AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Compose Your Policies! Improving Diffusion-based or Flow-based Robot Policies via Test-time Distribution-level Composition

Jiahang Cao, Yize Huang, Hanzhong Guo, Qiang Zhang, Rui Zhang, Weijian Mai, Mu Nan, Jiaxu Wang, Hao Cheng, Jingkai SUN, Gang Han, Wen Zhao and 6 more

ICLR 2026 · Reinforcement learning › Visual and representation RL · approach: Skill learning and diffusion policies · 8 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
compose DP with MP by convex score combinationDiffusion Policy · Robomimic+PushT averagesuccess rate (%): 39.19 -> 41.41Improved
compose Florence-Policy-D with DPFlorence Policy-D · Robomimic+PushT averagesuccess rate (%): 61.25 -> 66.76Improved
compose pi0 with Flow Policypi0 · Robomimic+PushT averagesuccess rate (%): 86.42 -> 88.94Improved
compose Florence-Policy-F with Flow PolicyFlorence Policy-F · Robomimic+PushT averagesuccess rate (%): 78.84 -> 86.39Improved
compose RDT with DP point cloudDP (point cloud) · RoboTwin 2.0 averagesuccess rate: 0.65 -> 0.72Improved
compose DP image with DP point cloudDP (point cloud) · RoboTwin 2.0 averagesuccess rate: 0.65 -> 0.70Improved
tune convex weight when both policies are moderately accurate (ablation)GPC DP img + DP pcd · Empty Cup Placesuccess rate: 0.62 -> 0.86Improved
compose with a much weaker policy (ablation)GPC DP img + DP pcd · Pick Apple Messysuccess rate: 0.26 -> 0.25No clear effect
compose with a much weaker policy using equal weights (ablation)GPC DP img + DP pcd · Block Hammer Beatsuccess rate: 0.76 -> 0.12Hurt
give stronger policy larger weight (ablation)GPC DP img + DP pcd · Dual Bottles Pick (Easy)success rate: 0.77 -> 0.85Improved
add a third base policy (FP+FP-F+pi0) (ablation)GPC · Robomimic Cansuccess rate: 0.99 -> 1.00Improved
compose DP and MP with logical AND superposition (ablation)GPC · Robomimic+PushT averagesuccess rate (%): +25.73%; SR 64.92Improved

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