ua>jrb8F!TSh+AAW4Ef
z;4hjAZ@F8AAnyXp+KgvEK!1MWB5;+ncwDz0{Gl!UdgwBD$;npM>=vbqhxY=T36?;$
zXh9r5LG0%sJGZ;iq`ux=CJuAUBbo2S0E`CYRdOTX3P&xJX$pL3$sWmjyzQ+Er>}T7
zZW)^yE_*z0j*vs=i-PhwaBb^}=(S6028DNeh|{15x@(+KT&*3yRe$Ns$^);=Mmf-w
zC~#zcAE1!tJ>!1d)sL=!cQzq5l@wlWpyy;DaGb?!i6p5Rx<P!%=<=|r(M!_|$#BPT
z9MJjY_}IZuwNt(9yr3gHiP>WLwd$JPaiEIS*`*Nl*oI+iYkgxPAS;cAW|9(Vjqqfr
zy=5D(&A)|>s6e%JdVhD43_%##_6qZuk1YPDtgG1ZLkcefa=YCGfEQ-NRb-9(*=*0R
zPA3b@AXWz2KdP#RH;(<}u~xZ?#VV7P5;z$p@30XJop|R>=7Io}Fpx<M_+E?!ptR$2
z5vp$)ysvo=>QJK<B9|`Mk6IYah(DUyz<_!V!Z}|%R+8w4h<|M)t89uZ@>;VV_OW7L
z!GNoPaQ(7Ag~M)Xl;MyG3`X)m-;6I}jDXx3m$96(pfqSHlZwEGn%-LlL&!dHB4*!z
z=iA(hLRvtJlCWfdOG_mPWi#;LqpGm|s7|DYv;{N5H_mc0Qc=&nnn!k?V+DPn5g_!}
zn-?B=l?%C1oqw67uWTC)d#mP%KT$=ZuyN4mju4@u62>arLB@ItwrE$3>Q*n^V%Gfe
zML)`ajUf4=ZiHXAiPJg$rEp8H82kk+b4oD=edR!v&E>F@y0rK{ppg5qCK@fk*L*~g
z2EA)slz~oEH9$a{65VJB#z(M2^P`<$oNQzsuP;72g@2l@Xhb%0d}uXvO|y@rov~Hq
zO0isb-9PZ4_DGl?jOtguZoSc{;rZ#HPWAN6s`>-D^oH`K<+2!$BFU{J5%G>aggvCm
zMR8YC`hk{EKbR(42&yT6>uNbuS0??TOX;uFhrq!Mx8KOMq0Vm5E{&O<@Uvda;tEv`
z^46P)>woLXPn3GJ!NjCLqw~22AHAe5N9!XeE`Tn-w+w`J6i(R#4duAme^R}YdJCfH
z!x#Fqqq`rIJ*&>AI^<1Cz`9-an~+UCc&|*WhxpU+j(|TWY%CBS^12skfr_3VbeN`u
z0_0lNX)Z)aBNcG#`o~zn-mf7|1n4a1pj60pn134#?op)Kfp3SVYNaWi;V6|#-+
z^k^Jmh5+<?LI{6pP+s<5+3sr-vj-IVQ@PWmV0s=hO?8>hSsajy?m2TW(^Shlwuc90
zD~tfJPM@uNRXus$s;W%Ok-i^9M)_gNqpODekXyA?^}ZTf>y63upyF7<83fhE$T<A9
zLVw-h$@)`It!q{+i;8&W?JNAAtl#nN1ahHT_0}3N3#cvz_&_^jtQBY=^p4a8EM-el
zt*qoF>lCfFAyM1gBQg-Ez)aT3c5>cxDqwp7krOL@0s->3Bf%v&m2R6Kq2*kU5A~qg
z?!Vi|dAtUBuJ^8D$-y%2BKW4FsQJ$Q%zv~m6Ln)t&`l0y*SenO=-UmY<zwU%{g|O@
zF0#8)zd+z$sO=%+U}T+eYQOK1H(nYOqSv#ZSa4Z~>f80}HS~^>#$chcqN#82_1zf2
z7il0MxgFjgnV#O?NUvCZ`OGmL^EMSaElvQn#r_Z@?4wun`sX{^7*iLhtg7HIiGQ|7
z75=#PwCdUV27;rN-O=xQ?kGG>Gd_<mz{BSm>wyX>z(VJp5FQP;zGqM`5TJYB`0Em~
zYM5OW?_dvo@(()hKq2(EZ)htw%4I6f<!yunYXh(dee_p><jj`qC>>s#-;M?w5-N9U
z6wbEBVj5yN+zUf$kxE7rnH7F=8-GX!!Z{NC<SlFmUXt5=Co_DI^7>4na&y_yIPwdx
zjYNkf1TMYEpHe6MdET$?X-s@U6LCL|PxaAB(my7(Na#n<!?InQxH2wvB;u(iqj3o9
z?|Lx-Z8BHK?yZi{sM}q}kY*MPYY`XmWj8;WDemlqw==EMJY)!-mO4F=RDZxBl7ixV
z5TF=i6^UOT`)y=U7AE>DL|u`pXyg9r1ev5W(6^O+K)PVE><xK8rr6=;TJ%1&9iwfW
zSlE*7jHwll?x)_XjJ8y?Wh8Z>@&xvc;uS!s++(c4JznNMC{pH~IWn3f##Ub&0e<+8
zkN0_uvH&VQ8{}>dSXHalmVdz_p;n!{M78<qj=7Do-0-Hn<mjBRsl%}T=*SO9ABAqW
zt!nBv%PH*b81TjBN{OY?srDZeYD;DxIuCVMuWv?g0y+ncWyVP9lVj;+#>ul}E24G#
zT6lYZ;_7289q?`RiJlkGFHsHS_A*U9Pug!nG}3Ggr#>B^vynlT*MB7&*K7kq#v``S
z0;H@;lT8-_qqW%~6#2`qHpSpZ3GC+XbG#-h6%ixr+2P0NJu1M}NxRj<@3SiLmM!^k
z$j{kMx4501KG*_nFo%pWmlj_}`u4jgnmSR2cPEb62~%eajH>HCvz=U$<;!p~Wh#G{
z<q|mdITT{fZvDb^mVX07P<?%ZUz6lP61ZxZ&N-%bUZqu}uG*!4yrplRl{35x{_w_a
zNwdy_UfiT>@9qaQcUA51SG8qXg|s<!>ZQ!!3VA>q3s5|e&2EJ13yTCH{wW4>i$5r?
zX`DYk57tSVfwK+#Ce=|CY|wKqpGmBTrvi&!ThMa?o2Ity-+w969&b6;`0W5!I6C4N
z)36|@Y^76&jp)}dVed5^0{^i97*9>n6;0oV!e;_Ub&(iXKpt!nNNb#?_wngZ6Wi9@
z!PZCY!toGHo{amkVJ6H%`!iqP%wFT0g1PN5L)eP5gWQR5oF{pJ9du9bk54*`Il-H~
z_hFc*ndJ<xcYmhHkqo(mNs!`Q+kIx8wR7XZi2Max4G|vSD+XJ_y%^uY@e~tAmECM|
zg-KJ{ONX+%W+LrYtmGFDTN)80UX*^(-N%V<Ebhuw$xo)KS(Tut6y_ymY*(k&s!O=W
zGiq+cD=yhI;+7vdt}SfsA}p<b2Os;37&Ji?@Ba6fFMrRIIDkyg#YU*Afwosp>c-0g
zo}Na6RU{Jd@0ZXt`QZ1k&egq8mQ<3;&rw9Uz2}_U1?@k~2C#n0E%Il_JU{c6vEf%P
z7s&E5#&QK3wWJ-!%7%8#Ad|{6u1o6bgDq;Zb$$2VFH8&=hUj!$?;TRAY|!zTm9_G6
z^no(jmVeRVgJiR#chlyOh4i)89y4!k8@imkVOAgM6~6UXI79eRmDO^}v~ZuY25GCN
zKU*s$I^)O7=*`zYLE6NX7gxxY4&Qp-lqg`_v&zn;1n|%JW4F%qf?l3CwmGeRPTy*^
zVHp3i6ck$cO`%**s8I>Gp}C<z2DAT`D_IV1wtwVIFT4s5aa0VF=nbiBX?;{0J1D+1
zye`=)6UpBp!`MiUFaF^~@U=iY6qH*j!i>r)eqbUV8XV>BE#<z7(8=valA%;uc5fLY
zzD)@Zeqawvy<Cwh1{<9s$@ns`F$(}pkg19f2cz23vQH&CYVHH;O;?X1Gbo|q1rr6k
z1Ao2wDGCAVX?Pk+sM6vTVfEb>7wbFj^5!jfM|h2|ua+hFFo*Qn0a->)_`75>v!adC
zX3tfp@<hm(h#diy$9{w0?_k(QwJ%2p4^_%lg?Ar4uic@mC>N8AzW&}Y34t7AQ3k2Z
z^<89FtzLov23Y>Ai#u{zf(O);N-UT0R)6u8p-^0P!6!s26HKgUf^4ju<H3GY#P>K*
z#CcUJZpiZ^<U&x`a#~D^z$-8iYK4pI1x@XkRyRZZQQ#{DHhV0ec(s`Lswq+c81pA;
z)mU(yVSk45-Fe2iMj16is38k53Vzzl-fKS2K`y`bt$x(Ijnd-j`?1#Ab2eH`X@3tu
z)j@B?yDG9+F9M>pZ^-QEt)W47)R7}&N}I|3dlrjn+%yduB;HiXFiUY9oV_wcgp(Cr
zx?6c?HBElp@bThK0Uw*Ah6tRs%x5EP{r*U#cvdM9y8#x|kQwTF&h2-SSk4m<9@K6D
zBP0u`QI!>D(whN<Nr+U^hLntRn1Aj|5b~A^)#ZmX<1C1b?Q+qLF#bmj+n}2#;x5JI
z6sn{a!bkgp86|Bec8C__6Qb$#kv40yT$LtWd3&ht#ad3i>lE-2(mBnmihB&-Yz*9k
z*_I(m(EQG;gq~!l>A-5lkhopOsyEZTgQ%7{d1c&9p5Pj<>H@X-EdJ6C7=MK#iTAjr
z;=Pa6lxm>njN#H9Cz$*2k>0nC!A#j)S|0}s6*xEdu=fr!Kxh*iIy!g5o480~7qdPi
z2167m4a&ZP`NtFEyfQq<IF(<oZq>k$iyyl`ZMrlZ`!p{yT44!a4@sOe&4!wm1^g{T
z6WzwdGMFufs}r3Zmy1etsDD(F=JV5h;PQhV;YfGpTH;2YR#;QJ<!L}pz$NZ+dz_3i
zOENCcLYm`O5ABc^@yl&-DtQ3I?PTEG$(I*q-A3yNZ6Vcoc4CFo<Qgm%R?_bu#~#x5
z=nVPJi}gq_M;&4k{!+kQ8L;{PFGmiz@yuxN?*}lerE%<ZJ!8;bVSh5oam&=hF9#r>
z9E}44h<iRaMS^z*{ess__(abKDinE@h`}~~G7dBkL-VKr@9GumCO*mSP}fGis8fb6
z92R+{cC76C?SUQud=ul$46B=>tONc0cat0yb(O((7~xbHE?36cD*WoKxITI^zcMgi
zojQ8cfFN1^n{sy{wtwMHHRv;5Dyir|f@%SD8kTr3KSA%yQRC!^HAb(YhLS{k+Gn=@
zmqr(+=$H!M^@HEZafQz&{kP`$M@JGDnpt~dL!EBhox4_nfiY;$vds0B$;~<0efNqx
zX`%|su}%n->7?a#8Zco1Onq%Kmv0c;+?^*Jd@po0Tk>%4ZhuTRZ9OW~X#5QGwv<>8
zr?MHjv+P+`reATyj=iv1TZPnh@noEfg6!vmwi)<{HFQkNe*Oac);F1xcVvB2Xm}q`
zW*Pi&&f$C$4>|6zh|Rk;eQLXQBaPC}K<vx$21k#+1!T5MU4(N?^3_g3Oz48oeo+QX
z>!JYEaa;c9hkxH9S&OVrbugE6{x*G?E@|nrxf`iTKVIEOOCsH%-NWO-pEphKtJz<N
z9XC2Ca#L)e6}ME{T%x;cf?L@}gl+4lgbB>|)LaT-+^6bji`@<-P2w|-*X@GKc8-#%
zF~7OxS^Z@N4pbhTAXP+woSQQZeOH$Hf?>xZV}vOxE`J=*VXTT7N>u0=&Y?GLRj97x
zX9>CaUA&*|HwA9-k+GYw$aT~WS0R}aC`A)`u*k8qTw`Xm4D387KF`nO7CFIBnBF2E
zkGC>BwuC2v1wbGV8AWkVB+*<o7@2_+n1BY#|7VggY#Sv+mZ@+r?sQU=6dT@sUK~v(
z`%<wxt$*Ga`t31?FX8#&!Q7wxFPgwa_CE7=>^@7+z4k2V<B`kJwkX|50<StjMn47x
zMw^C2-Pd0G)?OI>y~(P@^w-~)oG@Wke|=Ht>HCEWZ9H_}Ep8?Sh^z9nGF`jYXPzJi
z_cRU&nDW7FcILpqd`MFQ%ntHV^fp#dlSUGbx___RJfGAmEE%Z)N>I=C>;vFz7pW+R
z%1Pn`gVHW~UulU-g3hAfV>h^UF>v*r?3)(DI&YRq9hR%9rEL5V?^6nr@~Avtz(MT0
zNCCw-&7fNd;zpgcX1h<y&xJNG8(8Swuc3-1w25laMw+b2$mAvzdeFOPoR&2lxgc$#
zX@B4LtaKdQ=+14grH-89o&2nMfiS>YiGnq=d!cNGCi(PU7)VH_DTVaka7?OA1#93I
zQ2jXyg$aySdGwXMUBIxCtTfuwy$%=_o%;EmnL?MYav(c~`Y7FMpYGHN1dD_ICLZPy
z^B-@N`zMa?05@f9l&wxbBJqr~$Rd-+N`GmM8n=Z;p-Uw6U7#0G7eq$CwB;|Mh>J)B
z`yroS^H$(B@g=xh=CYCmQAnN?9UH-dG1E9PmU>R8Nu`X`v^roe<kQE&Q@<;?(in3J
zWREN_AM;=RUT!ZH)(qml_8kSoJya3|WLA#j`y4Jp=%nQ)&hT$coE^nw_kOGWyMIod
zKq&3?YhhH=;Te>4!9rI`ULauHX`D&9#D*I1*zZ*LwQ=(oHEdn~TEr*WUtQnz&k!j&
zA5G&KI1k$?M=-WtK#@8cD?}iyAedFnYZl#7{FL*hF9CV%Eh%Ep5%m)4WVTxpIM3i^
zbKqMj>KJ9%-gXTxAG4yH@*M?vhku^Nr|$V7?KvI`ZFjU<^%h~c6RnEn>YB2zA1T;D
zhumQ&)${%^lMe%VAPcn{lM$?kKVt3n{+vFdOdK&?!dTPd?7KM5md?Q9@rPZeXCyt=
zC9NM}_csPwl8G8)>7cCCy=BanC*OI8Yv^LjF!(CaNfgDP8)8SyF@`kqkbj;J<?(>e
z5oSF5=s_F`_!yd1wGY6Elp~9UmW-&O$|d|1kiF?IVJUqm#`EM+6bRN0fZDDt^(S%T
zmQivv*cRu=y+rO+i~YP7KKgS5?sInq8f+Pibn}~pAPep)U74b8csjS3!7&aIL`Ths
zUKm_V5F~K_Lp5KjF+a?HTz|3S?i0f0%hwS^<75{jWb+{5<TLfZXn--7XueU1fhH%B
z9*tvqz{Y)5E#6Ug4fPd{by{s)UiEssNez|Ag3lY#FWGmnjh3$NxJ<=lv;xrCa5`3>
z0PgwUy~3-DbbM}@dUkH=cdYta@REzIYJ%_BZS47(|LJCy<H3RTHGgV2tc^yw@u0!C
zC$3MoV29@`Cp&up&}a)bM=lfh)SSNH+&o(5m<GgyZ6>-(y2#MUMmct<#<F=3RK&5$
zLi`0?_obm4%0nF?8orujZpl|uU}}Loi7b1B615nOSA#3pDgc&q0WE_|!5@vmD34%>
zN%~n2i;Mojf|TTPhJWPyjkb)Lx=CQ@Gg8ftq78zi>B@^Qf<iHY5Eu%YlqEy`=?Nss
zCsp3pe_UP-oD_oh%#n6<Ls`0x%>7Sbfv;US#cXQ)Ii%hThiD^)k3jVv29PFVv;l2d
zfQ`}24QNS}hn|{u@xuj<<4fH-c0xSe1U#3E1l4Sv?YkyX&3}gCeM}>wnI4rDr)TKW
z3if=p)Ad1q9NGE2`FO@cyrxd!cb@A(mXhG$QehtyOvP1ygFdvQpyf~G>>V7~H%%Ya
zTs|gpHpEyC+nM?SSCu^EfQ8_WmF;<+CtRxj1Y(Fikj$Df05<zG?Uo~e{9*Jd^33*Q
zijBMT;QA|!9)CpBBHg;>V0|$b`mRw}fNxtq{q%k7GSMd~W+Wj9SpN8ugHk-`1rh<f
zeCVbG=-<8{&P2tkg5(MAG!*lHPaN-YDEe-sPfpjxBy`Dy+{(UkRZw#1koY-RA@CL}
z+u7agEW>0yzKEF_dKw3|vC&nD->eR2w!7<zO$h|7eSefp>!jU#)PK)znVg(r;-8~#
z9zrk9I7U!&auPx)yQ+N52cC^fBc?^%JCdX#S69(H1gSR?FOg{bCCa|FxpNebh+87b
zUoHR$M@Xmv$7a3JslG4dDiu3NM%~%$eh9K_@}k=+=2!72sPC_Mk~~eq34G6cFMC?_
zWnni<tA7Yc{={%^tdCbwjH<2%R(e9}(v%?;Kq@!W|0)wTYwLCKD5pCQ<%6Cv*k~31
zy;F1<G<W+UY76P}g7_e}kuI#wtjLc_U<JPK^ZG@jFs40DN&N{a0(v4#k58?D;4Pp#
ztsyrgfUM=l1X0(2m?G8i?}W8BR}d=?BnKp%qkqHIDn)NzCUwGZ!14Q8RVX!x@ga%v
zdj)j(NWRZkWx~E>>0rboTumXOcTY`kvoa*hy!lV}h=A2^NSjB)T$e61m`rf_{l?zl
zMI6jfw2wlML8C3_lc#pZOEDKXihIQurihWJ-~`Y}hd`ZWxIzn1=50p}P}KGNiGilj
zqks7}2kwnfcFv!Nl%Gv~;S4Tn__y%r@s;4z+!d%fZ(U{TJd-AAZ<;j8`OZAW34GKx
z3~JnY6?eF{-jq}00N4EmkHc<zM2mb*297TRh<h4`=;PGiPeM!R_|=Rft2-q$U1FLv
z#kh;R$&yd=Gd8?%BK5AUS3K^k8m&VUCV!Q$bjv05jRP$2rVjS^?W1mo9(K(sYx>S0
zt%+k+)!3+ku8dIS#<kWLd>e3(@q-$7yjb3XxZqICoYVn@bjb|Z0jOw`@RSU@&36GY
zBBtGg9To-4OP-uS4krTfZz5QFri%;&E>VYP8tFXUo(;e;7B{&EIZ;1AN4`GyA%E(e
z({n>fN2_IPyvO>T#Jd(EOfo|N(Rv_fiTDIF-Ju{<D`0+K|G_71z$S0b6{1Q`G9Ypg
z(T;*RD7ok1ZgGv7kr>`{+N^3*ns8cxJ0HKChad>Sikz<J`QP|gMGEQNuHj>(WZT!S
z;y$^-lRzzp0~19vWkKm0DcF=T=YM}9Nr}%ga?L`OzS!GL8Of})9;ym8&}*0acPrq`
zODo8{Kgd{p@|S4l^BzUn`q{p!k=&HdyQ5&8^o1mIomN&N5P|CVXo6D|JUGF{74-f~
zOCl2fDdz8^S+9=)4hdP#1~GSasoh;O5sN*UstDThmGTQD@p^d>aCU-Kfq$82$Ihm~
z*<mHmq54+`gT~q=^fQ_?9JWQISNpCLg#yG5nLE+DVQhc$Go->}K5J07Hdrtir6982
z8wJ9b1oV>CQtrJ+=ZX3iRvt6hLL>9{Ka1y57qAxahJj0G(xj*~_lC2kDCccQ>N%4!
z5U}nvJRRldtIAqmxAiOS&VRfxra}XKQ%6f4_TPvsGULDq`eBUCg~bY|fKU@p9v|7f
zl?bG7*aq<5UL&R&Z_;w|MoV5f9<$72-a?K=<fN3QFLC-{p-#vK=!vCeI6kf*U;dWz
zFqDvX09~_umX9w(**Eb|hXP%mNh{clRi8$IkrR_@p`alpObhiI^nd@M4zW=%UjWCg
z_Ac8*(@+ZZJIHTNldX|1)!jL|IK(1B1^&z>-&SG9J1-M4eC}CFGDY-EZMEIreY1d!
ze2W^VPH+8%lZ54dOhTFcBVp|O^XyJ)TV-*-?MawaS1MbW*+o?wwxz|06q|M=<U;-_
z-HYma1d(C1I#t421Ai^;5l;+QylOU(>xU-;iIO3w*epFt>40%+0U<uc-d2g3Dp%%O
zU&!P12|FpBL-t*wF$P`uQ-q|LEu%tm=8zbq39w_nwyrRM_eodex`527D5;#z4Za5A
z$&)`PqKzpeT1x#oAGT*GNk0SdkHupgjY^!<q=wO=Pg`wp`&WP)5GZ3zYZD*$9NPjW
z4`HZuv8=foyXvcBsz%j20|OC#G95dK6X2X#iTDi*!?0#g#(o1htu^eAw;9DWRd*a-
z|BA9&9tk|sVm<^fqpVyTgKh`N`Imtm8WWd@1|1~>IXE+yp+_4P1U5J`HJ5>Z8Yh3O
zwqtZ8Y_}#H+h(U?+eyc^osO-FZL4FWW81bnHaoU$Oh4~CXU?qiee-AL{HR)WZ(JMK
z-g^}}v7!pSkO|PpR2*pMOwY`~!~>AGv^8>ZQU=<}1G#1Cl}*iEq?`?H0DpDtaOC77
zj;4mrmOwjELuXSSfQG3FK-AP2z`}n5VCLrLh9d`v0PQ^-EzK>Q0aR+r8q~D3bpI*&
z#{^*H@gJJMo=%qLb^waMKU__1fcCbgcFunx{;wTXOicmK7N!6*OB+*wh=QV)l)MCh
zN<v-@AYp1}>iD+^MHeF*OJjhnrLn1<lPNX84Cn~3`S$~0474+`{HHW0hQEJULQVif
zfRnwcvE^SkQ+H!i`+rn)0DDtMTT3UWzrO*NP5^U9Lp$fcJ#YpBEbWYKTulB+;IG^a
z_^*WQ9f5x>Z2#*1f++%>oSlpvE$y8Ff1@gjivK%LXA48;e_}gX{#65jW`8YBfW|KW
zl=Ls%UzopIXG2RnCxEl5yYqiPv5ZUsCYDb2HijO5WB&!SceMOh8ZJ(jcIN+?0Uf~6
z)ZEa~#KzRg=`YM*?0>rRpLGKMx9Sb;?QJ~%<qiCo)&I)D(%H$>#*6`undNUxW9PrI
z%`NTV82_0UDLXSDfSKvvauXN(|IoRbI{vFcRR7Ej_1`26O@MYb9sqw6Q!_Y5d7$&(
zkN~RxJC+&#`$+z8p!mOq;QtnS|KG^{uN?hvOZ@-3&;J!#+{MO5-q7~%0Q~#T0RA2t
zLp#9VTLX{<{Bvw<3>^Xg92`sI|BtJot)-2}|H=8kT5Fj8JL~_6`Crxl4k2V`{+A(o
zW_AYle~T@h#4X)TO%#7EosBI3W`;I@oA)odnw^QMqm8AV>0grn)!)BPOicepr)pto
zY;E@s8|?p9ncA8B*Bbuv_OA*UMU_P5RitSD57+iz#)^NZ!dcbB-W2dZBWcJ1P5#&8
zA1GmApgX{eo|%IiK+nR-^miHmt|K=Ko6rAG*na^r|L0rI(Aj^{(jB0~#K6SF4EX!~
zAD{nx{rE31Vs^$rlYiz&#o5r#<nQSHFTp=*V;4uqzl{Iu8vL#NKi>aZ0#j3WQ)9T*
zC7>~HuvJ=Sx^pdd&hUh&PJb^m_;84QNrP%(zJE2a$1g2>o4dl`y(WaAaf;`&&13P@
z{$pR7=3ulJ%Z7ilC&lzKkHCkNx-TE~4@Czj?Li=;UI}9QnMSnN)q?C6gf<70`oYeG
zlCoa$I~3tm2dg8@)f2TpscXJJ2HB%MlCN=UeJo}tN+SqbZNcdoMdI<(Bjw=O$na3l
z4P@UwKJBkqSaM!NY~SQl;v{De1IPOY<4?#}S56mrH$;Cg^z;uG(f~V>G@wYG=iCd6
zKUGJ`UC@-SaH><oszoD{=hg7Ng$N%hAdDd7J%iIbK)jYT#yIr$8fCwq_W_xSAbFEk
zS}mi?Rnz~WVI>98Tbz-KHRD{${njv7!%&mWqJ>v{jY@f)FcV(oQP@SP`PdH+BOxz)
z*?D3(Z>E3eCs9qkOkYEhofu{O@mm5^{@50l+p%P^x7>*8Bn&I6O)xo{XME1<3CZKR
zqp)(diWwvSY2A80Q*njWYV@t=d%Xpxr0<k6MmKsM`kas6pj0KO_s9+7Qwwu*b-t+4
zUC)uoW*1J`uQiUDkSY$a%fw&bX$0SrmF^EtAT)nseqL7>@4%ug0(9fO#A;{mWy{m~
zi{BZrPeoe!jVc<ju^XnBgXUC2VPwk*Olag%apNo^ln{LlM%&>f+ZMn`=B+^S3P&-b
ziu|y2T5vGgw&1k9CJA3+mqb7Kd0_MLYjJvveT|T$<j;b?Yv^n^fEr=!WFzu{!$<zA
zE)9PtxlCPY|LS5?rswIulcG%dhLeea^_@enLkJ}NL|bPL<dAl;=~%q$zK;?#Qne&%
z%63F_u&=ro#^>ICnL(Vu-{a=E2+c#z4UroUY=P?ice%pg6#vyUT9eAkyOw5D0=a2L
zexFBu_S94nx!ke<S5%v{Z{dv%QZD7s(>#A7+5_9F##)gcccLatJUO!Xp8N%F7>4Ml
zT;0xKZk{qmLfD#bmN%9F)2;f%k>w6GGRy8N84Qmqe3t{|8d)rQPB5RfG-ExY>&Aih
ztaYK2{)7`yQCq=0Sg79!0wy^g7L8wX!z!lm^(;tguegaZvYalrp}E>9sMGiRBX@r)
zo$XO2f%So`8aQl5*KnPelbI(`k&5UqX)De6*AML$3!@um=6uI@ZeRiAh}kV)8g4@`
z?<X)DE#055<X0*~tTns<Y2jPsZ_JALfPk#|ne3u3G^}aaco@-q&y3Q1B15}=g!l^v
z<+2EkL9D=cwsi40koxa2I@w^>4%>gd+WK{w&d|<OIW5)z=hXA3gpgu0v}wb)iFbOn
zA(5gVe1tOP$UO3TQUWcSzwA7nxkhb(ovWDKPw0Jjx--eJ&rioksfD~GQr)H08vwlw
z{dOO6&&mou8A~%v7v5#0krHKee#1%hTi!~2T~R-5Cgcf@B<!`yu+I6Kc9(z6jkm>9
zYDd;w^ySTUyHf*H@i-rU4P)Hu*kuN)$krU`E-2tD!Z&9l+_c=t&H5&yRqrIf?={lc
zzmcO(Eu1J6T#Hf!5L34Y!+ox5@bnb1M)Z!5$z48q5U(T;#z10D71CkiCP#P~C{C<_
zkm?KRq)W_BgQm;mBs_5pEbf2U^X*8qZ9h9vggjNPB6;AKPC+$Ye@fnUMV#j0>qSsL
zK@Xc%lN@M}TVI!SR#ZxL%FhY$Yf75C)*`yGB)?zyR&d@WV2Hqf?0^S}s24j%rL~V%
z-!23>Ds}Wg=~D4}d-2Zvbm=@>c%Dx`*8Fz*+M^$L#n7zZwZ+9CTZ(_~8>h%`i~;A6
zILfzYIT-yb=ySA_^2v}4e&2NkVYbYQnR!RswOm;;(nqNVyG@<K9`%*77vElwvK6$I
z(gBLxg76G3HyJ0fG>E6L>czoC$Xy=<y)j|F=RrTz-5=|~H!|DtV>_?1ULbq>;H5h!
zLjh7HG0CY+l3#g^tTBIKsGFZ1JS91RSyOiP2FZ+7nYv9|+*@1S>WX-r(}S;|&8>xH
zUxc$rh(B4*3yIXo<AL`uhqs*%Avd<BnbkM}iKGl^5z?_SU3KM~uxQ`pFc*Wf@Vb`_
ziw*Ondx~^#ON^;2e@z<XPcbUn6fWm&f(J9aJ`fXY^{jq&IWK=l-SemPZZf*1evXso
z5@Rd)7GmhB1Qk&?)O5*sAPY;mJon4Zg!xWGX(Am0JjhI*JN`T>8C%P+ZYWj9a0+sX
zfmyPCh2Gt-QuQ+Q_GmbBaSUW-X<@kW)Yzh*ThXbM9dDzVJ3~3}lpNsnYaK}Yo8})c
z?M24y8<m{`;>~{(n$ysgQDXh29Q%)1N}mxUPk=#N+|d~lMggAMB$HN!0<Nxg^4!_N
zh_m;op0XhBMeC#QY9>0;GBAIRnxi)f!f4JJI+aS$Ev(kC3A1G&G)^97m>Kf~krT{}
zZ}FYAnvD*ZxiW(>zY%u$r+w?Mrj+%mk?=q?+j!7i&%J+>U2wNK+1_5=-39qJT)NCd
z5N{Qa1mSPyybT~dSbmMl7pQ@SURXp=)^JERsvgZYieMGV!kJa}?68qQ4uqd%f@z2u
z;+a)!|3gB6_I~0y#ld?^EDQJ&V-4NR5e~g(kg|(1sBaF^p@5bTcNNqlwdhJ*Xw^D(
zsj+>6cD{d?Pz_aYMGAwHUJNhF2uvl(7~)uvMB-Hk5wKfHR_29Sv79N-biv$n8;em2
z{vZzd!OyaNR0aMOxt^P}2O9S3REDhAJP<T9PjI^qMZmnkgY^(brRPsi$62q$L=39+
z!u1m+o$*Ja{05rr>p8bzh4F6KEDs|)2N_WZ0pEY}F1KOUWo|g??k_l&VQL13w2HFW
z2o~aw1vV-~U$@$DOe?K7)!sHlW?TtTu)dr5fto59OAvv{)Y(eK+#`g=D$7!X4}^9*
z(Sw^4^3h5(7hEVI8U#b;zCMB;wybM?Q0A{Qw<f-mNh2=g{xkRj7RzMxDYtS85C4?~
z2{?b%Rvg$#y|7IDfRTqC_+m_BWgH5%yXT7UlFGTh1tsgIzWya;PDxXSqA=YDn#5i0
zGPgO&UZG%QQZ)ylKXSYiRt~hQs?BFP1)AMJ5x!aIyJP?v{?MmTS;{rR%g^{S^<1c3
z=CL+lgom}<V9Lr9J~h5yM&*yR`FJI-lqP?>h*vGxS57Rdl|=%N@OxDoXedH?d7Bg|
zE{gKd)1IWsDkS;5-^qVf;`4ZlxT00PKU%MD)MaR@P_l&_F`Bjoc*V`}tEkoPnxmS=
zaBw8@7!0Yl!fShnfg{P@roX10yc>(V2_vTQvRa9ML#8Bot8{mbmh~JH7oa~5sU3f?
ze?w_QVk{#-u7-iHRw~0@YS{Q)&icLmY89gq@$HWG{a(UMyeIc9<Ipu7#+gTVMIojK
z&uVsfQ+1EY-~y=4kRSEZA%-E{;a|04-~_8(XDlGaJD!&bzYK{DoNREb>u8;pXpZsj
zk}P>P4+Ghe!=UgRz<PgL3^d*kfT@2D0~09jwZPH}bKIfBM}r*F`gYORqO8>OtN|p*
z8tOeD<~bgNAZV>-JsF)%oO)5{rSoVwIW5bnpB8L2Xet}RyY6-<uj<inpSxz-C*9Rz
z%BlW5=)M}_7}-Su%OCft_ebszdomSIH?#?C5dYewMw^Jl7Cc}F9mPfng_eI5hSZlz
zH;85*guP*x=<@rGyI|Zy147yNaS^lP3<4Pu9UfMg%vGxOssm-*t$9_C2}N^wO_6S@
zGlU#@X7eTAjHj!}?~g)76w!vXNh7;#l0qzOqqKa(C&j|sb{EqBB6Ss~(P^*kObZ|E
zhOU`BVEu{|G|?kxcSo7P46J|W6uY!TyOiK2h^*A|D2sHEz4#>kR!fRDXyEH@3OT9T
z*ckbXNr3m+a?tNfJCR&WhIPWj852AQF8%PG;t;VnrgI!j@z~#KZUc;tM7U%RH-Td;
z074XlAQ+-9p8$fIc!)R-6^&3E05w3$zu_eq0-jb?Bg^XrgU&ev^|aymmEI_Sq~`fX
zKlLZ9+7dH}{I?>S-<e>rF|q^=wYym%{zbS$kx2f-=Gvd?MKk7ix|}}H%qj}886K5t
zD_irkq0U3^8oT_=fq$y=RncE|3p5`v6AlTAXHyF<#f_}<g}BTrUJnue<X}2d?tZ^m
zAAZ{;t?uA#_b{s~2<q(r_<rAibU@)C@H#r~p%ryFl3s~uKezT^0_`>c;X0kUc;U5E
z0d<@?NCCa20>)h=ZLfgKsEE}2^y48Gol*i}XIqXPa&4~Gm#~6Ngb_8=(#n%>aD6+`
z^#pjKL1RxzE4S94Z}rkliiE)<ba>Prh9!X!UJwYsCM-e@Nm8<wL365qxF*&>o5*0N
z7&$Y_+Psgh6P$?KIO<u73md=~QPegO440N7?O%!N%pbUH0ap{j_+CJEbD!vh=@LW|
z75S;HjJ?Mz0#r<~)*w()j=dt6;%n14^;-27kCtBZE<Qi@1Dqd@l}r^nzs4)AnG6+%
zU(GDtze@E+WaK;>6>ctnU%o8OZw)$X7=uilav`PvAh=;-6G7TA8L~QTh5jr~kCdN4
zJeH`q(6UfnlpC1lJA~|9b6EH|Cdb$HDjec>D1so@ZVPF!S3k1kqN)pTZB@+7@b*rM
zv9m*whC$Hy_*x1^qboZYk8&~*5Z9+nWUe{B{E-<5(m?NIah@E1;qdt9=a1BrtEYoO
zTTsj`Q5z=)T#S;g)2hLzgQ0VE`L;vegWt1LA+IP=B3TZ@-AZNL{Uexrj2H_}oU-h>
zB|~IwQc|gNXM@wz5VMsyJIDH5TM)C?SFK(tW*n8=AK!pJH!=s?>=y65po)uNOJ5)C
zuWlu_Ym0=-y9Bj=)xpd)cS6m(EUfei1TGLT!H#jXTe^7Mwso#jQ$1XXPKU)s*#+<X
zN6$sVRy_HGOBWF_NX<&pS<Vqr>HZ&2*+AVeFs+zjl0B9Fn8U5gw;B5jSRoe7Z|ETL
zx}6Hr6Y;q+dR$G=utGg7p}*;|N9UoM@W9m}$A0HZd)R4zf6}^AddaQ0I=bqw_g2C?
zIBEC0Kd8idWe?bB(oxERlKH$?YwkwosM{Ci#}Ak9lFE?tiJi~jgY^|2cL%I-q4?q}
z(Zn{GF!npKGmXTSEzLBKu1N}BZ9PrUbC!LA_O@VsntPg~!Ke?O`SF)YBKWOi+uVTm
zMPtz6P!^tlF?3Gw&pea;fuac(AT>1G;e$gdW;|`1U56?2xNG4OQ#znBl-)$|75G*a
zo0G>Z%zfc9cRn_+cseD&sWk^M-W?J(53}HFQf3PmZ86t|db7C#Df#1h!ZwadO&B~G
z3FL)SGVCL+vmblMA&?mBPN!Ha-}Vpu%~8<7=Dmi0z`#?{I8$vDqRtmq5!~YFUAf9X
zj_J>M4HmwiUl@YHw$rAaJ~H@pCmT@PUjD+fDuSt_LM(SD3}O^5-h2=z(}KbME$vR!
zu#WXvYSL2<62t)MMRj7Q<U;VKT+@67#M(ChH`z8I_hX}1>Zh}o^j>A{<#JS9k)qoO
z6Efs~O}(Jg1tcS90S>>H4|saWt&v`CQ}eopHMXavBrmZu*bgyZwcU80OD!z-)3sPT
zTY(zxc8Q=csV=jE=B;JxO^pn&8fWuZAR9gaD<X!ft3y8ZN4lj?__>nO5n1NJM}b!4
z8VtcC21w`Ht#C)ZgOJMYL92=XYn<c$K`3B<`fYUh08R0hHax1kL)aJe5Mhd`_O>~G
zue9%v5<&@0APtzU#=&;soAytbHe=LLvGbC~FOCImJvR7AQYY@935H?+!A$K}c7{4t
zwhg}K4#wNaGz!}+#;NZ1RRwkpy#}xE5jZk3R>cXGK2FeOT<bVvbdqh?#yjf!!cm2P
zzby>yH<n8Ug#(vsEGWmHr#PnK#J9i=R);Z2KE17=wgF@Z8x=hE_L)p$?fPF7e0KmY
zXa!70^x43;7Bf$K3s-52?;h-9zA<HUNgD`OaonOHLx0v;JY)J!UVf5&_VP1<nAOuG
z3Ob#lMEmjAM1MY<3ng3{AbBDjn2-&BqDfOz52VFIX$&Yz+~IDi+J?%M&JwPrq0Dqx
zw3^=SGr55UvNegLWCl{_i7@7N!q$U<OGz3ht&hAO1Z>N!a<0pysO$XVmr1^u>+yyC
z)2vDrBAa9;c1`S8Y1PBQaRJD5-SFvVWZ`_0r1Ljf1?SH-o#;c%1RjWue1R5!S9xWl
zt{Wy;uO%^IoH?s((#J8sZ8tB81e0P$HnnL@)IYtOI7+&~J@T4;aUeSMBi?3~MI}#)
zLvBag5ExDx;W+G|A8wv@H^=WU+cgEj-y*~}2oqH$`_qf^KCGCrJdJQ84?};TtK+A=
z&=lceL%z4a(Tg45Wk0IxWzjW%MlHDvg0JYd-3m%_0$mt&z?x@o-?W2nI_%Op1Fy|A
z9%!}CC^dr@<d~HmXoG=<I|kI75|X`L`fO1s>B7S}qT)857jSf;<k<#wAhDYIGXkY)
z*RNC%+qG1d&u4ispO(_mo<s~=2kGFyE^1Nq=S#R4*(UL#96T+ZJ~bPEuSJX+yk@O#
z)X|CxH(Ul;!ySdZp3Mh@dH*zzk_*Ayz)Ag_Yeb`>F;y}xzNj<*E^ZVARj=och?uG$
zl^7@gwN~EL+U|s=K2tBZ<}=x^-k=ocU6j@2rOPEO*<P)PDJQo1t@u`rnaMAAV142Q
zn`mzo6MH@2w|15qlhuWPZ9kN^G#tWa`sZO-iX?f8%ybzM(3KOf3WWD?v2$E{7F$ZP
z7f5G645A(EkvKQGPxSy7ckRHafgU^4BS$XFADorf7zqFhB7Z78Lb%i_81u3{iqd&%
zskvUUYzMDTk|`jd!gxLzrg=q{TmS_E&p24?%|(TWZSZsRh6*-+Tk;^du}(xpJZq3K
zF{AyW;IHrM|J+$>8iGA8CYIPe3YwAQZW?Nzi8utSm}BntcvU@%d=2BC1-In^eOja+
zS6X%qm|xc@Vs$)qUD{_e&J(&+4?DzmzFF&6g6cF!Kpn|^I#beh?D!inowQ%g2^g~a
zHw+6_yUN}4{BbjXEAA@++X3c)g7Kee8lr!#g1kdnrW~YUTdHO(%tCwV>-oXXNxd&*
zaO&3dXrD$z*-wB`Cz0dPMJ7kmTj}IvE*Rm_W;3Y-;PpX$8OFfn(XL|#>V5*;gNA-B
zRHUr!#dcYm?R5kA+i|XAZ34<uNi1f`r9APP0<a4;XrLp13J^Gf;MBaJ;z1VMPfi;X
zauy_XHh#lGp?7G!F_rX??=lI#Sa<!zGp=b$MvL4in}o%X2G3GXr4o+3hH5OnXwETQ
ziP$ln)qZ*eG>`Is08ZgtBewvKq%Pkv5rprI9`?;CYZ~6fVgM*$?5Vq&2JxKcdbFJE
z^?4KSwyVN_ug`C;u;8DHlTAt_*vJCe=jTSZc4M!Y-3zxPKA;(;mM8-p#RfazgmjBE
zf98H!!MC>ytDF`MaAT^HswU!TsR$J46D0=lb)VS#)q>P7<%Q)CJ*Hp|pf32~UaMl}
zu9n>L5$4f|jpfcp(()4!lII$kWQn;Ab5-wnc=)J)Vb8^eKyGI(N~~=%*MpxmT=!W>
zWLkS=l;iq;PdhpWn}~0k#y#@H9Y0xX32KtE;}G};*%Z?=QiNAq>_StM(9=E~30l#t
z*sCJyokYE`XFI}IH(vOIN!w(XYuulrJ+sUPezQ153&mP4N^JWL%>t^G-5KKc9h-P1
zz&wV3=>scIeBBowKj?(IbAZ%0C(01s&S)*~*b?1zc;`o7Q=vr#2MzB?a!CpE05F*(
zK#&glk%Wv}cuK~xo+}U(*iDh0a#r#d!uk$L#MJV|0{5)@nKT>uV*6RCBzKM==Q`-G
zqPl0*t@IscZSJ(A`&U;NHDhRW9y0Ffs7l{|0B_Rh{oLj6Zt$Y{TT;<>jA1+nOTy~{
zX7t^1%%uHj<J2~FZOkc6Ez9)w1yuh@aNYq$0`9W*gebz-Jr$SgFUcc>+Qm`*Y(sSN
z)CDqo@0$op6x=a9y<Ca0v7~7!$WhD6c0u3ccR~H0h+OoczHIHD-PO^Nk1#mJXJTo8
zc`9jmR%TQhqC+T^#w*Xv@S2ZfPZ>W+X|=DSRwm_2()$JzVOILjCL_faXPzoQx+?dj
zI|j8VB?JVLiTqfvuc#^4=n+N?O+ufb0K*D&INk2kU8yj<A>)&ZAkn=R%aonJC$%}}
zH&Q)f`61sB|LHFzvuwh>y<`ZHZ~hN|tpV;!&Q0Wq`Y<h(3$FyS{B$EQX@)0LWZ7Dg
z8k!N?UV_^iEU}u94X2gFd@!*%QiagXtI|Ci`FW|~{<$@qs?R|pt(=Cm%pY6h5nkd;
z1Io>(MOd$hSgm$6_$Q5vzt!n*^Qfvd(L@(F%G7UWo^29MuqsPS6YMM3Td?kbyF6L~
zLkcuDl?cYLb(jO(%`0ABom@0J?=T9H?pmW)Ww5%S7P-uBrIODnoJ<J!7l9dc=fCq9
zq3`%zh>K@!bEEku8i6aSJ}lzwN25}L^y{<4u^su?3P2%J?^(>x^OEvxvGA;grQgAb
z3;_Wb9Rd?n94l*Ur8{idTb}NJejauCGai_#h2j%wFXwIL9kd?FcAx5N`2(fJO?HD+
z(|#3iL3sE)y6VsHBfyO`aV^~_ttslueO}_{N#Nd5U+W#qW)l12YC`?MdfXC6fu)5g
zx1AS5<b73!>b9sZlU@#;RCZQr)2biAfpw_K<}G*#PQ9SE4N2Ou+KFU;;?Cp;Wt^xo
z_-!n#gIogmH_1^SrU(>wZf+=I_ubPp*}|DHSFtvxfae{kXrc9lLf~M+W0#km+XgK)
zPffd@`()?I{R4OAcwt~`cHe!(I4k^`e^D_T7HaL;-gHs9A`ZuBnw!AT*Z2EdS#<To
zZL$}#dj!^}K&M!5t*sb;nfPsfPUm-G8pAl8qq$tqqLVNDg5dz_;pdS3Af2qJ9&s)S
z9|yT#4J@Yv&grmD0QVDZNt8?d`njkW4$Es`o_DuEstYVzXqsgxer#Pz-(~>2>?hzG
z!y^eCo<XkG)jm>l<U)yf8{&5mr1Fh`6>AgqpL<$kk;eqkJPS{MT;o@T6^}WAb`x)-
zQu4tE3%;vPF}no#hwD||w+L<SyijqyuJY*T#dtp{@Oh}He?4s6KeL&eAwQ^Wq@9GB
zBy~5p7jna6sZn*dg6o;K=^~MI1#oj6?aOh^pP+O^#T-aLERkh@CVfW=oxkA2Y**Eh
zAwes;|9xkigX+J3A((WME;2OBJLUAMRLQ0=!X<?B&~abSd;iUdCUjgB_J#smvmJ!b
z%cpwJZE^Ed>-nf8hJy$EjM^L8wrG(?9JqF_=67IxnMy~-EJWml9P2q_o1aoNYi+@<
zZ=v+WzhXRerM1eGVL=v>6yt0aN~V@yYOic_ZST%I$aS87MFrp6v)MayhF6Y}Y7E7(
zoC4A(sa&Zy?y~KpBa7@u6zntcV=jg%)Gzlq_yg20Tu-5+ajC<uC6t;<Dmbc2-PvB|
z&Qz}voBspRqt^7g@PPt50Zpz8RPx#^@Hhv1CT2YrGgEou13iuET8aJTEv+Z$K>d#M
zNMb@<&vlu9<Q7$Ci9kruqVR~mhvrxCZdr}47ypi@uWTF$z#F~G#cP=!2peg?NonS;
zW6Ax-P(@KBd?foD(d?%KOgB9lpE{9b`(AbY9X@M-9DFznyLx9d3MDFgA#I(HsvNH&
zEI|weM0F}&js9Zk1o?6b2lX{@>o*n$SgJR{ls7nk+LcRi$yIXXYGaa0Vf9&2G>8y!
zg7vg28n1t}fiZ)9<hS279Pvb_A3k-zB{4(2!ooF_l(!G~#bI+FiN-Cjcs_s4XNG1N
zpdil`fPFXI^B3PvM+E2ZGWZFu{@ZYGG%Wc7)M>c=swN3H<T_{bTqG$a{W}%Mn-*V5
zAIv&`W?$5JRI7NO@QgJnn+U^wNjnAFornvb#i`BUhkh@4^-KQydL>cjtoeJx2^)^s
zjBv5H@G$o3#)QRis7*OZRF=fvQM9*w$e;#@YNbR3boXXR$kxXk66jog%v_yzKw3cE
zkC~Crd9BQXc!XbkGo~2W)08%n`=u}EtG{f2GGO|#fwU2+J{9XR#&N~PFL_tj{-i1e
z3FOx-jxhLGYr7`rl4lWC0v;Z#g*n2}^JhCM@QLCPa8wiNgwq``m3$c;vUwx{?2;&l
zwLV6iWx=N}uyOAY_sd0Zzul~NQmNPjsXOKh0TcAMEd5VI%eq&gmY)&x@xL{BGsvHR
zVh2dxWENo)>;r2DJbo4x|3oUR38RYBuqHeSDIE|%AA0c6LHPXMs@xv!SIMm>YbHX!
zQg*D=_0jbP1A@aEs!7G{WMx0WGgLaljA;jsc?V@80RIJbhM0>5W%a@SmBrVyV8}bO
zPGf)I-SYkNun6C=i84HyixX41>gXwd9Pr&WjZBK$L!T?8{Jork-M*0TdM(SkvasNs
zHjNdL-hYfI*j&e<216<6F~1XQyGaq({{2^cPWX`~idh#4RUU>xODo6jYHkxD)xxw0
z6LXIJtzJ<O5~)e_H^kj^QE1qSCm0)RBfRSgbUbZ7jy`_g+;pn_z23>a*rua@WH?S1
zE7*Q!Y{*LHDSL?(yN#+s!O!fpm&8oZ@>=Lg(vh!e%-1pBm4urjJ#gQ~z8E21d+DX|
z@(vC1%bx-Dr}R2C-<vBCrCXBJ*l%C@!<3ByNIVCRD?Wm}@-td^=(}+BV7ioyJnlRo
z{*|lZkZO`OJ(dovVRU0$@9HRjp6RD8!hGdwx<onJYuwf+*Uts3sTJtWhkWyHb!gF>
zV*F2q+Zjf((`@%#?oZ2hI?$peeRM>r^1_|<;6L&@y(9C>)nxZTLZ88J2+sYR3M)ez
zKjEp0p%=+W(ay<5P(_qLFOO}U!9;&@M|K*<QNS2xXSa17xajmkV7kqJZ~LP9!JGMO
zl8j@$v6LzhRdhxy(r)(grF?4>q{!o^Q|7r!_hF~E<WPkflblNm_N8n<9lIX48s_cn
z-Rb;sBqvVfA^;(0-^5`Ifoha;G3Y6H3eE@+%?Ks;$t)BarDd@1POzDDgrrvpic5Bk
zy2aFL2h8mx22R9ny^$Y(`T<*v*6d&J7l!Q`FABtz^(jYaU|LeU82z0RLn>;`(<2OH
zs|-UNGkIhoRx-1?Hpk3#dPXl1dmc#&8!^>p>o6G4iSW7R|B%My9Ue`wT^3;t_Mbch
z=IX}kw4$n?robrTeaz*Tg%Yisy>xGZ(N=~F-`x24Kp-H!UN%^Nf`UF%1B)9~BY66t
zBqjpu61@_T!XnlC5W0;lrGLzHwIN*P_SR>%b(h&v9>lFMnRczh`qBN2)V(6eQy|(V
z^(zy0Pn_|ToD_C0-c|f5PoWh*&C<f5EM#9+^Q=mjwr)vOcN6Q3h=xs%Oa6-SnIZM{
zSf#-&RmB9+9x)4lB*lM8-Fb<Qb}Ha2hr_{vcP7Brr4ABX;%Odk(30o1hawg6Qm<I4
z4w|&!$S)=<Ag1Ex%@*hTTmi%5_iOOLIC>1#*3{^74;vSv`MNX-3yM`C)mnP%3qdnP
z9#yXO#ayV%iKGTM07gIy+V>-wX@E06xrE&3<#~#x#-t5@1gKfFtPZ$0NKS=`1ZtAr
zPi&;<5Mif97@LQ}z=m?BQC>a?r?g;J-1Kk0cSLYr>RhBB_0I!HZ93p@uJFlb6+I_z
z3q~ISf+KwgGk_p3*2YT|GrzHC%F>@LpWijk#-0Qu`BCX>l)JJb>K+<K2JrBQptuTl
zsPMc(g)u~bJbR2bpR@KD`%RKb(sU`42sHN};FI-r-~HcBOY|t&ve!=0=k~_y>>G74
zu;;Y0;hMo$haxuaG1Uip4tDp^>8wdV0ZW#L+RF<<@KP2U%2c++;2GZJpq7gYRDX29
zNTogmF>tnZOao#YX&~R+9ZFHUq-4Z%8yHTTL*`(A0q$lCwgaxclo;q3^~6sBPE-3i
zpZf~@TzkKLo(1J^e;%adZiHAb-N`LlC*WxskM0GhP7QPzn%W<53!2Xk8CA)LZ^vh~
z3?b)zx4pl+N4LS+1Xn88{=R$Os*IxJC)<8>4mHDHaW)CE;k>Hlx7|_<B4hIz&6?z8
z%5nF94fYp_-qH8wkk~IZ0<WnMVa5+(3pQMk^625Kpc02@pH0#R5m{qCNEk=(&RsZi
zM1K;?uKXb0l~f;Svh<}ZgG`Y5x>0dsO+9$-148_~6mQFRqqN%?DWh{@)CO)2I+1b{
zdMMdT`9uvJQv~epqp9*&d4-@fdb3){<Fp!oEfSGSpU0C^DsC(YUNA>Z>TrB$9DQdq
zqlZxb)~t`|AN^Yyuv7xStB->v8SKet>VD`Z*)(7)Hc!ZevJk)jma5gtW!TC;XmMgZ
z*@#WW3-^=FQH{?yQRXRL?@+F395vs5g(ZA}cG(h0GZT1RUR8Tr=%^BI7MLM6bD*bx
z7*os!_oayk+jzH0TQUauy<r{u#j^rg?)<2B1gJWA##~H0xFWU$$&^{8(LICbHeCwe
zLl9o*Xn025)(J_X6XCnSgxvdbydtoh0#oW|oGkHC!s>y>(@II9n&3h%$|v*k%vzG0
ztW`z+VS$xIuk86}k9Vc093g&i`G<;s=Fi<!bVQUmbc6j;f0jdKgbO4M71ukG?8SJx
zfzX;yqXI5#Y(_9;ZKRFZVNQeCkG4r({oaS0xXACxwA&4a6i^?*8}iA^W0k2@VW%B|
z2%40eC(u_4ro0*XC{**4qM@l577exrSE>5)1P<aV9ev$um<JOr#`CEsycFAiD>HqN
z<3Kr$A?_QuNk@$k!5C$B8d3odlV6u8uy~Hb{B=yrnmUN`Q`kC!Rr8KjnG?S?>{n_f
zp;=M#_J*o|b*&}CIg69xLq<uz<ek?@2rBQH(oO~7*1}6VixS<$ge#7>2Rm4XU=HwB
zjHQD!1sJMawuixb&`p)gr2Q&?S#PE9jI5U|-Kzvo-LB)CRW?9J;`h!e=O+DvI}C(A
z))|SvCRpF}3R87|9v;g}D`PNXD8M?4i44>a?sW@uu8!uY*>^!9<ffujn?};?ettQz
zEtMcsBcg3Qc!yF&29OE`(VU##*`^P{$|McA(b4{he`h4aFbw8d;68DGn#P-C<>YfF
zn1ESR>UgM>fV=kz4gi33mH6ww_%{f6y!J=^cr(D2MAb0!K`Z@~l%Yll_*s9AT70-?
z6Hsek*%68}Z<BMm;bF5-i%8(K<D_pq5g8}siz^u!Eve+Nj*QJuJgtS+6R%8N_q@>Y
z4R>I9pk#+RrV&$e#GoR7NzHG*&(UHe4>;v!91<sTU796#O&~He6^F>^&d2qR$l>2G
zaajwGGzen}M!IMD)9)h=4(a)1^jfZvCr2$fKuu0P0Ir<4dsE_<sxn<p`n+7gVQ1^t
zv+XBBRS*9{)2~=pCzD%`$%czg?`Ixz5A8Ekh$t!3urlLE9>i6D?25x~gULUEwnur=
zw;U8Zy)zBOT0OuMlsT7%m8B7$_>a+9Fk)U03NOsD2da&$A6q{(2&JP}PXGRAgJ!*$
zFiMlII6EAuSFEyHFzMiS59WAF;J^DKQBdT4OL4APu?k`$03J+Ai~Bs=r?xYe5Nn;=
zS|sRLvP`U!07SEYrr%Lu+464qG;vre4zol^Xc&D*pU4Jm(U~j-6Q-W-+W0Q1Le!gx
z7?tKz4kB9!RxVp@858m=C3GNK9!rsI&^jAzU!6&N;VlVN1or19cEj^VD~;81)(aR?
z*ZySo*1C#K;>3rs`F7dQPCROCQPFkz8HB2Bj@q08Ru?&cUMS9$vKzG7tnUX~TiqV7
zVLus0mgiX#YhWxP@yOj3PF&Nw=o>W^u8deYk9HL_D(=BR38+z0FI3_3ILBq+G}{J;
zG8w1ps0q?(5rLnOln>;}D)FfGTxC<)kzY@2|FK%QrAw;Y=0S3T-;nUhHw>m}Jx+6W
z!6}ZRvKX>|O@@1JkkGe$9S5)({x-K(Kz#<2XPcVAY3f}QJnvjlQ!5j!g`qaCVbWHg
zv2vV`e+K4?Vs<RHyN|ZzcF@*0D9(q@#ak#rTf=77M30=6sYRTH*JeH-S~;N7gKes~
zRfbRNPcx%Aw&gY^_2LxgWd)^jP#dN|heP|96BdPkO2lTi{SKwraIeTpznGSsTuuD^
za%*T0At6h6$QWOTu1b}vTZC}N5@h^)G`#pH<5Y4L8h4C{y!Qc~xM2s)9uo;9yWWOF
z6Rqkou&|6Be)Lz;`eYCbYH3C&ttN>J;iCo78GrQ$x4b*Q<WDU_kaXGs=0e<z4?Wek
zR0RNkj@@c2__%US;u497;BTP&;7{nRFP|YhWWBynK}7@4?luK2Ii<@l@A`h(Ga__w
zLY8WPxca&6D+OV1yNMH)*iu&2{ksWaN8>bDvea%VsU~d+v3j`=F^LC8RY<)$?xgId
z|Cg7KPsK84<(gyU6E!D@RW4o1-3)4bQ*T9oLXmvWM_ugtP4GpiDH-cHfn=wwk5pw-
zy4pmdmzXS~Z9(JDQ)WZ0xTl4?UahHd8)Cy7E7P6xBTJz&%Shsvks1244d8-al3Z7M
z%BOGpSJuM9O>d5$K+>l#K}qba&(K9kogA7T_#p4k3n}PASfWvmortaG7wnzHt%vS^
z<wQA#*h%h!p?UNvTct4ml-!0i>JFu^IJpx)Q}L~K-C{p2ozGxLMP{*{qR+M@wd(Gj
z&(*uEU!u43d1D77{3}GmN6QrCbmWJ&KOm<f;kOF5-ffeHfVcbhj}&_+LeUo@9#WCq
zb+tEVZ{o|`I8O9;H~a_H{uh^jVrSxiEq>dsQA&RcbI!5GPcnncfiaW?2kx(^(sSz-
z*wIqn)K80=9S9-LMQp9QQm9~f27d#O5kMr)r71Hgr~N+O(MG)YGVa)xtrO^-r7d)?
