Client selection in federated learning

• Grant number: 24/08223-3
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: June 01, 2024
• End date: May 31, 2026
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Agreement: MCTI/MC
• Principal Investigator: Miguel Elias Mitre Campista
• Grantee: Maria Victoria França Silva Ramos
• Host Institution: Instituto Alberto Luiz Coimbra de Pós-Graduação e Pesquisa (COPPE). Universidade Federal do Rio de Janeiro (UFRJ). Ministério da Educação (Brasil)
• Associated research grant: 23/00811-0 - EcoSustain: computer and data science for the environment, AP.TEM

Abstract

Federated learning's performance is affected by the time participants take to train models locally and submit them to the server. Thus, the main objective of this project is to analyze the influence of client selection on federated learning using the latency to respond to the server as a criterion. Thus, by distinguishing participants, we aim to reduce the impact of latecomers on the model's convergence time. However, as delays in response may result from factors beyond the participants' control, strategies that manage to benefit from late-arriving contributions will also be investigated.