Distributed No-Regret Learning in Multi-Agent Systems
This is a tutorial article that synthesizes existing research for the multi-agent systems community, offering no new incremental contributions.
The paper provides an overview of distributed no-regret learning in multi-agent systems, exploring challenges like dynamicity and incomplete feedback, and discusses implications for game modeling and algorithm design without presenting new experimental results.
In this tutorial article, we give an overview of new challenges and representative results on distributed no-regret learning in multi-agent systems modeled as repeated unknown games. Four emerging game characteristics---dynamicity, incomplete and imperfect feedback, bounded rationality, and heterogeneity---that challenge canonical game models are explored. For each of the four characteristics, we illuminate its implications and ramifications in game modeling, notions of regret, feasible game outcomes, and the design and analysis of distributed learning algorithms.