Mental health prediction from social media data

Abstract

This talk explores the computational task of predicting mental health conditions from social media data, conducted within the scope of the SetembroBR project (2022-2024). We discuss the creation of a dataset for detecting depression and anxiety disorders from Portuguese tweets, along with the development and initial results of computational models for these tasks. Finally, we address key challenges that remain for ongoing research, with particular focus on their relevance to the current AIM-Health project.

Video

Short Bio

Ivandré Paraboni holds a Ph.D. in Computer Science from the University of Brighton, United Kingdom (2003), and completed a postdoctoral fellowship at the University of Aberdeen, Scotland (2012). He is an Associate Professor (tenured) and researcher at the University of São Paulo (USP), with a full-time appointment at the School of Arts, Sciences, and Humanities (EACH). Research interests span a broad spectrum of human language processing, ranging from computational methods grounded in Cognitive Science to practical applications, such as the classification of web-based documents. He currently supervises research on stance recognition from text, the detection of mental health disorders from multimodal data, the identification of creative thinking, and arbitrary style transfer in natural language generation.