Robust Cross-Domain Speech-Based Alzheimer's Disease Detection via Iterative Adversarial Self-Training

2026-09-12

AI Summary

为解决跨域环境下阿尔茨海默病语音检测性能下降的问题，提出了一种迭代对抗自训练(IAST)方法，以学习鲁棒的、域不变的特征表示。

Abstract

As Alzheimer's disease (AD) has increasingly become a major global public health issue, speech-based AD detection has attracted widespread attention. However, most existing methods are trained and evaluated on a single dataset, often leading to severe cross-domain performance degradation due to reliance on dataset-specific artifacts rather than disease-related speech cues. In real-world applications, reliable Alzheimer's disease detection requires models that are robust to variations in recording environments, speakers and data collection conditions. To address this challenge, this paper adopts unsupervised domain adaptation to learn robust, domain-invariant feature representations in the absence of target-domain diagnosis labels. On this basis, a novel unsupervised domain adaptation method, Iterative Adversarial Self-Training (IAST), is proposed. Results demonstrate that IAST significantly improves the generalization ability and robustness under various cross-domain settings.

Problem

• Alzheimer's disease
• cross-domain performance
• unsupervised domain adaptation

Innovation

• Unsupervised Domain Adaptation
• Iterative Adversarial Self-Training
• Cross-Domain Robustness