English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

We study the problem of off-policy evaluation (OPE) in Reinforcement Learning(RL), where the aim is to estimate the performance of a new policy givenhistorical data that may have been generated by a different policy, orpolicies. In particular, we introduce a novel doubly-robust estimator for theOPE problem in RL, based on the Targeted Maximum Likelihood Estimationprinciple from the statistical causal inference literature. We also introduceseveral variance reduction techniques that lead to impressive performance gainsin off-policy evaluation. We show empirically that our estimator uniformly winsover existing off-policy evaluation methods across multiple RL environments andvarious levels of model misspecification. Finally, we further the existingtheoretical analysis of estimators for the RL off-policy estimation problem byshowing their $O_P(1/\sqrt{n})$ rate of convergence and characterizing theirasymptotic distribution.

More Efficient Off-Policy Evaluation through Regularized Targeted Learning