Deep Backtracking Counterfactuals for Causally Compliant Explanations

TL;DR

This paper introduces DeepBC, a novel method for generating causally compliant backtracking counterfactuals in complex models, offering a versatile alternative to existing counterfactual explanation techniques.

Contribution

We propose DeepBC, a practical approach for computing backtracking counterfactuals in deep generative causal models, with two implementation variants and demonstrated effectiveness on image datasets.

Findings

DeepBC effectively generates counterfactuals for high-dimensional data.

The method maintains causal laws, ensuring causally compliant explanations.

Experimental results show versatility and modularity of DeepBC.

Abstract

Counterfactuals answer questions of what would have been observed under altered circumstances and can therefore offer valuable insights. Whereas the classical interventional interpretation of counterfactuals has been studied extensively, backtracking constitutes a less studied alternative where all causal laws are kept intact. In the present work, we introduce a practical method called deep backtracking counterfactuals (DeepBC) for computing backtracking counterfactuals in structural causal models that consist of deep generative components. We propose two distinct versions of our method--one utilizing Langevin Monte Carlo sampling and the other employing constrained optimization--to generate counterfactuals for high-dimensional data. As a special case, our formulation reduces to methods in the field of counterfactual explanations. Compared to these, our approach represents a causally compliant, versatile and modular alternative. We demonstrate these properties experimentally on a modified version of MNIST and CelebA.