Séminaires VENISE (no1)

(VENISE Series Seminars)

LIMSI-CNRS, Université Paris-Sud
Bât. 508, B.P. 133, 91403 Orsay cedex.

Dans le contexte de l'action transversale VENISE
(Virtualité et ENvironnement Immersif pour les Sciences Expérimentales) du LIMSI-CNRS,
nous vous informons que dans le cadre du premier séminaire VENISE:

Catherine PELACHAUD


University of Rome "La Sapienza"
http://www.dis.uniroma1.it/~pelachau/
cath@dis.uniroma1.it

nous présentera ses travaux sur :

"Facial Expressions in Embodied Conversational Agents"

le 27 Décembre 2001 de 14h00 à 16h00,
en Salle de Conférence (rez-de-chaussée) du LIMSI-CNRS, Orsay (Accès).

P. Bourdot & J. Mariani

ABSTRACT
Our goal is to develop a Believable Embodied Agent able to dialog with a user, but we aim at making an Agent that can also combine facial expressions in a complex and subtle way, just like a Human Agent does. We are applying the metaphor of face-to-face communication to human-computer interaction. Face-to-face conversation is very complex phenomenon as it involved a huge number of factors: we speak with our voice, but also with our hand, eye, face and body. Our gesture modifies, emphasizes, contradicts what we say by words. It is therefore important to consider both verbal and nonverbal behaviors while building an embodied agent. The agent must have the capacity to decide which facial expressions to show, which words to say with which intonation. That is, the agent should be able to plan not only what to communicate, but also by what (verbal or nonverbal) signals, in what combination and how synchronized. In this talk, we will present our work on the creation of a multimodal believable agent able to dialog with a user and whose nonverbal behaviors reflect its affective state. We will pay particular attention on the representation of the ``agent's mind'' as well as on the translation of the agent's cognitive state into facial expressions.

We first review a taxonomy of communicative functions that our Agent is able to express non-verbally; but we point out that, due to the complexity of communication, in some cases different information can be provided at once by different parts and actions of an Agent's face. We will present our method to assess and manage what happens, at the meaning and signal levels of multimodal communicative behavior, when different communicative functions have to be displayed at the same time and necessarily have to make use of the same expressive resources.