z*_N_krkwk>Q09$`+4nfv27@-oP?EX4)_`<<_SWTY+>f>GFiu5<xk3#OQ|m{6bWbnN
zM5%Ce(%_)EkhE$x7C1G^si)*zcuhJXLB)CB4jyy&T>=5ooBX)sTbh@Q0mKycZWns-
z2k^L=le&IP<3keFhPBQkkO7R6JdE|Bg7&k*h*fGD#5upVoqNS-l&qNqw#jDi%8Tzi
zyME4Y`KY9n6O}%s#(`AN1<}BNar6Zo?|ny!&1A_=ZyQOhY3m$Kt+b<@`}-GRJDpmG
z{}Ne1S7jH)X1|nf)47hTh|G6v)wdPX{TcfC7p=1W#!t1|Jv9roYa;l^8ut<4B!$r&
z!PH8#4?2Tn5LKo4wy_m31%Vv9S!hE>J!Sq9*e4KA^RTx~Z*StfQG%p@eQc<plw#RB
znWph8Iu*k`|K#Zaq|b$+kzZQ`UwJ|JSQu%q25cVktj8Iju>LAfdyNzU9g8yTr=JM5
zcKCJc1@VRNI%izmK{xjjN|l@^!QgE6<1SzD(#j7zqkMtv6=?qO2uZ$UCnfBa<u8YG
z&YEdh0xk?J)<&q&gxrdM14ldPCKZa2xO|zeSui&$5CVnq=#M5f1DwsuO6LNjcnL_k
z0ho7WEI!4z2KMzw5tTm_wrQihJ#(>aQD0Ys&{2?>V~bwk1&h!Lb0Z673RuT|>h{}L
zy>hEk*Li=kshOh)X@%Un*>v%vt0jX(#z<w9aKVw9j8GGEA9#d+;#?@v;U5?$ZW=Jl
zJ%73|@CRCPnV$k+iG3t;{oWGIux|o?q|?$51c#tmgGOM4ED$4m5x)&_J!w1*jmf;&
z*fZ!(E5)hY(++)XaY@*?K<e^K9f<XXH|iTz;(@zRE5nx3uP{ZJpr4V^GHcC`Tvl?}
zeiN{B&*j-h3WuS8KsH?ZMF)F75pxVyIbNaY5F}g9LC(dnF>#IMnpsV_tr=3V1MM%J
zPJ-*Z%iaIu{EW^7yqO4{m#>80fCmEPgO2S^Rh%7MdZWp(TDE|}0AJ`uqv*K(5OTm+
zG1)GaKXO5q;FUjxCBSBpIoF7Chr8U`0F%PIT?~z!B*<=m_NJMgAeLMz5fS6ppYN3q
zZQ~@5d;uA5UZ;yZpl+JU`V!vBC(BuXZW&|wuZzu!39&rh!?{GJ(a?~9)-goyG7TJF
zA4IJ2%xTpaq@r+jgXp`%L&^N0Mt3wE;Sj%qypSKGlYmgq9q37uVi(J75Q<34J4ltY
zz;^qCh4J=(DxTh$pgG!mQ`$FgoqX}Ddv<z^P%3Bc7f5+VMGIY>=#&XQP6X=r<NWy)
zgx|^(*j9U1df$+Ws|+HBQ3Qf1b-%8IoDpzvi3VIgTW?{@$;6Qom-_V4$#SEoKn$kb
zRITnV_IcSo{OK-9UWcma+Heh7coq8ei04iPjFv5bs>md7?mFmO91l3u%B)t66%thr
zf|GK8xE=oN9ko|(8~i1gBv3=UCV#R)GwRxyXW5?S%^^`GwFXNHQcXFR6>@*z{$R8A
z#?U!_hZ;COt<#UQ?(kcP41JSVFiF5>-5x-(a~9%!bfnAo6($A5oEpO&QBx@T&2Rk8
z@g`S)^q?8S<Btj{USJ3PitdWb>yAyp^#?gem%RBbABQJl0glpXrkSedVc(owTGF{t
z&7*DOML|MhVN62+BoS*Z)49_Sgxx4Uyd}MD9ddFlE&~`94-l;LeT6NERo*?W6@~Tw
zzB;dmX1{aUa=$h`^K|c4PPL$k2Ah~AhcoPdo1BUR$lyvi_=+K$Y>*g~Cw@-)Q?cDD
zoqEP{YEO&oOrurmP`Jk%&DZ0;O5Q|7dysyTB%cHu<xvl*nIr3E5eNzcw=&(rP``w(
zjt6f^7lxl5!!(PS61hH*Dm5deN-FrYLM^A*cWnLoLnc-|=EXI4KBAGOf7Ys;M)uu*
z#mZ?-0q|{ZA7buZ<N|cXF+R0lp$IQSM%(?4<|<7gBCgV1jl*-A&5LH=Bl7ZoNrQZP
zDs2g%A}lV403M(@w&c)lfrlrR>TFRj<HsfdYf!f2tFAC2>WguBAeqa5TiTk(DFP3(
ztj9X69EY4b@yio#9<I0%DPeX30J^<@<T+#OosDPiI{dXP_;5Z`E-A0<C5bkFX_U;#
zN?FE8D5ux>2h(m6rRjnCo0i>_s^1^&3$LM)af@Z+NN|X3Hx+7{=l(IjSU$1>jMqe+
zlWV@Y5gDc)sS{DL!ft0(Hts7?!3me2Yhklor$4vpF8aZHuBVbrj3r_Oe(chJu3X5n
zau}VXUty#Yx*J7FUQjYuCk6m}YD7(<o3Bq1-J7cVC98FajX&SkWr21wYwgu_34Ii7
zF#8EUC^;M)zTp?G64$mKp%Yksf_N_EU6I6=aDN5^2sy{VQWCZvB$pT?%FC$Gzh0gi
z3_MMV7jE>KuuSa)l$$dL0(-B2N2=RNp))RpEIF)z`D8bqlS9pPpY(gbCrSCS72-OH
zf%t6>S2308ktKHdT*a<pVg-D^*fA?$g21MG1^lEP8<tgux>&NUX*7<Uo%WCKrB!RO
z@>3_)y>=#l-7omQR%otnhg5CeCl=$XjEYg`y1EcOTgd_9qsJeY)DmoeUM41KLbpUV
zgPJt=PmREuXd*1SYI_l3KSC-o8mo$B6eG)5vh9rpx8m(WNKt4FS)SoL)06%BA(j)|
z<|~cU>PSs(><<G2(|A}U<@WNgh_QmsVu82yw8Cj=v2phdmA$@x-#QHKOZT+I9sKTY
z@kVwr-o+Dv)2FtGV~iSq$(YGL(<lmsUzN`-?fw}($tzdp6s%00$y^$?ScV)uN(ADS
zOLO?Vi#b0L*GDK5rEMDlpdF>cdqukP?B@cV!2a(H%k|qxbdWUXYE^xK#W>tF=p^Qz
zKJ+Y!1$^xlC*kA1BHfN=Tneq}<kZffX~p&*JOQKq3T}>IH@BI84RdI9KaJYOWJ(X3
zxzZ#uZ}}<amfj(psu4)Su8YN~T3Lj?!*vd@Se%o8*LZu9x~UEyy-d4jq2`=86QAbb
zEC-cR5wx1mMe@E0jh5McqZFC)NNp$|Nrn}maomxV53*0S^~1t$o;_kImwCk`-BA4%
zVoZpI6~zltlLpy;3wpMjteRj=soE!d{**XqIeVucCtd}S_U8zWSN&S@wJ#3^;%d79
zsV^?LYy=Zalhy-znO0xwW_1QJ&w7~O-`DG_?8MJstKf@2>%+PEK!<ciiH4>_)z2Zp
z!q})T@LU21u#geZ6tCuhm#<%$mm%p3B6yI4MBI)s!Ct<9+2DSwOgj?7^r!tj#X*@l
z7l#K0au99vlo9rB&zn{*R3&T{{^&)^5bjEc?ew8J)w#lIK%VwYhlMyuW0%nf)^6Ef
zgvXn26o2pomx4-*b<{{7<#7d5o~Cg@hXudA7ks8vfPb$*N1SN@Jd3eNifQ4L#{eBo
z$z^>etp~<`i1{RJa5r5#&SDfCPw|wC0i($S=Dq=t1$?UVGyA}6iF!dGsk@2y$u={A
zc!a18C>2FGWWNcyida45=J596z|MBwYsm0Ce-lG4c9O9g*g40CKV`uvgcp&v@7Twt
zphUIKKzWRy(Or<eo|@mFd-V$-yjhcm05N4RqYF8I=7vE*i#b+hZvc=75NK5TNGP)I
z6-SbNlqvj1RKoH0uj>jOze&dKP=MWgM(&LUdkL*JOnD(0^{)t?n#{x}u#;897Xsgg
zp{O9z8FlE{58|9Uka!|WN7Fp^-;1_h;$oOl>oAF|nw$jXdnre_lJM0wcG}5E>H{?|
z?eBPhTkDWvm3s2aKkB~6ZJJ4`l@#^M)+J)SVO^mTB~*z!uRW0mP&MIX_)R~7DSfCD
zCj6{{BjcxhB2zT@Evai|+n-i>^jtWe1!aPXo)2S#lc6oGuBhNux^^w6G9W-LRnlG%
zw13|?-u%Iv{CJuT;NhQrjqsVj9WmkWu4dhTfm;5gQ{>d_)1(zAt0-8@%H?>jS7Th6
z)LlDfHkZtHY9l#E^6`wJDj)P~&YmB(cUY%d(+C&guGyzHz&pSb)Jz{MIVhg!ns6&^
zEO^}&i#m=!kRGr49Ldm6FtbszPS<qm{b5XYl2zakT}Q={yD9>TrTY>g9-9>F9Hh{H
zd>czvtIj6lo@iUV+|R-iZ@yndeb?TwOXW!>OfJ=LpyUA=M?8)sj96gfK}i=o7&s-+
zFyw*UvzBvICTm1T0h}d*nBR9I87UHPPRX0?hTjN^&Gu!7CoB@W7W-jJ*MdP_KZ}8a
za^9JeFwS=yG}fEk|28U@JUU+#qiQyPhkcsYcni-2pZkOj_a(iv7{8Dxu|-boFjbeb
z&<(Lj-u|%|g6^Gd6Sr&q(6J=>I}>4Zn;sP&ONI|Vw^{J_Bi$#*#1{MPoUhx5W6Z-{
zPUWAZS9hI0pz0i&{2k`O>aM!^kYFhOT<v}ac^)sD?*;ng-L|zwU=TpOWOU$v2!3*6
z$imjmkaKr@q`&J0Q`eCS%{VX1c)y)hmP5BPn!sczkR=u^eCu=Q;ZV9ncn%864xRIi
z%DNbrMb-{_KO@tnj|XFy1&~1SkRT?b*HTg|nr>8D%A#<#|NU(2TxPv*V%_eIoAO(J
zA{q?zu8*aF8tS&@+iZq{ZNU0}uM7I9nOjp+^9c0u48raOlONsu79m$4X{>A9LkgWj
z9#fqT0(#KDk3c=a)kY}`mDqtgQI`gM1awmsQ9@qzvndekr@Y^%hkcHRO585jO?xN*
z^xI=1+nkGGsu!rX1is_bi^0Z`y17Uc_yuE;W&zN1nio|A+OACQX<HqCW9nhwwBhe8
zQWbNq(YRVW=&Rt(ocBaRL|H}3v+Zml81CA$w|$hmaTl>qw@%_59%XSF-E;jK1#P{*
zulucnnWs5z!@Y1LN6I@N&ixeh0*6y_3BpSb$F&yih)XQzXhF7-QLxOsW+W49#c_4O
z7Dp>)PnWz!*B}(9NSTX&Mp<FG?fR~>iBHG&abZsDua6an1KR~kzJG0x3Ubmrr`&~n
z)e=4CDD;qum>DqW4M9yHU3&_<qsW{Qi50n6E^s%|+lMmHPJ6(X53Sd&Ev?~dGfaG5
zMEm@s;<8V<Fgu+v4D2(Qn=7CLGZk8HM73;m9!IG3q%Hp2@P}i6M85X<9aO>Ki+>IH
z)%(d(v2&S9wiv9)(;cPt6v}vs9N}QL#hOp>8rw@Z*IW}!Wm-YL5BHz|DP_o>J%mvp
z4@Jd+OL7Vv&U)XprYy#bpTo1aeIvlrthP^+T3B)O;L)y2=Sd1FkZS5n0Ai9O98uo=
zb1wS1r#=OKXxRjRkqQ+ZlB(z<Xbe`3$6!yK7(8&fKl0L#wH{!Dal_027CNru*Z}^U
zVpDkov?9mm-BKqt=j)ICa;R<AZ$8Bp`bC}iFzEVYh|UX<wr@r&+R-7)I)xJm&OvkS
zJZ8ts9JMjZWIwncJb@XyHiwgAeb@a(nFKwZQU@12w1vEXWl#u^60mU}*zs4q1aS}O
zKNWF_UMJ&xBO$@uiEj0mq|HCv9GMEOQhXeWh4h7(z)BLN9tVUpKm4$}ofw%G&=(R$
zWM+axC}0NbhfdBm$6{mWQAgEJ(S@r4v9G;Ec#sUFfLb4lY}GC5B);%n{eb*j_FG}L
z(+^bVWl83L(>&J+++j)Bt<BffZ{g+Q*vhTbsdp`pdTsGe7m0-xFSRiE?ogi-U?p>u
zG{<SYJ5}?J(nvngmN!294@dAq`gxYQO;?l_(#8@X5#8oPpwY3)oFv+s&O0{MYSfVO
zFt-Mw*1qQH&ez@v$bu35eLI!xLr0;KP-OMLpsf>sb$GVWOHIKv3l8^)KWLyl&VzMX
zM8tPaf%08yCpT1yvd|$_>lK`bRdm!wiQSijr&Hlo1FI)RRJD+=`d!CMik(E&!BYeS
zA$9cnI>QR6wsybMno^+*kerZD%s1Nq90tZf(p5B5sBLN8C6gtl5ixQU`+W^VnFFSU
zeSBemknoEHB2{4sgYoEn{h^Qv(`>uj&STL0`}Xcu|Aa(tL3>h|OUO(>OI=SQVc3#h
zYO~RjuB3xc);O}NlcIG84Lt&fu-&Q0q@MOyb$9kSqmi%a#SlO6<$8W7xOyI`z(=A}
z{&}s5s4Wn3K;i}vlT)a=v!##xj{v~{KL7E5qoufzYZV<>=c_9Uq)b3lwPUD79WevG
z1!3=lvk}t;ABUk{7d3@u;)!pJ%o*f&{U|297*!R(cBde_hEbQ1bO<IO4RC{8;y0v;
zKv=vKq*(NXXl=+7o8rjQmSrkK)QFa|g0uaMa+hbxl*e^v)z8?UG*XO(XQm4rbCL-~
z1~UQ@h={D1UW_j?!O$7yX>$_l1xA(Li<`$S8-)-+lcSl}&0*qlAp7pfzPkv4jDkg3
zTaAXOFqEBNjpZEAyIe{|;+KIP8WXo)M;${s0x~z3;e-(=f9+i9Pu$41|IT0GkMC&|
zT>B!8t^^?@kU&_%WEtf{0|t7g7pA+JB=@g>=bS3rWxL&>p+{F&dXsUvT(0FTzq3@i
z6OpluXC4Zh38^w}GAT?3|7*!Ip)(^yCS~TVgcq4NI+G^j0{iKVOXjwu&$#B$%fV=-
zG9}>dSf)5ke`-ahY{s2|XBg+T$l!m$U^&e*!I{rAtSlVdVOi<;mVtKVWoB@F2NbAG
zSa@$_CY<A$aTzSDGGjBzt;wtc%CyXoLaB6Sp<M}{StOCmEg(f24)1*?Ez_BUeir6&
zNHPq9C8Sp#ScJQQy8{-0IapFD80R_MVO|fcDDAg^e>A0FaSyC23+)>2K#w<>^1!?Y
z)-(viJGcYW9$C{M9L6&3U?69}1S|m+ky*z17H2@7HrPi&1t{l8C`y?#0~Bc-Wq~B1
zk29DXD9=!MFol3&oMSUk%z<Ed;-FBj@d*ULLD=vlK)W1!IM4uS5H5yfAt4}mY}0r-
zEMbIjfAq$qmQm*tMo4K<>(DN>#W(n5z1V^><BH*|hzCXUAp$^RSi%V91U&;<s2}1&
z3!GUYf)wZ%a0HjZv7k^u6X=3#7-2wOTMoKf2Fns$704SKK*kvoX)VZC;2Q^PAnDeF
zKmf(uaf#ePMF9B%D+mx2n$t#Y!3^A4kS0vye^5FoQs-d-3!^;;Wdf~U0FvN^N7dnQ
zPa1<U-je*G;$SL+IEVkXe*JYT|M26!$nulHZLyWVoK6<SWHHYq7~R=cepbw<53{Rc
z4i-oaN5%DUusi*p{m!7l0L38MuUjyD2ID|Z)a_sYqM7$$w(QqmvwR-}L~pPPn2S{q
zf2xU9px4H#L>6KzK(_3QI4heC3MSSO22_Vd^OM>1>bzKFzXL;i`&s^>_`b-#mh>#;
zl)DbA=6Sr<3?^(Z=c=Z@$yF*D)LuCs!6&vm)Nrg5WI~<1WLK&NAfuimCt=%i5_(f|
zs)_5BS&wjP2K%I%h%XgQLzHArv7Mu-f5wG4+GZ)UPwy}PD6RtW2V?XCtHlYOK!PuS
zB37x|VwK__fL{-w5}`c?Td6{Nkeg_vfy6|_O(-_S3IUIAM64aOO8pObuoSqdh*RMt
z5KZM&OFotAnrN*R^+c;ytRpA)k6QIp>Ffj~^|*+n^v~(EkR5eR=Gh)obs~#0e;cFu
zxy0#?K=sB13K^?jDbRc$5+$vdS<7;>SnmBBIYDV=PnAbUS^6JRR@31@rXCqS7^_UP
zO*lUz+PoO{JYJSMzaG7%%1U(D6SWhEm#VYjI}+!YI>j18Pt>{59XZL(>%1Yp9+~|-
z6HerW8%?uR22E(`O#6+%$#=wLe;*|#%hz%h7Pp$u(WmX`)5=W|y2-#`JHE!6P0_i^
z*T7~`YyuFDcj)&Uqkma2exjURC%1OuKS<~AhVyNnCFNIWW3c}JU;llQ0E)Gg+XCXZ
z-nOveFi;01ZA&QqwuAS|`)!|5iTo);t=7A319#4Zr3Q_sZg6otE|lO9fAN?Rd<53<
zVpJOPZum~(nGv)I@*~Ut(0j<yg${Kk!Wwuvc(%#>v?elzCr`0I-f{kN%+IcJ9-O1R
zu-ywZFCH%7&k9jJW8Ko(Egn}YdXM}eeQDqFPHSmXJySny*QSyS$qkQ1p-)LyNh7XN
z-c#PNj#t@k<Jbm&DDyafe>sOCAE4-hcc^gfK>e}g24E3OD-bB1QacGfLz%V%)!44}
zHgIMEC1vR?j-t{^18yXgTF~22CUd5x1&yW`XG(3W@@72;IzyR`GSY4PC8c9XTeswb
z2#q&Kt}d2<2)v9ziL5T2qct8;#{O~cYK_FvI+7(GIKx8}I|(*6e{mgdtDYDQ@E|E2
zrwFt5m0VHRD23*7D{j6s3NYp4UExu+G<CnSw?KO04NX5)N9Yqk!Woy_zP^!u;|+xd
zG)SBloR`FPjfHNZv<})bX$bT1Of`illt#kTp)|tdVjR)0$sNYJ7l1H{ZZD1s?Qv{q
zcLJR#wQy!7LXtYAe*;1%Es-Zpo>l{w&9`217#r%bKav6rIUoZaC$bXi+UceB$CMx}
zR}EWt@S~=VxBfTj+RZnOq0KhSP*$N_F-U?gRW-&yJc}j3M+X$)jj|_<Yq1P{qBuws
zu}qi<YkO`%?N+0vjbLa}4&h_CA(!Duiw*&49qJP12XHwue?D6~D=t+~JHZ{`0A7cA
z=!7kRU&vK0!$AlnAdrXB2ZnWA3wcC0=!6Ve8KA`CH!*a;FoYaEyjKh-Rnnp(@JNv|
zC@87XrUK}r<Ag=&Yg_ECVQZuDlcGg=U>o#QjLZXxHUhH<qY1rW%Q&;4w<JH|(9Lr+
z5rMt6LZfGx5j8Z6;5?!Wd1Uae17SVZRx_8>haEP5_2ZJHsiHlXq0K85B?3~Vx$rH9
zf7D}I)=~*!SPOJDN+3Lxq!|HuC~4HQa!9GN(PLyXxO?!-p&uk=9#QEa1%Uhy=Wxis
zX2`Pwe1cv<f6#6mj3Hl*=^f0&F(E-7>5Z|7ed85{ccKF`Mtn=o62l;VgeJR^Xfp;R
z3Ni+Ni8%ZUO8`bs=*i0=m_2AEj75oqBne4NEj-?WCk$okN}f}ubN2aI!%6d?V<-*k
z*m1NWFBOGrw00{9FD|r`6*5EV2pG2Z6p|!bKMP%$Lib|GkO7XkAd1Erf)nHDSgayg
z6HG0N9^#Cm`BApUVkdkJS~M72(y}Qz6($LPI$#*wG)Ek1Pk2z6VOX?RjidyfE2d(4
z2O49R!XKvq%#jhxLXff;#?b+Jf<0w7WN;MTgDH0FF=Gq)B9&=25)4|hAjYoQ<NU$B
z(Ni)zW$;V{w~rpqCLW^pF=3gymi9{RvGykMT|(Yd{Z#6$(iA$12SpbmUSP{evL^C>
z8X6h1WW(4m<W@T26MQnnsY(T*jmE~=L-I`mkMmc_wWyZYEM}rLqwQKE0-7vpBdI6a
zqt=E_MLma0cMetW&_hzdl!UCPkmPMdQrwf;pM(ONyjfk6|E-|*=wXkTl#1o8)t0&$
z9NXQ}eZtbN_+yOFvuAh0GS;xW&?B*bqtsRdfM_*51OT6;4FwlUZ74B_y5Q&p(Vf6L
z6oxdFCd>oD0i8Y`B{Pi3j-r_Lh(_bd{0X&AVdh~1NEEs}%rYFEQs9q&!csE&C`n%T
zXIryWWHBe|l=Y^$w*97+hw-%mM8Bb<2f=I|e<LL%K~+&gigj%;M|@5h=S)g}SP7#v
z9#}l$k`$|AiS&{=Y07&-8Co3Wg_v7nOhcdH9H;ClcneACD#iql`AP63G}1H50Yi4n
z=!RV!xh_ntc&s~vX=1q_haTk0!or*fkOnHKwB*PG<tC}Oxzf7a<cHmFtP6~2lClaM
z5n4+*5XH<?EuM}*;>fCZ=%Fls@~$+Mw7W~+23duAD0Mt3EsuH1V=_GY=^m45qq1=1
zM8`6ZI40wc)yHJa{q&d&S9t0%S$WiibCeg-=|w!|GPZH_RUUVh$7FZ}RULoDIvrP)
z$71maEj?a?e7UM)dnps**ZAs~EIl48@9FVbMSHB<C484OrZP#7)9AQA4CRe=xNG`~
zUuvT4(Qx6UA{LMS@B#Twbyyw2>BQ$_NAFm5R%%5!90~W7t%|*5J&Ql=w5C#b*%lxY
z+puq!!;2jufA^gxewG;$+?N;<eTys4I3IQ1i~xxefN0nZ!mh;zy|nDv2hOywE-y@1
zC95o<LVXmgji^OXglxpoyUIeWtNTTfrSpWuAel5ZCGB|g#95nQO7WhyS7Tz`#d`U^
z8WYi5zH6Hj<Z_u1jP6>nmMQ6WHq<dCdPx6NJ=?Yve=3hRCXpDEvSbjY#;dl{m@I&@
z{)UHN+xCk2)oggbn9jo2x8uPW-khFX?40iY^73f+6KC*ZG`O8-D)iW;k9jYEuNR0U
z&e0MyOq}n`uke!|Y^b*KmxKE^#qjoSfg9?r9GAg&FF3wDSPVwPtDVX1sK{XY^F=ZK
z4JLv&f1lzY5IM|oH<;n~z5mG%@=y7fe3pODf8_tp|6R<c{~L(eAHozo{Fb5lfwG^u
z^V5^Piz8xAFXv6M$eVKXQ|#ySCS@;gKbowFby#`*ZfEDyM`FeHs7Xgw@JIDl^XIc7
zY%eRQBvMh6CAAJq9}hns9KHsYUiCsxDB#H(f7Fi#>{(GWrC7=nOE@x&^>Cz8h9plJ
zx||KJicxX1xE$VwhZ%Omr@O)8+wi)`@9{f!*`<D1)b+)Es8>Uir$zDr{pJtj@nA8W
zPOgXZ`_Z5U4YQ6M&n^y6U%w?e_CN!m24ch?*&6|e3wZPap4TQOto-_7bgWm5QC3ZE
z+&bKRIDdTq_V7J$b9B&$3MnGenQ}jINOBd@U!WffELqXZ(#w1=Kg{3e@A9MkLp~VY
z-wpDsDQGV*?&rhNbi(suG01Q7+W|J-&WgdJnB{-ue-7^N2l;3)zPyH8F`wt-hkTL)
zMgUvW>-?{Lp3jH3FlLc2?r@I9Y&by52qX;|ihs5I-qHKxw>vu|!{vTY+9*M({Dvgx
z`$6%p$;w)OZ|C*Sm&1?5N*}n(=w$hZSaH0UmA(8`1Vezp5cE7kr3RIf8k^j%<r0rR
zzrA?<hPZVrJR-5n*`~1Xdf7!#1^_piTFXDZ*?+tH5~J{0pFvkqn6YQUvGP#H_k%Im
z<A3{caXZKd<m&${7U;dpI$2+SHJn{NjBiH8cThE0d0zbWFc{@G!@rCC7W6*K$DsFV
zKAj|yA@T9~{pev1iOjqh55okDyIE1>i*HkioSno*tmPyxzMQ_J#OACgui>RcCHq;<
z#vUb6*Dq0Iw84v9%UOLsIY0eyj=Y#2$bXl-a4$E44kvqgL2XugYBIH!+xqfl|I=rT
z$!9%YOOzfM{oy8<vi%AQ$R(g6CTkpaP$<{(R=c0}j^3OBljom%6F8r}p!^MU8f$gU
zCh+_;ccs`$@BA)5&d(t~{R|oEU^bh6BfF$pu*<<5!t-$cC!}9@`Q>0%r+!J^mw!XU
z7)Z&7lbhjWxcET<JP2|;xSubk`FJ=1VNM2@qd`6aXM)Mxd^DKf#bAJ(6tm(QV*GeG
z8qAu4Udw|XpPU>Xo|2%K+q-}uh#PLMFeCDO17g;uSIoQl{cJdHCV6r#|FyUK+nd)P
ziJ`uz4!ar3ScS%K-(nNM@hdpmBY(h@S`WJIa;etxY^Qr4KE8bY%hCDnQSaUw*u<X^
z3aKRit}jW<P0J1Vn$6z*O==oOrW~F%`B=-DogN&#JbnMm`Tp4v9$EA#40fH6nbTkO
zHJO*TW~FbZ*2ji!SU5}BAUnREew&2a!+n#jwcOY7#o^wky<d)p<I9Kn(SLOE?#1qO
zblsDoNO0G9l7)YP+@^bL(NL4tB1NMxeS2%x$T59$!|{r;%4u{YO%~U39A7SWFHZJT
z7SD>?hZUPdi9eTN`0K)yMcri4rYz!~dN~+V4p9&_hllxqGN^|pYiqfXx39h&pT3E#
zolVEbQ~&M-{sg^;JL3SJDt~^Ggpia5+@?eEX+|WN?^9vZ+8X|32RzD43P=0->-<fA
zQ0*X(^Y{5lewv>_kou5+#0)gQiW%m0UgR_#Y$k7pA=SK@KFlEByvy%?+}{-w5Zj+I
z^_&bLpvU~P+<#zv55DWKhhmNq2`bY?jA=96a29{VEc4-a+H{tQz<)#jcm6FL=(JM;
zy_U;+^X1_5;G(6K6_`V<2v=%Fs6;EcElISZd8!rOXIM@rO3~~oO1a2?1I5G)qDmJ)
zPnQp)QL(7h6?TBv`SlRO{CqgCv{#I;F%2jtbW3zOD%E*g%*KPs_2p=uYW9yF-7ei-
z6f+b>s`ym<E&anvVSgAs(Oex|l-raVAt9UT!O0S&tQA!DPWO(_50)SW)!K&^k6sw}
ziM)*#(>TYX$0R9`?W902-TbOY8}UiRN<a#rWDJU!gGv)~F7fdgqh3?>bbM55xM|kw
zDs4%~d^Ne6UJoa?Aj09z4S;)sf6eE2LG+I@X)E;AY$tG9N`G`^MugNLn#g_RClHPV
zhFus?9(W*Wk=%$s;xtH<WL2Uz;x{w05)F8br&v9Nhs02VhekEtkj_$NL5K!hLj^2S
zF=H%i8D8`q(I*5Cf&X4k<vs9UGN~;%6GA3l;_IUn<W-p!)p^p_t~~<uA1)X4h>d{x
z-NC$|zmV1W9Dmd>U;6iE=&yS14`=fQT1uwCe%=k@1MFqD@{hyo#oe6#-5ToqVR|tE
z)m$UkdXkq_6t%wLv<Z@Jg1F!fig-nlmtBJF`UJJU#Iy<0ZGv2%guL(4<JPCA{TZcA
zkZlvh5_-I=$uV7W{QBgyzDBgE@oj3n>mf&U$T4hva(~*N6I%3$4xo7hKA2}zfaYC_
z`1%yJ4(HnhbpUNmLXfTL@m>2eVtsO22kdQfI&kI<IHOD}`7F_xe1d$MXW(suI&kJT
zAt*6-*EKJr)+eWRdfg_cgJ;>W_EbEoOOIZkp61bSiyYbY&nZ3lP@-oEHBa!)&BNO^
zH61MD5r0fYjZSo>yVN|vH8)RP+tlclvN1eSRo4}oCwS)e`DmM<m7bY?BvBExL_SaO
z&8>69Ha&i2a5stTaUF8(6MS>)6tG2(;wuAJYFVsH&J!GS`?#)6PY2CTBm}uM_336B
z_6e@J_kgNRSqDhUG?0`STI^b8keTa3-BwQf{{;ZVw>&Sm;glV;4+S$aI5aYsnxq{b
z1v4@@G%}aHq#f>mOOJvu5XbNR6mvHQrnEp4<H1)XK6deuP2B7Og&@gF4d7wF{T6D}
zi0fld)BiC2^J_aT=U@T@bPzaLj00uBQ2=9L3mcGu%a!6FI0aRNCR1701VqbH1C>F6
zq}3wE;%6U$F_39NjQL?*RI*GWF*-pWm0!SEEj!EiL1UVK?AC{TBGYkm7R3ir&Fe+^
zCQJ5(%+!yC0-LFPlI2Br8)V=xVvK{MKy--lk%YxI2iMi5{d6P4B#ptfbz&ixX`*tG
zS|6)Y?e+_(qyB2@509ubxtmHcZcd^wsnc@2rlps#&KVX>@dT~?I_+@f`QGx1_V(U#
z>mDto)&=c<Gkfd2PTTW))46wp)+>VV8^rwX{!)4(7W@fd_jO?3^Sb_9V?LNo&@{~(
z4O>GLmklfU84<4#g;_@wb;S1|j&fpj<v{`2MCmM2qw+RZ(?>~|!zV3;GN=Nr77v!C
z!%<b<U;i;@+MQ{4t`*nHMjhBAcm9{VU7ZW)y9&A`tw5K6Ydx|pUj*%g+CZO3`D)c?
zb_3OCGA7#M1fptw0jwHiUzdR$8WWdbryX8VG&D6dMlv}#GD0^uH8w>!H8Da%Fhey$
zG%z+qHa0XiJ|H|aG&M9vGC4RhLN_-xHbpo!F+xKyLp4G)Fg8RsHZ(RqT?#KuWo~D5
zXdpE+IhPT39Vve$(#uO!VI0Qs_nF4a(HuKntj<NdUu;IL?537!GnzSNHK{X}HjNMw
zBvy-{qKkBaT9lhWWuUTzL>k&i1a4XcL5rwXEh1b7{Q<4)_xxt_;rqNZ?{MIZF~%$e
zj41;#Fs9r{<v4Qzo}DBE<SaQ&G9*QM;opC+Hws`{a;$%{MI-=Yf6{r5oFZwGB;CBB
z-zy+NnCfx;@)?pKJwUWs&x%MP5RT}yNHMI|S2{~c39Qtt&N8wX*3hUvRge_&tek{k
zjr1B=ue%LOo3eVtCQD#F&FJjo4XtHlDXht_2BG;eJzGJRBQ*C+=Sot6(Dz}TgJd;X
z1^Z!8&-#D)dquEEV>(^pz<%^mr%ycCPu}UQB9(9g4LaA7HE`?+o$JVPvYu=p8_5Z>
ziEJiY$W~HKwvjN2kQ!1;>d1DogX|>rWEa^@_K>}Bs{ZK{?1R(xMXLc$=AKp~oV!z6
zO>kZgXzhnH*Q9j-&fla~GhC}q>mc07b*&b-9lw9HTH)RbYaN0+c3CS5ceX|=2KPt)
z^|ZlVsL*PM=jNZ_FucYWS{?AB`L}umUguk_PI!YUtvI}!A+4kEKIA_=2ESxTs|$X7
z+Q9GnEhH+0p16>FEcAUA`o9V3pTgk0aQ=^Qu~N8HFAQH0u4aX6SA-ie;kF};%n0MP
z!t_T2A=@ZCZx<#L!c<!Sr^UpA|K^UI%}kj8vP5f73T19&b98cLVQmU!Ze(v_Y6^37
ZVRCeMa%E-;F*!6eH#0E`B_%~qMhZq9E|vfQ
delta 116825
zcmV(=K-s^y-**1gc90_jGch)ov6&kvf3+Ifj^wuS-Cr?z2%<rz6^WuQ8$qz%tdli%
zk-$5F9mD~mmN?VIsHL_J?~cDdbx|#Kb8I6(o>VMSWU*K``Q`m*fA}g*Z(<!LI#W0I
z4>u}K!z|BlvLXrNIJ>!TZvGX#dHwozKi=+C8U+u-ZCnKXkz3QjFz{tt)lUNUf1f_T
z{rCOXH#-$(NqMt_DN7ymRA1eyG^p%!c$ixr{^rxEt-5NmeK)#Ed5!t;!1JVOSj1W;
z5|TOtgQDQ^dG{KAoQBGFrs4kq@qo+p!?9`|YVIa8JXG~GSM7+?^3LP1OjFu<QG~k4
zf#@<+x#Cq`b$$1I>}MftocBjNe~v)uGC>)Znac40T^WUCnGv<663)HaBWd*{<BOXD
z#%J1%&td6ZT!e91@OVcjZjEV$l~o*=+c*uji~?*=+k5~{P_ydr8wIwTW;<Ez`w{;$
zuDfA?xb=9L9{S-HhVO{U5fFrXVLCn%0R#S!N78!6MjT8%e;1N$$D{Tre*(LG|B3sL
zePh~@yLjn*OEKcx$9@1|INEwmM8ZT2rtarqMDnwQ*Uc2{uHmiO29Gl2-uv4k3S>mp
zG=mvO`HdtLu$>|K02J<<03d!I?Y@-}ydxhw&Wyji-6drvL+v=(?vXLXqkeee)}b01
zI1Tm|cu@_ueP+yHT8oq*e++^|-*Fy>{_z$xV=^8_EUYYq{qSu~HFOLfdgRAP>8Je`
z$fwghF|MhmiXkhrG(h@b_jvsCWSWq%_XpQ+<eciAhn%%&(XNUDc2%`)UjzNP@giqB
zuxF~qCx+ng=}V+D<6t2EiHxkd^GedMPlL`cDKfW4CL2r&AbCL3f3)c&?lEYm(X<cJ
z-*3jHn?ne>4^4~lvEYZ<bQA17O9PMy9Sh$Pcr;*|V2XyD*P4siBWe=@qcKN@R>Da0
z^`2!9^RXi3kK#ZMj~gP~5r}!eplVP85j;dRW~|}N4YycM;D^0AR5;a-7L<+q@Sxn7
zYxb}8JSiM|RZ<3Qf3{HI$-XmxLhkz(71Z!A4`i9`D9wq*4h8}7Xj}<PMBreabUOIX
z0aoc))zcLH2KfwW7zV%^4t+D%GIN^o%+OQG$jY|;z|I4WCCN5=sdld1Pib&8laGJ_
z<yOU+n79}<zCZP}wV0OxT-T!%h{nkHmFXkI!%!U!E3&gBe>ejhE9e)wLcMX$=j=?j
zC-lbSZonxOFpT=icNXY7JyWS9RN$>;=xfhZS*UaE2CSU#^;|j*Rib}suH-xyONM8r
zBm-x8rmFW$6@&jxb2sGDkD}rwN$`D?MjXxn6mxhHCo#l2TN8JvP>o7(CIcR2nE-L5
zNQfHwyLKNif54$x@&Is}0evSh1Ho_v`x7$aev#GE>v$X;rW-eoJb?|F<H>^DVQhfS
zahe-z$TLCfqYP@yk&k8)1aAo;7tSF(T$p&}PNKHBuOxE@5&$+b2(bl5z$xdVh=?U!
zKtMQTTRS;hfatfKbXNn|{n@wiwi<4AgvM2+1#L>De|a#TU0qtZLrM!#5*V;}O8g1P
zV(4+GfJ5EHW2zbrqs(k&1~gjoA87ArcV5FstW4Q;#2RKgShGDHa3d42eAhqhfNwk4
zAP+r(FrsXCwG2b*`gxKGiN!FCSnM%##BY^v0|mf{`9-P1ZWje1-1IzP?C%&s<{bf#
zrNMy|f0&H2PT!(yzovxDVOEfpGjMul<aT4ek0Qlh9orIi6&r7n<%>>a#o~-<)H8qE
z6+4IO(RT_QIHVXj)O4l^oLSYd$p-e0+>22u_L#G=+gBqAu#=NkrYb~W0bi2|x{8Wj
z97f@I38x9}!jeZJbxzOWCv>tK1>mV;X>CpOe`xM_p!hoIbzd<TIzR>v<HX~b1rW#&
z(rY`Pd7#-6Ah)r=4)#__Tsnb?)?S0!YlY>ZwI}7kk*mPPcMKHFpx8wiH_l{!A~8P^
zCL5sH_QYaJp#)hUfYADYvXKFTW{p`30;3rgHNpu-p0MI3o}JfDfak$|DA_UsGjlD0
ze<%v_00=hgL&#k&2tjy{;=Wa0wHQey+5%IolmyWy(~S&r)Z`<|bTjFN7OIp^7D~(@
zJsdDV&j}Mmi~L;rAS!%fRf5LU_hizv<Ukr}pWw(mr_yjrGTZSWx%4oPlWYOZVH!NU
zG~T7~bBBswu^WT7Gbo-EIm^SMh$%%bf8vr?jBj^w1n2Y?!Ys<#U8v<8^eRKn;eX9}
zeC&AT&3L24m+?l(D~_aWV;E5ho;3~TncF)?fjT9Urt#h)O*|K-QM6DvlX8_Xr|~N7
z#Rjq-zZ731!0`Y=n@kuHE1QQ+=aka;9JQ~h5Kcmj0&8dp$q%MIF;uxcjPr@;e=7ZD
z<P!OE!!U3(lq?{KAur7bQRmLL_!5dHkUF}=JLKkDh$BqZaD$ULRs_l6l*^$M?_9wC
z(;Y)-VZ4i8p1an{moMqH65^P-Yp{ZVAa52%fJr`5MuAw`F6t4*Ki;(;F@VAo)Tu)6
z3^rp=`@yz&DFi8v9vMa%z%Q*2e|{b{_+t3pzuGp%QYFQqWQ`z(KvkAl2i)*`3%cFH
z?9Ft?W2JCY1mDafxXhNZkf7xl!-m{^dnv<b_q`@$Jedkj3r(#!uu{pAv@+ILZBpb>
zajdv7g$E0&OJgTSkQ-MR3O`Ana=HFTFt9s|GL&h|qDNT(o4ueV;M8C|f7_N7n<WiU
zKoJ=C&$;oRzmbs!s78m3mK%c1!H!S9rCl8W0`3~ZC{Ey^89R4Aa;Nx!GIs8@ywp1G
zb$w(JXjUS^d;VJsELm}y1m8*sn*^4UQr>(lW*WT)m{U$osgz|01|hbe4y_E5<1#=a
zGH2w8_Rl_37e~MS)9<K*fBn@C%zKhaQT1N-Tgcw>V0u0|Jt6e*)S4r3jtf{h9XNKZ
z!#~7VjGiivz0AaI2zZ%2JGc6ZM|FQR>@{?`@)}BqNtC*@?wLoZ1Tx1mPYTW<Rg~Q%
zVHT%6==VG*iB{8~%gS$dt5FK$oH+QW$5W{&jDUz}6@9B*LsWc3e+|i#OL@4ax`2%p
zOXmB@OORT2C8T^Oo6~DjVxV}*kpC!4WvdayL*+(%BXeXc2IXN~rfUZM*|TSW(qzLR
zS1@X%njb}BqV)FA_U=IPHAP8yE)e+EiWQvcLUJ{&YLOI3<}QK-;N!v82L{<N0Kls3
zTNtD{TFM0PJ-jHxe=JL3521Ef3ExLbc^D4^)BI*N?sLx}=3yBxC5P{BcUhJM`EStG
z=U&0mAXrNo#&(is0=59^be@G#lDqht)sCtMi+#dOpkicWT`iq6Tgp5pRHDy)oY9Je
z;1@WP{Tj~bFpV~xL1*UCfg~*QOBL8m1l!dM_gp>Zio;qjf8w8cCW$U}V@(2DxAOu0
zyz$63#$$7A8FXvb4H+e69Eo1f3F9s<A;`_jt(I)-Pzr>D%StiT$vcK5KWMl`t`xpU
zkdmYQiL-$?=Zaz;v^`Zl!MHxmT|GfU#g)of$|4>g1fp@O*gd$a4gw-vd^wVx6Z8zl
ziLVf{Om1IRf4{9wy9{1r@6zc!6SW@O7KOS<Rss-tBzGf4zaY2iuaw*L|01_9ll_Hf
z6&-3-{4Z|pf|Nxv(@T0bB8>Db>QCjOH27e-90j1i3Xpn2!DL-HB9YG%(9vQmY+R2Z
z16Jg&6z6VR5Vfug6;hD7<1tcww+(k)lQwck)xI2Se^t*&YNA}^?Q;Ai#yf+S`|!d_
zKn!1i0<2TgAPQ^}3UMKcaJL7D`3Nbzx%$aN2N!r!7aNg+m6?=@-30~pBXN&|F+6}&
z4e;Xl5cCA(IMWHz;ZOFsx-{HmR|mO-I9IDiM$iR_yPZfkM*&S|0RTh=dEY}A+*Wir
zfckK#e}-dqcfR!StW4D~x{XdF$Wv7G1%vPVwyTCSn$wqUFq&pE$=i-EX-5cZmtAKI
z|LG!^P7<N@@zEu*-2PVl8O#qe2qZ?XcZ{j847VpDy|{)~zvdcV(V7bHzXwUD=$uZ2
zmHaBk>gp^vONz5f{J#0*6=$j|aCs&!%iz^8e_^c}H(Xn!3z3lgVL21BxSEC)baLOl
zdP~NDhJaV-80yhm9FSUUo%+*^#dc~`CqD^84{0iSjtI2IKHbKAmlq|$ySMM(Fdmou
z2Es)<^iqNW`|^0W<6cf;0r?k9`UXlS0S#Gf#zayz$1hJG{`%_gSoqSk#yd#A_fphl
zf2@?!(%>EG5MW$#e8)|vN}B4Z8tnlbeBUn5BcMZ!0Zv0zc7N5tmEKjkD!nUU-IwGk
z%OY6S6g$R<n9`}N#1wvF6H&;eEdUjN#+gKl4_U%Htkgo*;V-1fHKlvofJtEoyWpsS
zQb(6(R^29PbfPSOud2f?yL4H=xFn%Ie+n*dk+odlYK~a6a4TfgJ(Ht3(L~JN)YVM4
zrZ{wNCwroz9dyZUz%%!8g|qfXIKG-;!4+l|fppU`MDzfmVHD|J<gP0K7%NKgqH-~V
zQL!?hcid;jlfTYWyQm!^qK_Qe0o33V2fi7}oG%3EV1#Av4+zg-c`!@%r%gCEfAZbQ
zqIjMVt1;$;_vLs?HxDGpNeZy5{XWr!SCTEN03gSpFl?3OBc97!CsSJ#K%xY&`{$RP
z4A^sdV*(JDwm`{jJNDf#So?fy>?OR^7o{cWA!#}r?sYV9$G^zGkema*-E<^pSU2hg
z!~9_6`rg%N|2LA9MZ&PmJ$nN{f5e6Q>}cG$_Sn-!4!~5fs#b3y0QZVNa+*@@0mMUJ
z&!ab#pxDFR69NDx1nYHt&jD?IZ^R=8^Uyo|-Wd05&tG6pH`#}0H=Ld-BuREw9<N}Z
z`B1pG4n=zLh@r@q%JJx;J>ox(B-YVu9O6uoZQP<qqUE*@?J-G_=6D4@f5&I|$#p&O
z7pxTQY+xZI6V3)d<#<bHK@CZW>3}l6=mm_o2J6}TMMbuh`$7cH6<KJkoemmtQ0Gxh
zJv-zN^<nW6&fvG=KR`ob{@HgmK*pz%i*$$<a+>cS<1}|tHvVv`z}app$x=7YgLigj
z3{Y5#w*~Z4S6;5^y+_YWe~My%-H!*$P1-^l$reb-dG?=E@&t`8ml@p+3mB(6?%>PA
zGH8Yhh4CY|=->wap@Ng>Bo-I*ot$koj##^}mo-}+4MiZa)kLun!W<*1DBn8jNP>3!
z@@J8@`N#-%yOB3p34PMNJclaI2?DFd^^*$HpK+5vU1SaNoDK2Lf4&3il9t^Iw@wm7
zk%sD&QXX)!9Cl#v#14zXD0bT}c>$F3k^ml9(^7ul$GQvC?&+0-`^iHAdJd5ws^=rH
zHI0Bkf3Oq@&rgM!XawBz<mJj><WhvD62mUeL!F5ney>z|yGx_!d<t2)ep`}9i{~gW
zQVWUMa>;M7-Fc!lf53YmlEHuA;0Jl`S^!O#l_Dy8LIUc|dO7I?L$ciU!^ng+8vFPi
zLlU1Kp%Iopu14>QMV_PJYeJt4>d0AhgdSJ-Owu{<<y*v(h+j^C5E^GYaRx(pVu07r
zZT5!8sn7Og=eVWAmk&~x`SRe8nxwLby?BC$sEzm1E}^bae|@Oi9_i8*!Re**TEDN*
zU`pSdpPOvnrDNag<$aqpIR%Yfp6L(m<>bZWuimZ^PA;odaTU8`@8)J25zbJm{32q1
zA?Uy}*c?deIgI$F(?}?GUftNEt(K_n>R}s<T|aD#TqcD9OYI!P@~BO&7e9M*|Ji>5
z9r7Q(3T19&mqnu;5EdXXAa7!73OqatFHB`_XLM*FH90sim$8`}D1Xg7S##tzdhh-f
z%3G_V<`MwUapgfCkG-~MvQsPVk*qgWY?4yb!zSB2R@VIbd>42L?eS5S%8MHO0D=I%
z`+&Z<|LhN6g~?UwrIC)V?jNqA$V)?Y73;+FO?-7<T>WD=9JAqiuak85=Jo5>ZIA6_
z_i(NJT~-aphq11oxqrXBR^jfabDdQU4?49)S##f^JZ9J0-~Fp<$Buimrr^G8*kAwi
z{@W}6YOj1)Kq?>Rc|v~&Q;vOgsOoBXW=PxA(t0|UGAcWtyY}2w*-$dzdq4E=%6zC=
z84&?QdGxKl_R=JQ4+frAveWzP1ZKY8o6ziNl_m^!ADZ@Bhkv`Lx-1^c8}1yARnIN{
z4FFOy;CtWKT|e-{zwkiHW*EDYhv9~n69l`Wf<Cp&`2_Ze3bQ!|UVf179@>r@k7ZMK
z*-$l)(y>4WL-KZPhO*<1d>FI3_qgZ&IQBQ(d@A`jYlbSXoH#i1VZtmz2GE+w0NQ!h
zaHB6f-Yq|7(tn@v&mrr}x@x48_o3A8VH{xZbUOIhm3^6Y`O#agB;l3(4Ut2LH9$;t
zRU%fVPZA1fffjezc*5<s-3I}m%ckeCOp1)1vFy2@mjw;Gj<U1Efl^(E)4|bJh#wkh
zR1C*Zejv0M743$$tpz@`3_6qppoL*Y)gubWsz1u0Zh!0tS@Uqb4tLM4uRe6`iOG3v
zIyi{zPz${iyY+N>z}MlGQt(!&k+I{%OH2e*uRWaxJdwVJd3}HPMV5bf?5<S;2ge)T
zNQJ$9TNzJJ_&8ShU1PL^*cd*^gC@g6>KTe$`T7lVqP&eHgbkq4mRkp!*f5<Vs+_b5
zqzMAV1b><_hmEW|R700_&-jRl#1%XOK6_#!9KwghZ4waY7V*b`a*%myk7v-nR=Cs~
z*8+Hdj77xYlHl;^m=X3h5K48*x(Y=W!XMgv>}l=rK<k3vyv|^iC_eIxSstMv(GJed
z*X0R0+5i*0g;f*MjK4rypgbQ$iO<9Ti$~~SF@FNc16wlUKd$!y5GWprFN5R7VC{Qx
zI8FDtW4gc#bYch4p~;}=3baaW5*!4C6h>FZi&V%1zjM$V?h|D4ele^ln-(Pz?w4gP
zGyre=V>|52CI_Sz0vw0)IEXZjy(rj-)1p0PEQXWa6NpU5Uc<nPO<=d;T%q>jX#ioa
zB!6tcocE-Uuytx#4<v2|0XWNX5&co@1u*Ga_O~Qh0+fwJJwTgG$O_0hD=JVf{8W^E
zE)SZe4FKEoP@Srer3|SaP-Z@q-82x$&4?_7KOi@I7p=7qEMMs0vGLH>HETsyK>%`X
z5#cOJMQI;?Ky5q=?pJ)YLAp3v__j7=g@4eu|6A~rE{G`hd|v^*DP?$&?;yv}3sI9U
zMgfKE7|f0imK^zjA0Pupt-qLDN1lpPka?aT#_KgC@Xa*A>1cW%B&?UFn>8#Ks=X);
z;VkD-+Rha&=0Zk{)pD}8JW4I85~fw@6^LMEM1X(qE8jt!p_eFiwFeVU*8g0$eShyb
zBH-yDnRpsG-@y#VOJ-7k=V0m7iz3twUXV`f(%LB?LRisi+^4_z;6yE)>Wfvcs0inn
z#{6a)0xn1VEk>bOTZp#hvSSbl=|(`V*0d6k&oYBqtSR!eA1&AP?~XM^AQXcaYXZ{N
zdK$l6mWqIo)qGq!yht)uo}t`wsDBHb=+_X-??7GfDb(rTi#qiZbu+^NV=0#%<a-e;
z0c>V32=u&(X<F!lQ%bd3o34K>$-+(aMNSMrQY?{V4`!v&R|f-<Te-I##fy1d+W-nm
zCbiFswddurfKRN!^2`skW@<f*4S7>HE;=)*7x?DaBmXz;G5H01oW51b@_#eI5#TSH
zaAd7&*&2*Ale44K-6=cw3=_K>V0=vGg3^Nl{lIP$4C&8VUdn5&oxF6~^<~D6qDr-W
z4V&a9W8(r}CUC6L|Ix9AUYsP$W5soqU`3$;O^cP~gI!R|`oZhUYV_Ojp`~%gj?2N2
zC@0@H@l5LUeZy32kW^-?*nekBC#MgbSmH5guxU-5E&o=hbK6UA4t@xD2vsvoGj#3v
zcywVz+=sCS+x;UM<!Ix9$MojM(c0+ngu*vX(Q8Ds_Y7ns)nB^4_(U;^3=JY0)Ve-v
z4ZA(QN>;6MTaW;^<`uRv`q55r9MD_@oj`swd=cHwz~4C?nRgX<rGGUnzZh=74G1JQ
z69n<@&m-8jb<I$uF72zBH4b*8C5=67$<_=6ckCq7^!y9Q-L+0XrWkYf;v!yg5bkd4
z;Y26@%x+mEZdqvEvPc%94aak+ZwlVvcyT<dyc%kmHOuJ?c>p?$u80#cb_$;K2sFm;
zKQn~rFI6vm+8c)kQ-4NzI*2~4!_dKrF6*$!Gk(BsP-IQmm%I(uBhc|i?rsbZur%#}
zj}8+q^!!w>Ot{)9M0^kii$RZ$sqkCtsopx|2)cpca;jMk!C7aGL%#$e2<=I|ymiX_
zc;2SS6m3FJC0iG;I+YDZ1gw*ibzNoyiw~dhXNbo+Lgfh7-G39?gUi2Qd!;6}zv0fO
zV?|7hO&R}GA(P?Q9>f(#3N*-+2aXnrPY1u5*bk9jS|-b-OaU{L$2rm~G5^a0GhlTj
zleK_BK@*5XAygQW(*zWGj&0}37>s0Iqnnx&$02~Cw<fUaGC&O{Xpt|knUABSer%PW
z@gtT)2qU>)^?xj0qg^%h496f^*5vXg9TIel^o>E^RF4(m8J`ww_FVk`O~Xvi;aP0u
zfJT?K17Zi9oc}k*P8_S1v6IeZ#QPDsRso+4BGc5;=7Fco#{SmQB~YG;wzBL+xW-Rs
z?_m+KL$Aam?{KA7`*YhA6dUseWnxqzV6uY{YDod$fPeC`e()In4rBys2<|G%PNE9%
zCtF4sxZj55E=D&DXss8D5}?UOHR3oX$)##EFH|FT9~VJp9y2Lu$0{0%XFtKfD0GEo
z2u@3lf~EZdP<6w<CEFE?ad4(5h7@IAJvO9i;fc;1ezU@@lQCe*shR#<<{XrCNSzR@
z>?oZ*iGLo*tHb?fL*}59AON3i;+Z+E<j7TU2ij#ahD<s5J(Z=z_LL9mBidN;BywQm
zZHFs?rr|dMgoj^t#1Q~prYoTH&<;m#-G0LZHnDx|;(S$vf~{Z@tK8?uJ1`1fk%(K)
zp=yATk%K+;Bgm%Vc&c&+6mZa}llG^~W)MK`K7aRv1bShy_io1l#NxooB*Hr;l^7L>
z21z^R^;ndWaylSUjq-}kLxK9`MqQmMUM=QSC>sU@1EQTXT;U@4w|C`fVC}D9U_xTN
z>rvPMr6GwX5hCFyod20eAk#S54hC-6F0pKv%xvd43Cn<_$|58tRPs~a(6Hh7mwp%<
z8h;Y+?#@~Bc$_{>ZxNcI@wOVd0n?__w5W1s`Eofs(eo}-{^mO%u&T~#+)Nto?v7b^
z%5J!ycLohIVJRB|L%_WC<)hfQPJ!BKzis<t<xEWgHUlFTGd6G!pL6qli=6gcWyQye
zA&$}7m(h3maU3oWlk5mRUNG^Q-uj_oseckF#=<!(D94pq0@0DbqYd5gDB(AM_cr_R
zA?vaqMuy{8Kj-Fqk~a9~sjTyl&mS(Hg-T@E(0~av?oOo}UjsW`qm4K`$~s2S$N?W2
zV*@9`Eqw=m;;C|aSa6C2R^UO$piME9#wuO2X@J3{#Ay7nU2Xk=^w#;|hLb@|D1RUp
z{(7y@sH}m-xyK^H<Nhx<k$y}sxg1Cs3XYqoY6Yfyu*pB5451W6;G?!6#X!OdpFz8z
zR~`5uJ<76>QERn_kVA<X5p1#ZG@$W4vk;18?Ci(O%p}thp70EeNMmf2v*xcy$q9<}
ziND;_Rqfa_0Jy)P_yIm3+W0g}VSh-2$CuN|gXEW3*NfC6moBK?AHl>YZRJ~@K`0Xf
zgUf+oL{FP>FUi;Silummu)Xa-N)GIql=2I2ea@|~$Es+Dp_;26E_y6r9D@dskVe9a
z1g-A1)B#8(b~h%QI%`+64Tn4<0Q<h!;mKqi%$A{JDWI=#!?D_Zg=a$P_<z`@kuxx7
zvHk+Pr)|favPBplR;;K3gU}OeYh(a)Q$m7$&*v)7uz1CVV9G^nV34}Zik_Q;fWWfu
zT3i&9Wx>#{T?#snfp<U$b%th#$T)ZbmV#9e*LyJ1p)oHCJ+-XAyc{%qnVIlN^R{WM
zVd19}4i<6=!Ig{_On^lP+<#YE7@iIS<#cMNec`mv42VE9<Y5K)kji&$N6pcZV|A9X
zqy)2L_Cwa(@C%~teulq*gjrF}1gzr_tUhxU8LUteE_e-JDt{b-BIk29{v4*gc;H6-
zeP<|F!oHNm#|EAJMbbUBC>BE_G>?rbh>j0{t-S281_Ga`U;!J8QhyVVA-DNaYFkjy
zx~fiAlpF>T@ni@d2b{se4XR1@V?UAx{sQL5j>I>PX6#cxtr$R3SLF6k++;wpNeNg$
zn=<Ouk}SQ#supRn$cDz)w1>$8FE~zF0QWm(|FYQ0gcRn?$&$baB&#pKes6UjKx~~5
zwu_R40X=CsI3R#Yo_~c@I&!)s1Z?@4gabl<RZt{FPOEfy|7t3}S!J|Aw+2c=U0?E|
zDwehZystf{S8u#vc9(FY?<X#$*IvIw$Ra=xu0n*Extc-^p2L?*yeZ1yt|w8k2Lj5D
z`8zNZP|gpo=%@0GF`v$JF2b6wJf||AA1hb%a}|g|*7V|V<9`R9odFqNnO~E8{K;_&
zk>>|f#LTt(RRZoeZ(fqTi%~OXS7?ifuKab8j1EIvs}TzOVGBf1K~%@F%z$ENT_x_I
zWLiRrfPy_n8n>5#9c;ja4)%ye;87#%u0eu5lI=7wjEZ9%=M)D^a{4s~wb}pnpgjIA
zm*HSbV*T5gxPSJHiZ`fUh3Fa8FGKX}5QrCsX7znC65z83ITTDdk?8{`L#6`2Fsd4}
zPd6Fgc(8hH-3hNR!_my=HzJLs6RmGoDUH(F3*+>cZ8BVKaz*r5N~2fhSl|1`=W3dt
z5d>O!$y^cf^C%8~5sJf={95PcRR|(14c4;OCWV5~On=G^pMr(OL!A!gcEdeS^1@Wu
zi$FYa21^RhWfS~3VqxtA8XAG+k|5czVGz66FkllJhg&uzVAX6ULzYvRejA)Bo0t!1
zL|qC&YfO)%zf`L0--66-UVej2QQdl>x^*a3p$HtLzFVXDGF1zQlWh+}o&+*u7@>mF
zh0777o_`6pdH=P+7f!i<Vmb+-Bd?KRv+my_?xK(NCp1B@x{^}QkH-OI5}6WNS8xYE
zJ@W&KkN^vWTX`n?16%fYke-noaI8BXyRLn_#?Whdksxj9t2Ut&JI~n}P!w7J!Q&=}
zm8hHKxQ+Q>Ljv7%KrFCxg<DU}pi}?&SoVWE*?&ssXuu$E`T@MGr9o~p+gx90)Mfje
zvrIO}0gWktK(kQ(;3eH}*7A2}jxvnGMRF|gxne#bC#zw=v~+v`<j%2=u9vYHcY;Sm
z204Pt)1fR1vG2EbRtok^5d3VyMSK3bZVwXzHsgYvnr<|S*qSDh!dz-H0<eXtBAY7@
z41Y%C3kv$9*x`{SH^9`UV2Q_%lX!(T_{m+f8jgE*U+kcR&|ol62Z7|G*%RTk4)#jE
zgu6x#JRv+|h;<BB7XhGcZ=Jx9e$1Vf9hqz_;~$Bq;nfk5%eea6acH@;xZ`0GECRB=
zeIsHE!S5v)P*Dx?>Hvr>-Cm9GDoAlNt$#S+3ne<hyl`Fbr|e5v+;BI{Ty@->%RTUE
zE-b+(jBAvILz!+?@#i;x{pRkCD>WPv@z|Um1*L(ApR2MiIH$HypxCpmi;z6spN6B{
zjqtY#F$_D{!`aLC)<(SSP8&>SVi<)V>{d(gwg~WLdpdODfPJ9oF#x|VI}YMC0)OC`
z0yMtEktHS(T!@j<65$iATx_kC;266x)6Xqi>!mn=hM<gp8uBu$k4(Z84)IMZ89(|)
zD^s+!9FSvabJ{NeCmLhm_`3<`1B61@T0&yjkUjar`{;p_bkgV|H=L=U7yF?)am4q+
z$vDe;)qKz3e_Sq_4|d=rNnNg4>VE~Z$N0)26yuGR5~ltx{x?^;Joe0F)&Il)g=M-w
zbC2Pl4(7lq);amT2_%zk7A0l^ra(w=A;g3)aJjxuxG1N~0D9|hNjVzdq`~5hO@W13
zsy@Ti_&JXN+P`5a#U3c&VFP7mXY&>`)MhF)ej0p*U!AHDEd-ruyw*aJ=zlVs{6(ra
zOov0ja2-!2_pYwsN%x_p3}z}ts|bTbAK_RrD($j`B|bS_l-gQz)_UwWmmLZUOJA_;
zZrS`oOR?k-6+X_w7&AqKFgn|NJr7w&@em^FSjA%HVK^NP-pdX~UW0@rot>l{b0r0o
zBy1mI+MF#%fXtCAgwJLRzkf;j5FAFrC6Q_jY#iz=|KLvRwLig-IIyP~#Nk|{fpD8S
zjM<9BewIjaut7LN`lFSHI8Y8F=Ab)5rp&*qW)JHvHn|G$V_w`h;ek)`%T62)<m5L-
zP+t@d#&&x~+%9LF@Jz&*#15B$!0%g(J2n!;OeABH*aQhQ#s*U)9e-cuLqOR)R<OeE
zLKJwBf9Q^kO7ssJsi(w;0})oZ_6?|q9f2F`U09n7pC+4&yE5mQlNk*3#Fvt=MRcQt
zYp`KWN`5+34M>>YYA?lRKK`D&vZyvlLimW`@{9REyc*0I4sB_f3=V^FA2g7#gSKLW
z{+uQOQ>z?~ZPC{4<9}3_08m?y=Sx8zf~rl8L1LvamE;A!V$$?4l;i~wXuV0P-Eke>
zI5^aaR$jU)!^62JL!jqxEObBgD*ho1`*;-$NV(!fsz&KkKU=cdM|gL0eR|(!d3&g`
zX8#2j7-K>99hfIz><T4dZT7qD%w4y`05YKG2dX-_dw0u>i+|%#3t<nT9vfB*xr!ph
zLW(2nP(v&4+P>doKCwNMDZ^m7b`aZ&j2QV;0%x|QoO%cOtvLf2aY~awqnmn;l+)IJ
z1S<E$+L}ypt{iMT5apa11dvpoU|)GUSKZtbS>?g6kzJ0%&+ePl);v!mHlwLdFTCK`
zrjdBn&=BIprhgRjSs%xBO*5nz$0zP<g%|E?ZOoh?43<T@z^jY0z98CG)#g-FAKBnD
z7NE`HDe9=RgI)G2x|A3X<jwDnE~TCh*MUl*GWhhddXyH}XJ`e1wUw71uXF9f*)7;x
z3CZ?arnMj9RPL7ePJb;<ZH)wua3ealHy4tHOXVz1y?^H{6w20XwA<xsJ%&h1>7dx5
zt)=8^_+UkyC8+49i%a}G#J38blsUpD!gA2BJtCV7!z}|hR0qx(^1o_u$aB7{4)WZJ
zZ7Nq_U!!#2S2sXk3Ex+bV!@@6`|5u?+86;$^A)UT{<jb|Pe%$J8*dscB7DUUx3n*V
zabmvAo__`+)ZewRVnsflzsrV2Y4{rsXE|@hE6-4^JJ4Gq*THUkHcG(`$gxhX;79TW
zY#}ZKK-z(>=G#;Rnk!o2K~lJtyzi*qxm>iM6G>X@c!&R;gF#I4a#ZkqkYDyzcqOm+
zCD1*iT-e7EgG@TBAbe8J6;R!798OseI)HdeNq=MZT_AKf9)A0licS^~J|r;xSXvNF
z$jVMfRg`gf$sn`*838?KgH*Femgpn3kxM=Y$~|u1wfy_U|KI?dtFw!I+K6vswo=n}
zN$jI-EoCbgB!F6!sYs2L1()1d)i5DoS>TD!RpVB4L}QjU!DpI{Mu~7|lH#u0R+8HM
zt$%*7f$z<&amsg3|AB%_j_KY6-2n^3-Q_|f8!Xi6TI~kLrr5EIB{sGw#%lacRGG$<
z{JW_&!+Fr}mih%}od(`VGF%*AavP-3qcrFL!C>!8tgkrw38*RkTQ5UzqRXjX>Q=Js
z0<og7LDJQFzKttmuf2F)%YS}zcaMWUdkJ&@+5ZCa!Ng6Ih$<!nHZU@mQ34t%f9*VJ
zZ`?+f-~B7JpF&|IhE>JW!C;ZZcHqDclC`i|U}k`@yD2qeHreJevgX(CyNXv=%ZigM
zc7cT;ur3y>UcLLhYIPC7pSnmAKaDQpEbvW`Tx<`|K0Nb-D2p}q&x+J#bP!N2l_#`*
z`TC$PUNp~sf<Gm*Uh}lqvw6R{e|h%ht0=w5e3&u5xVgQE!izZ3ew>(#o886VyjSJj
z&=!~1p$R?poS)w|L){nc?WGF5eA^Fs)p5sd(_$g=`k}4M`i@(3eq{Xkdk_WCSHM$o
zOIi2LY-Hpmm;bo=9+rHq0zZo~UI@O|6_+~l{&}e)uc&bXTu_*KO<UfTe|27AE%nMn
zephe@zEl;rJ&yf~oBDlGbLqBi4*UiVk=}-%FyX!*Us$Xs<@Ib`K!}3KyDRFV&HG}<
z)uBT`zyx2uf<v1CtxW8zB!+2ex}PsI4fEjE_Ao8Kus9h5r;;Xi{FP(M#E*1j2fX9C
zhTbuiPJC_0<@L_By37y9e<A<z@|qxUt+nxkAYC5Jf4+YA()EqN4*`i}3STDR#%mM$
zNv64Lm-l&B0I*e-(t*7xz8z|;W*+>fa?aSBbL-oox>R0YUcbp7094fWW6^GlW8XIQ
zG?8;(ksrmWUDn$-KRT8Uo6jadGNR&~_hrAAeW=I}GqVIlzG)ETe~g}hJ89r$0X~&?
z#MRY3H}*v*rF^>`T0ZoL%Qylk(dZu<?miTG2gKKLZJ%H2U^ct*HU*%vgH-zbJz`}+
zpfH5fDg!z@9NWbTsdHa&@1fZh)fbmh<aOMW?|59><!y<yg^g{q!Bq@hUl(1+Z#FIW
zmAy1<^J~|#j`_Cie;+Qd4NwJf3M?GkVpneaaw2SO551w@7HyvcZ;&@W40+v`e-8q^
z&HJ*ceQv(l7cwIeqRdJJ?OZ}#R5&E@np=FvBZG^T@f3G;`v#YD-*GkPfh|x9G7K7#
zAC6VQ?YlO|b>G+A#asnTYsQf+@EzA58CgFRhI&`D)dO)&e-i-^tGw%Izlh>EV%xNB
zvF)n|t_tKjuJt@m=v8juZ)o3uW4|Et{esJU3ULT;mmN0|^Ypp=3iriBBKZJ}y_=m?
zE%QuFa5f7Ug}<!XC!yQb<+TPpzAx~(#&GNzy28bs(%8>b`Wsni`WWkse4WNmVx6JS
z!uK)enfNN5e{s(<9MVJ}X2)SjrMP7h9|#A_a|7)sYn}_@1)vLqni>l#bBH4zP~_V^
zk0@(y;e*5iP>V|8gS||F7{^_|EF|q0KDXYu5gM2q$omZnJe|dS_8pet@R72x(UMfn
zRaiC5{ITa6Z9Nkd!-mB!#yO=A>C3SSmf=P-zcV_te=;1Kc<*K+_Av33&%?gWkH}9t
z9?{fxk|TE$pU{@TG^#5*rhsyX>vJq+gn-Uy^p=hbYrI1fSS1083)}OmLAu8x37Oeo
zC<vt!fZ<)>;u|N8=bN%34yz*o>OM#CMv2T5O;Ch3U<dHUP6mK9Z5Csb1#;lKgd9Hi
zy{@?}f3GV0R$DkxV+)+FW4xC2q`dxVsnAu*kd*+%CKRB0E<kYtq&jMSpw-djG9Jmm
z37ZpDO-F%OHO+f2iKu(v+OE9C6TU54{w6drTgMVp6qXwJY>N_;wHu^tR0))y#EEc{
z62zE;LjV_<X(3NwV3BXaaXi5*1k|ii7GP=Hf1tRm2}7VNI0Pkt^~dsnkOo$K#cd2M
zC;$=)lZi7wO|m7#tP&drZtjBZxRNjc#jH9sq|^kA@m&vcY$&^ZzCkgG&7y4w**Up5
zn;h{BBn`<%EP#OZjzEvfYY|oX0HA&Iz|VJEWY$zH<SCOFQUrmADpuoY2=Sg{L9&yl
zf2YTNQB~JYRX?xFjX*=|lw}te1*?Hbw|UJ?R{!UkOTWSPg?@wY7o8h#foMTtL|&C8
zf=^Twkqsq&re59jwoAq!`CYBrX66aLBquuwob#u`P(qp^xPg{cvYshWtmV1Z8*wPZ
zoU|r6+*fF@RU=-;7B8XxL?VX4Nj%>Ve^vn{PRT6`{V+0563BYg0_;X%(!)R)hoy!j
z8kCMz4yGNVH&Ku>fgh|%2(SaiNtz1yp};(Y+wO5MjnB1j5QDKm$B9p=1Gj3p#2?2J
zLjkD#cBo{sX7!;>;RK=RSY<Py=rM-m8$E~Cr5B}sl7^$X^uz0SH!i>_U>~C<e>2xy
z|D7klVKRXd@h*4@NWti@$dJIcZB^v5D!cmYEb!wfUP~3w3y}p|l>yXVFqGQn?lNN4
z2)jskq>>XSl+r$Y{wYWis&N?5YuoIGElD+)`nE;lZtCZ}3d`EARham3xYF9@s5<eX
zDz9@<K(MuLES=SNh$&d!7004pf4sG=0Qn}KQVAn$>WP$GMnIKs-(PQ<k1W3|AA=&;
zZbfx_UXCGlf$DL;FG-z*BI!lOs*_M~Iz+I8MFW=RXiH{{Q4<tOfB+x4e1AbmVy<qA
z1p<@#dCOg6gQz*!wn9a0dR4-vN%de;wI;TSJ^P|Ma_i8!!~m(OCh`vXf5`UG)g`DM
zd%q65yW#*E!xi^z3X7;vVO<FdOmBp_+!s3`cb;^u5XnHM4m4|qj;=t<zSxg|jW_$D
zyE^HoI*}nfi9ZE!{s4S1w@}1_>J!uWZAsb>c8fKjaUTO=t80OWq6czdgd$+U6X4JE
z9R-$Hmv~T*-2@qUe*4lFe>#kSN@4m)TOA19mpbyJQ1KT-Sa1OA`<mbUp}3bqLu*`2
z`-+x7`{*gHU)yh}*f)vzr@jT!1e|HfGS`8ySPce#g__k!yU;_Q3eqLC(SsoUyW#+2
zd&>i$1q~BBrvv3cdtg(e8ORx-ysjXAbHz1a9aVEDoeXqreo^2Gf9jHffGyfr^2&X|
z&ykZkEeQr$O<xz7)()7u53;OR=MCy<KUsW>qkcXf^>hpY9Y7SzAuo3mrrHaA<v6_<
zn$65>7m)vAl#=+O!*fk~;prMw5c#QI`Y+d7g&$1cdsB$FzywQAKx9Tw;(zf7=!*R;
zF)90ORgl1b{V)sdf6BFXOa&GL44jSol?R2{+^!&&EuRyq2b-8a7yCMhFN`0nm<Rnu
z289dPv>zh}o(=n}3`^$J0>-GHL~4byj!i}YDb1UV)|*^>88?|;EXG+k?qZ(?^o>y_
ze{?`A_JeQ)#P5VubRJSMbcYP6Xyz?G#TE4Rk9e>vfR>2&e_PAdpydj76QVdNw$Ol9
zR9ga9-dw888_-B+`hz~8pVKX=F!Lv)@}tXd23BM;sWPh-pvgb^om7^Aeza0xNbV60
zr{XShoDHCK^f{1S0c;Iflw?r`DhOyrCh-afMuHw1Yc;U|i(@BIl5BYt9ZxO%3XC>%
zegNwLfd(Ige?52H&TK0a#p(3>$^cra5yEdBqyR@0M`uW3vCN6yqh$!i0B#2>L*s^m
zS_YPeE?vi&q<);OQ3IebKNKXG8Mp)Ip%_50`Js@+fzNKp`;6u!3rjko9rxw3vdLbF
zM!WNtx$qSaxfks@Qod3v+^dUX_uNIuA*?3(oUJVwe+Gty%!-l?I-io%$+v;C$6EiP
zj63}kh{ktmIkEAxREdz~ap=+ub>i(B4%_OO91K_7iaG;64}~}dxmwmdlKgSlP!Boz
zC$LP!gaz_V9eW<ijLWIxi$QJuf>}b$e<}R26;v7V<v<0wLBU1w9w(Ecl5vzrXQ}rK
zA!gU3e@5%jcf?^f@qmoNG+%HL6#3?0nZ$f1W=P|^_u_J$&0<Zg7^54V+i#tV;|{gu
zO*8Z><%t$D(}>~SAX9`iF<|XQ48$E6G1O2qVwLzOikIDTKxr}ITaNSrg^IrR9owZP
za=L$7(+1jqq|<KDzyILCBix7qJy|||bD0L@e?|s9Yt;Io?6eE9LjVV*w&>6u$CX~@
zSe-npeOL0TQ)@*QfnF_QIU&G+94z?*w-Kv?=LZ1{4-vmu<{iOG1tp)tU0Jpi$2gG1
zF-Jeoc<R_HTFZnKRKSt6|6e<DY3{h-Xi~K~uq)_<Wne|1{d8O@>fIHW?j`hVBd}a!
