English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Human Activity Recognition (HAR) training data is often privacy-sensitive orheld by non-cooperative entities. Federated Learning (FL) addresses suchconcerns by training ML models on edge clients. This work studies the impact ofprivacy in federated HAR at a user, environment, and sensor level. We show thatthe performance of FL for HAR depends on the assumed privacy level of the FLsystem and primarily upon the colocation of data from different sensors. Byavoiding data sharing and assuming privacy at the human or environment level,as prior works have done, the accuracy decreases by 5-7%. However, extendingthis to the modality level and strictly separating sensor data between multipleclients may decrease the accuracy by 19-42%. As this form of privacy isnecessary for the ethical utilisation of passive sensing methods in HAR, weimplement a system where clients mutually train both a general FL model and agroup-level one per modality. Our evaluation shows that this method leads toonly a 7-13% decrease in accuracy, making it possible to build HAR systems withdiverse hardware.

Privacy in Multimodal Federated Human Activity Recognition