Statistical mechanics of extensive-width Bayesian neural networks near interpolation

• URL: http://arxiv.org/abs/2505.24849v1
• Date: Fri, 30 May 2025 17:46:59 GMT
• Title: Statistical mechanics of extensive-width Bayesian neural networks near interpolation
• Authors: Jean Barbier, Francesco Camilli, Minh-Toan Nguyen, Mauro Pastore, Rudy Skerk,
• Abstract summary: We study the supervised learning of a two-layer fully connected network with generic weight distribution and activation function.<n>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.<n>Our analysis uncovers a rich phenomenology with various learning transitions as the number of data increases.
• Score: 4.976898227858662
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: 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.

Related papers

• Generalization performance of narrow one-hidden layer networks in the teacher-student setting [40.69556943879117]
  We develop a general theory for narrow networks, i.e. networks with a large number of hidden units, yet much smaller than the input dimension.<n>Our theory accurately predicts the generalization error of neural networks trained on regression or classification tasks.
  arXiv  Detail & Related papers  (2025-07-01T10:18:20Z)
• Global Convergence and Rich Feature Learning in $L$-Layer Infinite-Width Neural Networks under $μ$P Parametrization [66.03821840425539]
  In this paper, we investigate the training dynamics of $L$-layer neural networks using the tensor gradient program (SGD) framework.<n>We show that SGD enables these networks to learn linearly independent features that substantially deviate from their initial values.<n>This rich feature space captures relevant data information and ensures that any convergent point of the training process is a global minimum.
  arXiv  Detail & Related papers  (2025-03-12T17:33:13Z)
• Diffusion-Based Neural Network Weights Generation [80.89706112736353]
  D2NWG is a diffusion-based neural network weights generation technique that efficiently produces high-performing weights for transfer learning. Our method extends generative hyper-representation learning to recast the latent diffusion paradigm for neural network weights generation. Our approach is scalable to large architectures such as large language models (LLMs), overcoming the limitations of current parameter generation techniques.
  arXiv  Detail & Related papers  (2024-02-28T08:34:23Z)
• An Entropy-Based Model for Hierarchical Learning [3.1473798197405944]
  A common feature among real-world datasets is that data domains are multiscale. We propose a learning model that exploits this multiscale data structure. The hierarchical learning model is inspired by the logical and progressive easy-to-hard learning mechanism of human beings.
  arXiv  Detail & Related papers  (2022-12-30T13:14:46Z)
• Neural networks trained with SGD learn distributions of increasing complexity [78.30235086565388]
  We show that neural networks trained using gradient descent initially classify their inputs using lower-order input statistics. We then exploit higher-order statistics only later during training. We discuss the relation of DSB to other simplicity biases and consider its implications for the principle of universality in learning.
  arXiv  Detail & Related papers  (2022-11-21T15:27:22Z)
• An Information-Theoretic Framework for Supervised Learning [22.280001450122175]
  We propose a novel information-theoretic framework with its own notions of regret and sample complexity. We study the sample complexity of learning from data generated by deep neural networks with ReLU activation units. We conclude by corroborating our theoretical results with experimental analysis of random single-hidden-layer neural networks.
  arXiv  Detail & Related papers  (2022-03-01T05:58:28Z)
• Data-driven emergence of convolutional structure in neural networks [83.4920717252233]
  We show how fully-connected neural networks solving a discrimination task can learn a convolutional structure directly from their inputs. By carefully designing data models, we show that the emergence of this pattern is triggered by the non-Gaussian, higher-order local structure of the inputs.
  arXiv  Detail & Related papers  (2022-02-01T17:11:13Z)
• Graph network for simultaneous learning of forward and inverse physics [0.0]
  We propose an end-to-end graph network that learns forward and inverse models of particle-based physics using interpretable inductive biases. Our approach is able to predict the forward dynamics with at least an order of magnitude higher accuracy.
  arXiv  Detail & Related papers  (2021-12-13T22:38:09Z)
• Reasoning-Modulated Representations [85.08205744191078]
  We study a common setting where our task is not purely opaque. Our approach paves the way for a new class of data-efficient representation learning.
  arXiv  Detail & Related papers  (2021-07-19T13:57:13Z)
• Statistical Mechanical Analysis of Catastrophic Forgetting in Continual Learning with Teacher and Student Networks [5.209145866174911]
  When a computational system continuously learns from an ever-changing environment, it rapidly forgets its past experiences. We provide the theoretical framework for analyzing catastrophic forgetting by using teacher-student learning. We find that the network can avoid catastrophic forgetting when the similarity among input distributions is small and that of the input-output relationship of the target functions is large.
  arXiv  Detail & Related papers  (2021-05-16T09:02:48Z)
• The Connection Between Approximation, Depth Separation and Learnability in Neural Networks [70.55686685872008]
  We study the connection between learnability and approximation capacity. We show that learnability with deep networks of a target function depends on the ability of simpler classes to approximate the target.
  arXiv  Detail & Related papers  (2021-01-31T11:32:30Z)
• Malicious Network Traffic Detection via Deep Learning: An Information Theoretic View [0.0]
  We study how homeomorphism affects learned representation of a malware traffic dataset. Our results suggest that although the details of learned representations and the specific coordinate system defined over the manifold of all parameters differ slightly, the functional approximations are the same.
  arXiv  Detail & Related papers  (2020-09-16T15:37:44Z)
• The large learning rate phase of deep learning: the catapult mechanism [50.23041928811575]
  We present a class of neural networks with solvable training dynamics. We find good agreement between our model's predictions and training dynamics in realistic deep learning settings. We believe our results shed light on characteristics of models trained at different learning rates.
  arXiv  Detail & Related papers  (2020-03-04T17:52:48Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.