zf0HK!q}q!88j(*Dt-qQaH`r3vB5D#p4Asi>T??OkxZdO)`@mx+<%-!OOn|U|`ORD0
z(G?HKxuY94yXKG+a}0>%@$30^U)IH07QlhuI1z_}iDUKz#>w2fk5#4{9B^j|cOE>4
z5gfBky!~b5J#5-iUW6SuG1qQOC!8_!e;v%ZXXoU;;$~9%l2qL4(4Yeye8t5V13x(|
z1BJIu3ruUfN7%@43K_u@G^@VcI#)4Npml`->UqFZJXVoSVk24DR^D<89ek3Io&C|J
z2?xR4Y3E{IP>Pay>=oxD7ZQDP9vX!X>-s1hA4@QHqCUsyT3W9YvBaaoi{HK%e=uNx
zK_bG&4`RLYiB9~AoPqG-L>5@Hanluh0xdtw%oEvD_(b*;{_kN=@srq7{3v^Zfs><n
zBhck^5DV)P<y2mk{~D9M77M;sA($pnv^=wq&Xwcd=g<xuD*_t^_BR_uw8yNHEP*lc
zjEe92W((>^kIp${jX>bLa%7HgfAacH+)VQ7i-$Ttl-tgRHvA|ur`dS#zI*dRG%<4Y
ziA$)dDoLNYazFy?pA7Z~hEkx-5)82$4ABA)7Qlj|%SKb~FUsYB0R;utzL1%r3n>-g
zlv!I5Pg>CmbwePQ#pi;+r4@jM11FmJsTqkyC5|nI@Y!&V$=b0t;G~H0f7xtZ&tBd<
z`!J?{niLGc&|E|*EHH~P_49B42rhQe34GW$U>n}k(8C3&E9BLxF5W%+i4#N@RAVyW
zt?;QR0XyINNL_YM;%A|TMW#m1`K6Nv)jZMesqEt1$%N-m29(Dv{e%2xFqaBH$yjEi
zgiN#^$1jipb`Z_n79%3We<z4={fP~n?6P%Em8jG;)nZjAQ+*|t#jlPn>%gCk@V~OI
zpj=zUyvb1;#u-5`dH(!4m16IE2d@i(gyCR<c&7Af%c(IH6E>z57GP>|8M9gFIA3I(
zW1#XJY_d<1dFtymnWM8H42<`K3py||h<^*7yLd+ACzIpqE9c>Zf4yx7ymm@x6E5>w
zU7{s0**w$4C>vE)U@^`m33e3ifqPg2Eyor|8^uikV5=x@SN_Z8!&3=tGkQ8PEKg}A
zw?m`mr$kXN9eVx4vD}KH1l40xJ=D#ioLY9pHh<uB57zX5Mr&l$lXWP{E);Gq6X+3L
zmoXBs<HS9O!On!if92p!m=-<a#~D5uX8-pX<sW4r0{WQv#AizoWzddEIoLMYZ(JXB
zV@@W$xK2T%iIe3Sk(0rquU6SI15R0gH+5C{#`=Gk%1!9&=(E7G*lp@3z{uuck+?f9
z<6f4B9ux9nU}G)|lf|c!V9EIqbMufDx*&sfhAShqxhs1ce~M_)BZE*QbT<;tT?Y$B
z=wS3;paWg!paa?h=oZ<ZU@7_h;n4#H>&dMa%3b0Xhb^a(Z%pP&azg(PLNDg@ZrJ-3
zR~h8-Ots~@KpO-2hXdnJ>>b;D3+z#P$o~kuaJz3BdlBTsabz8r#4yjW#Yn>o`g3)J
z6)~wOIWZm<e=gLnhqk<x=ABpv`NFU2(xC%ihfkEQzh!}4am+DjUP*R?1gHR8;?+t1
zye})K;7{iXluu)h-#gs~Q5?>f#u<=Gaxud|dv^n@weAb{8!5h01nqMNVqb7y-nk@T
z7=gj1K4p`er`qJn6K(S7Q#N`0DVsd~pR>u6xv(p5f8F6}4c6F+;c6*orO=sp<V^qO
ziirbjpKf0TPOfdHbb#?<z=>`%smb5F?A5?egLGO5-@9|)LzuKA^c0!U%EHLolzGQZ
zW0uE`d-%o&U$w(Wcc8>1>=JLgi}|3yOu}#(=$|0D;?8LrEi#qU$cm$?yrYasYNPmb
zx?q?oe=A#WhxSOW%dfa??T>^ePW+C<9|;BYN6vIGUd7_%zdm9TUHOp%LSJ+Klf4mD
z?)b_Ah|=9G44>g`!ed2QT_g9yJf1!;D@mcUqGg>zQ}eYhj=kjFd-0F7MJ4G!vn;7o
zkMo$HhjM6{mpA)}joh%KC-V+N)t7QJDU(}De;rpwcTnb`<AzK9evs~mvZ!`awvI;o
zLG7oOeH4dd*`n@@`845@r;YIsl`M54zz>l8p%C=O=^x|p>ui`S=$iU5`90xq^^EO-
zwPEd{C|GbDlbmECXyA!i3Y1FR5wS%1nZ1kXQa2FZ1|$PmxN1!6?Ro$CkrV9mo^R=e
ze-Z->Q!~p2)yQ2}){hMgf_l0q?aidKP$0D0P@u6YsX;H}(Y9e!(C1c`dDO}|9afN=
zbdh=GKvGgq!0vwe&09*swrS;Bq7V2Sx<L#hsO;#%5f0o9k`V{3eVdoH=u5HaOEFS*
zb8I>p52)@~KwBJv*m)4jO8z=ZMuaRUfBHjh{1_#!(Y=gqVXORX+8wEHB60D#%35@1
zy46B#v$XwHV%tfw>^gDR8JrFHwySl@0lOCW$x8~cavr6P&HlKIeMiBxt|z1|+CXeN
zvo=FaQ3pk&stUTDjj8cPe_zn$=Op4#2e$5JVOAMPA*(6HT~>uHXD;)drU9B%e-r0?
z;G%?ZfGQ}Ofl*=3jMrm&yY?to77JZ@C@Z?k2`~JX)Kf9<2_@jUY4|Et$oG(ZF3TMX
zEb>9}Z53`|Dbv2^8dIjmi|t!UCx><H3r@s!MzKoTa-um53C|ITn7=jN!@(X(@W+w%
zvQfwNoFDK(&X!3O>f^0&d7QEVf3fa!`L+E_e*O*zNY`0#O-I`MZYXz650)X-R*A)Q
zx0EP%$NhA(BjN*H;;EkH{2D%G=i%6U%q9^V6$>alI9~*7luthIcHU@A>_s9h(S>6y
z)Wsvar`y&#R4}T%+i&>dLnv9Qp*3m4<$Xx(7_YoR-`AZT%$Ki%+3K;qf0X{VAZ6>y
zc^S@`2oVJQMZCl=w@aXDx?Mt5x*W`7B>R{$tB1o;loHBha6_M802DzeF)*ky0fTT6
z^Z&6dS2QNft>krrSE8r+x}xr+*vP9did#2TjJe2)5_o(ry_Oca%LFA$e^*i!S0b|p
zl}ygX$rfj_Ko8_DKk~gMe}!tB`wsdEqh$o<12e|QTmT(`sJAXdz<HNBkor1NpW~1_
z%}krc30!VpFs!m@5luMf`%PDg#tUWWRCjt@BQlQ}X$qaWs1T_|d?QIYQ5enbFvj3d
zB1=}MF{x04eg@#ya6Tr~kHw_$KqzW33ra10mUgZ{fn}UL(MtW}f5Egenb%<Y)uA8b
z4Tm_w7!Vi}=ht-|6Ob^<VY2fW6~g{TwA7AyqPd}KKcr?`gEPZR7IY8dk^Kh=`H163
z{B&UKCJ6>s(@?Mgd8XU8HrZ3cfYGOjDgG0P>7$bZW*G`+JQRSmF=or6Q!bshW?Z2w
z4Uo+~5xR=!4v5F|e_Z49UW$+n`Prz^ovyOjUbKpw(iaDej*XS)Px<i_!T1DWso!al
zr4$UD6ILn9&w26PumRh5Ok@XUqBF8%mBrU1OSMcSmldm65L<??nxmP;?Pa|lI9C^v
z>})=~iW6Ax?~e+&l6&dH-#keE)uhF>`JHt(Un@*dO4rejf9axq%lD)w*8yH299;+4
zRgSj`OKCaeHF_*Wp`^Rh+#zegU#2K`i)+2cYs>DvZF)>q$%2`xLX}BX$$~Sk=OFOo
z@Uzmump^~y$LKejjT$v0_%z>|3;3or7Un8aM!)9+X6OWYz&LYqCs<rQtMpcA(Y~b8
zAe$)sJ8KBaf0e1BugapGd|VE9(8~{EA%W!-dj=oL;`drQ?tzmixWw0niEwc2%<_s|
zmKU{Kl^<Ui<0_8T4E5SB^!bNzIGBp%6jRFS_A33V4(ID0xrf{z_IGrsg~?jnTWp<E
zkfy=9tlPHjY1_7K+dXZ&|F&)0)3$Bfw#{knS^KPu6LHQ(Mts+GQD4=YnNPi7`Y8YH
z50F~V;`;N3$$cb2Ke6#mKVs7B(64v&tHhvfk6YGnZRb`S=5}ljymr>j{wttkG*aw9
zVRA}~Y*#Pi3unv)S8ELn!f$!KN_a$gA9l8y`*k<)`yBc|p?hZ+Co?14|9t%?b<fP&
zSTE#zfywg!N#V0|vizSE{<ZF9+@^Tq&rPrJu6!eh@!$TkARr4KH+G?}EWTM1?^Z2C
zstA^Y3=;Ij2EO;pTK(y;rH3i3)~LaKM#8l?l+MnMj*bhBVNcJ;t3Mn{9b<vC?Dq#J
zZV9DA#))Euw(;cq2j-n`mk*k|`c}<TlYZ{ctyzRUUM2(M46-<u082&LkElpA)R6UH
zs9g-MOofM38)Qf=V->P+Q1p=~#<>mVgf8$CW^8K72Fv<PMc-*vFKKjeEpvbNKunbR
z4VHowghEc;DsoeRQG=vhX$pj+O+2R@yvp`u7o04+j3Ys-qV%F=ft4=FUk$`O<8;?9
zO)3w#ss#<2q~b|?K)jEUtf32xOT=`#u*pZ`D#xs!Z)t$%?*dG2rho+qc1i?VTEj-b
z%W7}MFa|e3v%Jsz{D3s!1`}s3MBop`ZSL3ky3E!EHb<C71z7ql=?e3`#{lS_MItex
zupz%v_ltubDE3vkykXpxS!j*n2C>7^8WHPD18=YpiZulzK%8B17K~h0WzX|~-TWF?
z4z$(k1~elpIroaN4lfwgw7>NJ)J+SR&>!BBdGmPvJ>Zz)UN}gUj7r6JvgS1$G=7cN
zZ>S!ckehuRlQ3&tlh{Vmi+e@a-;tILdFSyMHysBrL=CN+lOPR)+sTj_%Fdh>&3`03
zzn2Fs>jAR_FhyQl`SFKSTYqqT?BIA5Byaqo#p+e}!rg4tciV9rov*d#v1ZN5S%qg{
zU=TKkT*xNWOgH{Dd6+Ld;=`su3u!TdF#IO6!?HG7yJm~q){w+7L|q7Docl4?9J|6S
z?G{mO-tn09w>tHHafKqA26^DIJKyU0HMialvO7oy0C@_iNgfSbX-F$Dnyt?;;pGfw
zMC12+reDq8=^qWb?Ppv46+7xVho~UnsK-ch!V;p3@)w)<5R#GFOp|~9k8RI+oaeHq
zW`nd+oaVO{S>-_|iow^uv_u^k9wPhMn$!LqdsyfzU1X~3Vp<Go`|+v6?H0kGj<XJH
zJMs#DfW@@z!9>d$yKQ@_9Aq`$jO?`g{_`J=TAW#sDum<)c~--!Idk4kM%AQs<44=9
zTUY}8(zT}B45*{z4=KEdo~qln!{<5x2_f^Ejd5dj1c{u^;3kL~%N4Mr*h|3KUG#h<
zH?Jiyhu`jFt@K7!0CYp1ld;VJ1w<a1k0xgmV7TWBW;OFwcjGN|G}dkvkH!-__n@W7
zfNZjBXJ#bTbs#g+SELYs^BBjSg}reaq3yDaTEOkNqiVu$4rVUf{WB{2k6yjih3(Fc
zGdguvVRguh=`%B^CN({`e5#&-Iy&ZK59H42Wo6q*C|m9+Y0WSB0Ga%Z@^<?psn5Ux
zK#Smek9E|5E7GKRqDy<t!}LdrJ@w@YYJ9p&yM2q;@NmOz<VETvk1icXC5gQhbDfjn
zj4HT>6-Zuw{mebuID;fk^>{WtFaBRm`>2)<3Su$2y6KN>Rakr&<QdJWl)PGHMG+!p
z>VCRG@U&*t+xPWnl6%d9prDZZJ;lerfZj+sfq9rUI2FE?tZ!)sqlZ-=L2?y`$FDXT
zj(kenbUGtTPBlV@pv*{v;3#qQ%rx||)ztIK(#y8s21&56!kItcxjI(5`<m(PIU+e%
z$7>QO7@%d=dmdWOpaRn<8vWXMq8$~Y37|S4r62ldh@P8iStcp1ZHd-)rqd@ofE~=|
zRD-9#t0p%o1Z6DC%8#UJJZ5+u-Z$;nYWd2no+x!B!`bf&alXJfdf0is$TM&^P%%f)
z%N#bVCv;=NaGjkD)^V;7BOq&?G+3nE8N%F5oqN}YIfahC%pvEza!5rHECtLrdh<tm
z4mu+a^>Sl=rqqATFPZXx%?BrR0=`z9wdQ2D^&Az9Pwr#TG%H@V#<DWVSV(WHM~Q%S
zNMp@0#u+BCW*vy@sP(4K2ng<1I*YWP&=#Mf#7?(Zhhc44=`9xJ$7D6)SdDU^?JZ6k
zl;&OJ{UqpCPMtv_*v<3vGLj9k6xz&4X8V-lf3s`~U#C<s&Xj(<itg?c0vg8WMY5`T
zZIT#dxx6fck|;hXEm5+FnwZu{)Tg&w*Wt>a-nygBzt2OuqX}XPMSqIvz9e!s!%9B^
zK-KTr2y+S;c$aDi6oHRX3nzhseZHj-j)3EK?!-OQq1i3Gai6&;6s#|X)V^9iXZigx
zg_)+7ooj?=RomzV-bOfc0LNU;>5?@id7p!v=6qksSP($g_7ayBYmgJKy9`M;tzU7N
zz@h6!H#4ip(%u_v(7fLR-A|n0IVeq8JoO;BysC8+UVD=hoAE{<#oJ}p(v>?D9$I1Q
zrP3!Z6EV$l%&Cmx9rfKD5*8Jdv+4IggYlyfM8bH}fCYd3TJ!A{U;?SHG>m>!sl<&8
zz|8uV%;oMf(eaQoQ|}#(g!^=8UuQcm-DBpFR&qHXkW-77y#=ROW!X<3XR|4Ok+IF@
z_b>P*#94@BHp*fufdV7&JNKLr3VAT~`rn_BA%wysPWP;D^vuv6%xO&UvBGo)ZA129
zo%RjptP*`M6-}=X08f^#X-KG34gm>6O`E3a6|IZYb{ajCYeV;Lx0X_GlwT#T!4zGD
zy3Z<}*Dq1kJFhdm6C<Lt_7j33zUyxq*rEivucQ))+H?!1gPe}&;e?Q3=na=^Gmjzr
z;Mpmt<6yfWA*ufaHSP`79Rp2=nJKr}BIfR|ck59|sn6*SFcNOs!t9nzMJOVId)95*
zZ~_f9Le3Cr9+?AIRRZ2a_0@JWsE$cB>Z_)8#0aC(c?a*jN$oGe@eEH<f=j>|o^w7U
zBuy?raH%4!r9Dl1<@JF8u_}GF)OxigC9x_3Bc{PEnbfWm`Ayn)F(+_Z({O6xSd!-I
zZu6dY4kT9$KrkYjk=QE@^ybxRG>jX%o$@f7Gf!26eLB@!wV6oBowuY_Log6~5Ye%j
z<Z?O;NoLO%+*5o8QLU|tOk;4MOz|IDYCB|2M3M48?KCD01VVS7Zn|aX3EauAh_}!O
z-x}T6g<E|(r*-49Ya2|X)T}#fUeYx)H?{&PHd>AXI4Vof3+TZfkD2AIT_dWsQ+-4-
zN^NzuaP1I}5nAxeqs%9}FV#VAH!%Q;R8a2he{^f<i7LP9L)YQvQFFvlX!@0PRLzLZ
zOj%zsbb)|5ZxFjcw*saz;)f8jhS>Yb`H~DVN+4>HbyvciqG)8A0mFiv228AIdImQ}
z*uhVLK}7KT!5UwBR_ofmXSu;XOuY0SK3q6%Y2XP2IvZNuFU*7)+$PyOTdU~kNdcpI
zM^s#iP(5eNez!R#^uv7He>nwv9UH-$<FZXe$;4G3sS+0``u&_N<Zp^0#kMu5!0@uY
zWd#q28t{*#*iRsyc+|suaW9^ZI_Nq1EC)Y88dp!pc7O>a4DJlW=(#D*M8P)142970
z+7M06k%hZPonZtAF~Jrd(lhP0jw3e1r?Ibw>U05@I(&rP1ZKcY;;~FJvYiOu9b0S9
z^ySG{P1P{v(2fROItQ9684UANUCx{pZSr>lvX^TbXB{en^UVuxO`nE<BO-Kg!eKpN
z;`RdK$epy8<Z6X;8knw_D?ec8bQO2gWO5{3T;v@iIylotbx>#^l$GQuU}Xa?_|9)_
z+M67AD38hZjC>e|&nNUBaDgQdMBKKw_QQk=J&B0~w!7(x4EsPIpF@Q*Aj*bN&<Kju
zC`6X>im#<vchEw+(|;Kj)Rr2!Kb8jY<eZgl2^aMt@c>`a(q%0;Ie|C>6?FsR-(#WG
z5DJR@&pZ|hsO(Pk(%5-1CZi~#Kb0m9=}#QO5m#;-BF5@^^WBW0P_7!rqFv~sUajWQ
zp7XCgkL2ImAj6SdLO&zsjuE2)t_R)3=_x3k7!1W?C1n7XzWNM?u(jBEcA5l;gSezV
z@1H>pgA}aDxr!dgrgWib$IBDF!0R7h1_p+hhf{O}3E42T70^A@^aE}2tv(WnY6Uvr
zy+y8%JE~J+ScU@kLPd|xbb9Gm2Z%|_g3G`QoNErCk}_F7CgFvJz8i-#Oo9GIrs1Kq
zJw7EW)kU(6L5D4ShS>(M(S-u^%4H*7ibBX#an%X(gP^f@;xSYg(gA3S<GQdsZlsxM
zYz|g$Z^(FyclDG5Zuy78)yhSX%(xE=%ng9@W}u8gYVSplhS<moei<zeU36jHn24~-
zQG{Av7*L?(r01?~{hfbS=H)Si#k<(<vD&(%iV244Ge~njX08zoftUx7$S?GfGo9#s
z{e|n#3GWz1an&qATTYa0U#acc%^Gf<ZWQRx^FEC<=*aw6_|T%xg~R(p1Py-RrLu!B
zMJ`xA<?QFTda{{#4O^bTT|Q<uSTx>X_vJ^~MEKxG(L{2y@1J$7vo92zDQ${JQG
z*4O91?lL$b@Hp&QpJ4}(^;p;k{j20($95W)MW?``HY3iBcu=$fn+7n@7(Fm(lhsa$
z>LQ^$`rOOLJ-5!{OB-orwbkF9-a&PE|D{XEQ)}GGQ2AR{7@7%0z5`1sJ=2wE6^V~Z
z>7#)OQy$fSpkmY<>S`Z%gk`Vu*6)+u(E4`d7$zt5z6>)cy$1z=HuHWjqOy{x@#Gch
zBKSpz108_79x9Y9-Dc|2YUOyaV%Of2`r`go{}U4xvS)MEBv#<O0Y`fNnTqNF+qNWg
zmZelPa%(AjZOnEyf6D&}=HrOYAb5;a^{oKDq#}_2?GU;<9G;4}lRG?=3t}&GLB;p-
zPN84}k#24itYiZ)D}+W<_(%ZDKyQAs=O1I0+|#2$rNlH>+8I|UJo~<jFY!7vzTYWU
zMlVjD&KJs>0Tp=6R)+g*%U<}ov+ad$U@R`?Mi)?n60GR_HD!*O6e$oPJ!?YF-fTck
zbFamI=3-o8!6`$VgQS;O`4pjiEJSW3etyIztry-d??eWWJwpf8d{;$iRzNU#OZ>~Y
z_$wEhq=Dl_)U>L^l70S@K)2SK&>tT>A!f#cM~CgKawwh2+Y_DYI3~4!tRBboZ~>U9
zY*e7j<<nP((R(llvoTsCI76^M*bc_$s?W7oApDoZ7}GiOA)D;LxNn(+WWRpN{i5sH
z4`CiGF#;_h@6d)=)9ErW@0YfWAx08#0DSmhaOC56IWXR!E4^p(lK(d=|2WGdql`sl
zL-?sr-YB=4;R3O49X;o2O6j3fKDsW^Uou+3Tv9G3kZ~<cjr!jRjzW6b1~gXA>Rc3H
zQCDv!L>7n$nO#DP+?H+-QH1nnM4rL&j4=sLtXzZu)c)V&qpAm@NPgQGDT=G$gKLS1
zk$o8HRd^{XSLT#Yl=MV8$?Z~Y4@0M|T43Z{X0#}h*ZvS-R_(T{sygBUNK5oyc`s6e
z4I>^St%ECa3Pk3$3ZOPwT@QaCSfailc<*3au;Y#@lWK>J@~6P0oDv^z@{z<RDNQ{9
zlw9+Gyh<+bw`cZRZ!5<Hab_f@1<~<=6(fr*yc}z=u$uN1;_WEt#OOGG)5yo#p&x&Y
zBZ{=}KWiyG+R}<E6P_n?!@Iba$48-pbz5TGb)YJ3Q3i3*ehArI8K_2<^;*UltiZ?g
z0CXe#dx<Hu@k5ALG#Jd?tu#C{Mz};|xFc8qBUZ9yFOLQ?sfdTxtU}9J!Q3_WPW@Sm
z>q6Aw9;&%!6Cu_U%==7V-k3YK%EZ1eN_xK>J&kPh3veO{v-+Jjt!%7RA03CwTs`Y}
zF}n7{bAjmKrs)Nvrc7UP8e-9fkYPHA05&ZO?|y18y_A+6Qbg*`t}9U-U#r}9B5e-f
ztYMdlZb#-8tL`;MzLu339O(R9&XIkD-fBvJU}Yv~cz`SupSzxI#i%%x&j?jI`{zi4
zKa}x%jFT@2x@rUII<F{&I-@W}(c9U^$hg`CJw26f9yqj_lVxi-q6IUs^=Pk~)1Q7f
zHLvZ(2|?9rA~t7`)X)yw=gdq?O(GibfI1X)-Py0FaZ@F98Scs0RPd$_ysY&4Ad(5!
z6j!mvGDV$pY+@agx3fo@hH39r^BAxFL5yvA8`)Jd2cL}H-`6U+>kHp&5v;4QXM}jm
zSu%Ep&5e|);sQjaxNx7r5|9tN#w-R|_nlfRN*gP_a;1VBv<n<N4kgg#cR344WS&SX
zwJH)UJ3%`pE(z}MqhYeGTNoVDRgJiS7dT#$BCWuugkps^J4Pb084=^W3m?<=`|fU!
z_C*#a!xAq&t_d6nH#LfZ(3(~xv>{L=gsVL~+wE6y!;@K8)8Cez_{WP21TTtC1j1C`
zw!nti@xysx5BIWbcbUx76FLW|uCTBSc<6%f*U2j#4zK;~Eg)``Koo+kJo>yl!J{Na
z2DCCMVzt#i)iFu)fXeMy;JKW#97ByY%EN0bx|7C-O@FF)``jd`SEFe@20gw%qBDC1
zeHJyb#&JZfbDkfeGU3horh){GO9(51%(!pSJ>n#Ntf`~eI};Q;V&MV;bJLU=i-?-W
z)65m8d67QVH58^h?dgS#|BhD#zc?QD``Y9j@~k4KEAhy+JGbXF!polj+lRht?DzEO
zj#n9W^(UGdh44a>wnZO^kAdt$LW7<n19`h_BcNqdv-cCSrA7kswO+7{9(H16_fGy2
zZit&Cp|+5epPJD~Dx?B<r&cUwg<Ndlr^nPamo3M#%@YDp;w*IG!&nCPS>yCbr@#sI
z@<aoEPy9yP+#HDvVqIblC~lO|m8bKi{8r`v7{aux9?QMs0^7fP;DRQ>pMIBJ4E{LC
zTqPd2K^53H#Wd#h^y-0)d$UA`G_s3A%!Qy{cBVNFw1Spr(fR-sz3O(nNFfla=N$5%
z+!@jR8=i=5PLx^1Y6}w+e94gv=b;*BjxAXsCv#7wx}zKxDym)dq#?|6-Xhp&hZtvt
z1$r3<F+i}rpTt>v-nxUIq>-J#7w9cbJ4b}xHrq#aS!|*T-0&Zjor|H;JfG@mz4s`O
zC<wZI26s}DGw1`RlfI#PWw3i?h8HWP|8jCX>Q1i8N=Sv>(K2o(9g+9{^f@GIdhq5G
zLZ<?j6}2U%_C35B6e8adj`hZs=q*P)(mAOWc_g=>4S=Z|ul^1!o5>W2u1`LdHm`nj
zuP^isN^zo+^DG)TOh9~rAJ6K-zR6AGGb4=JjrMpRse1y5YcV(MhtG{%O9fdhEF|LS
z)YAI>!P7kEoP)hKug>WU_=!_h1>Zviz;d(0`h0hBJL58$KBQ1S@2s~8@N$Kn3c2l)
z?ThLo8u03D$I_C){@}hs5X!K&_s8t6pPUs0d;olJH-}iE6qX5<5VI@?XLW_83`c7p
zuuz(k6$1gDXgc3)r;3`80vlR*7?p;u9bgus&CAdTH20fJQw=gIlN-P+=v;QP$5O?b
zb0kDQ8p>XnFNi-i#&-GQOukYN|7x-6f*AoS*};(Zcs-b-3A!*aTlM+oGgE01X(%!a
zh!`S&iHyhkjiOPY%p<rJe>Cqno=_HIs{88J0xbbk&%<|oMJzo(G=?e=F!9FI6*y2(
za>{>0!jlG16C?2LC*H3w@CC>qB}@rv&OEbkMGE_rWeK{X=6&(m|NPE;#qVtjYUJ@d
zi=i8&8k%5C^5L-wHZAz%fIsRR{(ooEER6rxFk02qAxuFO_@6-j3G|=9{t5h_ApQxm
zRY4TW`Wu9m^}oCr-Am^!_Wzw!6H2(^5Zsel8T_VEx=L<zW5;c`tEC=Sb2O)fU&T}2
zH!*v!?X_bS48)5j6KHTXOM?Ux<i+{=WIGect#8Eou4DN%i;^J9ZEmqS*k}NVn{`wC
zIn=`YrRC&i;e|h=O1t06-Rbr6YO&kx{gPLk%sp8s3vl**x!H*BUzC*e*vR?n_}kK9
zwb8IkIJ2rdOyl<b(Mh<U;_@Or6cWlb4SV|n{ye{RIhaT(6LR_Z@bJ(89Ze=m#&Yhp
zjpjOk;=5=(Yr(L~_hHhaLSINLlg5Lt`23W^j8Q^WbgsYJD)N)9{)+{m#%t_&qW&D<
zbvJZb04UNVuNjQ$Lt;+kb;ZBAh634{L`hhH;$5*IS&&sK`+<@#%HY-U?$*z)&5&2O
zx3##A88bz*IIx-aFK&6YXF3fqGAlkfn!2=*=;k+|xoH_J0c1KHM0S3CVvK^ckxFUo
z&dt^aNO6W4vv&`u`NL34W9QLjfUnXR6(2t50<e;Yk!g1blYg*-B7t&ZA-?!TGbrZ?
z9@n(6<5f8iqSh#SBHIXtuRr?2p;IwPdw2}W1~$=x_EWrTu8A^QMfyo*8rCs-VW&z8
z;aYrTTe*i49mx(|*pDXtF~&${G6ascdn}`nj$M);63T>y<HfZ9s4x=|CI;*pqldW&
z0g}moSHK)msR0Gm^+5;X$(;N>>m#bS{v@vh%P%zk>zq6;`&BIl=aWb#h`;7TQgON)
z`kjn1MS}czywbSd)aGix;laCw7COHmOCuz~KyEe*stZ)fr9%44VAagvMD4vEiVmI~
zCGO(c#Lx=Hg&2|6){T~0nTWl|>x(Yc5#WD|6@FzoSrfDfj5|FNad^jd?-+Jf{GDEV
zMy4|{YdG1?Nu~K}>Q~fPK$ajYPT-NXo2W9!@ziM>+QG1K>5NhPU@LHhzU=P1gW2Md
zWgkU3gee>qURM5fzn4?WKaPFq_#>zvcno`DFIk+B3_}h)--2u57|$sj$P7u30?6k6
zw)Y>gj^`o13){bcPhOEwt`i>(^rp6;OJ6*p&UM8cPhfQBe5W-F^dOQ$w=AVSWHB;#
zp7jrq7VH+VO9o#FeM#xx@Id2%;o-}fDV=JY&vcx!$OMqn3+htt>5MOy9J9Ak*SQ(!
zNHLTc2|XQS`@$S!b_FF%lo}$!05DY;eaUlab)8#IfLmfuA=~Bztj3&XOM{fwrkvd0
z_gkpBQI-m5E^(>#zwS=FsfoJ{zL}1E*OycW9XpLY)eT1jnpgK~$+ozp;;jGDIJ}7r
zOa;g(!|x<I6^xUU6~iTSjy0@a13=*70}c8jfY-`c`h?*YlA8vGfaMX^0QGphQrJ|=
z;#XD^lpmE7l6-R55fqf9DSVYNZ<v-v?e`$PfkE|8PxD30ZteD&^pr+;Pzm(EWFdda
z@((98=Tx3#|5?_Xf3`-ubxy6ZyQwxJ+m9NcPh#HkFP3JWM1OEp`fFA$Qs4nGGk3;j
zp!bc36$Z+5s1%~F0@br*2|yV*IEvV5)unz}ci9kq&=*+X9~~GWE|(*O@zqQW0P*O^
z>;}toT8zCJ|JQj~u{wXXW!bvQ*;G#Hc6R!mKidD!-fCM<(^`%cBr(c@%#>_#w;|!F
zn0IqKv+jIy`_d03V#(I-&0K70?#6XiX-&IXq7x?}fEbMOSh9OP37`f84gj$@Kf73(
z7M*p*uKfl-AWB?@{~J3@AIuyuldG7{C_ysEmN0CH{39lC3o;pcr}XQv!ClUBh!x`W
zXngDGBAZHg!^T9@!mPWzb$G3fCc(O@{TgQT?)usmSdQ@QoY_WOh!^=t#5bP%GAB5Y
zea)D?`e%PkQUDJd8W1V+lvs9BUoop<z^I335p{kLZ=LZj7zWHH92DV!7Q%i8Yp8o^
zT@fclgVwn83N8!_sWls|l$w!}#v$~tLpT*ROfZbi9}+KP3B5K*afBI|^jB){$cMhg
zKdR5)*OlT1nL!^?H&sQSR|m+k9_k$29O(hb5_7K6F*)+79<au<(nI2@y4b~|Ks@4S
z-}I{@GK+mTMQpn>zO%01)7jQ~pe1;C0IQUgIFrU8J^_ev?}HNCQiZA2`TUX5^0cB@
zfk~etXo6qR)rsE3{m6QzJ+12w17Tn0x1f~${`#RhLNe?8GK>3T0CskHcHHDad}3*=
z0>}!#iiAMyzmtkoS_MTp#k<yp6n}ZHykr<szVhO%U@80i%NJfz2mOPeW!zkO(Nm&y
zpI!i+ud<g_E_zU}AcPgVw7X>{!IsoGtL6s*EHrdU1<0Q~jW0em4Y8ifa)MP`FZQqj
zFvO3pGK|1=!8x6a`a2GfF73k^!w41vR{8MtMy;y!4}iOVhr$RY)7_wYXW?ecF+@FZ
z)wT+j2qO5*EAj5jvM=RYJhmwG=ll?1=4ZnBje~{U2SFh*ONjFyhiU6nDYya)?~7MS
zF{t02O1}<JEBRG)gl*L9X4flO(LyWnhB0B?>(<TB&I^8b*nLwU6Vdo6{8<k1ErUXC
zp++PZ_XiYNK|BHP%Rcl#rpQ`G@X1!SWXK6?=JS{RhV!ZntOU{g-PawY89P1;a*T=6
zE85cwExDf=ghd`X-THNEjr2?gUzmzxI%C))0B!@WYs$P_iZoE&pU~LmN(l-|t`{+@
z^<|^PES<ks+jC_pc8v9XU2g2GAk#6;19DQ=PzF%t$qw5?));><BrwM3lrBk=x(4?u
zD7*8E4h@lx*9(!Bs)DJ+S*GTHvekFTBE#&blnWT`Z>UWcn2-7#59|x~^8^)Z>dTX~
zdd3vxj}0i&y?e%|PB2Ygj=wGLYt6QNV??4+0Ig^IBX52;iiJ-?aX^B{>P-twB@cmP
zC=HPQ^mfLK8werZ^=|)S&3NCz@E(c5y~^;wbDKy-SO$rgM+06^W#!JLKsjn-!XUi(
z>$?UmPc$S&0Q{MF1s$HZ)R;tTtw>_%z>`T(J>7)V*anSS99pQq3GYl`)x35Gg%ijI
zVt#_H^93pr5IDg|y+C~Ury9~m2^1Y_9|jOFSWFK4!nVhIH+(QvHV|CW4}+HqAvUl;
zu3<Kv97(3*aR_0Jsy$Q3Uat5@<JX$Y8snNXrD_Vx!dmD<6oaeSj@_#&zxorts~fC~
zIsQM`-wRbuA_+?_65rTbZyKZXaO2co#(gh5h@wIu&=RIF&;`|NQlRsffIrYu9tZ%K
z%m4zI_Q7!{r&3Reqx-%!;CwEpsfS@Fe;&4Xk+WLZADQ5S&Pte=6%Lr@l+6oiO>^Ty
zO3%1DvV7%DSd%FdHd#m&!7S(1)FbuV&ZJK^)l9v82vIP14VDB7s}2ugDNQ%W#<foq
z3KoJR#sp`&>4<P5=u#gs%#gFO{|*oq1cjR|a)~aVh}wX&r&#AO0IDR6!<$khwR?&W
z-$NCxAF@C0ymkr-rr#N^4fq|M$>s%C#BwFU8c=!NL)E&X5R{IB8tcq_(82Nrq#GSW
zaJ`qX0uFxHmx%!<E5I+Doy4Py{X1G~>gkMnMCJn;??4&5tj2!pda`;tXbZ63mr?GZ
zb6heF+PRB**G%#R_J)fXAciAv7V{Fr;tmV50zUfr;Gok|2hQ`{TG#eba|Hs~>^8_a
z$D!MHDMP0<pEUvpT+-kM(b^>5Mk6TQr_Y(zznpajxA{68E3M|)*lONhVd17zU6VDD
zX$NzdUr858dm#$L@iYn)N&twX2xc5U(MEGmJxv4I6_qCnQ2o=p*?JDmsvaKDNO{ko
zUAY!K%jH3^Dx!v|cY2<;?yG4wcgZ!v411!2lxvDYaL0_J4IO}@Nj;YiIE#rGI^8mL
z)lDztI@`06#|MSZFO*^Bq-g2%>|nPJ4Z@+W1Uj#*IMQ@wJ<#;r%mv7BmT%V$KMrP2
z``+;~R_So8T60di?r0AsGb@nOz)p_njNWa19S4{w7nmF%VabgKD;RQ%jDpE8xJ04G
zop*MDD9rD4NNNB!KNs()9I(!VRZdd_o=q+aoMlMZhNaS`_g8W!X0*^9Bw~rs(Lz<j
z!oi!3jP(=`d)kWzW&pAKu~?J1jWWpvtXvDy?5yXvw3-(Ec*wOjsa?c{pOf3lj(ac7
zqdTfeJU$r7XhW!3C`{SCVCeh-hC2~Gq@2)>+|fhjpQ9e-DAiUH&0!<>$P&V6EsrO;
ze%>6)TCXBp-F8>O#65BS(;%<wp{B+hVBi1|)y3<qf`;nNQUDnF2KtkWZ8SAn^rpQ8
z4u!$kFu%n9FLTljwZ65HfIBEHIyu1{Z%<iwqGIxby2+2NMi555Xb!AsG$Wj(&2gm+
zE!4Yx4B2g9)$m3T4KaZ&h*<hu8bA`NPz;*Izs@1tu`D(f#Q|ECNB1`2!wcVX7Mn5f
z2@93)V47w?CxEfnQeHN$rQMdBhH=!t@=La#Q|^75>-T2IWi=|Bm%0lqN|d=zFFe>~
zpftG+@;Mk!`o^H*^RX>@(|&)xxU_g!oa01Im-tIpsAY&SCi??uDl}}0i@M%C?i%`q
z!Rk}g#Q|F8R@!Dkpqt00jo=2nVO5A1xeGE8dO~uX2w;F2W%1aKTLi1+iibLWsQDMC
zKuamdiX>;H^sH#-UE(jL#Jctt;woZwp&!W0QP7*SjBH-oZKF>MX`<b0OHbqO2@d!P
zNa)2!l^TuB=jE@cIlTzDgWP!x16#-RuemV03x9)@*aGm}0F)UMRB3rl?q;}W1blNF
zVrC)yV8C)%15No#9l2`Yp#i}}T&e@i08_n#E*)G(w9&h1WgUM=dkZ*cnPBDz=2FsA
z1?-~8ItbKq0~6!-@iCyc2d1AEgm;wcr6a%py42A;lCEm3y-}eLF4E=P$J>oA3ZEbJ
zU)jR0e`zj=DRg}+wJV?6a0-@M%i!<RjR&oN2|!}x2a;Qn3*YuOexb;=bcY?f3N5on
zVybUqT<kV-<uB|C68AK36FtQWx|-G|9@1O&UU)c5MVD?$P%_w?O3*$Wl?HB7&p_o(
z4@70anF{E5spt<!?Mmu?B|aN>jxb-%=8531IjKt%SWB@5E9zlzPGrcXQp_i?rxIaj
zIY6)5X(@aNGHmN0!~hW|R3-VJ(2c?>CzD=T*KSzr#leD#d%axfEne+jS#1oYaNDeK
z64VtEQci(&^r~HTs0C6U8+=CSoQI22=t{*TJCPh3vRBGk3TE6=X0=>3|FUd9+K6*t
zFgURpsIvGveeO6Q9{SXK2u%|E=@Bl-BtVolVNPdE+V+d?1Z(EgA|esGUPRA7_pq9g
zUEFM(>#?u;okE&<L8VS9%>SA);rt0dx_x+fgtt#u-k6b+KYN7UlB!*ANH4&~#JXyB
zlGMm9kA5*+*P`?X2PEzlif+YJSXhIKjT}fExSKXcryD=u?>mVWTG=_%35SBqAfRv|
ztUR%wC_@5$Z4gvds#JlVj<$wn^iAfNtWZ<#&YtCf!r2<M|M)3n3^YrU-xW1KL*7o9
zK7nN^&?C!XpEId8!fIk!O-d+~an4;sk<Kk&x`?K3LTV|NVBT@`MQ{Dtwq0nMiSt!m
z^A#hr`q7aP%`%ZM)R0MfPEIGE9uOMz>ktP13|H+)P$9$Y${!D^`E`lT{RWl|0bdPQ
z05VsC5cA?iW%LSd?eK%C5y^@bz9`F@?~|#n;*;km3`DZAIY>~l$C(&dnD#LXoHh|b
z?rSBVb|aK*yuP_loHrkq8mD~xDX-kWR?l`GnOQF6Yx++(%ve0{$<(p`6QKN9#4)wr
zhJdpr9*U{+@cE@|RpWMmU3(^J+dgu#ztE0$Qn@o#R?cf<z_%!u{EWVriJXT;u~e%O
zL#abnhNSzp&-4i*9xTjOV+SiQorZA=)Dw6?+ZR$pD5Seo2wSJ>>_xsdBbFsB#Vfhx
zT+I{p$hkv&&?pP}tA}|7A0Wd@H)>6@xe)BAxNbW+P&&}o;`S9taC0R|)Vy5Rh$t|{
zD~)$;8(&*{_Zu$Q>bGk|uBF1U6xX!8Ye)a{;O3vn_(&NN;*-3In8Po&UM5V|yNuDP
zyLD$DRt?U&mRTtaLs?U5^1~|<Nav;s)<)XcmHcTZ(X4xjfjXKPUchhnV28UL2`rPk
zf-){zx*T}aKau_p++>!PHep-acfhZ7M=}0U{7oaLQBwJS>~=ukpyrMr_uqVM^SM=b
z$U>5LB$WW$&476x#Zr7eB~y6X!LwEZ1f(lS4dYt0w|sD7ypehTfssJG6{$h_m=H-y
z5=Z+DzpMs`lo+$QAwa$e<BtdB*}7PgkT^vdG6G(C7t32m!bZ9Y1X8GdMLnHhA<geK
zR$FV-PVsHwMsQifPK?zsJ2aLhad8mA^m^g^!?+Y`16@u`#`IwpO-7<Y1oD5~R$il@
zF7G%$wjL*o&iGcukb5V{<V$1AYY-8yCVDS{i6sg%1(=_KKENhN0fPb-TDyLDUdy--
z?dMuDgf4`--?^_ykeoeT`F$S>60kqh_M1SD53YWYJ8s`S>q(Y`llAd-&p45`MpXEf
zVGW*P@gep{S*Voz%W~I6rWRW0Drm<+UB%TY(`XIki?oKA-2s4<P7hh>*qz;9M*T+=
zXNF&Bz%7pt53pwm;q5TN^_#L`sFEQ88xj1u_r9hsWWbsh*u6QozyA70e=FM22=7TI
zvV{RNVr*$NlvTF1alH}2k=WUaNUIs%>Zpc-7M+L0`E%|dC%{huvc9%_*pdxz!jjDc
zznIle$kqPgb(_DOh*4(a<drL&8HaURLm(*fOH=hF0>Dd0>7zjmWD<z%0aBf>*8~#h
zvGQgeU2sw~z!)hDJ7E~Y$eGkSAxS8l3`VcJ&LAfmsBf`}k@D=izpf`PrKe1zV>}5h
zRPZK)<%gCD7}6tNm}rF%D}p?{fDZ5I=vfbR4N_cjc5h$i&VDzd<Bc&a^IR=-qCg%S
zHjCBV09X|mmH<in+Obsf@ns=>1x}#)SqsUQH|_=)%@>Xq{UawjfR~w)exGAZxTG55
z8m6BJj%03L>h|k!p81|h)Fy-9j?gXO2v5}*KOOh4DW~TuV(F$nV;G7JLUf~fLqqPf
z`DT}5@Umgd)v32PqJ>F&&z(c@I!&r*w%u#i14^kS^mDm&mg)2>v$&%<J|LlS86mMd
z=0KQE?zfw+3gWcokx<MrZ&FLI+qV)(i-Cmg!cyyKRCsFG3s;bL0>ou*NfGwG0ly!J
zAn^crzfLh#ld}Qh;PoxC#}i+1?2kU&7L0ZJK9!Obf0RLX?rz9{hbpFH_Sh5=;e*{Q
z0ORz5SxG1Svl{E!vauX#WHcDuqBTVxBo}Xdms@OtqWcMZf<RWE$)ab8tZB$5QP$oD
z4Jutc#j$P^L((AvTQMnqrbQ^?M{(dSyfKMOA4JeY%2oRCq3}qlY%_MJF08^|7%rj?
zD|Zki^@bRTA!4n>7yYxGOIu!xUsy>zfDvn!Yb-SDsWXmEV~Zf5Yo~nexZQJOLIKY(
zbTmRsRKnmZ@U^6Eku>!lfsC4?1P6{W#tMb=34=(4s}wO2+1zPgATKuPP;AWpG6$5T
z9;GM#t3rQJAB3p&cEj$X4dn7BXs5vFx)OX1&}y*ew=X`e;lJaTtTDs3oS*e?04(b2
zX}&LR_?mXdQXA&Ui=$E;gl6aMjK|qET9ry@95-yYBJM2eXqGucWZx*^anERJN|$2V
zoUG=eFuRwCNVe9};YW2HVcojX2veY)?d{|18RaOTdZ#J(%cULe#tkG`2?`X^dH)T_
zJAMfwIaG4ar)1-i+JtY!y!suxv|=r23Q#uI|I*V_bab2#*;D;yYwtG4)G)%pp!FLR
zq-v9{ZM>4CI!ph!HLA%()rGK%QInyMzCU~0?a3+XJm!-)03{MR(C@aNZHCo;f9K5D
zN9Ieac|G2_6eVTJ9t&Yva3C&E|JY3Et-gF+tN)rWt+KW$pRVC!3p&rWD?%cND{Hs`
z&3yozvm<hZZhkzdhhz!ZkxrvEuWz3IvmyIcdCk@)w<H!3dCU0VH#dKL-}cG7*2Rq!
z&jK+UtPvSG)Q3+huG@?zwsUpDTA10`ALZ6ng;_)_+_zfb4G1?oVA|gOX>8klue5cK
z30wZ<BHkL_Nu2aW?K259;C7e!@0WSEsk;H5%$)%r-CvVW?puBd?SFFo-L%x2AlI&p
zXl#GlPjBW9*aClPxM?7?BJT!zgY(UfBs1!YW?4jC?zywd{spBn^BsUh+aGz6rHAZU
zDZMILXAi~g4dot6`~2$U%ytw*GYgnO@nGwISY?d%TnThf%pWVVij6(m$2SPWJW2zk
zVmj}OITDbqf3jifhg83Ps(k^C8n~gq?y)+`*o?z9$1tz8@EX|d8g4>xh=*l$48Jp!
z9@5)*9BPODgSXrN-juZe8>&jj{s+qC0^{|(MUjZigTC)I=TyfS9ZiY-CtYwQt6We5
znzfJ!jx7#=H(^&d&B921brOaDvQ-FR8&MJ!ZmFIgf@f1WlLb_U<?_&qc;J=Iyh3Zi
z@f@c<k@`5UFW35vN;tcGWDHi3>DQZtG0+iOAHJ*3(QfcHQ-`4(lP%LyID`zU_ubW9
zI(SG*i`N*?F-8SCj^6w;K34y}upum%9!bRCUo^4s^+qV5;XryA$QT3Rt6Ku<N)*bp
z`H=|{2SwB?&t$*)SI$J;)U4-NTxbQoswTHjuT4HE+@+KX8iuqGy=gce)CQBXc3$+q
zIMN|Busc~sF$f4);}^X@wixSMr)S<~Sc4U)YSviQqIGoE5zugRdHL5_?*h6)=CbTl
zO~N^03wd?uyO=Bm4j6;vZZ`mnEOxT#pgbgbym=Q5jer!&5a3y_)f_hr*=d?t+!v48
zoZFxW-tfnxX$I*)M*Lyqnabl~>XJZFhE;ARsq7&VWy2(dA!14HmE(%mIJQ4##cvtF
zcE;xZ*aL8Lfs3!XbdAB%utv}+Cu9WO*y&f{zLQ@&1hLKg>{3uMnI!=5d?p48e*)R%
za(;o&w?+k3wfNAYr>n`_JSg!Z>pquzPs~0#YOU)68)SNhQL%onZ!lvOZnZ2-8@@-~
zBP5&_u3QMagEUCq>)y>5!H(FnyJlMtMSi>Pj9I;PA`RjXG9qjdt8+<ViaaA9BLf(G
zQKO4_F9#aSZie42yx9Qfj5GGGgbXMg`JYItKzFm>?M}M6r%f*L=w9Q;n`G2upR8*T
zOLsRG%z>wu6RPoLMdT6=N2p|V8us1$-<yb-`oMX)p@%eJtyq2oP}+T(ra}Z|GQ7V=
zYdJ<+S_k3!8C_Ym)~I+2)W=YNnMj`T_dK}ztbim4<$224`w0M2rVaD-yGh1$)rX=9
z>zOzT^xk>p_g>MME+2Zg12=9m?rd<<+yBDl2dQ&q+V#0&K-*T@nE!JbIx&{73cwn3
zhQj6GZ73rVUVwL3z*Zc}ZR2_B!Iib1d1K#@r65gL@eGXyJJ76N_P*R80aoMaCRy!n
z7VxpEJ9w7i@tc5+0gHg+BL*YY7JnK<?}OlU8=gAx$LPywQ4GgEhNv;RBwk-jC1(#8
zZ?tj`V%K8k`}!gJx5CjHI+_jO{6YjFk^s&V88!c-nk)al%BlWJTHPwbkMhX-3EX90
zI=%%)r#y*;J}j}K@=u4K{zo~Y7*Ea(tGs;~A=e{~^lpHQ&%_@#*JBJnuPylto>dy>
z7$h&GkE~#12QFLFay*%>LYhFJi0~m~PW#fA`b)Mb572_&dAK4oW^f2ZHhQCS$mK1h
zk5Iw4@nSmI-F3r#zMhFb>utco49U4K>A9yyp84AC1k#)SR2epeclr=KU2owMPmg1<
z!xaEn2Y&z?HfyGOX>1mzgMlGXlaHnT@cM}+!kx{n{6e)SH&f39i%=>EiIJipS<{x^
zAQ`_kSzv1n5U3mns%VQ>E~T64*U7gFZ_BneaAa+CF44Xk5vy6NYO?!5Z-l#LSjpr>
zu8D*U!~>QDkmF~j#a8cQC3`%)2N>gD8&C%s5+}gK%gR(*<l^1@mUiPHev-k9TkDTn
z1>DgYM-vl&Rcu~@DScN$;W@FKl<EdN(S+qmA*N&_C}+CD+nXLN=Z<=Ci&1R0M&bla
zLa1)zb0`%?CU(9>%R9!<OG=hU3sA>L%YEU+zmSlRDnswq&*Rsqcb=bAkmj)k%!scm
z91Xy=jw_1_G9ia`tmp}+mfHP3_Y)<PP7&W+W6Ya7$=zp*vTCi~sal(~_7^zfs#E%o
zd<;GSqe$~R7itBm25+D3C*Lw`AB#9DGB5|ZuqCehTs#3<B<w<@UXWAA*mQp?Wi*zM
zt&$Nok;HdW+Ge{e1yLbet4b`m(2Xk`=pBH`h!7@SM@cm_*}q=|JjSFs$kcEAi2Aqa
zv+Y!;&r#hj*VD}cWT-!S3bw-7pKWZ=KS(iazX`ka!0|!Dq4>y8z{U2_hMPR_RrLmf
zz$z;WmMGsDLH79*sA<gyll%cX8Nr3%hemvN7*I$isag-FS_Cbdcr<|9|70=$RtIoz
z=>!j||08#c$bYO7FZzk$!wMss*yo5Jw+Hcj<GW)C=ih9j{biPt0KVrF1B@N-+Sh)7
zRG|pYtv|@9vTB<B*lsq>sShSvmlsM!LND0JEWsI8gx($%vJol4uy&Ho)Eq|qV)hl>
zK}-CMPq9(KwT32>a4w39`yxKJ+Xol{HvCKqyUCyw0+#cE_HAhl+7bxb-zKE>VDPgr
z)5xPn015n1Qcn?s1lEIyCBa8Hqc_wkP6O@#)rl;5dgd{!A1+$tfX=<0jZ;fi4aYh7
zBp>Yf(-mXf{eF@=PoP6ltP%Q(HWUPH_*>c|gBSybG*uRl=uQNwf!UA{@({4=f$2cR
zI%JH^g163nEJzK-q_lb?IzNz{mn)97@(jj!pxQ=?u9{Zhq2M-dF}#S>JiH}jT)6u0
zzT@CJ&dmT{mXGQ7lK-P`G|@)2^O}qET}+!%?;DCIF$uv;+Nc~TD2P=ZCY#?pNY_%M
zkpD48NYP>7Uqk{|_yhyJvIn?`TcmDuEnvkr)#?Ew9{WmOoWrYEx_L{^-6#I_-A-~f
z82r8DN|zBC0K^o(!;#Vt$?O+0?vwQRH{tHMD`9Nsy^O*LiTv+1gN6{d8Trjygtrvd
z6bN{$)S0-L&`{$TkK&)y+M(|?pb*D+k0LlN<UU&n20C>i5=^~_qyxZ$6}f}->RZ@2
zSQgOK48DNRTjaIXd;6%=K>G~g+DV^;b(|A24bT=Em#F)=u`w-O*N$I>LxKv_8$N=?
zy|v9_#yG9W8aGH^smM6jd1%uDw!X*jrbz3R7emN0p(NES!3lYI_O40{BgQ}4H~y4~
zo#;uReKmI(j2A$kN!9=a#kkZ>gd$puz6u)JNcrK_872Wdg~9Ssi;VFaAB@xPhA}ua
z$ULs^W=n@i#{FuZhC<mR2l;Y(F|(iB5*ne)e)#2GGH)K(#bKrsp$|6;u9~nzl%Rl>
z88@DYvjJXSlhUI0uN~tp|C`s3;vXgUV(AWGv==Owpzb9OPC5WB$(C%J6!6#t+AdfR
zyx|Sf+%p4gaTtx8<t?Vl%BDgqw(@W2s4t$H>(6orDJi6rb6)b&A}J=rFn2{7b4r3x
zj_}JUd%=yPFq60--MKR{bhfP$Moiww#H|ty2IXRM`iB_d04h9Li3za9gqh!PV^L6=
z39}^KyfJtsB<=tXw4}Wc(wbuU!gO#d<*75i#VXy^yUtW^H`~<=?R4-PS5L*XrhckE
z`_6Y6`T-w7{YKUtl36ugC%PV=Nalb9hPTh8>tXb~GH$~Ql~sbF7=gPvOgH^s*^l|A
z`A&DR<(&=ghkOQNyat{J2*}8Yy3$zw>*=0DKhNKXK@5O?6)<k>=Bw4uq5FU9y!Fck
z@j@}D{qoTx5`>X*zG&@O$*7YoMOsv<Y;?wa^e+m`Dm~&Ux@F{RSfP9Uy=CV1x1w8K
zeb#hfDDM>0W&B)NicTtd#LdOTh3E*PL51?&y4?a6tP$9Z#}sh}YT80D0!A1ff%V9-
zrCd20I1Zq<pkn=?gOW@^U6Cni-^LzikM9P~<C9Teg$Pm!oQN|Hc*Zz_Z13%fc<Gsz
zr*nu?df3pzq9AX+n9;Ys2bNPR1c>HfJfXy)Aeg1#*d^3u*mf|yQ1MOG89K7PGnLTs
z)b8YQ%9+=>Ro(B0x!VgvNRWrK0#fmD@7T?@eG{+<5t6WIPCQDVVf@ecF5_i=7hdCe
z#(ZBP{Ue1WG2R=_4?1;XYAoKDk@0RU2$5Q0Z3jY5&~1}BUK%oomaY?plY`%g6&LEx
zD7D`_udT2EoG4e=#HY1%5o=WvE!7_5@4{+Th15**B%C<bLHe<szzSC*c{Li4n{Uu4
zxDR-G+Es9Kpsw<d(YN^EA*HAx`9%{fZa`4njL<28+=<WFq)a48<y}iX1Jqx(nq$wO
zhU0NS@$`%H3buQOx8EID6E}&2*k>N52*-^_L>~0OHB#8{HkW~cS?Yj7)vSMCJT0LN
zB!+1$-JpS$%L7LR>5Wu_D+eM^UlIfza}i*aATfK1)=n45tLKRjcO-&JV5TA@CKQQH
zAG4Mf<_B3CS%oxr;>!OU`9+Gsw8SCz=R+OEXFl;nDD}_!`LGf2UHsm7aVKdF<Pej%
z@>UNK2smqP`HGM#S0vDW$Q&_ya14AJ1DY`!{$(-s*fTe}LYsO_!fCl?mPeJE?g${L
z#1xrU=^wwev+M|z|BOdU_)&gOlx8_moq*|=2X7|onyNAHJf`jrqjc>YmSegn^?)&<
zHRCoxA6+6|{-3T$aR!U5W1%Ybmyu;;<u-!G6KldKkjcbVgRoQuq5*tz9I5)L!O3jm
zkWd<utg<aNqyT13@$h~ow*$DCL=%A2kp!_(7%WyBA3AQxUneaS2>E-<jozQT`y2&~
z2BO$TH8XPFrS1p>yju9S{@6wbwhyN&l0~WfqUYSqP#H#)oCe6qgo)g(JR22kXH#Vp
zdHfW4MICAR&RIT{H3kr#HMP<#0UDl3h`-cP8!ZQhsMN(X%GeVRKt?v$z+eE7o6}Z0
z@-3_Zq<KlaKUJW1<%(iIBh$A$n{SU^0Qh737@+<<{QzR^Rg)@cNu)l(;8>%Nj|$$e
z@t`}KU+oE2zYi2@fWA06SogJJ5`8jPvfbb3bL$54f<@oReAe+*`lIot?T-IFyWTf`
zi<?VM5lF_1?7JP|K+h`w*cb>%?t~R9a`ckh9||@sqa5CVm75m@89IYv5w;JhFs|xF
zy~GHyiVyx{9MJgbj50c4B9%@14i&~J2t=CwTjXDc{FHz1MS_L}HNaZ($eJXe+SL-x
z)GnE(gpChDXS^dBX+#wMd>Zj=%Hh%PWZv)xl&X}+FE~Xi8m}3R#}Wo`t8ZFM8a|FS
zO96*^TGT$&BJ*T=`D>&`b*ZdFRoZBrW@?bL!t6Z+bGji_6VS{aP}ta)rG8q%8noV`
zprbgj*&q@oeA?3iLZ*X#qP*u5x=l3#j8+Q@+<|K2`AI&o)e#|s@w5tBD#SmGxFhI#
z-c)N$z$M^y^p}V3$FdEe%F==`WSc}e$JD3o;gTvwkC!INc_ZL>8lyP^NnQOD{b8j)
zvHC51^p9CAbl`QE$G?Cy4{tuNcv@nji*?pr@Vk@erhCi-z>Qa@yv~>r)N@`*OPeB_
z9##+C!@|4D753Mao(qJm31~6mzW(t88Xq8N@b8N0&?qh&s|o`EdrvY?;p*GAKlNQ8
z2_LXNBFO?X34xi<&n_DH_ln1Eg^*aY7HrW-V1nGQu-18VdR#3@55nb&lnm2lnF@=+
z4fnR=jKpA7FNze;WGg8y+0UqRqeeI7rB}o?13&^F<h2MmEoS=`<x|f-YzSWcQ|Cbs
zpJBU-lp53=zNZjC+`C8L$*JD%g*%*3><>4tTLuYIg0uHnLY^@}_Rx(dKD%7z_q!^S
z7zrVECo&^>T-Kjz8QHM>a9?(|P{l>8EzWn)<l{5>N_C6Cq;F#0+5Z85K!CsFX`p7Y
zdQW9bc4cKt_C=f9yCkBO77G}+!4{#&*OI1ul&iNcc*;cTRb4RPxqmS_6}}OdV!uF&
zg_jyKg4?u`>{zBWd|w)5sX+cr)Z>q1j>@MDx9(O_iK1KYCli{DCKTEQM+vfg{qHrL
zB}%v9iKGz#zLgvahY1H$f|X}YAh8IT*AJ)4%Q?&zn%ME!_FRKZ|KQuVRa>`=F92bI
zpU><ko+M1FXP-=j9e@98t8jrZY(9F?BSoqhuvJtvP!JqND}Tg4S#i>d6S_FaWhcdY
zLt+e%oEV^tf5(Um7v?S&NY3~5PhM{;&{=eaa!^^i@c&g#|598PEn&Kh^dbzO7a?$#
zZQt6R;dDe!8-Y~D<s29U0Fh*#8j1KEs}s?Kbf)SZ7pSi}eSeU_u2yr{mvJ4L?Z*lq
z+3V`kAEX2+<G<S|?wE{m4+<RNKu;wTJ&*NNLduol;aKNfNTMQf|Ea^q8}B4HU1FW#
zW*I`z;q*UDkIx>DMgPcXqE&h`g^jTK)$n!6cQMw{QV~HBIR?b0OT<M|^hWjoda|9U
zujn}4Di-K_`hTwh{)LG`0@T1iwQ`e=2^+9UVjZA^4zcMP)&NMD{HA~jKXIB|E&c~{
zR@dY*ib^F;lA!hOXTHjRAC%<A@Y+5Nq=uAKRA8e&(aTY=cUCG!q$EZq9#&Petcu_1
z9aAl?;vF03E+>!IPKjcVLw=Yn|J-m`s5aHJnIQxDLw}O>?g$1-kzG?sy944x1B-n-
z;}}qNl98x*P`H`uk6KDm5uc~Cj1_jCPQ$75DkT><xyZ4f_Lj~%l7uxTSzQ)|$(+N*
z11~G<61Om=hBJP2_mWbI)R^wUIU|Pl@hIlE_)w`jakry`!FFxgDsiczgoJ=ThJ>g+
zg@o>yB!63#=<%DmclkhAo3NrvqSPreQdLN&(#c3dOVF1Jdn6r@`x==?rEMJ4ipGSO
zx2&E((kx3>nH@+n{}B$>(HSnbh4wn^%(eeq4@hnC`CSGMt1kR?Q&Nh7e_0a8u822N
z4s(nDg_iL#lgX22<%u{tH%Hd}3rZp*>o47vwM`+17&xN?pId4iftkRaDM8jKsccCZ
zX(x+Xoxg*g(DHR@@GktjwsU`R;+kwtIl=mMbYVT=q%ctV+u<av-HjKOy<goweD(j)
z0C7H-e{KmU0yj37Q34t&e_LB~+qjZ`&#%xUowYc`mBj4BDcf-pCvj#@>^ZYrGgDi%
zL|fcYq=%O{&aXe+07y}0%a*fw+f+FOaYv*3>&6G_`0~Y1uT?TLq0wB=F0W@w&SDXT
zf}7dpa`s1Xd5}awdR1fxb8eD={p_BfvZgEA#=TruwM#VjWxLASe|+IymuZ_eS=;zF
zO}f7A`$-RY6nr^gD#+^eCi5Nr$4_&Ys7hb2Y149<)=S@EcEDBeB~RPDDt$+bx@vs$
zMb(vUR$m`5nDV0S(xM4n$#K@SzUQu8RO^1gw9XFxx_mc_W^)#WM)^gr@|#rwe{F^X
zYZ2t7dtOz=GIYtKe+Guio6p#jmQZ7vFVZ%1h3zVB`;^bR$?e?jl6`MqFefcKUn6ha
zb6wS$>lC{2Rh>W}@n@u>m+&9+`L!SB4i{aG>I{|SScp)%B}6}Diqo=Pr1vi8x~|rn
z6%aDIf~_rbJ1rI2A%C~3vH+>oIChAwUYuUOV9+$0F;p#`f5?zWe70D>`1?heC?jmy
zCS?oy20WC8&E}wu#$d@Ge?_w;l)zlV*u=9tTW>v6&}*b(NEfq<7ylgE&tb&*XvpJ<
zwsU4eneds8d1$0^EBgSeO6%g@t?VMZO>1WWEaD)c^U8{iaTqo@$?HV{+T@qInLA5F
zsA4hGI)=H5e}~iOu%dqYOva&*VwnD>6Cre%82NweL>j2c)c15U6_|!S(aB7!IAn=*
z@;a}US%CtHqTo$df<d>amQu?&_^ou-n76Czt|+s{it4A=Iv#b(5*0EjVOZPW|F|k!
z<1!r@q5B$V03>PG<X3r-xA#PkDhZ$NPW3hRnxtl+e`Ej8r?{=+N?ZHXVCqJ&hk7MV
zveQ7Pdc`7W0aMeGg`5jF=IP1FN#&QJMwJZ%L>Lwtg^}<e#5Hp%$kW97r3#hk+r4~A
zbS+rO1lx4YBckhMN7vvppVPNs)4oOY<g}mcnUS+*YM)upIpYb`R-BldK9~y_2L~XE
zrIq+>e=am=S){kUh1<%(Rg^w`@@wy<V!d-C-LzobUEAR+*g3|OOkY}0W+zU_oW-LN
zlq5PZn}xHT3y=0%@#*ooTQZ9U$H6r*A?C+}p^51*HyunGN`?3^+n6ZkYUn&Bq$$ME
zT2RF?wboCk|2n%kJ(w#U$9{s7T@u5vl9*M*e@2@aC&aWj^HRhfb4{6rNnqS~V{ww`
znCYkOVwcJBP1i<|JF!>S5U6li0*|+)da+n_5Jl?57_WpAsl}VU6qI5QilQl*#T$_l
zTg`cIHndg9&5m$Fu1a_~h+*jox*jR~!rq;p+cIA?=MZ&(x3FcSAQJQ`iS0V5ZHqSg
ze<UAWdSchc`eJFemrg3196HxTv5*@Z5xcgX^pz?BVa4jP7;JxEsbZKOIP7QnN=~Vf
z+QV1&o`FAw(FvhTn)~fwp(Z9EH=C&ysY<=P8KK93CsxxKM9YS&?{Pe|R7fw*G=PfI
zBZ&&qFn>sh5F1jK)4pR3f*nNwl-f#8e~gF?#RwsZ0fpLQKnnB0Fzv?3ScWlAzRN#K
zP3mi}E)?=aa_}XLbf9u${iK{WsG1P172Dt*W&Laq52+qs740Dh6Ow9|tRJcq2LZ(K
zNG;^92~*p<Q42)@{7!XZ-{BUml4z4W<^m-RJ%g8$@y?)im)MlrI0B+~3>f^Le^4S2
z@uR9B2^Z+y-1&!Jr!Fww?E)305gRWh>0?U2a=}T|9E`v`5=DX$5KB|hX61xV0iN!3
zfebh5d+3~P{^&V04^&VQIEPqINKMH_z<<Uk-;kP^X-Q2?)LB~09lcf;OILHmY+{w3
z*~IjO*(A`w@tleFv|ggd6dhzgf8;r#!oYJv@eZC7YI&X$d+ypfHMA!`1sD{PpEj{Y
zcVY(8o%r;0CniU9r>oJ^o%rVHPM0Fd!iVIijJL>7sU7(V*=rCx3o`fO`A<kWK|#GS
z|B3mv%ii#xkh#TwLg7fbP&x>gluu8Jf*L;{MIpzNqL6a5E+oE-1_c=Ce<_o{UjUG5
z?FvW9?;=8_Lc6AU(9xI#4=N2HC7VgWAszR;-p8|xOK`+OL?(E1esMyvoC+{fLpu(0
zXOD7=DhXq$D3(*<!9Llq69K2zJqSw47#xnp7C|W!Va5a{lPuYnpfmvRk1=a7b^Qa$
zzwPdR4Ee{ASWrKh{F4Bye_B7r{XNi0;y43gJ2-eatt&7?{`heQhK)BTdS-|KP+_Pv
zUA{sm%Ekv87DrJa7{sQNXiRW02QvBxy^8tJO>77P2dZc){#VM1Y&-*P;(#T{UOebS
z!b#?(CkHr0MpD!9B=nF#4++8La194%at%U)CJaL4FfU>PmK2ySe?p`8-Z~dr-k&9!
z2ts(W;bFIC5W&XveJCPE2k|HS@wX6NA$D;xZsZ1hF~|Lw1V8|E4C5Lef&VvDL@x1k
z?bHOSz?s`5#QN-k2^C2bBno%DmYpVSA_v6*9f%*$AiiWijt{qDK!6_>aKHpZZL*y?
