Cover Image

Transfer learning from speech to music: towards language-sensitive emotion recognition models

Main Authors: Gómez-Cañón, Juan Sebastián, Cano, Estefanía, Herrera, Perfecto, Gómez, Emilia
Format: Proceeding
Bahasa: eng
Terbitan: , 2020
Subjects:
sparse convolutional autoencoder
speech emotion recognition
music emotion recognition
unsupervised learning
multi-task learning
Online Access: https://zenodo.org/record/4076791
ctrlnum 4076791
fullrecord <?xml version="1.0"?> <dc schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd"><creator>G&#xF3;mez-Ca&#xF1;&#xF3;n, Juan Sebasti&#xE1;n</creator><creator>Cano, Estefan&#xED;a</creator><creator>Herrera, Perfecto</creator><creator>G&#xF3;mez, Emilia</creator><date>2020-10-05</date><description>In this study, we address emotion recognition using unsupervised feature learning from speech data, and test its transferability to music. Our approach is to pre-train models using speech in English and Mandarin, and then fine-tune them with excerpts of music labeled with categories of emotion. Our initial hypothesis is that features automatically learned from speech should be transferable to music. Namely, we expect the intra-linguistic setting (e.g., pre-training on speech in English and fine-tuning on music in English) should result in improved performance over the cross-linguistic setting (e.g., pre-training on speech in English and fine-tuning on music in Mandarin). Our results confirm previous research on cross-domain transferability, and encourage research towards language-sensitive Music Emotion Recognition (MER) models.</description><identifier>https://zenodo.org/record/4076791</identifier><identifier>10.5281/zenodo.4076791</identifier><identifier>oai:zenodo.org:4076791</identifier><language>eng</language><relation>info:eu-repo/grantAgreement/EC/H2020/770376/</relation><relation>doi:10.5281/zenodo.4076790</relation><rights>info:eu-repo/semantics/openAccess</rights><rights>https://creativecommons.org/licenses/by/4.0/legalcode</rights><subject>sparse convolutional autoencoder</subject><subject>speech emotion recognition</subject><subject>music emotion recognition</subject><subject>unsupervised learning</subject><subject>multi-task learning</subject><title>Transfer learning from speech to music: towards language-sensitive emotion recognition models</title><type>Journal:Proceeding</type><type>Journal:Proceeding</type><recordID>4076791</recordID></dc>
language eng
format Journal:Proceeding
Journal
author Gómez-Cañón, Juan Sebastián
Cano, Estefanía
Herrera, Perfecto
Gómez, Emilia
title Transfer learning from speech to music: towards language-sensitive emotion recognition models
publishDate 2020
topic sparse convolutional autoencoder
speech emotion recognition
music emotion recognition
unsupervised learning
multi-task learning
url https://zenodo.org/record/4076791
contents In this study, we address emotion recognition using unsupervised feature learning from speech data, and test its transferability to music. Our approach is to pre-train models using speech in English and Mandarin, and then fine-tune them with excerpts of music labeled with categories of emotion. Our initial hypothesis is that features automatically learned from speech should be transferable to music. Namely, we expect the intra-linguistic setting (e.g., pre-training on speech in English and fine-tuning on music in English) should result in improved performance over the cross-linguistic setting (e.g., pre-training on speech in English and fine-tuning on music in Mandarin). Our results confirm previous research on cross-domain transferability, and encourage research towards language-sensitive Music Emotion Recognition (MER) models.
id IOS16997.4076791
institution DEFAULT
institution_type library:public
library
library DEFAULT
collection DEFAULT
city DEFAULT
province DEFAULT
repoId IOS16997
first_indexed 2022-06-06T04:56:03Z
last_indexed 2022-06-06T04:56:03Z
recordtype dc
merged_child_boolean 1
_version_ 1739400927514722304
score 17.608969

Lihat Juga