外观
论文arXiv 2610.0913422 天前
世界模型梦见成功：诊断并修复机器人世界模型中的失败不敏感性
原文链接：World Models Dream of Success: Diagnosing and Repairing Failure Insensitivity in Robot World Models（arXiv 2610.09134 · 发布日期：2026-10-06）
World Models Dream of Success: Diagnosing and Repairing Failure Insensitivity in Robot World Models
针对机器人世界模型难区分成功与失败的问题，CureWM用执行验证的反事实失败微调，并在LIBERO与实体机械臂上验证。
核心要点
- 动机是机器人世界模型需区分成功动作与失败，但已发布检查点对动作变化敏感性弱。
- CureWM从成功演示构造不同严重度的替代动作，经仿真或硬件执行验证后用于微调。
- 在484个LIBERO保留失败反事实上，乐观预测由80%降至四个模型的30--43%。
- 实体机械臂评估中，成功式失败预测由90%降至33%；代码已开源。
验证范围与局限：验证范围包括两个架构家族的四个已发布检查点、LIBERO失败反事实和两次实体机械臂评估。
英文摘要 (Abstract)
Robot world models support policy evaluation, planning, and synthetic data generation, but these applications require predictions that distinguish successful actions from failures. Across four released checkpoints from two architecture families, we observe weak sensitivity to action changes and success-like predictions on verified failures. Although recent work incorporates failures into model training, which data can repair released checkpoints without changing their architecture or training objective still remains underexplored. To this end, we introduce CureWM, which constructs alternative actions from successful demonstrations across a severity grid, verifies their outcomes through execution in simulation or on hardware, and fine-tunes released models on the resulting failures and surviving successes alongside nominal demonstrations. This construction provides controlled action contrasts from shared starting contexts. On 484 held-out LIBERO failure counterfactuals, optimism falls from 80% after fine-tuning on the official data to 30--43% across four independently fine-tuned CureWM models (38% mean). In two separate evaluations on a physical robot arm, failure predictions scored as success-like by a latent-distance diagnostic decrease from 90% after fine-tuning on successful demonstrations alone to 33% with CureWM. With failure counts per task, successful replay data, and training budget matched, counterfactual failures yield a success--failure value gap of 0.124, compared with 0.014 for freshly collected on-policy failures. These findings support execution-verified counterfactual replay for post-hoc repair and show why reduced optimism must be evaluated alongside success--failure discrimination. Code is available at https://github.com/jiuyixu25/CureWM.
作者：Jiuyi Xu, Xiao Hu, Meida Chen, Peng Gao, Yang Ye, Yangming Shi
来源
本页事实对应的公开页面，以原文为准。