zE&~8OOeS*_24J^ee<&z>G^qhn6&M<Wr3Q#1#BQi$CjtlsAuQ*Qq4xd|fcH;KJPQH<
zG$!Gzpy~e|5Wo!{PbYnVYu&*BY-2zo2oPdoLrLy~0WrZWqt)+^0kH_xL=YSam6!W~
zfYih(0Zu1UKLZ3nkkiwEP@bWA3<M-`Qe?CUp9X=10D(x_f1}_7@raPg5eQ6=$6!6u
zY(O9(4v`>89StDB_;eZO2@r6^N7-a+Jd!*E5?D+SyRazBX*x&{#K{yE#rtI=a8J1#
zAsmw6u4?L{K+4BxPJvsS$RC0U+$69_^&Ct9-!;`HP5ZMQn4k<JPW&(>F87EE0Fx#{
z2k7PT^AI7Lf8rWbdn1C-Dom!@h+w}LAQ*xS#Hc_$2M}Oi6wwJ+PXhwQz3J#jMFj#v
zH$4sAVi(W#fDlcrOL9ci9uOFK_=s2^Qv)DOZzqtHoC3lFVrUBi%q|2FV7<{4IJHBh
zAt(^UDp)cE1ql&>DdB=b+(r;^LLsBr&GK-KXsUh;e*&n;UgJFj0R$t+LV?$%dm{jT
zHZzT;Fl;n?AOPT}CJ;c2XCZ*$)0n{CR_}=b;J;%@98VIn7Xnb;MI`7)_eKC7k&%r+
zaC`DxFuj`Nc<=Y45I{-dt_lJ-Yf#oQ5=Gp=NWAsiganF+E8zHuclZ_oVge!f2*;-(
z0Ybp{e})7?Q<p}ddG`9)0s_SAF$D-}ST@LD`sIL<?(z+H71TSL45Fp?urwf`fxxCc
zpS{}V%bsGj)C`#)g**akG;APfG>-wN6Hopo_#q!G;-o4UMb{w1N?b=I%(k2Z%xTX_
zT&^rhX-weR;Lith0UZ_DhTI%JW;H>bEwZw8f9Z7T=+1Is$yaD(zbJ(gq@xa465QO+
zkxE>;(si{Y*~zRZ^701iC=H!0U1Cv{*Hyhv3w%|<ZJjOig=3DfIQLIWS^D~IhO_<J
zx{T`DJx{;h7HM9RBl}N)H4BdFHvcn<_#(#<0qBD>=np1#lfm*rSIKH-=89J%O5By+
zf7aD))wqvqS&+JFr4{3y%Nu=*%7V5cIc3M>;i8~5!HK&9&rL(h@!gPEr(bgr7k=T1
zuVZq}x2f>Sc7FO=q>In<tLiH%K2p|}Uc@9efMr^`IDtObKD%;HCv_|2(B=Hn)h;m-
z3AA);qY3J&K9Esx=qWqC{dpf#1H(7=f0W7g9di^7XnBYVXLn@T4a(`MLELdHoXGZo
z1y;OJmS4N4(?)1pmBoF3?F-X+aG__bOYW5s5+hj(DSBCAN!~VD;j6FG#wCg@txIbq
z@Paopd1F5}$F_rX%O|LP>AzT4vgh%oz0o<uqzLX-`C{dA?|kV46~_A68hUF{e*u?$
zS9R6h_=a_Uvpolo*cH^lIe;Xy^=)3~3u{CS1(+81O>S>;pwxp6q?%RL;k^@9%1UpC
zkY8sF2xGCz>Zb2uNiF-fhCR!1@vbJQm(#~2=o-5kLt78lurLMBs$ORGcH?E<EV8<E
z9kg}V&=*b~E8Eq9vkT7LCO6iXfAm)>klU4YR^)zt92F_E!hoXRSGmh&e#4-+SyZ<^
z5An;D<;^2guyAdHB46ii<I+RI8C;QnOFbd3l`{=vbpW4NFuM`byG)BNge4jdu?2h5
z6p@Y_OtHG2`@oKC2tDQYBChT3p@pUH(Nsn5ZO&dofl=CfD9o_>Z4+JFf7L#7+iO|(
z5KYi8dI+VoDRowDRWvp_YhU3I){XqI;3kY9c##+a_n(a$?}v{TpUVozx+}8f%^G}W
zLpfwE?B&oLx{&@Xa8a$TLy4URXx;zX0x%f@C#UYC(#zl6`zyee)+O&U_gf1c+@|2y
zoKwJp>#kfn2Nefb_mm9Ie@>3hpj_y{J!2FlUO}>;cKNS!kfVhXc=_bb_0YZAi%n>A
z_VJC&_><#DX_&Ipgl9iZsI{nyBE71}y{N6=Nx2(12F3Vwff&x0pR#KgaAY<XAq3qw
zLc+rMnCYl!*UqzL-uujX5(mE?z_}*~{}bPm;HbRGiqsdq>2r^}f9Bi$vdBOCf}_6R
zRdtmwzGZ#>`##@CO4L$cWnJB5CAI+Db3H-A`W}SD4>{f)uMgeZN$84s1nb9WoIymr
ztiaR<3_ize_2qq;u1V463p6>#Tnq*EkBXbB23uKMp92M-GO!ky6ZEeD_RmTR3<NU7
z!)@N{1F3?%MvceIe|qfY?DX{14FEwY?hkA^?vE>V0V;mq=K9$_(PWFx)%TayRu+)B
zJb+jbQTYzR*I!`6dT|$5!-e$+^rsk75<eD*ng0yK4ix%H_}n=j68eue?ao8|(iK&M
zsL64g@u`E4xJj7d0t84LgUjq|UHDqCMGTTU0%LL|ZvY2ufBXSyh%f7kbf*xr^gjhN
z&N0FBWD^hL>#w7D*v8^kLXda`u|u6-p=a-~c>cQ1JXHQh2f^U)T8Q0RFU@~;Y3XC(
zjSbp1W7*UY9Jhz=%TuSa(DYx{a?-!MtS^EbZi4M{@Ho{?(fJymY#}7x=H+r@hfw=9
zM%#Vqah;c+f9tBa_J!}DfLIG0g>!qFz-pQY=e9j0oO~ehHeIhb!$8Jy=$`+8f9`XS
z)BY<CSU;Xiu=G>qMB23dP4&Z}5s#&muu%`A4$K$s95VQmxrq&Ea}2&_u_|4_gstoB
z7Bu_Ny`=Sj`A<y1afj5PM}3W-!SLby%0>TC^B;@me=q-4$Ng>gv$rGg28$vixAYh=
zpsCQ1zVvYd0P1D~@$EyFt<$o!Wp?^<D7wi06e2DL(PG~FzVL0@-T1tBF3%v$D*rV~
zRxG~Ve?FuZ=r*uF7-Hgm+LYM>G4ar4yb0aYo12IMg|qC|Wqt&3bzU9QZ67T(4?6YV
zPLcM}e-&x%#P<rqX)nnm|2w*__JNZ<AZ=~aNVc!4^;KQjlVI~X$K3^$Itu4jK$w1$
zK?H;0bLk7;h3<LU*RWH?#Cz-pbTBJ?RqwCxMR*|)pz;b1DbVmbtJYP0yF#d!(Y97%
z4xA29v=G?ZmyZ^5@^97-u>R}9F9XI)6r+8^e<Y?j|A2e0-~es775}Qb#T@MdHYkC9
zKySVP6!{Yxb9-l%q;Pd`|FH3)FZOezot!mXM3JJpYYrc{5nOe?$Xo9&awq#DUsaW_
zaqgmd>epr_Og5Zo??8O+_O2twNigyM9*hXz{7xDeqlN>T4W2OnEObu?L@a{c#P3<$
ze>}0kf8}E_zwd}x`kNu@_jCz&#Ejq=Dh$NTu_tDDzgYD4iFyc1*s2F5ib2_iC`&YK
zK#3Uyr9>xY042W_O`EpQ!>2a@?1wIP@EFpc&mr+Ph5c68z8oPC3tr)rSP$+Bf(QI<
ztFvfZ4i-HQhQONO0Mgo}lkO#V5zLl0f4!}Gjz>lQAQ|C|mW8i*)_TdX1iz&d_ywI`
zF(@*`$QpkWOD-Drd*wiDW$hip3Ts4GnIMXAWU2LcSoS`gpSU6&yh|6=l?yEGYdC+~
zw|dq2uNKp8c}Sv#m9Ve?P<7uBvheNPWeYEO$hm;CM)P?<KVR>C=$?Q^5k^2fe?c3-
zRUk8jC5s|$8XIL~K8HT#Hrb-iT6=ymu{&Rwc*HNdcTY0nkKcRV%EWz46bJWJ=Mn?!
z0O_{Q?9bUDnfss77aj;m_g?N$O9)JwZ*BU&Hh_oDK3I=sbnweP9K5PN`_yHRX62)$
zy863IyRV0?k#%NJ2$JJykT9L!K^@6Jog8dz$iQR;M8#Rw+cY-=fZ6hk7Hr2Qf9cv=
z2u=VxT{Z71p=1b}?OR12RlJ+Dk<gY`{6|lb{~Lq!Q;(NLf(a)BH#akvv6&kvf3;dm
zbK5u)zR$1lX^X0gI3&QQ_b??}iR0LwWG!cAcW3qhQ;3C_B2|1OBmesCMgtTnLsL74
z^}z%hZ}jWyhAtj{`0-two#lR>u;lFF%UK+qr7ZIUo}N9d&wlf+dbO^*PQ97Ud6awi
zK7D_Jzu7!VQ}43wt4=mwW?|rof1>L}Nj>B&jJ-d0rKm)w>Z&EGsMb{2)H4=)->K5o
zv;RDNI1A3^Vc_R+PD}qD1Pr^gIPkiS+@4dhRIr@r`bJV6zA)1_CC!9L6L#FAtcuN6
zG|#Gf@`puYpM_?T4|Ti2!^N=umH&4t+{}3F)p(-4{92pVkGA!fSrT~af2-<X@~@()
zt7p^W*ICZIVGgx5<O?E&dRLBfUTt6x&^O8g?}Lxu&|^WELop2U)O(mggSjre<YwJc
z=X)ti)pn|&>UFhNbQNpO1)S$z7IojW(zUMP1@jgKa!5Ru66@T1YuPpnVcmJ?5%-v<
zD*9fOqIjNP)PGQQDGSxYe+)GGPCU<D%K-<7C#!wi$!$Af2cteW;_0cBaBxO<5aM;j
zxpzeeLy?lycxS$=dTUCTja;kl4OOnUyQaqO<1x5H{DLr<!JX)wS5tq1v!n_2#KNNK
zsGzR^zO)*ZS4G|E(_!UtPv1}20WOy4sn@|}ZtyatZLc52qXKE(f4g?@un~IpOrP+v
zUh8Rdkg*7auPmtzI3*heGA^n9PF12Rl-5D$yaefJWZP0BZ9UhAa;#X9#&*OcEp%Gm
zY9jm#%7G_E(>nF&Ml@=RiWH@RxA5ohQRuOJAsSc^8xr87rLKyNo+aeW57Y20@Po*z
zJ*=kN=7FDahM%FWe=eW(mUdsvN*8uQEzDT>Wfr2Um9l{g7pB`a?Rg&Z(C0j#9-IX%
zKsn>EaX4$ryj9;Axq*LsqM=COcTqk8x4X?2)eZ?(#m=d^l-h17ZS>v{?H$fjH!~J^
zPSx1<r5O67hSX(0t`Fu6gr?oeqC25Hh6RlCfdyE$mz^elf1E{l*YWAAjvY6RlU&%@
zXfvFRZd#Z=*FJsWMp1J)<HL2)x#wx@r!hM^J>Ddb)A3F6IPu;!^_Ci(J_dTco%$E(
z37$lMud)Bh`zGdAdK~OC3D9_A#~<G%Z0roVpM)ul)}uZ(vk2H&NKowR2^R#Lq%8$}
z$x=W8&rYSqe+(q*@{TH(b@c^y#8b~`D~_$`Qp_@wExwzju_shhf%+OLSQG``gEbm)
z>>Z280}>lNoT%EJu*>b~XcIplrrhU1F=sO~@tFV4Si%!#6~=U%>N*^Mi#!kW&|9LL
z%0V=>G@y_mRc{Fk7u1hD8xjaPFdlk~wLqVa#0B-af0Ny&HUqzazUacLtBZ|jos87J
zk42&Pao2(Dzi3g=5G91m0L7QZrcwpU3-)T{u5Pr80IwyQe;Tin#GV<mu(b0q3Ynke
zS>Rgsx1#-z!6o~hG`IxpoGQ?Hz1GZI@!W#Js6-?7Ag>jIso1C#j~Pgo>Rx1N$O|GH
zO~dw!f3mOEva}s8efolfOnle=R%*LLXU=k~u6YDjK$PU;k!;=5#IFMVGe~FC2~~A_
zBD@8FvDNoV(|m6K+`BKO+)?Rm)ygd_=rmTh^0LQ!V6Z;87M3FmuoC8ZhUW`n@PU$O
zcRydR9<UgBO92WXTGOD7GJM~|DZj%dcCsSEe^NcUx7oSf*RARit!W)fBLZ<y`G?ZX
z!s5Rz?LsyNa-w~yW!IV>A8ZeO<yZj2D?kguW1B8_K7FmsVR`LjA*o3Oz~SoS)z$4=
z6tWB|vO@5C1cy@Zu15JFCS(aI57M<87mv78@G*vncSxm2C-%4nFJBD|y_e*iv=ZT=
zf4ZhG5Pmz^3WEBGt~c|Z-q2oFzS*L7AV46Agu<<qL!bGz=>|fbqBlm+MU9^;g)UXD
zsl#Fi&`ubac5Dw2uI9%4?;1()3n}`7sAxr>j6aJKxA707#825t<Bv0;_~BO@e=P8@
z?Tq21(UaZjT~&w6ME7vy78odk)--7Qe_8fq%T3?w3fMZRA`cFGYFj*~%McLMcG!Oa
zLU9xxnSZ$S4DCONU1}VF={M%_o!E%4MlS&Gkkm6HYL<Ly$wnCfg=|Ely#`VjK7C2d
z0R?Kdw6gg&RoiT2T{wb#(YM<BV;LCOHE5~;>9s`D7+1GB5ChPU5JN_O#&L6Ze;{EB
zAn`l~S8-5?7y_KjQox04kwh>&2ji6-dm9`zcpDsa<ZT=)j(krJz9)b?Nr^4sZVZE3
z2w47-=eZMA{R{^&<9`_}EmY|4d4jMP0@v@drC2H^u|!!bE$p-!<6NUItHH_vSTO!6
zHNg6{W`S^pVPs{`=&II%&IF+_e;62m!cR=VSqCOKzzh%kc))3}dKyGvipt&yF$+>Z
z0pvTm@1owiPt$EwDN1dbDB*$M8?s*cL5#7ouvBBy@{tx`?b3{982FhmBa2B2RduYi
z*^%gBG<^^~9z|2pqYmAssr#o5RhA>ifKfjUn&^TXN0EZ?se&LhW+$L=e-1a!lhe=`
zrqa-un{b;$z<}~Vq$9JZnPg_1#^DtS;Qt*BXR)7!?8qD@L5kKe39T`--6rzVH@nRY
z!^A{yA%gX8eEJ$pOckweuP-vo{U|ziGn5%HsT~-*9V-J8N+TBMR9&o1d}U^vMmqxT
z*yhpdohn@ZAddVPV!wk5f1zEtyM7o9Kx=^0-1HpZ!6$vm`0u+ci<dnn#5jO8?ea0w
zxljNibbI@(Oyi0gBg9e2BI>!86;a>Ip5(YR{Xf{r?p3kD=uR6PSL(zfIFL6Z4N$ld
zZ6ymG9z*R46*Isb6KbZ8RvAdkrD#fVPF)uC`Lu_aq@wbzdTd0qf2Rs!4g6kUiqRd`
zi{=^ng#0LJ0E8hgr)#uHD7f4f(nnh9vKYhkPN?nVPG|rb4->?LoiN1V^02{hW=>3j
z)+tk<2QdX&4%dq{%=4VeII<|mfan>5p|8D;#WL@`Pjt9XW}L^6A44q=Wx1yw_cejp
zwlno1XW&@xXMugcf2k+XPeE9aB{u(;V2H@ULWs4TgRO$2erb`8S>x`z)S*OzwyNz~
zD$@R;aM(qq(`!gmmD8(FIIrO}ZP?`)^T_AXpo9n+zR@7#9(N6b9pm$%@tm=M`#}P*
z&FK}HA7=cZ)3`mJ`8nXygpYFkpD5L>#QH@u5SF^veC3#Zf22H6X&P>Z22pi&JHM~j
zje6?k%kjXaBm$`Z8=_dzatJU*x<Tr_E5Z4_q4Hood63y89YU#hIL`vr{}RI@Jxq#L
z5p;msFxw=t4U?dHx&7a(2aNcYCrqLfaYD%bjQxo#j7*j&vRNW1EEbFFA8mq&e2bZ8
zvaeU+0dU^>f5gSLaZ_&ruoSOSg{!(yvU57S7*@!RogE+ro%)Fb>+DQ7K*>G_G18sy
z)uU7mK8ZwdpBT>KcQ6`QS;4V4ro+k-DLo5?t6l<V*MJnOmu3b^Wz4fA@n9FxujEl(
z5zVIVOPmX~tms2ptSv%BiE0$s{wJJ>NF%(Z>zWPGe+1aJL_kJGap)~7O8q)^Fvehr
zFc5Asej<UG`DwsUi$=mpG(fgUG=RYLqZg762}B^TuCH#EODbO*DW048k7fA~WAf0E
zdJH2iMA&vy)!$3GenR0lO-t1l+8}(B=<2w}PirdZ%z+ECsuReZ!3JN68i;y~vIapw
zWDz>`f1AkKs@U|DtOFRV{W%nF&`z7YJOM7GexzpR-4#T*YDBj&jrVpeT5|Cy%d$S6
zbtS;XNKy4pKtXvr2vM&DfsxFntPEXt^cAH~%_ad30CLFKXb8_nsR?72SZ2Nc$#nG;
z4!|j=Q7mRQFZR@Oz1@|nP#sZ`EnC&vltE-(e=%@|ox;ojIHi+rnkV4Ar0BqX6;wUs
z1{EgtVDI?}lgMH}$<s-MH_m$`uMHc}Y&zVIK8lVzt#kCag;IpGmZc)m9jxQQ{z0g$
zv4ackhEd?B$+1|baiYaCO(GK_f4aN9x8{(Bkc1j!wc6EnB@dzRD0_&{+Dq9`j!-{D
zf2zXtS=#p7x_UCS50<vtNC;Kd$K=b6tVXnxuelrpXAO5=USvhyJkdFNY_ej>Pdupr
z{U|{KloJV(Mf0dSgN8M%!9zY_oS>|S(mA&SB%SDX*LN>ENRH_~SO%8ooB9gTx7eD7
z8-LOPEScOsDT>S^okBw50q$q9jl*F~f87_dGZQU*m(j9d69C^d<EbALte738LXMI@
zj-!3viM4iL!dfWK0vqiF!w(L4m>-D5JSdETG1d_nV^71m7|!kbIe>#d6QDd+jC)xs
z^A3P0;!pAEtFyHoW$p9WJP6{$!cKt4AJhQbC@+%7ht~&6up{Y^?ayQTJBK*ne-p5V
zP50873LN4lxC&Jfe{Uf!rsVcGMN-(;Abfc~k>UlN&hwN08Ryd$GmGty1~8psy{{cU
zHP#ceh7ZS;eFUE7Fz_wenZrnJ)S9xtlEYvjFe;A5Mhzwi12k#I$(kOM{o?&QI8Bec
zl2Tnfu(BH9j|vJM5`g^Im1#6#f1nj)pu7i`w5IjOr?18ETC4uYoD{S0?cnZ!Lhl=B
zVr;GqE4PQFh!*X^GL3&9G)}*cZTYV{aW~2==%7xY8biW6d|WC4Kdd{`MDl-#3Sp(b
zZRwRx$BjNw&4UKd<G<$U`Kyy7<vZwV-6P(J@zqnB+@J-uSkS-DF~$0ie{<@(=-roG
zZ$&59lwq~M<1ugP=J_$RGXr>Ee!BOadX)KWP^0$(+#cFw8ua0oru!RJA28P}DV^mz
z#5fK8i1CpW;~*ZMTM^FvG(1AQSrj^DShJL##q>|HOv(cb^eRF54(PKn?SpNK!_2--
zFu=qYVlP{TIV$WmdJceYSl^Tr5}Rl1&e~N#q-!N@_<jo})5xW9mdkx5zzDY|PNIMH
zLkfp!5AFP-jPh8~pjvp(^lA!im-I{ut)wnMI-G-Pe4?Lf5s8=+>BWC|`|!hm0U2&v
zzL!vt2`2+LHZzy8nHwm7?K|6!+%}f){)*a%1*$RK9EsGmgL#NMb`Wp8vlDxe0g^1B
zs$zF3QAstDD(%kK&$&^zV%zN<J9$|&?4roS+u`B4^ZDCne|{Nft6b%YPF8PkS7D@8
zn6FY3D5LY$+uiDi_1lXqSQnehUR>!oSetA9eewJ?7xowN`r&AQ+hft%gY9k0)!Thj
zZf(aOw~f>`Cn?yAFkZidZsoz&eWxyddi(0xi?`4I`7B%o@E5Lhkg7B^t2BsJs>9Xx
z@a%`5g4GThU#$X_=IQF5Ko6@hPgI;5C|9dD&;Iu8xoj*|c^ZVcF_==p7)c(hI0<=U
zUlo1P+5X}xjDui*{pRa$zv896>Fu$T+UG^T-Ai@(o5fZB`6bLWjvPitg>epZ)1W^F
zI(~dFNgk*qNSAww^6X>w5*n4{u$e4Sah~wJ-<Gv4+Uh~pYT-2a%;H#QYZF{tMOhH9
z<3KhRs4&2Fu1uzMwz@(ZFg(=jrm4GS$MCG<6|CDxPt84lZ_0LCS=U!%UqgMz&OykX
zH5^AWpZ34MdGpqFjnAi*p3pIvulMWMZ*D%g+(*vC#43sO$J1843PY8{oOB>5fZ%}y
zkqRAkeR-kd_3>!y9bE$=oY`Hw(ppnDSg1-Pxoms9W?=?5rm7ko;=Zi!uqs^_<7fnO
zY#Y2pzm_|H=~L`>WnVURQE`Wz?TfM!=zd@HGEOV-tpx&sDSpNOin<rThQWHblw$$c
z^dgSeiod^{ts{k<^6d4F3(vV)Zg^FcwFAJK2brblev$sLjH0(YI^6)U5ddO^_t08k
zaW86E0ra|ixOxtEcE`1^o1LvH#*a>90Z{)xt%c=(^(F5O=CoY@_j+qvKuO&{aK+zF
zu<P>2Akf<a5y!RtzHLr-d;Yj-YG7UmLMhDQhHmOhZf|XOw6nS34knB%s1jVglfis@
zoE#${#27t0NBr*Xc3+oU{xrN#tS=A8$|C!e#be!Y<}}czIrZF$7%$erbjOBJh+lUW
zFkAG0z@DREeaTb`w|S&e6Q?Rk9d~<ib92*pXxW4+2qFh{Da)(3)oDk^gP*OvwXLnU
zyh}r8i}koNVhJY8V5(@2od)oC)d9?wPCd=m+lJ^JN*f>r#YLpSG>0}sNiXe;S2~pT
z*OEy6URrMDm&6ei=HiG>8)`v1oQ^y)<MkqcTr;^r&CnNFtrL8>J{EkJ{MENWZ_<aZ
zEbQadGDBEW0v~K-U=BEcI&S#bRK@m}t4;F`Kcn>liJn_vxkQ_QQ?A$+we*pS+vYTm
z$^A!)#M-YE|Gq4N+L&@Lxn>Ucf;+eN*tC7eb<lweh)wz>8iT?aHa(jK9Gt3N^aEml
z5%7CxT2V6?-%4l=*e!Goa*}@mvQCvb_76u70wxe15lndmj5G3xbE*wytoKdHc+7+K
zJ^p4n{I*h4H$4|vsA7re7l7d20u#A?kaox9@PqMCc0FGL0G^ODgLKVpe~-&9+*G71
zVd0$V;Rq9^N2?&VPK(A-no}ik8yOCNxSNr6;L<%OVQXf0Cepkqq9mf{u}Y7pTU)#9
zF1eC01;6{^LI(r0MW~&futeDQjsXikhb)<#g9cfef(j(;<JGAu3m=m)N{>)SxD~~n
zOn=)p2cEXuwy3fxt|xJdg$cBQ#dch~v9^}YP1uZFfDP^l*e9ziQLF;cwHR@K*(jY%
z7OiFP@Ly9B_G4m1u%@z;laE-OVqw>u;2^FxaE1chGO7j+g;2yPIzlr*+)H2Jp)JW%
z{E$sU^C&`w>G&pLsbm2rA_Aa^U}!~U$EB_Z+$e7SMn-pRc|k0TW5xT8Y&_*9nq)}{
z(6YB&Ribu^AiK7}=KGr4S#%|Tf+3SGuMIR+WshPC80R&Ee0Aamk%cK-i+0DlI?Z5#
z4$6&;x3OK%^~bgVSuG+cm8%BzCpjmmBfAFcu+f76k(hJkB9ECb<GuWXzrO~cJ>SWF
z#Z_o);BQeWF8mn8frdZ$4Y4L#C`|yeZ=BeI54O3WK#U#>8=3vY3=c|wq%z@eJB+cd
zimoFy$F?0;{oH^SMAiaK%GMe7*~@Ud67hEcp)lL%qcSVWy?8N+D0+%vzDHBia(O`~
zP{VtmK6|O^K(KOYyuk3m=ax*RD^?F{0g`7`<C4GKRd$27GEU(NUE}5foeCrzDZ>c7
zD{pZ_gXZGFz{d9X#79tn+%N{Qz(+5(FtG2=0gxIV3&Y0(L$<xxnfvl?U%@}7eyV}l
zM@cxBUm4iOgQ1=RhhT)@HPL$g3ek&4jh|*BBDtAliO)7+qjpf}$OX`FAE7H8jwJ$4
zvz#JZPC;d25dgDkn_^c9z-Z`jBcDAqfVl&3bfYy&pB^q35r`IlWd>^Jmy9^ITl!qY
z(dtsZd03d<JgQ1iR&<tI55uCxTEYTtoEMNX>xWLwv4?B&n}8JjLmrV$TMBHm`PnY4
z5gLLeHw*&1iJSnb7^IsG2>iL;6|IanUT0^u;2>4gNhg#&O^uA14pyfNl9}M;+u`gJ
zm8W3M14yD$u&+aZ73HEBZL3=uDVmIwsw_@c6EH-nXGWv4G@5uvv}ddcRy2VjW(eBy
zUl`N`Q_4)?bkh+w2t4EEDhdXVOx4_ZmK8=S)&tB@Ad(3lK;VaFhKNU39|x%)6H&gv
zMC@jK@a!#7Iv4<c_KcuanhXliJI_)Q6-4O_k>__iQ6`ChCoq9kcS`tF3q*7Ulmj**
zUMle>A!*5}Am0*Bc5hMtpVk!o<CTyyjLh>M3e|C>l0dr!K8ZqMMh`N8(BmZke-iyj
z5{>>{0-MpfVy+w+SVDsObEx7tnTr)|@<Au$FOd63awKYCjHKtDqbCnlVA2nvM?V!k
z+3$&-sjCEk_luGde_KmBS>#N|Z3UX?%(X;d<f(#R);0|KJ?e?;54sZqS%HA&QIry@
zMLFK_XDNd%6;@nCFa1ejg~Ci>WpS$`Wx{waR#>3~ZDiAsRs*G}A0!f2T^fy=+BJU@
zBH$OLTZ>w|PUDd_&QB!Fz$hGfqK~V>O$)Q}OpK3z1K{G5B`Hl0mTy|{bn1BUQil$b
z4|3;ZlmP>p;Gk<R@y+neo}_^PbU~a8MCkZQhzS4iOSKS-lS}oOSWG`D;&G|V`vAzN
z`+($2_W=qM;|;!kEekHZ_~*&_lMzVVn^C4>lh1F;eTf&u1Yq=zit{-N;y|V2;lA*0
zF^uwmk0H?C3)5F7Gfqs-Ne$_70;>}bwZV!4^T)FAAAFaMqDD`}!%rS^i&dtFb=;Fl
z>d}Q#W%5skqD&e1p1@suNg&RKb^YVw3N(;)`X7RVnbR(zSeHP4=UEu2L3s=aPePIR
z|K+nZfwD1*I00*Ie`+xTiPwV7K`2xd;N<duafreYR6>UhN>tg4o{B)^mB!zb^qRUs
ze8>j9EieO<f9=^(u975v1l6-$o#|VOm+<|*tn9cfR*m8mZdI}f3JvZKbPZP>nwtUW
zMW>^4z(f71p2>Zza&rHY;qPSx7Y7J~xvD^84ypn+lVKYUPKjUKI2BpWQ1?yGzhyUn
zEbby)yIy1yaCLdd)-qNLm&?~TFD`vPl#YGx(@+3=C)rx4dsc^fXL&3<C?x>2w(FhK
z-;FWsb#SFLXMUQ3xvrJ5iAY#!<q-2sWBn4Bl$tfVJz-c5NwGUZ#57lIn%x7JIMQTX
zQzX$;indY1y>t=P54*@{g^X27Gy}?iEdASI)`{C-(5*U({rAIg7@uUSmq9q(Tb-<I
zX3*A#<pU;<QW}yTW_KmZpqx(>qA>MExT73C;>6=IoSnhw-r&d&d&60R`cxEPEWw0D
z8{?T9WNkEXCpv`rjLmjmbbYy%2HPG6(pezYqjc&_1?iwlwO$60hWTg%(a<k{$r*>S
z<RQ>DONki{Iu0fW4IdKclk0Fg?%;I0x#>N6(Yk97>%YpuW(H;quu2;t=Ku9<?FwvW
z;DPWHnrtp47cvKg<dhMLwxG=f3e!{uB>XA#naW0)*v!Rt<TIB;RP%w}#$|nu)h{Gr
z6o&Jq1zwdoR0&ihLm<)JhDFMMNrD!zS?MfQ87y*<a$+VaK#|I`aGr27L&h*vnvvE+
zvuf`a!W1{2ndrTmW4u%xVF?jlF~DQ~G<4_xF1jPo7G?_Ell4zCn+x?W7%P-f{^Z?p
z0?k(wDAD1y9^hvu3_FO!bB41}2}kpx^<V)xLgu|P>HDzog+#Z1_UuD{1LD(RU4LPZ
zROs743<C~$!y%J&eSLjRrDXler<bJ$xNL-nYjQJEM>?ARQKnU-LtrYH)Fw>j26VC#
z<OjwB!aNbtALi-OzkrRDe3y$`9xJB)g-dQj`G`&PcoEi21)E2mGodmb3ALbvEL<W`
zd$FOS(fx-K%u6LEGUrr(ve3b^(HS8V1|K{zA4{M4f&-sLiy3Nzk%EE(kwzjA@w&6O
zly-2c<fNu*0RoU(fDM%Afi);bv4SKPou4O(6I6MPxf+WI3@5fl(YihzHYC`fOmZ21
z`-W3)peSjWQAA2e=R=JFvP%X;zw!$IeDv1vzS@%AiP3uVz`r|xm;Q+#NO;pkSEw=)
z8^SV+?s-NY4n=`wEDeF>Oa8u>K_ruB(4ywtEP<VzTchT^yfg6nE@p1-76`OCR1a<m
z{2Tx09O-zhe2>F~G9G8=<wBNOC<<Gqo$rY%Ir)H^xshY6E?UY>G&O=yqyP#H245MT
zCQw|>!~uo#LG&7bj5=H0%7C(@$qs26Mlx$p&_z}puJ77ncOrdDK1D}jFbb8s9fa#c
z@vc0a4nF(DDUe}siAumkMHBsNcWMV4W8&&$4b7v@pf3nS^HT(@ev*JqW<=t@6anT@
zjX=)60%f>eT28ibY0vi9xFTK>OuV3N)kA_6CtiHPpG&WQSu9>;>{5v$JO*Cbg&b=Y
zW_oUEk(miRhs;d;oy6sHiNEHU!b;~Zm68jUPHo!B7(1}oxSKmJtmUmz|1p~Y>v7|k
zT)y9z4(L$)g1e&#ZJsUqxIj1*XMylMoy6K*^76t0zN;msD~TCN2r{qT(KUcFY!1&R
z(P-7Oje??oupYer08`sj>&mmF55;xg)ch-_3uC#{r3ow-Bz{LE0iSs48%cwQ&*oN?
zrNl7=1n5Su>f9}-HSps2vnlR=LI<VhjL>1}k|YBqGMf@i3u}9~Ef=W5T<%3YL!3~L
zB)Nnxw22TyXw!LMgmOZTdKRq}z_OQuof3-kzy!H}qh=_PCkt=Xrn4<iyPNoKTz2`~
z&l!Vhl`i7}3#l>jbY6QTZLG5RlLcb4CkA5UCj?^6(!-oR=3zEb_9;EgA06g@%)^{~
z77z10W&rkU?ukM#!aY$Ces&B^%*pI&(k}ZyNV`SAE>5DwT0WC2e`b`5Op61MWx*oY
z^H=A8>A5|*)QP#S;K>x?^f8gj8|S_{9S)$;{^q2%#|w1nJM9!T=M1x1;YhRDy<TiN
z&CX*<pW>=cawZl_Q{fn#d_S*=pO%&sf0mq&C%FiowEhqB%0F-e_=?T9i4b2b>De|~
zG6L_sDK0wCKif`xMtY=@EJp@{6Ax2n9+w4w(3^2iu}7nCz#2^SQS$F8$cTOyWT5NM
zfsp=VI6upT>+j0lo#mRN1>KQeLm5`|zSZ1NH|;?P#tccT03#bbRlt7OK!oP027^F{
z0(9B_v~%qoV1R<8g8;}}m7&FvL1E_Vr<s~rX0Sl$ER*#7wa}SW;%s=Ra>>V(Q9E;g
zCXrWtaJ`Ga;9=m(JKl3^ow4Vv!6`C+a{+(qKu>2LjX8~;2QjLAbP(q%(Pz&}&s1GA
zrph_nZTVn#Vrt=Qt{f%6)YQ6ZY7j%MBp#7b94+6jV2uk2PIH!`Gh9aUScb-WIO4N4
zm(wITjvOj*7O-DhNn$#HBy@&GBv_7rW+F0T+(}v(QPp%Xp&WdqH4~xHkb4*{au2<x
z*pMyGx$zV^)wyxE04n9V09F=U)15YaqPaL&au@Q0is{Cno}c@44!Eb(9vXujA7liU
z<UmE}Qd*0?AMlXl^O~)IM2|d4SiBEIz@?WfpLn?d{w;=FJ>Cy`FQc_t?`?H|<j0pr
zo@Qd)a(ANfUG6p(5XmQgmdGzB-E$hSXKViE<8xW$TFc8ea-c!2*iBgM_WY^|%g*}d
zud?S~TXdyum*b+Zsar1g*f!&OWokn>ryDCi%mo5omUr}KkMGmM0TPK$Honvtb71G;
zJl|vZ^B0$4KFhu5XVIQCLal;6bC2|Vqy&V?pt|z2Zjbs3jAl$l#rq|J$zEPB%-<Q0
zdA^;kJ3Hryfbz0^e-<#lKecTyk7OlG_AdkkI?9K>qX~z;qY8(<qYH<=qYQ_>qYbyd
zqYmvE0yH$2Q34t%f89G>kK{IT@BS5&r|SWx6-w0SEc~$ET^nnhbDNt9lEdu<P^;O~
zx>ZYUeRw?n^{HYNDN4=RSuYMqE-&3=ks^y^eY5!8{Wt&meUe-iL6Ju3)&1criLbIK
z4}{FF?)O)J-MnuPs=p5NjcO}(d%cZhw)yv|Y{&ZWRJRWVfB31}t7iBERen&d>dUdd
z4#nmd{80q|poi(ssvn=O|91Z)Oq@FtM^Xk!dL?Mu*CTyY<McQ+Rmb|IZm9UkpmG?d
z6PMeL{?C2g^;G->uqfF)HfsN1zSwmW4?i9mcET0>;W|k+1LAhIWpuWo2#Ta2q#n%5
zU_e+|qE{vof0&L}l_r~jK=;Rbz*4+9sj?jiI3CN9N{{^IzN1cU#|>k-qe9(|o%xD8
zE1dFI7&nA{&|43{-E|~3(*aS!gVM0>K>v)d+EaBXPv=H4T8GYbp1AO12Xj*KT}x;F
z@I;;Id~e(tN4-sgHz=nKs@q{yB~F(Vc6O+xVpDbPf4&~;t}PpCp0R<-q3=%AZ>Zp#
z>KG9GKaX|81L!RS(6Pk%?e+ttS52+g2)(+NV6<rf*sBB5U^z7-wfurxRW_3vhzO(1
zAwrpDL73)-3O|Z4<6g1TD-DV~HhsQ@VT)uFc}P0Ug9w<5JC^O<Gh+yYW#K%bl%5Im
zpeVxCe}uv}SQ4a}>38cHC<?N9bt3fP90f^i$0(bjqix$yJ<^=!jW{U+34&t59F|j&
zOw!n#K(cA4(@s++05xR-Z0tr&2LQVJ;g4?#2()t^2>7VaY+$$=dC1PQ<|vGvMgQno
zbP*(B=IovXB2I+x$)G%tX=J+AElpcC4IxsMe}fu$d3s+Ex%#24KME05AQPJ>YCP5t
zM_RIGdLD<jkf428w@?eH>3bySj=#Y9NCqW%7{zj*aSNpo#J5s}o8hU2y&mg|;QMj}
zG{=EzNOZc_$*?d&k(cn_5fIoX5d>cFDgrF8VlN(2;0d=>0%3A8gACJdo*qUTX`%KL
zf6o!dX&QvNmHjPuNs-YqnQOHKEC>4##6?1#-gyohwkd{^(+f{K3pY{XIqYqm!}b=@
za(f$67it){%9Vndm7jQIS!_j9mJp!;=-ihToP&Xm^si~To0>g!+HvzhLR*tgty=D^
zg+BDsiUI4SN1)V^63mR~29PSNBZ2hke-cn14;GjZZm7p2PrDMI0I81Mpm+tuVOFg3
zo`7=06GrOHXcw4{Q{c^Wi4;VXpD}JwVBDYthQJ2sNh{5IKM~(&KpIm!!oZr-tI@)V
z>oV3$Fa@bSzyWlBI}X&U_X+;XN{>T5?c~GQRmXCGv#4jS4DkEw9FBs@N7bC^f3sfg
zr%L%2Zna3&ozBx(TKxqeQiJ_oZ(S<P<ia2x!nkF#rPJ>l)=x8cTAc>}t3yc_oJnep
zHdt&z;nSxu+pJHa)$G2|788|-FrY9EW>}VN0)>iy`FQAwphHQ_+7jVEJ%A6Y>{)w@
z&3_C;Cr~_fz2Y8(=MB}gj8BYqf5)AwA8J^WUM|%4^`3hk$*vG&<l#UdXht+d&D{0r
z45;upYlM3>fC&K1OM)OlVNnRssvrpjntMSOvB~@0HS7Wit2bS%Srka$OhGoR+Se6s
zy+^3v4yHjOb7MWD0E(H7vGn?UR&a8^QtdnX)ACpgkPc}8*n|N?cj6Lle;}52W%cQy
zhyCAEz2iCFe4zi*G%({dWz9V_Cq6&$UV*G`sC*<bxI}MBrU)TY!cwm)6k;AKD)(e^
zf&3bl)Px)CZ0`NFOg6ny7YHGqa*y#m^=A-Q&N(jk9&|lc^fBsv(muyTGBG2v$a37P
z%d;!OAQRSHd*6;z?b`Z{f0iUSG!Sk!4#e#JX&^96r1d_uknG6N8BwQtPz2Qm5CF%i
z=Vb!$sK7E9v<AmEa0Y<5ctrXjhmk9AF<^7XCu(2^0h)kiI-m!O0(8W4+L95$t6DO)
z87Gz-&~)_)Bn^HZN1Gjkdv#VQe;Te2^?{uZ=rkT-zK0{V&_V%<f3htg*!z+GD{cGW
zu!;aId+@W4S{ye4ps-nnxpcioDZ?f^V@O1yQN(k30L~t4e=>8ZBOBZo+H5gJxM6V&
zB_284R1J8Qz$|Ro&3$S3fGT%2sbnbf9)z2w9&t5O&vo<Rn|JtcgfIYJ|C#=4o_l;c
z*A>e>!~oTO-zx6?e^`%48bG6VffoplU87DQRXq#V?5j0Yg+wCRO~QYTHx`<51jNz*
z4G8d>{y@6}T~1BeQyq>1Bn)g#UzezMC1@UPkwTjmK&YBv8Conr`3`J_k(+03$La^?
z7_k0PB;xMP2lxvKP&2NNs%f?Z@S+)0ACokiR%6Lh!MpsOf2zu9VEs~$R#)H$Mp+nj
za{}RG<q;6p*b|DwQmYtq3yQNLk5i*CkuZH2<WWS^-<?&(=he(2MC6-0Eki_s;xN`b
z3llUXYX<5v;O-e;-#E1ZM@}OFCWUMNW)j4om0)i@c61h`c|wMCytKBL{$>c8(dqQ9
zZ#Dsbl{MNAf1Y7tFxnm0Her_wJQ~O24)V<U(!Y8zP6Gk{sEcv!bKC<7KAzL*m1nLP
z+ym?VKK-t_;?-QKJC~-}%MHS%*~577Oh^0Un@&2tq96|o>%NAbao{9GZav&rx3{;i
ze6|TFM=2M_YJyBTu(YNyiewRJoK-Qjpc8(%!~z`VfBE>{+DSi0aC%%HdPyJw;PfP6
zIb#PLRPTBt_K-ikNaZ5Pisbhp1w(sK%7Zj2T#Sgzw}}XYM64Y1d(E9;O5Si!Wi8%;
zqgn3(MDIBM>usPSeVI7I-G0wqK#yFVLv)~Bn1o}a(=j`?lm27d9ox2zFSebIZ5tii
zwr!h}Gv~}^7PY>+i}yZNCHRJ_%+C?q&o4itK>=c66oV`V0y)2AF!0mI58eG5rnMi4
z(IWVCJ8mMpi_#MlhldM?_)foAZ(PQwW~2?#{G}U8i1y#8!%@s~{|jF2i^56*4}~XG
zwcX!*(d;;x5ve+=4?c?YI)+4<o28MWU5*vBiHf4G<ktdU#kyk)_(b`F?7c=}5Fse2
z$fyMerGE@+XoUj*WrtrG)e|PmqA9Hb!*{R@2dY2dBfbE4ld%BO#)N;jku~hiUC*^6
z8NPK-th0nsbhRd5z58+WeRXN9JfG(qol28&Qr1fTEQLsAS0ka3E4#AOM~hp<aS{Ig
zc^fKvo0e%OX5XEiAjP}bA#4)hRJ3G}^IzU$EClW#k>K%MoC*;xmb?4C8UJ#D!&xsE
z&Y22TO{#oCI+bW?=X1)<<U?{0L+<JZLX`d}6UerrC+N>FUc`KcNP;m!$N&1q2u{Yv
zQ^BlhD@!U7`l4dM1Q}bkG$q{n-g1uMX&!uTb0C#s=^<yuP(*Z2Qog~IXXtcOhVa5J
zk3%8ZP@SIvCs)~5ww-2Op4bk6p_!QO3<v9MqqNL6u5_P&Z#T#Z(q8F%lBevR`);-0
zE&p34lek#!Y=%(sz`K1j?sT&VS?7`OcXGZ4r7FsaLov>jipaJRsctcIG9KQ3O2KuT
z_a7P8%c3H~t&m72O0S_FFiKHkxEoS%M^hZ5*S21Jp!3L#;EgA2DPRHc>zk1_A`UuK
z>4`!mXbUE}-#@S9AO+_Ii-Pb5k7I(f?eO`774(G=c{^&k#?E<^4^yc77zkgPzmT((
z?Eg=^E{U$OPA2@_pGx&9q28tKK`#R$O!^aaarvvPxMh9Q4C%XByj;vNoAOaxJxG9+
z<p%J)h8&|x4|!R!^xp>zE=;Q~bM;$WeiMg=8*F}`d9tqhwTCd*$KmI*5Sr`60WvI_
zTT{ezs^ScGV0rZ+t*X{y%p!x6=sW(rb=_ny!ht^3AaW0Pl4zvoUJK8K<Bnlpy==>$
z6)Xn#-Z_yk{~3ErwsPHME(OB(rPM>#QFiqE95kBP*D*qpsVf1@Gd*&UK7^DkbQ-yk
zw8E0fVQg3-h$VZLN9?S6B|v&gvz5l{!L%UT<DZhXGS$OHE%HOVYQ}O1ge<S`m<pGJ
zDjN&@bXkS`DAa@!aKh}z;Y0rJ(dd8+IJ&Ej8;zsZMBN=I`d-)+f9MkTd)+@F?&q#k
z7-BFGX0^Lgu}Z)%q6+kLl+5XVPo>9yjuN#;AJZvTS&e~UM$Wej<?6-ZMh++hE3QV^
zMj6wAM<Q$*IITX;$%cGsKPPi=v+G?1VWamoNCBSB2qDTTes^Q%`yz-Fs}x3ENY5<e
ze}c5cS-OT|L4M--h3F5(D@q-rG&7{qA}@#d!vrP}zL>!MP$c-oB|a`|i@a+^`;)4C
zC)0NP7gH`g!(~&H1ATAFRTZAa#M1N?>&cr5<?jcIggfZp-WQ<WBSn3ONP!~!#&9wM
z<&Gpmsw#LQuK5xgc}6E0?>X}0jDk(0bknfdflV>df(WNUy66k)wUk-$QtxoKF3`Q<
z)o4zePy)a+?~gk=$#GP9rvua1U(%goBk;x;E{>cT0#J_SIV}U?H?*}1csbea{Z0k;
zlmfb9L8fCp&BYN|`nKV@Mo8F8q0Ua$1+p3yf6b>dWffT#Jz!r^_bmxR>Fsj1_t6JB
zoM@9wvb=eYhL0gS-Rn|+_w>-h9R1>CK||JC;$H>Ev=L`z8YK^f*$-3)k|{G?<Pzup
zB^Pp|j{7mYlXL!(w69y(yU`G{Lm7Sn`J<50rXHnK&EXiHM@AOenU31v^8R8sn}(}a
z05hz2SL20Jk4HO0>fU3~{pWkB@+y1_?<ARJ>Z}SpH>OO6!@pAaJzpKxL7a?E^%TY!
z)?UDf8xeKS^34=Vz>r~7JXFNLSNAf{dM6>C9qENyalM+PD1E1BJsF{9zkm@7`@|%B
z9#=26VGAi*hS(A315%8(!Ub)NWCXmUssBVw_TQ7P;T^Cx7opRAcot2BV}57$ovfgC
zvVk%tWYtGnqot3^!>Kt};j{OG!cR-cqlp2Edy7RAm$3iYes|H(@^2x`=kC|WxS4}!
zTqn0P!Tok65dy`P-9BH+Xy`suKa8_WRtM#7%V))w^q|aeLa#VRq<3<FLx-A`74fQ_
zB`(5E?p3Q_Hbp6?UEU1(3h7)oEZW3i&EACCSV@2mQe5SxEb&TpBTY8tly}JxVKM-E
zcq|6Pn?wu|NM1;%GIB{AlEXOY)WZz~^5bnrzgsmC(k`GXf<QkUiBH)Z8#Y9nFw)!Y
zae#bTQYC?Mws&+5PQ5;1FL&fv<+30AlmelwP};06iQK|Zp3knwW2o58R!NMh^Ffwl
zZjXjZJgm3^3aP9nzWBLG+%o!6nFBzn8u?y@fZ5a5wp2Wf7Y+1Aa2iamb=5Y^f4|~R
z`54LMmzpg~N`9heFo93wYf5{ugq)qYX>@W#hB`KHuCav-_w*7wgfkna2$O3OU_0Uw
zzIl|q;&AspphBN3A!RS0@7)84|KXk?s9t7v2(wbT0;=%8IsbI&{La|IvbO@zeBmvB
zyH+gtu71rkSB=lM<UKrhUX*yN5)hgmjeB~L$F5nek5U+s>CmEI4k+ly+ryV0;KRW}
zqmrmgN|LCI&kiaD#5nCH4A~?0Fgyf3FUlGm2UnXCk*`$Lj4>IVir3Fda<%VWZ=6Em
zkj;)5k*&`4RYRZZej`zLp1J^U4W+i3<S|bw-Si<=T9!JN@QtLz6?zC_WU=0@OPmrV
z2~Lil^%j<_sR>o_%`$dgGfvoAs*IdATV1{2Ga-k(9XP*UU&_QV?LXBP)JlE+-L}TD
zq_(JmaoKj8slH>nC7Iv(F7#9Xt-)!kooAVV?vd=qF?&lm2g9Gw4(kACbd*%+0%!wg
zFvFZbeXwD+MyDY}QERWGjFNrm8UI2sr+o449*ESoek_{GMeq-%Fu3$up{S1bUQ;6r
zyP!mBZP0IfXRa}zZ00y;ebf(3)u-$pokZL?#$hsuI%aB~ZJ|$i`v6xwVMuAwPh*6L
zbMRjK2FUsDJGmi@dZxfE7=SwVYtvQj8L85as_)fh%uW<f*b!<lfo-6tRV_NS;%nEj
zxsHEaU6GB{AWNz@R-VwvVdNXIEET5Toh}_o9C(6z=qoK(cWGU^+i|S5l_pCIXnfN7
z^<VZJ$9N>!hxygpc<Icu*+HPa-6L8)CCr}ka2O&38=|*T@g0z{8Bz5sxKpO7ZBmbn
zM!ntZ99^L!BF}3E@8GT=yz_|E#8A+_&bfKGKDiz$Ky+4*lkbf!L2v>m@<Qp@D7M$i
z_EJ_{U~Sf^xG0xpWO-Hm$jV~!_)=q@6IoIB5+`dWB>~U+uMJkoM#g27dT7&FE<8^A
zl*qfws%_{vDmZ{rn}F)K83JC~;=WTCu=D+9=Pqe)coj^8*s*&0xVOE!`ksHGWsNuI
zrP_H4k?`I2L;Lvoh1#Ek#vK&f#3A0a^hf<3b+sKv312)+!jCh3DI_5G2*R{x49o$C
z!#c=eRNn8uR+#3Hzx1w+><bRMnm~uLtA&ig`?7b|*9UGvQT$axK5b8QDYOlDcC|5{
z930XfOt$ULivKJpWy9;u@B;neI_-nzKFemWbDsa<%~&DGgzx;p<<sGjTwGMOF(`IF
zU4yh>B*-{?Ti4+`DCJn9oLbjc$pB$2a;N0@i70l-zuL`fZ)j+Ge7r?w(4E(pakYk7
zlevZNVFieYG*Ge4r3sZvX2L3*O%o)gP44%K&1|$QnrFcaEO2@+r~El7RUn?d*GdWK
zar#}0<sEmDM|7tt`QD$!=PeQ$fXqEAAca`L{rB0r{8#sx>7g21b=LzoT@0^39F^zQ
zq!LfYj7qMpooK_1=#-<U!(hS$ypI#akB}d=7Ms8|q>}f<)(M5x%K-Sp_p{V#8_l~1
zD9Y<~gXQ?f&eJ0kK+c_noPC6#f6=MSz1g^53d`I}M`8>W<J!jCl+SM#KrjAh9y1G#
z8CT;=xV41*-axiVdT*L)qE_%;>D9Ho4m(u4pj5la6U7->;>mtvyJ)@15Gk8$%VaT2
zlMl2S+_jKDQkFg|NVqozgqUh?E+G%aiMdw#&Zh9DS@;?1<aWJ{q}XCWnd?uW;tw!k
z1ixwBI9x+0LXW!wfz_j%+-j|TPt9(^9=V&{cJkt?QSQ3@{-|(^O<J7+UVT@DTmRhe
zir(s)qhHN&`P2tB-OC?Y4v~E?UvcnNM}Vtk##2_r_W|PiOtPO{o|A_Z)!>N3s(#oT
z%)Nkl)(hRH{HetK+@SkUFHPg@t(6py|9m5+&i?S4UmE-@YIY;YtF=Yt*^%<OD0!lH
zy>n+_e5?4n+nb*+jM3}Pnz-X=#_;gEyZlW50{zVXWc~l;xhXbtNSNSk%>RcQ;*2@$
zH{EpVG7ZU#2!Rsz3>C$fd7E8qxZw}2_Cv2hhxeNuAH*8^H>iBRI_2YfMV8dX3(1z2
z;_S|tvEqD?ABc9#**G<XCBe<g*n8caMG!`Av;b`a*swbuxj%jGPe0^sH=7tfLHR~X
zTzoCrh&?|}1{Sz7ZR!XklskZz23Jj3{9!DQgTtYV_EVeV@c~&8i*nsvmUW+{v^wh7
zGHbK-zgf5oSHb)~KQ-n}w)zxxZue`g-}(}0P8T(B`Wl1GUnjRMaK_T(yThAaBw!$g
z51N0^DbKeyq9b9x2q%e&_d6|Y<Yr%AtVN@s4u_n@%wM&4SV&9Y^FIK0#mL>wnP>t<
zECv!>PV8O}^eQZa68NOB4l6xg;NrJcJ|Ns0R<`3*SM&v&oPRe~A?FfUhwG-R>F#)u
z*9XmG2knfh5tRN;?-xfXS-l+la(GBgUEw`W;>p-@;!q=5D`QFFb26~Tx-Bub?2e5%
zEugqrZv-<`@kRNA`R4%n?7k#%)5d9$it+8G5_v;3N<WPGD`9%aVV7d~B4l7$y1NNS
z;wSzEaM`d9tAdTArF%F!(*K>yiZNC+rv78!G?R-&4B~?DvUvZ5`4ef;yC~q_X}Hir
zdY-ETjWJGecjF)18<X<G$zM*QE#d)`f*t(ru}2N%w7BM~oN&7UE1r3Ftgt<X_zNCh
z8cablN%+uFB7aQci0Bw?D2vFE7resGyeKv3&!k^+I#4zlNUb~7<#nuxh+p-k9_`GZ
zGN0f3^9|hmv*5{v?eoy+9(fh6Wpp!*rdIaq?siJo<l8PIxO=|_K3DEEGZ<1f3mF)P
zyCfxnBN@~SSnV7@j1d2ar_`hXH%@w91&)Z`MPtZ;W-me;FQ7jJv;B|J%;Ki%<M!Zg
zW$=ML8WDE|nqWHywO%4bsgA|hP?I&SElBz+qTuf}!laRJw}f4=uH3Wo>2*NeF{QfA
zb<t_~>hh4On$^ua%Q+fWQ9X;-+vg!FVVFDkx&UE{FYPKI$ZHGF8|<Lkyu4c;!)E)A
zq@T*YDZpb1y?vb^2uTi-H3Jx0U<_U$)5hv}42k<e{;=}W=^;{T`e?d9Z2SXUzd+%{
zfi!}RK|o=DUP&uftoTr&rv~)newpXGz$@w;*Qg|C$*d{n1MJEoI{9FF%@tNmY7v+4
z?L(kRb98`F-7B~u7~5(Bw4U@u7Hn5%K`>FhuTf;g+Rh}2)7b0<k!gI{Ug^ba$BD$@
zdnpDpD5J9oJ<f*0As#JHNT*p0RL2jezKbnfp*`g}<fP_5;&6yu0oi<iifzr~c+&D#
z@APiaaPd{7Xs9>wwGL-c8;){?OwV&GrTn>_FRTF{&4q=a^~a=I2J;qtdCozDJ>gmu
ztscpJ#=MhX;y;y`6Ch%_(4*zM!-(=)pCJXKL9iBnAhcA$k|WHL-ZJ{6+o3|M|16q}
zhLY?Dfr1m`2Z(e2j7QN)2hfk__^7YI-}4fXJ;7fk(LMgu+cZ(@A1wuuD#G+^WxuE`
z>lXl0sk`tcZ1ts21o2?V1bl1AngfDpR1Ml`5TSz5CQn`pc}+SaXZhuB*5m|F(cxzK
zN5=b>{k(n=PkOHY-RI8-LCVqJ*p#5|Cw7<FlYIn<#k0{)N@4Zkt1Oq=va(WEU}x*I
z>e=fVG+HIm`+kd22?fFQToO1K`7hz(sm}r7%sH)3h&&_0{CiCnx0imHg2#R?L~&>m
zyI9)35kmCYqH;|%*vN;Bchlz{HSk6IH~#`T5r}g?tI-`R)$Cy61FaI*x~lZzzde&g
z{eKT&>zssd95Pc+#LvP}ib?o?hsQgr!H&`fCeWH4(=teS&7>6RcqnRknaQ(UnqC6M
zLb-w{sI}Px(T~siEyK7L>@z<-#%(glTv%jjwqO2WQ-pxrtBGN24=jsgg8W2yN;wTR
zS&QN_<(L>E9D#HT9&LMGnE@@4Qp#5IoOSeYif-Tvh9E}zQ4f}e-#=9@yN98zfoFq^
z&Z!$3(`}X7VcTAo69(#ppvCYk*{lul=a)>uAaFFZ5Ha?<E0{Bsc6s<{fm>L%t=A&$
zWMl_Y+~9(eLv!bm)kA7)oE!XvaIS@m-dou2E`?G#2&g3dgKRcT)=1}9JLBKgq;UYf
z<&kLfr0v<S0#<bC#GJaWJ&T~B=oQTb3dUod1v0D6!OcXy+GYphd6~E4zH|T>2bnYL
z`){=U7s85KC&BjBexgw{k|q{dYB|W{G|xWPf{g%=%g)tvWga|+a5Qw(Zq9{}ELhr<
z9r}xT{gonz_;)P?({+m?=<$|5&Iw2<@`>nXPVBat>1}*)W?=7df#YmDG!|Vs57Lzk
zbcjAT4KK35iXqDEnyxfc%rQ4G5n1Vmvt&UvI21KWTIk3UnFC7-qs89pb9X3%`>s%p
z#gGXh!RZkzcNrK5@so4OKZpLx11Enh?o%OdGQP*QM@%7Ne@>*j;JiZ0rt14!xH7}{
zG()5|Uth7|vFGuS-9}t2VhWB({k!ZASxEbU<_NCVhBcU`xZ^ovi^B&Hix<Km-OehV
z<l0hOm;2yv;*=)2u0M(YIvAH;x)?bL@VD|Ebx!rqQO|3GKjkz+$gd(<7F^o+p+#}L
zh$3A&B;to`!gc{dZs@A1O5iEBrWy>1cGGv-t7f_jvYWNk6jd)a?e9+d)Lh1z1@i<_
zk?r1>%$O^e+viK(9UlVhDU!Fb6fdgeBxr@^nEi1JfTddlKWkNSJ$_5@c6a{0V-1Zf
zOs7`Uje)*eP<}Z&a)vZ#BL26r?e+6@VJ&8<r`AY4qytUj$M&1VauPy=0mbX1L5Tau
zqX#P+8ekAr@Jp7a&YvP2jBZxUSoD4cg~z1BF`CDTyB8aWVvG~m2uWXYbUHfM+)bo%
zHI@{eN&CA%B5!yu`TAQlt)96MR|MZ%%O6n)yFh1M8`MCAf)qk`rej_b>q=(G^n%}Q
zEnlh>)v4--pC;x(as^B+oC1GsEJ#^3IhC3pr!Cx1o#U)vB;h_~{e;8?;*n5n_Bv=e
zUMHWd&~A)j_(5cVOj4`HK?z;(2UQ~<G`f?DY`Om*I^?rQdzA%L<+WsuxskZRfsrsZ
zK>Qjd+2I-DLsPdsp2L{b9w%@|ed2tvMrik$H1MLUyfbzshyF+i`~`gH3S1*rzFwA#
z$PeY#&_HT3*92iQVDH=yv)BHz<)reMxg;o2PPTL)h6WVi8~k^Ys1e%8kDnpN+B-~<
zcV?f$SFIQ7FD`@SqeJsEj1Y0?`L6*KHy6*jy-}~~549C)Qx#Tg<1}zH)wN*l#?(TI
zF<N#glI;Guus*vW9kWvT`%nLeNvrl~y3nA{{0;-|Rv1;;OO5({**dox2t}MY>fz1F
zy6q0yks}KrK7}&Rl?$)_gf6V9aI$ltB(cy50<;KU8x2h#&etcdcObM+WTF=;>TKzk
zc~O-|<P?7?_+NJsKkfe7fXCW^O#&hpJ1S`Dx(*V_FR1WqS1l*-RFXeknldOY2*18h
z+%iVL;h1!)c6qZ7>N}FgRYhcpDCZMUkbLQP#O)Y>LnfiofA9<s``3^8cSy8xbC2ad
zEc-V4bOsHBA~uAF?X780G&fLjgNmztmt%5=9Zv26EyeT{(4}z8A9PI{OvQgQFqFp9
zyKXt*)$NCML7T387bwcmVMp8S5dXQrmO%RqF}vg~b?F+8RZ7eRZ|0l76iV4z@evg?
zds40fM9Gri*e6NNo))-@Cqqa>hMUU!q7Apu3Tu6k!pL?uM+$!Wp9k50yWZ~@xUv;9
z|Hi{>*a6*+m8!-DzvrYHRjQ>3TYrtg75Yg}EA|Ka4hmRw@0QIxJW6ltg#{?t5_hE!
z6OQj56mH$qkxn-aeLMI0Wamqs7MBh5Z*4jOkFo!R|HQW0fGKmb_(@hLMl^7($_o1A
zVkQGUAdKhsO=+OoRaPb7nFUkJvUE*~Uuze=3ug<QS+*h3IQM-lUQYdlk&89oKb36N
zrP1+zX&^EkXUHyXxo5i2U;ZE=%j+=&a`j}rk4tBDr3UzvR9|}}qepl%kW%MP@q}~0
zk$d-M?CGC>BRmrGdvWlRXJPQ{J*OPtbTq6vkWVG!I~r2nQJC@xeZDFQbun1wXwJry
z!>t3hjEjP%KeC)6BY>7+LIZ~BMS|&~3gKj&K$$6$_#`=E-;JFTi7nl)vyX`fk60Wr
zLp#S2Cj%AsYl)iR20@ls!oh}*;1Yb`H3?q{jnsj3|MSnwt9zo6K*ba>bawp>*~Mcf
zlO+@TKB$Fc@0Z?}o5#S>jpKG@P?y`zoDv<m)H#utEQi{aBk#JGmq=w<?LV5e@eP9I
zwmVzCjIOVZlchuvS))sZ7Dg2MGjiME?8>xsS`c-DF>>DOkVFtHF<yi@@CrVlO^Jwz
zsQ+bMCDBBWUVs6Tr{&IC!XfKf`xPN9Ys?jwow*FN!93fD>x~g&N#*j6WV|Zc5Nt6u
z*KrkMdtz7dr*zb_Vm!VSJieQHQ~?&DQfFfEAUB-Tjo)`!09x;3Ic&we{g3cxY;|qw
zS$B;iUIc*-GeOhen5&(3=&F`LsTg+IO)wgUd>|a62kg02GPOFDdYO?|g3yo|R(Wom
zjoCAzUHDdCM&XF_m~HvX8THeJvaI$pWW1y+<}_E^Z5hoEvBax8RfE_$^-nj1M=-zt
z-dC1+zDIB~kx6KuR#xwuV~}3;a{nkF)qM1ue&UKeFP2du{Rr_UNLjB3z&(_btR@%Q
zXg*Zcb-3m-m9&t+Z0bej)K?T}L0S{1P~dAzte#8o8M}mvqz3~}>W~o|w6FJt=A>-X
zPZy^8y&}gLs34UeIL<34u%1-L!NhCM&3fF%`EIr`ap1v78Tq@$C&6}qO_yC`m~zJ4
zzje28N8EcuCW+gdgBIhHM;lGv=~ONA_q^S4HK)65Z7Bx1Oh5$x05=*T?i-L({zTLo
z?9vGd*+u#ma&yS;5BU<X|7)-FRuM=^|L;fVk?OOG8$%Hc!XSsu>4eQs+t<2=6q_wg
zY@>XPSemf!eB&!|lFHhp8_0s0aki~U6DXTewysT+0V5Q;Thi>rbO{yVdg9FS1mf_y
zmx2lk08&mOK4^|MI$$VZwgSFb0s|Td4vh7c7!dhF62Y(R<lc=hkjDYyuCH$QNVBB%
z+0yAIXLy4opg?#mf$(SCB+tz~PKi8qG+goRj3^Fp<LkI@h_$UHzYa+P<Gd6>0)^3(
zCE5|{J)Fa?9_osA=A<Gt-0L!Dm}4WUgl7(Vm1)RMq#T#%b7101K?~VE&|bG*=dPGt
zHkEq-3@yHkM?%L12?F3mSCkx94Gcf=BucxRvAnvZ+>bcJo67m+8o7YP6{eu&)$(Sa
zwAnXZnFjtNJ!PM;%E4@jgtiaN>h}Xxm#>+@t=_;x4ll)Ipi0zOpJK)U7qWQ{KfhOw
zdUS*wo2rD5tXK^%t!Z0ZtL(|8UI&FeBOUID*~A+iPVx>nk{YmZpNGB{Af)~1v{Cm;
zosNDMcn?(?@!U5dH5SXR(GPEM3UJnY^UCliwkYTZEB*u@S1k%wDTmh%h{l|@S=>&x
zh1OIig_WKrMN}e@R=q-(Of&DhYFKL=mpI?>GeD2$N7!aMb~bDCL5*k6Ng^;S|6>eJ
zX2!8U5(-WgP6<$y+HNfLeHgm?SxDw_12bi%MaRGvV?hQ~oBeu7r%4^pHe`3Fqc$8T
zUJD1OL5ny=fWCmk)$yJ~5Z0JZEt1VSSfdrn0ihiR$|(@%HRgG`;o>xwm@a9KfsSoN
zP(|&iEbPtl=GBf(wXXR_T+Oi&zR8AiaaMmPDSn)r2^_~9sD^<K4!(F}Szb9U!BwUO
z<Rzhvnv1)>#2MQV%gVX075x{4Cp$+Ad-xo0ML~*&XEp2FL?<aznA&_`fidn2upqA@
zDLJB6#JNgK>2TcrQ?w~Due2@t%qaM|VFR*?Vl!2OX2zSxd|RQ{(RrJ+4A!V3HlSab
z%6o8y1vs8>oCH;94YyRn{Y+HLBX#@CA2O^;pO}7R6bE`xI~4OBkgY|b2bMYxUbmxl
zVbUke6(Wr$)MCp2mRtN8)uCbUom{9;#hbISS_MCKFwN)r*I2TF^7wRYCN_~eR;Py+
z|H~jBkGl&~5_eL4hkZ|@F4=;MTQ5Jz)9?i&7-%T+K5(%tdcvAgGfh-MYoW&~?dbG{
zL6)q(c~4qkB=1g<=-&aGiI#_)n`Tnxg8O26_2n8~3W{x=$)pgN7}w>uWkJ?F?H{&5
z!Zxb&2Ov{n-1HO?t`PHA-5XAo_9W|~S8@NQ&p<8aG(H|5mM>jf8}fO9TuSg?5`rJ1
z0$8~qLLrceI*?c(VIsffqV4-u!QWXZQ<Gi;xgZ>A+0DVQihI08yu|9~V&29=6*7Oc
z*;~$Ox%2?f`u%zx`gu3$o$ph~&?F<?;x(PW!7(h@r~I2K-3I~Lxy&l#apZVTLpul1
z@0rNt=gM|N!1Y9(*BxJ-F*w%$PSrvRVo<*I^INFjLi-l_w=lki`7Nw(VSfv!xgZ8N
z<{X0cH#1ub5g`s5B-3w}?^G*$SX>IAvf0q-U~7>aM@g-8dA-(Jd7$Y6-*l)+yEZDE
zrpbB@Jsth+5Ri54;+ed7E4+HskGauAo$@;LI@B<<axj22_7FhTB*{so8xG4#^9zF}
zG}SlUKQA$|BrzSM6Jmg8tp)K~UD+8K3Bvacmc?iyL=mDOMCQjg1N#Ar1Fi<BfsC)O
zuV3$j=xcJhLc_B-Qh+3t#dH3qquT(~KVd<!TEFC`OpEe^uwnRbUUtA5ni}gt{WExp
zKC4la{P=wnIDY6^=p#bFjM11$j)sAk^dBgJ?(+Yb?wQ8P*|svW)<=k~WrXs_^P2}z
z)P(kZ=t2fD)Qk-q;H%_d_A&rqIl&M>qh0t7Bc@n$(wU|&e_`TxBOU7*?i@eAAlBGH
zq;g<=FK4=K1YZM!t;v=CF0?&oWAVly=z`Rgjq+c1ryL`srWl2oAv8P}MYpLkBuXuC
z=fuqB^jh4MKTwRTXZeb4ZeV%-E+_SM>)q?+k%>UZ9}Pm0lRn4S7zy}-pu&rBSqeQ(
zt!C@FAYpvv?VOu|GQTlEIe|duq~!*L{hpq=K|Mr$%RYHm&G!AO^c3RavU~!J*FI@Z
z0s9&|2MCRH=5CgHuSHgu*9vPBP=C{RQ&m!dq9zcNB4AA><KnY;3i9~68dUzHo@G9p
z6hX#HVDS<JHHV5jAIPYjGnD{Y_{tnb`0^(HViSMgBEAM9I=;U8Kjfl5o2A~ro_N0~
z8dB>U8#8hzj=<jEqCmZPnS%+tkGQ~NLErDD+1K8?6cDUwD>lBm;y+n5@Skm)1e+Gg
zzfKaInBVcj2#3!VG035c09-n}0orw6aySgx!9x&5p|dG>;A5-`jv16wP1DymyY17m
z0K~(Dyy)w?G*uHz)9(|-_hXBmaCo2Z1dx1SxQ%4CxE&>qZ}<usdeHTQ%a+<M*tK~Y
zxd0l7EiG@mkiNQMU{~5exlhtKT0rhAtj4oluM&Q+K_@(Y<?eU+rhc8#S|0D~?LbX`
zo~MVs=x%%gF<U^U(-*iXnvs>|?bIqI7q-7NFWUux{b`TO#l`Lg_g+2rwes-A{bhON
z$Iq6-^gfzb$z&UAY$5tr@wdw4c1x&>EgQ?eUX8YWrlfwS?l%E#Yq5;s+S+5O%qa9h
zI^`Mh)*H65mW4L)X~$Od+gO$PnizR2J+Y1AmvI1)7nR_`=PO_Rthrk1k$7~_F`mt!
