Wilka Carvalho

deep reinforcement learning (1) reinforcement learning (1) self-supervised learning (1) compositional generalization (1) zero-shot generalization (1) successor feature (1) sparse reward (1) goal-conditioned reinforcement learning (1) reinforcement learning agent (1) object representation learning (1) 3d environment (1) object dynamics model (1) language instruction (1) action decomposition (1) embodied environment (1) graph-based planning (1) state-space exploration (1) landmark-based graph (1)