multimodal reward

reinforcement learning (2) return-conditioned policy (1) conversation generation (1) multimodal encoder (1) large language model (1) goal generalization (1) dialogue agent (1) test-time scaling (1) reward decomposition (1) agent system (1)