Automatic Curriculum Design for Zero-Shot Human-AI Coordination
This work addresses the challenge of generalization in zero-shot human-AI coordination for real-world applications with unpredictable environmental changes, representing an incremental improvement over previous methods.
The paper tackles the problem of training AI agents to coordinate with humans in unseen environments without human data, by extending a multi-agent unsupervised environment design approach and proposing a new utility function and co-player sampling method, achieving high performance in human-AI coordination tasks in the Overcooked-AI environment.
Zero-shot human-AI coordination is the training of an ego-agent to coordinate with humans without human data. Most studies on zero-shot human-AI coordination have focused on enhancing the ego-agent's coordination ability in a given environment without considering the issue of generalization to unseen environments. Real-world applications of zero-shot human-AI coordination should consider unpredictable environmental changes and the varying coordination ability of co-players depending on the environment. Previously, the multi-agent UED (Unsupervised Environment Design) approach has investigated these challenges by jointly considering environmental changes and co-player policy in competitive two-player AI-AI scenarios. In this paper, our study extends a multi-agent UED approach to zero-shot human-AI coordination. We propose a utility function and co-player sampling for a zero-shot human-AI coordination setting that helps train the ego-agent to coordinate with humans more effectively than a previous multi-agent UED approach. The zero-shot human-AI coordination performance was evaluated in the Overcooked-AI environment, using human proxy agents and real humans. Our method outperforms other baseline models and achieves high performance in human-AI coordination tasks in unseen environments. The source code is available at https://github.com/Uwonsang/ACD_Human-AI