CeFlow: A Robust and Efficient Counterfactual Explanation Framework for Tabular Data using Normalizing Flows

TL;DR

CeFlow introduces a robust and efficient counterfactual explanation framework for tabular data using normalizing flows, overcoming speed and stability issues of previous VAE-based methods.

Contribution

The paper proposes CeFlow, a novel counterfactual explanation method employing normalizing flows to improve speed and stability for mixed-type tabular data.

Findings

Outperforms state-of-the-art methods in speed and stability

Provides stable counterfactuals for mixed-type features

Demonstrates effectiveness through numerical experiments

Abstract

Counterfactual explanation is a form of interpretable machine learning that generates perturbations on a sample to achieve the desired outcome. The generated samples can act as instructions to guide end users on how to observe the desired results by altering samples. Although state-of-the-art counterfactual explanation methods are proposed to use variational autoencoder (VAE) to achieve promising improvements, they suffer from two major limitations: 1) the counterfactuals generation is prohibitively slow, which prevents algorithms from being deployed in interactive environments; 2) the counterfactual explanation algorithms produce unstable results due to the randomness in the sampling procedure of variational autoencoder. In this work, to address the above limitations, we design a robust and efficient counterfactual explanation framework, namely CeFlow, which utilizes normalizing flows for the mixed-type of continuous and categorical features. Numerical experiments demonstrate that our technique compares favorably to state-of-the-art methods. We release our source at https://github.com/tridungduong16/fairCE.git for reproducing the results.