CodedPaddedFL and CodedSecAgg: Straggler Mitigation and Secure Aggregation in Federated Learning

被引:25
|
作者
Schlegel, Reent [1 ,2 ]
Kumar, Siddhartha [1 ,3 ]
Rosnes, Eirik [1 ]
Graell i Amat, Alexandre [1 ,4 ]
机构
[1] Simula UiB, N-5006 Bergen, Norway
[2] OHB Digital Connect GmbH, D-28359 Bremen, Germany
[3] Qamcom Res & Technol, S-41285 Gothenburg, Sweden
[4] Chalmers Univ Technol, Dept Elect Engn, S-41296 Gothenburg, Sweden
基金
瑞典研究理事会;
关键词
Servers; Codes; Resilience; Data models; Training; Decoding; Data privacy; Coded distributed computing; federated learning; gradient codes; linear regression; privacy; secure aggregation; straggler mitigation; MODEL;
D O I
10.1109/TCOMM.2023.3244243
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
We present two novel federated learning (FL) schemes that mitigate the effect of straggling devices by introducing redundancy on the devices' data across the network. Compared to other schemes in the literature, which deal with stragglers or device dropouts by ignoring their contribution, the proposed schemes do not suffer from the client drift problem. The first scheme, CodedPaddedFL, mitigates the effect of stragglers while retaining the privacy level of conventional FL. It combines one-time padding for user data privacy with gradient codes to yield straggler resiliency. The second scheme, CodedSecAgg, provides straggler resiliency and robustness against model inversion attacks and is based on Shamir's secret sharing. We apply CodedPaddedFL and CodedSecAgg to a classification problem. For a scenario with 120 devices, CodedPaddedFL achieves a speed-up factor of 18 for an accuracy of 95% on the MNIST dataset compared to conventional FL. Furthermore, it yields similar performance in terms of latency compared to a recently proposed scheme by Prakash et al. without the shortcoming of additional leakage of private data. CodedSecAgg outperforms the state-of-the-art secure aggregation scheme LightSecAgg by a speed-up factor of 6.6-18.7 for the MNIST dataset for an accuracy of 95%.
引用
收藏
页码:2013 / 2027
页数:15
相关论文
共 50 条
  • [1] Straggler-Resilient Secure Aggregation for Federated Learning
    Schlegel, Reent
    Kumar, Siddhartha
    Rosnes, Eirik
    Graell i Amat, Alexandre
    2022 30TH EUROPEAN SIGNAL PROCESSING CONFERENCE (EUSIPCO 2022), 2022, : 712 - 716
  • [2] Coding for Straggler Mitigation in Federated Learning
    Kumar, Siddhartha
    Schlegel, Reent
    Rosnes, Eirik
    Graell i Amat, Alexandre
    IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS (ICC 2022), 2022, : 4962 - 4967
  • [3] Straggler Effect Mitigation for Federated Learning in Cell-Free Massive MIMO
    Vu, Tung T.
    Ngo, Duy T.
    Hien Quoc Ngo
    Dao, Minh N.
    Tran, Nguyen H.
    Middleton, Richard H.
    IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS (ICC 2021), 2021,
  • [4] Verifiable and Secure Aggregation Scheme for Federated Learning
    Ren Y.
    Fu Y.
    Li Y.
    Beijing Youdian Daxue Xuebao/Journal of Beijing University of Posts and Telecommunications, 2023, 46 (03): : 49 - 55
  • [5] Quality Inference in Federated Learning with Secure Aggregation
    Pejó B.
    Biczók G.
    IEEE Transactions on Big Data, 2023, 9 (05): : 1430 - 1437
  • [6] SAFELearn: Secure Aggregation for private FEderated Learning
    Fereidooni, Hossein
    Marchal, Samuel
    Miettinen, Markus
    Mirhoseini, Azalia
    Moellering, Helen
    Thien Duc Nguyen
    Rieger, Phillip
    Sadeghi, Ahmad-Reza
    Schneider, Thomas
    Yalame, Hossein
    Zeitouni, Shaza
    2021 IEEE SYMPOSIUM ON SECURITY AND PRIVACY WORKSHOPS (SPW 2021), 2021, : 56 - 62
  • [7] HeteroSAg: Secure Aggregation With Heterogeneous Quantization in Federated Learning
    Elkordy, Ahmed Roushdy
    Avestimehr, A. Salman
    IEEE TRANSACTIONS ON COMMUNICATIONS, 2022, 70 (04) : 2372 - 2386
  • [8] SVFLC: Secure and Verifiable Federated Learning With Chain Aggregation
    Li, Ning
    Zhou, Ming
    Yu, Haiyang
    Chen, Yuwen
    Yang, Zhen
    IEEE INTERNET OF THINGS JOURNAL, 2024, 11 (08): : 13125 - 13136
  • [9] BalancedSecAgg: Toward Fast Secure Aggregation for Federated Learning
    Masuda, Hiroki
    Kita, Kentaro
    Koizumi, Yuki
    Takemasa, Junji
    Hasegawa, Toru
    IEEE ACCESS, 2024, 12 : 165265 - 165279
  • [10] Device Scheduling for Secure Aggregation in Wireless Federated Learning
    Yan, Na
    Wang, Kezhi
    Zhi, Kangda
    Pan, Cunhua
    Poor, H. Vincent
    Chai, Kok Keong
    IEEE INTERNET OF THINGS JOURNAL, 2024, 11 (17): : 28851 - 28862