A4 Refereed article in a conference publication
EA2: Energy Efficient Adaptive Active Learning for Smart Wearables
Authors: Alikhani, Hamidreza; Wang, Ziyu; Kanduri, Anil; Liljeberg, Pasi; Rahmani, Amir M.; Dutt, Nikil
Editors: Meinerzhagen, Pascal; Dev, Kapil; Yoo, Jerald
Conference name: International Symposium on Low Power Electronics and Design
Publisher: ASSOC COMPUTING MACHINERY
Publishing place: NEW YORK
Publication year: 2024
Book title : ISLPED '24: Proceedings of the 29th ACM/IEEE International Symposium on Low Power Electronics and Design
Number of pages: 6
ISBN: 979-8-4007-0688-2
DOI: https://doi.org/10.1145/3665314.3670840
Web address : https://doi.org/10.1145/3665314.3670840
Self-archived copy’s web address: https://research.utu.fi/converis/portal/detail/Publication/459155506
Mobile Health (mHealth) applications rely on supervised Machine Learning (ML) algorithms, requiring end-user-labeled data for the training phase. The gold standard for obtaining such labeled data is by sending queries to users and gathering responses for the corresponding label, which was conventionally done through triggering questions sent at random. Active Learning (AL) methods use intelligent query-sending policies by incorporating users' contextual information to maximize the response rate and informativeness of the collected labeled data. However, wearable devices' substantial battery drainage associated with the sensing of physiological signals underscores the need for developing an efficient sensing policy in addition to a query-sending policy. In this work, we present a co-optimization framework for both sensing and querying strategies within wearable devices, leveraging contextual information and ML model's prediction confidence. We designed a Reinforcement Learning (RL) agent to quantify different contextual parameters combined with model confidence to determine sensing and querying decisions. Our evaluation of an exemplar stress monitoring application showed a 76% reduction in sensing and data transmission energy consumption, with only a 6% drop in user-labeled data.
Downloadable publication This is an electronic reprint of the original article. |  |
Funding information in the publication:
This work was partially supported by NSF Smart and Connected Communities (S&CC) grant CNS-1831918, Nokia Foundation and Kaute Saatio, Finland.
