English

Multi-View Contrastive Learning for Robust Domain Adaptation in Medical Time Series Analysis

Machine Learning 2025-09-23 v2 Artificial Intelligence

Abstract

Adapting machine learning models to medical time series across different domains remains a challenge due to complex temporal dependencies and dynamic distribution shifts. Current approaches often focus on isolated feature representations, limiting their ability to fully capture the intricate temporal dynamics necessary for robust domain adaptation. In this work, we propose a novel framework leveraging multi-view contrastive learning to integrate temporal patterns, derivative-based dynamics, and frequency-domain features. Our method employs independent encoders and a hierarchical fusion mechanism to learn feature-invariant representations that are transferable across domains while preserving temporal coherence. Extensive experiments on diverse medical datasets, including electroencephalogram (EEG), electrocardiogram (ECG), and electromyography (EMG) demonstrate that our approach significantly outperforms state-of-the-art methods in transfer learning tasks. By advancing the robustness and generalizability of machine learning models, our framework offers a practical pathway for deploying reliable AI systems in diverse healthcare settings.

Keywords

Cite

@article{arxiv.2506.22393,
  title  = {Multi-View Contrastive Learning for Robust Domain Adaptation in Medical Time Series Analysis},
  author = {YongKyung Oh and Alex Bui},
  journal= {arXiv preprint arXiv:2506.22393},
  year   = {2025}
}

Comments

Published at the sixth Conference on Health, Inference, and Learning (CHIL 2025), PMLR 287:502-526, 2025. Models & Methods Track - Best Paper Award. https://proceedings.mlr.press/v287/oh25a.html