Embedded Files
Les Quais de Saône, Lyon
Hi!
I am a postdoctoral researcher at EPFL in the Information, Learning & Physics Laboratory (IdePHICS), working with Florent Krzakala. My research lies at the intersection of probability theory (especially random matrix theory and high-dimensional probability), high-dimensional statistics, and the mathematical foundations of machine learning.
Before joining EPFL, I completed my Ph.D. in Mathematics at ENS Lyon, co-advised by Alice Guionnet (CNRS and ENS Lyon) and Gérard Ben Arous (NYU Courant). My doctoral thesis is available online. Prior to that, I studied mathematics at ETH Zurich, where I worked with Dominik Schröder on my master's thesis. You can find more about my work on the Research page or in my CV.
I am a member of the European Women in Mathematics and a signatory of the Slow Science Manifesto in support of a more inclusive and sustainable academic environment.
Feel free to contact me at vanessa [dot] piccolo [at] epfl [dot] ch!
News
News
October 2026: New paper Symmetry-Aware Feature Learning: A Polynomial Separation for Multi-Index Models, with Jivan Waber, Yatin Dandi, and Florent Krzakala. We study how symmetry changes feature learning in high-dimensional Gaussian multi-index models.
Fall 2026: I am teaching Statistical Physics of Computation at EPFL. The course is organized into three main modules: inference and information-theoretic bounds, taught by Vittorio Erba; random matrix theory; and learning from data and neural networks. If you'd like access to the lecture notes or other course materials, feel free to get in touch.
September 2026: I joined the Simons Collaboration on the Physics of Learning and Neural Computation as a postdoctoral researcher.
August 2026: New paper Spectral Phase Transitions in Gaussian Multi-Index Models, with Florent Krzakala and Pierre Mergny. We establish sharp spectral phase transitions for matrix-valued spectral estimators and show that an optimal spectral method achieves the weak recovery threshold predicted by approximate message passing.
Page updated
Google Sites
Report abuse