Learning to Participate through Trading of Reward Shares
This addresses cooperation challenges in multi-agent systems for AI integration, though it appears incremental as it builds on existing reward-sharing mechanisms.
The paper tackles the problem of enabling autonomous agents to cooperate in social dilemmas by proposing a stock market-inspired method where agents trade reward shares, leading to cooperative policies and role development in tested Markov games.
Enabling autonomous agents to act cooperatively is an important step to integrate artificial intelligence in our daily lives. While some methods seek to stimulate cooperation by letting agents give rewards to others, in this paper we propose a method inspired by the stock market, where agents have the opportunity to participate in other agents' returns by acquiring reward shares. Intuitively, an agent may learn to act according to the common interest when being directly affected by the other agents' rewards. The empirical results of the tested general-sum Markov games show that this mechanism promotes cooperative policies among independently trained agents in social dilemma situations. Moreover, as demonstrated in a temporally and spatially extended domain, participation can lead to the development of roles and the division of subtasks between the agents.