AI Paper Atlas

Research map / Reinforcement learning / Offline RL

Speedup Patch: Learning a Plug-and-Play Policy to Accelerate Embodied Manipulation

Zhichao Wu, Junyin Ye, Zhilong Zhang, Yihao Sun, Haoxin Lin, Jiaheng Luo, Haoxiang Ren, Lei Yuan, Yang Yu

ICML 2026 · Reinforcement learning › Offline RL · approach: Diffusion models for offline decision making · 0 citations (OpenAlex)

Read the paper

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 SuP scheduler instead of fixed ds2 downsamplingACT · BiGym averagesuccess rate: 0.61 -> 0.67Improved
add SuP scheduler instead of fixed ds2 downsamplingDP · BiGym averagesuccess rate: 0.40 -> 0.51Improved
add SuP scheduler instead of fixed ds2 downsamplingpi0.5 · Libero averagesuccess rate: 0.928 -> 0.973Improved
add SuP scheduler instead of fixed ds2 downsamplingVLA-Adapter · Libero averagesuccess rate: 0.862 -> 0.937Improved
use SuP instead of ds3 downsamplingpi0.5 · Real-world averagesuccess rate: 0.356 -> 0.611Improved
use SuP instead of DemoSpeedup retrainingpi0.5 · Real-world averagespeedup: 2.07x -> 2.17xImproved
add SuP scheduler to base policypi0.5 · Libero averagesuccess rate: 0.969 -> 0.973No clear effect
add IQL-trained scheduler instead of greedy MPC (ablation)SuP · BiGym Sandwich Removesuccess rate: 0.54 -> 0.59Improved
add IQL-trained scheduler instead of greedy MPC (eps 0.02) (ablation)SuP · BiGym Sandwich Removesuccess rate: 0.35 -> 0.46Improved
raise deviation threshold from 0.015 to 0.02 (ablation)SuP · BiGym Sandwich Removesuccess rate: 0.64 -> 0.46Hurt
lower deviation threshold from 0.015 to 0.01 (ablation)SuP · BiGym Sandwich Removesuccess rate: 0.64 -> 0.59Hurt
add gripper action compensation to ds2 downsampling (ablation)pi0.5 · Libero averagesuccess rate: 0.842 -> 0.928Improved

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