Projects: Speech, Dialogue & Generation

The research on dialogue systems and natural language generation is currently concentrated around the following projects:

NaMuDDiS grant (Natural Multi-Domain Dialogue Systems)

This project focuses both on dialogue systems and language generation – it aims at developing more accurate language generation systems and dialogue systems trainable from limited data, domain portability, evaluation.

Student projects related to the grant also involved development of speech synthesis systems.

Teaching is an integrated part of the project – two new courses on dialogue systems (basic, advanced) were introduced and are now taught.

THEaiTRE grant (theatre play generation)

This is a collaboration with people from the theatre industry, with the aim of automatically generating a theatre play and performing it on stage.

METOD (industry collaboration with Agnostix)

This project focuses on modeling and visualizing dialogue systems behavior. The goal is to enable more sophisticated voice assistants for multiple domains such as banking, e-shops or logistics.

All Speech, Dialog & Generation Projects

All current and past projects related to Speech, Dialogue and Generation are listed below: