IPLS : A Framework for Decentralized Federated Learning
This framework addresses the problem of centralized reliance in federated learning for mobile device users concerned about data privacy.
This paper introduces IPLS, a fully decentralized federated learning framework that uses IPFS. It allows any party to initiate or join ML model training, scales with participants, is robust to connectivity issues, and converges to centralized FL accuracy with less than one per thousand drop.
The proliferation of resourceful mobile devices that store rich, multidimensional and privacy-sensitive user data motivate the design of federated learning (FL), a machine-learning (ML) paradigm that enables mobile devices to produce an ML model without sharing their data. However, the majority of the existing FL frameworks rely on centralized entities. In this work, we introduce IPLS, a fully decentralized federated learning framework that is partially based on the interplanetary file system (IPFS). By using IPLS and connecting into the corresponding private IPFS network, any party can initiate the training process of an ML model or join an ongoing training process that has already been started by another party. IPLS scales with the number of participants, is robust against intermittent connectivity and dynamic participant departures/arrivals, requires minimal resources, and guarantees that the accuracy of the trained model quickly converges to that of a centralized FL framework with an accuracy drop of less than one per thousand.