Single-channel speech enhancement using learnable loss mixup

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chang, Oscar, Tran, Dung N., Koishida, Kazuhito
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909056745078784
author Chang, Oscar
Tran, Dung N.
Koishida, Kazuhito
author_facet Chang, Oscar
Tran, Dung N.
Koishida, Kazuhito
contents Generalization remains a major problem in supervised learning of single-channel speech enhancement. In this work, we propose learnable loss mixup (LLM), a simple and effortless training diagram, to improve the generalization of deep learning-based speech enhancement models. Loss mixup, of which learnable loss mixup is a special variant, optimizes a mixture of the loss functions of random sample pairs to train a model on virtual training data constructed from these pairs of samples. In learnable loss mixup, by conditioning on the mixed data, the loss functions are mixed using a non-linear mixing function automatically learned via neural parameterization. Our experimental results on the VCTK benchmark show that learnable loss mixup achieves 3.26 PESQ, outperforming the state-of-the-art.
format Preprint
id arxiv_https___arxiv_org_abs_2312_17255
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Single-channel speech enhancement using learnable loss mixup
Chang, Oscar
Tran, Dung N.
Koishida, Kazuhito
Audio and Speech Processing
Machine Learning
Sound
Generalization remains a major problem in supervised learning of single-channel speech enhancement. In this work, we propose learnable loss mixup (LLM), a simple and effortless training diagram, to improve the generalization of deep learning-based speech enhancement models. Loss mixup, of which learnable loss mixup is a special variant, optimizes a mixture of the loss functions of random sample pairs to train a model on virtual training data constructed from these pairs of samples. In learnable loss mixup, by conditioning on the mixed data, the loss functions are mixed using a non-linear mixing function automatically learned via neural parameterization. Our experimental results on the VCTK benchmark show that learnable loss mixup achieves 3.26 PESQ, outperforming the state-of-the-art.
title Single-channel speech enhancement using learnable loss mixup
topic Audio and Speech Processing
Machine Learning
Sound
url https://arxiv.org/abs/2312.17255