Project description

This project was supported by Shenzhen government from 2017 to 2020.

The project designed and implemented a prototype of a human-computer dialogue system and applied it in medical services.

Solution

The project started from fundamental research in language representation, and successfully applied the theoretical innovations in the development of a natural language processing system with new and improved results in dialogue systems. 

The system successfully integrated a complete set of functions including speech recognition, language understanding, language generation and speech synthesis, and executed in Java platform with high efficiency.

The project designed and implemented a prototype of a human-computer dialogue system and applied it in medical services.

Key contribution

Theoretical: rigorous analysis of sparse models in general data representation.

Model: a unified representation scheme in natural language processing.

Practical: a lightweight human-computer dialogue prototype on Java platform.

Next steps

Further theoretical development of the sparse model for language processing.

Release of key models and algorithms to the public.

Incorporation of the techniques developed in the project into a real system.

Team/contributors

LI Wenye, YU Pengfei.