arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~

面向世界行动模型的完成感知引导

Completion Aware Guidance for World Action Models

Seungyeon Kim, Junhoo Lee, Baekseung Kim, Minkyu Kim, Nojun Kwak

arXiv 2610.01559首次发表:更新:

发表机构

Seoul National University; KAIST(首尔大学; 韩国科学技术院)

机构由 AI 辅助整理,请以论文原文为准。

AI 中文总结

针对世界行动模型因短块控制导致任务不完整想象的问题,提出无需训练的完成感知引导方法,在多个基准上将成功率提升至70%和75%,并大幅减少不完整想象。

AI 中文摘要

世界行动模型(WAMs)预测视觉未来和机器人动作,但它们仍然容易受到任务不完整想象的影响,在这种想象中,看似合理且与动作一致的预测忽略了任务完成所需的转换。在本文中,我们表明这种失败并非世界模型主干所固有,而是在适应短块控制时出现的,这种控制可能反复偏好合理的局部延续而非完成任务所需的转换。为了解决这个问题,我们引入了完成感知引导(CAG),一种无需训练的采样方法,引导生成朝向任务完成。在代表性的WAMs上,CAG在RoboTwin 2.0子集上将成功率从64%提高到70%,在零样本模拟中从69%提高到75%,同时将任务不完整想象从79%降低到40%。

英文摘要

World Action Models (WAMs) predict visual futures and robot actions, yet they remain susceptible to task-incomplete imagination, where plausible, action-consistent predictions omit the transition needed for task completion. In this paper, we show that this failure is not inherent to the world model backbone, but emerges when adapted for short-chunk control, which can repeatedly favor plausible local continuations over task-completing transitions. To address this, we introduce Completion Aware Guidance (CAG), a training-free sampling method that guides generation toward task completion. Across representative WAMs, CAG improves success from 64% to 70% on a RoboTwin 2.0 subset and from 69% to 75% in zero-shot simulation, while reducing task-incomplete imagination from 79% to 40%.

论文原文

arXiv 摘要页 · PDF 原文 · HTML 原文

↑