Enregistré dans:
Détails bibliographiques
Auteurs principaux: Zhao, Haixin, Yang, Kaixuan, Madhu, Nilesh
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2510.11395
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915858224250880
author Zhao, Haixin
Yang, Kaixuan
Madhu, Nilesh
author_facet Zhao, Haixin
Yang, Kaixuan
Madhu, Nilesh
contents To further reduce the complexity of lightweight speech enhancement models, we introduce a gating-based Dynamically Slimmable Network (DSN). The DSN comprises static and dynamic components. For architecture-independent applicability, we introduce distinct dynamic structures targeting the commonly used components, namely, grouped recurrent neural network units, multi-head attention, convolutional, and fully connected layers. A policy module adaptively governs the use of dynamic parts at a frame-wise resolution according to the input signal quality, controlling computational load. We further propose Metric-Guided Training (MGT) to explicitly guide the policy module in assessing input speech quality. Experimental results demonstrate that the DSN achieves comparable enhancement performance in instrumental metrics to the state-of-the-art lightweight baseline, while using only 73% of its computational load on average. Evaluations of dynamic component usage ratios indicate that the MGT-DSN can appropriately allocate network resources according to the severity of input signal distortion.
format Preprint
id arxiv_https___arxiv_org_abs_2510_11395
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Dynamically Slimmable Speech Enhancement Network with Metric-Guided Training
Zhao, Haixin
Yang, Kaixuan
Madhu, Nilesh
Audio and Speech Processing
To further reduce the complexity of lightweight speech enhancement models, we introduce a gating-based Dynamically Slimmable Network (DSN). The DSN comprises static and dynamic components. For architecture-independent applicability, we introduce distinct dynamic structures targeting the commonly used components, namely, grouped recurrent neural network units, multi-head attention, convolutional, and fully connected layers. A policy module adaptively governs the use of dynamic parts at a frame-wise resolution according to the input signal quality, controlling computational load. We further propose Metric-Guided Training (MGT) to explicitly guide the policy module in assessing input speech quality. Experimental results demonstrate that the DSN achieves comparable enhancement performance in instrumental metrics to the state-of-the-art lightweight baseline, while using only 73% of its computational load on average. Evaluations of dynamic component usage ratios indicate that the MGT-DSN can appropriately allocate network resources according to the severity of input signal distortion.
title Dynamically Slimmable Speech Enhancement Network with Metric-Guided Training
topic Audio and Speech Processing
url https://arxiv.org/abs/2510.11395