Simulating counterfactuals

TL;DR

This paper introduces a particle filter-based algorithm for simulating counterfactuals under complex conditions, enabling more accurate inference in scenarios like fairness analysis in credit scoring.

Contribution

It presents a novel algorithm for simulating counterfactuals with conditions on discrete and continuous variables, validated as asymptotically valid.

Findings

Algorithm effectively simulates counterfactuals under complex conditions

Application to credit-scoring fairness analysis demonstrates practical utility

Asymptotic validity of the particle filter approach

Abstract

Counterfactual inference considers a hypothetical intervention in a parallel world that shares some evidence with the factual world. If the evidence specifies a conditional distribution on a manifold, counterfactuals may be analytically intractable. We present an algorithm for simulating values from a counterfactual distribution where conditions can be set on both discrete and continuous variables. We show that the proposed algorithm can be presented as a particle filter leading to asymptotically valid inference. The algorithm is applied to fairness analysis in credit-scoring.