Handling Confounding for Realistic Off-Policy Evaluation

Abstract

Inverse Propensity Score estimator (IPS) is a basic, unbiased, off-policy evaluation technique to measure the impact of a user-interactive system without serving live traffic. We present our work on applying IPS to real-world settings by addressing some practical challenges, thereby enabling successful policy evaluation. In particular, we show that off-policy evaluation can be impossible in the absence of a complete context and we describe a systematic way of defining the context.