معرفی
Gabriele Perugini is a Lecturer in the Computing Science Department at Bocconi University since 2024. He holds a BSc (2013) and PhD (2017) in Physics from Rome University 'La Sapienza', followed by a PostDoc at Bocconi (2018–2023) focusing on machine learning and artificial neural networks.
His research integrates statistical mechanics with learning and optimization problems, aiming to bridge theoretical models with the behavior of deep neural networks. Key areas include the loss landscape analysis of neural networks and phase transitions in learning systems. He teaches courses on mathematical modeling in machine learning, Python coding, and foundations of physics.
Publications appear in top journals like Physical Review Letters and Machine Learning: Science and Technology, focusing on topics such as random-features models and flat minima in neural networks. No awards explicitly listed, but active in interdisciplinary computational research.
