AI Paper Atlas

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

Debiased Model-based Representations for Sample-efficient Continuous Control

Jiafei Lyu, Zichuan Lin, Scott Fujimoto, Kai Yang, Yangkun Chen, Saiyong Yang, Zongqing Lu, Deheng Ye

ICML 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Off-policy actor-critic with expressive policies · 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 InfoNCE mutual information loss (lambda_m > 0) (ablation)DR.Q · HumanoidBench and other representative tasksaverage returnImproved
remove InfoNCE loss (lambda_m = 0) relative to MR.Q (ablation)DR.Q without InfoNCE · representative tasksaverage returnNo clear effect
add the forget mechanism to LAP prioritized replay (ablation)DR.Q (only LAP) · 4 representative tasksaverage returnImproved
add TD-error prioritization to the forget mechanism (ablation)DR.Q (only forget) · 4 representative tasksaverage returnImproved
use DR.Q over SimBaV2DR.Q · DMC-Hardnormalized return: 15.5% gainImproved
use DR.Q over FoGDR.Q · HumanoidBench (w/ hand)normalized return: 58.9% improvementImproved
use DR.Q over MR.QDR.Q · DMC-Visualnormalized return: 26.8% leadImproved
use single shared hyperparameters across benchmarksDR.Q · Hopper-v4average returnHurt

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