Dafeng Chi

transformer architecture (1) reinforcement learning (1) offline reinforcement learning (1) imitation learning (1) policy learning (1) mutual information (1) variational autoencoder (1) distributional shift (1) language generation (1) latent code (1) dialogue system (1) joint training (1) user simulator (1) generative diversity (1) embodied instruction following (1) trajectory attention (1) compound exposure bia (1) contrastive dynamics learning (1)