Document understanding in the era of foundation models

Venerdì 13 Marzo 2026 09:00

Friday 13th of March 2026, from 09:00 to 13:00 and 14:00 to 18:00, in Building MO52, Dipartimento di Ingegneria “Enzo Ferrari”, Via Vivarelli 10, 411125 (Modena)
Prof. Dimosthenis Karatzas will give a lecture with the title "Document understanding in the era of foundation models". The lecture is particularly targeted to students of the Master on Artificial Intelligence Engineering and Computer Science, but is open to all interested people.

Abstract:

This lecture argues that understanding written communication is an emergent capability of large multimodal models. It reviews the evolution of document understanding systems and highlights key open challenges. The talk also examines critical issues such as privacy guarantees, adversarial robustness, and explainability in document processing. It concludes by outlining opportunities, limitations, and future research directions for building robust document understanding systems

Short bio:

Prof. Dimosthenis Karatzas is Full Professor at the Autonomous University of Barcelona and Associate Director of the Computer Vision Centre, where he leads the Vision, Language and Reading group. He also co-directs the ELLIS Unit Barcelona. His research focuses on computer vision and machine learning, particularly robust reading systems and vision–language modelling. He has received the IAPR/ICDAR Young Investigator Award, a Google Research Award, and two Amazon Research Awards, and has ranked among the top 2% most-cited AI scientists (Scopus) since 2019. He has led numerous research projects, co-founded the spin-offs TruColour (UK) and AllRead (Spain), and driven large-scale technology transfer in sectors such as banking and utilities. He also founded the Library Living Lab (L3), a quadruple-helix social innovation initiative. He has held leadership roles within IAPR, serves on the editorial boards of IJCV and ICDAR, and founded the Robust Reading Competition, a widely used international benchmark platform serving more than 60,000 researchers.

This is an event sponsored by Regione Emilia Romagna with the project "Seminari su temi avanzati di Intelligenza Artificiale" Rif.PA 2024-23126/RER, CUP: E93C24001690002.

Autore: - Ultima modifica: Martedì 28 Luglio 2026 09:15