混合专家VLA中的涌现组合技能
Emergent Compositional Skills in Mixture-of-Experts VLAs
浏览论文内容
中文总结 AI 辅助
研究从专家演示端到端学习组合机器人策略的问题，用简化MoE动作头训练VLA，发现其能将任务分解，学习到的专家可跨任务重用，MoE在展示专家专业化时与整体基线性能匹配，向模块化、可解释机器人策略迈进。
中文摘要 AI 辅助
我们考虑从专家演示中端到端学习组合机器人策略的问题，无需任何预先指定的任务分解或层次概念。我们探究用简化的混合专家（MoE）动作头训练的VLA是否能涌现出将任务分解为可重复使用、可解释原语的能力。我们发现学习到的专家在不同任务中被大量重用，且始终对应于定性不同的低级行为，这表明路由器隐含地学习执行高级排序，而专家作为组合原语。我们的MoE在展示有意义的专家专业化的同时，与整体基线的任务性能相匹配，朝着仅从数据中涌现的模块化、可解释机器人策略迈进了一步。
英文摘要
We consider the problem of learning compositional robot policies end-to-end from expert demonstrations, without any pre-specified notion of task decomposition or hierarchy. We ask whether a VLA trained with a simplified Mixture-of-Experts (MoE) action head can emergently learn to decompose tasks into reusable, interpretable primitives. We find that learned experts are heavily reused across tasks and consistently correspond to qualitatively distinct low-level behaviors, suggesting that the router implicitly learns to perform high-level sequencing while experts serve as compositional primitives. Our MoE matches the task performance of a monolithic baseline while demonstrating meaningful expert specialization, a step toward modular, interpretable robot policies that emerge from data alone.
