# CARE：面向视觉-语言-动作策略的经验引导原子纠错执行

- 来源：HuggingFace Daily Papers（社区热门论文）
- 发布时间：2026-09-21 08:00
- AIHOT 分数：39
- AIHOT 链接：https://aihot.news/items/cmucc8aul02q9roij58nn9iox
- 原文链接：https://arxiv.org/abs/2609.24118

## AI 摘要

研究者提出 CARE 框架，通过收集失败 rollout、建模分阶段失败后偏差分布，合成代表性失败状态与纠错示范，提升 VLA 策略偏离标称轨迹后的恢复能力。推理时 CARE 结合分阶段规划与物理落地的 3D 监控，触发原子级调整或重操作。

## 正文

Vision-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