# The Vote Hides the Failure: Aggregation Choice and Noise Robustness in Heart Murmur Detection

arXiv: https://arxiv.org/abs/2609.37161 · HTML: https://mlsanity.com/paper/2609.37161/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 5 of its 12 references that are in the map.
- First posted on arXiv 29 Sep 2026.

## Builds on (most cited first, 5 of 5)

- [Hidden Stratification Causes Clinically Meaningful Failures in Machine Learning for Medical Imaging](https://mlsanity.com/paper/1909.12475/) (2019)
- [Understanding and Mitigating the Tradeoff Between Robustness and Accuracy](https://mlsanity.com/paper/2002.10716/) (2020)
- [The CirCor DigiScope Dataset: From Murmur Detection to Murmur Classification](https://mlsanity.com/paper/2108.00813/) (2021)
- [Understanding Noise-Augmented Training for Randomized Smoothing](https://mlsanity.com/paper/2305.04746/) (2023)
- [A Markov-Switching Model Approach to Heart Sound Segmentation and Classification](https://mlsanity.com/paper/1809.03395/) (2018)

## Abstract (arXiv)

Noise robustness in automated phonocardiogram (PCG) murmur detection, and how it is measured, remains underexamined despite growing interest in low-resource screening. We evaluate two independently reimplemented pipelines, Hierarchical Multi-Scale Convolutional Network (HMS-Net)--CNN, and Bidirectional Long Short-Term Memory (BiLSTM)--LSTM, under controlled, multi-severity noise with noise-augmented fine-tuning and held-out generalization testing. Under matched aggregation, the complete BiLSTM pipeline outperforms the complete HMS-Net pipeline across all conditions in accuracy and Weighted Accuracy. A stable aggregate accuracy score can misrepresent what individual predictions show: HMS-Net's native aggregation degrades under salt-and-pepper noise far less than majority-vote (MV) aggregation at the same severity, a gap reflecting window-level disagreement its native rule absorbs, while BiLSTM's MV accuracy rises after noise-augmented training even though its individual predictions do not improve. HMS-Net's training effect is significant under one accuracy metric but not another. Noise-robustness conclusions can depend as much on evaluation choices as on the models themselves.