大会报告

Chin-Hui Lee:Recent Advances in Mispronunciation Detection and Diagnostic Feedback for Non-native L2 Learners of Mandarin


Speech generated by L2 speakers tends to have their syllable productions fall between two canonical categories, and their tone pronunciations prolonged in duration. These observations make conventional syllable and tone modeling for native speakers not directly applicable to non-native speech. Recent progresses in speech research have inspired us to:
(i) utilize posteriors extracted from speech attribute and memory-based features;
(ii) use soft targets to replace hard labels in handling potentially erroneous transcriptions in training; and
(iii) adopt memory-based feature and models to capture long-term context in model learning.
Together they have allowed us to established better phone and tone models to improve computer assisted pronunciation training (CAPT) system performances.
Furthermore, we will also discuss some recent work on building decision trees for corrective feedback. By asking questions directly related to articulatory gestures at each node of the tree that facilitates a natural diagnostic mechanism to provide speech production hints on how L2 pronunciations could be improved. With this knowledge-driven approach to directly relate key insights in big data to memory-based features and models, we hope to obtain sustainable research results that are not easily achieved via simple black-box learning tools.


Speaker: Chin-Hui Lee, School of ECE, Georgia Tech




Chin-Hui Lee is a professor at School of Electrical and Computer Engineering, Georgia Institute of Technology. Before joining academia in 2001, he had accumulated 20 years of industrial experience ending in Bell Laboratories, Murray Hill, as a Distinguished Member of Technical Staff and Director of the Dialogue Systems Research Department. Dr. Lee is a Fellow of the IEEE and a Fellow of ISCA. He has published over 500 papers and 30 patents, with more than 45,000 citations and an h-index of 80 on Google Scholar. He received numerous awards, including the Bell Labs President's Gold Award in 1998. He won the SPS's 2006 Technical Achievement Award for “Exceptional Contributions to the Field of Automatic Speech Recognition”. In 2012 he gave an ICASSP plenary talk on the future of automatic speech recognition. In the same year he was awarded the ISCA Medal in scientific achievement for “pioneering and seminal contributions to the principles and practice of automatic speech and speaker recognition”.