Short Bio

PhD Student
Computer Science Department, University of Turin
Parallel Computing group
Via Pessinetto 12, 10149 Torino – Italy
E-mail: samuele.fonio@unito.it
Samuele Fonio is a PhD student from Modeling and Data Science at UniTo funded by Leonardo Company.
He graduated in Mathematics in 2020 and in Stochastics and Data Science in 2022 with a thesis in Deep Learning comparing the behaviour of different geometries in Image classification task using prototype learning.
Fields of interest:
- Deep Learning
- Geometric Deep Learning
- Federated Learning
Publications
2023
- S. Fonio, L. Paletto, M. Cerrato, D. Ienco, and R. Esposito, “Hierarchical priors for Hyperspherical Prototypical Networks,” in 31th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN, Bruges, Belgium, 2023.
[BibTeX] [Abstract] [Download PDF]
In this paper, we explore the usage of hierarchical priors to improve learning in contexts where the number of available examples is extremely low. Specifically, we consider a Prototype Learning setting where deep neural networks are used to embed data in hyperspherical geometries.In this scenario, we propose an innovative way to learn the prototypes by combining class separation and hierarchical information. In addition, we introduce a contrastive loss function capable of balancing the exploitation of prototypes through a prototype pruning mechanism. We compare the proposed method with state-of-the-art approaches on two public datasets.
@inproceedings{23:esann:fonio, abstract = {In this paper, we explore the usage of hierarchical priors to improve learning in contexts where the number of available examples is extremely low. Specifically, we consider a Prototype Learning setting where deep neural networks are used to embed data in hyperspherical geometries.In this scenario, we propose an innovative way to learn the prototypes by combining class separation and hierarchical information. In addition, we introduce a contrastive loss function capable of balancing the exploitation of prototypes through a prototype pruning mechanism. We compare the proposed method with state-of-the-art approaches on two public datasets.}, author = {Fonio, Samuele and Paletto, Lorenzo and Cerrato, Mattia and Ienco, Dino and Esposito, Roberto}, booktitle = {31th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, {ESANN} }, address = {Bruges, Belgium}, keywords = {icsc}, title = {Hierarchical priors for Hyperspherical Prototypical Networks}, year = {2023}, note = {In print}, month = {oct}, url = {https://www.esann.org/sites/default/files/proceedings/2023/ES2023-65.pdf}, bdsk-url-1 = {https://www.esann.org/sites/default/files/proceedings/2023/ES2023-65.pdf}, bdsk-url-2 = {https://doi.org/10.14428/esann/2023.ES2023-65} }
- S. Fonio, “Benchmarking Federated Learning Frameworks for Medical Imaging Tasks,” in Image Analysis and Processing – ICIAP 2023 – 22th International Conference – FedMed, Udine, Italy, 2023.
[BibTeX] [Abstract]
This paper presents a comprehensive benchmarking study of various Federated Learning (FL) frameworks applied to the task of Medical Image Classification. The research specifically addresses the often neglected and complex aspects of scalability and usability in off-the-shelf FL frameworks. Through experimental validation using real case deployments, we provide empirical evidence of the performance and practical relevance of open source FL frameworks. Our findings contribute valuable insights for anyone interested in deploying a FL system, with a particular focus on the healthcare domain—an increasingly attractive field for FL applications.
@inproceedings{23:iciap:fedmed:ws:fonio, abstract = {This paper presents a comprehensive benchmarking study of various Federated Learning (FL) frameworks applied to the task of Medical Image Classification. The research specifically addresses the often neglected and complex aspects of scalability and usability in off-the-shelf FL frameworks. Through experimental validation using real case deployments, we provide empirical evidence of the performance and practical relevance of open source FL frameworks. Our findings contribute valuable insights for anyone interested in deploying a FL system, with a particular focus on the healthcare domain—an increasingly attractive field for FL applications.}, author = {Fonio, Samuele}, booktitle = {Image Analysis and Processing - {ICIAP} 2023 - 22th International Conference - FedMed}, address = {Udine, Italy}, keywords = {icsc, eupilot}, title = {Benchmarking Federated Learning Frameworks for Medical Imaging Tasks}, year = {2023}, note = {In print}, publisher = {Springer LNCS}, month = {sep} }
- G. Mittone and S. Fonio, “Benchmarking Federated Learning Scalability,” in Proceedings of the 2nd Italian Conference on Big Data and Data Science, ITADATA 2023, September 11-13, 2023, Naples, Italy, 2023.
[BibTeX] [Abstract] [Download PDF]
Federated Learning (FL) is a widespread Machine Learning paradigm handling distributed Big Data. In this work, we demonstrate that different FL frameworks expose different scaling performances despite adopting the same technologies, highlighting the need for a more comprehensive study on the topic.
@inproceedings{23:itadata:extabstract:mittone:fonio, abstract = {Federated Learning (FL) is a widespread Machine Learning paradigm handling distributed Big Data. In this work, we demonstrate that different FL frameworks expose different scaling performances despite adopting the same technologies, highlighting the need for a more comprehensive study on the topic.}, author = {Mittone, Gianluca and Fonio, Samuele}, booktitle = {Proceedings of the 2nd Italian Conference on Big Data and Data Science, {ITADATA} 2023, September 11-13, 2023}, address = {Naples, Italy}, keywords = {icsc, eupilot}, title = {Benchmarking Federated Learning Scalability}, year = {2023}, publisher = {{CEUR}}, note = {In press}, month = {sep}, url = {https://hdl.handle.net/2318/1933852} }