Interacting Dreaming Neural Networks

TL;DR

This paper analyzes the complex interactions of neural network agents with shared memories and different dreaming phases, revealing diverse equilibrium states and phases through replica methods and simulations.

Contribution

It introduces a novel model of interacting neural networks with dreaming phases, providing a detailed phase diagram and analytical insights into their collective behavior.

Findings

Identification of multiple equilibrium phases including student-professor and mutualism.

Discovery of a reinforced delusion phase where agents concur without memory overlaps.

Analytical and simulation results confirming the rich phase structure.

Abstract

We study the interaction of agents, where each one consists of an associative memory neural network trained with the same memory patterns and possibly different reinforcement-unlearning dreaming periods. Using replica methods, we obtain the rich equilibrium phase diagram of the coupled agents. It shows phases such as the student-professor phase, where only one network benefits from the interaction while the other is unaffected; a mutualism phase, where both benefit; an indifferent phase and an insufficient phase, where neither are benefited nor impaired; a phase of amensalism where one is unchanged and the other is damaged. In addition to the paramagnetic and spin glass phases, there is also one we call the reinforced delusion phase, where agents concur without having finite overlaps with memory patterns. For zero coupling constant, the model becomes the reinforcement and removal dreaming model, which without dreaming is the Hopfield model. For finite coupling and a single memory pattern, it becomes a Mattis version of the Ashkin-Teller model. In addition to the analytical results, we have explored the model with Monte Carlo simulations.