AI Paper Atlas

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

SRPO: Self-Referential Policy Optimization for Vision-Language-Action Models

Senyu Fei, Siyin Wang, Li Ji, Ao Li, Shiduo Zhang, Liming Liu, Jinlong Hou, Jingjing Gong, Xianzhong Zhao, Xipeng Qiu

CVPR 2026 · Vision-language and multimodal › Robot action and navigation · approach: Vision-language-action foundation models · 27 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
add online SRPO post-training to one-shot SFT policyOpenVLA*-One · LIBERO averagesuccess rate (%): 48.9 -> 99.2Improved
use online SRPO instead of offline SRPO (ablation)OpenVLA*-One · LIBERO averagesuccess rate (%): 92.5 -> 99.2Improved
add self-referential progress-wise rewards instead of sparse outcome rewards (vs RLinf/SimpleVLA-RL)GRPO-style VLA RL · LIBERO averagesuccess rate (%): 98.0 -> 99.2Improved
replace latent world representation reward with pixel-level reward (ablation)SRPO · progress reward benchmarkSpearman correlation: 0.998 -> 0.125Hurt
replace V-JEPA 2 latent world representation with ImageBind embeddings (ablation)SRPO · RL training success ratesuccess rate: plateaus around 85%Hurt
replace latent world representation with ImageBind on reward benchmark (ablation)SRPO · progress reward benchmarkSMD: 188.799 -> 18.111Hurt
add SRPO instead of GRPO outcome-only rewardVLA RL · LIBERO-Long and LIBERO-Object training curvestraining efficiencyImproved
apply SRPO online RL on one-shot SFT policyOpenVLA*-One · LIBERO-Plus (zero-shot perturbations)total success rate (%): 19.4 -> 59.6Improved
apply SRPO on one-shot SFT policy, compared to OpenVLA-OFT zero-shotOpenVLA*-One · LIBERO-Plus total (zero-shot)success rate (%): 69.6 vs 59.6Hurt
train SRPO in LIBERO-Plus perturbed environmentsOpenVLA*-One · LIBERO-Plus total (+Data)success rate (%): 30.7 -> 82.1Improved
apply progress reward shaping offline to real-world pi0 and pi0-FASTpi0 / pi0-FAST · five real-world X-ARM 7 tasksaverage success rate: +66.8% and +86.7%Improved

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