Paolo Frazzetto

I am a Postdoctoral Researcher at the Image & Signal Processing (ISP) Lab of the University of Valencia, Spain, working with Prof. Gustau Camps-Valls. My current research is part of the Thinking Earth Horizon Europe project, where I focus on developing and explaining AI models to better understand our planet's complex interconnected systems. Specifically, I am collaborating with the World Food Programme to model and analyze the impact of climate change on food security.

Previously, I was at the University of Padua as postdoctoral researcher for the SymboliG project within the Future of Artificial Intelligence (FAIR) group, focusing on how to apply symbolic and domain knowledge for Generative AI. During the PhD, I worked on AI Applications with Human Resources Data at Amajor SB S.p.a, under the supervision of Prof. Alessandro Sperduti within the Brain, Mind & Computer Science Doctoral School at the University of Padova. I additionaly spent six month in Bielefeld, Germany, visitng the HammerLab.

I have a broad interest in Statistical and Machine Learning, with a focus on Graph Neural Networks and eXplainable AI, and their applications to the Natural Sciences.

Best way to reach out to me is via e-mail.

Publications

• Categorical Explaining Functors: Ensuring Coherence in Logical Explanations

  Stefano Fioravanti, Francesco Giannini, Pietro Barbiero, Paolo Frazzetto, Roberto Confalonieri, Fabio Zanasi, Nicolò Navarin

  The 22nd International Conference on Principles of Knowledge Representation and Reasoning (KR 2025)

  DOI: 10.24963/kr.2025/30

  KR 25
• Enhancing Predictive Process Monitoring on Small-Scale Event Logs Using LLMs

  Alessandro Padella, Paolo Frazzetto, Nicolò Navarin, Massimiliano de Leoni

  Business Process Management Forum (BPM 2025)

  DOI: 10.1007/978-3-032-02929-4_16

  BPM 25
• Graph Neural Networks for Candidate-Job Matching: An Inductive Learning Approach

  Paolo Frazzetto, Muhammad Uzair Ul Haq, Flavia Fabris, Giovanni Da San Martino & Alessandro Sperduti

  Data Sci. Eng. (2025)

  DOI: 10.1007/s41019-025-00293-y

  DSE 25
• Exact Computation of Any-Order Shapley Interactions for Graph Neural Networks

  Maximilian Muschalik, Fabian Fumagalli, Paolo Frazzetto, Eyke Hüllermeier, Barbara Hammer, Alessandro Sperduti

  ICLR 2025, The Thirteenth International Conference on Learning Representations

  View paper

  ICLR 25
• Leveraging Deep Learning in Human Resources: A Systematic Investigation of Graph Neural Networks for Candidate-Job Matching

  Ph.D. Thesis, submitted in December 2024, defended on March 26, 2025.

  Download

  PhD 25
• From Text to Talent: A Pipeline for Extracting Insights from Candidate Profiles

  Paolo Frazzetto, Muhammad Uzair Ul Haq, Flavia Fabris, Alessandro Sperduti

  Presented at The 3rd Italian Conference on Big Data and Data Science (ITADATA), Pisa, Italy, 17-19 September 2024

  View paper

  ITADATA 24
• Physics-Informed Graph Neural Cellular Automata: an Application to Compartmental Modelling

  Nicolò Navarin, Paolo Frazzetto, Luca Pasa, Pietro Verzelli, Filippo Visentin, Alessandro Sperduti, Cesare Alippi

  Accepted and presented at The International Joint Conference on Neural Networks (IJCNN), Yokohama, Japan, June 30 - July 5 2024

  View paper

  IJCNN 24
• Beyond the Additive Nodes' Convolutions: a Study on High-Order Multiplicative Integration

  Paolo Frazzetto, Luca Pasa, Nicolò Navarin, Alessandro Sperduti

  Proceedings of the 39th ACM/SIGAPP Symposium on Applied Computing, 2024

  DOI: 10.1145/3605098.3636016

  SAC 24
• Improving Soft Skill Extraction via Data Augmentation and Embedding Manipulation

  Muhammad Uzair Ul Haq, Paolo Frazzetto, Alessandro Sperduti, Giovanni Da San Martino

  Proceedings of the 39th ACM/SIGAPP Symposium on Applied Computing, 2024

  DOI: 10.1145/3605098.3636010

  SAC 24
• Topology preserving maps as aggregations for Graph Convolutional Neural Networks

  Paolo Frazzetto, Luca Pasa, Nicolò Navarin, Alessandro Sperduti

  Proceedings of the 38th ACM/SIGAPP Symposium on Applied Computing, 2023

  DOI: 10.1145/3555776.3577751

  SAC 23

Academia