RICCARDO ZECCHINA
Courses a.y. 2022/2023
30554 MATHEMATICAL MODELLING IN MACHINE LEARNING
30558 STATISTICAL AND QUANTUM PHYSICS
30586 MACHINE LEARNING AND ARTIFICIAL INTELLIGENCE LAB
Courses previous a.y.
Biographical note
I am is professor in theoretical physics with a chair in Machine Learning. I did my PhD in Theoretical Physics at the University of Turin, where I had the luck of working with Tullio Regge. In 1997 I was appointed research scientist and head of the Statistical Physics Group at the International Centre for Theoretical Physics in Trieste. I 2007 I became full professor in Theoretical Physics at the Polytechnic University of Turin. In 2017 I moved to Bocconi University in Milan.
I have been multiple times long term visiting scientist at Microsoft Research (in Redmond and Cambridge MA) and at the Laboratory of Theoretical Physics and Statistical Models (LPTMS) of the University of Paris-Sud.
I am an advanced grantee of the European Research Council (2011-2015). In 2016, I was awarded (with M. Mezard and G. Parisi) the Lars Onsager Prize in Theoretical Statistical Physics by the American Physical Society.
About
We have just created a new department in Computing Sciences, an interdisciplinary center for research in fundamental and modeling problems in information and computation.
Our working paradigms are openness and collegiality.
Research interests
My current research interests lie at the interface between statistical physics, computer science and machine learning. My primary focus is on the study and the design of learning algorithms and processes, in modern AI and in biologically constrained models.
Selected research topics:
- Learning theory and learning algorithms
- Out-of-equilibrium dynamics in disordered systems
- Combinatorial optimization and discrete mathematics
- Probabilistic message-passing algorithms
- Computational neuroscience and computational biology
- Information theory
- Interdisciplinary applications of statistical physics
Selected Publications
Weight space structure and internal representations: a direct approach to learning and generalization in multilayer neural networks
Physical review letters 75 (12), 2432, 1995
Statistical mechanics of the random K-satisfiability model
Physical Review E 56 (2), 1357, 1997
Determining computational complexity from characteristic ‘phase transitions’
Nature 400 (6740), 133-137, 1999
Combinatorial and topological approach to the 3D Ising model
, Journal of Physics A: Mathematical and General 33 (4), 741, 2000
Analytic and algorithmic solution of random satisfiability problems
Science 297 (5582), 812-815, 2002
Survey propagation: An algorithm for satisfiability
Random Structures & Algorithms 27 (2), 201-226, 2005
Learning by message passing in networks of discrete synapses
Physical review letters 96 (3), 030201, 2006
Direct-coupling analysis of residue coevolution captures native contacts across many protein families
Proceedings of the National Academy of Sciences 108 (49), E1293-E1301, 2011
Subdominant Dense Clusters Allow for Simple Learning and High Computational Performance in Neural Networks with Discrete Synapses
PHYSICAL REVIEW LETTERS, 2015
Unreasonable effectiveness of learning neural networks: from accessible states and robust ensembles to basic algorithmic schemes
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, 2016
Entropy-SGD: biasing gradient descent into wide valleys
JOURNAL OF STATISTICAL MECHANICS: THEORY AND EXPERIMENT, 2019
Unveiling the Structure of Wide Flat Minima in Neural Networks
PHYSICAL REVIEW LETTERS, 2021