Federated Learning (FL) is experiencing a substantial research interest, with many frameworks being developed to allow practitioners to build federations easily and quickly. Most of these efforts do not consider two main aspects that are key to Machine Learning (ML) software: customizability and performance. This research addresses these issues by implementing an open-source FL framework named FastFederatedLearning (FFL). FFL is implemented in C/C++, focusing on code performance, and allows the user to specify any communication graph between clients and servers involved in the federation, ensuring customizability. FFL is tested against Intel OpenFL, achieving consistent speedups over different computational platforms (x86-64, ARM-v8, RISC-V), ranging from 2.5x and 3.69x. We aim to wrap FFL with a Python interface to ease its use and implement a middleware for different communication backends to be used. We aim to build dynamic federations in which relations between clients and servers are not static, giving life to an environment where federations can be seen as long-time evolving structures and exploited as services.

Efficiently Distributed Federated Learning

Mittone G.
First
;
Birke R.;Aldinucci M.
Last
2024-01-01

Abstract

Federated Learning (FL) is experiencing a substantial research interest, with many frameworks being developed to allow practitioners to build federations easily and quickly. Most of these efforts do not consider two main aspects that are key to Machine Learning (ML) software: customizability and performance. This research addresses these issues by implementing an open-source FL framework named FastFederatedLearning (FFL). FFL is implemented in C/C++, focusing on code performance, and allows the user to specify any communication graph between clients and servers involved in the federation, ensuring customizability. FFL is tested against Intel OpenFL, achieving consistent speedups over different computational platforms (x86-64, ARM-v8, RISC-V), ranging from 2.5x and 3.69x. We aim to wrap FFL with a Python interface to ease its use and implement a middleware for different communication backends to be used. We aim to build dynamic federations in which relations between clients and servers are not static, giving life to an environment where federations can be seen as long-time evolving structures and exploited as services.
2024
PhD Symposium held at the 29th International European Conference on Parallel and Distributed Computing
Limassol, Cipro
28 agosto - 1 settembre 2023
Euro-Par 2023: Parallel Processing Workshops - Euro-Par 2023 International Workshops, Limassol, Cyprus, August 28 - September 1, 2023, Revised Selected Papers, Part II
Springer
14352
321
326
9783031488023
9783031488030
https://link.springer.com/chapter/10.1007/978-3-031-48803-0_40
Distributed Computing, Federated Learning, HPC
Mittone G.; Birke R.; Aldinucci M.
File in questo prodotto:
File Dimensione Formato  
Efficient_FL.pdf

Accesso aperto

Tipo di file: PREPRINT (PRIMA BOZZA)
Dimensione 189.9 kB
Formato Adobe PDF
189.9 kB Adobe PDF Visualizza/Apri
978-3-031-48803-0_40.pdf

Accesso riservato

Descrizione: Efficiently Distributed Federated Learning
Tipo di file: PDF EDITORIALE
Dimensione 160.86 kB
Formato Adobe PDF
160.86 kB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/2318/2031766
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 1
  • ???jsp.display-item.citation.isi??? 0
social impact