2020/12/04 by Gautham Krishna Gudur, Gudur, Gautham Krishna, Satheesh K. Perepu +1 · 4 citations
Computer Science · Health Professions · #Activity recognition #Artificial intelligence #Computer science #Context-Aware Activity Recognition Systems #FOS: Computer and information sciences #Federated learning #Machine Learning (cs.LG) #Machine learning #Mobile Crowdsensing and Crowdsourcing #Mobile Health and mHealth Applications #Mobile device #Privacy-Preserving Technologies in Data #Transfer of learning #Variety (cybernetics) #World Wide Web #cs.LG
paper · pdf · doi:10.48550/arxiv.2012.02539
published in arXiv (Cornell University) (Cornell University) · 8 pages, 5 figures, Machine Learning for Mobile Health Workshop at NeurIPS 2020. arXiv admin note: substantial text overlap with arXiv:2011.03206
arxiv created 2020/12/04 · openalex publication_date 2020/12/04 · arxiv updated 2020/12/07 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/08
Various health-care applications such as assisted living, fall detection, etc., require modeling of user behavior through Human Activity Recognition (HAR). Such applications demand characterization of insights from multiple resource-constrained user devices using machine learning techniques for effective personalized activity monitoring. On-device Federated Learning proves to be an effective approach for distributed and collaborative machine learning. However, there are a variety of challenges in addressing statistical (non-IID data) and model heterogeneities across users. In addition, in this paper, we explore a new challenge of interest -- to handle heterogeneities in labels (activities) across users during federated learning. To this end, we propose a framework for federated label-based aggregation, which leverages overlapping information gain across activities using Model Distillation Update. We also propose that federated transfer of model scores is sufficient rather than model weight transfer from device to server. Empirical evaluation with the Heterogeneity Human Activity Recognition (HHAR) dataset (with four activities for effective elucidation of results) on Raspberry Pi 2 indicates an average deterministic accuracy increase of at least ~11.01%, thus demonstrating the on-device capabilities of our proposed framework.