Program

Date: September 4, 2025

Time: 2:45 PM – 6:00 PM 

Location: Room 101

14:45

Opening Remarks, AICS Scientific Commitee

14:50

Keynote: Franco Scarselli

Theory of graph neural networks vs complex data: some current results and perspectives

Complex information often includes patterns connected by relationships, which are naturally represented by graphs. Thus, machine learning models for graphs and, in particular, Graph Neural Networks (GNNs) are a powerful tool to implement applications for  complex data domains.  However, the complexity of modern applications is not restricted to the presence of relationships in domains, but it also depends on many other characteristics such as the heterogeneity of the patterns and the relationships, the need of different types of learning at the same time, the additional constraints related to expandability and so on.  In this context, theoretical studies are fundamental in order to better understand how GNNs can be adapted to face the mentioned peculiarities of complex applications. In this talk, I will focus  on theory of graph neural networks, recalling existing results and introducing some of those obtained in our lab. Moreover, I will discuss the perspectives and some of the open problems.

15:30

Characterizing Rider Behavior Patterns Through Contrastive Learning on Motorcycle Data

by Federico Pennino, Davide Sette, David Attisano, Maurizio Gabbrielli 

15:45

Atom Depth as an Informative Descriptor in Protein Stability Prediction: A Comparative Study with SASA across 3D-Structure and Graph-Based Models

by Sara Bacconi, Barbara Corradini, Filippo Costanti, Duccio Meconcelli, Giacomo Nunziati, Alessia Lucia Prete

16:00

Guided Molecular Generation through Logical Constraints

by Emma Meneghini, Paolo Frazzetto, Nicolò Navarin

16:15

Coffee break 

16:45

Predicting Non-commute Trip Attraction and Activity Hotspots with Explainable Machine Learning and Open Data

by Shaun Hoang, Elsa Arcaute, Howard Wong, David Arquati

17:00

Demystifying Graphons for Graph-Based Machine Learning

by Ben Cullen, Sara Bacconi, Veronica Lachi, Caterina Graziani

17:15

Bias and inconsistencies in LLMs: using complex systems approaches to peek into the black box

by Daniele Proverbio, Alessio Buscemi, Alessandro Di Stefano, The Anh Han, Pietro Liò

17:30

Y Social: an online Social Media Digital Twin

by Giulio Rossetti, Massimo Stella, Remy Cazabet, Katherine Abramski, Erica Cau1, Salvatore Citraro, Andrea Failla, Riccardo Improta, Virginia Morini, Valentina Pansanella

17:45

Nature-Inspired Local Propagation

by Alessandro Betti, Marco Gori

18:00

Closing