cs.RONov 11, 2026

Recover, Don't Just Continue: Learning from Feedback for Physical Failure Understanding and Correction

Authors: Miao Pan, Kaixiang Yao, Xiaohe Zhou, Xu Wang, Qinfeng Li, Wenqi Zhang, Xuhong Zhang

Organizations: Zhejiang University

Abstract

Multimodal large language models are regarded as the cognitive foundation of embodied intelligence, but planning from instructions alone cannot make them reliable. During execution, actions deviate from intended effects, constrained by the embodiment and environment. An embodied brain must understand why execution deviates, identify what can be reliably realized, and revise decisions using interaction experience. These correspond to causal failure localization, capability-constrained adaptation, and experience-driven revision. Current embodied foundation models largely lack these capabilities. To address this, we introduce Embodied-Corrector. We construct failure-recovery data by injecting perturbations at subtask boundaries of successful trajectories, replaying affected subtasks, and keeping failures that are observable from public evidence and recoverable from the current state. We align actions, consequences, failure causes, and recoveries into Judge-Cause-Sub-Fine supervision. Training uses evaluation-first SFT: the model learns Judge and Cause first, then Sub and Fine conditioned on accurate prefixes. Since deployment uses self-generated prefixes, evaluation errors propagate into recovery planning, and response-level rewards are too coarse to locate affected planning tokens. We therefore use OPD to internalize the planning distribution under accurate prefixes into that under self-generated ones: the model generates plans from self-generated prefixes, while a contrastive forward pass with accurate prefixes compares token probabilities to reinforce or suppress planning tokens. At deployment, this branch is removed. Experiments show stronger failure analysis and corrective planning: with a frozen VLA, composite task success improves by 6.1 percentage points; with fixed GPT-5.5 execution, manipulation and navigation tasks improve by 13.9 percentage points.

Figures & tables

Appendix figures & tables5 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Spotter: Let the Embodied Model Lead, and the VLM Reflect for It

    Sep 29, 2026Long Li, Qichao Zhao, Yue Yang +6Robot Failure RecoveryVision-Language-Action Models

  2. Replan for Error Recovery: Training Embodied VLMs through Verifiable Interaction with Action-Grouped Turn-GSPO

    Oct 19, 2026Zhuoran Wang, Keqing Fan, Yang Li +6VLMs for RoboticsRobot Task Planning

  3. From Failures to Supervision: DynamicEnvPlan for Robust Long-Horizon Embodied Planning

    Aug 1, 2026Hao Yuan, Yuxin Wang, Lei Ji +1Robot Failure RecoveryRobot Task Planning