Real-time End-to-End Federated Learning: An Automotive Case Study

被引:21
|
作者
Zhang, Hongyi [1 ]
Bosch, Jan [1 ]
Olsson, Helena Holmstrom [2 ]
机构
[1] Chalmers Univ Technol, Gothenburg, Sweden
[2] Malmo Univ, Malmo, Sweden
基金
瑞典研究理事会;
关键词
Federated Learning; Machine learning; Heterogeneous computation; Software Engineering;
D O I
10.1109/COMPSAC51774.2021.00070
中图分类号
TP39 [计算机的应用];
学科分类号
081203 ; 0835 ;
摘要
With the development and the increasing interests in ML/DL fields, companies are eager to apply Machine Learning/Deep Learning approaches to increase service quality and customer experience. Federated Learning was implemented as an effective model training method for distributing and accelerating time-consuming model training while protecting user data privacy. However, common Federated Learning approaches, on the other hand, use a synchronous protocol to conduct model aggregation, which is inflexible and unable to adapt to rapidly changing environments and heterogeneous hardware settings in real-world scenarios. In this paper, we present an approach to real-time end-to-end Federated Learning combined with a novel asynchronous model aggregation protocol. Our method is validated in an industrial use case in the automotive domain, focusing on steering wheel angle prediction for autonomous driving. Our findings show that asynchronous Federated Learning can significantly improve the prediction performance of local edge models while maintaining the same level of accuracy as centralized machine learning. Furthermore, by using a sliding training window, the approach can minimize communication overhead, accelerate model training speed and consume real-time streaming data, proving high efficiency when deploying ML/DL components to heterogeneous real-world embedded systems.
引用
收藏
页码:459 / 468
页数:10
相关论文
共 50 条
  • [1] Learning an interpretable end-to-end network for real-time acoustic beamforming
    Liang, Hao
    Zhou, Guanxing
    Tu, Xiaotong
    Jakobsson, Andreas
    Ding, Xinghao
    Huang, Yue
    [J]. JOURNAL OF SOUND AND VIBRATION, 2024, 591
  • [2] Real-time measurement of end-to-end available bandwidth: A feasibility study
    Miyasaka, M
    Iwai, T
    Kasahara, H
    [J]. IEICE TRANSACTIONS ON COMMUNICATIONS, 2006, E89B (02) : 401 - 412
  • [3] END-TO-END LEARNING OF COST-VOLUME AGGREGATION FOR REAL-TIME DENSE
    Kuzmin, Andrey
    Mikushin, Dmitry
    Lempitsky, Victor
    [J]. 2017 IEEE 27TH INTERNATIONAL WORKSHOP ON MACHINE LEARNING FOR SIGNAL PROCESSING, 2017,
  • [4] End-to-End Deep Learning Methodology for Real-Time Traffic Network Management
    Hashemi, Hossein
    Abdelghany, Khaled
    [J]. COMPUTER-AIDED CIVIL AND INFRASTRUCTURE ENGINEERING, 2018, 33 (10) : 849 - 863
  • [5] An end-to-end deep learning approach for real-time single image dehazing
    Jeong, Chi Yoon
    Moon, KyeongDeok
    Kim, Mooseop
    [J]. JOURNAL OF REAL-TIME IMAGE PROCESSING, 2023, 20 (01)
  • [6] End-to-End Learning to Warm-Start for Real-Time Quadratic Optimization
    Sambharya, Rajiv
    Hall, Georgina
    Amos, Brandon
    Stellato, Bartolomeo
    [J]. LEARNING FOR DYNAMICS AND CONTROL CONFERENCE, VOL 211, 2023, 211
  • [7] An end-to-end deep learning approach for real-time single image dehazing
    Chi Yoon Jeong
    KyeongDeok Moon
    Mooseop Kim
    [J]. Journal of Real-Time Image Processing, 2023, 20
  • [8] An end-to-end scheduling approach for real-time CORBA
    Tao, H
    Luo, ZG
    Liu, JD
    [J]. 2002 IEEE REGION 10 CONFERENCE ON COMPUTERS, COMMUNICATIONS, CONTROL AND POWER ENGINEERING, VOLS I-III, PROCEEDINGS, 2002, : 318 - 322
  • [9] Scalability in real-time systems with end-to-end requirements
    Yerraballi, R
    Mukkamalla, R
    [J]. JOURNAL OF SYSTEMS ARCHITECTURE, 1996, 42 (6-7) : 409 - 429
  • [10] End-to-End Real-Time IPTV Quality Monitoring
    Jelacic, Zlatko
    Balasko, Hrvoje
    Grgic, Mislav
    [J]. PROCEEDINGS ELMAR-2008, VOLS 1 AND 2, 2008, : 73 - +