WavesFM: Hierarchical Representation Learning for Longitudinal Wearable Sensor Waveforms

• URL: http://arxiv.org/abs/2605.09173v1
• Date: Sat, 09 May 2026 21:22:02 GMT
• Title: WavesFM: Hierarchical Representation Learning for Longitudinal Wearable Sensor Waveforms
• Abstract summary: We propose WavesFM, a foundation model utilizing a two-stage SSL framework for longitudinal physiological data.<n>Specifically, we decompose the learning problem into two stages: first, a segment-level encoder is pretrained to extract local embeddings from short waveforms.<n>This hierarchical approach overcomes the computational complexity of high-resolution, long-sequence data, allowing the overall model to capture both local signal semantics and the complex circadian and inter-day variations governing physiological dynamics.
• Score: 30.599420483017227
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Wearable sensors enable the continuous acquisition of high-resolution physiological waveforms, such as photoplethysmography and accelerometry, under free-living conditions. However, inferring health-related phenotypes from these signals presents significant challenges due to high sampling frequencies, multimodal dependencies, and extreme sequence lengths (e.g., weeks of recordings), compounded by a scarcity of ground-truth labels. To address these challenges, existing self-supervised learning (SSL) methodologies typically follow two paradigms: (1) learning rich morphological representations from short waveform segments while collapsing longitudinal dynamics through simple aggregation, or (2) modeling behavioral patterns from coarse, hand-crafted features (e.g. heart rate, step counts) spanning longer horizons but foregoing subtle, predictive signatures in raw waveforms. To bridge this gap, we propose WavesFM, a foundation model utilizing a two-stage SSL framework for longitudinal physiological data. Specifically, we decompose the learning problem into two stages: first, a segment-level encoder is pretrained to extract local embeddings from short waveforms; subsequently, a temporal encoder is trained to model the sequence of these embeddings across a multi-day horizon. This hierarchical approach overcomes the computational complexity of high-resolution, long-sequence data, allowing the overall model to capture both local signal semantics and the complex circadian and inter-day variations governing physiological dynamics. Pretrained on over 6.8M hours (N=324k individuals) of recordings for the first stage and 5.3M hours (N=10k) for the second stage, WavesFM demonstrates superior performance across 58 diverse tasks spanning demographics, lifestyle, health conditions, and medications.

Related papers

• Timesynth: A Temporal Fidelity Framework for Health Signal Digital Twins [0.0]
  Time Synth is a controlled benchmarking framework for health-signal digital twins.<n>It produces signals with analytically known ground-truth dynamics from parametric models fitted to real electroencephalography, electrocardiography, and photoplethysmogram signals.<n>Time Synth supplies the controlled preclinical stress test missing before models are coupled to patient data, with a reusable generator and diagnostics for fidelity-aware development.
  arXiv  Detail & Related papers  (2026-07-01T04:42:58Z)
• SL-S4Wave: Self-Supervised Learning of Physiological Waveforms with Structured State Space Models [30.242008207142575]
  We propose SL-S4Wave, a self-supervised learning framework that combines contrastive learning with a tailored encoder built on structured state space models.<n> SL-S4Wave consistently outperforms state-of-the-art supervised and self-supervised baselines in a challenging arrhythmia detection task.<n>We additionally evaluate SL-S4Wave on multiple EEG tasks, achieving superior performance over strong baselines, demonstrating generalizability of our approach beyond cardiac waveforms.
  arXiv  Detail & Related papers  (2026-06-18T07:46:34Z)
• MedMamba: Multi-View State Space Models with Adaptive Graph Learning for Medical Time Series Classification [60.41879345743941]
  We propose MedMamba, an end-to-end architecture that integrates state space models with domain-specific inductive biases.<n>MedMamba achieves state-of-the-art performance while maintaining linear computational complexity.
  arXiv  Detail & Related papers  (2026-05-24T09:23:39Z)
• TimeLesSeg: Unified Contrast-Agnostic Cross-Sectional and Longitudinal MS Lesion Segmentation via a Stochastic Generative Model [0.338424069286569]
  TimeLesSeg is a unified contrast-agnostic framework designed to segment MS lesions regardless of the presence of a temporal dimension in its inputs.<n>Our approach models pathological priors through lesion masks, which are processed together with the current scan.<n>Cross-sectional processing is enabled by exposing the model to training cases where no prior information is available.
  arXiv  Detail & Related papers  (2026-05-08T16:19:34Z)
• FreqPhys: Repurposing Implicit Physiological Frequency Prior for Robust Remote Photoplethysmography [54.32412212867472]
  Photoplethysmography enables contactless physiological monitoring by capturing subtle skin-color variations from videos.<n>FreqPhys is a frequency-guided rulation framework that explicitly leverages physiological frequency priors for robust signal recovery.<n>It yields significant improvements over state-of-the-art approaches, particularly under challenging motion conditions.
  arXiv  Detail & Related papers  (2026-04-01T06:25:42Z)
• PhysNeXt: Next-Generation Dual-Branch Structured Attention Fusion Network for Remote Photoplethysmography Measurement [50.524262997433546]
  rmography enables measurement of heart rate and other vital signs by analyzing color variations in facial skin induced by cardiac pulsation.<n>Current methods are mainly based on either end-to-end modeling from raw videos or subtle heartbeat-related map (ST) representations.<n>We propose PhysMapXt, a dual-input deep learning framework that jointly exploits video frames and ST representations.
  arXiv  Detail & Related papers  (2026-03-20T08:37:02Z)
• SpecMoE: Spectral Mixture-of-Experts Foundation Model for Cross-Species EEG Decoding [0.0]
  SpecMoE is a mixture of experts framework guided by a learned spectral gating mechanism.<n>It achieves state-of-the-art performance across a diverse set of EEG decoding tasks, including sleep staging, emotion recognition, motor imagery classification, abnormal signal detection, and drug effect prediction.
  arXiv  Detail & Related papers  (2026-03-17T16:20:14Z)
• Conditional Neural ODE for Longitudinal Parkinson's Disease Progression Forecasting [51.906871559732245]
  Parkinson's disease (PD) shows heterogeneous, evolving brain-morphometry patterns.<n>Modeling these longitudinal trajectories enables mechanistic insight, treatment development, and individualized 'digital-twin' forecasting.<n>We propose CNODE, a novel framework for continuous, individualized PD progression forecasting.
  arXiv  Detail & Related papers  (2025-11-06T20:16:33Z)
• PhysioWave: A Multi-Scale Wavelet-Transformer for Physiological Signal Representation [14.124553708665117]
  A novel wavelet-based approach for physiological signal analysis is presented.<n>Two large-scale pretrained models specific to EMG and ECG are introduced for the first time.<n>A unified multi-modal framework is constructed by integrating pretrained EEG model.
  arXiv  Detail & Related papers  (2025-06-12T05:11:41Z)
• PSDNorm: Test-Time Temporal Normalization for Deep Learning in Sleep Staging [63.05435596565677]
  We propose PSDNorm that leverages Monge mapping and temporal context to normalize feature maps in deep learning models for signals.<n> PSDNorm achieves state-of-the-art performance on unseen left-out datasets while being 4-times more data-efficient than BatchNorm.
  arXiv  Detail & Related papers  (2025-03-06T16:20:25Z)
• Transform Once: Efficient Operator Learning in Frequency Domain [69.74509540521397]
  We study deep neural networks designed to harness the structure in frequency domain for efficient learning of long-range correlations in space or time. This work introduces a blueprint for frequency domain learning through a single transform: transform once (T1)
  arXiv  Detail & Related papers  (2022-11-26T01:56:05Z)
• Interpreting Deep Learning Models for Epileptic Seizure Detection on EEG signals [4.748221780751802]
  Deep Learning (DL) is often considered the state-of-the art for Artificial Intelligence-based medical decision support. It remains sparsely implemented in clinical practice and poorly trusted by clinicians due to insufficient interpretability of neural network models. We have tackled this issue by developing interpretable DL models in the context of online detection of epileptic seizure, based on EEG signal.
  arXiv  Detail & Related papers  (2020-12-22T11:10:23Z)
• Video-based Remote Physiological Measurement via Cross-verified Feature Disentangling [121.50704279659253]
  We propose a cross-verified feature disentangling strategy to disentangle the physiological features with non-physiological representations. We then use the distilled physiological features for robust multi-task physiological measurements. The disentangled features are finally used for the joint prediction of multiple physiological signals like average HR values and r signals.
  arXiv  Detail & Related papers  (2020-07-16T09:39:17Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.