AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

From Reward-Free Representations to Preferences: Rethinking Offline Preference-Based Reinforcement Learning

Jun-Jie Yang, Chia-Heng Hsu, Kui-Yuan Chen, Ping-Chun Hsieh

ICML 2026 · Reinforcement learning › RL from feedback · approach: Preference-based RL reward learning · 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 preference-guided fine-tuning of FB representations on top of test-time search (ablation)FB-PbRL · DMC Cheetah averageepisodic return: 344.7 -> 621.7Improved
add fine-tuning on top of test-time search (ablation)FB-PbRL · DMC Walker averageepisodic return: 533.4 -> 762.9Improved
add fine-tuning on top of test-time search (ablation)FB-PbRL · DMC Quadruped averageepisodic return: 663.4 -> 846.9Improved
add fine-tuning on top of test-time search (ablation)FB-PbRL · DMC PointMass averageepisodic return: 69.1 -> 570.8Improved
FB-PbRL versus CLARIFYCLARIFY · DMC Walker averageepisodic return: 248.9 -> 762.9Improved
FB-PbRL versus CLARIFYCLARIFY · DMC PointMass averageepisodic return: 317.8 -> 570.8Improved
contrastive preference loss instead of BT reward model in FB fine-tuning (ablation)FB-BT-FT · DMC Cheetah aggregatedepisodic return: 536.6 -> 621.7Improved
contrastive preference loss instead of BT reward model (ablation)FB-BT-FT · DMC Walker aggregatedepisodic return: 600.6 -> 794.5Improved
contrastive preference loss instead of BT reward model (ablation)FB-BT-FT · DMC Quadruped aggregatedepisodic return: 714.1 -> 846.9Improved
use 2000 preferences instead of 10000 reward-labeled transitions vs FBFB · DMC Quadruped averageepisodic return: 561.7 -> 826.3Improved
preferences instead of ground-truth reward vs best RFRL baselinePSM · DMC Cheetah averageepisodic return: 626.0 -> 645.4Improved
preferences instead of ground-truth rewardFB · DMC Walker averageepisodic return: 719.9 -> 699.4Hurt

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