Statistical mechanics of extensive-width Bayesian neural networks near interpolation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Barbier, Jean, Camilli, Francesco, Nguyen, Minh-Toan, Pastore, Mauro, Skerk, Rudy
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913868060557312
author Barbier, Jean
Camilli, Francesco
Nguyen, Minh-Toan
Pastore, Mauro
Skerk, Rudy
author_facet Barbier, Jean
Camilli, Francesco
Nguyen, Minh-Toan
Pastore, Mauro
Skerk, Rudy
contents For three decades statistical mechanics has been providing a framework to analyse neural networks. However, the theoretically tractable models, e.g., perceptrons, random features models and kernel machines, or multi-index models and committee machines with few neurons, remained simple compared to those used in applications. In this paper we help reducing the gap between practical networks and their theoretical understanding through a statistical physics analysis of the supervised learning of a two-layer fully connected network with generic weight distribution and activation function, whose hidden layer is large but remains proportional to the inputs dimension. This makes it more realistic than infinitely wide networks where no feature learning occurs, but also more expressive than narrow ones or with fixed inner weights. We focus on the Bayes-optimal learning in the teacher-student scenario, i.e., with a dataset generated by another network with the same architecture. We operate around interpolation, where the number of trainable parameters and of data are comparable and feature learning emerges. Our analysis uncovers a rich phenomenology with various learning transitions as the number of data increases. In particular, the more strongly the features (i.e., hidden neurons of the target) contribute to the observed responses, the less data is needed to learn them. Moreover, when the data is scarce, the model only learns non-linear combinations of the teacher weights, rather than "specialising" by aligning its weights with the teacher's. Specialisation occurs only when enough data becomes available, but it can be hard to find for practical training algorithms, possibly due to statistical-to-computational~gaps.
format Preprint
id arxiv_https___arxiv_org_abs_2505_24849
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Statistical mechanics of extensive-width Bayesian neural networks near interpolation
Barbier, Jean
Camilli, Francesco
Nguyen, Minh-Toan
Pastore, Mauro
Skerk, Rudy
Machine Learning
Disordered Systems and Neural Networks
Statistical Mechanics
Information Theory
For three decades statistical mechanics has been providing a framework to analyse neural networks. However, the theoretically tractable models, e.g., perceptrons, random features models and kernel machines, or multi-index models and committee machines with few neurons, remained simple compared to those used in applications. In this paper we help reducing the gap between practical networks and their theoretical understanding through a statistical physics analysis of the supervised learning of a two-layer fully connected network with generic weight distribution and activation function, whose hidden layer is large but remains proportional to the inputs dimension. This makes it more realistic than infinitely wide networks where no feature learning occurs, but also more expressive than narrow ones or with fixed inner weights. We focus on the Bayes-optimal learning in the teacher-student scenario, i.e., with a dataset generated by another network with the same architecture. We operate around interpolation, where the number of trainable parameters and of data are comparable and feature learning emerges. Our analysis uncovers a rich phenomenology with various learning transitions as the number of data increases. In particular, the more strongly the features (i.e., hidden neurons of the target) contribute to the observed responses, the less data is needed to learn them. Moreover, when the data is scarce, the model only learns non-linear combinations of the teacher weights, rather than "specialising" by aligning its weights with the teacher's. Specialisation occurs only when enough data becomes available, but it can be hard to find for practical training algorithms, possibly due to statistical-to-computational~gaps.
title Statistical mechanics of extensive-width Bayesian neural networks near interpolation
topic Machine Learning
Disordered Systems and Neural Networks
Statistical Mechanics
Information Theory
url https://arxiv.org/abs/2505.24849