Thesis Details
Automatic Speech Recognition System Continually Improving Based on Subtitled Speech Data
Today's large vocabulary speech recognition systems are very accurate. However, tens or hundreds of hours of manually transcribed speech are needed in order to train such system. This kind of data is often unavailable, or they even do not exist for the desired language. A possible solution is to use commonly available but lower quality audiovisual data. This thesis addresses the methods of processing such data for semi-supervised training of acoustic models. Afterwards, it demonstrates how to continually improve already trained acoustic models by using these practically unlimited data. In this work is proposed a novel approach for selecting data based on similarity with the target domain.
Large vocabulary continuous speech recognition, semi-supervised training, time delay neural network, subtitled speech data, acoustic modelling
Beran Vítězslav, doc. Ing., Ph.D. (DCGM FIT BUT), člen
Horák Aleš, doc. RNDr., Ph.D. (FI MUNI), člen
Hrubý Martin, Ing., Ph.D. (DITS FIT BUT), člen
Janoušek Vladimír, doc. Ing., Ph.D. (DITS FIT BUT), člen
Rozman Jaroslav, Ing., Ph.D. (DITS FIT BUT), člen
@mastersthesis{FITMT22041, author = "Martin Kocour", type = "Master's thesis", title = "Automatic Speech Recognition System Continually Improving Based on Subtitled Speech Data", school = "Brno University of Technology, Faculty of Information Technology", year = 2019, location = "Brno, CZ", language = "english", url = "https://www.fit.vut.cz/study/thesis/22041/" }