| Item type |
SIG Technical Reports(1) |
| 公開日 |
2021-11-24 |
| タイトル |
|
|
タイトル |
Effective Integration of Transformer for Network-based Speech Emotion Recognition |
| タイトル |
|
|
言語 |
en |
|
タイトル |
Effective Integration of Transformer for Network-based Speech Emotion Recognition |
| 言語 |
|
|
言語 |
eng |
| キーワード |
|
|
主題Scheme |
Other |
|
主題 |
話者・感情認識 |
| 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 |
Yurun, He
Nobuaki, Minematsu
Daisuke, Saito
|
| 著者名(英) |
Yurun, He
Nobuaki, Minematsu
Daisuke, Saito
|
| 論文抄録 |
|
|
内容記述タイプ |
Other |
|
内容記述 |
The performance of a speech emotion recognition (SER) system heavily relies on deep representations learned from training samples. Recently, transformer has exhibited outstanding properties in learning relevant representations for this task. However, to better fuse it with conventional models, experimental investigations are still needed. In this paper, we attempt to take advantage of several integrations of transformer with two most widely used deep learning models - CNN and BLSTM. Experiments on the IEMOCAP benchmark dataset demonstrate that the proposed approaches can make a promising improvement. |
| 論文抄録(英) |
|
|
内容記述タイプ |
Other |
|
内容記述 |
The performance of a speech emotion recognition (SER) system heavily relies on deep representations learned from training samples. Recently, transformer has exhibited outstanding properties in learning relevant representations for this task. However, to better fuse it with conventional models, experimental investigations are still needed. In this paper, we attempt to take advantage of several integrations of transformer with two most widely used deep learning models - CNN and BLSTM. Experiments on the IEMOCAP benchmark dataset demonstrate that the proposed approaches can make a promising improvement. |
| 書誌レコードID |
|
|
収録物識別子タイプ |
NCID |
|
収録物識別子 |
AN10442647 |
| 書誌情報 |
研究報告音声言語情報処理（SLP）
巻 2021-SLP-139,
号 7,
p. 1-6,
発行日 2021-11-24
|
| ISSN |
|
|
収録物識別子タイプ |
ISSN |
|
収録物識別子 |
2188-8663 |
| 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-SLP21139007.pdf (1.2 MB)
