Davide Carbone

Ph.D. in Matematica Pura E Applicata , 36^th cycle (2020-2023)

Ph.D. obtained in 2024

Dissertation:

Generative Models as Out-of-equilibrium particle systems: the case of Energy-Based Models (Abstract)

Tutors:

Lamberto Rondoni

Profile

Research topic

Neural Networks as Interacting Particle Systems

Research interests

Mathematical Physics

Biography

From the very beginning of my student career I regard Mathematics and Physics as disciplines closely related to the concept of analogy. During my entire course I was fascinated in finding connections between very different topics. This belief led me to choose the bachelor's Degree in Physics at University of Turin. I graduated in 2018 with a thesis on Palatini-f(R) gravitational models (advisor: prof. L. Fatibene). Subsequently, I chose the master’s degree in Physics of Complex Systems at University of Turin, that guaranteed the opportunity to maximally enlarge my theoretical knowledge. I graduated in 2020 with a thesis on generalized T-Symmetry and Onsager relations (advisor: prof. L. Rondoni). I currently work on theoretical Machine Learning seen from the point of view of Statistical Mechanics as a system of interacting particles. As a side note, a graduated in Piano at Conservatory of Turin in 2019.

Teaching

Teachings

Bachelor of Science

Mathematical analysis II. A.A. 2022/23, INGEGNERIA INFORMATICA (COMPUTER ENGINEERING). Collaboratore del corso
Analisi matematica II. A.A. 2021/22, INGEGNERIA MECCANICA. Collaboratore del corso
Analisi matematica II. A.A. 2021/22, INGEGNERIA CIVILE. Collaboratore del corso
Analisi matematica II. A.A. 2021/22, INGEGNERIA CIVILE. Collaboratore del corso

MostraNascondi A.A. passati

Research

Research groups

Models and Methods in Mathematical Physics

Publications

Works published during the Ph.D. View all publications in Porto@Iris

Carbone, Davide (2024)
Generative Models as Out-of-equilibrium particle systems: the case of Energy-Based Models. relatore: RONDONI, Lamberto; , 36. XXXVI Ciclo, P.: 173
Doctoral Thesis
Carbone, Davide; Hua, Mengjian; Coste, Simon; Vanden-Eijnden, Eric (2023)
Efficient Training of Energy-Based Models Using Jarzynski Equality. In: 37th Conference on Neural Information Processing Systems (NeurIPS), New Orleans (USA), DEC 10-16, 2023, pp. 1-32
Contributo in Atti di Convegno (Proceeding)
Carbone, Davide; DE GREGORIO, PAOLO MARIO; Rondoni, Lamberto (2022)
Time reversal symmetry for classical, non-relativistic quantum and spin systems in presence of magnetic fields. In: ANNALS OF PHYSICS, vol. 441. ISSN 0003-4916
Contributo su Rivista
Carbone, D.; Rondoni, L. (2020)
Necessary and sufficient conditions for time reversal symmetry in presence of magnetic fields. In: SYMMETRY, vol. 12, pp. 1-14. ISSN 2073-8994
Contributo su Rivista