News

It's time to evaluate

Feb. 21, 2017, 10:20 a.m.

It's time for the second round of the evaluation of the first prototype. User partners are now involved in the evaluation sessions, presenting the system to potential end-users and health professionals. The evaluation is a key aspect of our project because it aims to gather expectations and feedback. The consortium is constantly working to improve the natural conversation.

Leo Wanner, main responsible and coordinator of the KRISTINA project, explains that users can conduct a basic dialogue with the system and this interaction is now supported by an overall structure able to exchange of information between each system component. The first prototype was not expected to support a flexible dialogue but partners are now on the right track and they will continue to work, improving the functionalities and the quality of each individual component. First of all it is necessary to improve the accuracy of the speech recognition, because the agent takes too long to analyse the sentences and it often misunderstand the user. Secondly KRISTINA should better process what the end-user is saying, understanding and providing the right information with an appropriate reaction. Finally, “flexibility” is the key to the future system development: the user should communicate with the virtual agent in a more flexible way in their native language.

Open results need open discussions! 


 


multilingual intelligent embodied agent social competence adaptive dialogue expressive speech recognition and synthesis discourse generation vocal facial and gestural social and emotional cues