Inductive biases for decision-making in multi-agent systems

Abstract

Multi-agent systems (MAS) are groups of autonomous agents that interact with each other to achieve common goals. Decision-making in MAS can be a challenging task, as it involves coordinating the actions of multiple agents with potentially conflicting objectives. Inductive biases are a set of assumptions or biases that guide the learning process of an agent, which can be used to improve decision-making in MAS. This paper discusses various types of inductive biases that can be used to improve decision-making in MAS.