diff --git a/notebooks/intro.html b/notebooks/intro.html
deleted file mode 100644
index be6a3d1ac0373796ac2fbab1be6a77c596e001cf..0000000000000000000000000000000000000000
--- a/notebooks/intro.html
+++ /dev/null
@@ -1,579 +0,0 @@
-<!DOCTYPE html>
-<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en">
-
-<head>
-
-<meta charset="utf-8" />
-<meta name="generator" content="quarto-99.9.9" />
-
-<meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes" />
-
-
-<title>Conformal Counterfactual Explanations – 1 ConformalGenerator</title>
-<style>
-code{white-space: pre-wrap;}
-span.smallcaps{font-variant: small-caps;}
-div.columns{display: flex; gap: min(4vw, 1.5em);}
-div.column{flex: auto; overflow-x: auto;}
-div.hanging-indent{margin-left: 1.5em; text-indent: -1.5em;}
-ul.task-list{list-style: none;}
-ul.task-list li input[type="checkbox"] {
- width: 0.8em;
- margin: 0 0.8em 0.2em -1em; /* quarto-specific, see https://github.com/quarto-dev/quarto-cli/issues/4556 */
- vertical-align: middle;
-}
-/* CSS for citations */
-div.csl-bib-body { }
-div.csl-entry {
- clear: both;
-}
-.hanging-indent div.csl-entry {
- margin-left:2em;
- text-indent:-2em;
-}
-div.csl-left-margin {
- min-width:2em;
- float:left;
-}
-div.csl-right-inline {
- margin-left:2em;
- padding-left:1em;
-}
-div.csl-indent {
- margin-left: 2em;
-}</style>
-
-<!-- htmldependencies:E3FAD763 -->
-<script id="quarto-search-options" type="application/json">{
- "location": "sidebar",
- "copy-button": false,
- "collapse-after": 3,
- "panel-placement": "start",
- "type": "textbox",
- "limit": 20,
- "language": {
- "search-no-results-text": "No results",
- "search-matching-documents-text": "matching documents",
- "search-copy-link-title": "Copy link to search",
- "search-hide-matches-text": "Hide additional matches",
- "search-more-match-text": "more match in this document",
- "search-more-matches-text": "more matches in this document",
- "search-clear-button-title": "Clear",
- "search-detached-cancel-button-title": "Cancel",
- "search-submit-button-title": "Submit"
- }
-}</script>
-
- <script src="https://polyfill.io/v3/polyfill.min.js?features=es6"></script>
- <script src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-chtml-full.js" type="text/javascript"></script>
-
-</head>
-
-<body>
-
-<div id="quarto-search-results"></div>
- <header id="quarto-header" class="headroom fixed-top">
- <nav class="quarto-secondary-nav">
- <div class="container-fluid d-flex">
- <button type="button" class="quarto-btn-toggle btn"
- data-bs-toggle="collapse" data-bs-target="#quarto-sidebar,#quarto-sidebar-glass"
- aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation"
- onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
- <i class="bi bi-layout-text-sidebar-reverse"></i>
- </button>
- <h1 class="quarto-secondary-nav-title"></h1>
- <a class="flex-grow-1" role="button" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar,#quarto-sidebar-glass"
- aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation"
- onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
- </a>
- <button type="button" class="btn quarto-search-button" aria-label="Search" onclick="window.quartoOpenSearch();">
- <i class="bi bi-search"></i>
- </button>
- </div>
- </nav>
-</header>
-<!-- content -->
-<div id="quarto-content" class="quarto-container page-columns page-rows-contents page-layout-article">
-<!-- sidebar -->
- <nav id="quarto-sidebar" class="sidebar collapse collapse-horizontal sidebar-navigation floating overflow-auto">
- <div class="pt-lg-2 mt-2 text-left sidebar-header">
- <div class="sidebar-title mb-0 py-0">
- <a href="/">
- Conformal Counterfactual Explanations
- </a>
- </div>
- </div>
- <div class="mt-2 flex-shrink-0 align-items-center">
- <div class="sidebar-search">
- <div id="quarto-search" class="" title="Search"></div>
- </div>
- </div>
- <div class="sidebar-menu-container">
- <ul class="list-unstyled mt-1">
- <li class="sidebar-item">
- <div class="sidebar-item-container">
- <a href="/index.html" class="sidebar-item-text sidebar-link">
- <span class="menu-text">Preface</span></a>
- </div>
-</li>
- <li class="sidebar-item">
- <div class="sidebar-item-container">
- <a href="/notebooks/intro.html" class="sidebar-item-text sidebar-link active">
- <span class="menu-text"><span class='chapter-number'>1</span>Â <span class='chapter-title'>`ConformalGenerator`</span></span></a>
- </div>
-</li>
- <li class="sidebar-item">
- <div class="sidebar-item-container">
- <a href="/notebooks/references.html" class="sidebar-item-text sidebar-link">
- <span class="menu-text">References</span></a>
- </div>
-</li>
- </ul>
- </div>
-</nav>
-<div id="quarto-sidebar-glass" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar,#quarto-sidebar-glass" ></div>
-<!-- margin-sidebar -->
- <div id="quarto-margin-sidebar" class="sidebar margin-sidebar">
- <div id="quarto-toc-target"></div>
- </div>
-<!-- main -->
-<main class="content" id="quarto-document-content">
-
-<header id="title-block-header" class="quarto-title-block default">
-<div class="quarto-title">
-<h1 class="title"><span class="chapter-number">1</span>Â <span class="chapter-title"><code>ConformalGenerator</code></span></h1>
-</div>
-
-
-
-<div class="quarto-title-meta">
-
-
-
-
- </div>
-
-
-</header>
-<nav id="TOC" role="doc-toc">
- <h2 id="toc-title">Table of contents</h2>
-
- <ul>
- <li><a href="#black-box-model" id="toc-black-box-model"><span class="header-section-number">1.1</span> Black-box Model</a></li>
- <li><a href="#conformal-prediction" id="toc-conformal-prediction"><span class="header-section-number">1.2</span> Conformal Prediction</a></li>
- <li><a href="#differentiable-cp" id="toc-differentiable-cp"><span class="header-section-number">1.3</span> Differentiable CP</a>
- <ul>
- <li><a href="#smooth-set-size-penalty" id="toc-smooth-set-size-penalty"><span class="header-section-number">1.3.1</span> Smooth Set Size Penalty</a></li>
- <li><a href="#configurable-classification-loss" id="toc-configurable-classification-loss"><span class="header-section-number">1.3.2</span> Configurable Classification Loss</a></li>
- </ul></li>
- <li><a href="#fidelity-and-plausibility" id="toc-fidelity-and-plausibility"><span class="header-section-number">1.4</span> Fidelity and Plausibility</a>
- <ul>
- <li><a href="#fidelity" id="toc-fidelity"><span class="header-section-number">1.4.1</span> Fidelity</a></li>
- <li><a href="#plausibility" id="toc-plausibility"><span class="header-section-number">1.4.2</span> Plausibility</a></li>
- </ul></li>
- <li><a href="#counterfactual-explanations" id="toc-counterfactual-explanations"><span class="header-section-number">1.5</span> Counterfactual Explanations</a>
- <ul>
- <li><a href="#benchmark" id="toc-benchmark"><span class="header-section-number">1.5.1</span> Benchmark</a></li>
- </ul></li>
- <li><a href="#multi-class" id="toc-multi-class"><span class="header-section-number">1.6</span> Multi-Class</a></li>
- </ul>
-</nav>
-<p>In this section, we will look at a simple example involving synthetic data, a black-box model and a generic Conformal Counterfactual Generator.</p>
-<section id="black-box-model" class="level2" data-number="1.1">
-<h2 data-number="1.1"><span class="header-section-number">1.1</span> Black-box Model</h2>
-<p>We consider a simple binary classification problem. Let <span class="math inline">\((X_i, Y_i), \ i=1,...,n\)</span> denote our feature-label pairs and let <span class="math inline">\(\mu: \mathcal{X} \mapsto \mathcal{Y}\)</span> denote the mapping from features to labels. For illustration purposes, we will use linearly separable data.</p>
-<p>While we could use a linear classifier in this case, let’s pretend we need a black-box model for this task and rely on a small Multi-Layer Perceptron (MLP):</p>
-<p>We can fit this model to data to produce plug-in predictions.</p>
-</section>
-<section id="conformal-prediction" class="level2" data-number="1.2">
-<h2 data-number="1.2"><span class="header-section-number">1.2</span> Conformal Prediction</h2>
-<p>Here we will instead use a specific case of CP called <em>split conformal prediction</em> which can then be summarized as follows:<a href="#fn1" class="footnote-ref" id="fnref1" role="doc-noteref"><sup>1</sup></a></p>
-<ol type="1">
-<li>Partition the training into a proper training set and a separate calibration set: <span class="math inline">\(\mathcal{D}_n=\mathcal{D}^{\text{train}} \cup \mathcal{D}^{\text{cali}}\)</span>.</li>
-<li>Train the machine learning model on the proper training set: <span class="math inline">\(\hat\mu_{i \in \mathcal{D}^{\text{train}}}(X_i,Y_i)\)</span>.</li>
-</ol>
-<p>The model <span class="math inline">\(\hat\mu_{i \in \mathcal{D}^{\text{train}}}\)</span> can now produce plug-in predictions.</p>
-<div class="callout callout-style-default callout-note callout-titled">
-<div class="callout-header d-flex align-content-center">
-<div class="callout-icon-container">
-<i class='callout-icon'></i>
-</div>
-<div class="callout-title-container flex-fill">
-Starting Point
-</div>
-</div>
-<div class="callout-body-container callout-body">
-<p>Note that this represents the starting point in applications of Algorithmic Recourse: we have some pre-trained classifier <span class="math inline">\(M\)</span> for which we would like to generate plausible Counterfactual Explanations. Next, we turn to the calibration step.</p>
-</div>
-</div>
-<ol start="3" type="1">
-<li>Compute nonconformity scores, <span class="math inline">\(\mathcal{S}\)</span>, using the calibration data <span class="math inline">\(\mathcal{D}^{\text{cali}}\)</span> and the fitted model <span class="math inline">\(\hat\mu_{i \in \mathcal{D}^{\text{train}}}\)</span>.</li>
-<li>For a user-specified desired coverage ratio <span class="math inline">\((1-\alpha)\)</span> compute the corresponding quantile, <span class="math inline">\(\hat{q}\)</span>, of the empirical distribution of nonconformity scores, <span class="math inline">\(\mathcal{S}\)</span>.</li>
-<li>For the given quantile and test sample <span class="math inline">\(X_{\text{test}}\)</span>, form the corresponding conformal prediction set:</li>
-</ol>
-<p><span id="eq-set"><span class="math display">\[
-C(X_{\text{test}})=\{y:s(X_{\text{test}},y) \le \hat{q}\}
-\tag{1.1}\]</span></span></p>
-<p>This is the default procedure used for classification and regression in <a href="https://github.com/pat-alt/ConformalPrediction.jl"><code>ConformalPrediction.jl</code></a>.</p>
-<p>Using the package, we can apply Split Conformal Prediction as follows:</p>
-<p>To be clear, all of the calibration steps (3 to 5) are post hoc, and yet none of them involved any changes to the model parameters. These are two important characteristics of Split Conformal Prediction (SCP) that make it particularly useful in the context of Algorithmic Recourse. Firstly, the fact that SCP involves posthoc calibration steps that happen after training, ensures that we need not place any restrictions on the black-box model itself. This stands in contrast to the approach proposed by <span class="citation" data-cites="schut2021generating">Schut et al. (<a href="#ref-schut2021generating" role="doc-biblioref">2021</a>)</span> in which they essentially restrict the class of models to Bayesian models. Secondly, the fact that the model itself is kept entirely intact ensures that the generated counterfactuals maintain fidelity to the model. Finally, note that we also have not resorted to a surrogate model to learn more about <span class="math inline">\(X \sim \mathcal{X}\)</span>. Instead, we have used the fitted model itself and a calibration data set to learn about the model’s predictive uncertainty.</p>
