File Download
There are no files associated with this item.
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1080/09298215.2017.1367820
- Scopus: eid_2-s2.0-85032688233
- WOS: WOS:000427945500004
- Find via
Supplementary
- Citations:
- Appears in Collections:
Article: Large Vocabulary Automatic Chord Estimation Using Bidirectional Long Short-term Memory Recurrent Neural Network With Even Chance Training
Title | Large Vocabulary Automatic Chord Estimation Using Bidirectional Long Short-term Memory Recurrent Neural Network With Even Chance Training |
---|---|
Authors | |
Keywords | Automatic chord estimation Deep learning Large vocabulary Music information retrieval Recurrent neural network |
Issue Date | 2017 |
Publisher | Routledge. The Journal's web site is located at http://www.tandf.co.uk/journals/titles/09298215.asp |
Citation | Journal of New Music Research, 2017, v. 47 n. 1, p. 53-67 How to Cite? |
Abstract | This paper presents an argument for the necessity of a large vocabulary in automatic chord recognition systems, on the grounds of the requirements of machine musicianship. It proposes a system framework with a skewed class-sensitive training scheme that leads to a preliminary solution to large vocabulary automatic chord estimation. This framework applies a bidirectional long short-term memory recurrent neural network architecture, which employs an ‘even chance’ training scheme to make up for the lack of uncommon chords’ exposure. The main drawback of this approach is the low segmentation quality, which inevitably lowers the upper bound of chord estimation accuracy. Under a large vocabulary evaluation, the proposed system can significantly outperform the baseline system in terms of the overall weighted chord symbol recall, and there is no significant difference between them in terms of average chord quality accuracy. The results demonstrate preliminary success in our approach, and also prove the even chance training scheme to be effective in boosting uncommon chord symbol recalls as well as the average chord quality accuracy. |
Persistent Identifier | http://hdl.handle.net/10722/260320 |
ISSN | 2023 Impact Factor: 1.1 2023 SCImago Journal Rankings: 0.389 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Deng, J | - |
dc.contributor.author | Kwok, YK | - |
dc.date.accessioned | 2018-09-14T08:39:41Z | - |
dc.date.available | 2018-09-14T08:39:41Z | - |
dc.date.issued | 2017 | - |
dc.identifier.citation | Journal of New Music Research, 2017, v. 47 n. 1, p. 53-67 | - |
dc.identifier.issn | 0929-8215 | - |
dc.identifier.uri | http://hdl.handle.net/10722/260320 | - |
dc.description.abstract | This paper presents an argument for the necessity of a large vocabulary in automatic chord recognition systems, on the grounds of the requirements of machine musicianship. It proposes a system framework with a skewed class-sensitive training scheme that leads to a preliminary solution to large vocabulary automatic chord estimation. This framework applies a bidirectional long short-term memory recurrent neural network architecture, which employs an ‘even chance’ training scheme to make up for the lack of uncommon chords’ exposure. The main drawback of this approach is the low segmentation quality, which inevitably lowers the upper bound of chord estimation accuracy. Under a large vocabulary evaluation, the proposed system can significantly outperform the baseline system in terms of the overall weighted chord symbol recall, and there is no significant difference between them in terms of average chord quality accuracy. The results demonstrate preliminary success in our approach, and also prove the even chance training scheme to be effective in boosting uncommon chord symbol recalls as well as the average chord quality accuracy. | - |
dc.language | eng | - |
dc.publisher | Routledge. The Journal's web site is located at http://www.tandf.co.uk/journals/titles/09298215.asp | - |
dc.relation.ispartof | Journal of New Music Research | - |
dc.rights | Preprint: This is an Author's Original Manuscript of an article published by Taylor & Francis Group in [JOURNAL TITLE] on [date of publication], available online: http://www.tandfonline.com/doi/abs/[Article DOI]. Postprint: This is an Accepted Manuscript of an article published by Taylor & Francis Group in [JOURNAL TITLE] on [date of publication], available online at: http://www.tandfonline.com/doi/abs/[Article DOI]. | - |
dc.subject | Automatic chord estimation | - |
dc.subject | Deep learning | - |
dc.subject | Large vocabulary | - |
dc.subject | Music information retrieval | - |
dc.subject | Recurrent neural network | - |
dc.title | Large Vocabulary Automatic Chord Estimation Using Bidirectional Long Short-term Memory Recurrent Neural Network With Even Chance Training | - |
dc.type | Article | - |
dc.identifier.email | Kwok, YK: ykwok@hku.hk | - |
dc.identifier.authority | Kwok, YK=rp00128 | - |
dc.description.nature | link_to_subscribed_fulltext | - |
dc.identifier.doi | 10.1080/09298215.2017.1367820 | - |
dc.identifier.scopus | eid_2-s2.0-85032688233 | - |
dc.identifier.hkuros | 291040 | - |
dc.identifier.volume | 47 | - |
dc.identifier.issue | 1 | - |
dc.identifier.spage | 53 | - |
dc.identifier.epage | 67 | - |
dc.identifier.isi | WOS:000427945500004 | - |
dc.publisher.place | United Kingdom | - |
dc.identifier.issnl | 0929-8215 | - |