zfhTS8wzd~8=^;`*4Uu47fmD-%JnvIL|Kf9O^yCuCJ4n^Ygmp3epjWa0Nl}ajD~9}(
z_(38R0It3ChsV>cVTH+Vzav<USAwEHO!g8Aso8n(>dzE?hKpVq$)t;h0!=T1Jvn0K
zNnH`p#b)Nkg7szwSTb5MC@F7lEf15ILLQ64uB>B*QF5x{Pb`M60p_Pyx`ZI*5i%6?
zK^`hm7e1jjcCSA<9qf?#CPw-v)ey-BG~_p=(QOu^Fnkb1<?aiGSEJ}$`UA3qc+AVD
zy9S3tv@ajp<zIbBYxbIsurkCb1K-{wOxl2<+oUhXc$E87hA3v<Wy?IXnb#&V({otK
zt?79yf;6p5W#qE*3H|(BrC}{4Ge6DE7PnKBsO+xhNVl3`U?Tyo!{+dMT2=N}rABXS
zpP`IK1$|!eg)O<|`d{T~Td(UqZ(7J*#z-E&y{WIiak)Gc9+*Nup_1z%j<=Zph28?F
zi7P&e4Y>I`|7jXRn>qI6ksq%}u?fgIvfb-FSnHVt*lPYkLntgb%qlJ*K%Hc{*9%>m
zW&bx^H=;?KM}9z1RSMDa#&u`vw2eG0#=e8TCO2iXIL6ie(?~P}UyZ5Z4vy8+{uwTI
zYf&}zjU(UrxzEKoxIBYp+j3kw++Yxxfg7iWb<+^9IsFC)tow^J<bd&G$riQ;Rp{G6
zjmeFOw9xHaQI^DViBsI;@zrwj4FSHbCx|L0yMsRV0TYg_hBV^t2miLguIllCXr0gX
zfRj$kkB%tI>(R){SHn@N&jnkn@(lR`yKOP+!5ilS4w;(IcJn31d~L~r5*ff_#(E>u
zpzmQ268<0aWuF;M&;@+@jV(h*E8(s(xLH1@k*bEO&BMA*g?Xe3TS3eDL8o4yngzK3
z{@qWNB{wPV&tI$yNXX?aMqm@#f}`b;L(wZP1D%=uuD$D*uN6;vqcuL7c2wI<IZ@*v
zGGTRk){ATxM;gNefz2@+I_CgA&p2NY`KJ9n6kCE<n+270>@TWAS%rjww<cYwF4sQV
zH|<sdXHkOu&7fqg%59qS1=a@T)-XB$M&Xg$*uQDNXAPzwQkNS0eL%Sa!L$fXxLz-6
z?rp9|M%OZNW0<fEtDh&dM7;Fg^3*saX%@cOhpSxnE(QksSmBOWuQ4E!7M<C=GOR2a
zKge3~ID~9|l18)(BF1J&eW1Su8QXDdXUU~m+Hh}jbiVLU!_+}>mDp4ITusYm=g^qc
zbGbh63Yq3<gS1Y0TCTbPeHd&U^2gz(hr&Mvl@0Z~e5^xq?BC}aIP%qjy^PqRZR4b~
ze#GZ{wvudk-ndQqGg3e?)wWjL6~9E0R8-&vrEX`A@|mF<ZZ`fcLyaU08Wl%-Eyq0_
zRyzN|>mW1#<!2HfAU`AjjYBUK>eF-ADHuJB<Ai|4;O73Nq;=7=W%u+c^f3neCwr|y
z-Tp_UJ&X?w+p=m{BI%*X3)G>`>Ze?hkxk8*?$^l(T8jWPK0CmyD?-i|UUR#-#F<zO
zuBHz!;9b~`p&G;I(*BmEYT&=)Yv5wpI81)gl!$qbl?Q}qGevZPJ4)jHjC9W+$KI0#
zeGX;>cNye;A(<ImlY@G5lL42HpMgBv$I}W}y+lq&I+8hRPy7;tlx~~<s|P)Hr>Zu-
zp4;=fu~RyqW&#YJFtgG%AQ|l)6TV6D?W9_~>dsT$apOu9ktDpo=KS>|Bg0n8jLrK+
z0HBJpnF>CD^U#4XKZ%@*rmXPAc%k%p)2@X4%h>_@<Xm&1iS|w9rCu7gxW5|MrSKU%
z$=g4*N<b&5#YrXXQYLeyuii26ts9O6DJfLTz5Z@cn*=xm{FkE9wH1jcTQBz1Bp>r4
z?#CWL_Nu!~QVE#G1LuC#9Q`dDVf;GIO2QQEwBPxhl)>gz+T!FZ7A0t?wT5j4?h_L$
zESGgC9>WMvhhakBOWVVCe+C6Jw8uhApv@DZY4a#x#y!GI5xbc-M&GUWtYEydHQr~d
z)hX4&mIFnynWsqc(|k(TtRkPxyJpT)Fn5eTtz1<#9v8cByXNOLneBG5iZi4u2g=%F
z5Kpky35tn@!!oKHTpi`e3a1EDP(8~YQAcaa`|92l$U%IK?syi74B1>T=XmbKi(A(+
zKJo~C5OSRQi_w>3Dg0Q&`ivAo0wgv&&q&?kAixi^lD7h+;l>bVJenEJ_wcv*%5bIB
zUN*tN>7LX2c|Y%uV9t40|0t5*YxYN5K$|+L_jfuOEWRWm%}oYt`H};xPppl%_0H^7
znZ1wmhTzcKUygsXw1W2dT`@Sd$(G#RKYTl2SZJ%n+<#)Cxnu9DkTiLic*SJ1?760~
z9s^<Wij-kEYf0@e?fhRRc_(IzF)dTVn|pWH^U^a_9lKQ1<|%6NB-e0SELuA4nkbR}
zet)l?VDJu0RAo9-c<uw$Pzkkt$t2)t_lQVyFE!kyU8N$yqA_f3xN`_rF?(Y->^pPC
z(tUB;s*dzs#NNlr1(K+JiaUkk?5G<fumO5;Vwzu3uf*?BQL&<YZyT76rJ^?DW>*1i
z6;1_BUR#zGM*KN0)vUr~pD7;GSIdN4A|)6s;|1z7#iqh`+M*VIjvuzjw$Y+~LIKwh
zE4&;NI5;Te^%#dLe4Uq&%E?x|bIb6NSs_mW^7O@h?-oqCJRHyg2o1Q|m{)_cw!npM
zsM-a-&np?p2WF4Vn96z0`)zeWTtO;W+s{W)wdfT;@+GL;|3?302%;MH$V_Ek4wR@x
z_lFm`x^M+=5Y8Ukf8OwBWGuL>3t-0N;C9|&v9~)vxEi!dJ?yf=h^JABCL_IS6pkFn
zbtwtU)puI^_|$IBYNXtyy~aKHbO9mDmpGwExC-s6?MdHx7GP)U{9VP5C+2lOjD8NC
zBY+~hl0T=jU&$ddpgOzaJr$odt@@V;0C7FLxDFzrORmax4Z~5H^?8Mae51dkgQBt6
zl*4Hxe)^0vp~dbOpJ;6IUGUdWk+4Q$4_g9_ic-tgyxQ^~JaTn4BaXHD!@yOu45%!B
zZ@K9nJak&KaF510XY=tVyuEpWFiiZkM{CiY8GVh(hIZ7Kl7Vkia{m3kMvj`(_W%@;
z#6n7e0HhNdfpmCz>M<>2gz)N@Vo7OC&d=8NdaSt>2c5Jf^FBd3C;BlCb=F~22VDgd
zOFq{)9)0}D<zBiH&N5sYWB^z1ictUQFvKFUzAsucTLj-7#a{9KVbK&q=jo4#-z|Px
z5jO7YqGm0jcW0v~q3fQ)2ruwYJZGUlIe`d|O;%e5_U7kCMmk7x03O5r3ls*?{5aeQ
zm+W)2UbJ_juyDP}2}u@Q8D=ww_la<0OvP`2l75cjGPbs5FP|HT24JzMTGRx!e*LEX
z8vi`+QAH`6ir7((JpiW(PLRc4Jn;_FZzbY5K1g}s(XY~ok?2~*ktJM`rEzv2h%Y`5
z9FWkGTtqJ*44R!{6|dLVW{gv+*vZ&UF~OZ+2rh_^bmcj!IdnIHIvB{E18*lhJ<@n@
z&sK6lk5AP6LH$)s4af@6laOXJU$Qmhf+C%<_J{_an{h^jmOR)#xD0xJq_ngP%S$R_
z84E<Di3NZVyfG1Bhfp68d|^qQOT0(iHyA6<e<tfSf%Tbd>M>=lPS(s}r$<rX#ciQ}
zel(x8RB2k;L+v!GHww6Opr=aZiIRJS{ri9_-wn*>^Ti;F0xI!wnUEY&>1O6a3Z(&l
z$+PqHhiR?gp=Y<hvGi~TNz8lpq7nsxrv^<16kZS-L*$C4T$GLOh*d#uf3^Ki24Zyz
z_&Upwxjb`o7dXG<2{FpUdXI$$Tz)9N<!4MhV#LQbMqU1`B`1Ik;|@Xkk2ghC_;U1b
zcbw(863Ex*7{L3hhrtXq-?NYh1TB9m$Jztr5S(e=st?N4pzT**7K2NxJXe{?l<@OV
z6gzu9xlOK|KrPa(yZFwoixXi!7oDw-=?3Sg4e!fn##n>`JSbwO44-8OD0TY_k$9H9
zU{KAYdd6P&lEcBIyOt@k%Xs8IhnbsB!f|Jv8D(G`8Nj}l9$yvlpHYI=)kl^Ki3G27
z?w*}-H*s-42|T`jIMsYTGC04oV^dZRlg_X;(tl?iDw||loUJ+$a((biosMj3Rfxpq
zP+gI1?pA*d4a}#vZ*vc7v3J$0%6><@PO8K#W`%e~;`1aEXcf4S(Z5$nx$qjDkCFC~
z*~?ml;Q+e53rBp~ULY#yP&BlXl<CxeoK<y|3N2(ajPY;v>+=?UR;e^mtqf;EbF#SQ
zme&#Gt>yz;g4%b7y~9}q-Gnnby5F5DbXtVzOE|e**zE*{p~B<KiiqlWmQeiP2euJ6
zWSZg*BX?0_&c1NBHkINFsp0P$ItcEMDxf~c4gjQ`pXoSK1Y$)3Fu_t6$431EhM{xP
z`?>$EE<izB2<@*UW$q^kF39_qeMbUrH9D!ZJyzhZW<#GbPU;FsQKU7{vJJ$-3dHW^
z4JafU!4r{W{Lr<Qz;u>lLH37^g$WfT&t>PGt~|Qs3PqUi69I*pfMJNZ;N3gYJjfvG
zPX&w#(g`nrK<52Y)=C(e{4ah~GCZhSCUz`EJ^<0m2&Tu{Q)`X2Qc!O|gW97f;~m=F
zpU6i_&BAli_<E1C%57I*O0W4MMec}_;Y3GX8A0%&mp)8d=r}l#6v7h*(G96tDvOB<
zUQ9lGrRJk{;s*q3q+MrI$s<(jb6v$pp)QbO1)KO^LdtvaxgvF+$B0u`;}3oCE+GPq
zO-`QcEi8iS9)1wS#d}ZBi?*+#{kPHWc>B~!w>|^XxKl2KC}aT=%)dxzW#gHKYaF8%
zR<?Z<QnW9NFRx>SG&0$|lDO)ezFHRxxPWWmwb7!I#hslH$%k+21lOrLVwWHR&0nBd
zs>1xmQ+i6AL$c6gWj*Hh>9`vd_mYmtnU@(dJ3|#-k_#z>Gbi9yfAUt8WI==Px$dd~
zsiZk7VeH0p{c9-iVvq*n8z|S!f31l&h)`MN5p@WkCoc&4Pn0>i(X9mo4)q^KHgeza
zHCW+l`3zs(;3eB<HfwX!&izTM0|DT)d&Te{l#?BwHj>#rRL$g7aY5<2A?kkyzLjN)
zx^oqbWDG?g?PL`TPbWw6y3)3}ah|1)6^Pxpwm3fmO){vdU^uaZn}qM^dY?eR|MmU)
zy!aIQ?ddPvvD>Dcx*A4%Hra&CpllF~XK;;z|G_|@@@-NMdto;JX_rk|WeN_+dhRYX
zjs&D6=hyfu)N0zm+6}~H3HzLC^VAwqBzRt5f!wL~(fnQ7Wh)^P#^yZzs51dOJ34y3
zvLJ9?Q_$e1uRHV_shq))6S}r8#gv}s<|u`*v`~I|;~6y^+0UV{BMU#-g|BGi<X(Bl
zBG&#HFUp)yTwh>theiMoEn^CN-JQPTGM=}gGZ$BQe)%208l`V%plyUOhvdCvmke~;
zM2&@|6VTu&nLA_jXEAJXiRq(W-nYDa96Bx-?)8YjApSxqDLK$A+vD3szjNJa-LP{=
zj-DS<;Nr*cDbgzOVi#H2pSq@gANe6ScO&o}Y2KQzUtYqw0}wlIwH^WvH!zdZn?~2s
zsXCEl#Tv}i{h0vKiUH-cgmt76zhRgA+mJXO1huW9;|_>g9N3V%#lGc`pdJS-0?7r-
zI}O^7pp8fy0G`Fjg}*B8yo=s5K^mX656%W+Uk2NXBkYQkQiuvxgc{wFr_sHCHr$}@
zaJj78zT)^Y!mO}W91A#t;E52J(Lkp{&gAO;9aepGt-6bq$w&REO+c?l22-ln2?yUG
z=4>-$JG6D7s7wA}rj_DmEk5inr5r3DdQ*I!fPE#Ve1dS;%d?T>`Nk%HwDm=c#wUFK
zt4*?QKCnLLt^7ufbz}B1;7faa$fd~duibfPQExy8uECw)Q5Fy)VGHabj~)S$g6g#(
z=Oul<H`QcjMY%Raqxl>jj&c+PJ*c{sS}5<xgc-qc$$PUg`)QCLNiGbK5ALpR&8!2k
z-?9CRP&M)9cs`;ps^;HJib+vh9d!2b2(xAuLJTZaVxW#>f>qqfA9F*LM0I>Y?-CmE
z$adlXbnEF={{do|gt}?lDkg-r={NP>T7P&a`>o%s*=a(j&uw@f^*+4#l$AJ@3&I~o
zq>tqpEGw9b7h4C3Jk|<O9lYGs>4XI~;~dA{Bnn|zRIwj+R(bG1owb`K%?Xq8gtnf;
zD~V5y|07((uj4yym%`<O4vo^3?LHf5mWf6#KOv(LU<0DWz5VQ-;gIpkjiUIwsc0=?
z|57fU2gwRP_wFhBDgyrO`z?(AnZ8sWsI{cB{f4bqN7+e}q#e$8Hi!;4H4e7}H@ycv
z;T7gpz5)^Y&C^yU0>)_r@MK&dxs6np+|GVil?VNg75_RIl;4*<0+5(rIOqAAT%^UN
z6Z{L$i~(ch7?WT$vM+m|J<fV-61~od(^vAm*4M)a#^VmDhJE5Jb~H!}yIM3V-W5&p
z#}k%X<(pnhwWU2!>d|FxNyc=Am~hQ74Qxa~@nmTpwyO!CX0-xOJ7W%l{Px==7-&?>
z-Y8CwoMiE2km<4!z;JJrt)*DHZ^Z1S-%5LT1AO6Z>*gx(+&J?kr_^^kttf}Xp6@}a
z^T`wXRyR{4lnE#v((NrY5)fuct?Xu07fRvG>-31+XBBe@-+9f@#nuG;iTqVBWxzmF
zH23$8!u_Fzo|TA_Q|oa4Nt?hhskGFdX3ypqrztegHme1Pof*GbJ(Z#eo1T8;_niaf
z26!LjshfK9yeP|gPlR^u_Ovi{1+OHNuT0(-HToox35@%)fs6j#n$PyN4P^(x3zn@9
zrOJ`oL(P4S7>Vm}zCIOHrWok-q~?iQw0{p>RXe6kCD=aLFe2z}U!F#n52j^3ev1|s
zc;aeUxOA1)W*AbnI*<WY$D3XO8C!cX3G}7uH@KK!Vgdn1oMT0=ZjFjN(|`*JUSAiI
z3QvMTX%dCW8~p5l$*@ISxjan&Je+%Nyq+~5mj)hIWi{A*_75OndEszZj2e?tjSglT
zrkosb8EWYy1EA4lWCWxHDSb99caAhVm`^PGWTrLF!Q4L%e?t{%Dp8^bq0>dm1Ng$X
zDsI-*w^Z}6E#2qZuTcLwDy7U#bXmR4<s$fM4to&p-2n@~*xqb^X^q4x+0A*@JbhoP
ziO*mS^N55O#!MS&`s!m$?K(Bq8kU11!}kCP?^mSCIP)tPm<-(BlmD`n@zVN+zZfyt
zqTd@9?2zbpMx7vM)Ea{2E7TR8fL&z@4_uAp?DCI`R&{gAOAzQeex#1qFm|qv9B3M^
zoLxhIZdu?l#n*0YS5IuL9ap{ksTOgl?jJi6X0vHp6Xa+Ui|@Z1KZUm~4U`@Bzp~L%
z*_|X>8=&$Q`&_y$XvZ@=Ys?>o*-a*TaQG+W@KA93PxL=TrXqO3iQxj4fzwtSeTee)
z0(tts@l|1!b!FupO~zOWVvx$YPYS?)H75TBS)9}fW`_)bja_i~phZlq3eFIX)(f>;
z1UC)(fywE?KMm}+)r>wWQM7d6{AS0{KB`3&mKlH5vm(72`z71FHFy?Fi!fj#IZE}=
zkId`i#JEYZy&Uh8OB3p|0j$iN53!>8M9g5>CYz68Nn@j*{kd)AJd)+G?kYoJW`Ma4
z_Aeh+YE@~kemy*%pu{nD*M!~%!0De6acR%x$GZ`7&MU*W;gIy$PYv~L8%qVAkn#$g
zuK4t^Ha#&=-z?*%33BJzjvOWg2DWnw?w<ILj(QkQJBux-nZp8TfN;(~7iA*h%+Su4
zY5M0*?w>wrK0^!u8lkC;*}pt$UP-u|hv%|hzq9+iTF6|Ke@^H!3FHAqs4#dqx$y)I
zRE3emT^v=lSMZLa!zD~GB!m*yg#uM3k3{6pgKwk_GP^5U#2&o?7j5;#BC&*%<k!->
z4jttTTJP56jpiEKfZO12*(E&J2`n66lHj|junf1qrf3-DS^jVIHgPZZa?o$WCh&7^
z&Goo2TY{XGX*U^r$?A6`_=i8jAZKGx1^RVg=;{Ro_yZ4q2%Q@8OQT@V{9%EKg8#k_
z2c(55`<51o6IzWrSlO_oidCeny+vp;@n7|mQGZzCspPG%0Kxrx*cK6<-P@nMn~8il
zHHs<8Jc;s11uMQucCE^}`4vYI#kyqlLBi5uEWiF0XuMl~5u^SybY`QXK^~P;EYb#F
zBktOc70|0`Ta}-4c7*7iwxnLy7B2C3&IsODhWC6s%0$E;adfPk$2?(8q(8qWAu611
zF+!^)_%3<W2h4p#DvQ(_^+CiiucFUeMJ2bHDr7{cFm)qSf3KTMc>)NhBTDyuA)e1j
zlI`ciMB2Y!dZjQxv0ob?1<0nuCg#2@nfI<SGtQlR^VC!?rO7A9WC0^JF%aRj8ufo-
zHzt3klkxhQd$X61dYR+OrHY+JUecZ)LHqR(8_3}_05z;I+Y(Uu6$OV_fuoof^~!lR
zufz|Z9MhGh&k<+ywtd>L@bLb<BP0xp4+W&bl$cPyXXc{^iR8!YO1qM11S*llA>|R_
zbqkicUxEpN)40w9Hv0PLTCJ#O65I1;vB&+81H&3{bg+Rm@J<f{eGw%xAx&NM2PT+0
zCFe+3zy#aH0c#M6%KLSflkSI|*qdFPDdQR!NbNc`<V`k8X*;iarZ;TRW4tN%``|_4
z!hed5t;G8BQUn)xEJMFHNk7qawLb;<z&dD|-PIaAC*`-qrCmBjkzJ9zddBg;q>esq
zyf$}9&EA?9K96A*?404{Qy|h|3)N4E$}YjN0MR*?xcW)`eJ$8J_KJPh%DwgQP;$pB
zCzL07X>l|amu6V1xqP_4IRn+eRQ?dve~wrkGMxpc({&s$L+g+%7S>8jsGBXSFB&i`
zj7Qv4|Gn8n?%5JObl->w300LVznJY+T!s<D6{#^<u9r)SxITqL%S*fj+`T%Jc)0W2
z0{70<zIWpy`_k=HvZL=5R3?nD5>qwPe>&4>z_$!K3$t==PmOAL3yA32{>%EMUGO-i
z&$)-P00$P9>{ij`4!N^3tS1olN@%Acc0!cFJIH;R$3A-7$69sixj}L{nNI$=Nr(0b
zpRuz}(Urgju4cT+5Q-t6RrA65i&wJi+y4;ThKyrVH`K7f_Pfq*-ZxjKnN#igk;t_Y
zWT`Hs+H`xPwjp&=@82lWK=A1SoFyf?)Z!^4?)u-2n*L1{K=aKQf*=l4=VR~3i;-Mg
zl&(yZ-1<;clI6UYMZqZQ#Zs95qDN8WvTlc{{V6=2IbsCyhg0wVf3f7nu^E}DmO!ZO
z`W52NT4vkTQ|AnUn%aX(W&g1}<t849vqB#R&6UkgC^kfmo$_(2kTx#<m&99Ub`Hge
zT0F;@plL7km4we4MH+Y=Tke+v!GG}r^yD0#O#czsE<@}70;$qz>$QXwrb_9eH|mNE
zrsW=o;&L!<;3=P`_VoVawoEsa+6lZ+UzH`?VPS1$gc9QDU$gZ@>ZCxL7>90pR1v4j
z=j&5}Hvaop_SwRMyk$XgM^FuC@R(*3Y*joJ$k}j^u7gO7eK8DsHZK)xz0!c(!AYfM
zLX!1)oOpsWxzLiGkH|cgId!jXXskg0KLA-kroYMc5Xt+L(Wu}ikCAD8w#qkmWOzhQ
ze-s>$5JE_$^+pc{9Y4F_5EU4PE2{ZC09|NcHX=_WBS-JCz5Cq0l3Nt43?xpDehdZR
zZ`Nps@rJ)H#9?dFzZb*YtQp%PNKMzX(QECPyJ7Mt?0E|94y<BNHt3;&*r;jYN|d|m
z_DyCZkA=~)5F}=mcha#-C<2hLemiN~e~0w;+jq9Op~GYa6!DjrI|op^BhZr!*Fhzh
zcXkXU#n*BlG_*mZQz&Ykz;f*jp8xiOIW!O7=N)cuvFFYB1)d5cKLq|r@rGAe-3sLf
zINy*}{we?1VN^-J3CMNQ4LdK7F#vaT_l?XMA79I1u}ir;<ZsPZ@Ut5@I$VQTe+caa
zZHR5fjp(U2^rX2TIJd4+`h=9lvZ!lQ#l3r0|D*-oMLH3gXgh!V_=bqW00o@cjt4{E
z$d4>;6g>}xb1ti2*C5cYE-?y_zQhzC@mOEy^itlphi-o%lJ6ckdzka{UXh8T?WeBZ
zUug5kY#V&Vx5j@OiW6Z*KVo6Ie@msJlCVxJv_Oy`cM8{*z28`g%_s|OHFBI_O`f}i
z!<h^ym;dSQ_w_t<x=_{C%6JPs7Q)kldw67keTNb`EKL()&Vma?RLSKWX+mD$g~>a%
zH9DiCQyjl@<u+~#cg^yf-_FN9s`VNyDyG^cBLNufYMWs@3bu1{$T{m3fBLUF0<klf
zUnH|JzO9^vreMh@&4OM>VN7kwRijMn5bVQ5SsEauP7ktt!UFnU8X9^|^_+-Q42;*R
zT>Ev@0;}%)?KjvSH+TfO$VI62-t>9`I@sao>)Fu)OyQ>P_%Cn)s8^l-1WIg^kS8{l
ze5;$p+x49R+59^EvaZYle^{0Y;k2v<Wjw&^U$(&JSK{}6BmvFd6aI_DrSxd<flO|S
zZ`aYQ)va0ri8iWV-I(()C>$pg{4jGI!Ke77yROn>ao@Y`vaESwBd^)`YBFb9x-Dr*
zx4F2*PEVnQ?V052F=4Aocc50KTgNQ21o-^T(p@;F(2S@z%4uMxf1(q9SN=V$R*GPN
z)jjXW_Un@;Y$3Dl9$tYwgU6lF8tlRAZy;y+Dtc0KI0scR>AqD0L%r&f(Dfti!lX{U
zjOTcnQ5L-lPME=`mjCERrs|uFLO8}bs!P)1D@?v+DD8W#yu5nF-?{B{d0jhB<{lqS
zT|^7tGwXEnuIL(yf1MNfxDX@ot<d-@R#K8kBL*fk?`Edm<rBurAyboz4GM}H{e6n7
zK^lI>h^iK4?3m1qu%?E*9Dk$zWl?Dl9984qJIFBt;gm=`Xqn#P;~#jSU-2;gbEg^X
zsOdkuewa)r#RnK17iQYQsR^OO{FE{&)#rBnbNNwBDA4~re~t$Ics9bcc{WrmjVlWz
z38P~P-pH3QXd#ag_|7~{DjJHW!4Yzvts&4;u6m>nawx$D+JW^O+LFs(J{`4Y;oKVW
z2q1$|$wM2D_w=EaR}{sc!u_dkQ4>*Re{*2FE3uxJz_OPOTOxO0`RSe#zy)v9<B5XD
zT4PyrOq&>~e{0>AvoMt9+v5Gm+-p!f!`7uhlu9zIRM~23Dov(R{Bq<}d6}B;OkI!8
zVNm{J4<3oVK5L2Mgk}r><=12jM)MAl35;lNx#_4*nFiGJWqCaZ3qNzVfkR3-16+&u
zWY8+u2TE5hR7)X4&)rvHJz0@$Ibxl;aP5Z1?Vi?;f7t77Q(#&|lOdT%O1#}fSZUG_
zuEmY^^r)y%fHaXX^|X5EhjN^mbkR?B1~iCJ_g<AQMN`4cq_I?;Mp>NN1q8k|IpKp9
zY<<q$TEu{sk!5`$+PgH}*pyR41^7@#<~-&C>k|0vbWF4oTrA@~V$Z3doj`qj6g<c!
zAIj_Qe_{<x{b<{UKd%C3E=n1!r4T;9Ep@QzT=kbjqq(mQYQD+q{j`loeoL<z6y>Os
z86pCeFP<C~l{hEt^2X}y?Jt($+xqjH_f2m{pQ#1hi^1<al60o;c~<!k->8WrFxTHn
z{n}4gl6k=qlQW7IMfAZFAor<Pb|uN+{usl(e`nPdFRY3Qz<VhikcrJw*!2A(c<K@N
zYF$MVqr;PZpW)_G-RCvZz4ixr_3u?%Siv$PM;&*F6<URL!03~@_k(T)-<`PQ==2xX
z8)Jdo2_cCZ&6A63g}L1qG`1g9Acg6E=dcm&;K_Ho`w(`y)T#HE*WT>4V}Q>0%lK;(
zf7y^7%MxxI3(#;Y+~zay&O#|T<lDE!w2xSiT^q|4scMhkTGXge)&SKG3Usj$h2My2
zfh!DrKJ%LpC`w0%TDVq7i+Nww&&7EGoo;r_5BaxhaqI5HURf>p4EvtDMaSL^E~(j3
zNDB$#OFkX?SYenU%k_OH-=S4Klg-44f4j%`w`F95cib2K$)2M?1OicLW!JxPpi7oV
z`%<IVi^}X%!ik-__KBD2MKI15{hK+{MA(Law+Ba0~}gWKE>we{BXC$9_`;7V0K
z!px*8_1q-D%<YR>fnVGnH)ZlJdofNdNiX*`Oy<%Ni)P@eNZM=GVVF1>e67<kf3Vz?
zEf5If^tI>%7KIeByJ*rCSVvbNE#QT5P5I?#%G}dIxvRr?P!>h~{7NtCyFUZK-#&Dx
zqYDvv8n3mWx_yE8+Oky#?u8n>&U)wXYPxIlT6<>5Sq+Kb0~XT0qnNf}hGxI|>)+Ag
z7O!=ntejQRC>(@*qe6*R9*t?If5*T|l(6SZ{=+QGXEZyu%;BO9y~Q~X=@Ib*a5UUO
zycLkSM_v@r?_Egj9n5W3rB~!ugf;h?QjOmeqMWh6{qkc$2n`e3v-F2`k8*PdE~Ieu
z<_g3Zod_}n%GAM1KP6MudxyB}4k-(<8DX~j1}Xd`Zr(HX7lp90jThx>f9rg<qgzd4
z3(GRgciI9vCy1)t>T@ruXaxTYFk2zH(@-PPW3x^AL~19B)D0}m<d!JefW>2UFj3bN
z-F&;|_n%@gPZ3l?)E&CeiIDskblj?C!z>G01qTG9`*Q0i=Tzk9Kyj@tH-d)xU9BLH
z+6vO>#QaT~{naxY<=K5Pf3wtsSKfn3iQkV%7x|uPsWy={<D-Fzx^s78^zqU-Yytj$
zR6;Oy<_+~^=l1VWmznQPr7K+%&^IJQchcnSY*pC#p?iUXi;gl%#>T9Ba1hcMv<7DJ
zwVV)W*nZU5DX_F1>~T16F3&Py+;VVnTTJxf*03l66V%CYCINIse?LaLq%8QJ<Zd5F
zS5_HyiSs6#@M!zvV4HAfdt4V{AIZ2;ICHr(0!Dkk58Of(mLpeAWvs~m(i|mc*v@dE
zwC%+8B)vp|fAlva%b0qSQ;fxk5k6V}PRUb~1&tce0yh=exr*oz>#lw=M+{59q4omD
z1+8e=Vi=f{jMKA|e+MJA$IRZ!og^ri#SIoep+sor@N4K&hmtx`0ORI6d}l_FIupZG
zOT+b(gUS3bPZmpG96(?ipSM*q*B0?!I8_OHg0=E@aAafQq6d{P6|W(o_|~T9H{=rR
z+X?Y|=!D4KnLL-cQOLX(wKbmmUM2^zl^*x6u-96VKE!_&f5H-#*jmvC5PTGf-|XRE
z9vwm@-JG2_@O3Q=UW%i@Os^9X3s8Yie)oYEgdeX4?sCJop&Gj!v6Da4-{ZtM*}~)l
zS#zhyGRbHf>O%|LU%uMox?(KL2SLgIIvCN}Gr3ucRs}bd#>2WFM~vI=BUX@mF6Zxw
zf<6BDtpMW4e|#-F647I6SXLOnD15MlU-+|D<TX$H%`hvy;Th1jW5H;7uN2N;9JHk@
zfc5w^&qzmL;D9~33ib|+aYv>UEt}G06mT^66Sg?Ir-HF>rc-=4-ZL$#ido1jq$6Oa
z{Mz+y|3sPCB@>zNsdQys-pEZBgRJDX=c7e!e;kPKe?KnnUaAw!Se1l&jp3#z2{MUS
zDRhK;DQ#);QZy|s-Z{WfS~(GSOh<^n1Pz^jsm7^0S))1OB4f=u*m>;;n^e~y;aFK|
zf=x%_uz*7%PVkflXF;BTaFgn4OX^eRZwn@umJY?eP>>uh-wZUGroV($#ZF#Nc|*BA
zL^1hUf77E0Iw9SE%p;hactmo8&qdiwxD97UZpQf;4aTUtHx=-VjAbK_@Jv0txkLS-
zt>Ag1^x9v*eoni;RINI6h$srEf5mY0la=1G*PK+e@zVfsc1ZYoEV}UBVy<@QbeQEi
z-eheDUx0CKd{5>*P%c3_Zn$q$s41O`B;7}wfAe2iE2gbdPJ@t-DyWMmL`f+0JW;F!
zD4p9jS@{5CB>o(~q|)JCBDemcR`+aa2YwrJ)Ox4{r`$QMMdTl}v-)5Z(cBnTYJDy@
zh(1z%JF<{gGAZlFXHU^1k~Hv;*B==`A|w6ju3LUn4c-NVJnuLG(dV7M`p$_;5hdmF
ze~9`+DE#fPLU$Wr#XH#+zn^EQ*$O{yWzagA>XgwyA$HAU&+Y2tdySV#a~?WV>n`_q
ziie??Zqo5IYhYa@*OpI1B}}s5clN&84(VQPyocbZIfe%ntmVYz&_`eb%=@2rvv?Bo
zR4ax4cMbf~gSY-5A8qXO<|43$g)l}Ee_uU|9vsO=O#*#B)87@so3|mSWOs-#JHsFS
zxM;Pc^U_BS#Ixh^RPA<@Dz&=~WhdK3sbt0*R$VXg*|~UXVjn|@kyh`niYy^*(Ctrr
zjm1TLjE0ByKAazq6CJCeGZN@%slsotmnQ6Sys^K<(s2yMV^<Pp2o-Y@AI|0cf5jB+
zC@(B^`U308*;}1{!Me=gG`+QanUNIp#gJgT^Da(CERjJC&o6Hb?L$>2D&}@*FE5A2
z@J9Chm7N;t6p7U8Wn!TddY?>FeuJnekDE&Po^{b|XcIl5lmS;ywI60nLx{kiVenNV
z8n;IO%Xco}X>8B)_r;#{zhEY_e?QE+O+oamrTWEOiIJ%B-#5Mr=$k4k3O$M+Lo$rE
z4Vil6N3+xc_vg!!uFP}HW>a<8HMEn}2(qSs)7$u|4Ve2hK2!A37DLqVW#p^bUupsc
zL!Vm7Ut51v>4j?_I$sj!tU>;{S|h)px+lhtWp+t!qs!md)1-vO8(^}Mf1s<-S&)u1
z11sUAq!6}j;+m4?CnKNYrj79LH@TMd+1Zn(#$JtZLt|!m9NJB5&D!_7?3y!9o0Q4a
zv61U6PbJrW=My}9h7kA?rYMBI<+gDIh;{N*xGI3cPj%D-<sS-%<Q(w5_^=^s{MNnn
zn~U=WphLb^)py{WFY9dqe;K6po$YlHWv=LbklziK2+9xzip+@Yc+hM5jIo4OyGb2j
zjtpF6tPvs1br>(p>4mO(YrSb0`vmj+Y+V?!!@_MGQXK+~(^!ekzlzcML*Nfu&Q<MK
z%})*#ko`F$b&FmPPlauZnX{=&2EW4RN$wzYRT2kE-Us13gA~V-e~;AgO=h><)U_+p
ze(R>8XQ^JXPLTV348-}($ao4gPJ$6OT%6PZtDIT1`+HSQ8`KYoW34xqFv27X--M>r
z$CAd&aODV8(smC8y}%K2V7X~W8GY}0C+@}k0thi`LiI1v^F4`%^*wuV5W7mCOy#RV
z!QwtlUiBU8y*T2Uf18J9ySp~oJC{kG;*V3ykSyw}VYrD?UvO~YcVBHja$&f+vrdXP
zl`WR*yEn1;yMHrX8a+ncl8YR61gY^*j-S29f&Bq!$Es(zRWSt4`;(@J(Yn$6v1wWd
zge^|eJ3VG-XKUZOVs<ao1sC+;PtK5&x%z+Ql_!MY*Tcy9f80ISk@+tL63dy;GnHgB
zq;Ujxx(7M0zmqDo1w5x&uwlGL5pQ|bzgdEfa#fGj<z_>i2NqF5vZdrYDiogt@0kou
zio1XDoZw>Oq+^!<_Rd`R&`O#{+LF9Ll}ANR<zyeU!NR84Ba42>$Bwk=`mDMPkC2;h
z*g<#F`J|X2e=SV1%@-OYEr@K5R7!R4O?Ho76D5)MBqFS>;P?ZG{Pgy~Q~u{-BhR{h
z(i#pV7jZHnqZAh~FtX}NO)7l$z7G(1wSK<q5P){!F@R68IBOhAPw;fWlbwLqH4@DO
z*DsP{5Yzrk*SC<(GkF{~&uFnR@H-50&n36?pm&U$e+-<N`zN_5Ls8=hbtxU0wBWV-
zaR8rau?8}u^^Tcy=Ebr)!=M)RH#i)kpNJ=s!Ml{c@jPuCm8y>WQ}x8c{?I-gl=@0r
z*qXP6suHv*a4fjLG*T;nnKhnxZXLS8ys0rPKg3_e@UkM|mRaoQl(r(0Csf(tA5T#?
z2im;Te<Py13%d)6ep$?0DziOJby{;#PfQiZj?<<$7K92!P@-Wtc8$O>LxHA+3MP1B
zze5b5F%sx5Fx<4`!DQY@2kVgpC+7V%XZi!f2Y(skja{@>ZH5V-!1kb=iC^1qH#(_>
zzulF*`VH1Et>$8@5!=oxozgf*0CXV@&Xgfhe>`1aGOE+U5p7rAV5L@qaLlH}q6fjh
zBJ3!doasZIF2H_O6w1grJQtkGJ_^6Szq>!VY@D{<gP&A&!1%H-YbMUk$QZ0=fIH`R
zq+SR+Au+pKAIfU}h~nkd+WWHtaYdZod4GkRR$R{lVYQ8qqcA3|C!w*7m$fLUry-oh
ze?&!7e00J>mfbN$=68N7eSCwLl{S1piSTn7w0K`o@0~8PaeaZkQ<u{L%Jt9P9{{AD
z$s(r(Q%<V-UW6{D5RCI*6>lG`^x^!*=f*w5!DhO@VS{r`p@y{hzSTEL(PUMXB$Q9I
zWgPoR<P}=h!e72PGf!os!Gy5=4NRytf6luYry51-B->=9Wh;O_9wF$i9I;Fs+@HPP
zeubj7u3zS;<Zt%Z<kNL)v^N{0_~=o0&QXDrWv(#E#Zm1$8LJowN6K@M<7TFeowwFH
z8K)ER`(6STh*c{~eAZi<gRI5q+xK(p8cH^F-cOc<QKDM^^yO)|!5e&9L2Oexf1+43
z>B$_Y5|NeyKZJ;3kAXt0%?f6(?a%nT)yAd*?tu0|4%Eo#iz3Uxmqp2r4h}l#X$7HZ
zxR5QY%a4Yo=-x}K*FBk*T!OLZ-nk{KZd0I6H6K-Qh>=`XRnSnOyn)U#v6Jixds_nQ
z2@%3vdMB4rgr^3BdXMW%rBzasf8LQVp?IdGk=2}!Wqwec0$ZRO6;&JoEa^e@oKck&
z-bUVVghi+EDuASJpAbnm;30aiZ=laUJp*#{H_HBej?w6${s}>z@ZIenQuNhoX{A{F
zomrCd?(+iGr6#P!m7qeCRWX#oTa;8>w|0rN)EW=+xctbmqGHl#7Wse~e_TTW?4Y17
znzteI;J9}>+igjGG(eW~Y`(w5*7GAl<O<dBGLoNR$?f1K<I*403X<|Cayzvw7t3_O
zl=%@c+^TRx+g2Y3_5(-<-av<NnQi~Cm+=g}lU%&>6IQ-4G*{Bl+0~Lt(YYowV!kx3
zNU^`u*CMvx80~qwCP8?`f3Ce%{1i(Ns8yLm$p*g40*KlDaLrs{p2Rc8V>o#kxzq4u
z!EKCd-o*sgqR?qE8Jq=X1KNH3B(;cnl*O%(<L5=AA^w$&W$H=sm5p_?*1Sh(1GZJU
zQN9TS8p}Y|CX8^m{4|7Af9xsm0zar}k5V79)F5>7L1INxczM(Ee~WQa3Y7KIzj5F@
z&~BdD;*Af{Heb0TI*`;RHJTDsch9CxR^p1!^ApxjaY`2^$`b4Iw;yATO4B{rW|AL4
zYaWS;XhCD3@&3)rli=MAz@sS)^=?m8wUi47$sN%wJ+CvGn_?`dVs?wCAS=onC1K#G
zl9;a}K3$AevDsiif2PEcZ>-!{Fpy1siW}3BYHFeE<=V<^hb>XW`v!@H6ye#@sJ%8k
zRQ|Uc&U9BB5+u-Oblk2r#Pvz2-R#Z547Z5%;-rfer-=^#y-qB8MDr=HMUnv`^LjpJ
zeK2pNKl&;K7d0&dQ(S{Su63XyBwixb_YK?|`hhfj7hh{;e=EW<iB76;K4<gRiyaPL
z!lPgm)|-walQPi~-6E+%Bv}QAjeu;^XZ3=~U^}g5ElhNlVY>eaN~3Ye{duch_@|F!
zA?1mlk(Ea~?vj*JCGXuKJfR(NYgLxJ^s#cq{4m8S3zLTtA2E)jDXK1mm4nOpaS;x8
zZajM|KyE^Te=r=XOSvQ=tws=`g6s>zi?bwySYUEAJZ8wAIp_?!1|4iPfS3;T$oLEJ
z;TtD|!;st<d)HQ=BhpMAzc5C=^KoI7*RX`+i|X$7hE)S5{tQD)vg*q2x_J{EpZ&6B
z%a2>wz5ZhejvS~3T33&?jFHS$C`->DewobV=%O~De?FO_msQyy*X`6{>l!1zL*_AX
zvwD|vn^oetj2AcDU!zAJ4hR@!FQ85p9ExyZ@Mj@c_wI6Jd^IKCXDj^iR#aX(wRYI0
z6ahiZ@p%3nI)k_<p&py^j-qq-X)wHY3^(+B4#zDJ#x~%3ous>oG*O_3hF14i>^@B-
zo2%sXe;CBmkYHyM(%*o0(a3BDJ8xH6$v^k>0x=&0M77|D{#X0d-NvpOZ#7QpvmT)M
z#sK38953hZEC$@owp{l2Qy7(5Eesj*odkV{!mPA0Il7tVeGh3Kb??<xkq;QvLnFX2
z&+0_}w5SSIm#CoaWh9}a<yC+@^v>}luD<P-e=NXbFum{QT36zi3R~~5&L;ho^EA_R
zai#~8skBu_mcn}#s=~OiIh%&(USu$Q<L!r)Xxb9#2y!dGdYyT83KRhAt_hnLLlOU=
zdwv$(!b1{oGd}TyVY%0pn;jJyWg1798@R?F+m2Vs?~USYwal)eD?@Qf6I$Py9gG9U
ze<ij|l58(Gbld}@Zln7>0qg5Jf5yF{Bqo-_`-s7uteo>!>rg=;{c90}Ok0IV8Z56&
ziX`nt85o64I?lJUuMdCIHTxusuBvS)l1vA|YyY;B`{jE2E2j$ua{hhgdU!pDRqG~~
zfn7*9z+xas{vAW+YF7#*zq#FXkDVo)fAKXi=J`O#gOO<=J4ev!k(}_)v}9LuLBL6E
zI~8w5@&34q+R%OPMQiCx8xNkqJ5DoT%VZxnWhi}1+c|cu;X;L^?VoG*I5h!coW@+L
zt}&u?xTjjE(1(}fvZ?TvM`@Dv@}4ShZL{#qV$KtZ?%4ii5i&n%e<)Ju*K)5Ef5hkI
z_vQ=2N!1IyZ&z0Q;Bogm`|C>-g`Z_<=&=O<u9U4p5%i#RX6wuUvgcjLT$P6}mu`=}
z{k?>mSuWIBV>%z)Ik-KLU|A%>ER|E^WfK|ul69nu(mR6{<@%t7?4ibp#GL-TL^hk9
zL81h<C0VfiHZkfN$D9K~OtBkzfA355V(7k?CwXEU6xL}!hgGJ`ue)q|9^MeQEGZiZ
z2UuD#V|Mi`Q@((UpRx*Cf_ostY6-@eC*Ea?mOu8KGv!bkx$4s4*5Z~X-FH^p(Ecna
zdgZWteu&@7;wFwT?^v==Y_co+aqls;aszqlGs(KNVXs?TK<#A!L%(mmfAujfhzkiv
z$3Fb&_#S?LyNPn{Y9yXi2Lr|3>-ZvAb*K`BHG<?ra^S@1P$@dUxwK}Dfn&!dBtG7)
z1x#jpNe_3NPm+)0KHSutHYtP-b||Z`bc5i3UEbMJL9~~zpsPc#oK0)&4<x5d0;!)k
z>NR5!+G}FHM%agzqFw(V1dnf!)t5n|9TS(Yx*R0~HZU@mAs`wQ1u`%+G&q-W2^%Mu
z7rPt{f2`QHZQrqN+qR90ZL5-0tcq=`k_x_h-nUPm9_Q=+(|vwiV_bV-F3h#p7@LAb
zNtHp^%)taG;b8B|z{1GP15mKCGjVfKaj;Wx;Fe=h0a~~L{>s_mC@4gofyS;@4)$Wk
zu0S4uCeREZ1~dh*vI1DRxw+ve0HO|#Ud~n)f0nKQYIPM&8ag`q|CIb=0x<FV56xds
z7b^>U0Oj8wcc86<qaD!R^)JN#wWBH!2ynFo0?e&!fdElOC2eU1NdUE^f;vDFXb*HY
zwgo7;nb=yH0_3brf%Yyy8i2WjGr;!W2Y{)Ay_wZNrMWQv#S(S_7z11!fu>e}-GH8^
ze?Z57RP+ExptGHoi_71CfRzit!r9o~^=}Vc9RODLrnYWo|0M8NZtn1}gdCk6{#w}m
z)%^uia&U2VF?F_bbOrp4sw5`y?>t>Cja~nV?PB#;4RA33YiZ_S>h@1b|I+=1`Kxs`
zwz78txB@+0|A}P+1ejU5INBO}{f+$>f6UR@>R)NNxmej-{AULA0B4|uv9p;i(8c91
z%wOz(y7QlP0{)lkjU63rz5eCx@Gq<XnS+(93((e_5sroRZ%k9yzp*W>?BSUHnH6b!
za|Zwm^S|Y0ZjS$<a|b&At3lNN%nZ%nB#g}*>}|aOW<YZ|CItuAzaatC|931i0sfbP
zejF2jUoP?g>wW&O&=PL8whG2}e+S^-I|lG~$r#%M{_Yuo9N?ctV{7dE|MM}nv$FO2
zKY9IEYfa$4)BQhyNxK^VZHutI#b0We8JYhrv~rQK@&uYGS-F~80?dtV|Mu)(a&>z%
zptG%&J@7BL|7t6MfrXj*zv$E~txRp~{~>{Y{og8}z1e@Q;4fYO%AZM8Nli^lneKnH
zVE<*P^mh<k)w~@4=Jr3Tk#{irpND^7L_{1s0X_^Y99#efRxXym_wR2)+^lSV|C^Bi
z0%G~kv%ImZvy~@6_ix?IEdQ$he|-OW(*G|o;`XKvX8(+is;jZR+23jVpMrnXrf$xE
z&VNb%*8=!k^?#iIHTOWEC(smbeZ|3)H`F>aC(E?~yI^cuOm}F21$->bv7%Y6v^cQN
zp+6uqa+kZt=)FFSvE?VvXQ$Wlx#P#74DHdx0G2IPe>(8Gh`^7GX0RA_i?W-O?kJea
zpaL=LLNnIqW=U=bLWhG{<7jVMS;e4#{2hw$XZJT}n41@xKr;8@Kn(I{M<jpKjHU$4
z9+Vakw1$%N3(Az!muISx$?>t#{yWIQLwvgY1Xv1QV{HGlbCOh7FC*v2X46l|H+N2#
z@5{OEnsult>K@+se^%Ux$)DnP(OI&tz7bh`34<Om3$2(W$a-8ygu}<PmL>~-&=da<
zomG=DXzW>$zF6b<b=f4qlWCK)#vN}3f7}wYBAH26%D(-HszEIzHL{eW6P{!yrNPXF
zlBKMz5yL;r$<~M`NFgqO6eH9SI9Z*#CsY<^(Cz9+1jFA4Qmd9Z+?aQv#4yHh1<nx1
za_kZP=)}h?B&2N}PJ0pi4E?5mQpFZ8xX*_NmcG@3B~m?O{e59<9THkFvo*dKwcbPa
z`heow8*T)u)RtWfF;A>?0mXywLz+Q*ziw||rZ{^r;-a9@?8#7zZ22CHSM!**RNoay
zr7WyKyh5QrTbj<r2hx&B{b)-Iy>85@-fj(kZkI@UPJcAVekw0v4&eEJh7D_<(p^x!
zDE$XwxM#Et(qjB*1Debv=#wFn1+(p$Oq~{N6M(l{Ck98?yy<sQw74Bug2~!mfZY4c
zCbZ}uFpU7P)EbY?!cNB+N(#pVSYLPdoaMyAf?7<ngI=QLk5IRPR-ZIbZ0-3nJ;&xO
z$IV%HsN9Y8M9WeWn+el@Izr8K2vBsj8G&BqRvG*%&jNQ(HJj~cxm@Kb2cR5<W-PC)
zFI|7Fq#0R_)@XoKQ<+?zSS`C9id|56rD2$7ql*U0v%6=ABL>S;vC=nT5G6nx_oEhy
z81jSgRjKChT2ge#=x<h_W-2&_q51pDLQr64g$lp?Av)@>o|6@S847D(#SX98?7z7f
z+Iw1IvF*hQd)+5&SO^SS)4>W{Dwi6gFn{eW4Hk*p3)Dj5gH9^w=by=i={?|g9ElxA
z-0xQvtcB|=DRk0^A}wYQWE#lmaWP_5GuRLFayy)vv5;?S1#H7Eu34PjY)cP?Y>9n*
zYmhBHpScQZK3RN!8OYPnF+=;#{2m|F7fDhDj-#ZR1NQUNSTEj})VRXm<%x??s}2i9
zv~2smLK3N*E^(I<x%#Qf+gYt4i5AEEC(U>3ozh%C2lknolv-AG7ETV{;@i!x523e&
zHxJ2+oL0gaP#XLJmv5LX8=L2TO>)`bV3SeS1oU&bcxy3#-{u<F>DX;kGX2+$IRUeD
z2aRT9r6nBjj|2tSkt`;ET<VnNb2Tm0olgf}zoGw`Z%_q$^6Bd;K5@AIEbPb)Sw_W@
zRc#6CNk$f0&cmc@e6kP~3nWGyKOrdFRv@4D0{EKGwIzT05>UjPs72lv_AgHmjrKx@
zW!fR=i@ODXpZyeSm-LBv`%y`GaVOckK{y5-xjp7Im^B|aLth##p_<NDX%Hm_EvG~f
zBx<P&wRHClJm6d`2a%wqhWBR_L<#DvFl^|mT8UDD@>;gSi{M*80q^u0Vl^Dw?j9^!
zdz2%jkr#zqFy~w^Z}RtatHz?uyMo(r#n7she&2h4b3%OXpg*IBAH!ws@T)mL*~i}B
z;dFk0K#B4W3<979(zj$z9l%J9-ak<bmVkP^xA2FtlJJP(CnWVhFp!V?SwieI6Mh}$
zIa%0xT<p&!V(ry9t&Rmv3~1<m<+!Y70f_du@9*q@X3i$BM!{QpVsG3ji@k&>L)hmU
z?YvxnBC+m%>+P7#qqq3`3&sSmW^&XIPfTkVho<atj@v$CU7s_YK2oOqpwJCE3zD=e
zKw7*-1n%0ORu|hmwT+|izsl3!a`*&HD@J}zLLanFd2^yG>FF%<#jQEP%`|&jClxen
zOjD*!k-7<$-Lu!>myA8A1>8$O0g1FCJ|lB~*OK_R3wR+wb9YzC2cK1NDS8#xFclr-
z&5&PZ;Nijir3ABkQ6p37cOJv<irGGX*Iua8`TL_(Yx*xPYlFg_OF_fF5t}I<wf|uA
z($4mGjGN&t;jv*bLyqT`rU0fJ@cO#;#7aG!XB$bK8fRcw_lQ|UmP~L4zhul%#qYI$
z3z>LO3Sfhi+l(5&6&h`)ERbIbH1?}SCij>H+A1W47dN@jeJ=4ZleCtJGj9Vnf-IL5
za3!Zt-UugNlD)MFEiy-oTT#P?Ph{J^d(uANMwJ3&O}%Ksu_vDWxf0=qSd8I1y@U|o
z?j0wwyhgDWpy*Pw<C1#yV)5YaPJ}RjqH67fKs!<RTAvU{`8V&IY2QeKqt5PH_w)61
zXpt?*!nD}f$vvk~%LAb-4bXpg)~UJB<G@Qx-)3@mvNyg|iQO&rhL%CgYO#V%%Un*$
z#^b=+Q^b5Zs=157XPTWz+(m%166qT-#g^^Q`QU1LJr}N2u%J5NQI>}OLCj)**{S|~
z;uSu!2YF!RR-q&SgRJj%pndzg_8t7-ZrXzhCFb6Yd-X;3u>SRkLDu4{PAFA7ho1Rs
z-r`Xkae5fVT1PZoI@QE81lx;?r>ic|hZnqcW2w`Hq=11Cmm@Q3%lS~^GHs6&G@w81
z5lS)W-giSW3A>LxI!*^jYOq;<%oxJ|8ds{3SxS9XUnvweOvasN60TDn9Dboc#$Rj6
z-I~E=52>|)+9SSVk7QJ7n;wSS(G0oV<=?|p>)Q$<uB|4@Y#xi}<FT1x(pBGri#Ux}
zX8}VVtN?OLHtM=JfPKKBw6EB@RO^V>u`Y<bj$_eFLX&CnYJSGILUIXz&iV#XT%J@1
zkDW)j!Xa$_Dv8yWNVPWC^KMRB7`|~vl6f>nVU-8pU$wn!2T+%l57rWf0mK%|yTOi}
zGoKKRD1s`+uKUDa+_5q(H)4ULg{o)sLb~pY($qcZ)0ny0DP2b+|Jv(I!re!1rDx;3
zu?XO7l@24A>^2&Z>pNY4gLl9XnF)7qIO&n4mkL=xL)V=-e8;?bUI4Qyf0W=w8^T!6
zx4c6i9rI<N;F6E`$!{j~+8;4;CE?1X+JBpf*tm@rE!*nR%<%Fio(B~zJ@a18nd=P@
zM1w>JF)r7_lV`>eaQX~fb`*i(B0IR;urpO@>s1rx$+79GNv*JdL#s{s%6b(Q5E$!f
zww(>KoNben#msc^nno+I^~E;c(f1p%hfP1y3eIe%X&{5NTP~2jX-^(G6f8_z;eX7r
zr}$0Mnbm5&tBs4rX5lfUnLmTwS{Z-|`DkfVT)haP*Mcz(oH>h09(Mp#^~s2|`$PNW
zf}EEmNx}}l<6(J!`kQ9c2dtMM<|n)n>Lemg&Mpt*Ub1N-v}<)8l!&Zzl)z$cEMviv
zR^JP|4q$}XH#(D-2Mw<NIQc|`qUzPT4Gtg8rxg0JlmBhQoCDV{O$VA4BP#pESOddM
zWzMwJ=wq~!;r9G9do7wg(a^7bGcaRdz2TK5`y8&`yKbp}QZFSVRs3}RsH%6dRo;~q
z4&TfBCl{w(L({W{8EZ!sLg~AwNGBATk)U_+QSmKJkY+H0$OBu)>u=-+Igb8qq(&TH
ze9T%Xb8rGZSd*?s26Ff3^pqs#(bj?!cw<&1ag;uqT}5;<+n(|0gp^*?A;oD~aUmV_
za)VM|^giN$U}z1KVXT&I48^Fvw(8jqMMUaYi>1i7tGNEJ<i2e3z9rGK4MD>5Hl8zu
z%Nx-;LQ8Qi$|)YH5slTeO6jU#dJiQ5-2$+J6Hv<#qB6#Tp%F<esy~Jd;Jv?{ZKn5R
zqOs`KyD8C~Rl8!eaD`8_D}H!kFOkca2#W8Rhc0e^a^U}pY$+^cG*B<PX;@QkImGF)
zMj~NxE{t(YZ~RE9zC*qD2!#5jW?QroxxA4na|VnhWT}0=T*N`Z^&$u>+?rPl*a(F1
zD&INR_##5h;UgT59C&apl|)Mub*V9*7J;XqfJ+{F10_rv5$<YwlJ7p3V8pS9!ftv<
zCdMm&;h&)=FqrNc_8WyC%0vDGeyHkP_sJ`^rX|Gw(66&Drd@7W<fv>*h3`zu1k+27
z?}=3x>8bJg<eZG1713nV|2RunK10`#yu)OD03~nSdD1sOFQ6)$g@sZ)U{N#(eQ5(1
z_vT<LN5MVe>b=OU`BXI*QXX|67h9a>Z0c!$OYAv&%jQ!`#jGX;?DJ|LG4}5V5sbYg
zlC^%(h<bCdn`ZG1uyWfGC#o&uq6jIPA5%ssM_w~fAKSBE%nsi!@mzT%0O`Ae{9-D=
zBgH)qi4Dq6b3eQxu>>2bLf-~~VszYz3vm-RFZO`>=0Da^$O2JHSu~99bp9K9Vufpe
z-X4f~1C;{AAKIC4EUGvVT88LDKJO=Q?TN^D3nave)aWCnsAQ~5tVsT&fN1^$LB^~M
zN`WAksV}s3MKhN6$>D|#PC5w>j?EJNL%-@c)!DvqyYvwNb3^@m<-8z6l<Lzdvf#u@
zycxMFS1#6PaRgzyS`l|(cXwnu(in$-ruWA;xs_B?(paXDJqag@_6xrJAbOLJ`Vbta
zbGR3SPq`pcg)dBpCj7f%C$c#}U*}%wDkbIC<Ia+6XL#zV|E(3iQ)RV0n;rOt#;%io
zg0a-Qh}?*?Pyk5x7eY%OWJ7B*7s>nVkf&9p;Z+^cj8QW_L3|`iL?-MdRBj1>jJO2)
z4zdNWI+k`qYX@nF;KA;|2R=vWOS!}CD>}nT9Bx9}mUylRAZ`o&^6}f!WFkIXy<BXC
zr6%fjOI|kpn+R;-oK|&cVmgZ%kgk{Sx7HE28{B;pVna-=(RZi7WG8JOAT*Je#2NwE
z+UfZq&yK!|^mllXr!lZS8{@2htpjsAt(r>&jIb`73Y73+vmnyUr1Gd<0(pxwBAu00
zol^+FkC7<cND@+Cyh&-b8H7~W2$vgC^qJ_=?C-0KvX?;$9<SA23Yxtb&z@xu>ntc7
z?+=UFsSQoMnWPvJ1!jgD&JoHczSQh_=I9yGZ18k+Rc(dLiV+nqQH9xm9OXvSnas{m
z93jadkWHe}%2g9zoO&JDKW#nRvhZ5VZdSJxVpyEe`HQB+>a}}AorRU=3ea1)iy*8Z
zmKHDhX%lBv+^SRpV);x8>&VNI^8K)Z=e=$-@7Wk55Q2*e6+CZ@Fh57Y!*8V`vLdVs
zI?(v9GGHs+8UU%hZ*QZ2c2kGNC}?D<IiOy%e#fm!K;NW(7OZPu^L+uuXSu8ItU!(*
z#BNvk<EVmQ7U$lSA7aiC82d#f=%;b@<dda#`}v#}X7TCxY<8Q@<<2rNn$WS?jEi}%
z^S;22B#!cFgfs;NiL-ldknx{Mbxx}yoad<2rrs8O&*uoi;Wt@-fdpeqy6u7c1Ujx}
z^9~*@DjeOJxK})}a{l<3LP+1Evy_QEZa#k2M?e3(WoyqgIj^n;=>>Z-_Z}?d1C5+3
zIGleUXjo*Ii-e~U?KC$bCyQ$$)hvnN<><?fR^LShKAtX4T(FZ|X;z2GC`NV$Fs=IV
z@^RP6_JwDYmpX}mxi>GS-#mK-cfloO56iwZNv07XkmG?qNWO&jhhn6yIuta0M~C*p
z(C`?~Ds=a}h!!Z>E|`O0Ub1a$ZW~UnUsK&7g#UR~sJim`V@<m|rxyM$yd2ji>b*nE
zD~r12mq(Cy?EJ9ptFlt!)UzNbPdz58T7}h=mZKa!m?>C)tQisBJ~^4*2KHL^6L$hX
zVZ9iIP}FhDhNPqp@euVcCNa4zn1LRKFhL}nl)aRIT#Jx}if2Mq3dee2Mfvev{bg^&
z{TOB>GLOv=CV9;&w0%waIer~-mf>~SD;G7NQAN?t&VXtqJF{|Q%p42N&T)(L3Afff
zhDGhw&(+|6yd8T_ugxnCZahL%z78*|1_^z$P8(ID7uTw=bTsF>k9LJ({;-L9t?1q7
z7J(7x<cUBMU${B`h`njj-XQF2r~}i~$~Ob&9}U5}!}T%e&Kf7bYfj#uytxkIp4$*T
z`85)C^gU7&<q$dzXe-$<luUL>8@1n@CF_N?5>dT>>9eyeq>>uh+!1du17&MFmg<CM
zFi;}xzv5Iv8${B=H_TDqiF}!!1aYB|Y(#e&*81bt6+V<9uK{)NTkBfbw&r&9GhIa`
z4V5#6Eh<8}vCsJVF#6d^X6@$#cD27fZxWskY<kTws>AnI`AJaSY9JVwznsD0#cE3E
zGT@VcoPYZ}qicnr*Ma~^^o7OKhOPN7NV%}-*Tb+l0b9pxrB)-k;`FSv|2M4wY
z8CwvQuE!@5f9t61N6si&^FxpiTG`KwGE^E=XPUZUc;#_BXn?Twk_V9H`PYF4kb$U<
z3_Y5HAwt22sm!iOFypigu)1v*eizB`;qS|T;P-zgo!sh|E}$jTdZzY{&64=QIzG$X
zbm!ISDvt~e9TG)m>b!5L3(b^ypevJWOrlcs39~XZtma$3z?4PCf1`!ZKI7NMl3L`o
zCEE|k#@X#qi@AGL*6Q-b#-B&GofALXqB^|i?>}nFHn-BCuJn?gL6pawfwIE7-Wr>K
z8oMnUcLrUWVW$YKbDZxutYuK~EOlOB&hqsIpSNv2N1T(t?U@}f_9(ne!`cI7tNKc5
zKBn|F*jCUDjoEIuiBKWW7mqQb$JRS_@{jrj`Ljc5K{b{Q=^29wUhgT<A#2E~?Wpq@
zV@1Z-nUk9RQYNdLyP|Gtgb)lBXG>#$0iALJr~h&L8+|UwnT7fZS|qA3?m;JDh|~OD
zy0PZ*qR~Y#kL%8Zg!;#lvYZ^$=V5lg5b^mvNBoYiqURs<Z^ltdLY=A$ml_ea<4hZ*
zjNud=oYLhoUo*x`#+G;*2nYBn4>3%I37DGJG}$mDhx8fKG*6I<db5+9dy3S5&yPA^
zJK3SQP^{?LFb(Esj>eeXe^|Sl9P5Ou*&bYpg|^ERvrJ=3X>O(S0`M6~4|?G^6lEqq
z%?3hjV8>kL@K{<T;dQ{}x<I^$j*|b#*Qmx)gHu~9zFH;)#l*oJrprFn?EYdw`7?K@
zU0(nE7F7fK3vQNBZm96^*AIbzG0u#RWCtW!iIk2FnWC$eb!p1!VbF^VoO4Cy>w_}S
zWX!<v;#;6c|Htxeu?n{H<DYBh*qc(Ni43!TF}Qwc4V+GTT3yqOc%^YU(E|hys+yDT
zK2@nG3@qic?Q%+{WknIjfHc)Xif8Z%B-~OgopC|dxht574@3xPR{fWMWnDDhfL`*Q
z1Ak^qoH=WFh5OgaV%~1k{;>YT&D)6+-OBTC#sNgp)BPFjk7YXK#Id6rdtWGcaMCo!
zJ?%YfOFdd>)k@BNUO-p=+a6ft*G%x&1fd75BddcH=K}@7r`iabSSt(C`DJq&2I^lT
z1a;_i2<$in$sa4^K#N6xW`wnO%H*HCnY+rhQP7`)qXF^X<mxIRO%{c_=%3O*q&sLJ
z063P#m>@|w9GEd;k!PM`cjY9X`M00QBm6e&47W?>eTljwYj<HTpS;D6YnWMVA;V&5
zU?{Leq!j#Y&+?EWAoQQ#ol5oLAeamv#Lk`BerWou8j*9!x}lqYoA0NOUTInkqmJ2%
zL7(oz*(LmH>bH?CI_Hx9`I416FrYa{sjV`r6s6t?LH`rWd=)(S(9!r4^~h5@kXf=?
zR~u(7*fC~D!R+A5T^kPD_;y|NzDC^D3<wuM6#41HT@Y^x%>0N^&Dktm(GBSS9BqdA
zSjL${;Ym6(=2@+OnMfF*6Sp(rz1~V8{^PF5^3%{mMAe0YEaUFy+fz&96@?7O?fkvj
z_G#k==GO+=@=8!fP!Z*?p)3X?(mx*{q^emIDKhU+7L&i_YZW{iA@XDS?rCGrYgcj6
z(q>}3fn9>OP5DH(GHTEZCHG?aWgzyu&T_i?Fj^8N<HK-&#u8YRt(THpdot!$B8^Cx
zzg4Hhe~nf-Hv9RtGI2@YDX!Rnh}CO)g?64d^uw}py0U~J?}s+4!OW$H5FgloCwCE+
z@jDHOfz#Uj2+;jFV?au^f}1zaz@QGSHpSX)m->L;JEvokzz(D(vG;<7e~Wmtq?6^7
z0ZX%|dF7ISiz!n|<A@+|SI#XAj$xrgTkV<_mR&A!<HEbS`*aw4q?a33-4*xl#oK?Y
zT=N9d%GJ-mUSLsy%|8cf<<h6R{<^VtK*8BCT806af*MqDuZ^)oIo+a1n8$eOp~oiq
zqBQZ)R)~g%L%F_qq$})O?R+jj0dShI!EOwbDj1)C#A`qY!NUC%nP~1FI_wN04%Fkf
zEX8J3m|oLGz=N&AFB9#rjA9oT2*t|mpm6lzoCC?~R$v%_jjfB5wTPQ+7s-=+e>Ro%
zMxpXnRipLj=#7*$ThsJ24k+}^imFk9r!Y|@YdYkQVM#b1llT+n1&?*^yf`SPqhLPs
z%nb2=_a$s_47Om+(EK22+11<>jb9xbW+!WP&KFPk0+R5gor4*rBXKtQWHaL<V2d>%
zP#)o30!Wrev+{_}bwqvn(0vV~Yt4#9q;{56B~xbn+PRk^uoj#?xabgW&c)i0%p<f!
