To function successfully throughout numerous contexts, robots should not solely carry out manipulation duties precisely but additionally adapt how their actions unfold to the duty, object, and interplay setting. We ask whether or not this execution-level variation could be discovered as a reusable behavioral issue shared throughout duties. We current MoMo, a two-stage imitation-learning framework consisting of a spatiotemporal motion tokenizer and a behavior-cloning transformer that takes process and a steady motion-mode situation as inputs. Throughout six real-robot manipulation duties, various this situation produces regular, dynamic, and intermediate behaviors that human raters can distinguish and that differ in joint velocity, acceleration, and end-effector method pitch. On duties demonstrated in just one mode, MoMo transfers the unseen requested mode whereas largely preserving process success. Collectively, these outcomes present proof of compositional generalization to unseen process–mode mixtures and present that movement mode could be reused throughout duties to manage how a manipulation ability is carried out.
