论文阅读“MV-UMI: A Scalable Multi-View Interface for Cross-Embodiment Learning“摘要Recent advances in imitation learning have shown great promise for developing robust robot manipulation policies from demonstrations. However, this promise is contingent on the availability of diverse, high-quality datasets, which are not only challengi