# Rethinking Batch Composition: Promoting Feature Diversity for Speech Enhancement

arXiv: https://arxiv.org/abs/2610.21803 · HTML: https://mlsanity.com/paper/2610.21803/

## Key facts

- No citing papers in the map yet (data through 19 Oct 2026).
- Placed in Speech enhancement and noise suppression, one of 11 directions in Speech and audio language models.
- Builds on 20 of its 35 references that are in the map.
- First posted on arXiv 15 Oct 2026.

## Placed in

- Speech and audio language models / [Speech enhancement and noise suppression](https://mlsanity.com/speech/speech-enhancement-and-noise-suppression/)

## Builds on (most cited first, 20 of 20)

- [Score-Based Generative Modeling through Stochastic Differential Equations](https://mlsanity.com/paper/2011.13456/) (2020)
- [SDR - half-baked or well done?](https://mlsanity.com/paper/1811.02508/) (2018)
- [Speech Enhancement and Dereverberation with Diffusion-based Generative Models](https://mlsanity.com/paper/2208.05830/) (2022)
- [A consolidated view of loss functions for supervised deep learning-based speech enhancement](https://mlsanity.com/paper/2009.12286/) (2020)
- [Analysing Diffusion-based Generative Approaches versus Discriminative Approaches for Speech Restoration](https://mlsanity.com/paper/2211.02397/) (2022)
- [Schr"odinger Bridge for Generative Speech Enhancement](https://mlsanity.com/paper/2407.16074/) (2024)
- [Interspeech 2025 URGENT Speech Enhancement Challenge](https://mlsanity.com/paper/2505.23212/) (2025)
- [FlowSE: Flow Matching-based Speech Enhancement](https://mlsanity.com/paper/2508.06840/) (2025)
- [Assessing the Generalization Gap of Learning-Based Speech Enhancement Systems in Noisy and Reverberant Environments](https://mlsanity.com/paper/2309.06183/) (2023)
- [Sparse Mixture of Local Experts for Efficient Speech Enhancement](https://mlsanity.com/paper/2005.08128/) (2020)
- [Universal Speech Enhancement with Regression and Generative Mamba](https://mlsanity.com/paper/2505.21198/) (2025)
- [Lessons Learned from the URGENT 2024 Speech Enhancement Challenge](https://mlsanity.com/paper/2506.01611/) (2025)
- [Personalized Speech Enhancement Without a Separate Speaker Embedding Model](https://mlsanity.com/paper/2406.09928/) (2024)
- [Improving Speech Enhancement with Multi-Metric Supervision from Learned Quality Assessment](https://mlsanity.com/paper/2506.12260/) (2025)
- [PercepNet+: A Phase and SNR Aware PercepNet for Real-Time Speech Enhancement](https://mlsanity.com/paper/2203.02263/) (2022)
- [Variational Autoencoder for Personalized Pathological Speech Enhancement](https://mlsanity.com/paper/2503.14036/) (2025)
- [Navigating PESQ: Up-to-Date Versions and Open Implementations](https://mlsanity.com/paper/2505.19760/) (2025)
- [FUSE: Universal Speech Enhancement using Multi-Stage Fusion of Sparse Compression and Token Generation Models for the URGENT 2025 Challenge](https://mlsanity.com/paper/2506.00809/) (2025)
- [Rethinking Training Targets, Architectures and Data Quality for Universal Speech Enhancement](https://mlsanity.com/paper/2603.02641/) (2026)
- [Influence of Clean Speech Characteristics on Speech Enhancement Performance](https://mlsanity.com/paper/2509.18885/) (2025)

## Abstract (arXiv)

Speech enhancement systems are commonly trained using randomly sampled mini-batches, which may contain limited diversity in intrinsic acoustic characteristics of the clean speech signals. Motivated by previous findings linking clean speech characteristics to enhancement performance, we propose a feature-stratified mini-batch sampling strategy that divides training examples into quantiles according to a clean-speech characteristic and enforces equal representation within each mini-batch. Experiments across two large-scale English and Spanish speech datasets and a smaller dataset containing neurotypical and pathological speech demonstrate improvements over standard random mini-batch sampling across multiple objective enhancement metrics. Ablation studies demonstrate that the gains are robust across training runs, while their magnitude depends on the acoustic characteristic used for stratification. The proposed strategy requires no modification to the enhancement model or training objective, providing a simple and complementary approach to speech enhancement training.