Optimizing Federated Learning Using Remote Embeddings for Graph Neural Networks

被引:0
|
作者
Naman, Pranjal [1 ]
Simmhan, Yogesh [1 ]
机构
[1] Indian Inst Sci, Dept Computat & Data Sci, Bangalore, Karnataka, India
关键词
D O I
10.1007/978-3-031-69766-1_32
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Graph Neural Networks (GNNs) have experienced rapid advancements in recent years due to their ability to learn meaningful representations from graph data structures. Federated Learning (FL) has emerged as a viable machine learning approach for training a shared model on decentralized data, addressing privacy concerns while leveraging parallelism. Existing methods that address the unique requirements of federated GNN training using remote embeddings to enhance convergence accuracy are limited by their diminished performance due to large communication costs with a shared embedding server. In this paper, we present OpES, an optimized federated GNN training framework that uses remote neighbourhood pruning, and overlaps pushing of embeddings to the server with local training to reduce the network costs and training time. The modest drop in per-round accuracy due to pre-emptive push of embeddings is out-stripped by the reduction in per-round training time for large and dense graphs like Reddit and Products, converging up to approximate to 2x faster than the state-of-the-art technique using an embedding server and giving up to 20% better accuracy than vanilla federated GNN learning.
引用
收藏
页码:470 / 484
页数:15
相关论文
共 50 条
  • [21] Learning Graph Dynamics using Deep Neural Networks
    Narayan, Apurva
    Roe, Peter H. O'N
    IFAC PAPERSONLINE, 2018, 51 (02): : 433 - 438
  • [22] Learning Graph Neural Networks using Exact Compression
    Bollen, Jeroen
    Steegmans, Jasper
    Van den Bussche, Jan
    Vansummeren, Stijn
    PROCEEDINGS OF THE 6TH ACM SIGMOD JOINT INTERNATIONAL WORKSHOP ON GRAPH DATA MANAGEMENT EXPERIENCES & SYSTEMS AND NETWORK DATA ANALYTICS, GRADES-NDA 2023, 2023,
  • [23] Learning Cortical Parcellations Using Graph Neural Networks
    Eschenburg, Kristian M.
    Grabowski, Thomas J.
    Haynor, David R.
    FRONTIERS IN NEUROSCIENCE, 2021, 15
  • [24] GDFed: Dynamic Federated Learning for Heterogenous Device Using Graph Neural Network
    Yoon, Ji Su
    Kang, Sun Moo
    Park, Seong Bae
    Hong, Choong Seon
    2023 INTERNATIONAL CONFERENCE ON INFORMATION NETWORKING, ICOIN, 2023, : 683 - 685
  • [25] GraFeHTy: Graph Neural Network using Federated Learning for Human Activity Recognition
    Sarkar, Abhishek
    Sen, Tanmay
    Roy, Ashis Kumar
    20TH IEEE INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND APPLICATIONS (ICMLA 2021), 2021, : 1124 - 1129
  • [26] FedHE-Graph: Federated Learning with Hybrid Encryption on Graph Neural Networks for Advanced Persistent Threat Detection
    Bahar, Athmane Ayoub Mansour
    Ferrahi, Kamel Soaid
    Messai, Mohamed-Lamine
    Seba, Hamida
    Amrouche, Karima
    19TH INTERNATIONAL CONFERENCE ON AVAILABILITY, RELIABILITY, AND SECURITY, ARES 2024, 2024,
  • [27] Community detection in networks using graph embeddings
    Tandon, Aditya
    Albeshri, Aiiad
    Thayananthan, Vijey
    Alhalabi, Wadee
    Radicchi, Filippo
    Fortunato, Santo
    PHYSICAL REVIEW E, 2021, 103 (02)
  • [28] Predicting vulnerability inducing function versions using node embeddings and graph neural networks
    Sahin, Sefa Eren
    Ozyedierler, Ecem Mine
    Tosun, Ayse
    INFORMATION AND SOFTWARE TECHNOLOGY, 2022, 145
  • [29] An Unsupervised Neural Prediction Framework for Learning Speaker Embeddings using Recurrent Neural Networks
    Jati, Arindam
    Georgiou, Panayiotis
    19TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2018), VOLS 1-6: SPEECH RESEARCH FOR EMERGING MARKETS IN MULTILINGUAL SOCIETIES, 2018, : 1131 - 1135
  • [30] Scalable out-of-sample extension of graph embeddings using deep neural networks
    Jansen, Aren
    Sell, Gregory
    Lyzinski, Vince
    PATTERN RECOGNITION LETTERS, 2017, 94 : 1 - 6