Presentation 2019-09-29
Hidden-word detection based on distributed representations and the topic model
Wei Yilun, Lao Yingying, Han Dongli,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In natural language analysis, there is a phenomenon called hidden-word that can be an obstacle. Most previous studies deal with hidden-words as one type of unknown words. However, the approach to detect hidden words as replacements for other words is not sufficient. In this research, by employing the topic model and word2vec to obtain distributed expressions of words, we have proposed a method to automatically detect hidden-words from a sentence by estimating the compatibility between each word and the topic of the sentence. Also, we have automatically created test sentences containing hidden-words and conducted an evaluation experiment. The experimental results have shown the effectiveness of our approach.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Hidden-word / Word2vec / Topic-model
Paper # TL2019-39
Date of Issue 2019-09-22 (TL)

Conference Information
Committee TL
Conference Date 2019/9/29(1days)
Place (in Japanese) (See Japanese page)
Place (in English) National Institute of Technology, Hakodate College
Topics (in Japanese) (See Japanese page)
Topics (in English) Theme 1 Arts and Language, Theme 2 Ba linguistics, Abduction of meaning, "Ba" in Co-Creation, Abduction and Innovation, Theme 3 Language and Thought
Chair Hiroshi Sano(Tokyo Univ. of Foreign Studies)
Vice Chair Tadahisa Kondo(Kogakuin Univ.) / Kazuhiro Takeuchi(Osaka Electro-Comm. Univ.)
Secretary Tadahisa Kondo(Kobe Gakuin Univ.) / Kazuhiro Takeuchi(Kyoto Inst. of Tech.)
Assistant Nobuyuki Jincho(Miidas) / Akinori Takada(Ferris Univ.) / Akio Ishikawa(KDDI Research)

Paper Information
Registration To Technical Committee on Thought and Language
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Hidden-word detection based on distributed representations and the topic model
Sub Title (in English)
Keyword(1) Hidden-word
Keyword(2) Word2vec
Keyword(3) Topic-model
1st Author's Name Wei Yilun
1st Author's Affiliation Nihon University(Nihon Univ.)
2nd Author's Name Lao Yingying
2nd Author's Affiliation Nihon University(Nihon Univ.)
3rd Author's Name Han Dongli
3rd Author's Affiliation Nihon University(Nihon Univ.)
Date 2019-09-29
Paper # TL2019-39
Volume (vol) vol.119
Number (no) TL-213
Page pp.pp.11-15(TL),
#Pages 5
Date of Issue 2019-09-22 (TL)