Presentation 2019-08-23
Estimating Exchange-level Annotations with Multitask Learning for Multimodal Dialogue Systems
Yuki Hirano, Shogo Okada, Haruto Nishimoto, Kazunori Komatani,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This study presents multimodal computational modelingfor estimating three labels: user's interest label, user's sentiment label, and topic continuance, which are independently annotated per exchange in spoken dialogue systems. A multitask learning technique is applied for binary classification tasks of the three labelsconsidering the partial similarities among them. Experimental results show the effectiveness of deep neural network trained with multitask learning framework.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Multimodal / Dialogue system / Multitask Learning / Sentiment analysis
Paper # HCS2019-32
Date of Issue 2019-08-16 (HCS)

Conference Information
Committee HCS
Conference Date 2019/8/23(2days)
Place (in Japanese) (See Japanese page)
Place (in English) Jikei Institute
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Masafumi Matsuda(NTT)
Vice Chair Tomoo Inoue(Univ. of Tsukuba) / Yugo Hayashi(Ritsumeikan Univ.)
Secretary Tomoo Inoue(Kanazawa Inst. of Tech.) / Yugo Hayashi(Osaka Electro-Comm. Univ.)
Assistant Tomoko Kanda(Osaka Inst. of Tech.) / Kazuki Takashima(Tohoku Univ.) / Ken Fujiwara(Osaka Univ. of Economic) / Kazunori Terada(Gifu Univ.) / Atsushi Kimura(Nihon Univ.) / HUANG HUNGHSUAN(Riken)

Paper Information
Registration To Technical Committee on Human Communication Science
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Estimating Exchange-level Annotations with Multitask Learning for Multimodal Dialogue Systems
Sub Title (in English)
Keyword(1) Multimodal
Keyword(2) Dialogue system
Keyword(3) Multitask Learning
Keyword(4) Sentiment analysis
1st Author's Name Yuki Hirano
1st Author's Affiliation Japan Advanced Institute of Science and Technology(JAIST)
2nd Author's Name Shogo Okada
2nd Author's Affiliation Japan Advanced Institute of Science and Technology(JAIST)
3rd Author's Name Haruto Nishimoto
3rd Author's Affiliation Osaka University(Osaka Univ.)
4th Author's Name Kazunori Komatani
4th Author's Affiliation Osaka University(Osaka Univ.)
Date 2019-08-23
Paper # HCS2019-32
Volume (vol) vol.119
Number (no) HCS-179
Page pp.pp.15-20(HCS),
#Pages 6
Date of Issue 2019-08-16 (HCS)