activity
20242026
collaborators

6 papers

cs.LG2026

Kernel Renormalization in Bayesian Deep Neural Networks: the Equivalent Wishart Ansatz in the Proportional Regime

Paolo Baglioni, Christian Keup, Vincenzo Zimbardo +4

The scaling limit where both the size of the training set and the width of a deep neural network grow at the same rate, the so-called proportional-width regime, has been in…

cond-mat.stat-mech2025

Unveiling the Dimensionality of Networks of Networks

Lorenzo Grimaldi, Pablo Villegas, Alessandro Vezzani +3

"Every object that biology studies is a system of systems." (François Jacob, 1974). Most networks feature intricate architectures originating from tinkering, a repetitive use of e…

physics.soc-ph2025

Pedestrian fluxes in confined geometric networks: entropic measures and robustness of accessibility in a university campus

Adamo Cerioli, Barbara Caselli, Lea Jeanne Marinelli +2

When discussing urban life, pedestrian accessibility to all main services is crucial for fostering social interactions, promoting healthy lifestyles, and reducing pollution. This i…

cond-mat.stat-mech2025

Rare Events and Redundancy in Random Walkers Target Search in a Finite Domain

Elisabetta Ellettari, Giacomo Nasuti, Alberto Bassanoni +2

Finding a target in a complex environment is a fundamental challenge across natural systems, from chemical reactions to sperm cells reaching an egg. A powerful strategy to reduce s…

cond-mat.stat-mech2025

Rare Events and Single Big Jump Effects in Ornstein-Uhlenbeck Processes

Alberto Bassanoni, Alessandro Vezzani, Eli Barkai +1

Even in a simple stochastic process, the study of the full distribution of time integrated observables can be a difficult task. This is the case of a much-studied process such as t…

cond-mat.dis-nn2024

Kernel shape renormalization explains output-output correlations in finite Bayesian one-hidden-layer networks

P. Baglioni, L. Giambagli, A. Vezzani +3

Finite-width one hidden layer networks with multiple neurons in the readout layer display non-trivial output-output correlations that vanish in the lazy-training infinite-width lim…