Machine Beats Machine: Machine Learning Models to Defend Against Adversarial Attacks

TL;DR

This paper introduces a two-layer machine learning approach to detect and defend against adversarial attacks, utilizing clustering algorithms like DBSCAN for improved detection accuracy.

Contribution

It presents a novel two-layer deployment framework and explores multiple feature sets and datasets, highlighting clustering algorithms' effectiveness in adversarial attack detection.

Findings

Clustering algorithms, especially DBSCAN, show promising results in detecting tampered data.

Using structural similarity index measures enhances detection performance.

The proposed method effectively distinguishes between tampered and genuine data in various datasets.

Abstract

We propose using a two-layered deployment of machine learning models to prevent adversarial attacks. The first layer determines whether the data was tampered, while the second layer solves a domain-specific problem. We explore three sets of features and three dataset variations to train machine learning models. Our results show clustering algorithms achieved promising results. In particular, we consider the best results were obtained by applying the DBSCAN algorithm to the structured structural similarity index measure computed between the images and a white reference image.