Spotlight

State Relevance for Off-Policy Evaluation

Simon Shen · Yecheng Ma · Omer Gottesman · Finale Doshi-Velez

[ Abstract ] [ Livestream: Visit Reinforcement Learning 12 ] [ Paper ]
Wed 21 Jul 6:45 a.m. — 6:50 a.m. PDT

Importance sampling-based estimators for off-policy evaluation (OPE) are valued for their simplicity, unbiasedness, and reliance on relatively few assumptions. However, the variance of these estimators is often high, especially when trajectories are of different lengths. In this work, we introduce Omitting-States-Irrelevant-to-Return Importance Sampling (OSIRIS), an estimator which reduces variance by strategically omitting likelihood ratios associated with certain states. We formalize the conditions under which OSIRIS is unbiased and has lower variance than ordinary importance sampling, and we demonstrate these properties empirically.

Chat is not available.