首页|Eindhoven University of Technology Reports Findings in Neural Computation (Reali zing Synthetic Active Inference Agents, Part II: Variational Message Updates)
Eindhoven University of Technology Reports Findings in Neural Computation (Reali zing Synthetic Active Inference Agents, Part II: Variational Message Updates)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
New research on Computation - Neural C omputation is the subject of a report. According to news reporting originating f rom Eindhoven, Netherlands, by NewsRx correspondents, research stated, "The free energy principle (FEP) describes (biological) agents as minimizing a variationa l free energy (FE) with respect to a generative model of their environment. Acti ve inference (AIF) is a corollary of the FEP that describes how agents explore a nd exploit their environment by minimizing an expected FE objective." Our news editors obtained a quote from the research from the Eindhoven Universit y of Technology, "In two related papers, we describe a scalable, epistemic appro ach to synthetic AIF by message passing on free-form Forney-style factor graphs (FFGs). A companion paper (part I of this article; Koudahl et al., 2023) introdu ces a constrained FFG (CFFG) notation that visually represents (generalized) FE objectives for AIF. This article (part II) derives message-passing algorithms th at minimize (generalized) FE objectives on a CFFG by variational calculus. A com parison between simulated Bethe and generalized FE agents illustrates how the me ssage-passing approach to synthetic AIF induces epistemic behavior on a T-maze n avigation task. Extension of the T-maze simulation to learning goal statistics a nd a multiagent bargaining setting illustrate how this approach encourages reuse of nodes and updates in alternative settings."
NETL
NSTL
