Skip to Main Content (Press Enter)

Logo UNITO
  • ×
  • Home
  • Pubblicazioni
  • Progetti
  • Persone
  • Competenze
  • Settori
  • Strutture
  • Terza Missione

UNI-FIND
Logo UNITO

|

UNI-FIND

unito.it
  • ×
  • Home
  • Pubblicazioni
  • Progetti
  • Persone
  • Competenze
  • Settori
  • Strutture
  • Terza Missione
  1. Pubblicazioni

Efficiently Distributed Federated Learning

Contributo in Atti di convegno
Data di Pubblicazione:
2024
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.
Tipologia CRIS:
04A-Conference paper in volume
Keywords:
Distributed Computing, Federated Learning, HPC
Elenco autori:
Mittone G.; Birke R.; Aldinucci M.
Autori di Ateneo:
ALDINUCCI Marco
BIRKE Robert Renè Maria
MITTONE GIANLUCA
Link alla scheda completa:
https://iris.unito.it/handle/2318/2031766
Link al Full Text:
https://iris.unito.it/retrieve/handle/2318/2031766/1422854/Efficient_FL.pdf
Titolo del libro:
Euro-Par 2023: Parallel Processing Workshops - Euro-Par 2023 International Workshops, Limassol, Cyprus, August 28 - September 1, 2023, Revised Selected Papers, Part II
Pubblicato in:
LECTURE NOTES IN COMPUTER SCIENCE
Journal
LECTURE NOTES IN COMPUTER SCIENCE
Series
Progetto:
Third party CINI - "The European PILOT - Pilot using Independent Local & Open Technologies" (H2020-JTI-EuroHPC-2020-1)
  • Dati Generali
  • Aree Di Ricerca

Dati Generali

URL

https://link.springer.com/chapter/10.1007/978-3-031-48803-0_40

Aree Di Ricerca

Settori (14)


PE6_11 - Machine learning, statistical data processing and applications using signal processing (e.g. speech, image, video) - (2024)

PE6_2 - Distributed systems, parallel computing, sensor networks, cyber-physical systems - (2024)

CIBO, AGRICOLTURA e ALLEVAMENTI - Farmacologia Veterinaria

CULTURA, ARTE e CREATIVITA' - Culture moderne

ECONOMIA, AZIENDE E ORGANIZZAZIONI - Sistemi e metodologie per la Qualità

INFORMATICA, AUTOMAZIONE e INTELLIGENZA ARTIFICIALE - Digitalizzazione della Cultura e della Creatività

INFORMATICA, AUTOMAZIONE e INTELLIGENZA ARTIFICIALE - Digitalizzazione della Società e della Pubblica Amministrazione

INFORMATICA, AUTOMAZIONE e INTELLIGENZA ARTIFICIALE - Industria X.0

INFORMATICA, AUTOMAZIONE e INTELLIGENZA ARTIFICIALE - Salute e Informatica

LINGUE e LETTERATURA - Linguistica

PIANETA TERRA, AMBIENTE, CLIMA, ENERGIA e SOSTENIBILITA' - Diritto dell'Ambiente

PIANETA TERRA, AMBIENTE, CLIMA, ENERGIA e SOSTENIBILITA' - Informatica e Ambiente

SCIENZE DELLA VITA e FARMACOLOGIA - Tecnologie Farmaceutiche e Cosmetiche

SCIENZE MATEMATICHE, CHIMICHE, FISICHE - Teorie e modelli Matematici
  • Utilizzo dei cookie

Realizzato con VIVO | Designed by Cineca | 25.4.2.0