Log In Sign Up

Anomaly Detection of Test-Time Evasion Attacks using Class-conditional Generative Adversarial Networks

by   Hang Wang, et al.

Deep Neural Networks (DNNs) have been shown vulnerable to adversarial (Test-Time Evasion (TTE)) attacks which, by making small changes to the input, alter the DNN's decision. We propose an attack detector based on class-conditional Generative Adversarial Networks (GANs). We model the distribution of clean data conditioned on the predicted class label by an Auxiliary Classifier GAN (ACGAN). Given a test sample and its predicted class, three detection statistics are calculated using the ACGAN Generator and Discriminator. Experiments on image classification datasets under different TTE attack methods show that our method outperforms state-of-the-art detection methods. We also investigate the effectiveness of anomaly detection using different DNN layers (input features or internal-layer features) and demonstrate that anomalies are harder to detect using features closer to the DNN's output layer.


page 1

page 2

page 3

page 4


Efficient GAN-Based Anomaly Detection

Generative adversarial networks (GANs) are able to model the complex hig...

Detecting Anomalous Inputs to DNN Classifiers By Joint Statistical Testing at the Layers

Detecting anomalous inputs, such as adversarial and out-of-distribution ...

A Uniform Framework for Anomaly Detection in Deep Neural Networks

Deep neural networks (DNN) can achieve high performance when applied to ...

Deep Anomaly Detection Using Geometric Transformations

We consider the problem of anomaly detection in images, and present a ne...

Phase Retrieval using Conditional Generative Adversarial Networks

In this paper, we propose the application of conditional generative adve...

When Not to Classify: Anomaly Detection of Attacks (ADA) on DNN Classifiers at Test Time

A significant threat to the recent, wide deployment of machine learning-...

G2D: Generate to Detect Anomaly

In this paper, we propose a novel method for irregularity detection. Pre...