Sensori Model Presented for Health from Wrist Movement
AI researcher shares new self-supervised model trained on wrist accelerometer data.
Tanishq Mathew Abraham posted about an arXiv paper titled Learning Human Health and Diseases from 24-hour Wrist Movement. The post describes Sensori, a self-supervised foundation model that learns general-purpose health representations from 24 hours of raw tri-axial wrist movement. The work draws on a dataset of 122,640 participants contributing 683,617 person-days of free-living recordings. Abraham, listed as founder and PhD holder, is CEO of SophontAI and posted the update with an attached screenshot of the paper.
Combined views
5.6K
2 posts, first seen 21h ago
Sensori Model Presented for Health from Wrist Movement
AI researcher shares new self-supervised model trained on wrist accelerometer data.
Tanishq Mathew Abraham posted about an arXiv paper titled Learning Human Health and Diseases from 24-hour Wrist Movement. The post describes Sensori, a self-supervised foundation model that learns general-purpose health representations from 24 hours of raw tri-axial wrist movement. The work draws on a dataset of 122,640 participants contributing 683,617 person-days of free-living recordings. Abraham, listed as founder and PhD holder, is CEO of SophontAI and posted the update with an attached screenshot of the paper.
Reactions
2 postsTM 
Tanishq Mathew Abraham, Ph.D.@iScienceLuvrLearning Human Health and Diseases from 24-hour Wrist Movement "Here, we present Sensori, a self-supervised foundation model that learns general-purpose health representations directly from 24 hours of raw tri-axial wrist movement." Dataset: 122,640 participants contributing…
- likes: 20
- replies: 4
- bookmarks: 9
- reposts: 2
TM Tanishq Mathew Abraham, Ph.D.@iScienceLuvrpapers like this are very interesting because a truly multimodal representation of patient health will include AS MANY MEASUREMENTS about the patient... in this case wrist movement (which you can get from a wearable) apparently adds signal for predicting clinical info!
- likes: 9
- replies: 2
- bookmarks: 1
- reposts: 2
