Personalized On-Device E-health Analytics with Decentralized Block Coordinate Descent

Guanhua Ye, Hongzhi Yin, Tong Chen, Miao Xu, Quoc Viet Hung Nguyen, Jiangning Song

Research output: Contribution to journalArticleResearchpeer-review

6 Citations (Scopus)

Abstract

Actuated by the growing attention to personal healthcare and the pandemic, the popularity of E-health is proliferating. Nowadays, enhancement on medical diagnosis via machine learning models has been highly effective in many aspects of e-health analytics. Nevertheless, in the classic cloud-based/centralized e-health paradigms, all the data will be centrally stored on the server to facilitate model training, which inevitably incurs privacy concerns and high time delay. Distributed solutions like Decentralized Stochastic Gradient Descent (D-SGD) are proposed to provide safe and timely diagnostic results based on personal devices. However, methods like D-SGD are subject to the gradient vanishing issue and usually proceed slowly at the early training stage, thereby impeding the effectiveness and efficiency of training. In addition, existing methods are prone to learning models that are biased towards users with dense data, compromising the fairness when providing E-health analytics for minority groups. In this paper, we propose a Decentralized Block Coordinate Descent (D-BCD) learning framework that can better optimize deep neural network-based models distributed on decentralized devices for E-health analytics. As a gradient-free optimization method, Block Coordinate Descent (BCD) mitigates the gradient vanishing issue and converges faster at the early stage compared with the conventional gradient-based optimization. To overcome the potential data scarcity issues for users local data, we propose similarity-based model aggregation that allows each on-device model to leverage knowledge from similar neighbor models, so as to achieve both personalization and high accuracy for the learned models. Benchmarking experiments on three real-world datasets illustrate the effectiveness and practicality of our proposed DBCD, where additional simulation study showcases the strong applicability of D-BCD in real-life E-health scenarios.

Original languageEnglish
Pages (from-to)2778-2786
Number of pages9
JournalIEEE Journal of Biomedical and Health Informatics
Volume26
Issue number6
DOIs
Publication statusPublished - 2022

Keywords

  • Analytical models
  • Computational modeling
  • Data models
  • Decentralized Optimization
  • E-health Analytics
  • On-Device Machine Learning
  • Optimization
  • Predictive models
  • Servers
  • Similarity-based Aggregation
  • Training

Cite this