Pie & AI - Shenzhen: Modeling of Paralinguistic Speech Attributes for Intelligent Speech Interaction

Pie & AI is a series of DeepLearning.AI meetups independently hosted by community groups. This event is hosted by Tsinghua SIGS.

Lecturer: Wu Zhiyong
(Associate researcher and doctoral supervisor of SIGS, Tsinghua University)

Research fields: intelligent voice interaction, voice processing, expressive visual speech synthesis, natural language understanding and generation, audio and video bimodal joint modeling.

With the development of artificial intelligence technology and the wide application of speech interaction products, intelligent speech interaction has been widely used in our daily life. The expectations for harmonious speech interaction are also increasing. The speech signal not only contains language semantic information, but also conveys a variety of rich paralinguistic information such as speaker, style, emotion, emphasis, and intonation. This paralinguistic information carries the underlying intention meaning of the speech, which is an important aspect of speech expressiveness, and is an indispensable factor for harmonious speech interaction. How to effectively analyze and model the paralinguistic information in speech is the key to build intelligent speech interaction technology and improve user satisfaction. It is also an important research topic in the field of speech and natural language processing.

This event will introduce the latest work of Tsinghua University’s Human-Computer Speech Interaction Laboratory in intelligent speech interaction, including speech emotion recognition, speech emphasis detection, user intention understanding, controllable emotional speech synthesis, and controllable emphatic speech synthesis etc.

Date & time: Fri, July 30, 2021

#DeepLearningAI #TsinghuaSIGS #grassrootAi #TsinghuaUniversity #AI