Learning Counterfactually Invariant Predictors
This work addresses the need for fair and robust predictors in real-world applications, though it appears incremental by building on existing conditional independence measures.
The paper tackled the problem of learning predictors that are counterfactually invariant for fairness, robustness, and generalizability, by proposing graphical criteria and a model-agnostic framework called CIP, which demonstrated effectiveness in enforcing invariance across simulated and real-world datasets.
Notions of counterfactual invariance (CI) have proven essential for predictors that are fair, robust, and generalizable in the real world. We propose graphical criteria that yield a sufficient condition for a predictor to be counterfactually invariant in terms of a conditional independence in the observational distribution. In order to learn such predictors, we propose a model-agnostic framework, called Counterfactually Invariant Prediction (CIP), building on the Hilbert-Schmidt Conditional Independence Criterion (HSCIC), a kernel-based conditional dependence measure. Our experimental results demonstrate the effectiveness of CIP in enforcing counterfactual invariance across various simulated and real-world datasets including scalar and multi-variate settings.