Emergent Communication under Competition
This work addresses a gap in emergent communication for competitive multi-agent systems, overturning previous negative results and inspiring future research in this domain.
The paper tackles the problem of learning communication between competitive agents in reinforcement learning, showing that communication can emerge in partially-competitive scenarios using standard algorithms, with key findings including that communication is proportional to cooperation and requires mutual benefit.
The literature in modern machine learning has only negative results for learning to communicate between competitive agents using standard RL. We introduce a modified sender-receiver game to study the spectrum of partially-competitive scenarios and show communication can indeed emerge in a competitive setting. We empirically demonstrate three key takeaways for future research. First, we show that communication is proportional to cooperation, and it can occur for partially competitive scenarios using standard learning algorithms. Second, we highlight the difference between communication and manipulation and extend previous metrics of communication to the competitive case. Third, we investigate the negotiation game where previous work failed to learn communication between independent agents (Cao et al., 2018). We show that, in this setting, both agents must benefit from communication for it to emerge; and, with a slight modification to the game, we demonstrate successful communication between competitive agents. We hope this work overturns misconceptions and inspires more research in competitive emergent communication.