Objective
General aim: This project will generalise the TRINDI and SIRIDUS "Information State Update"(ISU) approach to dialogue management to develop adaptive multimodal dialogue systems. It will also extend the use of abstract representation, in particular the use of domain ontologies, to increase robustness and reconfigurability of dialogue systems. While the project will build showcases for in-car and in-home information and control, the main aim is to advance our understanding of generic technologies that will extend the ISU approach to adaptive multimodal and multi-lingual interaction. The project will have scientific and technological impact as a foundation for adaptive, recon-figurable, multimodal dialogue systems. The delivery of information in an intuitive, conversational, user-adaptivemanner will result in more efficient and robust interaction with information technology applications, for many different individuals in society.
Multi-disciplinary research areas: There are four main research themes, each corresponding to a core work package:
- Unifying multimodality and multilinguality,
-Automatic generation and reconfiguration of multi-modal interfaces,
-Multi-modal presentation in the Information State Update approach,
-Adaptivity and learning.
Laboratory showcases: To enhance collaboration, technology development and transfer, and scientific impact, TALK will develop laboratory showcases in a variety of scenarios, including in-car dialogue systems and the automated home environment.
Building expertise in the European Research Area (ERA): TALK will build on current European expertise in multi-modal dialogue systems and will consolidate advances made in former European research projects including TRINDI and SIRIDUS.
The Consortium: TALK brings together internationally recognised, highly effective, and experienced research groups in the relevant disciplines, as well as several leading industrial partners.
Fields of science
CORDIS classifies projects with EuroSciVoc, a multilingual taxonomy of fields of science, through a semi-automatic process based on NLP techniques.
CORDIS classifies projects with EuroSciVoc, a multilingual taxonomy of fields of science, through a semi-automatic process based on NLP techniques.
Call for proposal
Data not availableFunding Scheme
STREP - Specific Targeted Research ProjectCoordinator
66041 Saarbruecken
Germany