Vulnerability Detection Through an Adversarial Fuzzing Algorithm

TL;DR

This paper introduces an adversarial fuzzing algorithm that enhances the efficiency of vulnerability detection by enabling fuzzers to explore more paths and find more bugs faster, making security testing more accessible.

Contribution

It presents a novel adversarial approach built on evolutionary algorithms to improve fuzzing efficiency and bug discovery rate.

Findings

Adversarial fuzzing outperforms existing methods significantly.

Increased number of crashes and bugs found.

Enhanced path exploration in fuzzing process.

Abstract

Fuzzing is a popular vulnerability automated testing method utilized by professionals and broader community alike. However, despite its abilities, fuzzing is a time-consuming, computationally expensive process. This is problematic for the open source community and smaller developers, as most people will not have dedicated security professionals and/or knowledge to perform extensive testing on their own. The goal of this project is to increase the efficiency of existing fuzzers by allowing fuzzers to explore more paths and find more bugs in shorter amounts of time, while still remaining operable on a personal device. To accomplish this, adversarial methods are built on top of current evolutionary algorithms to generate test cases for further and more efficient fuzzing. The results of this show that adversarial attacks do in fact increase outpaces existing fuzzers significantly and, consequently, crashes found.