People

Jean Barbier

I’m Jean Barbier, a Research Scientist (Tenured Associate Professor) in mathematical physics of signals and learning at the International Center for Theoretical Physics (ICTP), part of both the Quantitative Life Science and Mathematics sections. The ICTP is a UNESCO institute whose mission is not only the research excellence, but also teaching and scientific capacity building for the developing world: ICTP really is a special place to do top research and make a positive impact, worldwide. Prior to joining ICTP,  I did my PhD at École Normale Supérieure of Paris with Florent Krzakala followed by a postdoc at EPFL in Lausanne with Nicolas Macris. 

My research interests revolve around information processing systems such as appearing in machine learning, communication and error-correction, signal processing or computer science. I often study these systems and associated algorithms using statistical physics –the language used to describe phase transitions–, its close cousin information theory, and random matrix theory. I try to precisely quantify what is the optimal performance one can aim for when processing (big) data, as well as how close to optimality one can operate when using computationally efficient algorithms.

Website                ORCID

    Google scholar               Curriculum Vitae

Daria Tieplova

Dasha earned her bachelor and master degrees in mathematics at the National University of Kharkiv before obtaining her joint PhD on Application of large random matrices to multivariate time series analysis at the LIGM under Philippe Loubaton and the Verkin ILTPE under Leonid Pastur.

At the ICTP, her current research interests lie in applications of random matrix theory to obtaining information-theoretic limits for high-dimensional inference problems. In particular, she is presently studying high-dimensional time series using spin-glass methods and has recently obtained universality results for neural networks stemming from gaussian equivalence principles.

Google scholar

Francesco Camilli

Francesco Camilli got his bachelor and master degrees in Physics at the University of Bologna. He was a member of the Collegio Superiore, an institute of excellence of the University of Bologna.

He earned a PhD in Mathematics from the University of Bologna under the supervision of prof. Pierluigi Contucci, and a PhD in Physics in cotutelle with the École Normale Supérieure of Paris under the supervision of prof. Marc Mézard. His PhD thesis, titled "New perspectives in statistical mechanics and high-dimensional inference", was awarded the Indam-UMI-SIMAI national prize for best Italian PhD thesis in applied math.

His research interests range from spin glasses to information theory and high dimensional inference. 

Website                ORCID

Google scholar

Minh Toan Nguyen

Minh Toan received his engineering degree at École Polytechnique and his Master degree from the program "Mathematics of Randomness" at Paris-Saclay University in 2020. He got his PhD at Grenoble Alps University, under the supervision of Romain Couillet, with the thesis Replica method and asymptotic equivalence.

His research interests include information theory, random matrix theory, statistical physics and their applications on high dimensional inference.

Google Scholar 

Mauro Pastore

Mauro got his PhD in theoretical physics at the University of Milan, with a thesis on Replicas in complex systems: applications to large deviations and neural networks, advised by Sergio Caracciolo.

Interested in the interdisciplinary applications of statistical mechanics, he spent a few years in Paris as a postdoctoral researcher, first at the LPTMS working on spin glasses and neural networks with Silvio Franz, in the Simons collaboration on cracking the glass problem, then at the Physics Laboratory of the ENS, working with Simona Cocco and Rémi Monasson on inference problems in biology.

Among his current research interests, he would like to understand the feature learning regime of overparametrized neural networks and their connection with kernel methods, the role of data structure in machine learning, the problem of class imbalance in supervised learning.

Google scholar ORCID

Gibbs Nwemadji

Gibbs Nwemadji is currently pursuing a PhD in the Theoretical and Scientific Data Science group at the International School for Advanced Studies (SISSA) in Trieste, Italy. His research focuses on utilizing statistical physics to enhance our theoretical comprehension of machine learning systems.

Before joining SISSA, Gibbs excelled in the ICTP postgraduate diploma program in Quantitative Life Science, where he was recognized as the top student of his cohort. His admission to this prestigious program was greatly influenced by the connections he established during his master's studies in mathematical science at AIMS-Cameroon, where he graduated among the top five in his class. Prior to his time at AIMS-Cameroon, Gibbs pursued physics at the University of Douala, where he graduated as the second valedictorian of his master's program.

Outside of academia, Gibbs enjoys immersing himself in books, staying active through activities like basketball and running, and engaging in discussions on topics related to human nature and the purpose of life.

Rudy Skerk

Rudy Skerk obtained his bachelor's degree in Physics at the University of Trieste. He then enrolled in the International Master of Physics of Complex Systems, based partly in Italy and partly in France: at the International School of Advanced Studies (SISSA) and the International Centre for Theoretical Physics (ICTP) in Trieste, at the Politecnico in Torino, and at a consortium of Sorbonne Université, Université Paris-Cité, and Université Paris-Saclay in the Paris area.

He is now a PhD student in the Theoretical and Scientific Data Science group at the International School for Advanced Studies (SISSA) in Trieste. His research interests lie in the application of tools from statistical mechanics, information theory, and random matrix theory to tackle theoretical and numerical aspects of optimization, communication, inference, and learning.

Ali Umar Hussaini

Ali is currently a PhD student in the Theoretical and Scientific Data Science group at the International School for Advanced Studies (SISSA) in Trieste. Broadly, he is interested in understanding deep learning through the lens of network representations and weights using tools from the mathematical sciences.

Ali obtained his bachelor's degree in mathematics from Yobe State University and a master's degree in mathematical sciences from the African Institute for Mathematical Sciences (AIMS) in Ghana. Before joining SISSA, he completed a diploma program in Quantitative Life Sciences (QLS) at the Abdus Salam International Center for Theoretical Physics (ICTP), where he did his diploma thesis under the supervision of Jean Barbier on the numerical investigation of optimization algorithms for the maximum independent set on sparse Erdos-Renyi graphs.