Samuele Fonio
University of Turin, Corso Svizzera 185, Turin, Italy.
University of Turin
Dept. of Computer Science
Alpha Research Group
Hi! I am Samuele Fonio, a PhD student in Modeling and data science at the University of Turin. I am currently enrolled in my third year. My research focus is about:
- Deep Learning on non-Euclidean manifolds;
- Communication-efficient Federated Learning.
To tackle these challenges, I am focusing on Prototype Learning and other metric learning methods 
. However, I am very curious about many aspects of Machine Learning (ML), and eager to explore many aspects of it.
My background is in mathematics, I have a bachelor in pure Mathematics and a Master Degree in Stochastics and Data Science.
At the University of Turin, I am more focused on Machine Learning research, but I am part of the Parallel computing research group Alpha! My supervisors are Roberto Esposito and Marco Aldinucci.
On my website you can find my publications, my interests, my CV, and many other things. Of course, it is still in development, and I still need to get used to al-folio. Have fun on my website!
News
| Sep 22, 2025 | Our paper “Composable Sparse Subnetworks via Maximum-Entropy Principle” has been accepted at the Mechanistic Interpretability Workshop at NeurIPS 2025! Big Kudos and thanks to all the authors: Francesco Caso, Nicola Saccomanno, Simone Monaco and Fabrizio Silvestri. |
|---|---|
| Aug 29, 2025 | We got our Workshop on Federated Learning for Critical Applications (FLCA) accepted at AAAI 2026! Big thanks to the other organizers: Bruno Casella, Linara Adilova, Michael Kamp and Mirko Polato. Check out this site if interested :) |
| Jul 22, 2025 | Our paper “Personalized Aggregation for Federated Prototypical Learning” co-authored with Bruno Casella and Marco Aldinucci has been accepted at the 3rd Workshop on Advancements in Federated Learning (WAFL) co-located with the European Conference on Machine Learning (ECML) 2025! |
| Jul 22, 2025 | Our paper “Non-Euclidean Geometries for Prototype-Based Image Classification: a Reality Check” co-authored with Silvia Grosso and Sara Bouchenak has been accepted at the 4th Italian Conference on Big Data and Data Science (ITADATA) 2025! |
| Jul 11, 2025 | Our paper “Fed2RC: Federated Rocket Kernels and Ridge Classifier for Time Series Classification” co-authored with Bruno Casella, Lorenzo Sciandra, Claudio Gallicchio, Marco Aldinucci, Mirko Polato and Roberto Esposito has been accepted at the European Conference on Artificial Intelligence (ECAI) 2025! |
Selected Publications
- Non-Euclidean geometries for Prototype-Based Image Classification: a Reality CheckIn Proceedings of the 4th Italian Conference on Big Data and Data Science, 2025
- Personalized Aggregation for Federated Prototypical LearningIn Proceedings of the ECML-PKDD Workshops, 2025
- Fed2RC: Federated Rocket Kernels and Ridge Classifier for Time Series ClassificationIn Proceedings of the 28th European Conference on Artificial Intelligence (ECAI) 2025, 2025
- Hyperbolic Prototypical Entailment Cones for Image ClassificationIn The 28th International Conference on Artificial Intelligence and Statistics, 2025
- Fedhp: Federated learning with hyperspherical prototypical regularizationIn ESANN 2024 proceedings, 2024
- Federated AdaBoost for Survival AnalysisIn Proceedings of the ECML-PKDD Workshops, 2024
- Architecture-Based FedAvg for Vertical Federated LearningIn Proceedings of the IEEE/ACM 16th International Conference on Utility and Cloud Computing, 2023
- Hierarchical priors for Hyperspherical Prototypical NetworksIn ESANN 2023-Proceedings, 2023
- Benchmarking Federated Learning Frameworks for Medical Imaging TasksIn International Conference on Image Analysis and Processing, 2023