Guide Your Agent with Adaptive Multimodal Rewards
It addresses the problem of goal misgeneralization in imitation learning for agents, though it appears incremental as it builds on existing multimodal encoders and return-conditioned policies.
This paper tackles the challenge of agent adaptation in unseen environments by proposing the Adaptive Return-conditioned Policy (ARP), which uses natural language task descriptions and pre-trained multimodal encoders to calculate similarity-based rewards, resulting in superior generalization performance compared to existing text-conditioned policies.
Developing an agent capable of adapting to unseen environments remains a difficult challenge in imitation learning. This work presents Adaptive Return-conditioned Policy (ARP), an efficient framework designed to enhance the agent's generalization ability using natural language task descriptions and pre-trained multimodal encoders. Our key idea is to calculate a similarity between visual observations and natural language instructions in the pre-trained multimodal embedding space (such as CLIP) and use it as a reward signal. We then train a return-conditioned policy using expert demonstrations labeled with multimodal rewards. Because the multimodal rewards provide adaptive signals at each timestep, our ARP effectively mitigates the goal misgeneralization. This results in superior generalization performances even when faced with unseen text instructions, compared to existing text-conditioned policies. To improve the quality of rewards, we also introduce a fine-tuning method for pre-trained multimodal encoders, further enhancing the performance. Video demonstrations and source code are available on the project website: \url{https://sites.google.com/view/2023arp}.