Advances and Open Problems in Federated Learning

Достижения и открытые проблемы федеративного обучения
Phillip B. Gibbons, Dawn Song, Daniel Ramage, Graham Cormode, Yang Liu, Arjun Nitin Bhagoji, Mehdi Bennis, Martin Jaggi, Ben Hutchinson, Chaoyang He, Mehryar Mohri, Li Xiong, Aleksandra Korolova, Jakub Konečný, H. Brendan McMahan, Han Yu, Qiang Yang, Marco Gruteser, Ramesh Raskar, Sen Zhao, Peter Kairouz, Brendan Avent, Aurélien Bellet, Kallista Bonawitz, Zachary Charles, Rachel Cummings, Rafael G. L. D’Oliveira, Hubert Eichner, Salim El Rouayheb, David Evans, Joshua Gardner, Zachary Garrett, Adrià Gascón, Badih Ghazi, Zaïd Harchaoui, Lingxiao He, Zhouyuan Huo, Justin Hsu, Tara Javidi, Gauri Joshi, Mikhail Khodak, Farinaz Koushanfar, Sanmi Koyejo, Tancrède Lepoint, Prateek Mittal, Richard Nock, Ayfer Özgür, Rasmus Pagh, Hang Qi, Mariana Raykova, Weikang Song, Sebastian U. Stich, Ziteng Sun, Ajith Suresh, Florian Tramèr, Praneeth Vepakomma, Jianyu Wang, Zheng Xu, Felix X. Yu, Ananda Theertha Suresh
2020-12-02
SCID:  54.1/88v4shkt Discuss with AI
Central server orchestrationDecentralized training dataFederated learningFocused data collectionPrivacy-preserving machine learning
Abstract (AI)
Federated learning (FL) is a machine learning setting where many clients (e.g., mobile devices or whole organizations) collaboratively train a model under the orchestration of a central server (e.g., service provider), while keeping the training data decentralized. FL embodies the principles of focused data collection and minimization, and can mitigate many of the systemic privacy risks and costs resulting from traditional, centralized machine learning and data science approaches. Motivated by the explosive growth in FL research, this monograph discusses recent advances and presents an extensive collection of open problems and challenges.
Key Findings
1
By emphasizing focused data collection and minimization, federated learning can mitigate systemic privacy risks and costs associated with centralized machine learning and data science.
2
Federated learning enables clients such as mobile devices or organizations to collaboratively train models through a central server while keeping training data decentralized.
3
The monograph synthesizes recent advances in federated learning and identifies an extensive set of unresolved problems and challenges motivated by rapid research growth.
Research Object

Federated learning (decentralized collaborative training setting with many clients and a central server)

Research Subject

recent advances, open problems, challenges, and privacy-related benefits of federated learning

Publication Details
Publication Date
2020-12-02
Journal
Publisher
ISSN
Access Type
Author Information
Authors
Phillip B. Gibbons
Dawn Song
Daniel Ramage
Graham Cormode
Yang Liu
Arjun Nitin Bhagoji
Mehdi Bennis
Martin Jaggi
Ben Hutchinson
Chaoyang He
Mehryar Mohri
Li Xiong
Aleksandra Korolova
Jakub Konečný
H. Brendan McMahan
Han Yu
Qiang Yang
Marco Gruteser
Ramesh Raskar
Sen Zhao
Peter Kairouz
Brendan Avent
Aurélien Bellet
Kallista Bonawitz
Zachary Charles
Rachel Cummings
Rafael G. L. D’Oliveira
Hubert Eichner
Salim El Rouayheb
David Evans
Joshua Gardner
Zachary Garrett
Adrià Gascón
Badih Ghazi
Zaïd Harchaoui
Lingxiao He
Zhouyuan Huo
Justin Hsu
Tara Javidi
Gauri Joshi
Mikhail Khodak
Farinaz Koushanfar
Sanmi Koyejo
Tancrède Lepoint
Prateek Mittal
Richard Nock
Ayfer Özgür
Rasmus Pagh
Hang Qi
Mariana Raykova
Weikang Song
Sebastian U. Stich
Ziteng Sun
Ajith Suresh
Florian Tramèr
Praneeth Vepakomma
Jianyu Wang
Zheng Xu
Felix X. Yu
Ananda Theertha Suresh
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%