Scaling Empowered Agents

arXiv:2610.13002v1 Announce Type: new Abstract: Empowerment measures the capacity of an agent to actively control its environment. In this work, we explain the connection between skill learning and empowerment. While skill learning is not empowerment, we introduce a simple scalable recipe for empowerment maximization: empowerment corresponds to the number of executable skills from a state, so maximizing empowerment corresponds to reaching states that are good starting points for many skills. Our recipe applies to arbitrary underlying skill learning methods. Whereas most prior work on empowerment looks at tabular settings, this recipe enables us to study empowerment at a scale where we can realize theoretically-predicted but empirically-elusive properties of empowerment (e.g., gathering resources, environment modification, survival). Experiments show that empowerment maximization unlocks interesting emergent properties, such as building a house in Crafter. Website and code at href{https://empowerment-geometry.github.io/scaling-empowerment/}{this https url}.

Related

• The Geometry of Empowerment
• SKILL-DISCO: Distilling and Compiling Agent Traces into Reusable Procedural Skills
• Elastic Horizon: Discovering the Effective Interaction Frontier in Agentic Reinforcement Learning

Source: arXiv cs.AI | 2026-10-12