Ensemble of Counterfactual Explainers

TL;DR

This paper introduces an ensemble method for counterfactual explainers in XAI that combines multiple weak explainers to achieve comprehensive, model-agnostic, and data-agnostic explanations covering various desirable properties.

Contribution

It presents a novel ensemble approach that integrates weak counterfactual explainers to produce more complete and robust explanations in XAI.

Findings

Improves explanation quality by combining multiple weak explainers

Model-agnostic and data-agnostic due to autoencoder wrapping

Achieves comprehensive coverage of counterfactual properties

Abstract

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.