AI Paper Atlas

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

Bridging Successor Measure and Online Policy Learning with Flow Matching-Based Representations

Haosen Shi, Jianda Chen, Sinno Jialin Pan

ICLR 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Regularized scalable deep value learning

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 SF2 representation (gamma=0.99) to TD3 and SACTD3 / SAC · 7 DMC tasks, normalized AUCIQM normalized AUCImproved
use full successor horizon gamma=0.99 instead of transition-only gamma=0 (ablation)SF2 · 7 DMC tasksnormalized AUCImproved
SF2 gains for TD3 versus SACSF2 · 7 DMC tasksnormalized AUCMixed
add SF2 compared with successor-feature baselines in sparse reward tasksTD3 · sparse reward DMC tasksreturnImproved
increase EMA coefficient tau toward 1.0 (ablation)SF2 · AcrobotSwingupreturnHurt
vary denoising steps 1 to 4 (ablation)SF2 · AcrobotSwingupreturnNo clear effect
increase denoising steps (ablation)SF2 · AcrobotSwinguprun timeHurt
increase feature size (ablation)SF2 with TD3 · AcrobotSwingupreturnImproved
increase feature size (ablation)SF2 with SAC · AcrobotSwingupreturnNo clear effect
add SF2 to TD3 costs training timeTD3 · AcrobotSwinguprun time: 659 -> 1300 secondsHurt
more gradient steps for the feature updater (ablation)SF2 · DMC IQMIQM scoreImproved
use explicit mixture flow matching instead of standard FM (ablation)flow matching · 1D Gaussian mixture, gamma=0.99Wasserstein distanceImproved

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