Automatic Improvement of Acoustic and Language Models of Automatic Speech Recognition through Spoken Dialogue

2016 Fiscal Year Final Research Report

• Project/Area Number: 15K16051
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Intelligent informatics
• Research Institution: Osaka University
• Principal Investigator: Takeda Ryu 大阪大学, 産業科学研究所, 助教 (20749527)
• Project Period (FY): 2015-04-01 – 2017-03-31
• Keywords: 音声対話 / 音声認識 / 音響モデル / 言語モデル

Outline of Final Research Achievements

Our purpose is the development of a spoken dialogue system of which models used in speech recognition is maintenance-free. The main issues are the development of spoken dialogue systems on robots, and the development of essential technologies on acoustic model and language model. The main outcomes are 1) the development of the acoustic model based on DNN and its adaptation method, 2) the development of the un-supervised segmentation of phoneme sequences for spontaneous utterances, 3) the development of the dialogue strategy for unknown word acquisition using implicit confirmation requests, and 4) the development of sound source localization method based on DNN for human-robot interaction.

Free Research Field

音声対話