Competing Adaptive Networks

TL;DR

This paper introduces a decentralized competitive learning algorithm for adaptive networks, enabling teams of agents to compete and cooperate locally to achieve global objectives, with applications in training generative adversarial networks.

Contribution

It presents a novel algorithm for decentralized competition among adaptive agent teams, extending traditional cooperative learning frameworks.

Findings

Algorithm effectively models competitive interactions in adaptive networks.

Application demonstrated in decentralized training of GANs.

Analysis shows convergence properties of the proposed method.

Abstract

Adaptive networks have the capability to pursue solutions of global stochastic optimization problems by relying only on local interactions within neighborhoods. The diffusion of information through repeated interactions allows for globally optimal behavior, without the need for central coordination. Most existing strategies are developed for cooperative learning settings, where the objective of the network is common to all agents. We consider in this work a team setting, where a subset of the agents form a team with a common goal while competing with the remainder of the network. We develop an algorithm for decentralized competition among teams of adaptive agents, analyze its dynamics and present an application in the decentralized training of generative adversarial neural networks.