-</section>
-<section id="differentiable-cp" class="level2" data-number="1.3">
-<h2 data-number="1.3"><span class="header-section-number">1.3</span> Differentiable CP</h2>
-<p>In order to use CP in the context of gradient-based counterfactual search, we need it to be differentiable. <span class="citation" data-cites="stutz2022learning">Stutz et al. (<a href="#ref-stutz2022learning" role="doc-biblioref">2022</a>)</span> introduce a framework for training differentiable conformal predictors. They introduce a configurable loss function as well as smooth set size penalty.</p>
-<section id="smooth-set-size-penalty" class="level3" data-number="1.3.1">
-<h3 data-number="1.3.1"><span class="header-section-number">1.3.1</span> Smooth Set Size Penalty</h3>
-<p>Starting with the former, <span class="citation" data-cites="stutz2022learning">Stutz et al. (<a href="#ref-stutz2022learning" role="doc-biblioref">2022</a>)</span> propose the following:</p>
-<p><span id="eq-size-loss"><span class="math display">\[
-\Omega(C_{\theta}(x;\tau)) = = \max (0, \sum_k C_{\theta,k}(x;\tau) - \kappa)
-\tag{1.2}\]</span></span></p>
-<p>Here, <span class="math inline">\(C_{\theta,k}(x;\tau)\)</span> is loosely defined as the probability that class <span class="math inline">\(k\)</span> is assigned to the conformal prediction set <span class="math inline">\(C\)</span>. In the context of Conformal Training, this penalty reduces the <strong>inefficiency</strong> of the conformal predictor.</p>
-<p>In our context, we are not interested in improving the model itself, but rather in producing <strong>plausible</strong> counterfactuals. Provided that our counterfactual <span class="math inline">\(x^\prime\)</span> is already inside the target domain (<span class="math inline">\(\mathbb{I}_{y^\prime = t}=1\)</span>), penalizing <span class="math inline">\(\Omega(C_{\theta}(x;\tau))\)</span> corresponds to guiding counterfactuals into regions of the target domain that are characterized by low ambiguity: for <span class="math inline">\(\kappa=1\)</span> the conformal prediction set includes only the target label <span class="math inline">\(t\)</span> as <span class="math inline">\(\Omega(C_{\theta}(x;\tau))\)</span>. Arguably, less ambiguous counterfactuals are more <strong>plausible</strong>. Since the search is guided purely by properties of the model itself and (exchangeable) calibration data, counterfactuals also maintain <strong>high fidelity</strong>.</p>
-<p>The left panel of <a href="#fig-losses">Figure <span class="quarto-unresolved-ref">fig-losses</span></a> shows the smooth size penalty in the two-dimensional feature space of our synthetic data.</p>
-</section>
-<section id="configurable-classification-loss" class="level3" data-number="1.3.2">
-<h3 data-number="1.3.2"><span class="header-section-number">1.3.2</span> Configurable Classification Loss</h3>
-<p>The right panel of <a href="#fig-losses">Figure <span class="quarto-unresolved-ref">fig-losses</span></a> shows the configurable classification loss in the two-dimensional feature space of our synthetic data.</p>
-<div class="cell" data-execution_count="5">
-<div class="cell-output cell-output-display" data-execution_count="6">
-<div id="fig-losses" class="quarto-figure quarto-figure-center">
-<figure>
-<p><img src="intro_files/figure-html/fig-losses-output-1.svg" class="quarto-discovered-preview-image img-fluid" /></p>
-<p><figcaption>Figure 1.1: Illustration of the smooth size loss and the configurable classification loss.</figcaption></p>
-</figure>
-</div>
-</div>
-</div>
-</section>
-</section>
-<section id="fidelity-and-plausibility" class="level2" data-number="1.4">
-<h2 data-number="1.4"><span class="header-section-number">1.4</span> Fidelity and Plausibility</h2>
-<p>The main evaluation criteria we are interested in are <em>fidelity</em> and <em>plausibility</em>. Interestingly, we could also consider using these measures as penalties in the counterfactual search.</p>
-<section id="fidelity" class="level3" data-number="1.4.1">
-<h3 data-number="1.4.1"><span class="header-section-number">1.4.1</span> Fidelity</h3>
-<p>We propose to define fidelity as follows:</p>
-<div id="def-fidelity" class="theorem definition">
-<p><span class="theorem-title"><strong>Definition 1.1 (High-Fidelity Counterfactuals) </strong></span>Let <span class="math inline">\(\mathcal{X}_{\theta}|y = p_{\theta}(X|y)\)</span> denote the class-conditional distribution of <span class="math inline">\(X\)</span> defined by <span class="math inline">\(\theta\)</span>. Then for <span class="math inline">\(x^{\prime}\)</span> to be considered a high-fidelity counterfactual, we need: <span class="math inline">\(\mathcal{X}_{\theta}|t \approxeq \mathcal{X}^{\prime}\)</span> where <span class="math inline">\(t\)</span> denotes the target outcome.</p>
-</div>
-<p>We can generate samples from <span class="math inline">\(p_{\theta}(X|y)\)</span> following <span class="citation" data-cites="grathwohl2020your">Grathwohl et al. (<a href="#ref-grathwohl2020your" role="doc-biblioref">2020</a>)</span>. In <a href="#fig-energy">Figure <span class="quarto-unresolved-ref">fig-energy</span></a>, I have applied the methodology to our synthetic data.</p>
-<p>As an evaluation metric and penalty, we could use the average distance of the counterfactual <span class="math inline">\(x^{\prime}\)</span> from these generated samples, for example.</p>
-</section>
-<section id="plausibility" class="level3" data-number="1.4.2">
-<h3 data-number="1.4.2"><span class="header-section-number">1.4.2</span> Plausibility</h3>
-<p>We propose to define plausibility as follows:</p>
-<div id="def-plausible" class="theorem definition">
-<p><span class="theorem-title"><strong>Definition 1.2 (Plausible Counterfactuals) </strong></span>Formally, let <span class="math inline">\(\mathcal{X}|t\)</span> denote the conditional distribution of samples in the target class. As before, we have <span class="math inline">\(x^{\prime}\sim\mathcal{X}^{\prime}\)</span>, then for <span class="math inline">\(x^{\prime}\)</span> to be considered a plausible counterfactual, we need: <span class="math inline">\(\mathcal{X}|t \approxeq \mathcal{X}^{\prime}\)</span>.</p>
-</div>
-<p>As an evaluation metric and penalty, we could use the average distance of the counterfactual <span class="math inline">\(x^{\prime}\)</span> from (potentially bootstrapped) training samples in the target class, for example.</p>
-</section>
-</section>
-<section id="counterfactual-explanations" class="level2" data-number="1.5">
-<h2 data-number="1.5"><span class="header-section-number">1.5</span> Counterfactual Explanations</h2>
-<p>Next, let’s generate counterfactual explanations for our synthetic data. We first wrap our model in a container that makes it compatible with <code>CounterfactualExplanations.jl</code>. Then we draw a random sample, determine its predicted label <span class="math inline">\(\hat{y}\)</span> and choose the opposite label as our target.</p>
-<p>The generic Conformal Counterfactual Generator penalises the only the set size only:</p>
-<p><span id="eq-solution"><span class="math display">\[
-x^\prime = \arg \min_{x^\prime} \ell(M(x^\prime),t) + \lambda \mathbb{I}_{y^\prime = t} \Omega(C_{\theta}(x;\tau))
-\tag{1.3}\]</span></span></p>
-<div class="cell" data-execution_count="8">
-<div class="cell-output cell-output-display" data-execution_count="9">
-<div id="fig-ce" class="quarto-figure quarto-figure-center">
-<figure>
-<p><img src="intro_files/figure-html/fig-ce-output-1.svg" class="img-fluid" /></p>
-<p><figcaption>Figure 1.2: Comparison of counterfactuals produced using different generators.</figcaption></p>
-</figure>
-</div>
-</div>
-</div>
-<section id="benchmark" class="level3" data-number="1.5.1">
-<h3 data-number="1.5.1"><span class="header-section-number">1.5.1</span> Benchmark</h3>
-<p>Then we can simply loop over the datasets and eventually concatenate the results like so:</p>
-</section>
-</section>
-<section id="multi-class" class="level2" data-number="1.6">
-<h2 data-number="1.6"><span class="header-section-number">1.6</span> Multi-Class</h2>
-<div class="cell" data-execution_count="13">
-<div class="cell-output cell-output-display" data-execution_count="14">
-<div id="fig-losses-multi" class="quarto-figure quarto-figure-center">
-<figure>
-<p><img src="intro_files/figure-html/fig-losses-multi-output-1.svg" class="img-fluid" /></p>
-<p><figcaption>Figure 1.3: Illustration of the smooth size loss and the configurable classification loss.</figcaption></p>
-</figure>
-</div>
-</div>
-</div>
-<div id="quarto-navigation-envelope" class="hidden">
-<p><span class="hidden" data-render-id="quarto-int-sidebar-title">Conformal Counterfactual Explanations</span> <span class="hidden" data-render-id="quarto-int-navbar-title">Conformal Counterfactual Explanations</span> <span class="hidden" data-render-id="quarto-int-next">References</span> <span class="hidden" data-render-id="quarto-int-prev">Preface</span> <span class="hidden" data-render-id="quarto-int-sidebar:/index.html">Preface</span> <span class="hidden" data-render-id="quarto-int-sidebar:/notebooks/intro.html"><span class="chapter-number">1</span>Â <span class="chapter-title"><code>ConformalGenerator</code></span></span> <span class="hidden" data-render-id="quarto-int-sidebar:/notebooks/references.html">References</span></p>
-</div>
-<div id="quarto-meta-markdown" class="hidden">
-<p><span class="hidden" data-render-id="quarto-metatitle">Conformal Counterfactual Explanations - <span class="chapter-number">1</span>Â <span class="chapter-title"><code>ConformalGenerator</code></span></span> <span class="hidden" data-render-id="quarto-twittercardtitle">Conformal Counterfactual Explanations - <span class="chapter-number">1</span>Â <span class="chapter-title"><code>ConformalGenerator</code></span></span> <span class="hidden" data-render-id="quarto-ogcardtitle">Conformal Counterfactual Explanations - <span class="chapter-number">1</span>Â <span class="chapter-title"><code>ConformalGenerator</code></span></span> <span class="hidden" data-render-id="quarto-metasitename">Conformal Counterfactual Explanations</span> <span class="hidden" data-render-id="quarto-twittercarddesc"></span> <span class="hidden" data-render-id="quarto-ogcardddesc"></span></p>
-</div>
-<div id="refs" class="references csl-bib-body hanging-indent" role="list">
-<div id="ref-grathwohl2020your" class="csl-entry" role="listitem">
-Grathwohl, Will, Kuan-Chieh Wang, Joern-Henrik Jacobsen, David Duvenaud, Mohammad Norouzi, and Kevin Swersky. 2020. <span>“Your Classifier Is Secretly an Energy Based Model and You Should Treat It Like One.â€</span> In. <a href="https://openreview.net/forum?id=Hkxzx0NtDB">https://openreview.net/forum?id=Hkxzx0NtDB</a>.
-</div>
-<div id="ref-schut2021generating" class="csl-entry" role="listitem">
-Schut, Lisa, Oscar Key, Rory Mc Grath, Luca Costabello, Bogdan Sacaleanu, Yarin Gal, et al. 2021. <span>“Generating <span>Interpretable Counterfactual Explanations By Implicit Minimisation</span> of <span>Epistemic</span> and <span>Aleatoric Uncertainties</span>.â€</span> In <em>International <span>Conference</span> on <span>Artificial Intelligence</span> and <span>Statistics</span></em>, 1756–64. <span>PMLR</span>.
-</div>
-<div id="ref-stutz2022learning" class="csl-entry" role="listitem">
-Stutz, David, Krishnamurthy Dj Dvijotham, Ali Taylan Cemgil, and Arnaud Doucet. 2022. <span>“Learning <span>Optimal</span> <span>Conformal</span> <span>Classifiers</span>.â€</span> In. <a href="https://openreview.net/forum?id=t8O-4LKFVx">https://openreview.net/forum?id=t8O-4LKFVx</a>.
-</div>
-</div>
-</section>
-<section id="footnotes" class="footnotes footnotes-end-of-document" role="doc-endnotes">
-<hr />
-<ol>
-<li id="fn1"><p>In other places split conformal prediction is sometimes referred to as <em>inductive</em> conformal prediction.<a href="#fnref1" class="footnote-back" role="doc-backlink">↩︎</a></p></li>
-</ol>
-</section>
-
-</main> <!-- /main -->
-<script id = "quarto-html-after-body" type="application/javascript">
-window.document.addEventListener("DOMContentLoaded", function (event) {
- const toggleBodyColorMode = (bsSheetEl) => {
- const mode = bsSheetEl.getAttribute("data-mode");
- const bodyEl = window.document.querySelector("body");
- if (mode === "dark") {
- bodyEl.classList.add("quarto-dark");
- bodyEl.classList.remove("quarto-light");
- } else {
- bodyEl.classList.add("quarto-light");
- bodyEl.classList.remove("quarto-dark");
- }
- }
- const toggleBodyColorPrimary = () => {
- const bsSheetEl = window.document.querySelector("link#quarto-bootstrap");
- if (bsSheetEl) {
- toggleBodyColorMode(bsSheetEl);
- }
- }
- toggleBodyColorPrimary();
- const icon = "";
- const anchorJS = new window.AnchorJS();
- anchorJS.options = {
- placement: 'right',
- icon: icon
- };
- anchorJS.add('.anchored');
- const isCodeAnnotation = (el) => {
- for (const clz of el.classList) {
- if (clz.startsWith('code-annotation-')) {
- return true;
- }
- }
- return false;
- }
- const clipboard = new window.ClipboardJS('.code-copy-button', {
- text: function(trigger) {
- const codeEl = trigger.previousElementSibling.cloneNode(true);
- for (const childEl of codeEl.children) {
- if (isCodeAnnotation(childEl)) {
- childEl.remove();
- }
- }
- return codeEl.innerText;
- }
- });
- clipboard.on('success', function(e) {
- // button target
- const button = e.trigger;
- // don't keep focus
- button.blur();
- // flash "checked"
- button.classList.add('code-copy-button-checked');
- var currentTitle = button.getAttribute("title");
- button.setAttribute("title", "Copied!");
- let tooltip;
- if (window.bootstrap) {
- button.setAttribute("data-bs-toggle", "tooltip");
- button.setAttribute("data-bs-placement", "left");
- button.setAttribute("data-bs-title", "Copied!");
- tooltip = new bootstrap.Tooltip(button,
- { trigger: "manual",
- customClass: "code-copy-button-tooltip",
- offset: [0, -8]});
- tooltip.show();
- }
- setTimeout(function() {
- if (tooltip) {
- tooltip.hide();
- button.removeAttribute("data-bs-title");
- button.removeAttribute("data-bs-toggle");
- button.removeAttribute("data-bs-placement");
- }
- button.setAttribute("title", currentTitle);
- button.classList.remove('code-copy-button-checked');
- }, 1000);
- // clear code selection
- e.clearSelection();
- });
- function tippyHover(el, contentFn) {
- const config = {
- allowHTML: true,
- content: contentFn,
- maxWidth: 500,
- delay: 100,
- arrow: false,
- appendTo: function(el) {
- return el.parentElement;
- },
- interactive: true,
- interactiveBorder: 10,
- theme: 'quarto',
- placement: 'bottom-start'
- };
- window.tippy(el, config);
- }
- const noterefs = window.document.querySelectorAll('a[role="doc-noteref"]');
- for (var i=0; i<noterefs.length; i++) {
- const ref = noterefs[i];
- tippyHover(ref, function() {
- // use id or data attribute instead here
- let href = ref.getAttribute('data-footnote-href') || ref.getAttribute('href');
- try { href = new URL(href).hash; } catch {}
- const id = href.replace(/^#\/?/, "");
- const note = window.document.getElementById(id);
- return note.innerHTML;
- });
- }
- let selectedAnnoteEl;
- const selectorForAnnotation = ( cell, annotation) => {
- let cellAttr = 'data-code-cell="' + cell + '"';
- let lineAttr = 'data-code-annotation="' + annotation + '"';
- const selector = 'span[' + cellAttr + '][' + lineAttr + ']';
- return selector;
- }
- const selectCodeLines = (annoteEl) => {
- const doc = window.document;
- const targetCell = annoteEl.getAttribute("data-target-cell");
- const targetAnnotation = annoteEl.getAttribute("data-target-annotation");
- const annoteSpan = window.document.querySelector(selectorForAnnotation(targetCell, targetAnnotation));
- const lines = annoteSpan.getAttribute("data-code-lines").split(",");
- const lineIds = lines.map((line) => {
- return targetCell + "-" + line;
- })
- let top = null;
- let height = null;
- let parent = null;
- if (lineIds.length > 0) {
- //compute the position of the single el (top and bottom and make a div)
- const el = window.document.getElementById(lineIds[0]);
- top = el.offsetTop;
- height = el.offsetHeight;
- parent = el.parentElement.parentElement;
- if (lineIds.length > 1) {
- const lastEl = window.document.getElementById(lineIds[lineIds.length - 1]);
- const bottom = lastEl.offsetTop + lastEl.offsetHeight;
- height = bottom - top;
- }
- if (top !== null && height !== null && parent !== null) {
- // cook up a div (if necessary) and position it
- let div = window.document.getElementById("code-annotation-line-highlight");
- if (div === null) {
- div = window.document.createElement("div");
- div.setAttribute("id", "code-annotation-line-highlight");
- div.style.position = 'absolute';
- parent.appendChild(div);
- }
- div.style.top = top - 2 + "px";
- div.style.height = height + 4 + "px";
- let gutterDiv = window.document.getElementById("code-annotation-line-highlight-gutter");
- if (gutterDiv === null) {
- gutterDiv = window.document.createElement("div");
- gutterDiv.setAttribute("id", "code-annotation-line-highlight-gutter");
- gutterDiv.style.position = 'absolute';
- const codeCell = window.document.getElementById(targetCell);
- const gutter = codeCell.querySelector('.code-annotation-gutter');
- gutter.appendChild(gutterDiv);
- }
- gutterDiv.style.top = top - 2 + "px";
- gutterDiv.style.height = height + 4 + "px";
- }
- selectedAnnoteEl = annoteEl;
- }
- };
- const unselectCodeLines = () => {
- const elementsIds = ["code-annotation-line-highlight", "code-annotation-line-highlight-gutter"];
- elementsIds.forEach((elId) => {
- const div = window.document.getElementById(elId);
- if (div) {
- div.remove();
- }
- });
- selectedAnnoteEl = undefined;
- };
- // Attach click handler to the DT
- const annoteDls = window.document.querySelectorAll('dt[data-target-cell]');
- for (const annoteDlNode of annoteDls) {
- annoteDlNode.addEventListener('click', (event) => {
- const clickedEl = event.target;
- if (clickedEl !== selectedAnnoteEl) {
- unselectCodeLines();
- const activeEl = window.document.querySelector('dt[data-target-cell].code-annotation-active');
- if (activeEl) {
- activeEl.classList.remove('code-annotation-active');
- }
- selectCodeLines(clickedEl);
- clickedEl.classList.add('code-annotation-active');
- } else {
- // Unselect the line
- unselectCodeLines();
- clickedEl.classList.remove('code-annotation-active');
- }
- });
- }
- const findCites = (el) => {
- const parentEl = el.parentElement;
- if (parentEl) {
- const cites = parentEl.dataset.cites;
- if (cites) {
- return {
- el,
- cites: cites.split(' ')
- };
- } else {
- return findCites(el.parentElement)
- }
- } else {
- return undefined;
- }
- };
- var bibliorefs = window.document.querySelectorAll('a[role="doc-biblioref"]');
- for (var i=0; i<bibliorefs.length; i++) {
- const ref = bibliorefs[i];
- const citeInfo = findCites(ref);
- if (citeInfo) {
- tippyHover(citeInfo.el, function() {
- var popup = window.document.createElement('div');
- citeInfo.cites.forEach(function(cite) {
- var citeDiv = window.document.createElement('div');
- citeDiv.classList.add('hanging-indent');
- citeDiv.classList.add('csl-entry');
- var biblioDiv = window.document.getElementById('ref-' + cite);
- if (biblioDiv) {
- citeDiv.innerHTML = biblioDiv.innerHTML;
- }
- popup.appendChild(citeDiv);
- });
- return popup.innerHTML;
- });
- }
- }
-});
-</script>
-<nav class="page-navigation">
- <div class="nav-page nav-page-previous">
- <a href="/index.html" class="pagination-link">
- <i class="bi bi-arrow-left-short"></i> <span class="nav-page-text">Preface</span>
- </a>
- </div>
- <div class="nav-page nav-page-next">
- <a href="/notebooks/references.html" class="pagination-link">
- <span class="nav-page-text">References</span> <i class="bi bi-arrow-right-short"></i>
- </a>
- </div>
-</nav>
-</div> <!-- /content -->
-
-</body>
-
-</html>
\ No newline at end of file
diff --git a/notebooks/proposal.html b/notebooks/proposal.html
deleted file mode 100644
index b3bdcf7b5d29cb2434a6d3e723788b26ed6d0d73..0000000000000000000000000000000000000000
--- a/notebooks/proposal.html
+++ /dev/null
@@ -1,601 +0,0 @@
-<!DOCTYPE html>
-<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en">
-
-<head>
-
-<meta charset="utf-8" />
-<meta name="generator" content="quarto-99.9.9" />
-
-<meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes" />
-
-
-<title>Conformal Counterfactual Explanations – 1 High-Fidelity Counterfactual Explanations through Conformal Prediction</title>
-<style>
-code{white-space: pre-wrap;}
-span.smallcaps{font-variant: small-caps;}
-div.columns{display: flex; gap: min(4vw, 1.5em);}
-div.column{flex: auto; overflow-x: auto;}
-div.hanging-indent{margin-left: 1.5em; text-indent: -1.5em;}
-ul.task-list{list-style: none;}
-ul.task-list li input[type="checkbox"] {
- width: 0.8em;
- margin: 0 0.8em 0.2em -1em; /* quarto-specific, see https://github.com/quarto-dev/quarto-cli/issues/4556 */
- vertical-align: middle;
-}
-/* CSS for citations */
-div.csl-bib-body { }
-div.csl-entry {
- clear: both;
-}
-.hanging-indent div.csl-entry {
- margin-left:2em;
- text-indent:-2em;
-}
-div.csl-left-margin {
- min-width:2em;
- float:left;
-}
-div.csl-right-inline {
