Learning to Coordinate without Communication under Incomplete Information
This addresses the challenge of coordination in AI for scenarios where communication is infeasible, offering a practical solution with incremental improvements over existing non-communicative methods.
The paper tackles the problem of achieving coordination in cooperative games under incomplete information without communication, by enabling agents to interpret action sequences as intent signals, and shows that this method significantly outperforms uncoordinated strategies and closely matches direct communication performance.
Achieving seamless coordination in cooperative games is a crucial challenge in artificial intelligence, particularly when players operate under incomplete information. While communication helps, it is not always feasible. In this paper, we explore how effective coordination can be achieved without verbal communication, relying solely on observing each other's actions. Our method enables an agent to develop a strategy by interpreting its partner's action sequences as intent signals, constructing a finite-state transducer built from deterministic finite automata, one for each possible action the agent can take. Experiments show that these strategies significantly outperform uncoordinated ones and closely match the performance of coordinating via direct communication.