AI Paper Atlas

Research map / Reinforcement learning / Policy gradient and actor-critic

Model Selection in Batch Policy Optimization

Jonathan Lee, George Tucker, Ofir Nachum, Bo Dai

ICML 2022 · Reinforcement learning › Policy gradient and actor-critic · approach: Model-based RL with guarantees · 1 citations (OpenAlex)

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
Select actions pessimistically across all model classesComplexity-Coverage Selection · Synthetic fully realizable contextual banditregretImproved
Use the model class with the smallest hidden dimensionSingle-class base algorithms · Synthetic fully realizable contextual banditregretImproved
Select among model classes using SLOPESLOPE Method · Synthetic approximation error-complexity trade-offregretImproved
Select among model classes using hold-out regression lossHold-out Method · Synthetic approximation error-complexity trade-offregretImproved
Use hold-out selection instead of SLOPEModel selection for approximation error-complexity trade-off · Synthetic approximation error-complexity trade-off, small dataregretImproved
Balance approximation error, complexity, and coverage simultaneouslyBatch policy optimization model selection · Contextual bandits with nested linear model classesoracle-inequality competitivenessHurt
Ignore approximation error and select pessimistically across classesComplexity-Coverage Selection · Contextual banditsregretImproved
Ignore comparator-specific coverage and use nested classes with SLOPE value estimatesSLOPE Method · Contextual banditsregretImproved
Ignore comparator-specific coverage and select by hold-out regression lossHold-out Method · Contextual banditsregretImproved

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