- margin-left:2em;
- padding-left:1em;
-}
-div.csl-indent {
- margin-left: 2em;
-}</style>
-
-<!-- htmldependencies:E3FAD763 -->
-<script id="quarto-search-options" type="application/json">{
- "location": "sidebar",
- "copy-button": false,
- "collapse-after": 3,
- "panel-placement": "start",
- "type": "textbox",
- "limit": 20,
- "language": {
- "search-no-results-text": "No results",
- "search-matching-documents-text": "matching documents",
- "search-copy-link-title": "Copy link to search",
- "search-hide-matches-text": "Hide additional matches",
- "search-more-match-text": "more match in this document",
- "search-more-matches-text": "more matches in this document",
- "search-clear-button-title": "Clear",
- "search-detached-cancel-button-title": "Cancel",
- "search-submit-button-title": "Submit"
- }
-}</script>
-
- <script src="https://polyfill.io/v3/polyfill.min.js?features=es6"></script>
- <script src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-chtml-full.js" type="text/javascript"></script>
-
-</head>
-
-<body>
-
-<div id="quarto-search-results"></div>
- <header id="quarto-header" class="headroom fixed-top">
- <nav class="quarto-secondary-nav">
- <div class="container-fluid d-flex">
- <button type="button" class="quarto-btn-toggle btn"
- data-bs-toggle="collapse" data-bs-target="#quarto-sidebar,#quarto-sidebar-glass"
- aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation"
- onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
- <i class="bi bi-layout-text-sidebar-reverse"></i>
- </button>
- <h1 class="quarto-secondary-nav-title"></h1>
- <a class="flex-grow-1" role="button" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar,#quarto-sidebar-glass"
- aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation"
- onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
- </a>
- <button type="button" class="btn quarto-search-button" aria-label="Search" onclick="window.quartoOpenSearch();">
- <i class="bi bi-search"></i>
- </button>
- </div>
- </nav>
-</header>
-<!-- content -->
-<div id="quarto-content" class="quarto-container page-columns page-rows-contents page-layout-article">
-<!-- sidebar -->
- <nav id="quarto-sidebar" class="sidebar collapse collapse-horizontal sidebar-navigation floating overflow-auto">
- <div class="pt-lg-2 mt-2 text-left sidebar-header">
- <div class="sidebar-title mb-0 py-0">
- <a href="/">
- Conformal Counterfactual Explanations
- </a>
- </div>
- </div>
- <div class="mt-2 flex-shrink-0 align-items-center">
- <div class="sidebar-search">
- <div id="quarto-search" class="" title="Search"></div>
- </div>
- </div>
- <div class="sidebar-menu-container">
- <ul class="list-unstyled mt-1">
- <li class="sidebar-item">
- <div class="sidebar-item-container">
- <a href="/index.html" class="sidebar-item-text sidebar-link">
- <span class="menu-text">Preface</span></a>
- </div>
-</li>
- <li class="sidebar-item">
- <div class="sidebar-item-container">
- <a href="/notebooks/proposal.html" class="sidebar-item-text sidebar-link active">
- <span class="menu-text"><span class='chapter-number'>1</span>Â <span class='chapter-title'>High-Fidelity Counterfactual Explanations through Conformal Prediction</span></span></a>
- </div>
-</li>
- <li class="sidebar-item">
- <div class="sidebar-item-container">
- <a href="/notebooks/intro.html" class="sidebar-item-text sidebar-link">
- <span class="menu-text"><span class='chapter-number'>2</span>Â <span class='chapter-title'>`ConformalGenerator`</span></span></a>
- </div>
-</li>
- <li class="sidebar-item">
- <div class="sidebar-item-container">
- <a href="/notebooks/references.html" class="sidebar-item-text sidebar-link">
- <span class="menu-text">References</span></a>
- </div>
-</li>
- </ul>
- </div>
-</nav>
-<div id="quarto-sidebar-glass" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar,#quarto-sidebar-glass" ></div>
-<!-- margin-sidebar -->
- <div id="quarto-margin-sidebar" class="sidebar margin-sidebar">
- <div id="quarto-toc-target"></div>
- </div>
-<!-- main -->
-<main class="content" id="quarto-document-content">
-
-<header id="title-block-header" class="quarto-title-block default">
-<div class="quarto-title">
-<h1 class="title"><span class="chapter-number">1</span>Â <span class="chapter-title">High-Fidelity Counterfactual Explanations through Conformal Prediction</span></h1>
-<p class="subtitle lead">Research Proposal</p>
-</div>
-
-
-
-<div class="quarto-title-meta">
-
-
-
-
- </div>
-
-<div>
- <div class="abstract">
- <div class="abstract-title">Abstract</div>
- <p>We propose Conformal Counterfactual Explanations: an effortless and rigorous way to produce realistic and faithful Counterfactual Explanations using Conformal Prediction. To address the need for realistic counterfactuals, existing work has primarily relied on separate generative models to learn the data-generating process. While this is an effective way to produce plausible and model-agnostic counterfactual explanations, it not only introduces a significant engineering overhead but also reallocates the task of creating realistic model explanations from the model itself to the generative model. Recent work has shown that there is no need for any of this when working with probabilistic models that explicitly quantify their own uncertainty. Unfortunately, most models used in practice still do not fulfil that basic requirement, in which case we would like to have a way to quantify predictive uncertainty in a post-hoc fashion.</p>
- </div>
-</div>
-
-</header>
-<nav id="TOC" role="doc-toc">
- <h2 id="toc-title">Table of contents</h2>
-
- <ul>
- <li><a href="#motivation" id="toc-motivation"><span class="header-section-number">1.1</span> Motivation</a>
- <ul>
- <li><a href="#counterfactual-explanations-or-adversarial-examples" id="toc-counterfactual-explanations-or-adversarial-examples"><span class="header-section-number">1.1.1</span> Counterfactual Explanations or Adversarial Examples?</a></li>
- <li><a href="#sec-fidelity" id="toc-sec-fidelity"><span class="header-section-number">1.1.2</span> From Plausible to High-Fidelity Counterfactuals</a></li>
- </ul></li>
- <li><a href="#conformal-counterfactual-explanations" id="toc-conformal-counterfactual-explanations"><span class="header-section-number">1.2</span> Conformal Counterfactual Explanations</a>
- <ul>
- <li><a href="#minimizing-predictive-uncertainty" id="toc-minimizing-predictive-uncertainty"><span class="header-section-number">1.2.1</span> Minimizing Predictive Uncertainty</a></li>
- <li><a href="#background-on-conformal-prediction" id="toc-background-on-conformal-prediction"><span class="header-section-number">1.2.2</span> Background on Conformal Prediction</a></li>
- <li><a href="#generating-conformal-counterfactuals" id="toc-generating-conformal-counterfactuals"><span class="header-section-number">1.2.3</span> Generating Conformal Counterfactuals</a></li>
- </ul></li>
- <li><a href="#experiments" id="toc-experiments"><span class="header-section-number">1.3</span> Experiments</a>
- <ul>
- <li><a href="#research-questions" id="toc-research-questions"><span class="header-section-number">1.3.1</span> Research Questions</a></li>
- </ul></li>
- <li><a href="#references" id="toc-references"><span class="header-section-number">1.4</span> References</a></li>
- </ul>
-</nav>
-<section id="motivation" class="level2" data-number="1.1">
-<h2 data-number="1.1"><span class="header-section-number">1.1</span> Motivation</h2>
-<p>Counterfactual Explanations are a powerful, flexible and intuitive way to not only explain black-box models but also enable affected individuals to challenge them through the means of Algorithmic Recourse.</p>
-<section id="counterfactual-explanations-or-adversarial-examples" class="level3" data-number="1.1.1">
-<h3 data-number="1.1.1"><span class="header-section-number">1.1.1</span> Counterfactual Explanations or Adversarial Examples?</h3>
-<p>Most state-of-the-art approaches to generating Counterfactual Explanations (CE) rely on gradient descent in the feature space. The key idea is to perturb inputs <span class="math inline">\(x\in\mathcal{X}\)</span> into a black-box model <span class="math inline">\(f: \mathcal{X} \mapsto \mathcal{Y}\)</span> in order to change the model output <span class="math inline">\(f(x)\)</span> to some pre-specified target value <span class="math inline">\(t\in\mathcal{Y}\)</span>. Formally, this boils down to defining some loss function <span class="math inline">\(\ell(f(x),t)\)</span> and taking gradient steps in the minimizing direction. The so-generated counterfactuals are considered valid as soon as the predicted label matches the target label. A stripped-down counterfactual explanation is therefore little different from an adversarial example. In <a href="#fig-adv">Figure <span class="quarto-unresolved-ref">fig-adv</span></a>, for example, generic counterfactual search as in <span class="citation" data-cites="wachter2017counterfactual">Wachter, Mittelstadt, and Russell (<a href="#ref-wachter2017counterfactual" role="doc-biblioref">2017</a>)</span> has been applied to MNIST data.</p>
-<div id="fig-adv" class="quarto-figure quarto-figure-center">
-<figure>
-<p><img src="www/you_may_not_like_it.png" class="quarto-discovered-preview-image img-fluid" /></p>
-<p><figcaption>Figure 1.1: You may not like it, but this is what stripped-down counterfactuals look like. Here we have used <span class="citation" data-cites="wachter2017counterfactual">Wachter, Mittelstadt, and Russell (<a href="#ref-wachter2017counterfactual" role="doc-biblioref">2017</a>)</span> to generate multiple counterfactuals for turning an 8 (eight) into a 3 (three).</figcaption></p>
-</figure>
-</div>
-<p>The crucial difference between adversarial examples and counterfactuals is one of intent. While adversarial examples are typically intended to go unnoticed, counterfactuals in the context of Explainable AI are generally sought to be “plausibleâ€, “realistic†or “feasibleâ€. To fulfil this latter goal, researchers have come up with a myriad of ways. <span class="citation" data-cites="joshi2019realistic">Joshi et al. (<a href="#ref-joshi2019realistic" role="doc-biblioref">2019</a>)</span> were among the first to suggest that instead of searching counterfactuals in the feature space, we can instead traverse a latent embedding learned by a surrogate generative model. Similarly, <span class="citation" data-cites="poyiadzi2020face">Poyiadzi et al. (<a href="#ref-poyiadzi2020face" role="doc-biblioref">2020</a>)</span> use density … Finally, <span class="citation" data-cites="karimi2021algorithmic">Karimi, Schölkopf, and Valera (<a href="#ref-karimi2021algorithmic" role="doc-biblioref">2021</a>)</span> argues that counterfactuals should comply with the causal model that generates them [CHECK IF WE CAN PHASE THIS LIKE THIS]. Other related approaches include … All of these different approaches have a common goal: they aim to ensure that the generated counterfactuals comply with the (learned) data-generating process (DGB).</p>
