ObTIC- Workshop on Mental Health and AI

In the era of digital transformation, mental health stands at the intersection of new challenges and unprecedented opportunities. The integration of artificial intelligence, natural language processing, and speech technologies opens promising prospects for understanding, diagnosing, and managing mental disorders. This study day aims to bring together researchers, practitioners, and engineers to explore the interactions between innovative technologies and mental health, and to discuss recent developments in this rapidly evolving field.

Program

09:00 - 09:30 - Coffee break

09:30 – 10:30 - Mohamed Chetouani (Univ. Sorbonne)
Speech Social Processing for Mental Health

10:30 – 11:00
Frédéri Isel (Univ. Nanterre)
The Contribution of Large Language Models to Experimental Research on Natural Language Processing in Psycholinguistics

11:00 – 11:30
Jalal Altamimi (Univ. Paris-Cité)
Speech as a bio-marker guiding evaluation of cognitive impairment

11:30 – 12:00
Gudrun Ledegen (Univ. Rennes)
Characterization of Emotions in a Suicide Prevention Chat Corpus (2005-2025)

12:00 13:30 - Break

13:30 – 14:00
Valeria Lucarini (INSERM)
Voices and Dialogues Beyond Diagnosis: A Prosodic and Conversational Analysis Within the Psychosis Spectrum

14:00 –14:30
Rufaida Kashif (Univ. Sorbonne)
Detection of emotions in a conversational data for mental health patients

14:30 – 15:00
Eric Jordan (Univ. Sorbonne)
Audio Data Analysis in Clinical Interview Settings: Psychosis Detection and Automatic Transcription

15:00 – 15:30
Helena Gomez, Gemma Bel, Roberto Huerta (Univ. UNAM)
Exploring the effectiveness of combined lightweight machine learning models in the detection of mental health disorders

15:30 –16:00 - Closing

Registration: https://forms.gle/ZCo1GXKH3dFetryA8 
On-site: Free entry, registration required (limited seating).
Online: A link will be sent the day before to registered participants.
Presentations will be held in French or English (slides will be in English).

This conference is organized by the APAISE team, with the support of the Fondation de France, ObTIC, and SCAI.