Active Predictive Coding: A Unifying Neural Model for Active Perception, Compositional Learning, and Hierarchical Planning
Neural Computation · 12.12.2023 · 36 (1) · 1-32
DOI: 10.1162/neco_a_01627 ↗Kurz gefasst
Fachartikel in der Zeitschrift Neural Computation, online erschienen am 12.12.2023. Verfasst von Rajesh P N Rao und 2 weiteren Autor:innen. Beteiligte Einrichtung: Paul G. Allen School of Computer Science and Engineering and Center for Neurotechnology, University of Washington, Seattle, WA 98195, U.S.A. rao@cs.w…. Hauptthemen laut MEDLINE: Cognition, Deep Learning. Die Arbeit wurde bisher 28-mal zitiert (Crossref) und nennt 77 Literaturangaben.
Abstract (Originalsprache)
There is growing interest in predictive coding as a model of how the brain learns through predictions and prediction errors. Predictive coding models have traditionally focused on sensory coding and perception. Here we introduce active predictive coding (APC) as a unifying model for perception, action, and cognition. The APC model addresses important open problems in cognitive science and AI, including (1) how we learn compositional representations (e.g., part-whole hierarchies for equivariant vision) and (2) how we solve large-scale planning problems, which are hard for traditional reinforcement learning, by composing complex state dynamics and abstract actions from simpler dynamics and primitive actions. By using hypernetworks, self-supervised learning, and reinforcement learning, APC learns hierarchical world models by combining task-invariant state transition networks and task-dependent policy networks at multiple abstraction levels. We illustrate the applicability of the APC model to active visual perception and hierarchical planning. Our results represent, to our knowledge, the first proof-of-concept demonstration of a unified approach to addressing the part-whole learning problem in vision, the nested reference frames learning problem in cognition, and the integrated state-action hierarchy learning problem in reinforcement learning.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- Paul G. Allen School of Computer Science and Engineering and Center for Neurotechnology, University of Washington, Seattle, WA 98195, U.S.A. rao@cs.washington.edu.
- Paul G. Allen School of Computer Science and Engineering and Center for Neurotechnology, University of Washington, Seattle, WA 98195, U.S.A. gklezd@cs.washington.edu.
- Paul G. Allen School of Computer Science and Engineering and Center for Neurotechnology, University of Washington, Seattle, WA 98195, U.S.A. vsathish@cs.washington.edu.
Zitieren
APA
Rao, R. P. N., Gklezakos, D. C., & Sathish, V. (2023). Active Predictive Coding: A Unifying Neural Model for Active Perception, Compositional Learning, and Hierarchical Planning. Neural Computation, 36(1), 1-32. https://doi.org/10.1162/neco_a_01627
Vancouver
Rao RPN, Gklezakos DC, Sathish V. Active Predictive Coding: A Unifying Neural Model for Active Perception, Compositional Learning, and Hierarchical Planning. Neural Computation. 2023;36(1):1-32. doi:10.1162/neco_a_01627
BibTeX
@article{Rao2023,
author = {Rao, Rajesh P N and Gklezakos, Dimitrios C and Sathish, Vishwas},
title = {Active Predictive Coding: A Unifying Neural Model for Active Perception, Compositional Learning, and Hierarchical Planning},
journal = {Neural Computation},
year = {2023},
volume = {36},
number = {1},
pages = {1--32},
doi = {10.1162/neco_a_01627}
}
EUVIMED