SOTER is a generative foundation model designed for wearable physiological time‑series data. It integrates cross‑channel coupling, spectrum‑guided expert specialization, and continuous‑time latent evolution, using a spatial feature‑aware backbone, a PSD‑guided mixture‑of‑experts layer, and a neural controlled differential equation decoder. Trained on 226 billion time points from five public datasets, SOTER outperforms baselines in zero‑shot forecasting, classification, and imputation across six benchmarks, and remains robust to additive noise.
By Fangke Chen, Sirry Chen, Wei Chen, Zhongyu Wei
Time-series foundation models have demonstrated strong cross-domain transfer, yet their common architectural assumptions remain poorly aligned with wearable physiological signals, which are multichann...
BRIDGE-EEG is an efficient multi‑task EEG classification pipeline that leverages self‑supervised pretraining while dramatically reducing model size. It maps heterogeneous EEG recordings to a unified 62‑channel time‑frequency representation, pretrains an SE‑ResNet18 teacher with SimCLR, and distills it into smaller SE‑ResNet8 and SE‑ResNet4 students. The compact models achieve accuracy comparable to or better than larger foundation models on abnormality detection and emotion recognition, and they consume up to three times less energy on edge devices, enabling deployment on wearable hardware.
By Meghna Roy Chowdhury, Chengwei Zhou, Haotian Yu, Gourav Datta, Shreyas Sen
The study investigates ECG biometrics by training a Siamese ResNet with late multi-lead fusion on a large dataset from cardiopulmonary exercise tests. It evaluates the model under realistic conditions, including exercise-induced stress and cross-session variability, achieving an intra-session rest-to-peak EER of 1.7% and a state‑of‑the‑art 3.9% on the CYBHi dataset. The results demonstrate that an intrinsic cardiac signature remains robust to physiological and temporal drift.
By Luca Thiebaud (AMU, AMU SCI, DIAPRO, LIS), Paul Chauchat (AMU SCI, AMU, LIS, DIAPRO), Mustapha Ouladsine (AMU SCI, AMU, LIS, DIAPRO), St\'ephane Delliaux (AMU, APHM, C2VN)
WearableQA is a new benchmark that tests AI systems on health reasoning using real-world wearable data from 200 users, each with up to 500 days of daily measurements. It contains 4,084 ten‑option multiple‑choice questions derived from wearable time series, blood biomarkers, and demographics, and is organized into 16 question types that distinguish data‑driven computation from physiological interpretation and single‑signal from cross‑signal reasoning. Evaluation of 14 large language models shows wide performance gaps, indicating that the benchmark remains challenging and useful for diagnosing model capabilities.
By Ji Soo Lee, Xilun Chen, Pierce Chuang, Ashish Shenoy, Jason Wei, Dohwan Ko, Hyunwoo J. Kim, Benoit Corda
arXiv:2601. 04181v2 Announce Type: replace Abstract: Reliable long-term decoding of gestures from surface electromyography (EMG) is hindered by signal drift caused by electrode displacement, muscle fatigue, and/or posture changes.
By Nia Touko, Matthew O A Ellis, Cristiano Capone, Alessio Burrello, Elisa Donati, Luca Manneschi