Discrepancy-Aware Contrastive Adaptation in Medical Time Series Analysis

2025-08-07

AI Summary

Addressing two key challenges in medical time-series disease diagnosis—(i) overfitting due to scarce labeled data and (ii) poor generalizability of existing contrastive learning methods reliant on handcrafted positive/negative sample pairs—this paper proposes LMCF, a Learnable Multi-view Contrastive Framework. First, LMCF introduces a learnable multi-view contrastive mechanism to adaptively capture disease-specific representations across clinical conditions. Second, it couples an Autoencoder-Generative Adversarial Network (AE-GAN) to model target-domain distribution shifts, integrating the resulting disease probability prior into contrastive learning. Third, it enhances temporal modeling via multi-head attention jointly with cross-view and intra-group contrastive strategies. Evaluated on three real-world datasets—myocardial infarction, Alzheimer’s disease, and Parkinson’s disease—LMCF consistently outperforms seven state-of-the-art baselines, demonstrating superior generalizability and diagnostic robustness under low-labeling-cost settings.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Computer Vision: Multi-modal Vision

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

In medical time series disease diagnosis, two key challenges are identified. First, the high annotation cost of medical data leads to overfitting in models trained on label-limited, single-center datasets. To address this, we propose incorporating external data from related tasks and leveraging AE-GAN to extract prior knowledge, providing valuable references for downstream tasks. Second, many existing studies employ contrastive learning to derive more generalized medical sequence representations for diagnostic tasks, usually relying on manually designed diverse positive and negative sample pairs. However, these approaches are complex, lack generalizability, and fail to adaptively capture disease-specific features across different conditions. To overcome this, we introduce LMCF (Learnable Multi-views Contrastive Framework), a framework that integrates a multi-head attention mechanism and adaptively learns representations from different views through inter-view and intra-view contrastive learning strategies. Additionally, the pre-trained AE-GAN is used to reconstruct discrepancies in the target data as disease probabilities, which are then integrated into the contrastive learning process. Experiments on three target datasets demonstrate that our method consistently outperforms other seven baselines, highlighting its significant impact on healthcare applications such as the diagnosis of myocardial infarction, Alzheimer's disease, and Parkinson's disease. We release the source code at xxxxx.

Problem

• High annotation cost causes overfitting in medical data models
• Existing contrastive learning lacks adaptability for disease-specific features
• Need for robust framework to improve multi-disease diagnosis accuracy

Innovation

• AE-GAN extracts prior knowledge from external data
• LMCF adaptively learns multi-view representations
• AE-GAN reconstructs discrepancies as disease probabilities