Title:
LANGUAGE MODEL LEARNING DEVICE, INTERACTION DEVICE, AND TRAINED LANGUAGE MODEL
Document Type and Number:
WIPO Patent Application WO/2023/162513
Kind Code:
A1
Abstract:
This language model learning device, for providing a language model learning device whereby a large-scale language model can be learned with low computational cost independent of speech synthesis and speech recognition performance, includes a conversion means for converting text of natural language to output a symbol string of phonetic symbols, and a learning means for performing learning of a language model using the text and the symbol string outputted by the conversion means.
Inventors:
OH JONGHOON (JP)
ASAO YOSHIHIKO (JP)
TORISAWA KENTARO (JP)
MIZUNO JUNTA (JP)
OTAKE KIYONORI (JP)
ASAO YOSHIHIKO (JP)
TORISAWA KENTARO (JP)
MIZUNO JUNTA (JP)
OTAKE KIYONORI (JP)
Application Number:
PCT/JP2023/001155
Publication Date:
August 31, 2023
Filing Date:
January 17, 2023
Export Citation:
Assignee:
NAT INST INF & COMM TECH (JP)
International Classes:
G10L15/06
Domestic Patent References:
| WO2007097176A1 | 2007-08-30 | |||
| WO2020225999A1 | 2020-11-12 |
Other References:
SUGIYAMA MASAKAZU, YOSHIMURA RYOMA, YUTA TOMOMATSU, KOMACHI MAMORU: "Generation and evaluation of pseudo error data for error correction of Japanese speech recognition", THE 35TH ANNUAL CONFERENCE OF THE JAPANESE SOCIETY FOR ARTIFICIAL INTELLIGENCE, 1 January 2021 (2021-01-01), pages 1 - 4, XP093069918
ANONYMOUS: "AI Shift Advent Calendar 2021] Consideration of how to expand speech recognition error data", AI SHIFT, 26 January 2022 (2022-01-26), XP093087122, Retrieved from the Internet [retrieved on 20230929]
TANAKA YU, YUGO MURAWAKI, DAISUKE KAWAHARA AND SADAO KUROHASHI: "Building a Japanese Typo Dataset and Typo Correction System Based on Wikipedia’s Revision History", NATURAL LANGUAGE PROCESSING, LANGUAGE PROCESSING SOCIETY OF JAPAN, vol. 28, no. 4, 1 December 2021 (2021-12-01), pages 995 - 1033, XP093087124, ISSN: 2185-8314, DOI: 10.5715/jnlp.28.995
SAHASHI HIROYA, TOMOYOSHI AKIBA, HIJIRI NAKAGAWA: "Examination of rescoring of English-Japanese machine translation of scientific and technical paper abstracts and lecture audio", PROCEEDINGS OF THE 25TH ANNUAL CONFERENCE OF THE LANGUAGE PROCESSING SOCIETY, THE ASSOCIATION FOR NATURAL LANGUAGE PROCESSING., 4 March 2019 (2019-03-04), The Association for Natural Language Processing., pages 1165 - 1168, XP093087125
ANONYMOUS: "AI Shift Advent Calendar 2021] Consideration of how to expand speech recognition error data", AI SHIFT, 26 January 2022 (2022-01-26), XP093087122, Retrieved from the Internet
TANAKA YU, YUGO MURAWAKI, DAISUKE KAWAHARA AND SADAO KUROHASHI: "Building a Japanese Typo Dataset and Typo Correction System Based on Wikipedia’s Revision History", NATURAL LANGUAGE PROCESSING, LANGUAGE PROCESSING SOCIETY OF JAPAN, vol. 28, no. 4, 1 December 2021 (2021-12-01), pages 995 - 1033, XP093087124, ISSN: 2185-8314, DOI: 10.5715/jnlp.28.995
SAHASHI HIROYA, TOMOYOSHI AKIBA, HIJIRI NAKAGAWA: "Examination of rescoring of English-Japanese machine translation of scientific and technical paper abstracts and lecture audio", PROCEEDINGS OF THE 25TH ANNUAL CONFERENCE OF THE LANGUAGE PROCESSING SOCIETY, THE ASSOCIATION FOR NATURAL LANGUAGE PROCESSING., 4 March 2019 (2019-03-04), The Association for Natural Language Processing., pages 1165 - 1168, XP093087125
Attorney, Agent or Firm:
SHIMIZU, Satoshi (JP)
Download PDF: