DAAC: Discrepancy-Aware Adaptive Contrastive Learning for Medical Time series

Alan Liang (National University of Singapore) · Jiahua Dong (Mohamed bin Zayed University of Artificial Intelligence) · Yifan WANG (The Chinese University of Hong Kong, Shenzhen) · Hongfeng Ai (The Chinese University of Hong Kong) · ruiqi li (University of the Chinese Academy of Sciences) · Maowei Jiang (Tsinghua University, Tsinghua University) · Quangao Liu (University of Exeter) · ruiyuan kang (Technology Innovation Institute) · Zihang Wang (University of the Chinese Academy of Sciences) · ruikai liu (CAS) · Cheng Jiang (The Chinese University of Hong Kong) · Chenzhong Li (The Chinese University of Hong Kong)

adaptive contrastive learneralzheimer’s diseaseclinical datasetsdiscrepancy featuresdiscrepancy-aware adaptive contrastive learningdiscriminative representationsgan-enhanced encoder-decodermodel generalization multi-head attentionmulti-view contrastive frameworkmyocardial infarctionoverfittingparkinson’s diseasereconstruction errorstime-series data

Medical time-series data play a vital role in disease diagnosis but suffer from limited labeled samples and single-center bias, which hinder model generalization and lead to overfitting. To address these challenges, we propose DAAC (Discrepancy-Aware Adaptive Contrastive learning), a learnable multi-view contrastive framework that integrates external normal samples and enhances feature learning through adaptive contrastive strategies. DAAC consists of two key modules: (1) a Discrepancy Estimator, built upon a GAN-enhanced encoder-decoder architecture, captures the distribution of normal data and computes reconstruction errors as indicators of abnormality. These discrepancy features augment the target dataset to mitigate overfitting. (2) an Adaptive Contrastive Learner uses multi-head attention to extract discriminative representations by contrasting embeddings across multiple views and data granularities (subject, trial, epoch, and temporal levels), eliminating the need for handcrafted positive-negative sample pairs. Extensive experiments on three clinical datasets—covering Alzheimer’s disease, Parkinson’s disease, and myocardial infarction—demonstrate that DAAC significantly outperforms existing methods, even when only 10\% of labeled data is available, showing strong generalization and diagnostic performance. Our code is available at https://github.com/CUHKSZ-MED-BioE/DAAC.