LSM-2: Learning from Incomplete Wearable Sensor Data
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909639581368320 |
|---|---|
| author | Xu, Maxwell A. Narayanswamy, Girish Ayush, Kumar Spathis, Dimitris Liao, Shun Tailor, Shyam A. Metwally, Ahmed Heydari, A. Ali Zhang, Yuwei Garrison, Jake Abdel-Ghaffar, Samy Xu, Xuhai Gu, Ken Sunshine, Jacob Poh, Ming-Zher Liu, Yun Althoff, Tim Narayanan, Shrikanth Kohli, Pushmeet Malhotra, Mark Patel, Shwetak Yang, Yuzhe Rehg, James M. Liu, Xin McDuff, Daniel |
| author_facet | Xu, Maxwell A. Narayanswamy, Girish Ayush, Kumar Spathis, Dimitris Liao, Shun Tailor, Shyam A. Metwally, Ahmed Heydari, A. Ali Zhang, Yuwei Garrison, Jake Abdel-Ghaffar, Samy Xu, Xuhai Gu, Ken Sunshine, Jacob Poh, Ming-Zher Liu, Yun Althoff, Tim Narayanan, Shrikanth Kohli, Pushmeet Malhotra, Mark Patel, Shwetak Yang, Yuzhe Rehg, James M. Liu, Xin McDuff, Daniel |
| contents | Foundation models, a cornerstone of recent advancements in machine learning, have predominantly thrived on complete and well-structured data. Wearable sensor data frequently suffers from significant missingness, posing a substantial challenge for self-supervised learning (SSL) models that typically assume complete data inputs. This paper introduces the second generation of Large Sensor Model (LSM-2) with Adaptive and Inherited Masking (AIM), a novel SSL approach that learns robust representations directly from incomplete data without requiring explicit imputation. AIM's core novelty lies in its use of learnable mask tokens to model both existing ("inherited") and artificially introduced missingness, enabling it to robustly handle fragmented real-world data during inference. Pre-trained on an extensive dataset of 40M hours of day-long multimodal sensor data, our LSM-2 with AIM achieves the best performance across a diverse range of tasks, including classification, regression and generative modeling. Furthermore, LSM-2 with AIM exhibits superior scaling performance, and critically, maintains high performance even under targeted missingness scenarios, reflecting clinically coherent patterns, such as the diagnostic value of nighttime biosignals for hypertension prediction. This makes AIM a more reliable choice for real-world wearable data applications. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_05321 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | LSM-2: Learning from Incomplete Wearable Sensor Data Xu, Maxwell A. Narayanswamy, Girish Ayush, Kumar Spathis, Dimitris Liao, Shun Tailor, Shyam A. Metwally, Ahmed Heydari, A. Ali Zhang, Yuwei Garrison, Jake Abdel-Ghaffar, Samy Xu, Xuhai Gu, Ken Sunshine, Jacob Poh, Ming-Zher Liu, Yun Althoff, Tim Narayanan, Shrikanth Kohli, Pushmeet Malhotra, Mark Patel, Shwetak Yang, Yuzhe Rehg, James M. Liu, Xin McDuff, Daniel Machine Learning Foundation models, a cornerstone of recent advancements in machine learning, have predominantly thrived on complete and well-structured data. Wearable sensor data frequently suffers from significant missingness, posing a substantial challenge for self-supervised learning (SSL) models that typically assume complete data inputs. This paper introduces the second generation of Large Sensor Model (LSM-2) with Adaptive and Inherited Masking (AIM), a novel SSL approach that learns robust representations directly from incomplete data without requiring explicit imputation. AIM's core novelty lies in its use of learnable mask tokens to model both existing ("inherited") and artificially introduced missingness, enabling it to robustly handle fragmented real-world data during inference. Pre-trained on an extensive dataset of 40M hours of day-long multimodal sensor data, our LSM-2 with AIM achieves the best performance across a diverse range of tasks, including classification, regression and generative modeling. Furthermore, LSM-2 with AIM exhibits superior scaling performance, and critically, maintains high performance even under targeted missingness scenarios, reflecting clinically coherent patterns, such as the diagnostic value of nighttime biosignals for hypertension prediction. This makes AIM a more reliable choice for real-world wearable data applications. |
| title | LSM-2: Learning from Incomplete Wearable Sensor Data |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2506.05321 |