z{$nsOPj5eE^umC3?Cvs;;g+xN23f>me?EwR!DugfGKx4?&nyw_bSS!)39|;zVITVE
z=KGT>*6SkzKH>JgR0;2D2B=6K3Mc-niG@$*T^?=Q2Qz6f!#X-_k$vV%#%igia>>a{
z#}-GzZK;#*XwsSel|9}3%d(ov9^|7qOglE>g-~Tzmf~>*5DXC+JMwUX%r^kiNY|r(
zlF4$Cp@{HBUPUb0*RnU>@T^p0(bMx{{8os0VDh7IJngVwh^6TZa+dw5O{cP4G;L)y
zA9=&3tXvA=BEkZtk8ujQ>h-I;Z%;7!U2KonuD|8vE=y8^je39CG2m+kjuW5mPj5V3
zHq(fgq(&B;j7|X?*@7gKV>|0t6ub+60Q7?s?H=0y{;nW5^Ro;>&M{`hmq_+!e&w>x
zJbjjG-yb%a_LdBUU1*gs)U1n(^h8hc7SbrDL7X15HP=nzD2an73Vg!h2x8u)3C;oy
z@+v{<)@gPkQUv?p%B#=l*+z;#wwEiJBO}9=tVUexV75?31+mpHmkr&O4FczXv5fcI
zY`knm0iAK%ak}j@_nAw{PfHFdEj?S}uw6NYx~8kT;cRnKB|C3DthBU(CjCY$_oDd>
zN}K%;OOEs(nq#l-a8wOBBn=ow4C0+zV1?%OTbNyi!Z(4Vy@PeVgYyX+^k)pw?=D-K
z)Nbf3o6l2hzXd#^Yx^2MkPOj(>LtT*gDAwQ5Y@fc=inDsNeVYVjtyydYTVHp5%{dZ
z0*Nvt3AbzBk5U#Y9_kI0OcCr{4tJv+MS`9ztWmRx86=mh8d+fa4~Uim9;mrIXH#_I
zb3|DUUqm_^;1Ox$tlIO1%QqaTW=yhXKwq?YU=8Ig3_D0xcKT;fAo>!2#+(-uWOuhJ
z%C#p*+pG9$ajaWaE&XEJ-4>PX5DdwEPJa43Lyd#^gtb<!RlSmf#jkp`o23eQppwvg
z$Yp{P5)nL8-B@fmE<=`Tv{XX<iTxOV!Pz42?d0!;zR?vCkqX;kZdIuR`C>-EJVzne
zNWvzZ?ov|{0E7sQX$}N`%27Jtj-zY_u4+77K*h@zSscOE&4IO3zUs*%!M`OQQ7UJp
z!KT76b>bmznO6j15_Z1g(@b{RMobg@nz#s?ao(fe7+=r)p6>YAbJG4%|3I7>ERitC
zH+UH9)EST;keC!<O&ihCp6d=G6-`;U?JrZqTW?_872d?rPSXy5La_F(W#ARei$fIA
zGD0K`$-9l!=4l=hqQpckg|Ht$R9237)jB<qvE&Z3$B21YXcY)X-F0xh1OZ3fDu}_G
z?);NZLGVmW>b7z}vIG~7DgLL{tWoT`+@bm4yTGq=fd#$v&!~!7x@^h#xnFN{X|cK9
zfyco60c3kzy3DnIb4guTjy^;QP?{yOw>Er2>4`e+VHi%WtyaBo<pndZ2;-6KZ3#+C
z3++HaGs;))+a#Tw=r8ddfMxBUNT=VH#Bj#y;b3G@*!zxOHpwNTk3wgJ5P~7eTC1NE
zwLLW72Iqba-a5-R?XEJ#8ed1qRG0pSM_}Lt(Hyy1Eu0H~tcObQCfH8Cn)bh6K+?a4
z4V|d&MsX|Vr9{<xOVCA96&+;{v&|_gIk)hC?#B)fx%fUhy{FQ~+4^+A0<C5cz7hnX
zgQ97#`x_Ydr$>c_|BdtutZ_CIr?UYY)On;MTa?To$zYPhsRIhJE?TR;G(Tl@&rNd2
zTpk0<eCc+7ey9-JkVWj3(?<4;fSA`_&dfco_#O8tbP!&g<VZ&kTnVE{za?)VXtao8
ztg7_Z3_F0E{rB1~Cjtk&2Una<{sjKR^M!P4Ujc8A_a;5rCZvu2bIu>}iAnHTD{flU
z_HAC<-fsmRcrR>IywG?7A5d$6<#0mK(8n}8X)2|EcrY9_5PNs2@zQH=!Nu_t{ah3W
z_azz=3P*TdDB}x6a!TFK!nY1=i%H~*qGQ^U44jjt#~}-z48Gk{qxgGdwgygVWdoRu
zEpHi&=AE}F3d@*Jnj9X>&m8fYl#nX=5nF8Bh-?jvY%jPuNkS+|!C&Mm3MLLPBKS}K
z!VI^6NNb*~J4B<7OKxg5jLpf>QSw9lhb0Eb3#wiyf4<#B&l!N(Xi5z%$~VBcqk(D~
z%~)wm`w@)pXnqXzx9NkIEp<X=`Kj}J5{!l=^6o$Up#ARC_^SKK>wWV*L)vq}$=ssG
z`YgYnJF{@bZZL%yok0G^llUZtNq3oA;>m=6LX7B?LT=Zmg5dGp+RQeBUC+=rA()z`
z=HsT`yUqiL593Yu-DhvS*Jq=mQOvOPruRVtBig6|PFcSdksXuon=R))Q43R12e9e~
z5!Gw%y%6OUAyh~~cRYb1n^mIEOg)Y2F7V3AW$RbtErk+4&$r>DqQS1&*HWFldSTyx
z1uELv6x*Jm$JTDgKdOBy60JV~TR^10!6=jbEEHBn<iZx5+9uOS6r7lX8F;pWI@KnI
zv&mCc0?qXSuVkvS!ue*BmV`R4i8_60%+y1i?A`IAc^uzbJ*9Wfdw_mYx(~-Z0ryHD
z*U@OV(G}(JzNzM&kCzC`#{-32jFi5Zl*%`uf2w?g_sz=cH`c=m?mtFLdZS+QshPKQ
z8AGAWx(%!ysGj?srbyUSkxqVTc~vPhl3fmtrturi>jHUApyc7~z}T;(H7ew8Y`-UR
z6ujUGG}^z75R%UK%M$9Ur*eCiD`Ijth42|ix@}5x4+}?|xag}yo-W3xVHg@$3=eOa
ze}<N=Usx*<d^C1*cjG+|bbEvOOt_;OHPGQ=&PvT}$og-`B<yu@empqt|B!K@pZm<*
zv-sK!tO;j30zN(^F9)RFaI#{>NO`yL1ZK|a+DDGSv<7xOI1Z&d3??&$VMIoH=>%-w
zF{*ioaY$_-aS|&b0QSd_ykU<0<}n7`e-hc?g(OB{UTKET`a2&NQ^;rL3W7FZOw2-J
z8TdVPlG$jCI_s+Hur8r$lln;Pk+4xmZV+Tx5`uVEto&6=1zCmHIm?H(e8~WtoS{VS
zPmWe+*s)cpM%W+dOYbUxQ(J@VofUIYsqxC%jz7c*UW-T#+>tE}LSBUGekFJke;<bQ
zlBksW_?y}5J0f8FjtUH6`VtQed%na53qz^H$__+_gM8J)>St+HChN&=J&5araX?^k
z4$hFZB4HYOVy(5x-sNQug5-7ipughQ@9O>huG)+2m+i2$3Zk-WP=Uau-Q=~#<b8cI
zsH#N2-+K@p?qJXFI~@^~d?KUBe=x^3#>8qY3SG{Esgy&ZtTu&9>$*d_<n%7(EG0Q~
zy7jJtz49d4>mv3|#5?}KLxX4in3ZnhFe`Yrpuy**pNfh^cwwpQ>8Ii#_&)7m)_BkF
z770@wzwgf9KSPB|@=iL{k>hhQ*=|j3+~*!?+pRE~WcSm=I>?0jyeD#Cf5R>Q)T=bZ
z<i<0~_E+p^BLMRs9$uQjZeDvzHS|Yc*+;f)PUhZhFw8(cI+Setjd>@5r|x4&3a-u^
zuYw^emD))23i)!c4<GiUx)$R_PJA-#_!YGUCa0lz#VvC2P|9l<Jdwmdu!aL_juEx%
z`p2w^PNSmU3*Q$t->S%He_!tos3sUCk9RUSlJ4kb|8D+$QI0|>(N?pL3;qi$@6SS4
zQ0nKKyD1gGUV?xNZywm*Z7T)t>eN*r_OLqF&+wU?rk@zi{%_wfHa;WQUU9fX5eIH{
z>1IZ_K4?rU%^0Px6jdG11_j~FBZ(*R6rb<vTs6^rAuCju%7rsOe+4}M<8@~?E6sZO
zkWY)p*N3(^k&Jx!RdJ>^8<DGNKOTbPVBP~ak;V)4nj?2t_U1L4{5Xt)L22-j47Ihj
zq`&xNN!B0C@h3!5KbjR_eKXN7IjEFJ4(2M+AVu_B5$*!6ezqWwqQ-SeTViAhwxC4)
ztO4yKR3R-c`F!?0f7oF-uW%(C0@dSt8C~wnS=+6o7h550q2pEv@68#D#;r!Z57QF+
z#&Mw2oCYs|kVApS7+Kledtc_!n#>J31~x7C1>?agW2P?mno(LD(<RoVyLhaYrEp1>
zs9^Ezbhd|seoO)Sq6?U{VV!S>=^xOB`3B^<_X~ofyhq5Me`$2?YU#n4WX@oWv=TQw
zXdPZ}jd4BQI)|{$FVt!)Vg8Q5xJwzD$dv?8ROm}OTKbF~>NW>o{FaH^MIyvEQjV@T
z=L>B&Q8=0E=re)U8A1^#B>oiQ8g`>d*Jz?+iEcy2&bxuM{?Mhi#Xm-Uo(PbY=cF%^
z%(S|AFj$L6f7`!noCQUa1c^X;K}d<o5P?R|87wKJ`&y4@!GG95+yNzNHf~!l9ScGJ
zo-ShdyV|l%$gn=0GCr^rs`JLcfH;g<B?AIxuhRl!ym9F_)SGRAniW#B<Ev2DKDU}n
z`_Kzg%MMbS?@_v%w)p5p;o*l(YL1%qQXYPz@g<Ahf3&@b;%?~+8UJf}ufmCq5iiVx
zv`1ZIFHXrV>rSWCoX2-*J^ipJHC}>*#^TIz9AO}a5K@az`2HWk2#+qx!8q43RpYN}
zIO`*i7}{h|0C$wbcXysE>CG!MAFmo=MkH9?6a`Z4cRQ8KP~Ai1`7Q2&PR^YM1jr@H
z@9buTf9hQWB!O1WTuWSBSSV<J7=HYeFX1jiG+#SX@cYG6mlrS{a9Air>E<{&-56Xl
z0Q~7RCQ~9|ho>Vu=!7h7wCeq1xb=_#>l+<Yiy^Ly0I19~YmBfUYDv1kptbf_CU_43
z<+zxs-b?XG?BM;`zihYFk<bcNhwjr3Qa!MKe-|K?$_Sd|7$Xb~O{)qvav);(OxK5L
z@GE2_`81I<1PhB6P|qJ*lyQO5l=xJ?uSLw8fTiMihT!bid8lTAsJ!|+Wbn=&Z|Y9=
zR^tl>VjBc}bDbng6WVSeC#8SQOdRx#ta$V-x*=<nV@V2$R|8~Xy}&+CfieitiNg$<
ze_*1ZNqRgJHmnEMl3I<w<ON6YVXA%cc5Jna&ao!I-En|NZoEm52wOpCKS1PN*--M#
zT5#Md{Dd^ZYdIR_oaCU-14>F;)z5PW;alAf1A{Jf)>~O8`Vu@j{%~Gn$<2+UMWio8
zx0W7pNyOsI<r1?}MZA+wFyx7$Q%K|wf80MkvBVsXx=0}xt?>=hb+e%31EHB}@4ktZ
zu#b024CGoIpd&U2PU*|(AhtE+;vjA{<&2BfYya(>Cu#%ifaf37h2Jh>?rvI6_%0<)
z8bF4BbH&m7Be-+QELBtYWGFzaGr!vcy*mWzsI|iDPE84&wPP4NtL4}!79NbOe~&Y_
zS`f4(BJikt)Q1eGfU}yI-cz&!?>Owu<!iaZs(nQnv1A>~t>1k~yCXtkYC~u5J2yoZ
z+IwPSfz2``_$$(w7<&IpX>NB};for5(D{T8&!eiYLZ-z^8rAb>?J;v)z5Gz0H!N;B
z%J+z+r3Tt&Vo6RMjU;kVF$x|7fB4UwU?@lnyR9(g$wSPA-K5a*U~HfOD6!r)>I6X;
z;r`2`MJds0DYAp)7eF-O_SB?Qmq}h;$KJea&!)m%J}xYHX90vuVbuGLrkn+LoTd3a
z&a>*k>Xr0|nOYAf@~urL*XM6R*w-)?cE0TR{*#T|Y>W%CJCsPW6IahFe|hPf_p)Ro
zN&X$ON}Wn@aI1MXk28bnoF4FaNam(m?&us4sZx;S>}h}ELaHz6qac?8E)4t_sXxK-
zpq7W>8~U0UIjn(G6?AM+n5^vg-N?LDbh8uJ;X0<YE#;^`)~I3IDNRK5JC;mH6=CFu
zMB7{@<08?ax=Q`yPx;p-f1FDguni{pemmoP&m2(b712<?`4E0GEfBH`OUnOJ3a@n=
zHsG8#{eG~uytv6);;TD{IEMwzcwLxTX;Vs4rsOn=)&16uwymX0<F8&0t3sPN8Yr^k
z5i(QH5qach=3-m!p^CLm7bKEAHPrOr)To{pb3tzSR*|NnbQ_<be`B}n<?Ze^)9W1E
zc6W+#A1KA6KBUr6so?E4`IFz!zHxL_h5LgfUw%cGr|%xXt!)_vYszVBB9&Z=SJ|Ut
zeP9}vjoLqOCBW7o7X-*|2fNv~N%zH?{vj-9n9nvP>0+2;k|^f)fSjwcZiW5G#e%YP
zf_fGS<1mz^t}%H;f8)ZGwb>tEDXnF;8LDblvZlAAAKZJ`_v@9Z8e+bI@2Y^ucw(=|
z|G<dhsTc490fZ0TFbeklK)AWC*@Rb8cD>iK|K;!6vDk1oSjh|J=4?GrvCMZ3(Ws8N
z7&9stg?A`T6`3fD^9TDnz<S%C!SpATa&u{!Y+lY#zECi2f1vFlATH6rfCE=%Sdc7~
zSyM4FX2Gi4#0K)`j=A1V)FVkNR@;F@T{NpcDOO%XA$Q288Fpn$9kEMYi3}_R-agQt
znM;K(Y;u8{Lh42ht`tr#q{--4d0y1)xD<9LZ`x;i+Z}<mHW+0LQKs@hGAW9f^q?y(
zfoT-dmS2MPe;9oXz%TrlzVpifOco1L;gREMC6$DYiN0&Ixk?T-(~0+6{a6n{(EcCP
z#zu+Qb^K6glOfJ>0eAi2#%VC7h-!ML17T0<qQh6DJ{c|N4b$!Wo0BG%aa8-P6$lhv
zJ0cn?<k6M~%IHXzhbkT=Od(ZIU%g18RV1i0N=a+#e=1kXtUH~Py`ax>@xJm%-dWXe
z#CiPmQ*O5;dPId3#49=Dyob%HC}%fkcGs`cF0UE;;J8|SY4s3rN6;ec1p>zpKWCG9
zMec&45!@hG+@q)!yQ<O7fVCkE;7brOq_pv4mCs8}jeGu|>I`Q!-rcZMJ6$mjwQrm!
zWukHQf6rW2f0hY%G4;03H|XSpp}oHl?WKICv<f+K`t-q-oFv~d6%Ewj-30R#SHawE
zIFjt}m<N$NDU?Im7O!mKw4juu{KVM7DjAv2exR=pdp!&55@{=BTCrsn$%EcXNe*eO
z_X9IGBv%IszfD9y3E(;YF*=%1)zoh&Pkan5f2aSwLtL@ZU&KG2mMERUIXp#`zC1YF
zwH*4XJ&gI5V$QuQ6=%SkMG%J+U?D@G0LGAN@|e?<c9CJYp?2|XYFU~ckG>Qcm_yKO
zsiP=|$Ryy>Y!7~4&jiNa*Q!3l%g7S0@n={|UUR-IC)efzK~l>~^Z3<|N;dHBbA*i2
ze*kJ||8ytqyXLY03`_bBq#BW|XFdRa3HvZCL$^uaR1bl`c3O3;8Eao5#6F@lWhsEx
z7UZ{L^sW-opX35f@2gXzm*1+q@r@=@P+A?p83Hf5!6i-fpf~#U%#epR8qa3pNU40e
zEoLT)_Cr!BD7icKz2VqABtc}_g;R36e~G0-x-!)+C<iYWr#||(p4!J^Sf6EWvM7#y
zKTDLKkw}@zqO!SH_4g{09-g=4f*SMJYvEvH#`H-s#*^b^$`~(kxnEKzin)cb+}HiW
zjPOm>4YjqRHetMUkLn7hGiu!+u`yglooahdsL`27e6Rus_BK2VPbp7~O8T~Ue<QG>
zqHUJMP2<c?-rCgR(I>=ye-eFNRJ5ew#WWSsGX3hPwkV&{(pIA^Q>g5riDE~LdTd@U
z-7ChshDCpAm_<5OCA|y}n?#5qJ*1+N8zCY-YCZ@`_q?fK;#jn+NV~|2orH&+97+(Q
zbSRl*n7Xhx5>-o0hO%=Ww&9jge;Qzru>7zl9$fmQ*KoI5<&W)#A$GTz>MBpv>Fqe6
z&gMnBsQ&HRSe^r&Fm;s4)Pe~D5<CQ|A)Y8Tpx4;8Ml#l)LI3Hu&7Nc(4jNAhu?w@v
zsqH;LtKJP<<p`6~v4Zd1Bo5<E(avfgj1sD$*B{vywLK}f*Z=Ze#9D$Yf8v}{Xr*;C
z`P`Ywr9bDvnKFkVTfCnB<ppFDQJOaE7=x-iHjWvgw^5d#<5+WjhYB(#f>q4Ms|fuT
zy|di371n6gj9Q8e>Z9}>-rq3$TcU8L34CgKT$IuEZymR3UD})#^yydAuBUAcpPOnb
zL^BKF)jmr+O=7=EP%O>Ke+5~?3hDJqz*G}uAl(xit}wa?s#=c({*h?}cnBxWXn1g?
zQv8%gNQj#*F-E8VV%oZhInn!>2>SOE2HtP)3ft9dtfvs%we`n^B?uhicaX3PlQwZO
zkD`>-AT^fyh$SiJ--E{e1P4_hYd4Pdm-*|i!ssy5=wt$92`KVqe^DZ6RMZNIS{I%D
z_+a)wQ%|!gdXC50AruTNcLXg;nC71+GR{y=2$2?rr%h_eyda%^GGeX%Nm+WEEj6%Y
zEsH6De@A{6^o>iXNh<-j=xWYbGyXGfKU0e7;x~1)CpDL>T((|?VF8R=4Ti5}do-+P
z){<LEuF6f;JEs(Te*!H!H3L1>v*|LY+AI&}QB8V2KQWZ%jmX<KX)>|zN<$E)O(Vb`
zM%MCiEo0RrWxWug|ESSKy??21#I;!wnon$yI%&AhT9*Rz>to*eu@Qi3#XNAH&#ZA7
z<R2jHl5ij2IGe<RjR?|{?89(lGwu;9gTCP~Vb9f^zPOKRe;6~fle2*00VQr~`eOx&
z>UlbrAF(Ra8DVB1`q!QN#Whc5#!UBf#mhDCFkfKkcQO3{N!2>&)AKfGgdDw|)9kkL
zYiTl_#@BG|fi6i_gP(pGVpj*CsWlVIfw5t}efUw4tciTu*Fj;hcJ$v1L8|NJ_k|ym
z*hb>#uROA&e=0zfIvZ(&CM$9_w=PgT1~rTq)Zu`Kv{jpdb4Vn<5tq|}><QJPSK-~z
z*SEhX&#}<3e1L>!HNSdxG)WsGvB_LBVYaL8$rmrPw%1Z9W%KGpD93FbhB9R4u&ZB5
zMZ=I4a(T|;vh~5cG{PB0PCyZ;pb08>QwURhfvWAgfBn0Qpb%MHDz^1921-4IcTo6q
z*x93my=&g>xc&YOz5=Vf(zje<K4ChrQ`-2@I#(qr!_nreEVz{kyqcFBTVP~%FHE@h
z*PEbWy4ky=%Yt$M<D)v27U?6e%HX$*C~%81<mHV3IPV@2-Me$pB64E-UjP{3vFI+a
z!hnh`fA81%A$XOLAFZZYlRRT!UV~L_YFpUQ$|Ca&Logl(+>PE9a-;PZ!WxH`@8+Sl
zj09qomwwYq5@MmrG=eTrFc}K}^>I!6+$eDKPGqD{<>G8ckYI#Fd|v<Yhmp+!VW(4_
zC_wscw1YKL(GbYpzU@yxwN=JvQ{D`^_!!hee-OXS?Re~K0;dPpro=$zglUu;2hs^!
z%Jzue!a%IeKpTLRdU7cYojG_{C}fG%UwrTbOnGD-0Uj&_m(<^yPqnYYhKG@4MyS`b
zV>R4ZfU2+1ZzkqSCwdcRtEm}*7nW;-n9;q?LXSuT6^AcMKl}cC(FGC2%ml~QG3rgM
zf83K*FK8f*1IktRvoQ`;^)}r;F6&aPkidj&8cly(A?>N5aQGEo=f^C*N9_eBGhdnr
z1ky`YRD3%#T9xc?cofqU(~|B~((SAGd)B2!2V*ukmc1O#eP`N(Ih`NDqlt=`Sr45W
z$!hG|G>5A>iUCZJJU2ajlst&X83sgOe_qEZ`D7zgM6whUXjX-NISNgnnbJ~|F>F}D
zyvtlTv}6JoA7Jbz)`-c=FXS}9+wKVoecsFlD}<FoKrL?{t`a`_XxLoTiwt^4Kq}mA
z4tng8TR@P*B9)7<kS8%<9DdhOYnfsH2Xu-(m|Ttc9ABKk1;*6%@5a>MM<ia1e;Wg-
z@!KZ6h(B(Mi{aI-A%7Y5IvPO6h_~4sHU?Jz=HCbY)Cu6l8y-LBxb(F+HV8fuLHQYH
z9mkaMh}eqkUu2=vCYv+;jWiHJ(j_SmA0jGb7&_Tx2?ftlZ2sCAkEK(Mg5@4fzU6!Q
zudsCN2k(%%`3J92#~%US$}8!re_Pc=$rVC}ge$S<;2}oF4&p1{P}ishmim@d;d{^z
zUjY=^UF-;SQJENM%%d0L4Y?V>+y@Fsw1bMs_wtNS8L8=g&Tp3PzX@Yi<LjGDWoj&c
z%`p2IUPtw-#>8osKMNl(MLQq|TZt8BC_->q^Ih68R1(!5mxz1ZJ50emf7<+NyhY3|
zhnaig#s_kruW(KoxrUG7URhJ+vUggEE&Oo^<<)-y#c79V`lQaL*_6~2S4=vtmv$z#
za3e~jDBk|MnuRA;F2qW5j2G^c$J@ll*1uE)3CY?Nd)`ICUsc|9xJPgU2>`iixTi`Q
zHv?;*@{d!$NmnS1{pu<3e>*=ZM^cyYXZL{`Wy;#W3lvofuY;T1<)UEns*zS|hUe^8
zF5YsGH0=h;Pv<*C4QzXS{X`EGLl==S-J-FJ4EF8K1-F`^`XYr3(&6IPSG2;i{jC|L
zGG$DK4aIuCz+#^2>|;HLl$Ah*5a5R?(9{tswS)3%WVnVz%1Yh}f9N1G(5;Fkm68ay
zp)+w3<W5wnc4GXMEp}R|z#kSCfc~)KcEV=(P<%Kz=%v!kN$rAlECr_r$w;m@if$9P
zHJmdbuqW>2ry7fAUS}Mi2CX8Zxl^B+zA^|KSSY}(Ej$o%6VLv_ICbg>`bX-Hbx9R=
zHl-rk1pz5*Z~5dYe|^nn&@W-t>4-Bv@M1JM40)4A{cO;W9Bd6gwBza#GU@$-BmFT<
zq6RQMl?VMXTW8y}kSJE5se*C~PEe~(EsBz8SB%SeWn&M@2FZD;r?26lDr!J-jd|4F
z46gAJ)OR`=g+2V0Tt}@Zc)I3Xy%lc*TkPJ5x!_XX={8WQf4?{A(AA*GzNNpA&pGg?
zxZ>&|$#Lp;Z0tNXe!a79CY4;PTR8YP?8d>JQ*L5Xianp+LYcjR0%druBR|{)7{KZ~
zEav-_G;Y<2t-gbwv=UN{+DO42gs-tR3W=8_+<MrY+oqLtb6})jXk~QpOua!te?%%(
z5O4V>yNBP{e+@PyTm#9R)0?Qfq0FL}ipvw9%FRY^&uQ3<>k*p{QJJgN*e1c3j=_&4
zP3cJC+u<qlnb&pMrEf%8og^G?ZYU8~<_qe>ow!f%n(ANiz9aK$wvDWYUtL4HzrPdF
zCF&M&-#zeC-n%>+!`jT?F9P~{-xFmHCS~VjCe-LZf68vBMDwTz_GZtyVm2^CnH#Ip
z;3)%Fa%6T`Y_gc9T-m(4Zkf)jB^V*38cV*`*p?lWe81dT%uL2QpuIWHLtt6@k_tYm
z^>x&@N!!*!(e}dNhLyJX;YtwBUG-QweuFF%_kF*7!dcE4sG&K@Le(0K2qq7?+^^Dq
z7p8i1f6t-zT#=D@2u>J`(lcJaA&THP7irEyiDYdi7u2?u>=a+qew<oLVk`A!cEo&M
z11-3qDBgf2hgo&7Xh62f6K%RNr|mq(Bm_kpE<V-9%G=Gl0ceMxtW!lel<e}Ce4r4k
z&aM!*o5F9%r!xx!Nf=HQwbToV0_ja@F2J4`f3RuxvtWJ<F4mv!J#Ica;BqK*j$3c6
zxrqW^axa4B=>i>*Sm{d{5G$CHT-**@=@`S{=<4P0zMcUd7OdtEwHe_#gm&xWUAz^4
z@MtY;zVv3@iN2|9&>rG5ay8Xy=(U)ahcTFjX;$D}7nXbPVa0?xZ`YRj>XBPR`UKj4
ze@|YjG%09^QpU4Y%QnzARBq^kRG4_42+GmOgMt;LMtb79B8tYMlBW<anj`j*(2kyn
z&@?{PoyJM1{!M&RbDb4sT`b(_s|g>oBjBD(S+mv#Vvbq}y6_O|!ql9o5eU_~{mlsz
zQC74BvbQ))sm*i3qfT?1#ES}6jwDk!fBB09IAdRv*~8wh<tEWFjiCKKT`u3o2P|43
zJA7*Xz4?^rJ^t9btV7a2BXX2{EHK;TrjY9l#<kN%^Lq_knVPi|gkLEt_R&B^s@Prz
zk=He%U(YV!n?e$;zQ;sl3#Nt+sMa#$SFx~MKeQ)VIXd0WdU(mpb;d9}O-MT+e~doV
zG}!S6)(2KvyGom|Sc77P5BMC3jSI_>%jWxSi@yi=Yn2}|TZ|O?wGaLCI8C8ku8uJ%
zUblhtd1w}3MFyErdsl?}!jh|rjw2@?Pafe22x0(mThNWSAkDOKVc?k>MJsHKuyKVR
z$idI#DwA8CL2j@*G(Zw^x>So2e}SOoq%5kZ`_Z5oFD(Z^JlB1yxh6H$x5Pa73KT%M
zA=OObnBs&xC9M1;<YmIgRzBzGX%B5hbG#P9jCm!VM0&O~_!uM7%+|8Y&sY`lbW%n$
z>GFA5Sd7pe&XB~>vO<t??zcV4Fz!JRCk_kT-W$RoUAGh?y-pFc+7fO1e~?FSk4Wjx
z>f}jBwXEcY+dNPzEhb(23c;0}k!RVY9NC==v10qy#vI6*1diY?T-qonK}@z6U#-t7
zU^w)E@+RO^Y@(MuCa3uxo$A9IM)35r#KEcq0i+F~7%{$^-d?C*3{L8rqyz?UgXW;j
zRKLg|khDGIG*adSA%I|Tf7U*VPu8Y=J)C4P&g8opbEKm4;?B?OH^Pu1^2N5xO1eKR
zT(NQ*fkw&ifRQ^n1anq4c33afP;ebRYHfT=+2<m*#JAN!$H0(Ll#lVZ0GFR0)*b&}
z0N)NE@h-)b?W8DlNe8{j=&YUT<H9_q+=DJEKmV~o6ElABxZO7Xe^v77aDB38g5szZ
zj|-FK_W0TqEi9b-{H)&{M?V{5Zh3!BZ~*Y)$nv(DpZkoAS$xQU$}Hv5<8Oib;QKWR
z+MBotM(~oJe|3h|@X9+4L=@I^`$i%^ZJLGnm@187tF8V?Mp-#rLK=^~3tw<)0%e?4
z0Y5AVz2xQTGc4^~f3|b<()z3akA*irT3w4gNSm$#-RF~j1oi1fsRXC_sxH->J(!YH
zTkt^@#}SX$%y|J-Xm2TUg6u&~!DwbCE6A_`CX~ZXV$q1nRpr!xH-Z5!4Ydg?2>!eZ
zw{RQkYYOhIR45tMk)om=8W)*GKwU3`qy}sfJgr9o|Aar+e<4e&u3z5U=U^)%%{nBM
zPR#f^WiMYPx1@Ffg;kXtRK|foC4<zv0Ul?j1J^T$km#(2_V$lRW4c9KjKS;ms<>|{
z(ZSxL-u4eQfwM3?o5z!qpary6;D?Pc>bsw80Os?e#8^@RMBcj5E^@jC>Nu}uie@9!
zM772*k4^KUfA}fhUB;2V@la{Z4BbT*p9#;2Zh;#pR7UvzJUNwPptV(9nA#7fRAh3p
zSF#E+@!PLRP#{5gF=2yJl09gVLFm2Ga186+<AprkWmvg(!ly2-v({u4=giYV1>W?n
zQRte@d3K6=&<?V2V)(`1TaQ&D&m^d8f`Y#=L&A`7e_^h9d-MO0vp$oLnr6=SMM&K?
z-ECbPbdOs8NV1WH27Zj9(r$FUou(~;S})MU;h~ypdML^LC8Lya;nylVp=%QBc?Ca-
zRh_F)sdvvu?q7G$0x~`Nqy_c;{d*L3b-;{ZDnhjEZ;eDLcfl_0y5P(|uC~^RKrB)P
zqxEvze+(seKXZv>j5;GZPS>0sD??X7-ginMvexUSMhnuY5%B|<_)<!LuF0kdQSFKh
zxHSSRjvWoD!1yuNx!ZeomwJU6QXJMg?U#6&VJLCE5Lxi*s{K72gG+-m+uGrobT(LI
zUf<6tJ1U1Og3zJJP~r3!mao|9<iPdxo%KnEe{ugl8$n|jMjl2&eZn_+^`_{EpHE5f
z0V=2!Lq_BLD`F~jtviw_0vyRm0&GHP)HFAI{Uh>b;i@Hzx9d6Y+{h9JG}QF6C!fBD
z0XmIn`SNO&->&M4K)!`+nm88B93zSAqemdHcRIPNC)bd1P7`vEYiUIF+_5=Ji`h0p
zf0+Jx>GIUiiJ*MZn%43E8zP^p;*cjV20!np#^bJFrejkMk2q~$0%Q#E`85Dhn0h21
zpr!bS@Wtq`%D3QQuj{OGF}kN2xhGC>z)!cVf=ajeqHPXbM+HlO=po5~!yBYC1Alc8
z72<^9R@??Sd%U-HJT~T#XMO7@44m3Qf83LVWO__)_!N_2x%1qjP(41_ep!UPU^eU#
z0o4r=DCFJ+uA(Y|gs!P-pi3)_XbzjRWW~cGd0#-MdS9?*2(Vac);ydynrz$q9-(67
zEJBhpJ|5@^@E|hQa0sNI)tF7XrDD}l209)AaK*jTsDlvNb|FkKu;;Btpol=ye>1Oc
z91Jow_ohk7h~$Uo1=-D>3`RMCoaj;?zI{z5iak=rX$gFZ?5RkUrN?=LpHj-?6`>Fx
z;0keWAk%*&|GbT?W?$tsrEwHpyx3H<1z0X9C*4NAMAfxVBibk1%2?t@%Nie$UWOPb
zaW+|pF<KmZdZUvvi-U#ZFm8>2e{Q#`<Ij#Al(j=qS%~KI&zpMjAkTGafzj+tHpNzC
z$W9*vCE;nlh^dnq9W{ynonXk2AqYWQUQi_pEODPn4t4WD2|<N~VMcX;C!3{8H)sRp
zk-0m-qa5{CsiB}Pbgv<0KFZg%^!m9?bw>zk#H{Y#3b`ZxP$MW}djMhCe<Y$f7@W<5
z-27;o@j4~ZPRw5}9%KQ;JDWlE=@jg2;7e50SEBOQ5U(h%pwY~IE?>lN&w29ryt#hv
z{s}Q|8?c|d85#mt#sHG+ZO?QnAD9OjI*b_oE~+doONSd3iabUBvApRw@S<s>PC~cQ
zQMpe-z}HrO^Z0J-&<h@De?JJQr3oK}rNh{5_d{WDV-lvoWKsP6?Hb+oxJQ39X9eAA
z#8rWVZN+{3CbQPf0rAL&w@ob?VzU%Ue-;yt$rMA-B$HL@{<FBdaHcaD?*a=>Yp(xu
zk8(zopy9EDq!f20@+2I<;o^0_aXO-7+S5nY{8(|D%nFK9Rou=-f6@T=cI|maIsij-
zarn0{oe4l#<^c}74u)DVf;1LMUhrkb4MO?#O_WABjl07YytB#4{_r1&FF1sFIW=pI
zu#aCz)hj_;k)AJBfN0O>dT28vk5E0+CG6M2EBx7CF(5rM%i#OGmY$K(OM_xkE5SHE
z@p}~PK2>&4KpNi^e{1U|ff5*jVcMKg#i$HiayYAq!Z=%-=okubjKdCi<tV|BR`}M!
zn?QEZRE~O}gozxfXsjS}<yT^;V~3B1$*Qbqr9jzpA(6<|9{OM-Lyl6yque_Y2c?u1
zSq<^%Bx|Prngu3W+MulA1Gxczu=SIMGz_1bZL)RbG+40Ne=+gHOq^tZ=F+Vb<8d{0
z;gG-F3B+`cJ@hFF*cc@oxh7&M*{G`ra|K(Uik|Sy`TtlWVrxiJ{^eNVgR&!Doe2DZ
z!u*5mr*_tpRL#79Mk{6i-4WfObgvT3OMe}*=c`XPrMkTkxQoRgV2ZpaW(c+!lF2E+
zM)*PmwF3Qlf7RVk%e<ONK!t$UVf<bpA5|mU?t;_U8Hg1^H97(g8hzlYQ-dm3$4~`L
zyfAg&fT*IJVLs|91{Hu*=eSas&{tN*cj?b9=+Fd<UoQKpy*6`Y5Ew5YKW5_@+2r-I
zMS&z-9IVxy>;V&&cVu3Qui{J_M|V0ue^eJmSDf@%e?pk@pnF*zkm;Gnc5v6e=!W9N
z<#lUJpWSVJ`>-sP%Kzvmt6`lP5V<jmX*C#AnGe_%Hc~$okuQFB=(Jlh2Cyn~r!m7M
zpuz&sF7c*yf*pPgd{$C)rCdq+;4H!`1-P2JMB3D_X1+PM^Ef9;A@_$l<VtyRYLKB$
ze+B;ne^(AyVnZiH_g+kl62VvLCoBn4^2}5Jzp=CoKZriaS7@Qtf<}SLa^bi-T*z4d
zlx$bghM)0Ty3hzebgVl4G)+z0rDKQr&=ayb;|Rn7!a?-s$LEv2v4w8$cBub(7-Y^u
zKi`Y_124QtEE5<LuRS&C%m)aesnW#GUFhc=e>6tng0Fu?3v?a~cQX1`FBgM7azY_N
zSwj)H0U!)1FRdZ!sgSevsr=^)W_O?z())grLb~0r4g9eCgUlclsp;6+&>aIzA`iFb
zl-&LkEModBE$ZC%d*7=wW}6L3Mi%L@j7q83=;<rArziZwIW0^Z!(jx_5Nq*8-n$a0
ze<t1Jy_!Ua4fZGZj7D8*{>UkC$Vz1O8Q;~Ol`S>F+Nh#F7`=G@?u=wzX|(Y={GNSG
zAdK>4>3LAGZgrh`r*Za3V0|IKXE2+{V8GdJ7{Z8g`+9gkM*ZlbTec!pm0|wlK=|9p
zU1UV^9~%4j<Wcd5W2-f4*hLQR$pL|pe>sCGdW)z#3kjKW1PIPXeYbChTkMBKxSE9X
z^`Eel7_+9kEL4=f6|5I3LyT}0yF;?u)FluafbM_o&?e%`g*a2#=QuD#8RnbCI8ZGZ
z?4%K7W2b~QAA``F8Jowp5&hD5=3;r+F;M$auL|C}1skRCPu_K6`2aPdI}_jYf0Qxo
zJ+m6hAE{y0PC$7&z<Lox;BC<`lXe1pCv}C3l!OH_HV2P_PnU{_i+@$Q7n-4P8IAX@
zxDYKp;z<@^@<wjycX93vH*$Z2b$f{}_7X$8iDyqVHD<AmL(Bz$g}5HKVoPIPgEc&+
zt6fcq$|EgZu`SZs3d@X^$v$4?e;LauTUyxg++{NeWTfC<2ww=eEjjKV!;>B5q~Gc4
z3`6@OXYBCskU?9uB2TMOTaSy9hWs<ygKEvGHcCsP<;_<MVIJ|&hS#HaJ3uhINS*q=
z9W7SS;-<+YA^#)tI&`JEKn4rks0J3DZacL{-%z%J6xg%;?)M+XD13!Ae@Rmt5r~7)
zz_Ew5#zsYhvqPZ`3c3R7OT3rbExRHnPKbNxVsvob%V2N}QJ-k!5#)OHA>-9ewdhYs
zWN`(`uFUPgx0uzV2*2)M(6MpV)s*g+=G<E1y~sZ0?O{JY3-k{B9>@iJ+32bIZ`?uU
zLA<T&^z^p34<y(G(09nOf0Uy8NuVq8J@qS15m-ke_#l^2+0P*9A`*eFV3vi}-mmN4
zp}%nj!xRg;0ODRlWkfZ@4phtxUfzAlYmz&=?<U>@7{sor@3%`wr4jKFl%PnT0LL+d
zVq4a@{kSNYV`t(A@GicDD^uni*0#?IYx<#OVxHSn&ZZ(feY(X%f8?YRw*fZB$i$|4
zM@#usG6B)yAS4wNSex{C0YJz#4l@Q|^2P3Ssid}^3+}vH=9A3C;$J1C$I8r@yp%zz
zCDPcLCPRr4UR)ofjbE@8@Xh!i5ur8Y$(Xw^;B%Ayt3$mQB&V>F2%0}3+HmkGMCjxK
z=1S&NyJO}<cL;~?f82<?BkL*xP4Q`_wNB3rwT)3$ja#?>UWor4yYl<!fD+1ZA)dBt
z^%84P$=jyOFO^fWn}xoW3KEnCrZchI4P*EKK&^zLkLnQAl~I3TcrUu?we!C-rX9q;
zIWgHS3T+9t6yYf=qi;n0G=olYss+J}x4%7zE!KP&S#Igle@~v|6z+KB7Fhw!R~5!X
zBvN`e!;Ry-fAxIaJ&z<K+kc3T06IsJ&;8OT_>7Ez#4M5zvM~B?f48Y>W#b7!bsHeQ
zI?^*}!MZVwN1R4}dNMc$SORmhFY*CuOg|;wt8Tj6cRR71AJ<ROE$zcj0NnUTQ%8aC
zf=oeWHI5d=T{%6eE9l@nwVeBDg{B}u7cjlbsNGEFd+MEY`&i8)=QC|tR%*pkmR7m2
zY4Qcer;`b@hn*P|s901%QU3GX3rsJ{c-2??VO*zhJ?HAF(6u^^wR*WXXIEY551yAn
zqa72MZwehH12!@+mmwe;6a_IhG&eYxaS0nIf33G;lx1DBC7ibHJZamuZQHhOqtdLj
zZQEICtJ1bAjjrc?`}XZ|zpwvv-yeIAy<)DI5wT{h80U}?DXP#3nK~Goi96W4(lOC9
zas%Y8?2O%9lpXBk9k^ual+7&Mq+E?`0e>+z7&0;uXEP&LD+hZ~BUdwSfQFeVK-A0x
zf56NPVB+H9f*}KlI5>JaTUl7T0w~p#HK=H4X#dml&lZ5O*MA^?JzcCU>;dF|Kiti1
z9USe<>|Os-{9ik&n3(}wEzJPtR<>pU5d}ppDR~J1rG&g1K*G%4%=vE-if+cXRwe*h
zD-$z&7c(k=xq~yn_TLMDiG#hV)jy@Vf6)Ks5^@0;0bCr-OsxL8nR%L+IsOCD0vyep
z?W|l}{(b|jTmTl%M)t0M&%o6IU}bM&>t^~-0)Or14*yEX(b?he4!ghDUou4p7grY(
zXDdfnz~88fqT>I~)78?*^`F=-R)1lDgZbZ`rVb`<|CIDE>@Uq<)YZt!-UZ-lf9C1>
zPb_0IfT@*>qpgwG-`Ic29G$KHm4=&(mA%D(W<U#YHnT8tHnlZ#arsO0m;0a7`Oi85
z|6BD&j*hlo|MGVD*RKDSgO#g`nXNfJ3={L;m?o}&V_R6+!!Z2QFH-jA4ge;`f7?yn
z9RCAzH*@~i0a5<bGgN<*Ffw(pf4B7ln3|cxFvvT&{tXGB{J&F~{=ZM;{|1TwTLk`Z
zk@x>i-2bZ4|8|N0U+?pOr51OywUsxr``ZBjo-=^IgT}}n@ORb#WC8yS8(Sl1z&``W
z%H;p!YGh|+>-B$f{;#_=%>JGA|K$9y@V`R{*<1Woh>nSkp6%aeD;IGqe@`=0MJrbm
zOMtnN?cc-umtD=?)Xdq|%HHg+$p1RuzfO#d{{>UEv@)@=|3?kBe?w;WrvJ5uzq<Xa
z0tO{{WpOPbn*YPJ{nuv2zg^*~>g8w#_@9w9<Qz=@*Ww>4VPOYPfDauLI~RbCnS=4~
zW&C>`xtLk~{(r*$i-_qzf7fzGuFh7T03AkpMn)#U-}iq!|GCopFEnEICJv_m^pT3I
zk-h2P*85+Ee_#_gXXn3)|7#ljt^7aM|GETbW}apyFzYK0COo0m8QGbx^_Y31Q=&S9
z{Y;>vVUA@@s>OwYwGMp&8IikORfZpRVf4*2++Q7D%jb@thtkwXf8+fawiJD7X4eIH
zek4=_g~-3jyEtf$f*JJ7;4?2YVtsCwWOu-{*%{T3_NJ7S^-Dh>@MpSMoS|->sRBvd
z3j@(epBxeVP0||^(7TbEfl%s;&M(N5PoJMCh9|~GNBZu-2M%#*ekMSZ@fczHr=Am~
zxOy2nKQ@_sfxo$Pf4ISV!26(Pe!7ta*%N0th%|UFyfOz;c9lJZ%;*ZIyELs^HZywP
zj6Pb5@R0x_@WMW_IDCV|>q+8G!tZa94hMPuATr{mZqrDsW%amg2L5SUOM~|nXP{)s
zx|H&~H_F#A(qy%4;}PGWRNlnThE;hIc2jCS4ZuN5D#%@Rf1MgFnCpuptZS4RXezN6
zqxf<9o&-@iv4i1pDp~3)H>Nrb#e!@bN`~V7Bk%2u`03JFSh-fkoPqzmVY862w90xt
z_TKxW(UL>be?}Rt7qtL&!B2l!sv6jL?2h5Njj6S^P}KOL??hy~2dm=e2K!uC4Litn
z@=soB!S__9f5)RUFpc=Q+uG7SXryI;?hhZa`ngBh%1r*!4|>dVk#>ILswPa#rrFhy
z1=VmU*-AW9YPodmM9U~8cz?t3PFTr~B@p68Yhaw>akQ9{01TZrEOgc#7%iV^{MUpP
z(NBJE=tA6jtUeQeV+1Mri%?z-oh>I|W3;_ocs@|rf9RjJ<&nhK>1&<eJq*fp+=CBN
z6lnxl*>D)V?D}0oK#^zKItxI@G|Mfg;ysT86u{A{Wic~$W1_<Ywf#_jkB+PK;&_2x
zcc&#NUUDArTsR<0lpm9o3d1w}H?t@$Dr+BFnlVXaW>tj)UX8gkGbLnls{))c9n${A
zceV)me-v@&1%xz5b~VlQB7L5OE$BFMq(Ax!m$;zlW1sVNyF<CSD;V&h>j<pwtN>;^
zjmcxHU23FOy)`muUNyLGN6K}w7<3#Uej6Dk`h>U5L!CLBLT7_XXTYL%f(6hJlW+p2
zd0v*y-wUHEX0VOS2x@QG$?>urZujB&+DOQ=e~*V_4=UZAF=fGx!7LhBtj4!6-PhCE
zXED)=sIM7ot+=;;IxUyRx6Cd0P9Ho#g2>==+rBkCMqWS8ptjn2zuw4hR7P0pcmUGE
z_lN{cinxHFoW;4^l5Z4@S=k>@qJ`dB<%NVs_JeRguIQC3qBMpvf<IU@#S?)VdE<3*
ze?e@VcKfvr8nRs>T`BY0Yyht5m(NLIrRFHJM(<M}bZR3aC3<}LGL?wj^7>K&ZJIyr
zy<IuSZ5_JT(Yc;c2Of0iQlVd-PfyZ|d5EQY%c-^i`dJ2@eq`R&ReUm5=ICxbs|aId
z%BcKC)2R16)dsqv0ho-4Q|u|28`Tlre~WdUZrfY$%jZ<iEcvLb+ne_1hREWHet{Y$
z*tH3(^pw%<dD1-)4sUP-uEy9I`O({rErjd7DFM87(wLKoG3S;p<O=R3X#()+yTg%w
zH+4AriWp=1Cx~QjU)=CFl1CFj@#hMeP>IuHJoMyeHV)wGOPM4qOfJJ_t7OF7e~I+W
zo|ucB2s9mW-AF>-s@Bomuq)@ln(lFu4?R)m1-SZA6wi>O=C#B}8e}%NW!+WPQr+?k
zLj0PN7Vh=%9?Yp9SN>HT4@qbuu%CONAtLIf&M_IC<F)roA<jx&eGs~oJib0Wb8&9n
z7fUaTnWvfr=Wl%mi8u7E20c5RfAq5DsQ!tH{3d8HPRZkZ2Uf$eKSRF8yD6THNMR4%
z*Wl(WteKehwB0L}C8Pb6>M%Rh$sLj3DEe_7^(oqc+bNtN$Sm<MVDeM3lFLK5i)&t;
zOocoRfKZ#07W@7fgnI^J9QjA*I)CmKR5uFb?jF7N=4B~>t0bqmR7mnGe{YaBCyn&-
zvw^0i1~F;MuHPY;vnW$_Xp8%5t6Sd?{o?TAE9&rQV?Gq&C=udMmGeO$G4}eybDYQ1
z$p@F8P}j<0l7v89fv^nj+?=VpMj$LYFg?mi?<&0MBg1UVH0_xt-QN~(=FVT20Ujqt
zX_v<7x{LS69A^N?)J8pLe~?Yi$65FCIkT6Pc4d(LthLO<8n%rHawbVd)B`zF@`nR6
zbEQ1j>)n*aUQ>B8EgUS+T!AO<A~Gp!+o*0hW!GpLVws_NszH_h!_RW{3e?V67z=SU
zL}h7Vm|y7$C2>2k=@eb><Jo&7d9W1hpmZBu2#4DipRb)ICTv^Pf87G&ty7w_kkv6_
zgXQdpPdQ3oQDe`5VLR;cIbsF@?)ntdc7-C&o=vj+`QoUHkC?uS5UypL6J9k_9cdY;
z-zTlH+jtSwm-O9AWvG_c8<_aHGGH2Kf99AN3Iq|8%uVibUA0<`k5@UfL(vKFy8<%^
z25Tu~{puvV;LW%Ge`s#zKgce5+Mez1uJ7*y5sa3va^wB5{(%4%XzscTAURrnizyVS
zgN9sMMoralN;R$>&oz!>5y`=tSN86*l|T%Joo0k;iXY*gSM2;vjEC}Z<~_sCb5A4-
z_!eUc-_8>bzonP5k1=d)4bdTol#g^5)F-j*NnUE#I(MtHe|v^>eUwlQS8qp%fRSE~
zEXfK^C(asSUy?-NQ3n#RUrSZyfm*YgtIT#oKk%4{R|@?k3e)3f-aV-S{f^$uPdNaN
zcyp;h)NdUMnOnrW--N(pTH?m|6G5pTNJq=jsKiJFto6znhm^^nmn^@9BKvm9C0J#$
zA2H9(z{XBWf7pe`x4O?|lyjXQiM;<4hIy2Vo<5_hA|Z;IsB4Lp65ijVJ`&wp>s__K
z1D*+6LKI}+ZgHrt2FeOZU^;!iS~33wZn?&)-0%~w(_ZxG?u=}_8pRD8LWmm9h-qK|
zua7n7)&Q94`@*Az?`+zb6LIhYwusp(6?Mj=lH4nBe{D$uMztLadRjjs8#id|&mL?k
zx`{Ftx!S`^6|banzJF2KrkQ_WS%pi=jFBi*?~x`^Pp8a%UaC(x2&q)v5wKpKZ_?V4
zc1^9tJcmH52QVCgrGZ<P1O0EhG)gPE7FhW?f5yHmwd(?wCbY<iwmWoLdHm<*kL#Gi
zu?|0<f7G?|R5$UOCCBQiWwnZEhZEd>)h254a2}o(Me?hX0@RFWY0@f5KHm?rpVheB
z-XiWOH6Kql>st+3nkp2mVJ8e`9YH>c3;Zf-_4^jcX7TLo$=rq`s_n4az7e1ZviF&9
z8D}3R;vT~A89Xf3;sl5k#P8Lf?y<7o6XF7Nf2U#fLyqrA9S95+c!;%7u(e7Rm@7?N
zla(yIoj2=f&G7FJG#`%==Hh+%?^(z0nNY6Wx@!vYeK^+hquZ(njD}YZ+Vq7nuU%qj
z(p`ZyYlbe++6^WGQary3vSC-jF&(CxJQ}*%XC+$WeS0L!UMwPj_T<pW1BNg@o|l77
ze-4A7Y9l}dO8YG_v?84MXmL@%N3;m82HKRB`d%~~@N!1_kBGQWC&2L9Ygx|5=aXk%
z75Zttn$FHE@)~Ca+YMVPMsRL>ohobk3_2HX84pSJwHWhizm9ruM%c&pkw6N6`85V2
z4n{niiDw$w1~-X+Z&RU6MPmvcu>p@`f5L@BN(zG;NM#zvvJFGuu}Sm<OkyvY^icy*
z418Y2uekz&N5w`)6sK~QYrW|}nDlDiG@?ULAK#K^n(2%nMqk)|%eUg_D)I*+Q4&UV
zAZ%01u9~J13EL{I-tkE>^K?9f4Zcd<Bx-a!YP-_FhI*iCrViP>A%slz$=N?pe<U$E
zG;)Yt+oN1daN$K)Yk5^fJIP*skr33ApbQ)O`<j7It2Q@B|6~;4d9fM}_|{G)6O&<?
z@^VE7&4bB2{vbbw?~m{P1)_Kw=(4Z{LQ5=Mc7UD4J`n^aiiQ^o)=-ECMnyD2l!%Oi
zuMPJa3I<D~s*&UKibm_2g?!$0fBHse99H*oXOJETt+v7hBu`L6J(	WRU5RKK4S
z7FdEk5{(czYN7q5UNUF#pv&P0$)uu?kmXgawzji4AMQHxp|Q`;6#TokP!;uczew{B
zdeSjo>3n+8wYafOp%ABe)!Q-L?>uy8ihbUz&C&O5lG-kgPA~I@qLA*ve^1`WmLqZ}
zfw%ErURp87W0}?Pjtd)qOd&mnz}#oEm#=*Gsvu6Yhshy#R6w{&q#YHo859xvpY{GE
zpi)S{?d{63fp09-`{P%UiZCFDTUmSa4R7uyyPr8+X;3>-(8z5J7Fxfyk|3aQ3mu<y
zMqo&wMHU6aZU~EzffJW)e`HagD{hE2(InH`D@M<av$P(f>Vzg^H;;RlV?zfqM3r<*
zg~DW{Ne5OVyYdIGTEf&tF?<w}-aRI}pu2?-$3%Z=D`Ot;h&U*w*=XP?DJR^JN%3_U
znE9;xipNTC_?BLt1^_OPC(34uUEh8vZJ3S}N8Ze>JibZwM`h)`e;5~UuU@~dEba_D
zYnT8{opT~&>fzlnvWg&VnT}W=w?lrFW=6|T!JkT0U1?dWF3Szg@*RVBZ#XS|o|56}
z`V^1wJC%TuX?KJ*IjWynaZ)x!wzn&0XZiZ3#M|2=NkhSDe15NlqR>_x{fcoh6%aR|
zNM@=#z1GVP25O>nf3dtwjdFVW9jBLmcJq8RYzK_KBWmkHkBwH=b6zw2d^B>YF5hv?
zb2K?W6ZVD_Ba-7Z+N)H-H8_TTz<{>o!Xe9+Up7M8AtjZ*a4|eP3pQVkwRdX3xdS$T
zd(-ZdX3k#C^-191cPDeS%Vzn(1FX0Vvhw}O_U2J$x3P@Bf4YxXUmMC)_aM}|&&)!X
zgy#kZ73!Qwv!jc{W!K;?HPgqL>~dUMl3Vn_fAUfyY|ULbymA#4kI<?lo#Pr6lNtE=
zoa>+)0iqQ@N_?O)7=OG|{XXY-1uevkPJjyZL$_N&dg@2Mj6P=zB(zW;bND13=J+B+
z3l69{_{3zsf3%mq_7{yig^%2tyR*B&W`8xTlZ*DC=O2{>pWGo^O<D>$U{b$V8_oUb
zJaxyC!XKlR`y?`Cd}5b#xF7??r@cWNoJjt-O4JEWrVN8FY>Z<G6)SVC;~SEKH#^T$
zbQ~35!2N9)Ul!gLC{XId7Xkbgl5hc=n6`Jo1F>kde^?a77xdjz{Btj)zagkY1xSp{
z_xNCtN*T^O<~N}#ydK&(#gvXHjbyjs`~?VV67mX|gt@NV7A_|i70+knx3v}kCi^3z
z77><wEy}EsqHPx15bw4(KxMzZ&scvUQ{jhBM?3JqC>ae9H8@UuVBtwj^k$N8RPF`_
zPqG(vf3f;*z|nJ8HP6)=hpF>LRE4%V`&O^>|3VLBxCMzkEG!N~W8Le}&Kw(levl2S
z@2q^~UKc^vQ6Z9h5C$@i6>mLClxaibBFK0UHELjak(&0F0|zohcvYR+E4vcBtJJhu
z1G2FzoFv@^<bQ7UOU1bWNkF#0YDphd*I%#3B$g<8jDImAg5Na?x?F)Xa1>$j`}l!o
zhTR+M=eM+OYS>_UTS@W|xq|44`K#^!;J((v@I2p0u(uPa<LZ<MiID0sFKXRcwb|Cl
z0;zMgNN`}q1z<$QQ}%Spr|V@}`9)qTDV>mJAAJ^SMQ=diO``#IZ`=!aH985Y+#j`@
z2EHXaAAcT&17_dHM~_ex?`a}qdb@=Efsf&480+s_e;kw#{8oZ1qYkD9vC}x(Eq>RI
zgX%Cr9v8bTYyM_m($;5%jV5v73YnrG4IIwaeq*C=P-WfXYwcpVkIo>s%VC)5?Oa!2
z)6j47;f=zQk+CjKs`hh%tl-?lnxK{JxHZ{RKYtXCDW0@6a@<-i7ZeU&t+S-~^)kag
zlPJCeYPdd%M*QV#4Y3O#J=&_`c67{Uoai+8Cg*zqa6&4eGoa1~C$^b;J6gI+Tk?9b
zP58%GETn9~StoLd0*(CMWcH39ID3sF{p#mu1Tt@=gBNr;M~V&LuZ#Wqb19T`ZHV9v
zcYkC`I*cMsMKzT1143g+QQ`r6N7XJ|rhFcMBLiu!%d*|<;gHb-B$%~D94R}Psz8Jx
zzZ<#{1XN1WBxQ5#?I>thW}RbGCQV)EC%;VU)k2>?^zT+x$}rg!bFo{ZfNJYLcJ?bk
zw)>V}F9S2jvm|Yx={hKXzS-0Oe73`p*nikJaA}QCF7l>PlFdd6Bie<l$~Ijh)BA4g
zibyC4dUQ*N)>Py3hpDrq2h0<X`8PYf(;)nPc129;v^e-~tSz3=v@w>`9_sP-d2j2F
z!&Up15ZHUTA5Fr9HK~DgqC9`r%$c9ZxDZDnKT*|jGhV4na4^9?I^XHUP9Jig)PMDJ
zXq#hJ+=fBdbUW?^B{>}27<52d=kMROL+-llGdY59%{BheXkSohhAzo5DLc`GIvDL4
zQf*5}_V*aD#vo-1k79|6+j?KY(1w%c8a4nWXd283lxN(&QG)H(Q(C=T6hM7hNymB<
z((fE)g8sa!M>1F}<78l+#))zAwtsT@(rmsJF>dmix4u(HDJkA^8)k`g7V>$q7!v0B
z-8xPt1a$`^6}QlgLP>3=WLA3BV8JVH90Jj(?+Fi|ZV;24DF3}t+0x$Wf}%dxD7WD^
zJ*eKKl;~TM)8eDcDJ<DptB5WqwoOoaug1g}kUz9Jb%sfJFpiG788oS#qkqO|ePuTY
z;VTUTx1ITQ9FZnTmL@Y>LFnMlfl~v-bG+RBOL`ttO0wU9)^QX_JJc(AVfv8r4@}~%
z6N3h7!d#ymnJj;3PC;`tz(El4OW_I3tzN-|hxJL6)>})>{f2orbaR?i0S+0;``IYN
zC%Wv)K_K{oou$!2RCv@DH-Eoiq-eXX0E`RcOhm-H4iOzaHXsJ}_M!3jgO#Qc$kTFi
znf;TXIT`l0k@kg%Q>cms`hK5J&5OwQDE38YM}dQ1oAlFK+nyoQ+Xi`pj<>E`=X};>
zQjh9!m)IVGjX^cAPHPmziOiQP1#Q<}pdsT~=goqE5ldjxs9>$T+<#r)Zx8d*fijRi
zhdf{q{tHbb)bDkm4+yKYqYO+d)vTp?NFM|J0O$p&kEJXQ-MT*Q^QaidDG;g@G923I
z)Mz?uoxJQNV;q`XMwK9(0f=v-c$fm3P4r;hFMwyr$j_y!w2gy=9xL;MUI2e5)@_1q
zP-Qx?<vf{`H%?0sW`D5;HDpu~9ESra6%Vj@h~@6H%hr^fB{8jSz^G980}4-kH68ee
zOp-sw!ywU|dxnznG8fV|ekr)&i<C>bgfowk8nZu&YdmK%W_)*TfIc4ell*UhOC;ym
zJ%Bx>$A3Zuj+ep9u{CW&!<R@50O62u?y06hw4k{eD<^w<*?)q)>#p$gi@+Tk^h<HN
zMTr;_Q6Ts7()iwf;tjob>3+-)IIG+WX^6ela1RupcA5J3!cS}1&Q4*K^O7MhbX5}7
zWE?FOfg%IK<RHG@Gsl2>pvIMgh&;llH1r|lr2y<(RrLJzvU@)K0&1~|{P}1aems1#
zd}GrbF^^Hs+J8MSFF!TRg@iEh-JE5Kjcukz(2J(q0ZWN&8=tI7>_Fa(lT(nXA1$-k
zC*Ih<&NkXYTBPjR1qi@f;`_!*aOz9lsOysYI!B{{t6CNNRYZMLsFn`w#`qe373(o-
zo9=V|3MB8$uCRsOE=|)yvXP4s+kHo|glK1T1-t*iB!5~9vWRE=#3&Ho^oPX_Iiu<x
zBJt0QF@m)>-Y7V=LNy!R(;H|hwya{O<{3*ZD`OgRNF@#uq=kGUCgl>Ik#TP13<h@S
zCC^Q}D0>fM`2Z(mZ2M-0dC~n!nU8+8`>IxwyM&W-9}ZMeJ+SUo;)U8+IPdEH+0#SC
z5FT5Ah<|-Pt}-yhlQRCea6Q=zTC#XgBHD>IisNL3e_O<ax?hQ&au{op-l49IKBK8+
zmD#z395@ZiGo*;eRneIgga3A*;#T`Dc>-6zJZ_L{gi4maMC#~!7e#@DJ%OX2FEKHZ
zGAji>ZdKhW=zsblXwVmxk2*4ttKGN1K0fvt0e_?TLL@CuDGkfQgiK9%4589|<DDH@
z_j&3q6Cf$A_FdA>s9a6**kmfqLig2TthnaNT@yfC<GJ!cuNI>O2S+?rnBemrGvgjR
z#(<`Y?-vqeRD}wo+k3t*6@fEia#j^0deCN-w)c0Ww&qPDG{RRN^NsMI|3om)#XmSm
z1%DGE2>jC?<hkP7LI!UD)mFXqh9}EUI|h|ubT&hps}-%G8MW&pxU0dOpb6e|UQNUY
zm4GEx4B5Ia-M3X(kRBPBU$?FL5+c&hVMN39x$`T^M|@>Sx%IpR;|(68-JTlvta*7-
zoff-*vQ`sCba|^n{ci5XHrW)Ty1YEev448A4db!Lt1URJNMl<GZvs<?Dag~J>h;aV
zO{4n(tr+2<J$79NqX%M{)BIj4^^)Ah6#sD9A&d5MvVZ~df!~d&blxsMmVc_*VNKPK
zS)A=;TuP8`bDk)ns}NJcL5Re69{uaGtTI<DGG}RJG8CRZDCnw7V2YA`ZDXT+kAF3H
z$J;Z&tD$ht3thEXd@AGhvZJz##w*qSOMRnosJyhrewcDLpz1vY2bWt{{RMW+VJkyi
zOE*SqhU)r|hv;S6;oxMT{egKqg>88~sc~pCafQ9e%2Jfe-iJQ=u_jA(S5%i#KaW-_
zH>bR1Jpk{>Cfs!U9yAQAQBd2KIDg|*?MyOxZ~BuWQB)c9J^|WEE(w%Ca@>zG3dxg;
z3xddV|2#vscrL<Stb;M=We*}&XfvtUVL0ij$H(4di-wB3t~0=My8G<$4_EfD;^6k&
zfyby{EU+7aC8eww$n_Tovn7>^SnT5&9s(oZypQ*?sOrbNq_3opa4gTkE`JHWT08MF
zKX&;!TtA4YjS{g=7V^DI&c1PrMuVtEU&0PUbaG<)#5pDWoaBBsF`o~)W<t9FJkPWx
zk**CI7h>Yst!^C(e0v4b-Jn^+GpxdK6B^P6wu9JYzW@aEPsA`dhWT1IhX}3FOJ(96
z@Vr0>m0Lk;Hm2%v2U-)+r+;{m+)K}#zix`Fp7Mh2r{2e<<U@~^{MTLL_wjI#H*0+F
z;W|8dAQJoC<xww7apK5vxhbiB{@HqbVYM(v{G+m!aTZ~k(%afu%ms^~M%mpCs&Cez
zi$L5H#Kn1XD95>YhSU`ke<T66LYn)P!ix~Tc*Td_sj4GGj8gJA`F~)Nha9*im~xgW
zGBVFI<MO6d&8jfQDTMW>>#>pNk-(Tb{Ff;79XY0EClH^HU+sa%^7gsb%Sl;0J2&VB
zl`o`S$uhOL!^Wjrz>&#yIxQ)a5TOfVg7=(VVOq(&jU}6brP4G1n#stG);eRBC23en
zysLFMsaj#Vqq6O-qkkvQFz00sC2W7+cK_T3P9<8p2?YCU8qk2Ga<%?1w_QISS;PRs
zP`@d?g?PsBfc(?YPvC$^eTA;(l`i|Xa4Jfv(3l!^S4Wu#GyP^v{!e(Xdb8W&KjfH6
zC~`f(lDFo;r+Juj@tX<g*~(L&s2P;EN^Gz18GRu~>JJ<z5`R<T`tGa5_sB9UctV1f
z#V2%q)IUS_E9!K8`1id1WfOq_zNkHJKC5&N(9w?D6y}~fR$TA&HRL72C$f|9=5bC?
