Cargando…
Online Continual Learning in Acoustic Scene Classification: An Empirical Study
Numerous deep learning methods for acoustic scene classification (ASC) have been proposed to improve the classification accuracy of sound events. However, only a few studies have focused on continual learning (CL) wherein a model continually learns to solve issues with task changes. Therefore, in th...
Autores principales: | , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
MDPI
2023
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10422258/ https://www.ncbi.nlm.nih.gov/pubmed/37571676 http://dx.doi.org/10.3390/s23156893 |
_version_ | 1785089162385293312 |
---|---|
author | Ha, Donghee Kim, Mooseop Jeong, Chi Yoon |
author_facet | Ha, Donghee Kim, Mooseop Jeong, Chi Yoon |
author_sort | Ha, Donghee |
collection | PubMed |
description | Numerous deep learning methods for acoustic scene classification (ASC) have been proposed to improve the classification accuracy of sound events. However, only a few studies have focused on continual learning (CL) wherein a model continually learns to solve issues with task changes. Therefore, in this study, we systematically analyzed the performance of ten recent CL methods to provide guidelines regarding their performances. The CL methods included two regularization-based methods and eight replay-based methods. First, we defined realistic and difficult scenarios such as online class-incremental (OCI) and online domain-incremental (ODI) cases for three public sound datasets. Then, we systematically analyzed the performance of each CL method in terms of average accuracy, average forgetting, and training time. In OCI scenarios, iCaRL and SCR showed the best performance for small buffer sizes, and GDumb showed the best performance for large buffer sizes. In ODI scenarios, SCR adopting supervised contrastive learning consistently outperformed the other methods, regardless of the memory buffer size. Most replay-based methods have an almost constant training time, regardless of the memory buffer size, and their performance increases with an increase in the memory buffer size. Based on these results, we must first consider GDumb/SCR for the continual learning methods for ASC. |
format | Online Article Text |
id | pubmed-10422258 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2023 |
publisher | MDPI |
record_format | MEDLINE/PubMed |
spelling | pubmed-104222582023-08-13 Online Continual Learning in Acoustic Scene Classification: An Empirical Study Ha, Donghee Kim, Mooseop Jeong, Chi Yoon Sensors (Basel) Article Numerous deep learning methods for acoustic scene classification (ASC) have been proposed to improve the classification accuracy of sound events. However, only a few studies have focused on continual learning (CL) wherein a model continually learns to solve issues with task changes. Therefore, in this study, we systematically analyzed the performance of ten recent CL methods to provide guidelines regarding their performances. The CL methods included two regularization-based methods and eight replay-based methods. First, we defined realistic and difficult scenarios such as online class-incremental (OCI) and online domain-incremental (ODI) cases for three public sound datasets. Then, we systematically analyzed the performance of each CL method in terms of average accuracy, average forgetting, and training time. In OCI scenarios, iCaRL and SCR showed the best performance for small buffer sizes, and GDumb showed the best performance for large buffer sizes. In ODI scenarios, SCR adopting supervised contrastive learning consistently outperformed the other methods, regardless of the memory buffer size. Most replay-based methods have an almost constant training time, regardless of the memory buffer size, and their performance increases with an increase in the memory buffer size. Based on these results, we must first consider GDumb/SCR for the continual learning methods for ASC. MDPI 2023-08-03 /pmc/articles/PMC10422258/ /pubmed/37571676 http://dx.doi.org/10.3390/s23156893 Text en © 2023 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). |
spellingShingle | Article Ha, Donghee Kim, Mooseop Jeong, Chi Yoon Online Continual Learning in Acoustic Scene Classification: An Empirical Study |
title | Online Continual Learning in Acoustic Scene Classification: An Empirical Study |
title_full | Online Continual Learning in Acoustic Scene Classification: An Empirical Study |
title_fullStr | Online Continual Learning in Acoustic Scene Classification: An Empirical Study |
title_full_unstemmed | Online Continual Learning in Acoustic Scene Classification: An Empirical Study |
title_short | Online Continual Learning in Acoustic Scene Classification: An Empirical Study |
title_sort | online continual learning in acoustic scene classification: an empirical study |
topic | Article |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10422258/ https://www.ncbi.nlm.nih.gov/pubmed/37571676 http://dx.doi.org/10.3390/s23156893 |
work_keys_str_mv | AT hadonghee onlinecontinuallearninginacousticsceneclassificationanempiricalstudy AT kimmooseop onlinecontinuallearninginacousticsceneclassificationanempiricalstudy AT jeongchiyoon onlinecontinuallearninginacousticsceneclassificationanempiricalstudy |