AI Paper Atlas

Research map / Reinforcement learning / Offline RL

Expert-Supervised Reinforcement Learning for Offline Policy Learning and Evaluation

Aaron Sonabend, Junwei Lu, Leo Anthony Celi, Tianxi Cai, Peter Szolovits

NeurIPS 2020 · Reinforcement learning › Offline RL · approach: Simple practical offline RL

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use posterior hypothesis testing to decide when to depart from the behavior policyESRL · Riverswimmean test reward per episodeImproved
Use posterior hypothesis testing to decide when to depart from the behavior policyESRL · Sepsisexpected cumulative inverse SOFA valueImproved
Replace posterior hypothesis testing with mean-based ensemble action selectionESRL · Riverswimmean test reward per episodeHurt
Use behavior cloning instead of ESRL policy learningRiverswim · mean test reward per episodemean test reward per episodeNo clear effect
Use random ensembles instead of the DQN ensemble baselineRiverswim · mean test reward per episodemean test reward per episodeImproved
Use random ensembles instead of the DQN ensemble baselineSepsis · expected cumulative inverse SOFA valueexpected cumulative inverse SOFA valueNo clear effect
Estimate policy value by posterior MDP sampling rather than sample-based importance samplingESRL value estimation · Riverswimmean squared errorImproved
Estimate policy value by posterior MDP sampling rather than non-parametric model estimatesESRL value estimation · Riverswimmean squared errorImproved
Increase the offline dataset sizeESRL value estimation · Riverswimmean squared errorImproved
Deviate from physician behavior without accounting for uncertaintyoffline policy learning · Sepsisexpected cumulative inverse SOFA valueHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy