ECE: Ensemble of Counterfactual Explainers

Riccardo Guidotti, Salvatore Ruggieri
Department of Computer Science, University of Pisa, Italy
riccardo.guidotti@unipi.it, salvatore.ruggieri@unipi.it

In eXplainable Artificial Intelligence (XAI), several counterfactual explainers have been proposed, each focusing on some desirable properties of counterfactual instances: minimality, actionability, stability, diversity, plausibility, discriminative power. We propose an ensemble of counterfactual explainers that boosts weak explainers, which provide only a subset of such properties, to a powerful method covering all of them. The ensemble runs weak explainers on a sample of instances and of features, and it combines their results by exploiting a diversity-driven selection function. The method is model-agnostic and, through a wrapping approach based on autoencoders, it is also data-agnostic

References

[1] R. Guidotti, S. Ruggieri. Ensemble of Counterfactual Explainers. Discovery Science (DS 2021). 358-368. Vol. 12986 of LNCS, Springer, October 2021.

How to install required packages

Linux / macOS

python3 -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt

Adding Scamandar package directory to pythonpath so that we can import the experiments as is

cat >> .venv/bin/activate <<'BASH'
export PYTHONPATH="${PYTHONPATH:+${PYTHONPATH}:}$(cd "$(dirname "${BASH_SOURCE[0]}")/../.." && pwd)"
BASH

In case the venv is active, you have to run deactivate and source .venv/bin/activate again to take effect.