Attracting Sets in Perceptual Networks

TL;DR

This paper introduces a model for perceptual networks that optimize mutual information between inputs and attractors using genetic algorithms, inspired by the interface theory of perception, and provides an accompanying Python implementation.

Contribution

It presents a simple optimization method for perceptual networks based on mutual information and offers a practical Python tool for experimentation.

Findings

Optimizes mutual information in noisy networks

Models nodes based on interface theory of perception

Provides an accessible Python implementation

Abstract

This document gives a specification for the model used in [1]. It presents a simple way of optimizing mutual information between some input and the attractors of a (noisy) network, using a genetic algorithm. The nodes of this network are modeled as simplified versions of the structures described in the "interface theory of perception" [2]. Accordingly, the system is referred to as a "perceptual network". The present paper is an edited version of technical parts of [1] and serves as accompanying text for the Python implementation PerceptualNetworks, freely available under [3]. 1. Prentner, R., and Fields, C.. Using AI methods to Evaluate a Minimal Model for Perception. OpenPhilosophy 2019, 2, 503-524. 2. Hoffman, D. D., Prakash, C., and Singh, M.. The Interface Theory of Perception. Psychonomic Bulletin and Review 2015, 22, 1480-1506. 3. Prentner, R.. PerceptualNetworks. https://github.com/RobertPrentner/PerceptualNetworks. (accessed September 17 2020)