Robustness of neural networks to random noise perturbations of their inputs

Mark Levene, Martyn Harris
6/30/2026
cs.LGcs.NE

No official code found. Best implementation: martyn-harris-bbk/NN_Perturbation_Robustness (0 stars, last commit 2026-04-21).

Abstract

We investigate the problem of the robustness of a trained neural network to the perturbation of its input values. More specifically, we examine the interplay between the accuracy of the network, as measured by the mean squared error, and robustness. Accordingly, we present a robustness measure, which, with high probability, suggests an upper bound on the mean squared error of the network, with respect to an input data set, for a given perturbation of the input values of the network. The measure we propose is both simple and efficient to compute, treating the neural network as a black box. We provide experimental results on several real-world data sets showing the efficacy of the proposed method. We also introduce the concept of robustness curves, which allows us to further analyse robustness within and between data sets.

View on arXivView PDF

Code Implementations(1)

martyn-harris-bbk/NN_Perturbation_RobustnessCommunity64%

A data repository for the paper titled "Robustness of neural networks to random noise perturbations of their inputs"

00Apr 21, 20266 months ago

Similar Papers

Lipschitz-Margin Training: Scalable Certification of Perturbation Invariance for Deep Neural Networks

84% match1 repo352018

Towards Certified Probabilistic Robustness with High Accuracy

83% match1 repo2023

Certified Robust Accuracy of Neural Networks Are Bounded due to Bayes Errors

83% match2024

Estimating Neural Network Robustness via Lipschitz Constant and Architecture Sensitivity

83% match2024

An Estimator for the Sensitivity to Perturbations of Deep Neural Networks

83% match2023

How Does Bayes Error Limit Probabilistic Robust Accuracy

83% match2024

Cite this paper

@article{levene2026robustness,
  title  = {Robustness of neural networks to random noise perturbations of their inputs},
  author = {Mark Levene and Martyn Harris},
  year   = {2026},
  eprint = {2606.31581},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2606.31581v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2606.31581.svg)](https://codeofpaper.com/paper/2606.31581)

Discussion