A meta-learning method with reduced data requirements for training and updating deep learning models on wearable devices
Amirshahi A, Toosi MH (equal contribution), Mohammadi S, Albini S, Schiavone PD, Ansaloni G, Aminifar A, Atienza D (EPFL ESL; University of Tehran; IPM; Lund University, per PubMed)
Summary
Meta-learning lowers how much data a new wearable must collect up front, and a prototypical updating mechanism changes class prototypes instead of retraining. Case studies: AF detection from ECG (PhysioNet/CinC Challenge 2017) and epileptic seizure detection from EEG (Siena, TUSZ). Backbones: MobileNetV2 and ViT.
Key facts
- License
- Not stated
Updated: Oct 12, 2026
Related Research
- A Decodability-Grounded Audit of Clinical Prediction Claims in PPG Foundation Models
Audits frozen representations of four encoders on 37,784 MC-MED encounters via concept decodability,
- A Physiology-Aware Hierarchical Foundation Model for Long-Context Waveform Representation Learning (Physio-HNet)
Causal hierarchical waveform FM replacing fixed patches with physiology-aware dynamic chunking (land
- Agentic AI-enabled discovery across large-scale sleep physiology (AI Sleep Co-Scientist)
Expert-guided environment where scientists direct specialist agents for hypothesis development, sign