[Submitted on 27 Dec 2021] · arXiv.org

View PDF HTML (experimental)

Abstract:A central concept in active inference is that the internal states of a physical system parametrise probability measures over states of the external world. These can be seen as an agent's beliefs, expressed as a Bayesian prior or posterior. Here we begin the development of a general theory that would tell us when it is appropriate to interpret states as representing beliefs in this way. We focus on the case in which a system can be interpreted as performing either Bayesian filtering or Bayesian inference. We provide formal definitions of what it means for such an interpretation to exist, using techniques from category theory.
Comments: 11 pages + 26 pages appendix, to be published in the proceedings of the 2nd International Workshop on Active Inference 2021
Subjects: Artificial Intelligence (cs.AI); Neurons and Cognition (q-bio.NC)
ACM classes: I.2.0
Cite as: arXiv:2112.13523 [cs.AI]
  (or arXiv:2112.13523v1 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2112.13523

arXiv-issued DOI via DataCite

Submission history

From: Martin Biehl [view email]
[v1] Mon, 27 Dec 2021 05:44:47 UTC (1,400 KB)

Read the original on arxiv.org ↗