Sampling the space of solutions of an artificial neural network

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zambon, Alessandro, Malatesta, Enrico M., Tiana, Guido, Zecchina, Riccardo
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912498795413504
author Zambon, Alessandro
Malatesta, Enrico M.
Tiana, Guido
Zecchina, Riccardo
author_facet Zambon, Alessandro
Malatesta, Enrico M.
Tiana, Guido
Zecchina, Riccardo
contents The weight space of an artificial neural network can be systematically explored using tools from statistical mechanics. We employ a combination of a hybrid Monte Carlo algorithm which performs long exploration steps, a ratchet-based algorithm to investigate connectivity paths, and coupled replica models simulations to study subdominant flat regions. Our analysis focuses on one hidden layer networks and spans a range of energy levels and constrained density regimes. Near the interpolation threshold, the low-energy manifold shows a spiky topology. In the overparameterized regime, however, the low-energy manifold becomes entirely flat, forming an extended complex structure that is easy to sample. These numerical results are supported by an analytical study of the training error landscape, and we show numerically that the qualitative features of the loss landscape are robust across different data structures. Our study aims to provide new methodological insights for developing scalable methods for large networks.
format Preprint
id arxiv_https___arxiv_org_abs_2503_08266
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Sampling the space of solutions of an artificial neural network
Zambon, Alessandro
Malatesta, Enrico M.
Tiana, Guido
Zecchina, Riccardo
Disordered Systems and Neural Networks
Statistical Mechanics
Probability
The weight space of an artificial neural network can be systematically explored using tools from statistical mechanics. We employ a combination of a hybrid Monte Carlo algorithm which performs long exploration steps, a ratchet-based algorithm to investigate connectivity paths, and coupled replica models simulations to study subdominant flat regions. Our analysis focuses on one hidden layer networks and spans a range of energy levels and constrained density regimes. Near the interpolation threshold, the low-energy manifold shows a spiky topology. In the overparameterized regime, however, the low-energy manifold becomes entirely flat, forming an extended complex structure that is easy to sample. These numerical results are supported by an analytical study of the training error landscape, and we show numerically that the qualitative features of the loss landscape are robust across different data structures. Our study aims to provide new methodological insights for developing scalable methods for large networks.
title Sampling the space of solutions of an artificial neural network
topic Disordered Systems and Neural Networks
Statistical Mechanics
Probability
url https://arxiv.org/abs/2503.08266