titleCARE: Experience-Guided Atomic Corrective Execution for Vision-Language-Action Policies
authorsJunlan Xiao, Junwei Jiang, Zaibin Zhang, Yifan Wang, Zhongbo Zhang, Huchuan Lu, Lijun Wang
date2026-09-21
venuearXiv (preprint, not peer reviewed)
idarXiv:2609.24118
affiliationDaLian University of Technology
upvotes24 (Hugging Face Daily Papers, 2026-11-01)
linksabs pdf hf code
primerVision-language-action models
abstractVision-Language-Action (VLA) policies achieve strong performance in robotic manipulation but remain brittle once execution deviates from nominal trajectories. We propose CARE (Corrective Atomic Robotic Execution), a framework that improves recovery by learning from failures encountered during execution. Instead of generating corrective data from manually designed or random perturbations, CARE collects failed rollouts, models stage-conditioned post-failure deviations, and uses the resulting empirical distributions to synthesize representative failure states and corrective demonstrations. At inference time, CARE combines stage-wise planning with physically grounded 3D monitoring to trigger atomic adjustments or re-operations while preserving task progress. We further introduce the Failure State Recovery Benchmark (FSR-Bench), which evaluates recovery from intermediate failure states under local deviations and structural anomalies. Experiments across multiple VLA backbones, simulation benchmarks, and real-world dual-arm tasks show consistent improvements, with average task-success gains of 14.5 points in simulation and 15.9 points in the real world. Code, models, and data are available at https://github.com/xiaojunlan/care

cite

@misc{xiao2026care,
  title={CARE: Experience-Guided Atomic Corrective Execution for Vision-Language-Action Policies},
  author={Junlan Xiao and Junwei Jiang and Zaibin Zhang and Yifan Wang and Zhongbo Zhang and Huchuan Lu and Lijun Wang},
  year={2026},
  eprint={2609.24118},
  archivePrefix={arXiv},
  url={https://arxiv.org/abs/2609.24118}
}