• Home
  • Search
  • Branching time active inference: Empirical study and complexity class analysis
  • Cite Icon10
  • https://doi.org/10.1016/j.neunet.2022.05.010Copy DOI Icon

Branching time active inference: Empirical study and complexity class analysis

Show More
  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Active inference is a state-of-the-art framework for modelling the brain that explains a wide range of mechanisms such as habit formation, dopaminergic discharge and curiosity. However, recent implementations suffer from an exponential (space and time) complexity class when computing the prior over all the possible policies up to the time horizon. Fountas et al. (2020) used Monte Carlo tree search to address this problem, leading to very good results in two different tasks. Additionally, Champion et al. (2021a) proposed a tree search approach based on (temporal) structure learning. This was enabled by the development of a variational message passing approach to active inference (Champion, Bowman, Grześ, 2021), which enables compositional construction of Bayesian networks for active inference. However, this message passing tree search approach, which we call branching-time active inference (BTAI), has never been tested empirically. In this paper, we present an experimental study of the approach (Champion, Grześ, Bowman, 2021) in the context of a maze solving agent. In this context, we show that both improved prior preferences and deeper search help mitigate the vulnerability to local minima. Then, we compare BTAI to standard active inference (AcI) on a graph navigation task. We show that for small graphs, both BTAI and AcI successfully solve the task. For larger graphs, AcI exhibits an exponential (space) complexity class, making the approach intractable. However, BTAI explores the space of policies more efficiently, successfully scaling to larger graphs. Then, BTAI was compared to the POMCP algorithm (Silver and Veness, 2010) on the frozen lake environment. The experiments suggest that BTAI and the POMCP algorithm accumulate a similar amount of reward. Also, we describe when BTAI receives more rewards than the POMCP agent, and when the opposite is true. Finally, we compared BTAI to the approach of Fountas et al. (2020) on the dSprites dataset, and we discussed the pros and cons of each approach.

Similar Papers
  • Research Article
  • Citations24

A Polynomial-Time Algorithm to Obtain State Machine Cover of Live and Safe Petri Nets

  • Feb 14, 2019
  • IEEE Transactions on Systems, Man, and Cybernetics: Systems
  • Andrei G Karatkevich +1
  • Book Chapter

Postoptimal Analysis in Nonserial Dynamic Programming

  • Jan 01, 2008
  • Oleg Shcherbina
  • Research Article
  • Citations114

Sophisticated Inference.

  • Mar 01, 2021
  • Neural Computation
  • Karl Friston +4
  • Conference Article
  • Citations17

Guaranteed-Safe Approximate Reachability via State Dependency-Based Decomposition

  • Jul 01, 2020
  • Anjian Li +1
  • PDF
  • Research Article
  • Citations695

Active inference and learning

  • Jun 29, 2016
  • Neuroscience and biobehavioral reviews
  • Karl Friston +5
  • Conference Article
  • Citations1

MapReduce guided approximate inference over graphical models

  • Dec 01, 2014
  • Ahsanul Haque +3
  • Research Article
  • Citations14

Fast polyhedra abstract domain

  • Jan 01, 2017
  • ACM SIGPLAN Notices
  • Gagandeep Singh +2
  • Research Article
  • Citations2

Learning dynamic cognitive map with autonomous navigation.

  • Dec 11, 2024
  • Frontiers in computational neuroscience
  • Daria De Tinguy +2
  • Research Article
  • Citations115

The Emperor's New Markov Blankets.

  • Oct 22, 2021
  • Behavioral and Brain Sciences
  • Jelle Bruineberg +3
  • Conference Article
  • Citations4

Probability Convergence in a Multithreaded Counting Application

  • Jan 01, 2007
  • Chad Scherrer +5
  • Research Article
  • Citations7

A Comparative Study of Three Heuristic Functions Used to Solve the 8-Puzzle

  • Jan 10, 2016
  • British Journal of Mathematics & Computer Science
  • Anca-Elena Iordan
  • Conference Article
  • Citations7

Structural Bayesian network learning in a biological decision-theoretic intelligent agent and its application to a herding problem in the context of distributed multi-agent systems

  • Oct 07, 2001
  • F Sahin +1
  • Supplementary Content

The emergence of autism in the Bayesian brain

  • Mar 02, 2017
  • Figshare
  • Colin J Palmer
  • Conference Article

Parameterised Counting in Logspace

  • Mar 16, 2021
  • Anselm Haak +3
  • Conference Article
  • Citations9

Non-autoreducible Sets for NEXP

  • Jan 01, 2014
  • DROPS (Schloss Dagstuhl – Leibniz Center for Informatics)
  • Dung Tien Nguyen +1
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.