arXiv cs.AI· Wensi Zhang, Tomas Teijeiro, J\'er\^ome Thevenot, David Atienza·· 4 小时前AI 评分34
基于机器学习的咳嗽模型为何无法泛化:结核病筛查的跨数据集系统评估
Why ML-based cough models do not generalize: a systematic cross-dataset evaluation for tuberculosis screening
AI 导读
研究对经典机器学习和深度学习咳嗽结核病分类器在三个独立数据集上做跨数据集评估,数据集内 ROC-AUC 最高仅 0.76±0.06,外部性能常低于 0.6,两类流程均无法泛化。音频表征按录音设备和数据集而非结核病状态聚类,预测概率随 CODA 中各国患病率变化,设备不匹配会降低迁移效果。临床变量基线泛化更稳定(ROC-AUC 0.65-0.71),说明采集环节的差异是泛化差的主要驱动因素。
正文 · 原文
Abstract:Cough acoustics are promising for non-invasive tuberculosis (TB) screening, yet whether machine learning (ML) models capture disease-related acoustics or artifacts of data collection remains unresolved. We evaluated the cross-dataset generalizability of classical ML and deep learning (DL) cough-based TB classifiers across three independent datasets. Despite moderate within-dataset performance (ROC-AUC up to $0.76 \pm 0.06$), both pipelines fail to generalize, with external performance frequently below 0.6, indicating a possible limitation of the data. We further observed that audio representations are organized by recording device and dataset rather than TB status, predicted TB probability tracks country-level prevalence in CODA, and device mismatch degrades transfer while device-diverse training improves it. Additionally, a clinical-variable baseline generalizes more consistently (ROC-AUC $0.65-0.71$), indicating acquisition-specific variability in acoustics is a stronger driver of poor generalizability. High within-dataset performance is not enough. External validation is essential before cough-based TB models are clinically ready.
| Subjects: | Audio and Speech Processing (eess.AS); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Sound (cs.SD); Signal Processing (eess.SP) |
| Cite as: | arXiv:2608.25846 [eess.AS] |
| (or arXiv:2608.25846v2 [eess.AS] for this version) | |
| https://doi.org/10.48550/arXiv.2608.25846 arXiv-issued DOI via DataCite |
Submission history
From: Wensi Zhang [view email]
[v1]
Wed, 26 Aug 2026 14:22:32 UTC (6,844 KB)
[v2]
Tue, 17 Nov 2026 16:04:39 UTC (8,661 KB)
来源:arXiv cs.AI · arxiv.org