In wireless mobile edge computing platforms, such as those supported by vehicular networks, location specific machine learning models can be trained by distributing computations using resources available at the edge. The computational and communication resources at the wireless edge are heterogeneous and unreliable, which can lead to straggler effects that significantly slow down recursive learning tasks, such as gradient descent. In this paper we propose a coded computation framework, which utilizes statistical knowledge of resource heterogeneity to determine optimal encoding and load balancing of training data using Random Linear codes, while avoiding an explicit step for decoding gradients. Results show that the proposed coding framework speeds up the training time for linear regression models, by up to ten times, when compared with repetition coding schemes.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Coded Computing for Distributed Machine Learning in Wireless Edge Network


    Contributors:


    Publication date :

    2019-09-01


    Size :

    448043 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Coded Distributed Computing For Vehicular Edge Computing With Dual-Function Radar Communication

    Nguyen Thi, Hoai Linh / Le Hoang, Hung / Luong, Nguyen Cong et al. | IEEE | 2023


    Learning Distributed Coded Caching Strategy in a Cellular Network

    Doshi, Yash / Bharath, B. N. / Garg, Navneet et al. | IEEE | 2021


    Data Sharing in Virtual Edge Computing using Coded Caching

    Pannu, Gurjashan Singh / Ucar, Seyhan / Higuchi, Takamasa et al. | IEEE | 2023


    Dynamic Coded Distributed Convolution for UAV-based Networked Airborne Computing

    Zhou, Bingnan / Xie, Junfei / Wang, Baoqian | IEEE | 2022


    Machine Learning-Based Workload Orchestrator for Vehicular Edge Computing

    Sonmez, Cagatay / Tunca, Can / Ozgovde, Atay et al. | IEEE | 2021