開放世界技能發現於未分段示範中

Open-World Skill Discovery from Unsegmented Demonstrations

March 11, 2025

作者: Jingwen Deng, Zihao Wang, Shaofei Cai, Anji Liu, Yitao Liang

cs.AI

摘要

在開放世界環境中學習技能對於開發能夠結合基本技能處理多種任務的智能體至關重要。線上示範視頻通常較長且未經分段，這使得它們難以被分割並標註技能識別符。與依賴序列採樣或人工標註的現有方法不同，我們開發了一種基於自監督學習的方法，將這些長視頻分割成一系列語義感知且技能一致的片段。受人類認知事件分割理論的啟發，我們引入了技能邊界檢測（Skill Boundary Detection, SBD），這是一種無需註釋的時間視頻分割算法。SBD通過利用預訓練的無條件動作預測模型的預測誤差來檢測視頻中的技能邊界。該方法基於這樣一個假設：預測誤差的顯著增加表明正在執行的技能發生了轉變。我們在《我的世界》（Minecraft）這一擁有豐富線上遊戲視頻的開放世界模擬器中評估了我們的方法。我們通過SBD生成的片段，將條件策略在短期原子技能任務上的平均性能提升了63.7%和52.1%，並將相應的分層智能體在長期任務上的性能提升了11.3%和20.8%。我們的方法能夠利用多樣的YouTube視頻來訓練遵循指令的智能體。項目頁面可在https://craftjarvis.github.io/SkillDiscovery找到。

English

Learning skills in open-world environments is essential for developing agents capable of handling a variety of tasks by combining basic skills. Online demonstration videos are typically long but unsegmented, making them difficult to segment and label with skill identifiers. Unlike existing methods that rely on sequence sampling or human labeling, we have developed a self-supervised learning-based approach to segment these long videos into a series of semantic-aware and skill-consistent segments. Drawing inspiration from human cognitive event segmentation theory, we introduce Skill Boundary Detection (SBD), an annotation-free temporal video segmentation algorithm. SBD detects skill boundaries in a video by leveraging prediction errors from a pretrained unconditional action-prediction model. This approach is based on the assumption that a significant increase in prediction error indicates a shift in the skill being executed. We evaluated our method in Minecraft, a rich open-world simulator with extensive gameplay videos available online. Our SBD-generated segments improved the average performance of conditioned policies by 63.7% and 52.1% on short-term atomic skill tasks, and their corresponding hierarchical agents by 11.3% and 20.8% on long-horizon tasks. Our method can leverage the diverse YouTube videos to train instruction-following agents. The project page can be found in https://craftjarvis.github.io/SkillDiscovery.