AI Paper Atlas

Research map / Reinforcement learning / Offline RL

RvS: What is Essential for Offline RL via Supervised Learning?

Scott Emmons, Benjamin Eysenbach, Ilya Kostrikov, Sergey Levine

ICLR 2022 · Reinforcement learning › Offline RL · approach: Offline-to-online fine-tuning · 3 citations (OpenAlex)

This paper tried a common change: Add dropout

Add dropout regularization · Kitchen complete · evaluation return Improved

What 795 accepted papers report for this change

See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.

Read the paper

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
Condition on goal states instead of reward valuesRvS on AntMaze v2 · AntMaze v2 averagenormalized score: 25.6 -> 53.5Improved
Condition on goal states instead of reward valuesRvS on Kitchen v0 · Kitchen v0 averagenormalized score: 1.0 -> 54.0Improved
Use reward conditioning rather than goal conditioningRvS on Gym v2 · Gym v2normalized scoreImproved
Use a simple MLP with supervised likelihood training instead of a Transformer sequence modelRvS-R · Gym v2 averagenormalized score: 54.7 vs 57.0No clear effect
Use RvS rather than the strongest value-based baselineRvS-G · AntMaze v2 averagenormalized score: 50.6 -> 53.5Improved
Use RvS rather than the strongest value-based baselineRvS-G · Kitchen v0 averagenormalized score: 48.2 -> 54.0Improved
Use reward-conditioned imitation rather than CQLRvS-R · Gym random-v2 averagenormalized score: 10.1 -> 3.8Hurt
Add dropout regularization (ablation)RvS · Kitchen completeevaluation returnImproved
Add dropout regularization (ablation)RvS · Hopper medium-expert and AntMaze medium-playevaluation returnHurt
Add dropout regularization (ablation)RvS · Pusherevaluation returnNo clear effect
Use a categorical distribution over discretized actions instead of a unimodal Gaussian (ablation)RvS-G · GCSLgoal success scoreImproved
Increase network capacity (ablation)RvS · GCSL and D4RL tasksevaluation returnImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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