Scaling Wearable Foundation Models
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866912075458019328 |
|---|---|
| author | Narayanswamy, Girish Liu, Xin Ayush, Kumar Yang, Yuzhe Xu, Xuhai Liao, Shun Garrison, Jake Tailor, Shyam Sunshine, Jake Liu, Yun Althoff, Tim Narayanan, Shrikanth Kohli, Pushmeet Zhan, Jiening Malhotra, Mark Patel, Shwetak Abdel-Ghaffar, Samy McDuff, Daniel |
| author_facet | Narayanswamy, Girish Liu, Xin Ayush, Kumar Yang, Yuzhe Xu, Xuhai Liao, Shun Garrison, Jake Tailor, Shyam Sunshine, Jake Liu, Yun Althoff, Tim Narayanan, Shrikanth Kohli, Pushmeet Zhan, Jiening Malhotra, Mark Patel, Shwetak Abdel-Ghaffar, Samy McDuff, Daniel |
| contents | Wearable sensors have become ubiquitous thanks to a variety of health tracking features. The resulting continuous and longitudinal measurements from everyday life generate large volumes of data; however, making sense of these observations for scientific and actionable insights is non-trivial. Inspired by the empirical success of generative modeling, where large neural networks learn powerful representations from vast amounts of text, image, video, or audio data, we investigate the scaling properties of sensor foundation models across compute, data, and model size. Using a dataset of up to 40 million hours of in-situ heart rate, heart rate variability, electrodermal activity, accelerometer, skin temperature, and altimeter per-minute data from over 165,000 people, we create LSM, a multimodal foundation model built on the largest wearable-signals dataset with the most extensive range of sensor modalities to date. Our results establish the scaling laws of LSM for tasks such as imputation, interpolation and extrapolation, both across time and sensor modalities. Moreover, we highlight how LSM enables sample-efficient downstream learning for tasks like exercise and activity recognition. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_13638 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Scaling Wearable Foundation Models Narayanswamy, Girish Liu, Xin Ayush, Kumar Yang, Yuzhe Xu, Xuhai Liao, Shun Garrison, Jake Tailor, Shyam Sunshine, Jake Liu, Yun Althoff, Tim Narayanan, Shrikanth Kohli, Pushmeet Zhan, Jiening Malhotra, Mark Patel, Shwetak Abdel-Ghaffar, Samy McDuff, Daniel Machine Learning Artificial Intelligence Human-Computer Interaction Wearable sensors have become ubiquitous thanks to a variety of health tracking features. The resulting continuous and longitudinal measurements from everyday life generate large volumes of data; however, making sense of these observations for scientific and actionable insights is non-trivial. Inspired by the empirical success of generative modeling, where large neural networks learn powerful representations from vast amounts of text, image, video, or audio data, we investigate the scaling properties of sensor foundation models across compute, data, and model size. Using a dataset of up to 40 million hours of in-situ heart rate, heart rate variability, electrodermal activity, accelerometer, skin temperature, and altimeter per-minute data from over 165,000 people, we create LSM, a multimodal foundation model built on the largest wearable-signals dataset with the most extensive range of sensor modalities to date. Our results establish the scaling laws of LSM for tasks such as imputation, interpolation and extrapolation, both across time and sensor modalities. Moreover, we highlight how LSM enables sample-efficient downstream learning for tasks like exercise and activity recognition. |
| title | Scaling Wearable Foundation Models |
| topic | Machine Learning Artificial Intelligence Human-Computer Interaction |
| url | https://arxiv.org/abs/2410.13638 |