RT Journal Article
T1 Dynamic and adaptive fault-tolerant asynchronous federated learning using volunteer edge devices
A1 Morell Martínez, José Ángel
A1 Alba-Torres, Enrique
K1 Internet
AB The number of devices, from smartphones to IoT hardware, interconnected via the Internet is growingall the time. These devices produce a large amount of data that cannot be analyzed in any datacenter or stored in the cloud, and it might be private or sensitive, thus precluding existing classicapproaches. However, studying these data and gaining insights from them is still of great relevanceto science and society. Recently, two related paradigms try to address the above problems. On theone hand, edge computing (EC) suggests to increase processing on edge devices. On the other hand,federated learning (FL) deals with training a shared machine learning (ML) model in a distributed(non-centralized) manner while keeping private data locally on edge devices. The combination ofboth is known as federated edge learning (FEEL). In this work, we propose an algorithm for FEELthat adapts to asynchronous clients joining and leaving the computation. Our research focuses onadapting the learning when the number of volunteers is low and may even drop to zero. We propose,implement, and evaluate a new software platform for this purpose. We then evaluate its resultson problems relevant to FEEL. The proposed decentralized and adaptive system architecture forasynchronous learning allows volunteer users to yield their device resources and local data to train ashared ML model. The platform dynamically self-adapts to variations in the number of collaboratingheterogeneous devices due to unexpected disconnections (i.e., volunteers can join and leave at anytime). Thus, we conduct comprehensive empirical analysis in a static configuration and highly dynamicand changing scenarios. The public open-source platform enables interoperability between volunteersconnected using web browsers and Python processes. (...)
PB Elsevier
YR 2022
FD 2022
LK https://hdl.handle.net/10630/24281
UL https://hdl.handle.net/10630/24281
LA eng
NO José Ángel Morell, Enrique Alba, Dynamic and adaptive fault-tolerant asynchronous federated learning using volunteer edge devices, Future Generation Computer Systems, Volume 133, 2022, Pages 53-67, ISSN 0167-739X, https://doi.org/10.1016/j.future.2022.02.024
NO This research is partially funded by the Universidad de Málaga, Spain, Consejería de Economía y Conocimiento de la Junta deAndalucía, Spain and FEDER, Spain under grant number UMA18- FEDERJA-003 (PRECOG); under grant PID 2020-116727RB-I00 (HUmove) funded by MCIN/AEI/10.13039/501100011-033, Spain; and TAILOR ICT-48 Network (No 952215) funded by EU Horizon 2020 research and innovation programme. José Ángel Morell is supported by an FPU grant from the Ministerio de Educación, Cultura y Deporte, Gobierno de España, Spain (FPU16/02595). Funding for open access charge is supported by the Universidad de Málaga/CBUA, Spain. The views expressed are purely those of the writer and may not in any circumstances be regarded as stating an official position of the European Commission. Funding for open access charge: Universidad de Málaga / CBUA
DS RIUMA. Repositorio Institucional de la Universidad de Málaga
RD 20 ene 2026