Concept-driven Off Policy Evaluation

TL;DR

This paper introduces concept-based estimators for off-policy evaluation that leverage human-interpretable concepts to reduce variance, improve accuracy, and enable targeted interventions, with proven unbiasedness and practical algorithms for learning concepts.

Contribution

It develops a family of concept-based OPE estimators that are unbiased and variance-reducing, and proposes an end-to-end method to learn interpretable concepts for real-world applications.

Findings

Concept-based estimators significantly improve OPE performance.

Learned concepts are interpretable, concise, and diverse.

Estimators enable targeted interventions and are unbiased.

Abstract

Evaluating off-policy decisions using batch data poses significant challenges due to limited sample sizes leading to high variance. To improve Off-Policy Evaluation (OPE), we must identify and address the sources of this variance. Recent research on Concept Bottleneck Models (CBMs) shows that using human-explainable concepts can improve predictions and provide better understanding. We propose incorporating concepts into OPE to reduce variance. Our work introduces a family of concept-based OPE estimators, proving that they remain unbiased and reduce variance when concepts are known and predefined. Since real-world applications often lack predefined concepts, we further develop an end-to-end algorithm to learn interpretable, concise, and diverse parameterized concepts optimized for variance reduction. Our experiments with synthetic and real-world datasets show that both known and learned concept-based estimators significantly improve OPE performance. Crucially, we show that, unlike other OPE methods, concept-based estimators are easily interpretable and allow for targeted interventions on specific concepts, further enhancing the quality of these estimators.