AI Paper Atlas

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

Multi-timescale Reinforcement Learning by Value Reconstruction

Zhan Su, Peixi Peng, Xinyu Hu, Cong Li, Yisen Zhao, Zhuojian Li, Yonghong Tian, Fanqi Shen

ICML 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Temporal-difference and discounting theory

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 multi-timescale critics with reward decoder and attention fusion to SACSAC · DMControl 6 tasksaverage score: 206.8 -> 475.2Improved
add multi-timescale critics with reward decoder and attention fusion to PPOPPO · DMControl 6 tasksaverage score: 167.7 -> 445.7Improved
add multi-timescale framework to DrQ-v2DrQ-v2 · DMControl 6 common tasks, 500K stepsaverage score: 888.0 -> 918.3Improved
add multi-timescale framework to DrQ-v2DrQ-v2 · DMControl 5 hard tasks, 500K stepsaverage score: 406.8 -> 528.4Improved
add multi-timescale framework to DrQ-v2DrQ-v2 · DMControl 5 hard tasks, 1M stepsaverage score: 564.0 -> 686.6Improved
add multi-timescale framework to DrQ-v2DrQ-v2 · CARLA highway (4 weathers)average episode score: 223.6 -> 343.1Improved
increase number of critics from N=2 to N=4 (ablation)Ours · Quadruped Walk 1Mscore: 797 -> 830Improved
increase number of critics from N=4 to N=8 (ablation)Ours · Walker Run 1Mscore: 717 -> 726No clear effect
vary decoding horizon T among 10, 15, 20 (ablation)Ours · Quadruped Walk 1Mscore: 850 -> 846 -> 857No clear effect
widen DrQ-v2 critic to match parameter budget (ablation)DrQ-v2 · Quadruped Walk 1Mscore: 732 -> 760No clear effect
share trunk across critics (compact variant) (ablation)Ours · Quadruped Walk 1Mscore: 846 -> 830No clear effect
add reconstruction loss, reward loss, attention fusion, and neural decoder (each removed in ablation) (ablation)Ours · Quadruped Walk and CARLAlearning curveImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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