[BibTeX] [RIS]
Towards a Meaningful Communication and Model Aggregation in Federated Learning via Genetic Programming
Publicatietype: In proceedings
Citatie:
Boektitel: Proceedings of the 17th International Conference on Agents and Artificial Intelligence: ICAART
Deel: 3
Jaar: 2025
Maand: Februari
Pagina's: 1427-1431
Uitgever: Springer
Locatie: Porto, Portugal
ISSN: 2184-433X
ISBN: 978-989-758-737-5
URL: https://www.scitepress.org/Pap...
DOI: 10.5220/0013380400003890
Samenvatting: Federated Learning (FL) enables collaborative training of machine learning models while preserving client data privacy. However, its conventional client-server paradigm presents two key challenges: (i) communication efficiency and (ii) model aggregation optimization. Inefficient communication, often caused by transmitting low-impact updates, results in unnecessary overhead, particularly in bandwidth-constrained environments such as wireless or mobile networks or in scenarios with numerous clients. Furthermore, traditional aggregation strategies lack the adaptability required for stable convergence and optimal performance. This paper emphasizes the distributed nature of FL clients (agents) and advocates for local, autonomous, and intelligent strategies to evaluate the significance of their updates—such as using a ''distance'' metric relative to the global model. This approach improves communication efficiency by prioritizing impactful updates. Additionally, the paper proposes an adaptive aggregation method leveraging genetic programming and transfer learning to dynamically evolve aggregation equations, optimizing the convergence process. By integrating insights from multi-agent systems, the proposed approach aims to foster more efficient and robust frameworks for decentralized learning.
Trefwoorden: Communication Efficiency, federated learning, Genetic Programming, Models Aggregation, Multi-Agent Systems
Auteurs Pacioni, Elia
Fernández de Vega, Francisco
Calvaresi, Davide
Toegevoegd door: []
Totaalscore: 0
Bestanden
  • paper.pdf
Aantekeningen
    Onderwerpen