Beyond Sensor Data: Foundation Models of Behavioral Data from Wearables Improve Health Predictions

2025/06/30 by Eray Erturk, Erturk, Eray, Fahad Kamran +17 · 15 voices · 8 citations
Computer Science · Engineering · Psychology · #Advanced Sensor and Energy Harvesting Materials #Context-Aware Activity Recognition Systems #Emotion and Mood Recognition #cs.AI #cs.LG

paper · pdf · doi:10.48550/arxiv.2507.00191

openalex publication_date 2025/06/30 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Wearable devices record physiological and behavioral signals that can improve health predictions. While foundation models are increasingly used for such predictions, they have been primarily applied to low-level sensor data, despite behavioral data often being more informative due to their alignment with physiologically relevant timescales and quantities. We develop foundation models of such behavioral signals using over 2.5B hours of wearable data from 162K individuals, systematically optimizing architectures and tokenization strategies for this unique dataset. Evaluated on 57 health-related tasks, our model shows strong performance across diverse real-world applications including individual-level classification and time-varying health state prediction. The model excels in behavior-driven tasks like sleep prediction, and improves further when combined with representations of raw sensor data. These results underscore the importance of tailoring foundation model design to wearables and demonstrate the potential to enable new health applications.

Citations

• Chronos: Learning the Language of Time Series
• Mamba: Linear-Time Sequence Modeling with Selective State Spaces
• TimeGPT-1
• Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture
• On the Opportunities and Risks of Foundation Models
• Attention Is All You Need

Cited by

• CLEF: Clinically-Guided Contrastive Learning for Electrocardiogram Foundation Models
• Wrist Photoplethysmography Predicts Dietary Information
• HiMAE: Hierarchical Masked Autoencoders Discover Resolution-Specific Structure in Wearable Time Series
• Generalist vs Specialist Time Series Foundation Models: Investigating Potential Emergent Behaviors in Assessing Human Health Using PPG Signals
• Beyond the Clinic: A Large-Scale Evaluation of Augmenting EHR with Wearable Data for Diverse Health Prediction
• SlotFM: A Motion Foundation Model with Slot Attention for Diverse Downstream Tasks
• PPG-Distill: Efficient Photoplethysmography Signals Analysis via Foundation Model Distillation
• Speech Foundation Models Generalize to Time Series Tasks from Wearable Sensor Data

Discussions

• Beyond sensor data: Foundation models of behavioral data from wearables [hn, 230 points, 54 comments]
• [2507.00191] Beyond Sensor Data: Foundation Models of Behavioral Data from Wearables Improve Health Predictions arxiv.org/abs/2507.00191 - 연구자들은 16만 2천명의 애플워치 사용자로부터 25억 시간의 행동 데이터를 사용하여 WBM이라는 기초 모델을 [bsky, 1 points, 1 comments]
• New AI behavioral models to predict changes in health from wearables. #Medsky arxiv.org/pdf/2507.001... [bsky, 1 points, 0 comments]
• Foundation Models of Behavioral Data from Wearables Improve Health Predictions [hn, 1 points, 0 comments]
• Apple Watch wearable foundation model https://arxiv.org/abs/2507.00191 https://news.ycombinator.com/item?id=44973375 [bsky, 0 points, 0 comments]
• Apple Watch wearable foundation model https://arxiv.org/abs/2507.00191 [bsky, 0 points, 0 comments]
• Beyond sensor data: Foundation models of behavioral data from wearables https://arxiv.org/abs/2507.00191 [comments] [167 points] [bsky, 0 points, 0 comments]
• Apple Watch wearable foundation model https://arxiv.org/abs/2507.00191 (https://news.ycombinator.com/item?id=44973375) [bsky, 0 points, 0 comments]
• ⚡ Hackernews Top story: Apple Watch wearable foundation model [bsky, 0 points, 0 comments]
• Beyond sensor data: Foundation models of behavioral data from wearables https://arxiv.org/abs/2507.00191 (http://news.ycombinator.com/item?id=44973375) [bsky, 0 points, 0 comments]
• Apple Watch wearable foundation model https://arxiv.org/abs/2507.00191 (https://news.ycombinator.com/item?id=44973375) [bsky, 0 points, 0 comments]
• https://arxiv.org/abs/2507.00191 ウェアラブルデバイスからの行動データを用いたFoundation Modelが、健康予測を向上させるという研究。 センサーデータだけでなく、行動データを活用することで、より高度な予測が可能になる。 論文では、具体的なモデルの構造や実験結果について詳しく解説されている。 [bsky, 0 points, 0 comments]
• Beyond sensor data: Foundation models of behavioral data from wearables view on hacker news [bsky, 0 points, 0 comments]
• https://bsky.app/profile/hackernews.com.web.brid.gy/post/3lwwgvxfph3i2 [bsky, 0 points, 0 comments]
• Apple Watch wearable foundation model #HackerNews https://arxiv.org/abs/2507.00191 [bsky, 0 points, 0 comments]

Related

• Towards a General Intelligence and Interface for Wearable Health Data
• Toward Foundation Model for Multivariate Wearable Sensing of Physiological Signals
• OpenMHC: Accelerating the Science of Wearable Foundation Models
• HealthCAT: An Interpretable Encoder-only Transformer Framework for Health Indicator Prediction and Temporal Interpretation of Wearable Sensor Data
• Scaling Wearable Foundation Models