AI Paper Atlas

Research map / Reinforcement learning / Offline RL

From Static Constraints to Dynamic Adaptation: Sample-Level Constraint Relaxation for Offline-to-Online Reinforcement Learning

Lipeng Zu, YU QIAN, Shayok Chakraborty, Xiaonan Zhang

ICML 2026 · Reinforcement learning › Offline RL · approach: Diffusion models for offline decision making · 0 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
Add DARE sample-level exchange to Cal-QLCal-QL · D4RL total, 0.2M online stepstotal normalized score: 1272.7 -> 1415.4Improved
Add DARE sample-level exchange to IQLIQL · D4RL total, 0.2M online stepstotal normalized score: 1085.3 -> 1253.0Improved
Add behavior-consistent exchange to DARE-I with SAC objective removed (ablation)DARE-I w/o SAC · Adroit human-v0 totaltotal normalized score: 106.67 -> 113.90Improved
Add exchange to DARE-I with PEX removed (ablation)DARE-I w/o PEX · Adroit human-v0 totaltotal normalized score: 8.96 -> 56.02Improved
Add exchange to TD-only DARE-I (no SAC, no PEX) (ablation)DARE-I w/o S&P · Adroit human-v0 totaltotal normalized score: 104.25 -> 122.15Improved
Use posterior-induced exchange instead of monotonic Gaussian or direct distance exchange (ablation)DARE-C · D4RL locomotion (replay and medium datasets)normalized score stabilityImproved
Use behavior-consistent exchange rather than reversed (curiosity-style) assignment (ablation)DARE-I · D4RL locomotion total, diffusion behavior modeltotal normalized score: 791.1 -> 819.8Improved
Use behavior-consistent exchange rather than reversed assignment, online actor as behavior model (ablation)DARE-I · D4RL locomotion total, policy behavior modeltotal normalized score: 766.5 -> 812.1Improved
Add exchange mechanism to DARE-C and DARE-I (ablation)DARE-C, DARE-I · D4RL locomotion replay and medium datasetsaverage score differenceImproved
Use online policy actor instead of offline diffusion model as behavior model (ablation)DARE · D4RL locomotiontotal normalized score: 819.8 -> 812.1Hurt
Use DARE-I in place of PEX on hopper-medium-expertIQL group · hopper-medium-expertnormalized score: 90.2 -> 104.5Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 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