外观
论文arXiv 2504.179591 个月前
Civil：因果且直观的视觉模仿学习
原文链接：Civil: causal and intuitive visual imitation learning（arXiv 2504.17959 · 发布日期：2026-09-18）
Civil: causal and intuitive visual imitation learning
提出CIVIL，让人类用标记和语言说明任务相关特征，在仿真与真实实验中改进视觉模仿学习。
核心要点
- 标准视觉模仿学习只观察人类做什么，难以理解决策原因，易受视觉干扰物影响。
- 人类教师为任务相关物体附加标记，并用自然语言描述状态表示，提供增强示范数据。
- CIVIL过滤视觉观测并提取与教师一致的因果特征，再训练基于Transformer的策略。
- 仿真、真实实验和用户研究显示，CIVIL优于先进基线，并减少总训练时间。
验证范围与局限：验证包括仿真、真实机器人实验和用户研究，但摘要未说明具体任务或机器人数量。
英文摘要 (Abstract)
Abstract Today’s robots attempt to learn new tasks by imitating human examples. These robots watch the human complete the task, and then try to match the actions taken by the human expert. However, this standard approach to visual imitation learning is fundamentally limited: the robot observes what the human does, but not why the human chooses those behaviors. Without understanding which features of the system or environment factor into the human’s decisions, robot learners often misinterpret the human’s examples (e.g., the robot incorrectly thinks the human picked up a coffee cup because of the color of clutter in the background). In practice, this results in causal confusion, inefficient learning, and robot policies that fail when the environment changes. We therefore propose a shift in perspective: instead of asking human teachers just to show what actions the robot should take, we also enable humans to intuitively indicate why they made those decisions (i.e., what features are critical for the desired task). Under our paradigm human teachers attach markers to task-relevant objects and use natural language prompts to describe their state representation. Our proposed algorithm, CIVIL, leverages this augmented demonstration data to filter the robot’s visual observations and extract a feature representation that aligns with the human teacher. CIVIL then applies these causal features to train a transformer-based policy that — when tested on the robot — is able to emulate human behaviors without being confused by visual distractors or irrelevant items. We show that, by harnessing and combining existing language-grounding, visual-masking, and policy-learning techniques, CIVIL realizes this richer teaching paradigm and incorporates human-provided task relevance directly into imitation learning. Our simulations and real-world experiments demonstrate that robots trained with CIVIL learn both what actions to take and why to take those actions, resulting in better performance than state-of-the-art baselines. From the human’s perspective, our user study reveals that this new training paradigm actually reduces the total time required for the robot to learn the task, and also improves the robot’s performance in previously unseen scenarios. See videos at our project website: https://civil2025.github.io .
作者：Yinlong Dai, Robert Ramirez Sanchez, Ryan Jeronimus, Shahabedin Sagheb, Cara M. Nunez, Heramb Nemlekar, Dylan P. Losey
来源
本页事实对应的公开页面，以原文为准。