My name is Marion Ristorcelli. I’m PhD student in computer science at Aix-Marseille University, in the LIS lab. I started my PhD in November 2022 in the TALEP team (Traitement Automatique du Langage Écrit et Parlé).

My research is based on the use of ACAs (animated conversational agents), or more commonly, virtual characters, to improve social skills such as public speaking.

My research topic concerns the effect of virtual audience characteristics (composed of virtual characters) on user public speaking performance in virtual reality.

As part of my thesis, I am interested in the effect of virtual audience gender and social attitude on the perceived difficulty of the task and on the measured performance of the participants. The first step will be to collect the human-virtual audience corpus. More specifically, we’ll be asking participants to speak in front of a virtual audience that varies in gender and social attitude.

At the same time, we’ll record the participants’ non-verbal, verbal and physiological behavior. They will also have to complete different questionnaires regarding their user experience (emotions, ease of public speaking, anxiety…). We’ll also ask experts in the domain to rate the participants’ performance. At the end, we will use subjective data (questionnaires and expert ratings) as well as objective data (physiological and behavioral measurements) to feed a machine learning model. This will then enable us to identify the behavioral and physiological cues that correlate with performance and difficulty perceived by users.