-<div id="def-plausible" class="theorem definition">
-<p><span class="theorem-title"><strong>Definition 1.1 (Plausible Counterfactuals) </strong></span>Formally, if <span class="math inline">\(x \sim \mathcal{X}\)</span> and for the corresponding counterfactual we have <span class="math inline">\(x^{\prime}\sim\mathcal{X}^{\prime}\)</span>, then for <span class="math inline">\(x^{\prime}\)</span> to be considered a plausible counterfactual, we need: <span class="math inline">\(\mathcal{X} \approxeq \mathcal{X}^{\prime}\)</span>.</p>
-</div>
-<p>In the context of Algorithmic Recourse, it makes sense to strive for plausible counterfactuals, since anything else would essentially require individuals to move to out-of-distribution states. But it is worth noting that our ambition to meet this goal, may have implications on our ability to faithfully explain the behaviour of the underlying black-box model (arguably our principal goal). By essentially decoupling the task of learning plausible representations of the data from the model itself, we open ourselves up to vulnerabilities. Using a separate generative model to learn <span class="math inline">\(\mathcal{X}\)</span>, for example, has very serious implications for the generated counterfactuals. <a href="#fig-latent">Figure <span class="quarto-unresolved-ref">fig-latent</span></a> compares the results of applying REVISE <span class="citation" data-cites="joshi2019realistic">(<a href="#ref-joshi2019realistic" role="doc-biblioref">Joshi et al. 2019</a>)</span> to MNIST data using two different Variational Auto-Encoders: while the counterfactual generated using an expressive (strong) VAE is compelling, the result relying on a less expressive (weak) VAE is not even valid. In this latter case, the decoder step of the VAE fails to yield values in <span class="math inline">\(\mathcal{X}\)</span> and hence the counterfactual search in the learned latent space is doomed.</p>
-<div id="fig-latent" class="quarto-figure quarto-figure-center">
-<figure>
-<p><img src="www/mnist_9to4_latent.png" class="img-fluid" /></p>
-<p><figcaption>Figure 1.2: Counterfactual explanations for MNIST using a Latent Space generator: turning a nine (9) into a four (4).</figcaption></p>
-</figure>
-</div>
-<blockquote>
-<p>Here it would be nice to have another example where we poison the data going into the generative model to hide biases present in the data (e.g. Boston housing).</p>
-</blockquote>
-<ul>
-<li>Latent can be manipulated:
-<ul>
-<li>train biased model</li>
-<li>train VAE with biased variable removed/attacked (use Boston housing dataset)</li>
-<li>hypothesis: will generate bias-free explanations</li>
-</ul></li>
-</ul>
-</section>
-<section id="sec-fidelity" class="level3" data-number="1.1.2">
-<h3 data-number="1.1.2"><span class="header-section-number">1.1.2</span> From Plausible to High-Fidelity Counterfactuals</h3>
-<p>In light of the findings, we propose to generally avoid using surrogate models to learn <span class="math inline">\(\mathcal{X}\)</span> in the context of Counterfactual Explanations.</p>
-<div id="prp-surrogate" class="theorem proposition">
-<p><span class="theorem-title"><strong>Proposition 1.1 (Avoid Surrogates) </strong></span>Since we are in the business of explaining a black-box model, the task of learning realistic representations of the data should not be reallocated from the model itself to some surrogate model.</p>
-</div>
-<p>In cases where the use of surrogate models cannot be avoided, we propose to weigh the plausibility of counterfactuals against their fidelity to the black-box model. In the context of Explainable AI, fidelity is defined as describing how an explanation approximates the prediction of the black-box model <span class="citation" data-cites="molnar2020interpretable">(<a href="#ref-molnar2020interpretable" role="doc-biblioref">Molnar 2020</a>)</span>. Fidelity has become the default metric for evaluating Local Model-Agnostic Models, since they often involve local surrogate models whose predictions need not always match those of the black-box model.</p>
-<p>In the case of Counterfactual Explanations, the concept of fidelity has so far been ignored. This is not altogether surprising, since by construction and design, Counterfactual Explanations work with the predictions of the black-box model directly: as stated above, a counterfactual <span class="math inline">\(x^{\prime}\)</span> is considered valid if and only if <span class="math inline">\(f(x^{\prime})=t\)</span>, where <span class="math inline">\(t\)</span> denote some target outcome.</p>
-<p>Does fidelity even make sense in the context of CE, and if so, how can we define it? In light of the examples in the previous section, we think it is urgent to introduce a notion of fidelity in this context, that relates to the distributional properties of the generated counterfactuals. In particular, we propose that a high-fidelity counterfactual <span class="math inline">\(x^{\prime}\)</span> complies with the class-conditional distribution <span class="math inline">\(\mathcal{X}_{\theta} = p_{\theta}(X|y)\)</span> where <span class="math inline">\(\theta\)</span> denote the black-box model parameters.</p>
-<div id="def-fidele" class="theorem definition">
-<p><span class="theorem-title"><strong>Definition 1.2 (High-Fidelity Counterfactuals) </strong></span>Let <span class="math inline">\(\mathcal{X}_{\theta}|y = p_{\theta}(X|y)\)</span> denote the class-conditional distribution of <span class="math inline">\(X\)</span> defined by <span class="math inline">\(\theta\)</span>. Then for <span class="math inline">\(x^{\prime}\)</span> to be considered a high-fidelity counterfactual, we need: <span class="math inline">\(\mathcal{X}_{\theta}|t \approxeq \mathcal{X}^{\prime}\)</span> where <span class="math inline">\(t\)</span> denotes the target outcome.</p>
-</div>
-<p>In order to assess the fidelity of counterfactuals, we propose the following two-step procedure:</p>
-<ol type="1">
-<li>Generate samples <span class="math inline">\(X_{\theta}|y\)</span> and <span class="math inline">\(X^{\prime}\)</span> from <span class="math inline">\(\mathcal{X}_{\theta}|t\)</span> and <span class="math inline">\(\mathcal{X}^{\prime}\)</span>, respectively.</li>
-<li>Compute the Maximum Mean Discrepancy (MMD) between <span class="math inline">\(X_{\theta}|y\)</span> and <span class="math inline">\(X^{\prime}\)</span>.</li>
-</ol>
-<p>If the computed value is different from zero, we can reject the null-hypothesis of fidelity.</p>
-<blockquote>
-<p>Two challenges here: 1) implementing the sampling procedure in <span class="citation" data-cites="grathwohl2020your">Grathwohl et al. (<a href="#ref-grathwohl2020your" role="doc-biblioref">2020</a>)</span>; 2) it is unclear if MMD is really the right way to measure this.</p>
-</blockquote>
-</section>
-</section>
-<section id="conformal-counterfactual-explanations" class="level2" data-number="1.2">
-<h2 data-number="1.2"><span class="header-section-number">1.2</span> Conformal Counterfactual Explanations</h2>
-<p>In <a href="#sec-fidelity"><span class="quarto-unresolved-ref">sec-fidelity</span></a>, we have advocated for avoiding surrogate models in the context of Counterfactual Explanations. In this section, we introduce an alternative way to generate high-fidelity Counterfactual Explanations. In particular, we propose Conformal Counterfactual Explanations (CCE), that is Counterfactual Explanations that minimize the predictive uncertainty of conformal models.</p>
-<section id="minimizing-predictive-uncertainty" class="level3" data-number="1.2.1">
-<h3 data-number="1.2.1"><span class="header-section-number">1.2.1</span> Minimizing Predictive Uncertainty</h3>
-<p><span class="citation" data-cites="schut2021generating">Schut et al. (<a href="#ref-schut2021generating" role="doc-biblioref">2021</a>)</span> demonstrated that the goal of generating realistic (plausible) counterfactuals can also be achieved by seeking counterfactuals that minimize the predictive uncertainty of the underlying black-box model. Similarly, <span class="citation" data-cites="antoran2020getting">Antorán et al. (<a href="#ref-antoran2020getting" role="doc-biblioref">2020</a>)</span> …</p>
-<ul>
-<li>Problem: restricted to Bayesian models.</li>
-<li>Solution: post-hoc predictive uncertainty quantification. In particular, Conformal Prediction.</li>
-</ul>
-</section>
-<section id="background-on-conformal-prediction" class="level3" data-number="1.2.2">
-<h3 data-number="1.2.2"><span class="header-section-number">1.2.2</span> Background on Conformal Prediction</h3>
-<ul>
-<li>Distribution-free, model-agnostic and scalable approach to predictive uncertainty quantification.</li>
-<li>Conformal prediction is instance-based. So is CE.</li>
-<li>Take any fitted model and turn it into a conformal model using calibration data.</li>
-<li>Our approach, therefore, relaxes the restriction on the family of black-box models, at the cost of relying on a subset of the data. Arguably, data is often abundant and in most applications practitioners tend to hold out a test data set anyway.</li>
-</ul>
-<blockquote>
-<p>Does the coverage guarantee carry over to counterfactuals?</p>
-</blockquote>
-</section>
-<section id="generating-conformal-counterfactuals" class="level3" data-number="1.2.3">
-<h3 data-number="1.2.3"><span class="header-section-number">1.2.3</span> Generating Conformal Counterfactuals</h3>
-<p>While Conformal Prediction has recently grown in popularity, it does introduce a challenge in the context of classification: the predictions of Conformal Classifiers are set-valued and therefore difficult to work with, since they are, for example, non-differentiable. Fortunately, <span class="citation" data-cites="stutz2022learning">Stutz et al. (<a href="#ref-stutz2022learning" role="doc-biblioref">2022</a>)</span> introduced carefully designed differentiable loss functions that make it possible to evaluate the performance of conformal predictions in training. We can leverage these recent advances in the context of gradient-based counterfactual search …</p>
-<blockquote>
-<p>Challenge: still need to implement these loss functions.</p>
-</blockquote>
-</section>
-</section>
-<section id="experiments" class="level2" data-number="1.3">
-<h2 data-number="1.3"><span class="header-section-number">1.3</span> Experiments</h2>
-<section id="research-questions" class="level3" data-number="1.3.1">
-<h3 data-number="1.3.1"><span class="header-section-number">1.3.1</span> Research Questions</h3>
-<ul>
-<li><p>Is CP alone enough to ensure realistic counterfactuals?</p></li>
-<li><p>Do counterfactuals improve further as the models get better?</p></li>
-<li><p>Do counterfactuals get more realistic as coverage</p></li>
-<li><p>What happens as we vary coverage and setsize?</p></li>
-<li><p>What happens as we improve the model robustness?</p></li>
-<li><p>What happens as we improve the model’s ability to incorporate predictive uncertainty (deep ensemble, laplace)?</p></li>
-<li><p>What happens if we combine with DiCE, ClaPROAR, Gravitational?</p></li>
-<li><p>What about CE robustness to endogenous shifts <span class="citation" data-cites="altmeyer2023endogenous">(<a href="#ref-altmeyer2023endogenous" role="doc-biblioref">Altmeyer et al. 2023</a>)</span>?</p></li>
-<li><p>Benchmarking:</p>
-<ul>
-<li>add PROBE <span class="citation" data-cites="pawelczyk2022probabilistically">(<a href="#ref-pawelczyk2022probabilistically" role="doc-biblioref">Pawelczyk et al. 2022</a>)</span> into the mix.</li>
-<li>compare travel costs to domain shits.</li>
-</ul></li>
-</ul>
-<blockquote>
-<p>Nice to have: What about using Laplace Approximation, then Conformal Prediction? What about using Conformalised Laplace?</p>
-</blockquote>
-</section>
-</section>
-<section id="references" class="level2" data-number="1.4">
-<h2 data-number="1.4"><span class="header-section-number">1.4</span> References</h2>
-<div id="quarto-navigation-envelope" class="hidden">
-<p><span class="hidden" data-render-id="quarto-int-sidebar-title">Conformal Counterfactual Explanations</span> <span class="hidden" data-render-id="quarto-int-navbar-title">Conformal Counterfactual Explanations</span> <span class="hidden" data-render-id="quarto-int-next"><span class="chapter-number">2</span>Â <span class="chapter-title"><code>ConformalGenerator</code></span></span> <span class="hidden" data-render-id="quarto-int-prev">Preface</span> <span class="hidden" data-render-id="quarto-int-sidebar:/index.html">Preface</span> <span class="hidden" data-render-id="quarto-int-sidebar:/notebooks/proposal.html"><span class="chapter-number">1</span>Â <span class="chapter-title">High-Fidelity Counterfactual Explanations through Conformal Prediction</span></span> <span class="hidden" data-render-id="quarto-int-sidebar:/notebooks/intro.html"><span class="chapter-number">2</span>Â <span class="chapter-title"><code>ConformalGenerator</code></span></span> <span class="hidden" data-render-id="quarto-int-sidebar:/notebooks/references.html">References</span></p>
-</div>
-<div id="quarto-meta-markdown" class="hidden">
-<p><span class="hidden" data-render-id="quarto-metatitle">Conformal Counterfactual Explanations - <span class="chapter-number">1</span>Â <span class="chapter-title">High-Fidelity Counterfactual Explanations through Conformal Prediction</span></span> <span class="hidden" data-render-id="quarto-twittercardtitle">Conformal Counterfactual Explanations - <span class="chapter-number">1</span>Â <span class="chapter-title">High-Fidelity Counterfactual Explanations through Conformal Prediction</span></span> <span class="hidden" data-render-id="quarto-ogcardtitle">Conformal Counterfactual Explanations - <span class="chapter-number">1</span>Â <span class="chapter-title">High-Fidelity Counterfactual Explanations through Conformal Prediction</span></span> <span class="hidden" data-render-id="quarto-metasitename">Conformal Counterfactual Explanations</span> <span class="hidden" data-render-id="quarto-twittercarddesc">We propose Conformal Counterfactual Explanations: an effortless and rigorous way to produce realistic and faithful Counterfactual Explanations using Conformal Prediction. To address the need for realistic counterfactuals, existing work has primarily relied on separate generative models to learn the data-generating process. While this is an effective way to produce plausible and model-agnostic counterfactual explanations, it not only introduces a significant engineering overhead but also reallocates the task of creating realistic model explanations from the model itself to the generative model. Recent work has shown that there is no need for any of this when working with probabilistic models that explicitly quantify their own uncertainty. Unfortunately, most models used in practice still do not fulfil that basic requirement, in which case we would like to have a way to quantify predictive uncertainty in a post-hoc fashion.</span> <span class="hidden" data-render-id="quarto-ogcardddesc">We propose Conformal Counterfactual Explanations: an effortless and rigorous way to produce realistic and faithful Counterfactual Explanations using Conformal Prediction. To address the need for realistic counterfactuals, existing work has primarily relied on separate generative models to learn the data-generating process. While this is an effective way to produce plausible and model-agnostic counterfactual explanations, it not only introduces a significant engineering overhead but also reallocates the task of creating realistic model explanations from the model itself to the generative model. Recent work has shown that there is no need for any of this when working with probabilistic models that explicitly quantify their own uncertainty. Unfortunately, most models used in practice still do not fulfil that basic requirement, in which case we would like to have a way to quantify predictive uncertainty in a post-hoc fashion.</span></p>
-</div>
-<div id="refs" class="references csl-bib-body hanging-indent" role="list">
-<div id="ref-altmeyer2023endogenous" class="csl-entry" role="listitem">
-Altmeyer, Patrick, Giovan Angela, Aleksander Buszydlik, Karol Dobiczek, Arie van Deursen, and Cynthia Liem. 2023. <span>“Endogenous <span>Macrodynamics</span> in <span>Algorithmic</span> <span>Recourse</span>.â€</span> In <em>First <span>IEEE</span> <span>Conference</span> on <span>Secure</span> and <span>Trustworthy</span> <span>Machine</span> <span>Learning</span></em>.
-</div>
-<div id="ref-antoran2020getting" class="csl-entry" role="listitem">
-Antorán, Javier, Umang Bhatt, Tameem Adel, Adrian Weller, and José Miguel Hernández-Lobato. 2020. <span>“Getting a Clue: <span>A</span> Method for Explaining Uncertainty Estimates.â€</span> <a href="https://arxiv.org/abs/2006.06848">https://arxiv.org/abs/2006.06848</a>.
-</div>
-<div id="ref-grathwohl2020your" class="csl-entry" role="listitem">
-Grathwohl, Will, Kuan-Chieh Wang, Joern-Henrik Jacobsen, David Duvenaud, Mohammad Norouzi, and Kevin Swersky. 2020. <span>“Your Classifier Is Secretly an Energy Based Model and You Should Treat It Like One.â€</span> In. <a href="https://openreview.net/forum?id=Hkxzx0NtDB">https://openreview.net/forum?id=Hkxzx0NtDB</a>.
-</div>
-<div id="ref-joshi2019realistic" class="csl-entry" role="listitem">
-Joshi, Shalmali, Oluwasanmi Koyejo, Warut Vijitbenjaronk, Been Kim, and Joydeep Ghosh. 2019. <span>“Towards Realistic Individual Recourse and Actionable Explanations in Black-Box Decision Making Systems.â€</span> <a href="https://arxiv.org/abs/1907.09615">https://arxiv.org/abs/1907.09615</a>.
-</div>
-<div id="ref-karimi2021algorithmic" class="csl-entry" role="listitem">
-Karimi, Amir-Hossein, Bernhard Schölkopf, and Isabel Valera. 2021. <span>“Algorithmic Recourse: From Counterfactual Explanations to Interventions.â€</span> In <em>Proceedings of the 2021 <span>ACM Conference</span> on <span>Fairness</span>, <span>Accountability</span>, and <span>Transparency</span></em>, 353–62.
-</div>
-<div id="ref-molnar2020interpretable" class="csl-entry" role="listitem">
-Molnar, Christoph. 2020. <em>Interpretable Machine Learning</em>. <span>Lulu. com</span>.
-</div>
-<div id="ref-pawelczyk2022probabilistically" class="csl-entry" role="listitem">
-Pawelczyk, Martin, Teresa Datta, Johannes van-den-Heuvel, Gjergji Kasneci, and Himabindu Lakkaraju. 2022. <span>“Probabilistically <span>Robust</span> <span>Recourse</span>: <span>Navigating</span> the <span>Trade</span>-Offs Between <span>Costs</span> and <span>Robustness</span> in <span>Algorithmic</span> <span>Recourse</span>.â€</span> <em>arXiv Preprint arXiv:2203.06768</em>.
-</div>
-<div id="ref-poyiadzi2020face" class="csl-entry" role="listitem">
-Poyiadzi, Rafael, Kacper Sokol, Raul Santos-Rodriguez, Tijl De Bie, and Peter Flach. 2020. <span>“<span>FACE</span>: <span>Feasible</span> and Actionable Counterfactual Explanations.â€</span> In <em>Proceedings of the <span>AAAI</span>/<span>ACM Conference</span> on <span>AI</span>, <span>Ethics</span>, and <span>Society</span></em>, 344–50.
-</div>
-<div id="ref-schut2021generating" class="csl-entry" role="listitem">
-Schut, Lisa, Oscar Key, Rory Mc Grath, Luca Costabello, Bogdan Sacaleanu, Yarin Gal, et al. 2021. <span>“Generating <span>Interpretable Counterfactual Explanations By Implicit Minimisation</span> of <span>Epistemic</span> and <span>Aleatoric Uncertainties</span>.â€</span> In <em>International <span>Conference</span> on <span>Artificial Intelligence</span> and <span>Statistics</span></em>, 1756–64. <span>PMLR</span>.
-</div>
-<div id="ref-stutz2022learning" class="csl-entry" role="listitem">
-Stutz, David, Krishnamurthy Dj Dvijotham, Ali Taylan Cemgil, and Arnaud Doucet. 2022. <span>“Learning <span>Optimal</span> <span>Conformal</span> <span>Classifiers</span>.â€</span> In. <a href="https://openreview.net/forum?id=t8O-4LKFVx">https://openreview.net/forum?id=t8O-4LKFVx</a>.
-</div>
-<div id="ref-wachter2017counterfactual" class="csl-entry" role="listitem">
-Wachter, Sandra, Brent Mittelstadt, and Chris Russell. 2017. <span>“Counterfactual Explanations Without Opening the Black Box: <span>Automated</span> Decisions and the <span>GDPR</span>.â€</span> <em>Harv. JL & Tech.</em> 31: 841.
