Distributional Counterfactual Explanations With Optimal Transport

TL;DR

This paper introduces distributional counterfactual explanations (DCE) using optimal transport to capture distributional properties of data, offering broader insights into black-box models beyond traditional point-based counterfactuals.

Contribution

It proposes a novel distributional counterfactual explanation method leveraging optimal transport, enabling the alignment of counterfactual and factual data distributions with statistical confidence.

Findings

DCE effectively captures distributional characteristics of data.

The method provides deeper insights into model decision-making.

Experimental results demonstrate its practical utility.

Abstract

Counterfactual explanations (CE) are the de facto method for providing insights into black-box decision-making models by identifying alternative inputs that lead to different outcomes. However, existing CE approaches, including group and global methods, focus predominantly on specific input modifications, lacking the ability to capture nuanced distributional characteristics that influence model outcomes across the entire input-output spectrum. This paper proposes distributional counterfactual explanation (DCE), shifting focus to the distributional properties of observed and counterfactual data, thus providing broader insights. DCE is particularly beneficial for stakeholders making strategic decisions based on statistical data analysis, as it makes the statistical distribution of the counterfactual resembles the one of the factual when aligning model outputs with a target distribution\textemdash something that the existing CE methods cannot fully achieve. We leverage optimal transport (OT) to formulate a chance-constrained optimization problem, deriving a counterfactual distribution aligned with its factual counterpart, supported by statistical confidence. The efficacy of this approach is demonstrated through experiments, highlighting its potential to provide deeper insights into decision-making models.