Migrating Models: A Decentralized View on Federated Learning

Kiss, Péter; Horváth, Tomáš

Abstract

Federated learning (FL) researches attempt to alleviate the increasing difficulty of training machine learning models, when the training data is generated in a massively distributed way. The key idea behind these methods is moving the training to locations of data generation, and periodically collecting and redistributing the model updates. We present our approach for transforming the general training algorithm of FL into a peer-to-peer-like process. Our experiments on baseline image classification datasets show that omitting central coordination in FL is feasible.

Citation

Kiss, P., & Horváth, T. (2021, September). Migrating Models: A Decentralized View on Federated Learning. Presented at ECML PKDD 2021, Online

• Presentation Conference Type: Conference Paper (published)
• Conference Name: ECML PKDD 2021
• Start Date: Sep 13, 2021
• End Date: Sep 17, 2021
• Online Publication Date: Feb 17, 2022
• Publication Date: 2021
• Deposit Date: Apr 8, 2024
• Publisher: Springer
• Pages: 177-191
• Series Title: Communications in Computer and Information Science
• Series Number: 1524
• Series ISSN: 1865-0929
• Book Title: Machine Learning and Principles and Practice of Knowledge Discovery in Databases: International Workshops of ECML PKDD 2021, Virtual Event, September 13-17, 2021, Proceedings, Part I
• ISBN: 9783030937355
• DOI: https://doi.org/10.1007/978-3-030-93736-2_15
• Keywords: Federated learning, Peer-to-peer, Neural networks
• Public URL: http://researchrepository.napier.ac.uk/Output/3587425
• Related Public URLs: https://ecmlpkdd.org/2021/