# Sensitivity Analysis of Deep Neural Networks

arXiv: https://arxiv.org/abs/1901.07152 · HTML: https://mlsanity.com/paper/1901.07152/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 22 Jan 2019.

## Built on by (most cited first, 6 of 6)

- [Breach in the Shield: Unveiling the Vulnerabilities of Large Language Models](https://mlsanity.com/paper/2504.03714/) (2025)
- [Investigating Weight-Perturbed Deep Neural Networks With Application in Iris Presentation Attack Detection](https://mlsanity.com/paper/2311.12764/) (2023)
- [Data Provenance via Differential Auditing](https://mlsanity.com/paper/2209.01538/) (2022)
- [Top-down Green-ups: Satellite Sensing and Deep Models to Predict Buffelgrass Phenology](https://mlsanity.com/paper/2310.00740/) (2023)
- [Exploring and Exploiting the Asymmetric Valley of Deep Neural Networks](https://mlsanity.com/paper/2405.12489/) (2024)
- [Delta-AttnMask: Attention-Guided Masked Hidden States for Efficient Data Selection and Augmentation](https://mlsanity.com/paper/2508.09199/) (2025)

## Abstract (arXiv)

Deep neural networks (DNNs) have achieved superior performance in various prediction tasks, but can be very vulnerable to adversarial examples or perturbations. Therefore, it is crucial to measure the sensitivity of DNNs to various forms of perturbations in real applications. We introduce a novel perturbation manifold and its associated influence measure to quantify the effects of various perturbations on DNN classifiers. Such perturbations include various external and internal perturbations to input samples and network parameters. The proposed measure is motivated by information geometry and provides desirable invariance properties. We demonstrate that our influence measure is useful for four model building tasks: detecting potential 'outliers', analyzing the sensitivity of model architectures, comparing network sensitivity between training and test sets, and locating vulnerable areas. Experiments show reasonably good performance of the proposed measure for the popular DNN models ResNet50 and DenseNet121 on CIFAR10 and MNIST datasets.