Speech Technologies for Language Learning - Supporting Speaking, Listening, and Conversation Skills
Explore how speech technologies can enhance language learning in this hour-long lecture by Professor Nobuaki Minematsu from the University of Tokyo. Discover innovative applications of speech synthesis, recognition, and voice conversion to improve speaking, listening, and conversational skills for students and immigrant workers. Learn about text-based prosody prediction for explicit prosody training, High Variability Phonetic Training (HVPT) enhanced by speech analysis and voice conversion, and the use of adversarial training for robust listening skills. Examine the potential of speech recognition in shadowing assessment for improved parallel processing in conversations. Gain insights into a novel speech assessment framework focusing on comprehensibility rather than native-likeness, and explore recent findings in objectively measuring learners' speech comprehensibility.