Google Research has introduced SensorFM, a large foundation model for wearable health data. Pre-trained on over a trillion minutes of sensor data from five million individuals, SensorFM learns a general representation of human physiology that can be applied to 35 different health prediction tasks. This model aims to move beyond single-purpose health algorithms, offering a more integrated approach to interpreting wearable device data for personalized health insights and potentially serving as the core for a Personal Health Agent. AI
IMPACT Sets a new standard for interpreting wearable health data, potentially transforming preventive and personalized healthcare by enabling more sophisticated health agents.
RANK_REASON Frontier-lab model release with system card
- Daniel McDuff
- Fitbit
- Google Pixel Watch
- Google Research
- Personal Health Agent
- SensorFM
- Xin Liu
- accelerometer
- altimeter
- Eda
- photoplethysmogram
- skin temperature
- ViT-1D
- GPT-4
- Oura
- Oura Advisor
- Whoop Coach
AI-generated summary · Google Gemini · from 6 sources. How we write summaries →