Behavior cloning for robot manipulation relies on expert demonstrations. However, for tasks that require dynamic stability, precise contact timing, or dexterous coordination, human operators may find it hard or even impossible to collect data. We study this infeasible-demonstration regime and propose GLIDE: Guardrails...
ProgSpec, a benchmark that evaluates final outputs against requirements explicitly stated in the initial request and additional requirements supported by the available task evidence, is introduced and RepoMAS, an issue-driven multi-agent framework inspired by open-source project management, is proposed.
Yu-Chen Song, An-Dong Chen, Wen-Xin Zhu et al.· 0 citations
DeVA is introduced, a Decoupled Video-Action model with specialized video and action experts, multi-level feature transfer, and physically salient guidance, enabling rich information exchange while making policy learning more tractable.