zy>z5}>V%S=2em&Qa9M)nU?Z8?)VpJmD3IBTX&U@g<#>#s@#4Y2YSVG*43^8M$X3(X
zscs#1CNbDS(tmyNW_&?WZrno4Zc?Myn^RPZYcEP-frN;XY-UwacmiV$P3Rq?2`1~<
ze-NI3`ZY{SqKErLL~1B0?;i1sL+3vc{<6B^{`$F?9iC-~gt$-y!fSL8D88Es56a(T
z7ze68X>>3ik$MH}GTM1lmx3L3o40)_l9HCmOUeGO#eY{e0JVue5c4ahU3@@z&W40l
zg#NLtlN{wi#0|&t+;;fWpr5SvweVxJnlO9b;-l$|6-#VRxYSp86mxxR%5pT^wvsp|
zN8;cl)>l4kSOZA4S|SRvcRMU>=W_u8c%d<Vp+P$+Bd9@dZtQDOE4%0i+)utaGc?Ru
z3R}s;@_*OM^`Ev`P=lBbG*Rh(Rh#i9iKV5l1vj^WBq~KoWVg)DP`DTy`=*zY7g5#%
zUS8|PdBU-a7ketO$>LEklvA1bvt3Ztd|6$x1;jyYl1Rt(e#RUXq35sAi63B(t0nJ~
z9yWXFlx)FNT?@s4DY|>+!RL`x-J5W$uc*ZzlYg2#S!B-%L&Wbg%g{-V!SzF4aV4d3
z2*q^~l!+QP_-A3|LjtHHf4p?yzIfY}J7WW?x%6etMd;QlPL+B-d)}dduvo%1DVbcX
z9jCZQ%Ey?{?LpBWAWQ{dzacK*^D!W-KiR%>`1+QNc;+^#9glq5c(0F3aGhHyB2zgz
z(0`R{PM#wHyzUvKQe0jJoMDw8mGo?m#eBCLIX2bBMVB-gEbw%J6WpN|I!<+HN+C~$
z-59$qir9`HKY!#!o@gSO_YhMSpc%Hcv+u9xx8PGQ&5AHG<vHH#mxLgYn8p&o?`Mia
zLQg$I+1eQ6+)kn5X!EfT@blznQXU@kPk$dIw49{Ea4=g#4>DnbS2N8xO03y$)f5YU
z<z~DlXM0!HLr#;7eb1u5P57@R-IeHr`Zo{63-LHgul%a)(jdEz3u-*4)2ZWat%8?s
zOHpIHe;tfaHUS`TA3d%43G&F#Y2l#m!!&~EQZR6PasvfcuZx4LN!ImQIk800PJeKI
zs3Uo2p0^3}RjTO{=4o$m*__?J6s@OMp|&3LEqXMd#BPi6KNs(28OzSHK5}|KuiEQC
zik1z~5~j-wcQ=CS6?FSX7gnmt9sq^EfZpL<2DTJehc$n}QkFt4laioZl8GRTC;?xe
z+PZ>>#&Jb=8zquM8Rh17^c=bA^nZh)do1qyBL~2m2Wk@k!gyycS0JqFj#{SK9^gwO
z=ny0?;HOpQzRC1sqqAaHg_@9DND1|)XhNR2{bfDM)7`(<t#=|PPUt28Cg<3~ZUTmE
zoOU(rEqD&f01(XzC-ciL78<9ackE5FopuJNQwT{+b&k15*Xjf;>?H?JC4cU`lN|>*
zv>9(WzCA9D+BaVni76XUj8Q|irS~udx+I5H)mvmn870&hg*j()%Yv<C=k#n(nCtY7
zU&Hr35f?Y3tIao{F<cVjaxVTRi7z-lnPI&y!5AJqdjTvoOf+c4)IQIEkpJ+rkY5!_
zwrTazy?2PUHd-R^;Nt@V1Aq5<-C_<2`AQEiZB~uq9)OUT3T{aDNkWK-Rv&=tHMWx0
zo9pR-yUFiw%<kx|u%kFiTw^rrS%(gwjf>X3!7ET8+$9O95cW)-^Ol?zb}ijkjFTtV
z`Z3Gg#;z>nSW)+)N}I81Mc8nc;0lj|Nrz4LhW3>u_5D<%!6jA02!GZYHSa)z`<lM@
z8XN0U#8(M}g$3)1hp9^yBDTWaI@+Wq&*KO|BI2W7wN@K4ZOL9(N?Jrj$;Fc^&iAzj
zg2Nxs<b`(f6t1nQ(c>BMONjdW+B700L4`zX<$WLw#Rzd+x!xarsUa_#3e?ac3R2K<
z5W!pnl;PPe?6IK0TYoe?K4U0I&4zh>$g@duCPE}wljLz~D@}(GGb75xA{+`joFjwc
z`dK)w4ZZ4aaQmYxisM@6D)VG;5kO+w1$%dcOFFOUJ#}9+{sa&l8#tN+g!r&DUn7|Z
zOuSH($F+U&YFtb_3rO-K)72^W<U}?6X&M{C!5x9%EZU>Q@qY;yMicSwGv0p5Ibawx
zO(o9IrAWckJp2QjYM}cO_+eJ2Pr;hIagMrh@T<YGSr-j+K`R%g6?A<hYU>eQeW>qf
z{}7edhU5#dVs)&&x+DZEWvQV|X;%uG<x2)^wX8t-TL*+h>QfL6YgflCD504e{N2;3
z9H~c2Mm)cX{(rnRYyleJX})AP<law#hKklm^c>_ebC~yasKC#8FzNRqD1RS!l$O60
zX0!4jw``MyqiHgJ5Sl(S)MaGmc*G@WF+XBlBlp&bAVcwn#>v>ku#z%d-{8IUKDr1!
z$Y?Dc_aOlnP`<~h^V3&2<5;aJ-*yI`(4)pgDO5o43x9%(&EwL1minQX1QLmdemvE;
zCuv~u*Kqk5_fb{TxQvxJ>^eYfq@O4NZ`;*0|LV(VypDi<Uq%7{YM~u=?YB?zbhbED
zg%U3YMR2W&QX9@8skyL`HhY<_P}U`4KPr(C)}&Mcj=|{)oAT<%pLxNyuNkHK332x^
zv8AM|4u7Ilf*&Fct4UsD(iT|N(ZUR&qpkab84d=7Q4!pvSwtC`POwDRaVCiMxgt$a
z#BDnkqpXg?IfX>`EO;`VMQ7)7HMS}}h7hCGMV1tD-7)j!w@l9S3J8=d<IdX~iEVkv
zuO}yh@tl#9#1qyJKHV~GNq13c!M2Vg=YDPOD1SRQM8rK<&Eo0Fa~#mDSbsp>ZaUFr
zLyHS3o`5LC(({{MzQ(cU?n(xnsNKHhTk{r{ZHw<wi-50Vp6@_bAYbEVdX(-+#^;#m
zet^0yrCl?Mv8f4*D<(l|n6X%lLGgv0ouGT|-D`8E@NB@YK)fBL-fRT1*?J$q9w#>z
zrGGqI-GV8AqgPz>q*YC4>+Osf3xm!3|FL_~$ifwFP51DTM{tIPava+>Y+*CtG92XC
zF}fWwLf!??Ohp{1@d*v^d3#lR`?wOY>aVlzSF~0Yb|mPH(Ht&RusYMHj?+No?xcqd
zrEA_>3kty6vpKL&N=z#7h@D#2=Fcapy?-RiOEObLBH))@s?WB5)}r005#<g`kN_v7
z%VSIP8DisXM8V%1{&L2pZFfON0A`}YKe|~7A?Se{Gw(4cqzz4}=3$~SLiCgFv2mIy
z%jW7eom(W(dJrS+&wF#I{K?>o6slMUboP!(P1Y>})EH=0j_WcL*emG=7u#@ydVj<N
z6>0&wjh4tX&SV^3z##o;aaDiCV=F^^A|D$k9FYT$hS#C4QP!o#soC=~!}4yEE7Y#)
z?8bcPl?OkoR|XF3p6$M5N3r93$BeZoKyMHo4-!7Az#PJLy^vh!%f)<^s)A3ec){6@
zW%kK1NZxCSzp^Eooy$qG+Pe=UAb*>8zlp~AU|h_gVb2U_*cmyX8}k^HS~pf!jUox-
zo3J>=RPC}sI#Vw9gQJ{*P{eSWXt3u=oWH@r0;NHlkJb8tJ#vE5&u*yEvAZpPNe!Dy
zYive9OOOYfRc;iJtkrocg%N6{?0WY&?mtRwztz(Ktx&j_1?J#kNG3T?M}M>Og7zoo
zAoDq_p@UUCj6DAnq6KViG|-nwl-n1A-X7X-rYP9a>O%t*Zf~iX!ZCFs(AYnMPBJmE
zVjJHY)EQ<`BSGG~>cjN-jwkfe^P_I#QEPY>Zwf-rsfm#yWL9%Dae5$CgG>Z!`E9*M
zZ72Rd>lU_2_j;LKh-6*rkAL0LOvagPnf_zJUL#wKeB>5lbW>}nH_gazRfR$QbWVW>
zHGGP}QQDyQF*NPj0H@lvcn#UKGNfOC-=q%=D<gEj`YGb59ER`94^|h%ug58aA6+2r
zy6794>?C<9D!S6*`eWSs>Etkem;RmOQ=<`4VLA%Y@R7EmqtV(+`+r>7l;e2u;FBxY
zVH`H)YPcSGw+xPshg2|@)whlOr{o)VtM$zFp~`0|LeQ(*GZ~-zwGo!OOQ5`e$nO@s
z$1%V%nVE>x)#6z~dP@<*ae)H&Jne_r!S~laa0@`*vFi0*nh||*_br8+pPCfIxxjyn
zkm>&#Y)$Fv$xIskU4I)Z7L6|Yg+Tb*VUw1tn2aMbV_paQmBjEMJ4guZjQzKZ6$Al7
z^GcdjPjslg&7x?vc8SFHqb1|5Y=_<4kh*7k6VSBW>29%S2!XZXWx*gM(QbMB=7PK&
zIO%ttS?J%tcUXw+KD};?SG+s2%h(CJGT)<e*hbHw=&6#lPJe)$UMj_I%luhmM9q9I
zQl717kDRi8qOGKSMEMf@R)w1SC&yQ~Mz1$8D)QO_9#A}VP0m-N@%W_+j?kE_WqDz|
zHN(v|GK}HhGA68w)+np-FWr0Fa8R6043_$`ujl1@eq`a8g(#ImaRUtJ&!EAy@J0;)
zs}Rzs@r@md9e<7p%O4o;_nXf|%UNGKj-P`4ZMaL`uaL2VHQ6E;s~b23V!gLaj2?cr
z1OOTClu#S&tJ!u_tYGb%lSd3qeNh)M4-QR~OA;>DqYo+WwjmQW0?xb#ET}rUxHCo~
zEJ-Z2s23pnXkaMJg_^RVw8?=h@g?^zouFi*^Mu;-mVXnmz*D1pR}WTa0^uqS-xQB8
zUh-Ex{}u2cbzzBMVm)J6(c~XAD60iuer_1HSM)a^*BQ+=TNDm3hF(fv!&{iGMgF*o
zOVw|sc4k|^$T0`@qj2KH3I$y04tW3zr>d$c{NYO?^ZAu%L2(D*^soJ9$5W*1Ev~l2
zX|%`M41a>!z4Hm%T(6MBm(KPgIE=f)>tin}NP@JLs+<yvaG%4C5-+TcvzcpPPqdZ#
zq|7A4izCcH^3O8N5?&sIPa}OuM#Jvd>jx^a=!+E{f|MN%Uxo4=C$f4jErtiI*%c+j
zrddXEoIDK96dcyZpzcAD>ZTFKOgJ{8$yaQgpnr2hrAyzTEQM$_KqyrPd&{Do-Z#!T
zee+rXB}YJljIH=#Jc=0Q)Fqh7s=7EU(GPa*v_N;MVlSm<c$@0k6nM5mr`X=6BnDp;
zO>Jo6YT-KLcH7#Zw#H`oL(&^yU+gK~n#?bgC~#b4q9qv==qE(HYK=6tC{M_H6xcqR
zx_<_GhS9{}RoeY<wlAJ<eT=Ea+XUYxuRk-(;+{cQdKa0yInrXx07BTSe9y98iOo
zzy)pW;fPq<$166f&q^lRc%Hf7iK(!&N!p=RmmfgQ^Nb4E8(*^%8I|5WR|vGd3)L&h
zU4p_wtg}@!n0#({^)3L9@*re>;(=eWz<<$sGXP!SL49xY-&)xLR?e&GNP4hWD93_R
zx!YQ@j6d0lnieBZ6NSufY1WHtV<VofIsKh1;-|0`QMEBZuw`Xdmtn2dJ$^FGcA{Zl
zfz<`aj_*3)sA05m&dxNiGg7D~&Zky}LNv*L$6zEcevN><YeyKn7#L;?$B+AV@qenf
zRdi>oQd}(Z3{-ePwo43aR?>D_p6l-w=h%enxSeyfue^N{B1MY=b$PA*n58~X+3z-<
zWl2wUjeu8SIozzgc`rL@5-l13(5gQ!CIM@ufG4C|tG9`&XI{N|&9+CM0(!zvj=Ip8
zt4BRXS)-ssPTzIiI%?<xQ^f&v#D5D4G=y<^By7aN{)(i+(^I-{l16j?v1Kq^>hmjs
z(ilab72XU-JP&IB(+7|eltc99ZudF(haxDBg(#{v8mMHzeU6*)i&Yu(_I7Be@xIQG
zrT9Z)_sIqc>w+ItzBFtohz~U$r378ga5m94>8Wv<-?R~n4T^c0ba)JBvVX8~RV7<u
zTAgbR(&G$3CLfA}*r@1fHF|39PwFHV4%$$sOTic%#{e6ll)Raf%Ob*7x0i+u+@bo(
z>BvBkOT4=#&UhLqvR9o#dWOIFyro3S%XlWd)H$Q(aK^)?Sp+&~u~ubNH`|NWc>kqO
zQK_akuu%j8_RCA_aBS2rzJI^;_}pD?@rb$HM*eJ51Ny7wC6RBE*T3$2Ei83YD$Z<o
zz;p8)jZ4LBH|hNG4GR3JP?x;~r8yTwd1gZGDcIU&C0Ir>nNY8DPbK?6Gt>o&aAU(&
z@o{rv!<&0*2Qy1E)i4-@3bfzOreaZmqfwvao6hxdD9u{O6^|B&&wng~vdD%}LmB<J
z++J5N-38@{zs@_9+se5thzdUFU`FQ$nf|_)AQ27=tanI3?YJ#B4}#d?6OUZX%=wq8
zD&^4Ry<WKXZ<!|I$k~{o29Vv}@~%Ps$qkl+r<Hrz=EnU>(p90u5BO>(EY!~SAT_;@
zV9IN?%ARKtonCveOn(MqK`UbLrS@b}Rr8g8YB|1(S6rdK(<)n{^&QHM2onjai*azC
zoH!gs?d5%~Sq3!*oL%7pumTFh55@X?Vui4=&nt4)sVsbK<+Dzw3`DbwiNa>~y&}E`
zh{Z7h95Ibffggnn3OpuxS;%|I3!O}?XALzC>P`mb&Jff20Dm6Yg!iUJzU0ar)|Z4b
zEx%Ha+TP6GYLM>uvlQmCscVxBG#Rbyo`h$1_eV^<HJ*#cAllzCoQEdhxYm4}2{7B7
zU!n=st&i*SF<8HgnqkvHZ%X~(cd&eke&yIE_OuUKHBz6oY;ROkHd5x0e|^FbwV<0z
zKftyjSNq1eg?}jyqe$X{yv8AkLxD~IUMW_e3x5<?)s7Zv$q$#CV)UJ!Df&Q<a+kb9
z#WmA|vH!%OIe(9{9;S=nR>=NQO#8MdmHzuc(3)aGW<PW-$8!(K5BK;54k06g@b}JZ
z&Zv#MP`Irx#F(Cxf7T!C1d@~W&sR!<9V!Z^bR{&**?(HK<t$rbV#}7C6Ei7?!@gkF
z^Z{DK;#EPcB8~Gf8g~N&DguB$LjzyzqIRDOz;tv8E4j)^mZod+hxFZV)fXcU^V~ip
z^|#azCj5H7bsN0?wf5Q08VCdLEB##asBGnMWZ=kV-@8wTQ&BJG-iCv(xfwaz4@eZA
z&0JEyTz`f<t+^vX7Tr=df?on*j$%p)v>bGDKoi6uL8iRqN1b7vD;ZsW*!U-X4PSTq
zu19b<w;Lp6d5_4L_Y5(Fp)uTgAKkyG&)#Yy{*)p&*zrgIakde3)^2#m_ODU*v{!$D
zc0o$MjCb5ik<2EXGnHp2bz_4w93qPM4u<5JlYbamF2Wt?1+^mO&*N3(B$5gUu%^&j
z-P(G{O{v~9(GqDUj<S(}-u8>9<-0H0%Dk^Tk}|dy8szhULg+$T;&`nzF14ACJJvii
zo8*jDCer?zVU`4%qUy4%(E@!)=cOt~o;D)Y{{9+)_F+4mK!!0nlPxR&|8fctIkX^i
z?tc)FlnI)LOKDfItv=ErEqXKRd^GYLp86TQLO5BAI|(yK@`HQGPpF8d4C3r)fiJJu
z4YAWK@+4Tw^@4ORyuJyrJaSiNmTaN8El<)zJLU%Qd)G`1UeX$9cM0n~mOKOp6r$Mb
zwdi-!^XYSO;?P)+qd-&UI8ad&(Z`0*RDaCAnXOW{5UD6yW{92}lhpw6uXLM04GOfn
zJG+B{Nl=FLFCNzf{sCmOSx~6JwqdTHuG7!y-0O1Xrnum7c3P{!E*CJ{w~piH($1*N
zn*Fagbwm9q71Uu!y!#aZNC_bwrPI8D!$^o(WQ28~y9~MfnCLSbN3PFzB}k4anSWbt
z<^6C7*L~v)bzmcbv1q}QI8C@r;v?~@wE8dtdy_K(3Kz~b-tUiNHh}=a)O2`y`ngB|
zvNAHNKU|4$nNBe=pp^iYX;?Lc95f%(morhI1NK^l8A<+Nu5d}p2u36|ps28<LI6C)
zik;S)9_p;>!15W`9|pTzC17LUY=6Dhw08SxD0vhE85e)a60j1+IFAW@_55r*TFi@y
z0f%f?aTbxTws~;AshTa*IHf0)Yl>Z=3kKP^U`|TxueyrFz3DU;?bu~@<DVjT9=Ih`
zR_YB&(jLfP=b8jt*N(%zMJLF4Voh<1+Wt#PnhA8`t3eUb$1rUZ*E)UA?0<YKr;Ukd
zJfePQ@Jmt@OQ8tgimO``H>FG~hN}K&6C^34biIdHJtHc|qz4_xA+|oB6!NN2#nYXQ
z11kZSO@X#t_Y}qV#UZCihw358syK1l7p&<4Rj!m_&FY?(iT~v;o)g;8=XT*&YvroF
zm71qkcq=l`C%-WDzLDtnFMo%)6NYputf?yNzzDQ{6R_V|2v*PGJ0%SxBB-D1BQ~7Q
zqo5@xD-}ehkTEnM?0m_Xm-S6*Qrwcn68fO>qT1ziEpD$tVrncrZ-TN)VW&-uie<NT
zHt>I<1poe$WXxRFaXuI;FUzoINToelwuv1WWVVN<+~B}qv3>qg^nYAmc7`c%rT2<R
z0L&8lsW|n~0N`UE$SOzUYo<GHpG*;y6$IMG?^^X|(EMc~X{Hq&#MjIsRLRk|->7)n
zF3)Q8^)lK@(2)G2_hwX04G}9fxj}y<33+}qWpw=Y<8d8Jfx)3H7^|v+6@1+0lzsp5
zv^G_g4IB}2vYW6*+<y%A3O*+)d}F2pB!=IjpOjVpiBx7vG+^3C3~8M<Dk60VLs8bJ
z4;f3nKu@ctHEhJ-xuw4v#e1u}4sKSD`1s-{V@MD#xA>0wR!6NeQq_Yn0)z{cX7Ssh
zH^Nd(C@^KH8jP?}Jod8`&yj!PH^_8t=})-U!yekiP7KcLJAZ4|if&!hy}MHiII9(f
zquE??v;x9T=aKCg;~l+m6)CmL{_H@75ni-cgFXN{it2JB3{;k;zQ{02p!U->o4!_;
z<?luq5Wfvee)_Sw%i{LpkLUPrW<u&{Bf=&ww9gsvsfG3IhAIR%IbGNbE#OeXw0HOT
zV%Lc2w~T_w&VPb~i0NlGwN=i-F79zpQe#e|LUi*zh#AwVl~7EU>An&;2S(P#Tj^f6
z97j`MWaqFQv0?aQX)0)v$UxBhSU<-u^XLQBQw7AzeKd5cKzJW<{PMfja^idH%8TxH
z92I8$aSDQdRa~#%S5tM>!r<lm3o3QHS1$!i5p;R>5r5_JXpHU@gfo*Rkml<9A-odd
zav;@+r5F7aXHJYCvuxxfX^c9OS9#Hls_*Q6dE2vuN6a)4CY$F2Q=<dUbMagBkyb!;
z#dA@91E|d?KaICCDS@fcNF~aYSz{vNNHq}?nqq0CdCOl)fYgOuf?`fxhj(Lyjx=pN
zdT~%Q$A4&dg;fMQz03kp17KC7NX0J-t<v+Hbj*#fm!iIaAD#KCVBntms%<ngc{y|F
zc%UTtMfRiy@v#MG4E!7FelkqAcT6>`jP4?&0@u7JM*Y$G298NjB!W(UTG-u#*1Ibo
za8z}qNDEwEX`CAQmGaz_&s$ch)#T;XMrFSan}3@jNU+F+M#4%((uLy-)~yVjU4D!m
z{8lyi680Lvqm4k_lL$Zg0FkA41}g|eD*IT_8C$2@fqvRL`=AxZ&Hma_=CHN8Pzf$9
z7=xN5NX&D3Pt0C$SYT>^&@!wk=)xygB?#&3cm!p#0LF#jMFBUgcAg!%UZ))XvRPT*
zK7ZN$cD0rat$i}3@AhT_jwmYNN~KHWA60KhVw`;OeuEDEg_;l2|9W(BBlJqe86YMU
zPjwSFw~e<UJ@=aO6f!K4yV&3A!2O=SZ6vtKR=4>ml~Ks1>DcD7eZNFo2(uJQnoCfu
zOJOH*YYSQyLq?785(l4P{|>3f&B-CV^ncr~ILr*^9nayiJiC=f&T9;Dj`kS4QiHI<
zQkS#jL5M>wq3fV!tm)R?qjy57iLGJ?oZ+Gz`pA%!H6D8449~sV_C5@ySkVSL;Vd4z
z6YdQ0tT?m)DKDeGayjJ79B!az;F&GAEX)jVzAgAXiR%Us06To;8mz}Q$iE5ZV1F-=
z-F%^{+YmC-KQ!f7Gll%Rw~jxfQ^Bm~CuNJReaw6N+31{_Tk5|5JT|qslP#$g$R73r
zRzFH65EC!&2F+jUF%T9NxHNUqNP-*0m2hrR6Cb2G{my0Ltw=R?^|L#N0pG+-fVK^#
z$10_LfDq#iYw>{o=S>+I>GLAr{(sQ@mkX&DisetG;i2ce6&|i&mXZ^(ykeZb*~fxx
zQ;*K1>q30@5g;)q-#apfpYl41yEt`OF!@=7aTZ&AcPcn5x$AVBiW8G8QmT~~!hFTf
zq^<9iZO2vlg^v(JaM?#nuNA1WW@?@h0g&f&9DhE{Yg&jEC2TOtaA<<AtAD)PW9(Q7
zdJz`6zqD78(NhBq<ZC@{>QsU-Ftn;8q1uJ#6{VJDEqw&7E@2LL%nJ&EkUv`>Vh1Ri
zfi}=+&L@$spp%NyRBPcR<u4O%$n7o15JNby5{1Zpv*c>3#ss@=5e+{Ll1s1If0Enu
z*dLLTM%JQaypQJ?G=-#bGk*hMPUl-6(x<~<MngKriZ{_P9vgZMpvxMfL?ne3KC?{D
zv>+}EF!?$r2AV;ULruDNpd>#seM?hSYrue-&VNgmHqx)4b}_uN3o+fGCqlFA21byM
z)i7$pxuM&L`!|4gd7|`(>9=60_59EZwx?RxgT1&w_^9@%di0kQqJIa!0`mM2d_I=f
z_1q!=ea0xxp0NP#2h8rfnxpMM2gnk-59u5I(dk=dSsg4NXmd>K-LGLBAB7D>p$G>p
z)`ALv$_#1$`JGQv1p3GCO2By(YFx$gV1#phh0`ko!R;DD_fE6eZPznk5890fX$1*&
zZF20jrE!2B@^;XxXn&<S(HYVN&9xx~%7$z~pQ;@GPbJAfYhN@&k2`&pp3>YsWPN-Z
zq<pY{$tg9*k5ecJ)*(0X(3|_9-n7yWtlZH0sgk{n7Km4w-^@<<zanmO-^B^ygqTG@
zP{R+Z$Jt=cu;k+>6rQg_+%NR;UZ?G8BZ>ChmohL~fb`!NLx0gFffu4<?RD?mDGw=Y
zxVj9{s=G?<X<o%&E(f!eo?cWd-nk)}!nX+#UP<C3AGgP28q{Y4g}Q4<WuzJ2$UyC$
z9pO$KqkHFjz271>J3M}~k1wHZP`TBSa-M;PdQ;4H!9*;M{8Fh@05>EO(I^6|F$GG#
z2TJ$Z)krPjm|fYAsCR+^>`&@K6w&N7#@Ah~L1B6klxP323T&dPq16#oi`thCC-i!G
zjRM<g!un}If+wtsfOXWC%WNjaF}Z@|dd!d2A#`up6rA1yY3=#Os(YF$@IL?@j*W|#
zL8BcLx2sGYLpcI8GncWM8z_J6Tv>11$d-Qhui)bh8Zm2MfW`&9v=cjV;<d9h@*vx^
zt!^!DN$w=~*Uvepie!-#70cGeV1PluVzD-!<vU9i<%o=BJo8YPOh}b+ok;;zlWEB^
zp)(@{v}DfkOu|4dGifp|gv+GPxD;$ldYC0tCLJsS%P5g?%VefF+`WI!l+Czv@C@_3
z)tMHVV2Wj$X95<;w1qcwxWj7|Q(N#Rp<x{Z?>Ee4MrFc5zmb{nOl1aM&6&-Ng}e8e
zg;xt}G7F?h=(NzTl**t<YH79{1CoJg>oaMAZXiWESjPcl&YR2u%iaQuFwc9qJ77V>
zIG-s6^E`(;tP5RAIkA7`ffWT3^9t|K4lHV*2bNOa=*-(pYhVJp;O;XZQ5y%sV8F9O
zQ8Nxqp*WbxKvHUpgSf2rI7VRwB+4`ph@xdcp>sIP0!csw2bp-qu?4$9RvG6w)q{{Y
zm-r;H)c6Df;N0R9>Vd-y?m!7#zzo5VECFLd@YseIiku4=D;$54C}4(U8cP@}C2&Qn
zO9RAn!hqv?i%XaR+5-e3(TdhX1SrFyO~DN1uvbC5rg2D*7PzuP1Stn^LteDSxd1dk
z6TU;klK}y4;Q))GLK#|^kYR%X0U?pr0<e%j>p%m5S`JDE(88M>cTi#r03}1tow6t)
zr~qQOG>);TGPr*u!~i86?tmqDlZR>sw0enl0CxrQ6WC#}#1;$KM1^<{#2BE5G0nGr
z{dFt9`0-F=`N`m}*vj{&lSMIE%rgl_cea(E74zxi?7EnP1yaLtaWfq3PQPb=FlaD9
zvGwBX7EGVPJdhI&``5o{<pWqP`}Nl>KL7zy4^{zlu?l}eHL(gBwy_FE?aEdK{Km0}
z#j@F;U}7C%Ky`R&elnY0pBIbl4`68jAj>a`?~CkfNzYPFx$Us3U&d?AV8Zrtu4?L=
zT&0phjhE}80?T%X8jf|1kO_58$m&SdfJ>7vk(21iNo-6`HF3Q%>k&@PV4qYI@ui|^
zh*F?A#dd#=rdrDB(}$}+i|at};TXEqY7xR9klNl)L@7JgCHbaAnVQUU5T(+2ibNWs
zCvVl3R6}X0F8eT=Dz?;MSKJW%*ha+NL9W#Q0dJNvo{ByNV*=OIht)buC9@`SYhgdr
z!WHYt$!!Xj313T%{kbG}EJ;1hl9c|rQ?iget-^ozb7^)Rn$5;&ekpOfBT&6DfkHl`
z)oO<U`IFYm3}?B8EDs8eR3UXQ&yhz*S^7VutfudS1U@o+I98e7h#xef&5L0#<7KHI
z?9p4QtVD-B!94MOsX80JBk_Z&f2=X|Or0Cuk(1oK&Ku(Ek=f4^<U~%e88l|844Tk_
z?X7>`2%LOJO!i4)vV1L9VQ~lg5`Ef^KCRpop_>d0w&QE8*%Y0te2vMh*aRRP@6hiz
zM*p&4{7gAJ)uz*l|3y0gFr07mEGfT28-w@%|Mh=Q5<sz*a$7+B*4q~L9tP@wq-_bM
z-*(Wiyx;a2mB^nm)M~xkHgM-mSZdIC7{`C`*b|n4Eu+TMNAM9?m$Ze%(vWw<J_!K;
zTLk%$<^QN3(siLjU5W4xyc|5+WPN%kGKFVJaXj8}{c_IFu5uk*qr9-)3p6htF5u4!
zQ9WZFFQN~O3Hy;hq%R#?-sxT1RL?XH+qJ3WLUO~?QW#UxRnmy>DDNq6SjVeuw{d@N
zgFlpcT)$kykPlFFK_4nyJ5YZtxdB+j(h3Ahr_@fu$WW&3KsB~&y$zgMKuKA8i?gV-
z(tsNYr54m1%4E&7w4l-S;!LS+Ro<-UKxZh^Sw=d5aJO^}Y3r6;5TWtr$koLX5P_F5
zD3R5rbM%fUlyQ7qyLw0BXdTItb9#S>Vkg1ICa%M6)f1x$9wepX5@FT8k}Db-rO<rc
zikt7u0!;aMS9pRgO+Bt0Es&mgL(@;y5yk|NaK`1fuWzK^c%x%@4jr&8I4_CsH5P`2
z(mH6%q#?}1Gu0HLP#OtKhtdd-FXM=QP3|z(y#RzsbbE1DXpeJ4yA$X{sfB+tD-n`3
zD4h@nX^A{(^0XSbZ1#G|w^gji@kk0V<bVuxoXAS3Yp0jipHhObTs3Xo$xm83-uipe
zcQ<<)Q=4s;p{zo=Vv+=1s%nmbcos{5j}9op8)Z)z*J2sQL~)QNVwo@z-tD;swOfsv
zHiDr^IfRelhFpdtEjk3Gb*O(!SRcUU$oOpSthiJ`?F4s#19%<gVGy<eej!)23@0Iw
zfIuEf9~jp0UC1N4K__I$$^a#R977HmrjVnDe#LN6B`rDvj}$3`f|43-Du6yZPIxJP
zgi~>}w)mA1(TMWEHW;ZGnFkVW1ZELN6MDgxab-ijBtPNM&2uynfuk0+LZfGxH8hLh
zJfaJEV(_j5VLjJYE0-IM9X5ZB<C3JQqCJ<P%_|ip0#c>5uouHW>M<>Ase~}B1-cp~
z5FSd>ihw+nH0oJ7q*U4HG4R7J_JL;(<M7iSqS8SM0Qn!T;gEmLkY@$>1iga(pxro_
zL%tf*J6MNfLV`R}kFkh-;}wN>q60HV>?LQ3VGuYMWLFYx#(+dY#vp$Y&<v(PobX8~
zH4ee-K`UV{N*p9fNLp&)@fJK`C{tJRk}_Sh&&L{0ng<<2X;8<GqYZhfC|skpTS0hn
zp`EOd8A?aMu(hX<B+>d=7{U~~7ej^&aKsm)Xq+K9F^-PKDuOk^)S~Di&L~<RWos-B
z!fw!_!Pt_PP06XSNYH-)!{DYl;z)bKgTf5MqP=P)CFoo+71KM=7_$`qI0ayij93<e
zl*KTP4#*SiDZ3$qqwpR~v0INBTgVrwOtX<-(2@l)cFi8w5AKbglGQ1LXCk<L^l&!u
z5Vem9uc>S4sMH>7ZxZ_w@}BFbQg4-}&`~@nx)AXKTSk&Kk=K9F$e1M?#&IFH(g~m7
zlOax3DhO>fH?AI%ZxVQ1ze=t}^?J=>CR#Jvt|cO%#iBNndZImQZRk|gYq)ghP>l{F
zBn3=K$chR{dLxqJp49#%6xgI^bxHoWg4&a(Jz-HQmbX@0>Sl0k4@>t6ONR=~i)m!f
z?u2EmVRvCfVn=_etp)(m+wc$oe3CX4Tqw1n#31T|qZ34T0`H+Pq^UGv9S9ES^zkT}
zVMKNm#jHm(8c*g=sC5c64--J5(B)y4;pmjYL249UC8Lj$<aK|xHA_VnbD~aJZ<=e{
zZ(4bnUmHO58#;Or%+~RzQ$i9{6(yut*9LRMmy~hMq=bJrVU)%Lizi%?VpS}WUNR?5
zc~2-qi=(^{b4!eA=rf$-lsyG+At_zOnBXx#37&*zdPX^5$Zi?kaEK$<g{c*fb!RY5
zEcfFuf?Qd6G3NoKfeI=uIr2ccN$PE`v@SRKVfP#B0wbEFtO93*)=~~cF*8+*rz4O!
zvl<;nD2so*D@`Ts?$X{Mt56T6jwhw%F;96+hDSf$V=`@27LJ_gSmp`GWZbd(n2fof
z9+Tl4o_kDI9yQ?_<%M*55s$fyZ5(}-$6e(y86H7Z$6v8d$5rLASUf^YkJmi@{xH_D
zy_5;@YkYM~mL89l_w@LzqCM8_6240sQ<<d4X>>qbhVsTb+%<j0FEvq)Xt;1v5sOEE
z_<($;I;@W1bmH@Ipm(e~E43mVj)Z&4R>fZOK8wHiw5C#b*%lxY+i+}`36UKkf45H)
zKg$dW?n?}bzQvVioKHG$Mu0>KKs0OyVb@}VIxTzgfivx^%L~(0$tp{zP#?u=BWe*8
zAscbjS6PU4^|&aqbe)hGB#Wk|q#bXbIBOG3Dc;leYEG=XSTFmlIT5{OU)z)*m&=4;
zbk~BlOi8!1p^hohL;9~8*|w!ne|fqwiNu(cC4(q6UbU6RWC4`*H$42>wqMMzXTyiZ
zbQZq8y&sIB=k(-q=XCd%z2n_aoI%HEa5v9X7_mzq^Iid8uMkO`qa|oqINzCH<0n1X
zP;KRVgNHZ8@a}$r8|tkbUxR(GICdT`2BYEi&g5=XWU&1Cq8R@U3qjAPe>e$54r|;G
zX83*Yzw*QUQ~o8N<=^uk`M>jj7qjVq2VxF}umlgkWoUk&>}T%$^ko0?n3&Vcc~dO%
zrW~e9=H<Le*~{CHCM#kcR$jl`+4=O5Sg}27(vcN`F--H9vm$ISE2tz=QIjRL4oe@8
zJ{}&u29|#7g`QBrlQ*aze+}5PqGn35lqZ&OWEkt=NTm!(o-%Yb8(bHo;&yR0ybBLA
z9EMN#gT=SuO_4v~ckFUV{jjL(i-%CJrY28|<N?ObAIIatVmO`L4CfD{K?@pY9XXy|
z9-Y2^OLFXi20#tOh`+rz0z9uxOj!B##pqbC7^AG3+_-hPxj6dpfA;7DaC3avhYBeo
z(&-ocD;<(th4feGhXPAh^s=;<@8?JP+x%UAoL}UF(Zl^9zn+5j^5S7W98D)YFBXIR
zHoqHS<K3(nEQ(qFXa3jV;bD-E2IH$6xE1qxK7Pz6IbZ~^HNDCI&gc1jcn5P9`Qjee
zSj>h4q>Mn)kfB)1fA1ZCc>i{1hh(_i?@1daD3#xk1bsg!-ZfcS%kS;H-uZI$kyz;i
zcNv{5-w-Q~_p-8|{}#azATR_yk5H*WrKH9tw`;k?<Iis|U%w%4-3pIL>~gj#?7Loe
z5tISIO{UiJPj3$1?(Sg}KI=2+Dhf09A~;qa%J_aT27CN4e=hC@`G8#gU&R8wcUdRv
z%ddyC>&Nl!sQ3=51}o2tzaIyq{C4<Hk>7#dNBJ1^KFz0-Br+sEK7SZJ&LNSR7vo`A
zU~xYyihS{H3X!vu*od{9<mH#sca+$i_2f0Yl&EAs%h}kIB<lJlii|dRk!v}t&nM@n
z7w5=}>4AL7e+&0=Bj|9lmlxD#rKcuSYq_m2Uk*Nf#+ZE8<F!QTfzcmsf+^dtpnzNg
zDq^z6VF!hBEpN5^Y5(}m88CVNxi^9H*(=K5FsHFr*K7jM&vRFbt@O_C^7r{U<foq@
zLmkX!({E&#R10=Fm_v9T&i{h+>ps64%<9xH$@_9>e;5NP`EYVOoD3H~D1Zk+jt39(
z#WWueCm_tp;A%9;C*VvlnVXLW^ZOVKkdtCo+(3*U4@ZMpQ_yR9(Dx@NM@Oe5=;ih<
zAP9Q;n@h}yJl}wrwdob}ZvHSEj+;rIT+4s$@BaSg^+#f;FRH_ChB8+C^H=M{CV=Bt
zaI{Z=e<`&dblc@pt>xKH_b)!~z5eC+eD}C_Zw+kX&j^K75`Wj1B<7~&27JwC@BSt=
z4I@(y&zgL!<;+eG5BE+#{BnM9c8o_BJqm+eCuHXM7s1+OUfP<KzMWbh8@gfPEM<f2
z_-6WT5^9eRO}5r@U+*uE_CM|a@_snJdYm6me<$x=?M_EGJsFAwca0}m4u5o)?yW^b
zO<Ic-jl%Tptyv?-^vw;&Z<JL|qa$guxR&Gia=CkXa*(okR@^<V*d$8)xeUW!7p5%g
zCW|&@5%<)~!I*N0f~Yw>$_JD|JvLce%Z0rA{loeGVPx%WI(|R(?_S|g(0jNu4&bTc
ze<w)@Nol}sIuxH~M1uJ~6*jG{;ZJtJqwG;QI>=w=Z}P)x2l+n#ke}qI`56SMi~J*I
zp!s#oFmLiAr{!QXc{>cL=I!)x2KnZFe*fd)zL<d6{)(yRWC#I0=AY&M1LJ$}U4K7L
z7crF0aGzQH1C!^&@3hw}<Nss+PyQ{Of8(?>0lk)|+dH|q_;lJ*#tNvRGK4FYAylFa
z+>Ip4&^%QJ@3SYT6J2Qb99>-Izk@zv;!tG?pq}0>!qwwwR4gj>g&pBdelvuyKOfF3
zZ5HDjObdz$-4fl7O7-3qv+-bZb2XZ$n*Os#*GqRB#SCSX0wF~M;Gvjfh7y1Qe{TxY
z=!w?q0H)lw)X)jpO%H&UKxn-n^ZxAM^Ue|op}zZI<k2hBf4I?xk!hT3(PNS%$aa#T
zmu`mDla2Z$f+#=;;4=nA%t57zd6)QT82R0Dh5t;~bbeH-v}xe$YH3N#{B3eQy%|pK
zKxV_+TR`&!|IX(RLC76t(pGS*fB8)Cvy|w{pa_~l7LoDDPaqr#4BIfEJn%q9LKZ{(
z5vM_fWEUk&BVaQlE75@ec$WPqCJ(7h!|EY+BvupLH|q43RN2(w0~j7EXOXG^16|7s
zqa%q9BDjkToMh!n@P9w4%{dc7CNAUala%LGnKIQ$)Yq<Yf*m|wE$9&&f58BD2lImd
zTC4M|sA0?Wf7H<5fgKEI^97nsrofoq4dMfAbhq-4!<)tZoc^yH8hbIloPgSH5QsgQ
z%_@poA9&gX$u>b;@CHS^qR7iGL3VwDTAyj!1nD+Gu1`YVcj<BK)6@RO(gA@5e)=ZJ
zwh3YhJ>J#im@YYfeR5i#D%#ZeHh(qV^^hYv<QTR-Iqh!^EqX);(7XX3%tk6e^Daev
zeTrHq_ichYfVL(f$kz1uu6-G?J~^$^_cl2lIP(UaQKprAmS{{qLq5#|@-{&oICGm2
zl$g8gnwL@QlhZodZj;l&vus#<DjwCPN3Ty$^IW(^j_ms9lpcI2(X)h_XMcF-=E-iG
znhuune5RsCC%V#IYM$Ymn@6&3YV=Ck7@nxA>k7>?JahZ-v`x@T&rH9ds0dmjpJ({y
z)*)k?9=|fUo5c0F4mtK2zPWWo*dj;qm4Pd@EY>CG8IHMqme;1IgJvfZf?S&VbTbY6
z4A<Oydex?^10-b{NJ<PX4i7Ce$jo)2ZY!t#KZYSoVV6Op9Tc}QnjN4I1u`@-GBcOf
zsU02#GBhzVGne_P9qxZyZ-OuwhVT0;zWc?%*8|XGSu$}ZWat!i$LxYlq9l_Vz>EF+
z6>6LjHV;?r`_jHVZNF9suz>@*2m%-vKp6;>z&SY50p#EbtprFxK@UQkYuB>@kulxT
z4oW1~*TkBBuM@EPDl3UK*?g3>tV|?UH!hOu3s|qS?rhY*8m51%k4-U9S=fn^#fH@L
zdP(t_r|VMX`o~g(McPmEqV$(>4lXBF7;golORT4KUM>spJTvqz`&9EZTY%>nW2$CZ
zs=Y+7kJr}C%^9fS<HMvEY*8b%8Y&rf2FW~a(h5AJrTeh%FD%;f1X|~OT7NX1K93J*
z+xJ$m|7a;?d$fP2?5%q~?RXgVUv67ygEIbVAr}AcFQq$TDNX?PE&vV!fBbMY=|=;!
zZM&sl8;EkhV5Rs)!~;ZWHxXqMaTG6-f>^%VXds7ZohQ0i-Nt(QeL64TjaCA%UNunD
zFk`Z<s<YoO3Fjt=*C!6bC}EU?rs2oZ#b4=8eJ)_`Dwrj*0y6|Wb}Y7f5ljx60COVM
ztJR$8$GXncf|!UMh#vC;>FH#2mqDW)6PKK<9bQpIH#tQ%H843bLPJADK|wS&G($Ny
zF*P(rI5IXyHa9~)AUs7kIYl-#FgY<oLqkMCK{PfrLpe4vH8e#yGB!pwH$y&M3NK7$
zZfA68AT=~Hm*H9+DSsu>OKePG7zW_~oz}F+m>GvsJ(s#wTgtdrb&8qR&~~aARV`&Y
zMOe5jBod7@nh<Q%LZgc&6}OOBR3fop#X{&pQW0BLB4VW*k|yJOzu7!_-!JDo`IG-Q
z#+ZD-m@tTeF;OFFl65k`*+G&e{iKhaAV)|K{NZ!DF^9E?1b<-cIA`Gja){(eFX`sT
zMKH;?`iZU;!_@xJdW>`d&F{39kRVW((0Z6yq!d=NMrR>17gkS))-ag|>%4AE7Ly=n
zQ4)bQ95b*UJvRu&XLQ3R^I_d-(t4B|nii4;upU+$M80QrwuqD=^5(bJa<Uka>1$d~
zk_xf}_VtfCJAcmqmB1ePrqv}5>{~ar`ox2MyKwrYq!JbRFItzAWpJYPT33(*WF@I0
zt4M}alNwS>R+BZPj;tl?NQ|r}8^}hoiEJiYNIls~8pt*{{#`vmBb?M}O>sE6N=;30
zt|m2Yhjah0rX6rzf7P@T&aXe35^#f4nv!tq$}}~@ZGXM2sRizp_nLOW9Sdnn!F~Bb
z({8xavzqq6oqeIH6<$T*1a0sd3zydpucbj#8eV&wrVe<y!cFgmH~36bC%jiNP5a;n
z3s<operwjiPmc<nQ$ofFT?wIkNa(#G^o<DFd&0nkaO$&g=BIGh6E0+gi}GIlOaFwy
zPr^`67$BVyo}`7byf7XTCd!2eRr*d#i3R_0N@h>{&3_#pY!R1@uN@x?Gc!3c3MC~)
GPeux7_dh)V
diff --git a/paper/appendix.tex b/paper/appendix.tex
index 68abf60c..9a44cb0e 100644
--- a/paper/appendix.tex
+++ b/paper/appendix.tex
@@ -100,7 +100,7 @@ The core part of our code base is integrated into a larger ecosystem of \texttt{
\subsection{Experimental Setup}\label{app:setup}
-Table~\ref{tab:params} provides an overview of all parameters related to our experiments. The \textit{GMSC} data were randomly undersampled for balancing purposes and all features were standardized. \textit{MNIST} data was also randomly undersampled for reasons outlined below. Pixel values were preprocessed to fall in the range of $[-1,1]$ and a small Gaussian noise component ($\sigma=0.03$) was added to training samples following common practice in the EBM literature. All of our models were trained through mini-batch training using the Adam optimiser (\citet{kingma2017adam}). Table~\ref{tab:perf} shows standard evaluation metrics measuring the predictive performance of our different models grouped by dataset. These measures were computed on test data.
+Table~\ref{tab:params} provides an overview of all parameters related to our experiments. The \textit{GMSC} data were randomly undersampled for balancing purposes and all features were standardized. \textit{MNIST} data was also randomly undersampled for reasons outlined below. Pixel values were preprocessed to fall in the range of $[-1,1]$ and a small Gaussian noise component ($\sigma=0.03$) was added to training samples following common practice in the EBM literature. All of our models were trained through mini-batch training using the Adam optimiser (\citet{kingma2014adam}). Table~\ref{tab:perf} shows standard evaluation metrics measuring the predictive performance of our different models grouped by dataset. These measures were computed on test data.
Table~\ref{tab:genparams} summarises our hyperparameter choices for the counterfactual generators where $\eta$ denotes the learning rate used for Stochastic Gradient Descent (SGD) and $\lambda_1$, $\lambda_2$, $\lambda_3$ represent the chosen penalty strengths (Equations~\ref{eq:general} and~\ref{eq:eccco}). Here $\lambda_1$ also refers to the chosen penalty for the distance from factual values that applies to both \textit{Wachter} and \textit{REVISE}, but not \textit{Schut} which is penalty-free. \textit{Schut} is also the only generator that uses JSMA instead of SGD for optimization.
diff --git a/paper/bib.bib b/paper/bib.bib
index daed3cba..e31a8866 100644
--- a/paper/bib.bib
+++ b/paper/bib.bib
@@ -1,17 +1,23 @@
-@TechReport{kingma2017adam,
- author = {Kingma, Diederik P. and Ba, Jimmy},
- date = {2017-01},
+@TechReport{xu2022conformal,
+ author = {Xu, Chen and Xie, Yao},
+ date = {2022-06},
institution = {arXiv},
- title = {Adam: {A} {Method} for {Stochastic} {Optimization}},
- doi = {10.48550/arXiv.1412.6980},
- note = {arXiv:1412.6980 [cs] type: article},
- url = {http://arxiv.org/abs/1412.6980},
- urldate = {2023-05-17},
- abstract = {We introduce Adam, an algorithm for first-order gradient-based optimization of stochastic objective functions, based on adaptive estimates of lower-order moments. The method is straightforward to implement, is computationally efficient, has little memory requirements, is invariant to diagonal rescaling of the gradients, and is well suited for problems that are large in terms of data and/or parameters. The method is also appropriate for non-stationary objectives and problems with very noisy and/or sparse gradients. The hyper-parameters have intuitive interpretations and typically require little tuning. Some connections to related algorithms, on which Adam was inspired, are discussed. We also analyze the theoretical convergence properties of the algorithm and provide a regret bound on the convergence rate that is comparable to the best known results under the online convex optimization framework. Empirical results demonstrate that Adam works well in practice and compares favorably to other stochastic optimization methods. Finally, we discuss AdaMax, a variant of Adam based on the infinity norm.},
- annotation = {Comment: Published as a conference paper at the 3rd International Conference for Learning Representations, San Diego, 2015},
- file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/1412.6980.pdf:application/pdf},
- keywords = {Computer Science - Machine Learning},
- shorttitle = {Adam},
+ title = {Conformal prediction set for time-series},
+ doi = {10.48550/arXiv.2206.07851},
+ note = {arXiv:2206.07851 [cs, stat] type: article},
+ url = {http://arxiv.org/abs/2206.07851},
+ urldate = {2023-07-22},
+ abstract = {When building either prediction intervals for regression (with real-valued response) or prediction sets for classification (with categorical responses), uncertainty quantification is essential to studying complex machine learning methods. In this paper, we develop Ensemble Regularized Adaptive Prediction Set (ERAPS) to construct prediction sets for time-series (with categorical responses), based on the prior work of [Xu and Xie, 2021]. In particular, we allow unknown dependencies to exist within features and responses that arrive in sequence. Method-wise, ERAPS is a distribution-free and ensemble-based framework that is applicable for arbitrary classifiers. Theoretically, we bound the coverage gap without assuming data exchangeability and show asymptotic set convergence. Empirically, we demonstrate valid marginal and conditional coverage by ERAPS, which also tends to yield smaller prediction sets than competing methods.},
+ annotation = {Comment: Strongly accepted by the Workshop on Distribution-Free Uncertainty Quantification at ICML 2022},
+ file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/2206.07851.pdf:application/pdf},
+ keywords = {Statistics - Machine Learning, Computer Science - Machine Learning, Statistics - Methodology},
+}
+
+@Article{kingma2014adam,
+ author = {Kingma, Diederik P and Ba, Jimmy},
+ title = {Adam: A method for stochastic optimization},
+ journal = {arXiv preprint arXiv:1412.6980},
+ year = {2014},
}
@TechReport{xiao2017fashion,
@@ -41,17 +47,18 @@
}
@InProceedings{altmeyer2023endogenous,
- author = {Altmeyer, Patrick and Angela, Giovan and Buszydlik, Aleksander and Dobiczek, Karol and van Deursen, Arie and Liem, Cynthia},
- booktitle = {First {IEEE} {Conference} on {Secure} and {Trustworthy} {Machine} {Learning}},
- title = {Endogenous {Macrodynamics} in {Algorithmic} {Recourse}},
- file = {:altmeyerendogenous - Endogenous Macrodynamics in Algorithmic Recourse.pdf:PDF},
- year = {2023},
+ author = {Altmeyer, Patrick and Angela, Giovan and Buszydlik, Aleksander and Dobiczek, Karol and van Deursen, Arie and Liem, Cynthia CS},
+ booktitle = {2023 IEEE Conference on Secure and Trustworthy Machine Learning (SaTML)},
+ title = {Endogenous Macrodynamics in Algorithmic Recourse},
+ organization = {IEEE},
+ pages = {418--431},
+ year = {2023},
}
%% This BibTeX bibliography file was created using BibDesk.
