发表机构
Princeton University; UC Berkeley(普林斯顿大学; 加州大学伯克利分校)
机构由 AI 辅助整理,请以论文原文为准。AI 中文总结
本研究通过将赋权最大化与技能学习相结合,提出新的几何框架,解答了赋权与结构性中心性的长期疑问,并区分信息与奖励几何,为可扩展赋权方法奠定理论基础。
AI 中文摘要
赋权(Empowerment)衡量的是智能体主动控制其环境的能力。虽然作为一个信息论量在概念上颇具吸引力,但赋权与那些能够提供广泛未来结果访问权限的结构性中心状态之间的联系仍是一个悬而未决的问题。在本工作中,我们将赋权最大化方法与技能学习方法联系起来,为解释和分析赋权提供了新的几何结构。我们的分析回答了关于赋权与结构性中心性之间联系的长期悬而未决的问题。我们的分析还揭示了信息几何与奖励几何之间的区别,强调了构建可扩展的赋权最大化方法的重要理论意义。网站和代码可在该 https URL 找到。
英文摘要
Empowerment captures the capacity for an agent to actively control its environment. While conceptually appealing as an information-theoretic quantity, the connection between empowerment and structurally central states that provide broad access to future outcomes has remained an open question. In this work, we link empowerment maximization and skill-learning methods to provide new geometries for interpreting and analyzing empowerment. Our analyses answer longstanding open questions on the connections between empowerment and structural centrality. Our analyses also reveal distinctions between information and reward geometries, highlighting important theoretical implications to build scalable empowerment-maximization methods. Website and code can be found at https://empowerment-geometry.github.io/.
Comments34 pages, 12 figures