On the Generalization Behavior of Deep Residual Networks From a Dynamical System Perspective

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Huang, Jinshu, Sun, Mingfei, Wu, Chunlin
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912923247443968
author Huang, Jinshu
Sun, Mingfei
Wu, Chunlin
author_facet Huang, Jinshu
Sun, Mingfei
Wu, Chunlin
contents Deep neural networks (DNNs) have significantly advanced machine learning, with model depth playing a central role in their successes. The dynamical system modeling approach has recently emerged as a powerful framework, offering new mathematical insights into the structure and learning behavior of DNNs. In this work, we establish generalization error bounds for both discrete- and continuous-time residual networks (ResNets) by combining Rademacher complexity, flow maps of dynamical systems, and the convergence behavior of ResNets in the deep-layer limit. The resulting bounds are of order $O(1/\sqrt{S})$ with respect to the number of training samples $S$, and include a structure-dependent negative term, yielding depth-uniform and asymptotic generalization bounds under milder assumptions. These findings provide a unified understanding of generalization across both discrete- and continuous-time ResNets, helping to close the gap in both the order of sample complexity and assumptions between the discrete- and continuous-time settings.
format Preprint
id arxiv_https___arxiv_org_abs_2602_20921
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle On the Generalization Behavior of Deep Residual Networks From a Dynamical System Perspective
Huang, Jinshu
Sun, Mingfei
Wu, Chunlin
Machine Learning
Deep neural networks (DNNs) have significantly advanced machine learning, with model depth playing a central role in their successes. The dynamical system modeling approach has recently emerged as a powerful framework, offering new mathematical insights into the structure and learning behavior of DNNs. In this work, we establish generalization error bounds for both discrete- and continuous-time residual networks (ResNets) by combining Rademacher complexity, flow maps of dynamical systems, and the convergence behavior of ResNets in the deep-layer limit. The resulting bounds are of order $O(1/\sqrt{S})$ with respect to the number of training samples $S$, and include a structure-dependent negative term, yielding depth-uniform and asymptotic generalization bounds under milder assumptions. These findings provide a unified understanding of generalization across both discrete- and continuous-time ResNets, helping to close the gap in both the order of sample complexity and assumptions between the discrete- and continuous-time settings.
title On the Generalization Behavior of Deep Residual Networks From a Dynamical System Perspective
topic Machine Learning
url https://arxiv.org/abs/2602.20921