%% https://bibdesk.sourceforge.io/
-%% Created for Anonymous Author at 2022-12-13 12:58:22 +0100
+%% Created for Patrick Altmeyer at 2022-12-13 12:58:22 +0100
%% Saved with string encoding Unicode (UTF-8)
@@ -61,6 +68,7 @@
@Article{abadie2002instrumental,
author = {Abadie, Alberto and Angrist, Joshua and Imbens, Guido},
title = {Instrumental Variables Estimates of the Effect of Subsidized Training on the Quantiles of Trainee Earnings},
+ doi = {10.2139/ssrn.195733},
number = {1},
pages = {91--117},
volume = {70},
@@ -87,6 +95,7 @@
author = {Ackerman, Samuel and Dube, Parijat and Farchi, Eitan and Raz, Orna and Zalmanovici, Marcel},
booktitle = {2021 {{IEEE}}/{{ACM Third International Workshop}} on {{Deep Learning}} for {{Testing}} and {{Testing}} for {{Deep Learning}} ({{DeepTest}})},
title = {Machine {{Learning Model Drift Detection Via Weak Data Slices}}},
+ doi = {10.1109/deeptest52559.2021.00007},
pages = {1--8},
publisher = {{IEEE}},
date-added = {2022-12-13 12:58:01 +0100},
@@ -97,6 +106,7 @@
@Article{allen2017referencedependent,
author = {Allen, Eric J and Dechow, Patricia M and Pope, Devin G and Wu, George},
title = {Reference-Dependent Preferences: {{Evidence}} from Marathon Runners},
+ doi = {10.3386/w20343},
number = {6},
pages = {1657--1672},
volume = {63},
@@ -153,16 +163,11 @@
year = {2022},
}
-@Unpublished{angelopoulos2021gentle,
- author = {Angelopoulos, Anastasios N. and Bates, Stephen},
- title = {A Gentle Introduction to Conformal Prediction and Distribution-Free Uncertainty Quantification},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2107.07511},
- eprinttype = {arxiv},
- file = {:/Users/FA31DU/Zotero/storage/RKSUMYZG/Angelopoulos and Bates - 2021 - A gentle introduction to conformal prediction and .pdf:;:/Users/FA31DU/Zotero/storage/PRUEKRR3/2107.html:},
- year = {2021},
+@Article{angelopoulos2021gentle,
+ author = {Angelopoulos, Anastasios N and Bates, Stephen},
+ title = {A gentle introduction to conformal prediction and distribution-free uncertainty quantification},
+ journal = {arXiv preprint arXiv:2107.07511},
+ year = {2021},
}
@Misc{angelopoulos2022uncertainty,
@@ -187,6 +192,7 @@
@Article{angelucci2009indirect,
author = {Angelucci, Manuela and De Giorgi, Giacomo},
title = {Indirect Effects of an Aid Program: How Do Cash Transfers Affect Ineligibles' Consumption?},
+ doi = {10.1257/aer.99.1.486},
number = {1},
pages = {486--508},
volume = {99},
@@ -230,6 +236,7 @@
@Article{ariely2003coherent,
author = {Ariely, Dan and Loewenstein, George and Prelec, Drazen},
title = {``{{Coherent}} Arbitrariness'': {{Stable}} Demand Curves without Stable Preferences},
+ doi = {10.1017/cbo9780511618031.014},
number = {1},
pages = {73--106},
volume = {118},
@@ -242,6 +249,7 @@
@Article{ariely2006tom,
author = {Ariely, Dan and Loewenstein, George and Prelec, Drazen},
title = {Tom {{Sawyer}} and the Construction of Value},
+ doi = {10.1017/cbo9780511618031.015},
number = {1},
pages = {1--10},
volume = {60},
@@ -254,6 +262,7 @@
@Article{arrieta2020explainable,
author = {Arrieta, Alejandro Barredo and Diaz-Rodriguez, Natalia and Del Ser, Javier and Bennetot, Adrien and Tabik, Siham and Barbado, Alberto and Garcia, Salvador and Gil-Lopez, Sergio and Molina, Daniel and Benjamins, Richard and others},
title = {Explainable {{Artificial Intelligence}} ({{XAI}}): {{Concepts}}, Taxonomies, Opportunities and Challenges toward Responsible {{AI}}},
+ doi = {10.1016/j.inffus.2019.12.012},
pages = {82--115},
volume = {58},
date-added = {2022-12-13 12:58:01 +0100},
@@ -297,6 +306,7 @@
@Article{bechara1997deciding,
author = {Bechara, Antoine and Damasio, Hanna and Tranel, Daniel and Damasio, Antonio R},
title = {Deciding Advantageously before Knowing the Advantageous Strategy},
+ doi = {10.7551/mitpress/3077.003.0044},
number = {5304},
pages = {1293--1295},
volume = {275},
@@ -310,6 +320,7 @@
@Book{berlinet2011reproducing,
author = {Berlinet, Alain and Thomas-Agnan, Christine},
title = {Reproducing Kernel {{Hilbert}} Spaces in Probability and Statistics},
+ doi = {10.1007/978-1-4419-9096-9},
publisher = {{Springer Science \& Business Media}},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
@@ -319,6 +330,7 @@
@Misc{bernanke1990federal,
author = {Bernanke, Ben S},
title = {The Federal Funds Rate and the Channels of Monetary Transnission},
+ doi = {10.3386/w3487},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
publisher = {{National Bureau of Economic Research Cambridge, Mass., USA}},
@@ -390,6 +402,7 @@
@Article{borch2022machine,
author = {Borch, Christian},
title = {Machine Learning, Knowledge Risk, and Principal-Agent Problems in Automated Trading},
+ doi = {10.1016/j.techsoc.2021.101852},
pages = {101852},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
@@ -397,20 +410,18 @@
year = {2022},
}
-@Unpublished{borisov2021deep,
- author = {Borisov, Vadim and Leemann, Tobias and Se{\ss}ler, Kathrin and Haug, Johannes and Pawelczyk, Martin and Kasneci, Gjergji},
- title = {Deep Neural Networks and Tabular Data: {{A}} Survey},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2110.01889},
- eprinttype = {arxiv},
- year = {2021},
+@Article{borisov2022deep,
+ author = {Borisov, Vadim and Leemann, Tobias and Se{\ss}ler, Kathrin and Haug, Johannes and Pawelczyk, Martin and Kasneci, Gjergji},
+ title = {Deep neural networks and tabular data: A survey},
+ journal = {IEEE Transactions on Neural Networks and Learning Systems},
+ publisher = {IEEE},
+ year = {2022},
}
@Article{bramoulle2009identification,
author = {Bramoull{\'e}, Yann and Djebbari, Habiba and Fortin, Bernard},
title = {Identification of Peer Effects through Social Networks},
+ doi = {10.2139/ssrn.965818},
number = {1},
pages = {41--55},
volume = {150},
@@ -423,6 +434,7 @@
@Article{bramoulle2020peer,
author = {Bramoull{\'e}, Yann and Djebbari, Habiba and Fortin, Bernard},
title = {Peer Effects in Networks: {{A}} Survey},
+ doi = {10.2139/ssrn.3534495},
pages = {603--629},
volume = {12},
date-added = {2022-12-13 12:58:01 +0100},
@@ -445,6 +457,7 @@
@Book{brock1991nonlinear,
author = {Brock, William Allen and Brock, William A and Hsieh, David Arthur and LeBaron, Blake Dean and Brock, William E},
title = {Nonlinear Dynamics, Chaos, and Instability: Statistical Theory and Economic Evidence},
+ doi = {10.2307/2234554},
publisher = {{MIT press}},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
@@ -476,6 +489,7 @@
@Report{card1993minimum,
author = {Card, David and Krueger, Alan B},
title = {Minimum Wages and Employment: {{A}} Case Study of the Fast Food Industry in {{New Jersey}} and {{Pennsylvania}}},
+ doi = {10.3386/w4509},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
school = {{National Bureau of Economic Research}},
@@ -486,6 +500,7 @@
author = {Carlini, Nicholas and Wagner, David},
booktitle = {2017 Ieee Symposium on Security and Privacy (Sp)},
title = {Towards Evaluating the Robustness of Neural Networks},
+ doi = {10.1109/sp.2017.49},
pages = {39--57},
publisher = {{IEEE}},
date-added = {2022-12-13 12:58:01 +0100},
@@ -506,6 +521,7 @@
@Article{carrell2009does,
author = {Carrell, Scott E and Fullerton, Richard L and West, James E},
title = {Does Your Cohort Matter? {{Measuring}} Peer Effects in College Achievement},
+ doi = {10.3386/w14032},
number = {3},
pages = {439--464},
volume = {27},
@@ -518,6 +534,7 @@
@Article{carrell2013natural,
author = {Carrell, Scott E and Sacerdote, Bruce I and West, James E},
title = {From Natural Variation to Optimal Policy? {{The}} Importance of Endogenous Peer Group Formation},
+ doi = {10.3982/ecta10168},
number = {3},
pages = {855--882},
volume = {81},
@@ -539,6 +556,7 @@
@Article{cascarino2022explainable,
author = {Cascarino, Giuseppe and Moscatelli, Mirko and Parlapiano, Fabio},
title = {Explainable {{Artificial Intelligence}}: Interpreting Default Forecasting Models Based on {{Machine Learning}}},
+ doi = {10.2139/ssrn.4090707},
number = {674},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
@@ -572,6 +590,7 @@
@Article{chetty2011adjustment,
author = {Chetty, Raj and Friedman, John N and Olsen, Tore and Pistaferri, Luigi},
title = {Adjustment Costs, Firm Responses, and Micro vs. Macro Labor Supply Elasticities: {{Evidence}} from {{Danish}} Tax Records},
+ doi = {10.3386/w15617},
number = {2},
pages = {749--804},
volume = {126},
@@ -596,6 +615,7 @@
@Article{crawford2019variable,
author = {Crawford, Lorin and Flaxman, Seth R and Runcie, Daniel E and West, Mike},
title = {Variable Prioritization in Nonlinear Black Box Methods: {{A}} Genetic Association Case Study},
+ doi = {10.1214/18-aoas1222},
number = {2},
pages = {958},
volume = {13},
@@ -608,6 +628,7 @@
@InProceedings{dai2022counterfactual,
author = {Dai, Xinyue and Keane, Mark T and Shalloo, Laurence and Ruelle, Elodie and Byrne, Ruth MJ},
title = {Counterfactual Explanations for Prediction and Diagnosis in Xai},
+ doi = {10.1145/3514094.3534144},
eventtitle = {Proceedings of the 2022 {{AAAI}}/{{ACM Conference}} on {{AI}}, {{Ethics}}, and {{Society}}},
pages = {215--226},
date-added = {2022-12-13 12:58:01 +0100},
@@ -638,6 +659,7 @@
@Article{dehejia1999causal,
author = {Dehejia, Rajeev H and Wahba, Sadek},
title = {Causal Effects in Nonexperimental Studies: {{Reevaluating}} the Evaluation of Training Programs},
+ doi = {10.1080/01621459.1999.10473858},
number = {448},
pages = {1053--1062},
volume = {94},
@@ -650,6 +672,7 @@
@Article{dell2010persistent,
author = {Dell, Melissa},
title = {The Persistent Effects of {{Peru}}'s Mining Mita},
+ doi = {10.2139/ssrn.1596425},
number = {6},
pages = {1863--1903},
volume = {78},
@@ -663,6 +686,7 @@
@Article{denhengst2020reinforcement,
author = {den Hengst, Floris and Grua, Eoin Martino and el Hassouni, Ali and Hoogendoorn, Mark},
title = {Reinforcement Learning for Personalization: {{A}} Systematic Literature Review},
+ doi = {10.3233/ds-200028},
issue = {Preprint},
pages = {1--41},
date-added = {2022-12-13 12:58:01 +0100},
@@ -675,6 +699,7 @@
@Article{deoliveira2021framework,
author = {de Oliveira, Raphael Mazzine Barbosa and Martens, David},
title = {A Framework and Benchmarking Study for Counterfactual Generating Methods on Tabular Data},
+ doi = {10.3390/app11167274},
number = {16},
pages = {7274},
volume = {11},
@@ -717,6 +742,7 @@
@Article{epstein1979stability,
author = {Epstein, Seymour},
title = {The Stability of Behavior: {{I}}. {{On}} Predicting Most of the People Much of the Time.},
+ doi = {10.1037/0022-3514.37.7.1097},
number = {7},
pages = {1097},
volume = {37},
@@ -741,6 +767,7 @@
@Article{falk2006clean,
author = {Falk, Armin and Ichino, Andrea},
title = {Clean Evidence on Peer Effects},
+ doi = {10.1086/497818},
number = {1},
pages = {39--57},
volume = {24},
@@ -776,6 +803,7 @@
@Article{fehr2000cooperation,
author = {Fehr, Ernst and Gachter, Simon},
title = {Cooperation and Punishment in Public Goods Experiments},
+ doi = {10.2139/ssrn.203194},
number = {4},
pages = {980--994},
volume = {90},
@@ -832,6 +860,7 @@
@Article{galizzi2019external,
author = {Galizzi, Matteo M and Navarro-Martinez, Daniel},
title = {On the External Validity of Social Preference Games: A Systematic Lab-Field Study},
+ doi = {10.1287/mnsc.2017.2908},
number = {3},
pages = {976--1002},
volume = {65},
@@ -876,6 +905,7 @@
@Article{gilbert1998immune,
author = {Gilbert, Daniel T and Pinel, Elizabeth C and Wilson, Timothy D and Blumberg, Stephen J and Wheatley, Thalia P},
title = {Immune Neglect: A Source of Durability Bias in Affective Forecasting.},
+ doi = {10.1037/0022-3514.75.3.617},
number = {3},
pages = {617},
volume = {75},
@@ -888,6 +918,7 @@
@Article{gneezy2006uncertainty,
author = {Gneezy, Uri and List, John A and Wu, George},
title = {The Uncertainty Effect: {{When}} a Risky Prospect Is Valued Less than Its Worst Possible Outcome},
+ doi = {10.1093/qje/121.4.1283},
number = {4},
pages = {1283--1309},
volume = {121},
@@ -901,6 +932,7 @@
author = {Goan, Ethan and Fookes, Clinton},
booktitle = {Case {{Studies}} in {{Applied Bayesian Data Science}}},
title = {Bayesian {{Neural Networks}}: {{An Introduction}} and {{Survey}}},
+ doi = {10.1007/978-3-030-42553-1_3},
pages = {45--87},
publisher = {{Springer}},
date-added = {2022-12-13 12:58:01 +0100},
@@ -943,6 +975,7 @@
@Article{goodfriend2005incredible,
author = {Goodfriend, Marvin and King, Robert G},
title = {The Incredible {{Volcker}} Disinflation},
+ doi = {10.3386/w11562},
number = {5},
pages = {981--1015},
volume = {52},
@@ -955,6 +988,7 @@
@Article{graham2017econometric,
author = {Graham, Bryan S},
title = {An Econometric Model of Network Formation with Degree Heterogeneity},
+ doi = {10.1920/wp.cem.2017.0817},
number = {4},
pages = {1033--1063},
volume = {85},
@@ -977,6 +1011,7 @@
@Article{grether1979economic,
author = {Grether, David M and Plott, Charles R},
title = {Economic Theory of Choice and the Preference Reversal Phenomenon},
+ doi = {10.1017/cbo9780511618031.006},
number = {4},
pages = {623--638},
volume = {69},
@@ -1000,9 +1035,11 @@
@Unpublished{griffith2020name,
author = {Griffith, Alan},
+ date = {2020-08-01},
title = {Name {{Your Friends}}, but {{Only Five}}? {{The Importance}} of {{Censoring}} in {{Peer Effects Estimates}} Using {{Social Network Data}}},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
+ doi = {10.1086/717935},
year = {2020},
}
@@ -1029,6 +1066,7 @@
author = {Gupta, Neha and Granmo, Ole-Christoffer and Agrawala, Ashok},
booktitle = {2011 10th {{International Conference}} on {{Machine Learning}} and {{Applications}} and {{Workshops}}},
title = {Thompson Sampling for Dynamic Multi-Armed Bandits},
+ doi = {10.1109/icmla.2011.144},
pages = {484--489},
publisher = {{IEEE}},
volume = {1},
@@ -1058,6 +1096,7 @@
@Article{hamzacebi2008improving,
author = {Hamza{\c c}ebi, Co{\c s}kun},
title = {Improving Artificial Neural Networks' Performance in Seasonal Time Series Forecasting},
+ doi = {10.1016/j.ins.2008.07.024},
number = {23},
pages = {4550--4559},
volume = {178},
@@ -1071,6 +1110,7 @@
author = {Hanneke, Steve},
booktitle = {Proceedings of the 24th International Conference on {{Machine}} Learning},
title = {A Bound on the Label Complexity of Agnostic Active Learning},
+ doi = {10.1145/1273496.1273541},
pages = {353--360},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
@@ -1080,6 +1120,7 @@
@Article{hansen2020virtue,
author = {Hansen, Kristian Bondo},
title = {The Virtue of Simplicity: {{On}} Machine Learning Models in Algorithmic Trading},
+ doi = {10.1177/2053951720926558},
number = {1},
pages = {2053951720926558},
volume = {7},
@@ -1112,6 +1153,7 @@
@Article{hershfield2011increasing,
author = {Hershfield, Hal E and Goldstein, Daniel G and Sharpe, William F and Fox, Jesse and Yeykelis, Leo and Carstensen, Laura L and Bailenson, Jeremy N},
title = {Increasing Saving Behavior through Age-Progressed Renderings of the Future Self},
+ doi = {10.1509/jmkr.48.spl.s23},
issue = {SPL},
pages = {S23--S37},
volume = {48},
@@ -1125,6 +1167,7 @@
author = {Ho, Tin Kam},
booktitle = {Proceedings of 3rd International Conference on Document Analysis and Recognition},
title = {Random Decision Forests},
+ doi = {10.1109/ICDAR.1995.598994},
pages = {278--282},
publisher = {{IEEE}},
volume = {1},
@@ -1148,9 +1191,11 @@
@Unpublished{hoff2021bayesoptimal,
author = {Hoff, Peter},
title = {Bayes-Optimal Prediction with Frequentist Coverage Control},
+ url = {10.3150/22-bej1484},
archiveprefix = {arXiv},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
+ doi = {10.3150/22-bej1484},
eprint = {2105.14045},
eprinttype = {arxiv},
file = {:/Users/FA31DU/Zotero/storage/IQK27WVA/Hoff - 2021 - Bayes-optimal prediction with frequentist coverage.pdf:;:/Users/FA31DU/Zotero/storage/K8EAZA25/2105.html:},
@@ -1191,8 +1236,10 @@
@Article{hsee1996evaluability,
author = {Hsee, Christopher K},
title = {The Evaluability Hypothesis: {{An}} Explanation for Preference Reversals between Joint and Separate Evaluations of Alternatives},
+ doi = {10.1006/obhd.1996.0077},
number = {3},
pages = {247--257},
+ url = {10.1006/obhd.1996.0077},
volume = {67},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
@@ -1203,8 +1250,10 @@
@Article{hsee2004music,
author = {Hsee, Christopher K and Rottenstreich, Yuval},
title = {Music, Pandas, and Muggers: On the Affective Psychology of Value.},
+ doi = {10.1017/cbo9780511618031.033},
number = {1},
pages = {23},
+ url = {10.1017/cbo9780511618031.033},
volume = {133},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
@@ -1215,6 +1264,7 @@
@Article{hsieh2016social,
author = {Hsieh, Chih-Sheng and Lee, Lung Fei},
title = {A Social Interactions Model with Endogenous Friendship Formation and Selectivity},
+ doi = {10.1002/jae.2426},
number = {2},
pages = {301--319},
volume = {31},
@@ -1249,6 +1299,7 @@
@Article{innes2018flux,
author = {Innes, Mike},
title = {Flux: {{Elegant}} Machine Learning with {{Julia}}},
+ doi = {10.21105/joss.00602},
number = {25},
pages = {602},
volume = {3},
@@ -1283,6 +1334,7 @@
@Article{jackson2007meeting,
author = {Jackson, Matthew O and Rogers, Brian W},
title = {Meeting Strangers and Friends of Friends: {{How}} Random Are Social Networks?},
+ doi = {10.1257/aer.97.3.890},
number = {3},
pages = {890--915},
volume = {97},
@@ -1306,6 +1358,7 @@
@Article{johansson2005failure,
author = {Johansson, Petter and Hall, Lars and Sikstr{\"o}m, Sverker and Olsson, Andreas},
title = {Failure to Detect Mismatches between Intention and Outcome in a Simple Decision Task},
+ doi = {10.1126/science.1111709},
number = {5745},
pages = {116--119},
volume = {310},
@@ -1319,6 +1372,7 @@
@Article{johnsson2021estimation,
author = {Johnsson, Ida and Moon, Hyungsik Roger},
title = {Estimation of Peer Effects in Endogenous Social Networks: {{Control}} Function Approach},
+ doi = {10.2139/ssrn.3043404},
number = {2},
pages = {328--345},
volume = {103},
@@ -1331,6 +1385,7 @@
@Article{jolliffe2003modified,
author = {Jolliffe, Ian T and Trendafilov, Nickolay T and Uddin, Mudassir},
title = {A Modified Principal Component Technique Based on the {{LASSO}}},
+ doi = {10.1198/1061860032148},
number = {3},
pages = {531--547},
volume = {12},
@@ -1348,15 +1403,11 @@
year = {2021},
}
-@Unpublished{joshi2019realistic,
- author = {Joshi, Shalmali and Koyejo, Oluwasanmi and Vijitbenjaronk, Warut and Kim, Been and Ghosh, Joydeep},
- title = {Towards Realistic Individual Recourse and Actionable Explanations in Black-Box Decision Making Systems},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {1907.09615},
- eprinttype = {arxiv},
- year = {2019},
+@Article{joshi2019realistic,
+ author = {Joshi, Shalmali and Koyejo, Oluwasanmi and Vijitbenjaronk, Warut and Kim, Been and Ghosh, Joydeep},
+ title = {Towards realistic individual recourse and actionable explanations in black-box decision making systems},
+ journal = {arXiv preprint arXiv:1907.09615},
+ year = {2019},
}
@Unpublished{jospin2020handson,
@@ -1365,6 +1416,7 @@
archiveprefix = {arXiv},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
+ doi = {10.1109/mci.2022.3155327},
eprint = {2007.06823},
eprinttype = {arxiv},
year = {2020},
@@ -1402,6 +1454,7 @@
@Article{kahneman1990experimental,
author = {Kahneman, Daniel and Knetsch, Jack L and Thaler, Richard H},
title = {Experimental Tests of the Endowment Effect and the {{Coase}} Theorem},
+ doi = {10.1017/cbo9781139175197.009},
number = {6},
pages = {1325--1348},
volume = {98},
@@ -1414,6 +1467,7 @@
@Article{kahneman1992reference,
author = {Kahneman, Daniel},
title = {Reference Points, Anchors, Norms, and Mixed Feelings},
+ doi = {10.1016/0749-5978(92)90015-y},
number = {2},
pages = {296--312},
volume = {51},
@@ -1459,6 +1513,7 @@
author = {Kaur, Harmanpreet and Nori, Harsha and Jenkins, Samuel and Caruana, Rich and Wallach, Hanna and Wortman Vaughan, Jennifer},
booktitle = {Proceedings of the 2020 {{CHI}} Conference on Human Factors in Computing Systems},
title = {Interpreting Interpretability: Understanding Data Scientists' Use of Interpretability Tools for Machine Learning},
+ doi = {10.1145/3313831.3376219},
pages = {1--14},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
@@ -1488,6 +1543,7 @@
@Article{kihoro2004seasonal,
author = {Kihoro, J and Otieno, RO and Wafula, C},
title = {Seasonal Time Series Forecasting: {{A}} Comparative Study of {{ARIMA}} and {{ANN}} Models},
+ doi = {10.4314/ajst.v5i2.15330},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
year = {2004},
@@ -1530,6 +1586,7 @@
archiveprefix = {arXiv},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
+ doi = {10.1007/978-3-030-93842-0_6},
eprint = {2111.02244},
eprinttype = {arxiv},
year = {2021},
@@ -1610,17 +1667,16 @@
}
@Article{lecun1998mnist,
- author = {LeCun, Yann},
- title = {The {{MNIST}} Database of Handwritten Digits},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- shortjournal = {http://yann. lecun. com/exdb/mnist/},
- year = {1998},
+ author = {LeCun, Yann},
+ title = {The MNIST database of handwritten digits},
+ journal = {http://yann. lecun. com/exdb/mnist/},
+ year = {1998},
}
@Article{lee2003best,
author = {Lee, Lung-fei},
title = {Best Spatial Two-Stage Least Squares Estimators for a Spatial Autoregressive Model with Autoregressive Disturbances},
+ doi = {10.1081/etc-120025891},
number = {4},
pages = {307--335},
volume = {22},
@@ -1633,6 +1689,7 @@
@Article{lerner2013financial,
author = {Lerner, Jennifer S and Li, Ye and Weber, Elke U},
title = {The Financial Costs of Sadness},
+ doi = {10.1177/0956797612450302},
number = {1},
pages = {72--79},
volume = {24},
@@ -1645,6 +1702,7 @@
@Article{list2004neoclassical,
author = {List, John A},
title = {Neoclassical Theory versus Prospect Theory: {{Evidence}} from the Marketplace},
+ doi = {10.3386/w9736},
number = {2},
pages = {615--625},
volume = {72},
@@ -1685,6 +1743,7 @@
@Article{madrian2001power,
author = {Madrian, Brigitte C and Shea, Dennis F},
title = {The Power of Suggestion: {{Inertia}} in 401 (k) Participation and Savings Behavior},
+ doi = {10.3386/w7682},
number = {4},
pages = {1149--1187},
volume = {116},
@@ -1703,15 +1762,22 @@
year = {2008},
}
-@misc{manokhin2022awesome,
- author = {Manokhin, Valery},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- title = {Awesome Conformal Prediction}}
+@Software{manokhin2022awesome,
+ author = {Manokhin, Valery},
+ title = {Awesome Conformal Prediction},
+ doi = {10.5281/zenodo.6467205},
+ note = {{"If you use Awesome Conformal Prediction, please cite it as below."}},
+ url = {https://doi.org/10.5281/zenodo.6467205},
+ version = {v1.0.0},
+ month = apr,
+ publisher = {Zenodo},
+ year = {2022},
+}
@Article{manski1993identification,
author = {Manski, Charles F},
title = {Identification of Endogenous Social Effects: {{The}} Reflection Problem},
+ doi = {10.2307/2298123},
number = {3},
pages = {531--542},
volume = {60},
@@ -1724,6 +1790,7 @@
@Article{markle2018goals,
author = {Markle, Alex and Wu, George and White, Rebecca and Sackett, Aaron},
title = {Goals as Reference Points in Marathon Running: {{A}} Novel Test of Reference Dependence},
+ doi = {10.2139/ssrn.2523510},
number = {1},
pages = {19--50},
volume = {56},
@@ -1745,6 +1812,7 @@
@Article{mccracken2016fredmd,
author = {McCracken, Michael W and Ng, Serena},
title = {{{FRED-MD}}: {{A}} Monthly Database for Macroeconomic Research},
+ doi = {10.20955/wp.2015.012},
number = {4},
pages = {574--589},
volume = {34},
@@ -1769,6 +1837,7 @@
@Article{migut2015visualizing,
author = {Migut, MA and Worring, Marcel and Veenman, Cor J},
title = {Visualizing Multi-Dimensional Decision Boundaries in {{2D}}},
+ doi = {10.1007/s10618-013-0342-x},
number = {1},
pages = {273--295},
volume = {29},
@@ -1781,6 +1850,7 @@
@Article{miller2019explanation,
author = {Miller, Tim},
title = {Explanation in Artificial Intelligence: {{Insights}} from the Social Sciences},
+ doi = {10.1016/j.artint.2018.07.007},
pages = {1--38},
volume = {267},
date-added = {2022-12-13 12:58:01 +0100},
@@ -1812,6 +1882,7 @@
@Article{mischel1988nature,
author = {Mischel, Walter and Shoda, Yuichi and Peake, Philip K},
title = {The Nature of Adolescent Competencies Predicted by Preschool Delay of Gratification.},
+ doi = {10.1037/0022-3514.54.4.687},
number = {4},
pages = {687},
volume = {54},
@@ -1825,19 +1896,20 @@
author = {Mittelstadt, Brent and Russell, Chris and Wachter, Sandra},
booktitle = {Proceedings of the Conference on Fairness, Accountability, and Transparency},
title = {Explaining Explanations in {{AI}}},
+ doi = {10.1145/3287560.3287574},
pages = {279--288},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
year = {2019},
}
-@Book{molnar2020interpretable,
- author = {Molnar, Christoph},
- title = {Interpretable Machine Learning},
- publisher = {{Lulu. com}},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- year = {2020},
+@Book{molnar2022interpretable,
+ author = {Christoph Molnar},
+ title = {Interpretable Machine Learning},
+ edition = {2},
+ subtitle = {A Guide for Making Black Box Models Explainable},
+ url = {https://christophm.github.io/interpretable-ml-book},
+ year = {2022},
}
@Book{morgan2015counterfactuals,
@@ -1865,6 +1937,7 @@
author = {Mothilal, Ramaravind K and Sharma, Amit and Tan, Chenhao},
booktitle = {Proceedings of the 2020 {{Conference}} on {{Fairness}}, {{Accountability}}, and {{Transparency}}},
title = {Explaining Machine Learning Classifiers through Diverse Counterfactual Explanations},
+ doi = {10.1145/3351095.3372850},
pages = {607--617},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
@@ -1922,6 +1995,7 @@
author = {Nelson, Kevin and Corbin, George and Anania, Mark and Kovacs, Matthew and Tobias, Jeremy and Blowers, Misty},
booktitle = {2015 {{IEEE Symposium}} on {{Computational Intelligence}} for {{Security}} and {{Defense Applications}} ({{CISDA}})},
title = {Evaluating Model Drift in Machine Learning Algorithms},
+ doi = {10.1109/cisda.2015.7208643},
pages = {1--8},
publisher = {{IEEE}},
date-added = {2022-12-13 12:58:01 +0100},
@@ -1971,6 +2045,7 @@
@Article{pace1997sparse,
author = {Pace, R Kelley and Barry, Ronald},
title = {Sparse Spatial Autoregressions},
+ doi = {10.1016/s0167-7152(96)00140-x},
number = {3},
pages = {291--297},
volume = {33},
@@ -2014,6 +2089,7 @@
@Article{pearl2019seven,
author = {Pearl, Judea},
title = {The Seven Tools of Causal Inference, with Reflections on Machine Learning},
+ doi = {10.1145/3241036},
number = {3},
pages = {54--60},
volume = {62},
@@ -2037,6 +2113,7 @@
@Book{perry2010economic,
author = {Perry, George L and Tobin, James},
title = {Economic {{Events}}, {{Ideas}}, and {{Policies}}: The 1960s and After},
+ doi = {10.5860/choice.38-4002},
publisher = {{Brookings Institution Press}},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
@@ -2089,6 +2166,7 @@
@Article{qu2015estimating,
author = {Qu, Xi and Lee, Lung-fei},
title = {Estimating a Spatial Autoregressive Model with an Endogenous Spatial Weight Matrix},
+ doi = {10.1016/j.jeconom.2014.08.008},
number = {2},
pages = {209--232},
volume = {184},
@@ -2154,6 +2232,7 @@
@Article{romer1989does,
author = {Romer, Christina D and Romer, David H},
title = {Does Monetary Policy Matter? {{A}} New Test in the Spirit of {{Friedman}} and {{Schwartz}}},
+ doi = {10.3386/w2966},
pages = {121--170},
volume = {4},
date-added = {2022-12-13 12:58:01 +0100},
@@ -2165,6 +2244,7 @@
@Article{rudin2019stop,
author = {Rudin, Cynthia},
title = {Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead},
+ doi = {10.1038/s42256-019-0048-x},
number = {5},
pages = {206--215},
volume = {1},
@@ -2177,6 +2257,7 @@
@Article{sacerdote2001peer,
author = {Sacerdote, Bruce},
title = {Peer Effects with Random Assignment: {{Results}} for {{Dartmouth}} Roommates},
+ doi = {10.3386/w7469},
number = {2},
pages = {681--704},
volume = {116},
@@ -2189,6 +2270,7 @@
@Article{sadinle2019least,
author = {Sadinle, Mauricio and Lei, Jing and Wasserman, Larry},
title = {Least Ambiguous Set-Valued Classifiers with Bounded Error Levels},
+ doi = {10.1080/01621459.2017.1395341},
number = {525},
pages = {223--234},
volume = {114},
@@ -2204,6 +2286,7 @@
author = {Satopaa, Ville and Albrecht, Jeannie and Irwin, David and Raghavan, Barath},
booktitle = {2011 31st International Conference on Distributed Computing Systems Workshops},
title = {Finding a" Kneedle" in a Haystack: {{Detecting}} Knee Points in System Behavior},
+ doi = {10.1109/icdcsw.2011.20},
pages = {166--171},
publisher = {{IEEE}},
date-added = {2022-12-13 12:58:01 +0100},
@@ -2247,6 +2330,7 @@
@Article{simonson1989choice,
author = {Simonson, Itamar},
title = {Choice Based on Reasons: {{The}} Case of Attraction and Compromise Effects},
+ doi = {10.1086/209205},
number = {2},
pages = {158--174},
volume = {16},
@@ -2259,6 +2343,7 @@
@Article{sims1986are,
author = {Sims, Christopher A and others},
title = {Are Forecasting Models Usable for Policy Analysis?},
+ doi = {10.21034/qr.1011},
issue = {Win},
pages = {2--16},
volume = {10},
@@ -2300,16 +2385,11 @@
year = {1974},
}
-@Unpublished{spooner2021counterfactual,
- author = {Spooner, Thomas and Dervovic, Danial and Long, Jason and Shepard, Jon and Chen, Jiahao and Magazzeni, Daniele},
- title = {Counterfactual {{Explanations}} for {{Arbitrary Regression Models}}},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2106.15212},
- eprinttype = {arxiv},
- shortjournal = {arXiv preprint arXiv:2106.15212},
- year = {2021},
+@Article{spooner2021counterfactual,
+ author = {Spooner, Thomas and Dervovic, Danial and Long, Jason and Shepard, Jon and Chen, Jiahao and Magazzeni, Daniele},
+ title = {Counterfactual explanations for arbitrary regression models},
+ journal = {arXiv preprint arXiv:2106.15212},
+ year = {2021},
}
@Article{srivastava2014dropout,
@@ -2339,6 +2419,7 @@
@Article{sturm2014simple,
author = {Sturm, Bob L},
title = {A Simple Method to Determine If a Music Information Retrieval System Is a ``Horse''},
+ doi = {10.1109/tmm.2014.2330697},
number = {6},
pages = {1636--1644},
volume = {16},
@@ -2351,6 +2432,7 @@
@Article{sunstein2003libertarian,
author = {Sunstein, Cass R and Thaler, Richard H},
title = {Libertarian Paternalism Is Not an Oxymoron},
+ doi = {10.1017/cbo9780511618031.039},
pages = {1159--1202},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
@@ -2381,6 +2463,7 @@
@Article{thaler1981empirical,
author = {Thaler, Richard},
title = {Some Empirical Evidence on Dynamic Inconsistency},
+ doi = {10.1016/0165-1765(81)90067-7},
number = {3},
pages = {201--207},
volume = {8},
@@ -2393,6 +2476,7 @@
@Article{thaler2004more,
author = {Thaler, Richard H and Benartzi, Shlomo},
title = {Save More Tomorrow{\texttrademark}: {{Using}} Behavioral Economics to Increase Employee Saving},
+ doi = {10.1086/380085},
number = {S1},
pages = {S164--S187},
volume = {112},
@@ -2405,6 +2489,7 @@
@Article{tversky1981framing,
author = {Tversky, Amos and Kahneman, Daniel},
title = {The Framing of Decisions and the Psychology of Choice},
+ doi = {10.1007/978-1-4613-2391-4_2},
number = {4481},
pages = {453--458},
volume = {211},
@@ -2418,6 +2503,7 @@
@Article{ungemach2011how,
author = {Ungemach, Christoph and Stewart, Neil and Reimers, Stian},
title = {How Incidental Values from the Environment Affect Decisions about Money, Risk, and Delay},
+ doi = {10.1177/0956797610396225},
number = {2},
pages = {253--260},
volume = {22},
@@ -2442,6 +2528,7 @@
author = {Ustun, Berk and Spangher, Alexander and Liu, Yang},
booktitle = {Proceedings of the {{Conference}} on {{Fairness}}, {{Accountability}}, and {{Transparency}}},
title = {Actionable Recourse in Linear Classification},
+ doi = {10.1145/3287560.3287566},
pages = {10--19},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
@@ -2451,6 +2538,7 @@
@Article{vanboven2000egocentric,
author = {Van Boven, Leaf and Dunning, David and Loewenstein, George},
title = {Egocentric Empathy Gaps between Owners and Buyers: Misperceptions of the Endowment Effect.},
+ doi = {10.1037/0022-3514.79.1.66},
number = {1},
pages = {66},
volume = {79},
@@ -2484,6 +2572,7 @@
@Article{verstyuk2020modeling,
author = {Verstyuk, Sergiy},
title = {Modeling Multivariate Time Series in Economics: {{From}} Auto-Regressions to Recurrent Neural Networks},
+ doi = {10.2139/ssrn.3589337},
date-added = {2022-12-13 12:58:01 +0100},
date-modified = {2022-12-13 12:58:01 +0100},
journal = {Available at SSRN 3589337},
@@ -2493,6 +2582,7 @@
@Article{wachter2017counterfactual,
author = {Wachter, Sandra and Mittelstadt, Brent and Russell, Chris},
title = {Counterfactual Explanations without Opening the Black Box: {{Automated}} Decisions and the {{GDPR}}},
+ doi = {10.2139/ssrn.3063289},
pages = {841},
volume = {31},
date-added = {2022-12-13 12:58:01 +0100},
@@ -2534,6 +2624,7 @@
@Article{widmer1996learning,
author = {Widmer, Gerhard and Kubat, Miroslav},
title = {Learning in the Presence of Concept Drift and Hidden Contexts},
+ doi = {10.1007/bf00116900},
number = {1},
pages = {69--101},
volume = {23},
@@ -2543,20 +2634,17 @@
year = {1996},
}
-@Unpublished{wilson2020case,
- author = {Wilson, Andrew Gordon},
- title = {The Case for {{Bayesian}} Deep Learning},
- archiveprefix = {arXiv},
- date-added = {2022-12-13 12:58:01 +0100},
- date-modified = {2022-12-13 12:58:01 +0100},
- eprint = {2001.10995},
- eprinttype = {arxiv},
- year = {2020},
+@Article{wilson2020case,
+ author = {Wilson, Andrew Gordon},
+ title = {The case for Bayesian deep learning},
+ journal = {arXiv preprint arXiv:2001.10995},
+ year = {2020},
}
@Article{witten2009penalized,
author = {Witten, Daniela M and Tibshirani, Robert and Hastie, Trevor},
title = {A Penalized Matrix Decomposition, with Applications to Sparse Principal Components and Canonical Correlation Analysis},
+ doi = {10.1093/biostatistics/kxp008},
number = {3},
pages = {515--534},
volume = {10},
@@ -2583,6 +2671,7 @@
@Article{yeh2009comparisons,
author = {Yeh, I-Cheng and Lien, Che-hui},
title = {The Comparisons of Data Mining Techniques for the Predictive Accuracy of Probability of Default of Credit Card Clients},
+ doi = {10.1016/j.eswa.2007.12.020},
number = {2},
pages = {2473--2480},
volume = {36},
@@ -2607,6 +2696,7 @@
@Article{zhang2003time,
author = {Zhang, G Peter},
title = {Time Series Forecasting Using a Hybrid {{ARIMA}} and Neural Network Model},
+ doi = {10.1016/s0925-2312(01)00702-0},
pages = {159--175},
volume = {50},
date-added = {2022-12-13 12:58:01 +0100},
@@ -2703,56 +2793,35 @@
}
@Book{murphy2023probabilistic,
- author = {Murphy, Kevin P.},
- date = {2023},
- title = {Probabilistic machine learning: {Advanced} topics},
- publisher = {MIT Press},
- shorttitle = {Probabilistic machine learning},
+ author = {Murphy, Kevin P},
+ title = {Probabilistic machine learning: Advanced topics},
+ publisher = {MIT press},
+ year = {2023},
}
-@TechReport{artelt2021evaluating,
- author = {Artelt, André and Vaquet, Valerie and Velioglu, Riza and Hinder, Fabian and Brinkrolf, Johannes and Schilling, Malte and Hammer, Barbara},
- date = {2021-07},
- institution = {arXiv},
- title = {Evaluating {Robustness} of {Counterfactual} {Explanations}},
- note = {arXiv:2103.02354 [cs] type: article},
- url = {http://arxiv.org/abs/2103.02354},
- urldate = {2023-03-24},
- abstract = {Transparency is a fundamental requirement for decision making systems when these should be deployed in the real world. It is usually achieved by providing explanations of the system's behavior. A prominent and intuitive type of explanations are counterfactual explanations. Counterfactual explanations explain a behavior to the user by proposing actions -- as changes to the input -- that would cause a different (specified) behavior of the system. However, such explanation methods can be unstable with respect to small changes to the input -- i.e. even a small change in the input can lead to huge or arbitrary changes in the output and of the explanation. This could be problematic for counterfactual explanations, as two similar individuals might get very different explanations. Even worse, if the recommended actions differ considerably in their complexity, one would consider such unstable (counterfactual) explanations as individually unfair. In this work, we formally and empirically study the robustness of counterfactual explanations in general, as well as under different models and different kinds of perturbations. Furthermore, we propose that plausible counterfactual explanations can be used instead of closest counterfactual explanations to improve the robustness and consequently the individual fairness of counterfactual explanations.},
- annotation = {Comment: Rewrite paper to make things more clear; Remove one theorem \& corollary due to buggy proof},
- file = {:artelt2021evaluating - Evaluating Robustness of Counterfactual Explanations.pdf:PDF},
- keywords = {Computer Science - Machine Learning, Computer Science - Artificial Intelligence},
+@InProceedings{artelt2021evaluating,
+ author = {Artelt, Andr{\'e} and Vaquet, Valerie and Velioglu, Riza and Hinder, Fabian and Brinkrolf, Johannes and Schilling, Malte and Hammer, Barbara},
+ booktitle = {2021 IEEE Symposium Series on Computational Intelligence (SSCI)},
+ title = {Evaluating robustness of counterfactual explanations},
+ organization = {IEEE},
+ pages = {01--09},
+ year = {2021},
}
@Article{guidotti2022counterfactual,
- author = {Guidotti, Riccardo},
- date = {2022-04},
- journaltitle = {Data Mining and Knowledge Discovery},
- title = {Counterfactual explanations and how to find them: literature review and benchmarking},
- doi = {10.1007/s10618-022-00831-6},
- issn = {1573-756X},
- language = {en},
- url = {https://doi.org/10.1007/s10618-022-00831-6},
- urldate = {2023-03-24},
- abstract = {Interpretable machine learning aims at unveiling the reasons behind predictions returned by uninterpretable classifiers. One of the most valuable types of explanation consists of counterfactuals. A counterfactual explanation reveals what should have been different in an instance to observe a diverse outcome. For instance, a bank customer asks for a loan that is rejected. The counterfactual explanation consists of what should have been different for the customer in order to have the loan accepted. Recently, there has been an explosion of proposals for counterfactual explainers. The aim of this work is to survey the most recent explainers returning counterfactual explanations. We categorize explainers based on the approach adopted to return the counterfactuals, and we label them according to characteristics of the method and properties of the counterfactuals returned. In addition, we visually compare the explanations, and we report quantitative benchmarking assessing minimality, actionability, stability, diversity, discriminative power, and running time. The results make evident that the current state of the art does not provide a counterfactual explainer able to guarantee all these properties simultaneously.},
- file = {Full Text PDF:https\://link.springer.com/content/pdf/10.1007%2Fs10618-022-00831-6.pdf:application/pdf},
- keywords = {Explainable AI, Counterfactual explanations, Contrastive explanations, Interpretable machine learning},
- shorttitle = {Counterfactual explanations and how to find them},
+ author = {Guidotti, Riccardo},
+ title = {Counterfactual explanations and how to find them: literature review and benchmarking},
+ pages = {1--55},
+ journal = {Data Mining and Knowledge Discovery},
+ publisher = {Springer},
+ year = {2022},
}
-@TechReport{mahajan2020preserving,
- author = {Mahajan, Divyat and Tan, Chenhao and Sharma, Amit},
- date = {2020-06},
- institution = {arXiv},
- title = {Preserving {Causal} {Constraints} in {Counterfactual} {Explanations} for {Machine} {Learning} {Classifiers}},
- doi = {10.48550/arXiv.1912.03277},
- note = {arXiv:1912.03277 [cs, stat] type: article},
- url = {http://arxiv.org/abs/1912.03277},
- urldate = {2023-03-24},
- abstract = {To construct interpretable explanations that are consistent with the original ML model, counterfactual examples---showing how the model's output changes with small perturbations to the input---have been proposed. This paper extends the work in counterfactual explanations by addressing the challenge of feasibility of such examples. For explanations of ML models in critical domains such as healthcare and finance, counterfactual examples are useful for an end-user only to the extent that perturbation of feature inputs is feasible in the real world. We formulate the problem of feasibility as preserving causal relationships among input features and present a method that uses (partial) structural causal models to generate actionable counterfactuals. When feasibility constraints cannot be easily expressed, we consider an alternative mechanism where people can label generated CF examples on feasibility: whether it is feasible to intervene and realize the candidate CF example from the original input. To learn from this labelled feasibility data, we propose a modified variational auto encoder loss for generating CF examples that optimizes for feasibility as people interact with its output. Our experiments on Bayesian networks and the widely used ''Adult-Income'' dataset show that our proposed methods can generate counterfactual explanations that better satisfy feasibility constraints than existing methods.. Code repository can be accessed here: {\textbackslash}textit\{https://github.com/divyat09/cf-feasibility\}},
- annotation = {Comment: 2019 NeurIPS Workshop on Do the right thing: Machine learning and Causal Inference for improved decision making},
- file = {:mahajan2020preserving - Preserving Causal Constraints in Counterfactual Explanations for Machine Learning Classifiers.pdf:PDF},
- keywords = {Computer Science - Machine Learning, Computer Science - Artificial Intelligence, Statistics - Machine Learning},
+@Article{mahajan2019preserving,
+ author = {Mahajan, Divyat and Tan, Chenhao and Sharma, Amit},
+ title = {Preserving causal constraints in counterfactual explanations for machine learning classifiers},
+ journal = {arXiv preprint arXiv:1912.03277},
+ year = {2019},
}
@TechReport{antoran2023sampling,
@@ -2780,32 +2849,23 @@
}
@InProceedings{welling2011bayesian,
- author = {Welling, M. and Teh, Y.},
- date = {2011-06},
- title = {Bayesian {Learning} via {Stochastic} {Gradient} {Langevin} {Dynamics}},
- url = {https://www.semanticscholar.org/paper/Bayesian-Learning-via-Stochastic-Gradient-Langevin-Welling-Teh/aeed631d6a84100b5e9a021ec1914095c66de415},
- urldate = {2023-05-15},
- abstract = {In this paper we propose a new framework for learning from large scale datasets based on iterative learning from small mini-batches. By adding the right amount of noise to a standard stochastic gradient optimization algorithm we show that the iterates will converge to samples from the true posterior distribution as we anneal the stepsize. This seamless transition between optimization and Bayesian posterior sampling provides an inbuilt protection against overfitting. We also propose a practical method for Monte Carlo estimates of posterior statistics which monitors a "sampling threshold" and collects samples after it has been surpassed. We apply the method to three models: a mixture of Gaussians, logistic regression and ICA with natural gradients.},
- annotation = {[TLDR] This paper proposes a new framework for learning from large scale datasets based on iterative learning from small mini-batches by adding the right amount of noise to a standard stochastic gradient optimization algorithm and shows that the iterates will converge to samples from the true posterior distribution as the authors anneal the stepsize.},
- file = {:welling_bayesian_2011 - Bayesian Learning Via Stochastic Gradient Langevin Dynamics.html:URL;:welling2011bayesian - Bayesian Learning Via Stochastic Gradient Langevin Dynamics.pdf:PDF},
+ author = {Welling, Max and Teh, Yee W},
+ booktitle = {Proceedings of the 28th international conference on machine learning (ICML-11)},
+ title = {Bayesian learning via stochastic gradient Langevin dynamics},
+ organization = {Citeseer},
+ pages = {681--688},
+ year = {2011},
}
@Article{gill2010circular,
- author = {Gill, Jeff and Hangartner, Dominik},
- date = {2010},
- journaltitle = {Political Analysis},
- title = {Circular {Data} in {Political} {Science} and {How} to {Handle} {It}},
- doi = {10.1093/pan/mpq009},
- issn = {1047-1987, 1476-4989},
- language = {en},
- number = {3},
- pages = {316--336},
- url = {https://www.cambridge.org/core/journals/political-analysis/article/circular-data-in-political-science-and-how-to-handle-it/6DF2D9DA60C455E6A48FFB0FF011F747},
- urldate = {2023-05-15},
- volume = {18},
- abstract = {There has been no attention to circular (purely cyclical) data in political science research. We show that such data exist and are mishandled by models that do not take into account the inherently recycling nature of some phenomenon. Clock and calendar effects are the obvious cases, but directional data are observed as well. We describe a standard maximum likelihood regression modeling framework based on the von Mises distribution, then develop a general Bayesian regression procedure for the first time, providing an easy-to-use Metropolis-Hastings sampler for this approach. Applications include a chronographic analysis of U.S. domestic terrorism and directional party preferences in a two-dimensional ideological space for German Bundestag elections. The results demonstrate the importance of circular models to handle periodic and directional data in political science.},
- file = {Full Text PDF:https\://www.cambridge.org/core/services/aop-cambridge-core/content/view/6DF2D9DA60C455E6A48FFB0FF011F747/S1047198700012493a.pdf/div-class-title-circular-data-in-political-science-and-how-to-handle-it-div.pdf:application/pdf},
- publisher = {Cambridge University Press},
+ author = {Gill, Jeff and Hangartner, Dominik},
+ title = {Circular data in political science and how to handle it},
+ number = {3},
+ pages = {316--336},
+ volume = {18},
+ journal = {Political Analysis},
+ publisher = {Cambridge University Press},
+ year = {2010},
}
@InProceedings{liu2023goggle,
@@ -2820,18 +2880,274 @@
shorttitle = {{GOGGLE}},
}
-@TechReport{du2020implicit,
- author = {Du, Yilun and Mordatch, Igor},
+@Article{du2019implicit,
+ author = {Du, Yilun and Mordatch, Igor},
+ title = {Implicit generation and generalization in energy-based models},
+ journal = {arXiv preprint arXiv:1903.08689},
+ year = {2019},
+}
+
+@InProceedings{krizhevsky2009learning,
+ author = {Krizhevsky, A.},
+ date = {2009},
+ title = {Learning {Multiple} {Layers} of {Features} from {Tiny} {Images}},
+ url = {https://www.semanticscholar.org/paper/Learning-Multiple-Layers-of-Features-from-Tiny-Krizhevsky/5d90f06bb70a0a3dced62413346235c02b1aa086},
+ urldate = {2023-06-21},
+ abstract = {Groups at MIT and NYU have collected a dataset of millions of tiny colour images from the web. It is, in principle, an excellent dataset for unsupervised training of deep generative models, but previous researchers who have tried this have found it dicult to learn a good set of lters from the images. We show how to train a multi-layer generative model that learns to extract meaningful features which resemble those found in the human visual cortex. Using a novel parallelization algorithm to distribute the work among multiple machines connected on a network, we show how training such a model can be done in reasonable time. A second problematic aspect of the tiny images dataset is that there are no reliable class labels which makes it hard to use for object recognition experiments. We created two sets of reliable labels. The CIFAR-10 set has 6000 examples of each of 10 classes and the CIFAR-100 set has 600 examples of each of 100 non-overlapping classes. Using these labels, we show that object recognition is signicantly improved by pre-training a layer of features on a large set of unlabeled tiny images.},
+ annotation = {[TLDR] It is shown how to train a multi-layer generative model that learns to extract meaningful features which resemble those found in the human visual cortex, using a novel parallelization algorithm to distribute the work among multiple machines connected on a network.},
+ file = {Semantic Scholar Link:https\://www.semanticscholar.org/paper/Learning-Multiple-Layers-of-Features-from-Tiny-Krizhevsky/5d90f06bb70a0a3dced62413346235c02b1aa086:text/html;Full Text PDF:http\://www.cs.toronto.edu/~kriz/learning-features-2009-TR.pdf:application/pdf},
+}
+
+@Misc{becker1996adult,
+ author = {Barry Becker, Ronny Kohavi},
+ date = {1996},
+ title = {Adult},
+ doi = {10.24432/C5XW20},
+ note = {Type: dataset},
+ url = {https://archive.ics.uci.edu/dataset/2},
+ urldate = {2023-06-21},
+ publisher = {UCI Machine Learning Repository},
+}
+
+@InProceedings{tolomei2017interpretable,
+ author = {Tolomei, Gabriele and Silvestri, Fabrizio and Haines, Andrew and Lalmas, Mounia},
+ booktitle = {Proceedings of the 23rd {ACM} {SIGKDD} {International} {Conference} on {Knowledge} {Discovery} and {Data} {Mining}},
+ date = {2017-08},
+ title = {Interpretable {Predictions} of {Tree}-based {Ensembles} via {Actionable} {Feature} {Tweaking}},
+ doi = {10.1145/3097983.3098039},
+ note = {arXiv:1706.06691 [stat]},
+ pages = {465--474},
+ url = {http://arxiv.org/abs/1706.06691},
+ urldate = {2023-06-21},
+ abstract = {Machine-learned models are often described as "black boxes". In many real-world applications however, models may have to sacrifice predictive power in favour of human-interpretability. When this is the case, feature engineering becomes a crucial task, which requires significant and time-consuming human effort. Whilst some features are inherently static, representing properties that cannot be influenced (e.g., the age of an individual), others capture characteristics that could be adjusted (e.g., the daily amount of carbohydrates taken). Nonetheless, once a model is learned from the data, each prediction it makes on new instances is irreversible - assuming every instance to be a static point located in the chosen feature space. There are many circumstances however where it is important to understand (i) why a model outputs a certain prediction on a given instance, (ii) which adjustable features of that instance should be modified, and finally (iii) how to alter such a prediction when the mutated instance is input back to the model. In this paper, we present a technique that exploits the internals of a tree-based ensemble classifier to offer recommendations for transforming true negative instances into positively predicted ones. We demonstrate the validity of our approach using an online advertising application. First, we design a Random Forest classifier that effectively separates between two types of ads: low (negative) and high (positive) quality ads (instances). Then, we introduce an algorithm that provides recommendations that aim to transform a low quality ad (negative instance) into a high quality one (positive instance). Finally, we evaluate our approach on a subset of the active inventory of a large ad network, Yahoo Gemini.},
+ annotation = {Comment: 10 pages, KDD 2017},
+ file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/1706.06691.pdf:application/pdf},
+ keywords = {Statistics - Machine Learning, 68T01, I.2.0, I.5.1},
+}
+
+@TechReport{dandl2023counterfactuals,
+ author = {Dandl, Susanne and Hofheinz, Andreas and Binder, Martin and Bischl, Bernd and Casalicchio, Giuseppe},
+ date = {2023-04},
+ institution = {arXiv},
+ title = {counterfactuals: {An} {R} {Package} for {Counterfactual} {Explanation} {Methods}},
+ note = {arXiv:2304.06569 [cs, stat] type: article},
+ url = {http://arxiv.org/abs/2304.06569},
+ urldate = {2023-06-21},
+ abstract = {Counterfactual explanation methods provide information on how feature values of individual observations must be changed to obtain a desired prediction. Despite the increasing amount of proposed methods in research, only a few implementations exist whose interfaces and requirements vary widely. In this work, we introduce the counterfactuals R package, which provides a modular and unified R6-based interface for counterfactual explanation methods. We implemented three existing counterfactual explanation methods and propose some optional methodological extensions to generalize these methods to different scenarios and to make them more comparable. We explain the structure and workflow of the package using real use cases and show how to integrate additional counterfactual explanation methods into the package. In addition, we compared the implemented methods for a variety of models and datasets with regard to the quality of their counterfactual explanations and their runtime behavior.},
+ file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/2304.06569.pdf:application/pdf},
+ keywords = {Statistics - Machine Learning, Computer Science - Machine Learning, Statistics - Computation},
+ shorttitle = {counterfactuals},
+}
+
+@TechReport{laugel2017inversea,
+ author = {Laugel, Thibault and Lesot, Marie-Jeanne and Marsala, Christophe and Renard, Xavier and Detyniecki, Marcin},
+ date = {2017-12},
+ institution = {arXiv},
+ title = {Inverse {Classification} for {Comparison}-based {Interpretability} in {Machine} {Learning}},
+ doi = {10.48550/arXiv.1712.08443},
+ note = {arXiv:1712.08443 [cs, stat] type: article},
+ url = {http://arxiv.org/abs/1712.08443},
+ urldate = {2023-06-21},
+ abstract = {In the context of post-hoc interpretability, this paper addresses the task of explaining the prediction of a classifier, considering the case where no information is available, neither on the classifier itself, nor on the processed data (neither the training nor the test data). It proposes an instance-based approach whose principle consists in determining the minimal changes needed to alter a prediction: given a data point whose classification must be explained, the proposed method consists in identifying a close neighbour classified differently, where the closeness definition integrates a sparsity constraint. This principle is implemented using observation generation in the Growing Spheres algorithm. Experimental results on two datasets illustrate the relevance of the proposed approach that can be used to gain knowledge about the classifier.},
+ annotation = {Comment: preprint},
+ file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/1712.08443.pdf:application/pdf},
+ keywords = {Statistics - Machine Learning, Computer Science - Artificial Intelligence, Computer Science - Machine Learning},
+}
+
+@TechReport{delaney2021uncertainty,
+ author = {Delaney, Eoin and Greene, Derek and Keane, Mark T.},
+ date = {2021-07},
+ institution = {arXiv},
+ title = {Uncertainty {Estimation} and {Out}-of-{Distribution} {Detection} for {Counterfactual} {Explanations}: {Pitfalls} and {Solutions}},
+ note = {arXiv:2107.09734 [cs] type: article},
+ url = {http://arxiv.org/abs/2107.09734},
+ urldate = {2023-06-23},
+ abstract = {Whilst an abundance of techniques have recently been proposed to generate counterfactual explanations for the predictions of opaque black-box systems, markedly less attention has been paid to exploring the uncertainty of these generated explanations. This becomes a critical issue in high-stakes scenarios, where uncertain and misleading explanations could have dire consequences (e.g., medical diagnosis and treatment planning). Moreover, it is often difficult to determine if the generated explanations are well grounded in the training data and sensitive to distributional shifts. This paper proposes several practical solutions that can be leveraged to solve these problems by establishing novel connections with other research works in explainability (e.g., trust scores) and uncertainty estimation (e.g., Monte Carlo Dropout). Two experiments demonstrate the utility of our proposed solutions.},
+ file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/2107.09734.pdf:application/pdf},
+ keywords = {Computer Science - Machine Learning, Computer Science - Artificial Intelligence},
+ shorttitle = {Uncertainty {Estimation} and {Out}-of-{Distribution} {Detection} for {Counterfactual} {Explanations}},
+}
+
+@InProceedings{casanueva2020efficient,
+ author = {Casanueva, Iñigo and TemÄinas, Tadas and Gerz, Daniela and Henderson, Matthew and Vulić, Ivan},
+ booktitle = {Proceedings of the 2nd {Workshop} on {Natural} {Language} {Processing} for {Conversational} {AI}},
+ date = {2020-07},
+ title = {Efficient {Intent} {Detection} with {Dual} {Sentence} {Encoders}},
+ doi = {10.18653/v1/2020.nlp4convai-1.5},
+ location = {Online},
+ pages = {38--45},
+ publisher = {Association for Computational Linguistics},
+ url = {https://aclanthology.org/2020.nlp4convai-1.5},
+ urldate = {2023-06-27},
+ abstract = {Building conversational systems in new domains and with added functionality requires resource-efficient models that work under low-data regimes (i.e., in few-shot setups). Motivated by these requirements, we introduce intent detection methods backed by pretrained dual sentence encoders such as USE and ConveRT. We demonstrate the usefulness and wide applicability of the proposed intent detectors, showing that: 1) they outperform intent detectors based on fine-tuning the full BERT-Large model or using BERT as a fixed black-box encoder on three diverse intent detection data sets; 2) the gains are especially pronounced in few-shot setups (i.e., with only 10 or 30 annotated examples per intent); 3) our intent detectors can be trained in a matter of minutes on a single CPU; and 4) they are stable across different hyperparameter settings. In hope of facilitating and democratizing research focused on intention detection, we release our code, as well as a new challenging single-domain intent detection dataset comprising 13,083 annotated examples over 77 intents.},
+ file = {Full Text PDF:https\://aclanthology.org/2020.nlp4convai-1.5.pdf:application/pdf},
+}
+
+@TechReport{liu2019roberta,
+ author = {Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin},
+ date = {2019-07},
+ institution = {arXiv},
+ title = {{RoBERTa}: {A} {Robustly} {Optimized} {BERT} {Pretraining} {Approach}},
+ doi = {10.48550/arXiv.1907.11692},
+ note = {arXiv:1907.11692 [cs] type: article},
+ url = {http://arxiv.org/abs/1907.11692},
+ urldate = {2023-06-27},
+ abstract = {Language model pretraining has led to significant performance gains but careful comparison between different approaches is challenging. Training is computationally expensive, often done on private datasets of different sizes, and, as we will show, hyperparameter choices have significant impact on the final results. We present a replication study of BERT pretraining (Devlin et al., 2019) that carefully measures the impact of many key hyperparameters and training data size. We find that BERT was significantly undertrained, and can match or exceed the performance of every model published after it. Our best model achieves state-of-the-art results on GLUE, RACE and SQuAD. These results highlight the importance of previously overlooked design choices, and raise questions about the source of recently reported improvements. We release our models and code.},
+ file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/1907.11692.pdf:application/pdf},
+ keywords = {Computer Science - Computation and Language},
+ shorttitle = {{RoBERTa}},
+}
+
+@TechReport{kunstner2020limitations,
+ author = {Kunstner, Frederik and Balles, Lukas and Hennig, Philipp},
date = {2020-06},
institution = {arXiv},
- title = {Implicit {Generation} and {Generalization} in {Energy}-{Based} {Models}},
- doi = {10.48550/arXiv.1903.08689},
- note = {arXiv:1903.08689 [cs, stat] type: article},
- url = {http://arxiv.org/abs/1903.08689},
- urldate = {2023-05-16},
- abstract = {Energy based models (EBMs) are appealing due to their generality and simplicity in likelihood modeling, but have been traditionally difficult to train. We present techniques to scale MCMC based EBM training on continuous neural networks, and we show its success on the high-dimensional data domains of ImageNet32x32, ImageNet128x128, CIFAR-10, and robotic hand trajectories, achieving better samples than other likelihood models and nearing the performance of contemporary GAN approaches, while covering all modes of the data. We highlight some unique capabilities of implicit generation such as compositionality and corrupt image reconstruction and inpainting. Finally, we show that EBMs are useful models across a wide variety of tasks, achieving state-of-the-art out-of-distribution classification, adversarially robust classification, state-of-the-art continual online class learning, and coherent long term predicted trajectory rollouts.},
- file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/1903.08689.pdf:application/pdf},
- keywords = {Computer Science - Machine Learning, Computer Science - Computer Vision and Pattern Recognition, Statistics - Machine Learning},
+ title = {Limitations of the {Empirical} {Fisher} {Approximation} for {Natural} {Gradient} {Descent}},
+ doi = {10.48550/arXiv.1905.12558},
+ note = {arXiv:1905.12558 [cs, stat] type: article},
+ url = {http://arxiv.org/abs/1905.12558},
+ urldate = {2023-06-30},
+ abstract = {Natural gradient descent, which preconditions a gradient descent update with the Fisher information matrix of the underlying statistical model, is a way to capture partial second-order information. Several highly visible works have advocated an approximation known as the empirical Fisher, drawing connections between approximate second-order methods and heuristics like Adam. We dispute this argument by showing that the empirical Fisher---unlike the Fisher---does not generally capture second-order information. We further argue that the conditions under which the empirical Fisher approaches the Fisher (and the Hessian) are unlikely to be met in practice, and that, even on simple optimization problems, the pathologies of the empirical Fisher can have undesirable effects.},
+ annotation = {Comment: V3: Minor corrections (typographic errors)},
+ file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/1905.12558.pdf:application/pdf},
+ keywords = {Computer Science - Machine Learning, Statistics - Machine Learning},
+}
+
+@TechReport{botev2017practical,
+ author = {Botev, Aleksandar and Ritter, Hippolyt and Barber, David},
+ date = {2017-06},
+ institution = {arXiv},
+ title = {Practical {Gauss}-{Newton} {Optimisation} for {Deep} {Learning}},
+ doi = {10.48550/arXiv.1706.03662},
+ note = {arXiv:1706.03662 [stat] type: article},
+ url = {http://arxiv.org/abs/1706.03662},
+ urldate = {2023-06-30},
+ abstract = {We present an efficient block-diagonal ap- proximation to the Gauss-Newton matrix for feedforward neural networks. Our result- ing algorithm is competitive against state- of-the-art first order optimisation methods, with sometimes significant improvement in optimisation performance. Unlike first-order methods, for which hyperparameter tuning of the optimisation parameters is often a labo- rious process, our approach can provide good performance even when used with default set- tings. A side result of our work is that for piecewise linear transfer functions, the net- work objective function can have no differ- entiable local maxima, which may partially explain why such transfer functions facilitate effective optimisation.},
+ annotation = {Comment: ICML 2017},
+ file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/1706.03662.pdf:application/pdf},
+ keywords = {Statistics - Machine Learning},
+}
+
+@TechReport{sharma2021sketching,
+ author = {Sharma, Apoorva and Azizan, Navid and Pavone, Marco},
+ date = {2021-02},
+ institution = {arXiv},
+ title = {Sketching {Curvature} for {Efficient} {Out}-of-{Distribution} {Detection} for {Deep} {Neural} {Networks}},
+ doi = {10.48550/arXiv.2102.12567},
+ note = {arXiv:2102.12567 [cs] type: article},
+ url = {http://arxiv.org/abs/2102.12567},
+ urldate = {2023-06-30},
+ abstract = {In order to safely deploy Deep Neural Networks (DNNs) within the perception pipelines of real-time decision making systems, there is a need for safeguards that can detect out-of-training-distribution (OoD) inputs both efficiently and accurately. Building on recent work leveraging the local curvature of DNNs to reason about epistemic uncertainty, we propose Sketching Curvature of OoD Detection (SCOD), an architecture-agnostic framework for equipping any trained DNN with a task-relevant epistemic uncertainty estimate. Offline, given a trained model and its training data, SCOD employs tools from matrix sketching to tractably compute a low-rank approximation of the Fisher information matrix, which characterizes which directions in the weight space are most influential on the predictions over the training data. Online, we estimate uncertainty by measuring how much perturbations orthogonal to these directions can alter predictions at a new test input. We apply SCOD to pre-trained networks of varying architectures on several tasks, ranging from regression to classification. We demonstrate that SCOD achieves comparable or better OoD detection performance with lower computational burden relative to existing baselines.},
+ file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/2102.12567.pdf:application/pdf},
+ keywords = {Computer Science - Machine Learning},
+}
+
+@TechReport{amini2019spatial,
+ author = {Amini, Alexander and Soleimany, Ava and Karaman, Sertac and Rus, Daniela},
+ date = {2019-05},
+ institution = {arXiv},
+ title = {Spatial {Uncertainty} {Sampling} for {End}-to-{End} {Control}},
+ doi = {10.48550/arXiv.1805.04829},
+ note = {arXiv:1805.04829 [cs] type: article},
+ url = {http://arxiv.org/abs/1805.04829},
+ urldate = {2023-06-30},
+ abstract = {End-to-end trained neural networks (NNs) are a compelling approach to autonomous vehicle control because of their ability to learn complex tasks without manual engineering of rule-based decisions. However, challenging road conditions, ambiguous navigation situations, and safety considerations require reliable uncertainty estimation for the eventual adoption of full-scale autonomous vehicles. Bayesian deep learning approaches provide a way to estimate uncertainty by approximating the posterior distribution of weights given a set of training data. Dropout training in deep NNs approximates Bayesian inference in a deep Gaussian process and can thus be used to estimate model uncertainty. In this paper, we propose a Bayesian NN for end-to-end control that estimates uncertainty by exploiting feature map correlation during training. This approach achieves improved model fits, as well as tighter uncertainty estimates, than traditional element-wise dropout. We evaluate our algorithms on a challenging dataset collected over many different road types, times of day, and weather conditions, and demonstrate how uncertainties can be used in conjunction with a human controller in a parallel autonomous setting.},
+ annotation = {Comment: Originally published in Neural Information Processing Systems (NIPS) Workshop on Bayesian Deep Learning 2017},
+ file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/1805.04829.pdf:application/pdf},
+ keywords = {Computer Science - Artificial Intelligence, Computer Science - Machine Learning, Computer Science - Robotics},
+}
+
+@InProceedings{lecun1989optimal,
+ author = {LeCun, Yann and Denker, John and Solla, Sara},
+ booktitle = {Advances in {Neural} {Information} {Processing} {Systems}},
+ date = {1989},
+ title = {Optimal {Brain} {Damage}},
+ publisher = {Morgan-Kaufmann},
+ url = {https://proceedings.neurips.cc/paper/1989/hash/6c9882bbac1c7093bd25041881277658-Abstract.html},
+ urldate = {2023-06-30},
+ volume = {2},
+ abstract = {We have used information-theoretic ideas to derive a class of prac(cid:173) tical and nearly optimal schemes for adapting the size of a neural network. By removing unimportant weights from a network, sev(cid:173) eral improvements can be expected: better generalization, fewer training examples required, and improved speed of learning and/or classification. The basic idea is to use second-derivative informa(cid:173) tion to make a tradeoff between network complexity and training set error. Experiments confirm the usefulness of the methods on a real-world application.},
+ file = {Full Text PDF:https\://proceedings.neurips.cc/paper_files/paper/1989/file/6c9882bbac1c7093bd25041881277658-Paper.pdf:application/pdf},
+}
+
+@TechReport{martens2020optimizing,
+ author = {Martens, James and Grosse, Roger},
+ date = {2020-06},
+ institution = {arXiv},
+ title = {Optimizing {Neural} {Networks} with {Kronecker}-factored {Approximate} {Curvature}},
+ doi = {10.48550/arXiv.1503.05671},
+ note = {arXiv:1503.05671 [cs, stat] type: article},
+ url = {http://arxiv.org/abs/1503.05671},
+ urldate = {2023-06-30},
+ abstract = {We propose an efficient method for approximating natural gradient descent in neural networks which we call Kronecker-Factored Approximate Curvature (K-FAC). K-FAC is based on an efficiently invertible approximation of a neural network's Fisher information matrix which is neither diagonal nor low-rank, and in some cases is completely non-sparse. It is derived by approximating various large blocks of the Fisher (corresponding to entire layers) as being the Kronecker product of two much smaller matrices. While only several times more expensive to compute than the plain stochastic gradient, the updates produced by K-FAC make much more progress optimizing the objective, which results in an algorithm that can be much faster than stochastic gradient descent with momentum in practice. And unlike some previously proposed approximate natural-gradient/Newton methods which use high-quality non-diagonal curvature matrices (such as Hessian-free optimization), K-FAC works very well in highly stochastic optimization regimes. This is because the cost of storing and inverting K-FAC's approximation to the curvature matrix does not depend on the amount of data used to estimate it, which is a feature typically associated only with diagonal or low-rank approximations to the curvature matrix.},
+ annotation = {Comment: Reduction ratio formula corrected. Removed incorrect claim about geodesics in footnote},
+ file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/1503.05671.pdf:application/pdf},
+ keywords = {Computer Science - Machine Learning, Computer Science - Neural and Evolutionary Computing, Statistics - Machine Learning},
+}
+
+@TechReport{fong2021conformal,
+ author = {Fong, Edwin and Holmes, Chris},
+ date = {2021-06},
+ institution = {arXiv},
+ title = {Conformal {Bayesian} {Computation}},
+ doi = {10.48550/arXiv.2106.06137},
+ note = {arXiv:2106.06137 [stat] type: article},
+ url = {http://arxiv.org/abs/2106.06137},
+ urldate = {2023-07-19},
+ abstract = {We develop scalable methods for producing conformal Bayesian predictive intervals with finite sample calibration guarantees. Bayesian posterior predictive distributions, \$p(y {\textbackslash}mid x)\$, characterize subjective beliefs on outcomes of interest, \$y\$, conditional on predictors, \$x\$. Bayesian prediction is well-calibrated when the model is true, but the predictive intervals may exhibit poor empirical coverage when the model is misspecified, under the so called \$\{{\textbackslash}cal\{M\}\}\$-open perspective. In contrast, conformal inference provides finite sample frequentist guarantees on predictive confidence intervals without the requirement of model fidelity. Using 'add-one-in' importance sampling, we show that conformal Bayesian predictive intervals are efficiently obtained from re-weighted posterior samples of model parameters. Our approach contrasts with existing conformal methods that require expensive refitting of models or data-splitting to achieve computational efficiency. We demonstrate the utility on a range of examples including extensions to partially exchangeable settings such as hierarchical models.},
+ annotation = {Comment: 19 pages, 4 figures, 12 tables; added references and fixed typos},
+ file = {arXiv Fulltext PDF:https\://arxiv.org/pdf/2106.06137.pdf:application/pdf},
+ keywords = {Statistics - Methodology, Statistics - Computation},
+}
+
+@Book{hyndman2018forecasting,
+ author = {Hyndman, Rob J. and Athanasopoulos, George},
+ date = {2018-05},
+ title = {Forecasting: principles and practice},
+ isbn = {9780987507112},
+ language = {en},
+ note = {Google-Books-ID: \_bBhDwAAQBAJ},
+ publisher = {OTexts},
+ abstract = {Forecasting is required in many situations. Stocking an inventory may require forecasts of demand months in advance. Telecommunication routing requires traffic forecasts a few minutes ahead. Whatever the circumstances or time horizons involved, forecasting is an important aid in effective and efficient planning.This textbook provides a comprehensive introduction to forecasting methods and presents enough information about each method for readers to use them sensibly.},
+ file = {Google Books Link:https\://books.google.com/books?id=_bBhDwAAQBAJ:text/html},
+ keywords = {Business \& Economics / Forecasting, Business \& Economics / Statistics, Computers / Databases / Data Mining, Computers / Mathematical \& Statistical Software},
+ shorttitle = {Forecasting},
+}
+
+@InProceedings{xu2021conformal,
+ author = {Xu, Chen and Xie, Yao},
+ date = {2021-07},
+ title = {Conformal prediction interval for dynamic time-series},
+ language = {en},
+ pages = {11559--11569},
+ publisher = {PMLR},
+ url = {https://proceedings.mlr.press/v139/xu21h.html},
+ urldate = {2023-07-24},
+ abstract = {We develop a method to construct distribution-free prediction intervals for dynamic time-series, called {\textbackslash}Verb{\textbar}EnbPI{\textbar} that wraps around any bootstrap ensemble estimator to construct sequential prediction intervals. {\textbackslash}Verb{\textbar}EnbPI{\textbar} is closely related to the conformal prediction (CP) framework but does not require data exchangeability. Theoretically, these intervals attain finite-sample, {\textbackslash}textit\{approximately valid\} marginal coverage for broad classes of regression functions and time-series with strongly mixing stochastic errors. Computationally, {\textbackslash}Verb{\textbar}EnbPI{\textbar} avoids overfitting and requires neither data-splitting nor training multiple ensemble estimators; it efficiently aggregates bootstrap estimators that have been trained. In general, {\textbackslash}Verb{\textbar}EnbPI{\textbar} is easy to implement, scalable to producing arbitrarily many prediction intervals sequentially, and well-suited to a wide range of regression functions. We perform extensive real-data analyses to demonstrate its effectiveness.},
+ file = {Full Text PDF:http\://proceedings.mlr.press/v139/xu21h/xu21h.pdf:application/pdf},
+ issn = {2640-3498},
+}
+
+@Book{howard2020deep,
+ author = {Howard, Jeremy and Gugger, Sylvain},
+ date = {2020-06},
+ title = {Deep {Learning} for {Coders} with fastai and {PyTorch}},
+ isbn = {9781492045496},
+ language = {en},
+ note = {Google-Books-ID: yATuDwAAQBAJ},
+ publisher = {"O'Reilly Media, Inc."},
+ abstract = {Deep learning is often viewed as the exclusive domain of math PhDs and big tech companies. But as this hands-on guide demonstrates, programmers comfortable with Python can achieve impressive results in deep learning with little math background, small amounts of data, and minimal code. How? With fastai, the first library to provide a consistent interface to the most frequently used deep learning applications.Authors Jeremy Howard and Sylvain Gugger, the creators of fastai, show you how to train a model on a wide range of tasks using fastai and PyTorch. You’ll also dive progressively further into deep learning theory to gain a complete understanding of the algorithms behind the scenes.Train models in computer vision, natural language processing, tabular data, and collaborative filteringLearn the latest deep learning techniques that matter most in practiceImprove accuracy, speed, and reliability by understanding how deep learning models workDiscover how to turn your models into web applicationsImplement deep learning algorithms from scratchConsider the ethical implications of your workGain insight from the foreword by PyTorch cofounder, Soumith Chintala},
+ file = {Google Books Link:https\://books.google.nl/books?id=yATuDwAAQBAJ:text/html},
+ keywords = {Computers / Data Science / Machine Learning, Computers / Image Processing, Computers / Computer Science, Computers / Machine Theory, Computers / Data Science / Neural Networks, Computers / Programming / General, Computers / Languages / Python, Computers / Data Science / Data Visualization},
}
@Comment{jabref-meta: databaseType:biblatex;}
+
+@Comment{jabref-meta: keypatterndefault:[auth:lower][year][veryshorttitle:lower];}
+
+@Comment{jabref-meta: saveActions:disabled;
+all-text-fields[identity]
+date[normalize_date]
+month[normalize_month]
+pages[normalize_page_numbers]
+;}
diff --git a/paper/body.tex b/paper/body.tex
index 294a7e68..750f31d5 100644
--- a/paper/body.tex
+++ b/paper/body.tex
@@ -9,7 +9,7 @@
Counterfactual explanations provide a powerful, flexible and intuitive way to not only explain black-box models but also help affected individuals through the means of algorithmic recourse. Instead of opening the black box, counterfactual explanations work under the premise of strategically perturbing model inputs to understand model behaviour~\citep{wachter2017counterfactual}. Intuitively speaking, we generate explanations in this context by asking what-if questions of the following nature: `Our credit risk model currently predicts that this individual is not credit-worthy. What if they reduced their monthly expenditures by 10\%?'
-This is typically implemented by defining a target outcome $\mathbf{y}^+ \in \mathcal{Y}$ for some individual $\mathbf{x} \in \mathcal{X}=\mathbb{R}^D$ described by $D$ attributes, for which the model $M_{\theta}:\mathcal{X}\mapsto\mathcal{Y}$ initially predicts a different outcome: $M_{\theta}(\mathbf{x})\ne \mathbf{y}^+$. Counterfactuals are then searched by minimizing a loss function that compares the predicted model output to the target outcome: $\text{yloss}(M_{\theta}(\mathbf{x}),\mathbf{y}^+)$. Since counterfactual explanations work directly with the black-box model, valid counterfactuals always have full local fidelity by construction where fidelity is defined as the degree to which explanations approximate the predictions of a black-box model~\citep{mothilal2020explaining,molnar2020interpretable}.
+This is typically implemented by defining a target outcome $\mathbf{y}^+ \in \mathcal{Y}$ for some individual $\mathbf{x} \in \mathcal{X}=\mathbb{R}^D$ described by $D$ attributes, for which the model $M_{\theta}:\mathcal{X}\mapsto\mathcal{Y}$ initially predicts a different outcome: $M_{\theta}(\mathbf{x})\ne \mathbf{y}^+$. Counterfactuals are then searched by minimizing a loss function that compares the predicted model output to the target outcome: $\text{yloss}(M_{\theta}(\mathbf{x}),\mathbf{y}^+)$. Since counterfactual explanations work directly with the black-box model, valid counterfactuals always have full local fidelity by construction where fidelity is defined as the degree to which explanations approximate the predictions of a black-box model~\citep{mothilal2020explaining,molnar2022interpretable}.
In situations where full fidelity is a requirement, counterfactual explanations offer a more appropriate solution to Explainable Artificial Intelligence (XAI) than other popular approaches like LIME~\citep{ribeiro2016why} and SHAP~\citep{lundberg2017unified}, which involve local surrogate models. But even full fidelity is not a sufficient condition for ensuring that an explanation faithfully describes the behaviour of a model. That is because multiple very distinct explanations can all lead to the same model prediction, especially when dealing with heavily parameterized models like deep neural networks, which are typically underspecified by the data~\citep{wilson2020case}.
@@ -39,7 +39,7 @@ Here $\text{yloss}(\cdot)$ denotes the primary loss function, $f(\cdot)$ is a fu
The baseline approach, which we will simply refer to as \textit{Wachter}, searches a single counterfactual directly in the feature space and penalises its distance to the original factual. In this case, $f(\cdot)$ is simply the identity function and $\mathcal{Z}$ corresponds to the feature space itself. Many derivative works of~\citet{wachter2017counterfactual} have proposed new flavours of Equation~\ref{eq:general}, each of them designed to address specific \textit{desiderata} that counterfactuals ought to meet in order to properly serve both AI practitioners and individuals affected by algorithmic decision-making systems. The list of desiderata includes but is not limited to the following: sparsity, proximity~\citep{wachter2017counterfactual}, actionability~\citep{ustun2019actionable}, diversity~\citep{mothilal2020explaining}, plausibility~\citep{joshi2019realistic,poyiadzi2020face,schut2021generating}, robustness~\citep{upadhyay2021robust,pawelczyk2022probabilistically,altmeyer2023endogenous} and causality~\citep{karimi2021algorithmic}. Different counterfactual generators addressing these needs have been extensively surveyed and evaluated in various studies~\citep{verma2020counterfactual,karimi2020survey,pawelczyk2021carla,artelt2021evaluating,guidotti2022counterfactual}.
-The notion of plausibility is central to all of the desiderata. For example, \citet{artelt2021evaluating} find that plausibility typically also leads to improved robustness. Similarly, plausibility has also been connected to causality in the sense that plausible counterfactuals respect causal relationships~\citep{mahajan2020preserving}.
+The notion of plausibility is central to all of the desiderata. For example, \citet{artelt2021evaluating} find that plausibility typically also leads to improved robustness. Similarly, plausibility has also been connected to causality in the sense that plausible counterfactuals respect causal relationships~\citep{mahajan2019preserving}.
Consequently, the plausibility of counterfactuals has been among the primary concerns for researchers. Achieving plausibility is equivalent to ensuring that the generated counterfactuals comply with the true and unobserved data-generating process (DGP). We define plausibility formally in this work as follows:
@@ -83,7 +83,7 @@ In doing this, we merge in and nuance the concept of plausibility (Definition~\r
\subsection{Quantifying the Model's Generative Property}
-To assess counterfactuals with respect to Definition~\ref{def:faithful}, we need a way to quantify the posterior conditional distribution $p_{\theta}(\mathbf{x}|\mathbf{y}^+)$. To this end, we draw on recent advances in energy-based modelling (EBM), a subdomain of machine learning that is concerned with generative or hybrid modelling~\citep{grathwohl2020your,du2020implicit}. In particular, note that if we fix $\mathbf{y}$ to our target value $\mathbf{y}^+$, we can conditionally draw from $p_{\theta}(\mathbf{x}|\mathbf{y}^+)$ by randomly initializing $\mathbf{x}_0$ and then using Stochastic Gradient Langevin Dynamics (SGLD) as follows,
+To assess counterfactuals with respect to Definition~\ref{def:faithful}, we need a way to quantify the posterior conditional distribution $p_{\theta}(\mathbf{x}|\mathbf{y}^+)$. To this end, we draw on recent advances in energy-based modelling (EBM), a subdomain of machine learning that is concerned with generative or hybrid modelling~\citep{grathwohl2020your,du2019implicit}. In particular, note that if we fix $\mathbf{y}$ to our target value $\mathbf{y}^+$, we can conditionally draw from $p_{\theta}(\mathbf{x}|\mathbf{y}^+)$ by randomly initializing $\mathbf{x}_0$ and then using Stochastic Gradient Langevin Dynamics (SGLD) as follows,
\begin{equation}\label{eq:sgld}
\begin{aligned}
@@ -214,7 +214,7 @@ For the \textit{Circles} data, it appears that \textit{REVISE} performs well, bu
The results for our real-world datasets are shown in Table~\ref{tab:results-real-world}. Once again the findings indicate that the plausibility of ECCCos is positively correlated with the capacity of the black-box model to distinguish plausible from implausible inputs. The case is very clear for \textit{MNIST}: ECCCos are consistently more faithful than the counterfactuals produced by our benchmark generators and their plausibility gradually improves through ensembling and joint-energy modelling. Interestingly, faithfulness also gradually improves for \textit{REVISE}. This indicates that as our models improve, their generative capacity approaches that of the surrogate VAE used by \textit{REVISE}. The VAE still outperforms our classifiers in this regard, as evident from the fact that \textit{ECCCo} never quite reaches the same level of plausibility as \textit{REVISE}. With reference to Appendix~\ref{app:results} we note that the results for \textit{Schut} need to be discounted as it rarely produces valid counterfactuals for \textit{MNIST}. Relatedly, we find that \textit{ECCCo} is the only generator that consistently achieves full validity. Finally, it is worth noting that \textit{ECCCo} produces counterfactual images with the lowest average predictive uncertainty for all models.
-For the tabular credit dataset (\textit{GMSC}) it is inherently challenging to use deep neural networks in order to achieve good discriminative performance~\citep{borisov2021deep,grinsztajn2022why} and generative performance~\citep{liu2023goggle}, respectively. In order to achieve high plausibility, \textit{ECCCo} effectively requires classifiers to achieve good performance for both tasks. Since this is a challenging task even for Joint Energy Models, it is not surprising to find that even though \textit{ECCCo} once again achieves state-of-the-art faithfulness, it is outperformed by \textit{REVISE} and \textit{Schut} with respect to plausibility.
+For the tabular credit dataset (\textit{GMSC}) it is inherently challenging to use deep neural networks in order to achieve good discriminative performance~\citep{borisov2022deep,grinsztajn2022why} and generative performance~\citep{liu2023goggle}, respectively. In order to achieve high plausibility, \textit{ECCCo} effectively requires classifiers to achieve good performance for both tasks. Since this is a challenging task even for Joint Energy Models, it is not surprising to find that even though \textit{ECCCo} once again achieves state-of-the-art faithfulness, it is outperformed by \textit{REVISE} and \textit{Schut} with respect to plausibility.
\subsection{Key Takeways}
--
GitLab