Multi-Agent Empowerment and Emergence of Complex Behavior in Groups

arXiv cs.AI / 4/25/2026

📰 NewsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper studies how intrinsic motivation—specifically empowerment—can emerge from agent-environment interactions, extending the concept to settings with multiple agents.
It proposes a principled multi-agent formulation of empowerment and shows that the resulting quantity can be computed efficiently.
The authors find that empowerment induces distinct, characteristic group-organization behaviors in two different environments.
The demonstrated behaviors occur both in a simple two-agent system coupled by a tendon and in a controllable Vicsek flock, suggesting scalability beyond individual agent control.

Abstract

Intrinsic motivations are receiving increasing attention, i.e. behavioral incentives that are not engineered, but emerge from the interaction of an agent with its surroundings. In this work we study the emergence of behaviors driven by one such incentive, empowerment, specifically in the context of more than one agent. We formulate a principled extension of empowerment to the multi-agent setting, and demonstrate its efficient calculation. We observe that this intrinsic motivation gives rise to characteristic modes of group-organization in two qualitatively distinct environments: a pair of agents coupled by a tendon, and a controllable Vicsek flock. This demonstrates the potential of intrinsic motivations such as empowerment to not just drive behavior for only individual agents but also higher levels of behavioral organization at scale.