# Locally optimal detection of stochastic targeted universal adversarial perturbations

arXiv: https://arxiv.org/abs/2012.04692 · HTML: https://mlsanity.com/paper/2012.04692/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 8 Dec 2020.

## Built on by (most cited first, 1 of 1)

- [A Subspace Projection Approach to Autoencoder-based Anomaly Detection](https://mlsanity.com/paper/2302.07643/) (2023)

## Abstract (arXiv)

Deep learning image classifiers are known to be vulnerable to small adversarial perturbations of input images. In this paper, we derive the locally optimal generalized likelihood ratio test (LO-GLRT) based detector for detecting stochastic targeted universal adversarial perturbations (UAPs) of the classifier inputs. We also describe a supervised training method to learn the detector's parameters, and demonstrate better performance of the detector compared to other detection methods on several popular image classification datasets.