ISTI-Talk: A Path Through Image, Audio, and Language: Where Do We Go Next?

Day - Time: 22 October 2025, h.11:00

Place: Area della Ricerca CNR di Pisa - Room: C-29

Speakers

• Alì Reza Omrani

Referent

Giulio Del Corso

Abstract

Applying AI to understand complex human behaviors from diverse data like audio, movement, and language remains a significant challenge. This presentation showcases a research journey across multiple projects to demonstrate the power of multimodal models. The work combines custom deep learning for behavioral analysis with large-scale language models for industrial applications. Key outcomes include the successful automated measurement of stereotyped behaviors in a clinical context and the development of an efficient AI agent for HR. Ultimately, this talk highlights the possibility of how Large Language Models can serve as a unifying framework, paving the way for AI that holistically understands human behavior.