Research Article
Intelligent Automated Penetration Testing Using Reinforcement Learning to Improve the Efficiency and Effectiveness of Penetration Testing
@INPROCEEDINGS{10.1007/978-3-031-34497-8_3, author={Mohammed Y. A. Aqra and Xiaoqiang Di}, title={Intelligent Automated Penetration Testing Using Reinforcement Learning to Improve the Efficiency and Effectiveness of Penetration Testing}, proceedings={Mobile Wireless Middleware, Operating Systems and Applications. 11th EAI International Conference, MOBILWARE 2022, Virtual Event, December 28-29, 2022, Proceedings}, proceedings_a={MOBILWARE}, year={2023}, month={5}, keywords={Machin learning Deep Reinforcement Learning IAPTS HRL-GIP}, doi={10.1007/978-3-031-34497-8_3} }- Mohammed Y. A. Aqra
Xiaoqiang Di
Year: 2023
Intelligent Automated Penetration Testing Using Reinforcement Learning to Improve the Efficiency and Effectiveness of Penetration Testing
MOBILWARE
Springer
DOI: 10.1007/978-3-031-34497-8_3
Abstract
A penetration test is a process that involves planning, generating, and evaluating attacks that are designed to find and exploit vulnerabilities in digital assets. It can be used in large networks to evaluate the security of their infrastructure. Despite the use of automated tools, it can still be very time consuming and repetitive. The goal of this paper is to develop an intelligent automated penetration testing framework that uses reinforcement learning to improve the efficiency and effectiveness of penetration testing. It utilizes a model-based approach to automate the sequential decision-making process. The framework’s main component is a partial observed Markov decision process that is solved using an external algorithm.
One of the biggest challenges in performing penetration tests on large networks is finding and evaluating clusters of vulnerabilities. This paper presents a method that combines a hierarchical network model with a cluster-based approach. It allows for faster and more accurate testing compared to previous methods. The results of the study show that the IAPTF method outperforms other approaches in terms of time, accuracy, and human performance. One of the main advantages of IAPTF is its ability to perform repetitive tests, which is typically not possible with traditional methods. This method could potentially replace manual pen testing.
