AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

PolicyTrim: Boosting Intrinsic Policy Efficiency of Vision-Language-Action Models

Xianghui Wang, Feng Chen, Wenbo Zhang, Hua Yan, Zixuan Wang, Changsheng Li, Yinjie Lei

ECCV 2026 · Vision-language and multimodal › Robot action and navigation · approach: Vision-language-action foundation models · 3 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
apply PolicyTrim (two-stage RL post-training) to the VLApi0.5 · LIBERO-Objectphysical steps: 125.0 -> 64.3Improved
apply PolicyTrimpi0.5 · LIBERO-Objectsuccess rate: 99.1 -> 98.5No clear effect
add reliable chunk extension alone (ablation)pi0.5 · LIBERO-Spatialphysical steps: 108.3 -> 113.8Hurt
add reliable chunk extension alone (ablation)pi0.5 · LIBERO-Spatialend-to-end speedup: 1.0 -> 2.86Improved
add step-saving reward alone (no regularization) (ablation)pi0.5 · LIBERO-Spatialsuccess rate: 97.8 -> 93.7Hurt
add step-saving reward alone (ablation)pi0.5 · LIBERO-Spatialphysical steps: 108.3 -> 81.7Improved
add group-anchored regularization to the step-saving reward (ablation)pi0.5 · LIBERO-Spatialsuccess rate: 93.7 -> 97.5Improved
add group-anchored regularization to the step-saving reward (ablation)pi0.5 · LIBERO-Spatialphysical steps: 81.7 -> 61.6Improved
use dynamic horizon exploration instead of fixed gamma=1.0 (ablation)pi0.5 · LIBERO-Object (H=20)success rate: 94.4 -> 98.8Improved
increase fixed acceptance ratio from 0.25 to 1.0 (ablation)pi0.5 · LIBERO-Object (H=20)success rate: 99.2 -> 94.4Hurt
increase GRPO group size from 8 to 16 (ablation)PolicyTrim pi0.5 · LIBERO-Spatialsuccess rate: 97.8 -> 97.3No clear effect
apply PolicyTrim step reduction only to OpenVLA-OFTOpenVLA-OFT · LIBERO-Spatialend-to-end speedup: 1.0 -> 1.79Improved

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