Nonmodular architectures of cognitive systems based on active inference

Baltieri, Manuel and Buckley, Christopher L (2019) Nonmodular architectures of cognitive systems based on active inference. 2019 International Joint Conference on Neural Networks (IJCNN), Budapest, Hungary, 14-19 July 2019. Published in: 2019 International Joint Conference on Neural Networks (IJCNN): 2019 Proceedings. Institute of Electrical and Electronics Engineers ISSN 2161-4407 ISBN 978-1-7281-1985-4

[img] PDF - Accepted Version
Download (1MB)


In psychology and neuroscience it is common to describe cognitive systems as input/output devices where perceptual and motor functions are implemented in a purely feedforward, open-loop fashion. On this view, perception and action are often seen as encapsulated modules with limited interaction between them. While embodied and enactive approaches to cognitive science have challenged the idealisation of the brain as an input/output device, we argue that even the more recent attempts to model systems using closed-loop architectures still heavily rely on a strong separation between motor and perceptual functions. Previously, we have suggested that the mainstream notion of modularity strongly resonates with the separation principle of control theory. In this work we present a minimal model of a sensorimotor loop implementing an architecture based on the separation principle. We link this to popular formulations of perception and action in the cognitive sciences, and show its limitations when, for instance, external forces are not modelled by an agent. These forces can be seen as variables that an agent cannot directly control, i.e., a perturbation from the environment or an interference caused by other agents. As an alternative approach inspired by embodied cognitive science, we then propose a nonmodular architecture based on active inference. We demonstrate the robustness of this architecture to unknown external inputs and show that the mechanism with which this is achieved in linear models is equivalent to integral control.

Item Type: Conference Proceedings
Keywords: modularity, separation principle, active inference, Bayesian inference, optimal control, embodied cognition
Schools and Departments: School of Engineering and Informatics > Informatics
Research Centres and Groups: Evolutionary and Adaptive Systems Research Group
Related URLs:
Depositing User: Lucy Arnold
Date Deposited: 30 Apr 2019 14:19
Last Modified: 18 Nov 2019 14:32

View download statistics for this item

📧 Request an update
Project NameSussex Project NumberFunderFunder Ref
Distributed neural processing of self-generated visual input in a vertebrate brainG2144BBSRC-BIOTECHNOLOGY & BIOLOGICAL SCIENCES RESEARCH COUNCILBB/P022197/1