Title:
VECTOR GENERATION DEVICE, SENTENCE PAIR LEARNING DEVICE, VECTOR GENERATION METHOD, SENTENCE PAIR LEARNING METHOD, AND PROGRAM
Document Type and Number:
WIPO Patent Application WO/2019/168202
Kind Code:
A1
Abstract:
The present invention makes it possible to generate an accurate word vector even without limiting the vocabulary of a word vector data collection. A vector generation device 10 generates a vector representing an input sentence P. When a definition sentence considering context encoding unit 280 generates a sequence of vectors representing the input sentence P on the basis of a vector for each word included in the input sentence P, the definition sentence considering context encoding unit, on the basis of a dictionary database 230 storing a set of an entry word "y" and a definition sentence Dy of the entry word "y", generates a sequence of vectors representing the input sentence P for a word, among words included in the input sentence P, that is an entry word stored in the dictionary database, such generation carried out using the definition sentence Dy of the entry word "y".
Inventors:
NISHIDA KOSUKE (JP)
NISHIDA KYOSUKE (JP)
ASANO HISAKO (JP)
TOMITA JUNJI (JP)
NISHIDA KYOSUKE (JP)
ASANO HISAKO (JP)
TOMITA JUNJI (JP)
Application Number:
PCT/JP2019/008473
Publication Date:
September 06, 2019
Filing Date:
March 04, 2019
Export Citation:
Assignee:
NIPPON TELEGRAPH & TELEPHONE (JP)
International Classes:
G06F16/332
Other References:
BAHDANAU, DZMITRY ET AL.: "Learning to compute word embeddings on the fly", CORR, ABS/1706.00286, 2017, pages 1 - 12, XP080766832, Retrieved from the Internet [retrieved on 20190411]
HILL, FELIX ET AL.: "Learning to understand phrases by embedding the dictionary", TRANSACTIONS OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, vol. 4, 2016, pages 17 - 30, XP055633933
CHEN, QIAN ET AL.: "Enhanced LSTM for Natural Language Inference", PROCEEDINGS OF THE 55TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, 2017, pages 1657 - 1668, XP055555645, doi:10.18653/v1/P17-1152
HILL, FELIX ET AL.: "Learning to understand phrases by embedding the dictionary", TRANSACTIONS OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, vol. 4, 2016, pages 17 - 30, XP055633933
CHEN, QIAN ET AL.: "Enhanced LSTM for Natural Language Inference", PROCEEDINGS OF THE 55TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, 2017, pages 1657 - 1668, XP055555645, doi:10.18653/v1/P17-1152
Attorney, Agent or Firm:
TAIYO, NAKAJIMA & KATO (JP)
Download PDF: