返回全部动态
REFACTOR-VLA：无监督学习类型化电机程序库
原标题：REFACTOR-VLA: Unsupervised Library Learning of Typed Motor Programs
AI 摘要
Apple 研究团队提出 REFACTOR-VLA 系统，通过“唤醒/睡眠”架构学习可复用的机器人技能，利用行为等价核和潜在世界模型进行聚类，并生成类型化 lambda 程序。在 LIBERO 基准测试中发现，增大世界模型规模反而降低性能，而添加 InfoNCE 对比损失能显著提升技能聚类质量。
以上摘要由 AI 生成，可能存在误差。事实请以原文为准。
正文节选
Most current vision-language-action (VLA) models—such as OpenVLA, π0, RT-2, and RDT-1B—are “monolithic.” This means they generate raw motor commands or very short sequences of actions, without organizing behaviors into reusable, well-defined abstractions. As a result, these models perform poorly on long-horizon (multi-step) tasks, and it’s difficult to interpret what they have learned. Existing approaches for discovering skills often avoid the core problem of deciding when two action sequences a
发布时间：2026-09-02 08:00
抓取时间：2026-09-02 23:58
来源机构：Apple