Homer Rich Walke

robot manipulation (3) domain generalization (2) offline reinforcement learning (2) imitation learning (2) goal-conditioned policy (2) representation learning (2) generalization ability (1) policy training (1) vision-language model (1) reward shaping (1) embedding alignment (1) subgoal planning (1) visual input (1) skill adaptation (1) language interface (1) offline dataset (1) prior datum (1) transformer-based policy (1) policy finetuning (1) demonstration datum (1)