dc.contributor.author | Choi, K | en_US |
dc.contributor.author | Fazekas, G | en_US |
dc.contributor.author | Sandler, M | en_US |
dc.contributor.author | Cho, K | en_US |
dc.date.accessioned | 2017-12-15T11:45:49Z | |
dc.date.available | 2017-10-23 | en_US |
dc.date.issued | 2018-03-23 | en_US |
dc.date.submitted | 2017-12-11T05:04:03.339Z | |
dc.identifier.other | 2 | en_US |
dc.identifier.other | 2 | en_US |
dc.identifier.other | 2 | en_US |
dc.identifier.uri | http://qmro.qmul.ac.uk/xmlui/handle/123456789/30523 | |
dc.description | date-added: 2018-06-06 23:32:25 +0000 date-modified: 2018-05-06 23:32:25 +0000 keywords: evaluation, music tagging, deep learning, CNN bdsk-url-1: https://arxiv.org/pdf/1706.02361.pdf bdsk-url-2: https://dx.doi.org/10.1109/TETCI.2017.2771298 | en_US |
dc.description | date-added: 2018-06-06 23:32:25 +0000 date-modified: 2018-05-06 23:32:25 +0000 keywords: evaluation, music tagging, deep learning, CNN bdsk-url-1: https://arxiv.org/pdf/1706.02361.pdf bdsk-url-2: https://dx.doi.org/10.1109/TETCI.2017.2771298 | en_US |
dc.description | date-added: 2018-06-06 23:32:25 +0000 date-modified: 2018-05-06 23:32:25 +0000 keywords: evaluation, music tagging, deep learning, CNN bdsk-url-1: https://arxiv.org/pdf/1706.02361.pdf bdsk-url-2: https://dx.doi.org/10.1109/TETCI.2017.2771298 | en_US |
dc.description.abstract | Deep neural networks (DNN) have been successfully applied to music classification including music tagging. However, there are several open questions regarding the training, evaluation, and analysis of DNNs. In this article, we investigate specific aspects of neural networks, the effects of noisy labels, to deepen our understanding of their properties. We analyse and (re-)validate a large music tagging dataset to investigate the reliability of training and evaluation. Using a trained network, we compute label vector similarities which is compared to groundtruth similarity. The results highlight several important aspects of music tagging and neural networks. We show that networks can be effective despite relatively large error rates in groundtruth datasets, while conjecturing that label noise can be the cause of varying tag-wise performance differences. Lastly, the analysis of our trained network provides valuable insight into the relationships between music tags. These results highlight the benefit of using data-driven methods to address automatic music tagging. | en_US |
dc.format.extent | 139 - 149 | en_US |
dc.language | English | en_US |
dc.publisher | IEEE | en_US |
dc.relation.ispartof | IEEE Transactions on Emerging Topics in Computational Intelligence | en_US |
dc.subject | deep learning | en_US |
dc.subject | CNN | en_US |
dc.subject | evaluation | en_US |
dc.subject | music tagging | en_US |
dc.title | The Effects of Noisy Labels on Deep Convolutional Neural Networks for Music Tagging | en_US |
dc.type | Article | |
dc.rights.holder | © 2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. | |
dc.identifier.doi | 10.1109/TETCI.2017.2771298 | en_US |
pubs.author-url | http://semanticaudio.net/ | en_US |
pubs.issue | 2 | en_US |
pubs.notes | Not known | en_US |
pubs.publication-status | Published | en_US |
pubs.publisher-url | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8323324 | en_US |
pubs.volume | 2 | en_US |
dcterms.dateAccepted | 2017-10-23 | en_US |
qmul.funder | Fusing Semantic and Audio Technologies for Intelligent Music Production and Consumption::Engineering and Physical Sciences Research Council | en_US |
qmul.funder | Fusing Semantic and Audio Technologies for Intelligent Music Production and Consumption::Engineering and Physical Sciences Research Council | en_US |