Presentation information

Oral presentation

General Session » [General Session] 10. Vision / Speech

[2N1] [General Session] 10. Vision / Speech

Wed. Jun 6, 2018 9:00 AM - 10:20 AM Room N (2F Sakurajima)

座長:辻川 剛範(NEC)

9:40 AM - 10:00 AM

[2N1-03] Incremental estimation of user's turn-taking state

〇Shinya Fujie1,2, Katsuya Yokoyama2, Tetsunori Kobayashi2 (1. Chiba Institute of Technology, 2. Waseda University)

Keywords:Spoken Dialog System, Turn Taking, Speech Recognition

Turn-taking state estimation to determine utterance timing of a spoken dialog system is discussed. We propose the recurrent neural network based method to estimate user's turn-taking state incrementally. The proposed method utilizes acoustic feature extracted using a spectrogram autoencoder as well as linguistic feature extracted from a partial speech recognition result using a neural network based language model. The article shows an example of estimation result and discuss the performance of the proposed method.