Abstract
In this study the problem of Federated Learning (FL) is explored under a newperspective by utilizing the Deep Equilibrium (DEQ) models instead ofconventional deep learning networks. We claim that incorporating DEQ modelsinto the federated learning framework naturally addresses several open problemsin FL, such as the communication overhead due to the sharing large models andthe ability to incorporate heterogeneous edge devices with significantlydifferent computation capabilities. Additionally, a weighted average fusionrule is proposed at the server-side of the FL framework to account for thedifferent qualities of models from heterogeneous edge devices. To the best ofour knowledge, this study is the first to establish a connection between DEQmodels and federated learning, contributing to the development of an efficientand effective FL framework. Finally, promising initial experimental results arepresented, demonstrating the potential of this approach in addressingchallenges of FL.