Stefano Vigogna

Associate Professor

University of Rome Tor Vergata, Department of Mathematics

Rome Center on Mathematics for Modeling and Data Sciences (RoMaDS)



  Machine Learning, Statistical Learning Theory, Harmonic Analysis


  [last name] [at] mat [dot] uniroma2 [dot] it


  Via della Ricerca Scientifica 1, 00133 Roma, Italy



Research

Fractal and Regular Geometry of Deep Neural Networks

S Di Lillo, D Marinucci, M Salvi, S Vigogna


A Lipschitz spaces view of infinitely wide shallow neural networks

F Bartolucci, M Carioni, J A Iglesias, Y Korolev, E Naldi, S Vigogna


Spectral complexity of deep neural networks

S Di Lillo, D Marinucci, M Salvi, S Vigogna


Neural reproducing kernel Banach spaces and representer theorems for deep networks

F Bartolucci, E De Vito, L Rosasco, S Vigogna


A Quantitative Functional Central Limit Theorem for Shallow Neural Networks

V Cammarota, D Marinucci, M Salvi, S Vigogna


How many samples are needed to leverage smoothness?

V Cabannes, S Vigogna


A Case of Exponential Convergence Rates for SVM

V Cabannes, S Vigogna


Multiclass learning with margin: exponential rates with no bias-variance trade-off

S Vigogna, G Meanti, E De Vito, L Rosasco

Understanding neural networks with reproducing kernel Banach spaces

F Bartolucci, E De Vito, L Rosasco, S Vigogna

Park: Sound and efficient kernel ridge regression by feature space partitions

L Carratino, S Vigogna, D Calandriello, L Rosasco

Construction and Monte Carlo Estimation of Wavelet Frames Generated by a Reproducing Kernel

E De Vito, Z Kereta, V Naumova, L Rosasco, S Vigogna


Estimating multi-index models with response-conditional least squares

T Klock, A Lanteri, S Vigogna

Conditional regression for single-index models

A Lanteri, M Maggioni, S Vigogna

Multiscale regression on unknown manifolds

W Liao, M Maggioni, S Vigogna

  Google Scholar

Teaching

Fall 2024

Spring 2025