| Item type |
SIG Technical Reports(1) |
| 公開日 |
2024-06-07 |
| タイトル |
|
|
タイトル |
An Experimental Study on Model Simplification for End-to-end Automatic Speech Recognition System Based on Transformer |
| タイトル |
|
|
言語 |
en |
|
タイトル |
An Experimental Study on Model Simplification for End-to-end Automatic Speech Recognition System Based on Transformer |
| 言語 |
|
|
言語 |
eng |
| キーワード |
|
|
主題Scheme |
Other |
|
主題 |
ポスターセッション2 |
| 資源タイプ |
|
|
資源タイプ識別子 |
http://purl.org/coar/resource_type/c_18gh |
|
資源タイプ |
technical report |
| 著者所属 |
|
|
|
The University of Tokyo |
| 著者所属 |
|
|
|
The University of Tokyo |
| 著者所属 |
|
|
|
The University of Tokyo |
| 著者所属(英) |
|
|
|
en |
|
|
The University of Tokyo |
| 著者所属(英) |
|
|
|
en |
|
|
The University of Tokyo |
| 著者所属(英) |
|
|
|
en |
|
|
The University of Tokyo |
| 著者名 |
Yang, Cao
Daisuke, Saito
Nobuaki, Minematsu
|
| 著者名(英) |
Yang, Cao
Daisuke, Saito
Nobuaki, Minematsu
|
| 論文抄録 |
|
|
内容記述タイプ |
Other |
|
内容記述 |
This paper focuses on how to simplify the model of Transformer-based Automatic Speech Recognition System. For this purpose, attention mechanism in encoder parts is focused on and a scaling method for latent elements is introduced. Experimental results show that the training time for the proposed model has been reduced compared with that for a naive Transformer model. |
| 論文抄録(英) |
|
|
内容記述タイプ |
Other |
|
内容記述 |
This paper focuses on how to simplify the model of Transformer-based Automatic Speech Recognition System. For this purpose, attention mechanism in encoder parts is focused on and a scaling method for latent elements is introduced. Experimental results show that the training time for the proposed model has been reduced compared with that for a naive Transformer model. |
| 書誌レコードID |
|
|
収録物識別子タイプ |
NCID |
|
収録物識別子 |
AN10438388 |
| 書誌情報 |
研究報告音楽情報科学（MUS）
巻 2024-MUS-140,
号 44,
p. 1-5,
発行日 2024-06-07
|
| ISSN |
|
|
収録物識別子タイプ |
ISSN |
|
収録物識別子 |
2188-8752 |
| Notice |
|
|
|
SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc. |
| 出版者 |
|
|
言語 |
ja |
|
出版者 |
情報処理学会 |
IPSJ-MUS24140044.pdf (1.3 MB)
