AI Paper Atlas

Research map / Reinforcement learning / Offline RL

A Closer Look at Offline RL Agents

Yuwei Fu, Di Wu, Benoit Boulet

NeurIPS 2022 · Reinforcement learning › Offline RL · approach: Simple practical offline RL · 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
replace AWR actor loss with SAC loss plus behavior log-likelihood term (RIQL)IQL · D4RL total over 9 taskstotal normalized return: 690.04 -> 759.05Improved
RIQL vs IQLIQL · hopper-medium-v2normalized return: 64.64 -> 91.58No clear effect
RIQL vs IQLIQL · halfcheetah-medium-v2normalized return: 47.43 -> 55.93Improved
RIQL vs IQLIQL · walker2d-medium-replay-v2normalized return: 73.53 -> 81.46Improved
augment IQL with model-generated data (MIQL) without sample selectionIQL · D4RL total over 9 taskstotal normalized return: 690.04 -> 334.16Hurt
augment TD3+BC with model-generated data (MTD3+BC) without sample selectionTD3+BC · D4RL total over 9 taskstotal normalized return: 676.82 -> 488.99Hurt
add uncertainty-based sample selection (USS) (ablation)MIQL · D4RL total over 9 taskstotal normalized return: 334.16 -> 699.98Improved
add uncertainty-based sample selection (USS) (ablation)MTD3+BC · D4RL total over 9 taskstotal normalized return: 488.99 -> 702.29Improved
add USS (ablation)MIQL · hopper-medium-v2normalized return: 8.89 -> 80.79Improved
add USS (ablation)MTD3+BC · walker2d-medium-v2normalized return: 85.33 -> 84.80No clear effect
add USS (ablation)MIQL · walker2d-medium-expert-v2normalized return: 87.96 -> 108.25Improved
compare USS agents with model-free counterpartsMIQL-USS · D4RL tasksnormalized return: 690.04 -> 699.98Mixed

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