Samsung Research America has developed two AI foundation models, xMAE and HiMAE, designed to analyze biosignal data from wearable devices. These models utilize self-supervised learning to extract insights from unlabeled health data, aiming to support continuous, personalized, and preventive healthcare. xMAE focuses on the temporal relationship between photoplethysmogram (PPG) and electrocardiography (ECG) signals, enabling analysis of cardiovascular health through passive PPG measurements. HiMAE learns health patterns across various time scales in wearable data, facilitating on-device processing with limited resources. AI
IMPACT These foundation models could enable more sophisticated on-device health monitoring and personalized healthcare insights from wearables.
RANK_REASON Research publication at academic conferences (ICML, ICLR) for AI models.
Read on Mastodon — fosstodon.org →
- photoplethysmogram
- Samsung
- electrocardiography
- HiMAE
- International Conference on Learning Representations
- International Conference on Machine Learning
- Samsung Research America
- Sharanya Desai
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →