ECCCo

Energy-Constrained Counterfactual Explanations.

This work is currently undergoing peer review. This README is therefore only meant to provide reviewers access to the code base. The code base will be made public after the review process.

Inspecting the Package Code

This code base is structured as a Julia package. The package code is located in the src/ folder.

Inspecting the Results

All results have been carefully reported either in the paper itself or in the supplementary material. In addition, we have released our results as binary files. These will be made publicly available after the review process.

Reproducing the Results

To reproduce the results, you need to install the package, which will automatically install all dependencies. Since the package is not publicly registered you will need to install it from source. To do so, start Julia from within this folder by executing julia --project from the command line and then enter Pkg mode by typing ]. Then execute the following commands:

(ECCCo) pkg> resolve
(ECCCo) pkg> instantiate

Next, you may need to activate, resolve and instantiate the environment in experiments:

(ECCCo) pkg> activate experiments/
  Activating project at `~/code/ECCCo.jl/experiments`
(experiments) pkg> resolve
(experiments) pkg> instantiate

After that is done, you can exit Julia and proceed below.

Sequential

The experiments/ folder contains separate Julia scripts for each dataset and a run_experiments.jl that calls the individual scripts. You can either run these scripts inside a Julia session or just use the command line to execute them as described in the following.

To run the experiment for a single dataset, (e.g. linearly_separable) simply run the following command:

julia --project=experiments/ experiments/run_experiments.jl -- data=linearly_separable

We use the following identifiers:

• linearly_separable (Linearly Separable data)
• moons (Moons data)
• circles (Circles data)
• california_housing (California Housing data)
• gmsc (GMSC data)
• german_credit (German Credit data)
• mnist (MNIST data)
• fmnist (Fashion MNIST data)

To run experiments for multiple datasets at once simply separate them with a comma ,

julia --project=experiments/ experiments/run_experiments.jl -- data=linearly_separable,moons,circles

To run all experiments at once you can instead run

julia --project=experiments/ experiments/run_experiments.jl -- run-all

Pre-trained versions of all of our black-box models have been archived as Pkg artifacts and are used by default. Should you wish to retrain the models as well, simply use the retrain flag as follows:

julia --project=experiments experiments/run_experiments.jl -- retrain data=linearly_separable

Multi-threading

To use multi-threading, proceed as follows:

julia --threads 16 --project=experiments experiments/run_experiments.jl -- data=linearly_separable threaded

Multi-Processing

To use multi-processing, proceed as follows:

mpiexecjl --project=experiments -n 4 julia experiments/run_experiments.jl -- data=linearly_separable mpi

Multi-processing and multi-threading can be combined:

mpiexecjl --project=experiments -n 4 julia experiments/run_experiments.jl -- data=linearly_separable threaded mpi

Reproducing Figures

To recreate the exact figures shown in the main paper you can use two notebooks:

• experiments/notebooks/figure2.qmd: Figure 2 (gradient fields)
• experiments/notebooks/figure1and3.qmd: Figures 1 and 3 (MNIST examples)