Automatic adaptation framework of neural network language model
Project/Area Number |
18K11354
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Multi-year Fund |
Section | 一般 |
Review Section |
Basic Section 61010:Perceptual information processing-related
|
Research Institution | Kyoto University |
Principal Investigator |
AKITA Yuya 京都大学, 経済学研究科, 教授 (90402742)
|
Project Period (FY) |
2018-04-01 – 2023-03-31
|
Project Status |
Completed (Fiscal Year 2022)
|
Budget Amount *help |
¥4,420,000 (Direct Cost: ¥3,400,000、Indirect Cost: ¥1,020,000)
Fiscal Year 2020: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2019: ¥1,820,000 (Direct Cost: ¥1,400,000、Indirect Cost: ¥420,000)
Fiscal Year 2018: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
|
Keywords | 音声認識 / ニューラルネットワーク / 言語モデル |
Outline of Final Research Achievements |
In automatic speech recognition, common models trained with general texts have limited performance for specialized topics, such as those in classroom lectures and academic talks. To deal with this problem, language model adaptation is often conducted. In this study, we investigated automatic adaptation framework of neural-network-based language models by using texts relevant to the topics in the target speech, and incorporate it into our system of automatic captioning, which produces captions for both of recorded audio and real-time audio, with the adapted language models.
|
Academic Significance and Societal Importance of the Research Achievements |
音声認識はコミュニケーションの支援技術として社会的な重要性が増大しているが,専門的な内容を含む音声に対してニューラルネットワークのような高度なモデルを適用することには技術的な困難がある.本研究により,非専門家がより性能の高い音声認識を容易に取り扱えるようになることには,大きな意義があると考えられる.
|
Report
(6 results)
Research Products
(2 results)