-</div>
-</div>
-</section>
-
-</main> <!-- /main -->
-<script id = "quarto-html-after-body" type="application/javascript">
-window.document.addEventListener("DOMContentLoaded", function (event) {
- const toggleBodyColorMode = (bsSheetEl) => {
- const mode = bsSheetEl.getAttribute("data-mode");
- const bodyEl = window.document.querySelector("body");
- if (mode === "dark") {
- bodyEl.classList.add("quarto-dark");
- bodyEl.classList.remove("quarto-light");
- } else {
- bodyEl.classList.add("quarto-light");
- bodyEl.classList.remove("quarto-dark");
- }
- }
- const toggleBodyColorPrimary = () => {
- const bsSheetEl = window.document.querySelector("link#quarto-bootstrap");
- if (bsSheetEl) {
- toggleBodyColorMode(bsSheetEl);
- }
- }
- toggleBodyColorPrimary();
- const icon = "";
- const anchorJS = new window.AnchorJS();
- anchorJS.options = {
- placement: 'right',
- icon: icon
- };
- anchorJS.add('.anchored');
- const isCodeAnnotation = (el) => {
- for (const clz of el.classList) {
- if (clz.startsWith('code-annotation-')) {
- return true;
- }
- }
- return false;
- }
- const clipboard = new window.ClipboardJS('.code-copy-button', {
- text: function(trigger) {
- const codeEl = trigger.previousElementSibling.cloneNode(true);
- for (const childEl of codeEl.children) {
- if (isCodeAnnotation(childEl)) {
- childEl.remove();
- }
- }
- return codeEl.innerText;
- }
- });
- clipboard.on('success', function(e) {
- // button target
- const button = e.trigger;
- // don't keep focus
- button.blur();
- // flash "checked"
- button.classList.add('code-copy-button-checked');
- var currentTitle = button.getAttribute("title");
- button.setAttribute("title", "Copied!");
- let tooltip;
- if (window.bootstrap) {
- button.setAttribute("data-bs-toggle", "tooltip");
- button.setAttribute("data-bs-placement", "left");
- button.setAttribute("data-bs-title", "Copied!");
- tooltip = new bootstrap.Tooltip(button,
- { trigger: "manual",
- customClass: "code-copy-button-tooltip",
- offset: [0, -8]});
- tooltip.show();
- }
- setTimeout(function() {
- if (tooltip) {
- tooltip.hide();
- button.removeAttribute("data-bs-title");
- button.removeAttribute("data-bs-toggle");
- button.removeAttribute("data-bs-placement");
- }
- button.setAttribute("title", currentTitle);
- button.classList.remove('code-copy-button-checked');
- }, 1000);
- // clear code selection
- e.clearSelection();
- });
- function tippyHover(el, contentFn) {
- const config = {
- allowHTML: true,
- content: contentFn,
- maxWidth: 500,
- delay: 100,
- arrow: false,
- appendTo: function(el) {
- return el.parentElement;
- },
- interactive: true,
- interactiveBorder: 10,
- theme: 'quarto',
- placement: 'bottom-start'
- };
- window.tippy(el, config);
- }
- const noterefs = window.document.querySelectorAll('a[role="doc-noteref"]');
- for (var i=0; i<noterefs.length; i++) {
- const ref = noterefs[i];
- tippyHover(ref, function() {
- // use id or data attribute instead here
- let href = ref.getAttribute('data-footnote-href') || ref.getAttribute('href');
- try { href = new URL(href).hash; } catch {}
- const id = href.replace(/^#\/?/, "");
- const note = window.document.getElementById(id);
- return note.innerHTML;
- });
- }
- let selectedAnnoteEl;
- const selectorForAnnotation = ( cell, annotation) => {
- let cellAttr = 'data-code-cell="' + cell + '"';
- let lineAttr = 'data-code-annotation="' + annotation + '"';
- const selector = 'span[' + cellAttr + '][' + lineAttr + ']';
- return selector;
- }
- const selectCodeLines = (annoteEl) => {
- const doc = window.document;
- const targetCell = annoteEl.getAttribute("data-target-cell");
- const targetAnnotation = annoteEl.getAttribute("data-target-annotation");
- const annoteSpan = window.document.querySelector(selectorForAnnotation(targetCell, targetAnnotation));
- const lines = annoteSpan.getAttribute("data-code-lines").split(",");
- const lineIds = lines.map((line) => {
- return targetCell + "-" + line;
- })
- let top = null;
- let height = null;
- let parent = null;
- if (lineIds.length > 0) {
- //compute the position of the single el (top and bottom and make a div)
- const el = window.document.getElementById(lineIds[0]);
- top = el.offsetTop;
- height = el.offsetHeight;
- parent = el.parentElement.parentElement;
- if (lineIds.length > 1) {
- const lastEl = window.document.getElementById(lineIds[lineIds.length - 1]);
- const bottom = lastEl.offsetTop + lastEl.offsetHeight;
- height = bottom - top;
- }
- if (top !== null && height !== null && parent !== null) {
- // cook up a div (if necessary) and position it
- let div = window.document.getElementById("code-annotation-line-highlight");
- if (div === null) {
- div = window.document.createElement("div");
- div.setAttribute("id", "code-annotation-line-highlight");
- div.style.position = 'absolute';
- parent.appendChild(div);
- }
- div.style.top = top - 2 + "px";
- div.style.height = height + 4 + "px";
- let gutterDiv = window.document.getElementById("code-annotation-line-highlight-gutter");
- if (gutterDiv === null) {
- gutterDiv = window.document.createElement("div");
- gutterDiv.setAttribute("id", "code-annotation-line-highlight-gutter");
- gutterDiv.style.position = 'absolute';
- const codeCell = window.document.getElementById(targetCell);
- const gutter = codeCell.querySelector('.code-annotation-gutter');
- gutter.appendChild(gutterDiv);
- }
- gutterDiv.style.top = top - 2 + "px";
- gutterDiv.style.height = height + 4 + "px";
- }
- selectedAnnoteEl = annoteEl;
- }
- };
- const unselectCodeLines = () => {
- const elementsIds = ["code-annotation-line-highlight", "code-annotation-line-highlight-gutter"];
- elementsIds.forEach((elId) => {
- const div = window.document.getElementById(elId);
- if (div) {
- div.remove();
- }
- });
- selectedAnnoteEl = undefined;
- };
- // Attach click handler to the DT
- const annoteDls = window.document.querySelectorAll('dt[data-target-cell]');
- for (const annoteDlNode of annoteDls) {
- annoteDlNode.addEventListener('click', (event) => {
- const clickedEl = event.target;
- if (clickedEl !== selectedAnnoteEl) {
- unselectCodeLines();
- const activeEl = window.document.querySelector('dt[data-target-cell].code-annotation-active');
- if (activeEl) {
- activeEl.classList.remove('code-annotation-active');
- }
- selectCodeLines(clickedEl);
- clickedEl.classList.add('code-annotation-active');
- } else {
- // Unselect the line
- unselectCodeLines();
- clickedEl.classList.remove('code-annotation-active');
- }
- });
- }
- const findCites = (el) => {
- const parentEl = el.parentElement;
- if (parentEl) {
- const cites = parentEl.dataset.cites;
- if (cites) {
- return {
- el,
- cites: cites.split(' ')
- };
- } else {
- return findCites(el.parentElement)
- }
- } else {
- return undefined;
- }
- };
- var bibliorefs = window.document.querySelectorAll('a[role="doc-biblioref"]');
- for (var i=0; i<bibliorefs.length; i++) {
- const ref = bibliorefs[i];
- const citeInfo = findCites(ref);
- if (citeInfo) {
- tippyHover(citeInfo.el, function() {
- var popup = window.document.createElement('div');
- citeInfo.cites.forEach(function(cite) {
- var citeDiv = window.document.createElement('div');
- citeDiv.classList.add('hanging-indent');
- citeDiv.classList.add('csl-entry');
- var biblioDiv = window.document.getElementById('ref-' + cite);
- if (biblioDiv) {
- citeDiv.innerHTML = biblioDiv.innerHTML;
- }
- popup.appendChild(citeDiv);
- });
- return popup.innerHTML;
- });
- }
- }
-});
-</script>
-<nav class="page-navigation">
- <div class="nav-page nav-page-previous">
- <a href="/index.html" class="pagination-link">
- <i class="bi bi-arrow-left-short"></i> <span class="nav-page-text">Preface</span>
- </a>
- </div>
- <div class="nav-page nav-page-next">
- <a href="/notebooks/intro.html" class="pagination-link">
- <span class="nav-page-text"><span class='chapter-number'>2</span>Â <span class='chapter-title'>`ConformalGenerator`</span></span> <i class="bi bi-arrow-right-short"></i>
- </a>
- </div>
-</nav>
-</div> <!-- /content -->
-
-</body>
-
-</html>
\ No newline at end of file
diff --git a/notebooks/proposal.ipynb b/notebooks/proposal.ipynb
deleted file mode 100644
index d1b73ccb806be37956fed1fe0df7119efe0a9b35..0000000000000000000000000000000000000000
--- a/notebooks/proposal.ipynb
+++ /dev/null
@@ -1,218 +0,0 @@
-{
- "cells": [
- {
- "cell_type": "raw",
- "metadata": {},
- "source": [
- "---\n",
- "title: High-Fidelity Counterfactual Explanations through Conformal Prediction\n",
- "subtitle: Research Proposal\n",
- "abstract: |\n",
- " We propose Conformal Counterfactual Explanations: an effortless and rigorous way to produce realistic and faithful Counterfactual Explanations using Conformal Prediction. To address the need for realistic counterfactuals, existing work has primarily relied on separate generative models to learn the data-generating process. While this is an effective way to produce plausible and model-agnostic counterfactual explanations, it not only introduces a significant engineering overhead but also reallocates the task of creating realistic model explanations from the model itself to the generative model. Recent work has shown that there is no need for any of this when working with probabilistic models that explicitly quantify their own uncertainty. Unfortunately, most models used in practice still do not fulfil that basic requirement, in which case we would like to have a way to quantify predictive uncertainty in a post-hoc fashion.\n",
- "---"
- ],
- "id": "0675cd89"
- },
- {
- "cell_type": "code",
- "metadata": {},
- "source": [
- "include(\"notebooks/setup.jl\")\n",
- "eval(setup_notebooks)"
- ],
- "id": "8a310cd5",
- "execution_count": null,
- "outputs": []
- },
- {
- "cell_type": "markdown",
- "metadata": {},
- "source": [
- "## Motivation\n",
- "\n",
- "Counterfactual Explanations are a powerful, flexible and intuitive way to not only explain black-box models but also enable affected individuals to challenge them through the means of Algorithmic Recourse. \n",
- "\n",
- "### Counterfactual Explanations or Adversarial Examples?\n",
- "\n",
- "Most state-of-the-art approaches to generating Counterfactual Explanations (CE) rely on gradient descent in the feature space. The key idea is to perturb inputs $x\\in\\mathcal{X}$ into a black-box model $f: \\mathcal{X} \\mapsto \\mathcal{Y}$ in order to change the model output $f(x)$ to some pre-specified target value $t\\in\\mathcal{Y}$. Formally, this boils down to defining some loss function $\\ell(f(x),t)$ and taking gradient steps in the minimizing direction. The so-generated counterfactuals are considered valid as soon as the predicted label matches the target label. A stripped-down counterfactual explanation is therefore little different from an adversarial example. In @fig-adv, for example, generic counterfactual search as in @wachter2017counterfactual has been applied to MNIST data.\n"
- ],
- "id": "ea1eb4c1"
- },
- {
- "cell_type": "code",
- "metadata": {},
- "source": [
- "# Data:\n",
- "counterfactual_data = load_mnist()\n",
- "X, y = CounterfactualExplanations.DataPreprocessing.unpack_data(counterfactual_data)\n",
- "input_dim, n_obs = size(counterfactual_data.X)\n",
- "M = load_mnist_mlp()\n",
- "# Target:\n",
- "factual_label = 8\n",
- "x = reshape(X[:,rand(findall(predict_label(M, counterfactual_data).==factual_label))],input_dim,1)\n",
- "target = 3\n",
- "factual = predict_label(M, counterfactual_data, x)[1]\n",
- "# Search:\n",
- "n_ce = 3\n",
- "generator = GenericGenerator()\n",
- "ces = generate_counterfactual(x, target, counterfactual_data, M, generator; num_counterfactuals=n_ce)"
- ],
- "id": "2bc6d1d2",
- "execution_count": null,
- "outputs": []
- },
- {
- "cell_type": "code",
- "metadata": {},
- "source": [
- "image_size = 200\n",
- "p1 = plot(\n",
- " convert2image(MNIST, reshape(x,28,28)),\n",
- " axis=nothing, \n",
- " size=(image_size, image_size),\n",
- " title=\"Factual\"\n",
- ")\n",
- "plts = [p1]\n",
- "\n",
- "counterfactuals = CounterfactualExplanations.counterfactual(ces)\n",
- "phat = target_probs(ces)\n",
- "for x in zip(eachslice(counterfactuals; dims=3), eachslice(phat; dims=3))\n",
- " ce, _phat = (x[1],x[2])\n",
- " _title = \"p(y=$(target)|x′)=$(round(_phat[1]; digits=3))\"\n",
- " plt = plot(\n",
- " convert2image(MNIST, reshape(ce,28,28)),\n",
- " axis=nothing, \n",
- " size=(image_size, image_size),\n",
- " title=_title\n",
- " )\n",
- " plts = [plts..., plt]\n",
- "end\n",
- "plt = plot(plts...; size=(image_size * (n_ce + 1),image_size), layout=(1,(n_ce + 1)))\n",
- "savefig(plt, joinpath(www_path, \"you_may_not_like_it.png\"))"
- ],
- "id": "93ab4114",
- "execution_count": null,
- "outputs": []
- },
- {
- "cell_type": "markdown",
- "metadata": {},
- "source": [
- "{#fig-adv}\n",
- "\n",
- "The crucial difference between adversarial examples and counterfactuals is one of intent. While adversarial examples are typically intended to go unnoticed, counterfactuals in the context of Explainable AI are generally sought to be \"plausible\", \"realistic\" or \"feasible\". To fulfil this latter goal, researchers have come up with a myriad of ways. @joshi2019realistic were among the first to suggest that instead of searching counterfactuals in the feature space, we can instead traverse a latent embedding learned by a surrogate generative model. Similarly, @poyiadzi2020face use density ... Finally, @karimi2021algorithmic argues that counterfactuals should comply with the causal model that generates them [CHECK IF WE CAN PHASE THIS LIKE THIS]. Other related approaches include ... All of these different approaches have a common goal: they aim to ensure that the generated counterfactuals comply with the (learned) data-generating process (DGB). \n",
- "\n",
- "::: {#def-plausible}\n",
- "\n",
- "## Plausible Counterfactuals\n",
- "\n",
- "Formally, if $x \\sim \\mathcal{X}$ and for the corresponding counterfactual we have $x^{\\prime}\\sim\\mathcal{X}^{\\prime}$, then for $x^{\\prime}$ to be considered a plausible counterfactual, we need: $\\mathcal{X} \\approxeq \\mathcal{X}^{\\prime}$.\n",
- "\n",
- ":::\n",
- "\n",
- "In the context of Algorithmic Recourse, it makes sense to strive for plausible counterfactuals, since anything else would essentially require individuals to move to out-of-distribution states. But it is worth noting that our ambition to meet this goal, may have implications on our ability to faithfully explain the behaviour of the underlying black-box model (arguably our principal goal). By essentially decoupling the task of learning plausible representations of the data from the model itself, we open ourselves up to vulnerabilities. Using a separate generative model to learn $\\mathcal{X}$, for example, has very serious implications for the generated counterfactuals. @fig-latent compares the results of applying REVISE [@joshi2019realistic] to MNIST data using two different Variational Auto-Encoders: while the counterfactual generated using an expressive (strong) VAE is compelling, the result relying on a less expressive (weak) VAE is not even valid. In this latter case, the decoder step of the VAE fails to yield values in $\\mathcal{X}$ and hence the counterfactual search in the learned latent space is doomed. \n",
- "\n",
- "{#fig-latent}\n",
- "\n",
- "> Here it would be nice to have another example where we poison the data going into the generative model to hide biases present in the data (e.g. Boston housing).\n",
- "\n",
- "- Latent can be manipulated: \n",
- " - train biased model\n",
- " - train VAE with biased variable removed/attacked (use Boston housing dataset)\n",
- " - hypothesis: will generate bias-free explanations\n",
- "\n",
- "### From Plausible to High-Fidelity Counterfactuals {#sec-fidelity}\n",
- "\n",
- "In light of the findings, we propose to generally avoid using surrogate models to learn $\\mathcal{X}$ in the context of Counterfactual Explanations.\n",
- "\n",
- "::: {#prp-surrogate}\n",
- "\n",
- "## Avoid Surrogates\n",
- "\n",
- "Since we are in the business of explaining a black-box model, the task of learning realistic representations of the data should not be reallocated from the model itself to some surrogate model.\n",
- "\n",
- ":::\n",
- "\n",
- "In cases where the use of surrogate models cannot be avoided, we propose to weigh the plausibility of counterfactuals against their fidelity to the black-box model. In the context of Explainable AI, fidelity is defined as describing how an explanation approximates the prediction of the black-box model [@molnar2020interpretable]. Fidelity has become the default metric for evaluating Local Model-Agnostic Models, since they often involve local surrogate models whose predictions need not always match those of the black-box model. \n",
- "\n",
- "In the case of Counterfactual Explanations, the concept of fidelity has so far been ignored. This is not altogether surprising, since by construction and design, Counterfactual Explanations work with the predictions of the black-box model directly: as stated above, a counterfactual $x^{\\prime}$ is considered valid if and only if $f(x^{\\prime})=t$, where $t$ denote some target outcome. \n",
- "\n",
- "Does fidelity even make sense in the context of CE, and if so, how can we define it? In light of the examples in the previous section, we think it is urgent to introduce a notion of fidelity in this context, that relates to the distributional properties of the generated counterfactuals. In particular, we propose that a high-fidelity counterfactual $x^{\\prime}$ complies with the class-conditional distribution $\\mathcal{X}_{\\theta} = p_{\\theta}(X|y)$ where $\\theta$ denote the black-box model parameters. \n",
- "\n",
- "::: {#def-fidele}\n",
- "\n",
- "## High-Fidelity Counterfactuals\n",
- "\n",
- "Let $\\mathcal{X}_{\\theta}|y = p_{\\theta}(X|y)$ denote the class-conditional distribution of $X$ defined by $\\theta$. Then for $x^{\\prime}$ to be considered a high-fidelity counterfactual, we need: $\\mathcal{X}_{\\theta}|t \\approxeq \\mathcal{X}^{\\prime}$ where $t$ denotes the target outcome.\n",
- "\n",
- ":::\n",
- "\n",
- "In order to assess the fidelity of counterfactuals, we propose the following two-step procedure:\n",
- "\n",
- "1) Generate samples $X_{\\theta}|y$ and $X^{\\prime}$ from $\\mathcal{X}_{\\theta}|t$ and $\\mathcal{X}^{\\prime}$, respectively.\n",
- "2) Compute the Maximum Mean Discrepancy (MMD) between $X_{\\theta}|y$ and $X^{\\prime}$. \n",
- "\n",
- "If the computed value is different from zero, we can reject the null-hypothesis of fidelity.\n",
- "\n",
- "> Two challenges here: 1) implementing the sampling procedure in @grathwohl2020your; 2) it is unclear if MMD is really the right way to measure this. \n",
- "\n",
- "## Conformal Counterfactual Explanations\n",
- "\n",
- "In @sec-fidelity, we have advocated for avoiding surrogate models in the context of Counterfactual Explanations. In this section, we introduce an alternative way to generate high-fidelity Counterfactual Explanations. In particular, we propose Conformal Counterfactual Explanations (CCE), that is Counterfactual Explanations that minimize the predictive uncertainty of conformal models. \n",
- "\n",
- "### Minimizing Predictive Uncertainty\n",
- "\n",
- "@schut2021generating demonstrated that the goal of generating realistic (plausible) counterfactuals can also be achieved by seeking counterfactuals that minimize the predictive uncertainty of the underlying black-box model. Similarly, @antoran2020getting ...\n",
- "\n",
- "- Problem: restricted to Bayesian models.\n",
- "- Solution: post-hoc predictive uncertainty quantification. In particular, Conformal Prediction. \n",
- "\n",
- "### Background on Conformal Prediction\n",
- "\n",
- "- Distribution-free, model-agnostic and scalable approach to predictive uncertainty quantification.\n",
- "- Conformal prediction is instance-based. So is CE. \n",
- "- Take any fitted model and turn it into a conformal model using calibration data.\n",
- "- Our approach, therefore, relaxes the restriction on the family of black-box models, at the cost of relying on a subset of the data. Arguably, data is often abundant and in most applications practitioners tend to hold out a test data set anyway. \n",
- "\n",
- "> Does the coverage guarantee carry over to counterfactuals?\n",
- "\n",
- "### Generating Conformal Counterfactuals\n",
- "\n",
- "While Conformal Prediction has recently grown in popularity, it does introduce a challenge in the context of classification: the predictions of Conformal Classifiers are set-valued and therefore difficult to work with, since they are, for example, non-differentiable. Fortunately, @stutz2022learning introduced carefully designed differentiable loss functions that make it possible to evaluate the performance of conformal predictions in training. We can leverage these recent advances in the context of gradient-based counterfactual search ...\n",
- "\n",
- "> Challenge: still need to implement these loss functions. \n",
- "\n",
- "## Experiments\n",
- "\n",
- "### Research Questions\n",
- "\n",
- "- Is CP alone enough to ensure realistic counterfactuals?\n",
- "- Do counterfactuals improve further as the models get better?\n",
- "- Do counterfactuals get more realistic as coverage\n",
- "- What happens as we vary coverage and setsize?\n",
- "- What happens as we improve the model robustness?\n",
- "- What happens as we improve the model's ability to incorporate predictive uncertainty (deep ensemble, laplace)?\n",
- "- What happens if we combine with DiCE, ClaPROAR, Gravitational?\n",
- "- What about CE robustness to endogenous shifts [@altmeyer2023endogenous]?\n",
- "\n",
- "- Benchmarking:\n",
- " - add PROBE [@pawelczyk2022probabilistically] into the mix.\n",
- " - compare travel costs to domain shits.\n",
- "\n",
- "> Nice to have: What about using Laplace Approximation, then Conformal Prediction? What about using Conformalised Laplace? \n",
- "\n",
- "## References\n"
- ],
- "id": "9f0a2e10"
- }
- ],
- "metadata": {
- "kernelspec": {
- "name": "julia-1.6",
- "language": "julia",
- "display_name": "Julia 1.6.5"
- }
- },
- "nbformat": 4,
- "nbformat_minor": 5
-}
\ No newline at end of file