deLearyous

deLearyous

deLearyous
Project information
Abstract: 

The goal of the deLearyous project is the development of a 3D serious video game where players can practice their communication skills by interacting with a virtual character. The trainee talks to the virtual character by typing in sentences in Dutch. The software then has to interpret these sentences and determine the appropriate reaction for the virtual character.

The virtual character's response follows a communication model known as the Interpersonal Circumplex (AKA "Leary's Rose", after Timothy Leary who developed the framework in 1957). Every utterance in conversations or negotiations can be position on this Rose according to the speakers state of mind (is the speaker friendly or hostile, dominant or submissive...). The state of mind of the speaker also has a strong influence on the listener's state of mind. With a thorough understanding of the Interpersonal Circumplex, it is thus possible for the speaker to guide the listener to a disposition that is more favorable to his goals. deLearyous will offer trainees an interactive and fun way to acquaint themselves with Leary's Rose and to develop their interpersonal communication skills.

Natural language processing plays an important part in the deLearyous project, as the position of the speaker on Leary's Rose has to be determined automatically on the basis of the Dutch text input. It is also necessary to extract and track topic information so the virtual character can be guided down the correct conversation path. Automatic emotion classification especially is an new and interesting research topic, and most of the research in the deLearyous project focuses on this topic.

Period: 
01/01/2010 - 31/12/2012
Sponsor(s): 

TETRA/IWT

CLiPS team participated in competition on emotion detection

During the summer, members of the CLiPS Computational Linguistics group participated in the 2011 Medical NLP Challenge, where we submitted results in (fine-grained) emotion detection from suicide notes.

 

Rankings will be revealed during the workshop in October.

Vaassen, F., & Daelemans W. (2011).  Automatic Emotion Classification for Interpersonal Communication. 2nd Workshop on Computational Approaches to Subjectivity and Sentiment Analysis (WASSA 2.011). PDF
Vaassen, F., & Daelemans W. (2011).  Automatic Emotion Classification for Interpersonal Communication. Presented at the 2nd Workshop on Computational Approaches to Subjectivity and Sentiment Analysis (WASSA 2.011).
Vaassen, F., & Daelemans W. (2010).  Emotion Classification in a Serious Game for Training Communication Skills. Presented at the 20th Meeting of Computational Linguistics in the Netherlands (CLIN20), Utrecht, The Netherlands.
Syndicate content