Analysis of Autonomous Penetration Testing Through Reinforcement Learning and Recommender Systems

Conducting penetration testing (pentesting) in cybersecurity is a crucial turning point for identifying vulnerabilities within the framework of Information Technology (IT), where real malicious offensive behavior is simulated to identify potential weaknesses and strengthen preventive controls. Given...

Full description

Saved in:
Bibliographic Details
Main Authors: Ariadna Claudia Moreno, Aldo Hernandez-Suarez, Gabriel Sanchez-Perez, Linda Karina Toscano-Medina, Hector Perez-Meana, Jose Portillo-Portillo, Jesus Olivares-Mercado, Luis Javier García Villalba
Format: Article
Language:English
Published: MDPI AG 2025-01-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/25/1/211
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850072358375653376
author Ariadna Claudia Moreno
Aldo Hernandez-Suarez
Gabriel Sanchez-Perez
Linda Karina Toscano-Medina
Hector Perez-Meana
Jose Portillo-Portillo
Jesus Olivares-Mercado
Luis Javier García Villalba
author_facet Ariadna Claudia Moreno
Aldo Hernandez-Suarez
Gabriel Sanchez-Perez
Linda Karina Toscano-Medina
Hector Perez-Meana
Jose Portillo-Portillo
Jesus Olivares-Mercado
Luis Javier García Villalba
author_sort Ariadna Claudia Moreno
collection DOAJ
description Conducting penetration testing (pentesting) in cybersecurity is a crucial turning point for identifying vulnerabilities within the framework of Information Technology (IT), where real malicious offensive behavior is simulated to identify potential weaknesses and strengthen preventive controls. Given the complexity of the tests, time constraints, and the specialized level of expertise required for pentesting, analysis and exploitation tools are commonly used. Although useful, these tools often introduce uncertainty in findings, resulting in high rates of false positives. To enhance the effectiveness of these tests, Machine Learning (ML) has been integrated, showing significant potential for identifying anomalies across various security areas through detailed detection of underlying malicious patterns. However, pentesting environments are unpredictable and intricate, requiring analysts to make extensive efforts to understand, explore, and exploit them. This study considers these challenges, proposing a recommendation system based on a context-rich, vocabulary-aware transformer capable of processing questions related to the target environment and offering responses based on necessary pentest batteries evaluated by a Reinforcement Learning (RL) estimator. This RL component assesses optimal attack strategies based on previously learned data and dynamically explores additional attack vectors. The system achieved an F1 score and an Exact Match rate over <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>97.0</mn></mrow></semantics></math></inline-formula>%, demonstrating its accuracy and effectiveness in selecting relevant pentesting strategies.
format Article
id doaj-art-72b8277cba2642eaa2f3c5c06af36375
institution DOAJ
issn 1424-8220
language English
publishDate 2025-01-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-72b8277cba2642eaa2f3c5c06af363752025-08-20T02:47:06ZengMDPI AGSensors1424-82202025-01-0125121110.3390/s25010211Analysis of Autonomous Penetration Testing Through Reinforcement Learning and Recommender SystemsAriadna Claudia Moreno0Aldo Hernandez-Suarez1Gabriel Sanchez-Perez2Linda Karina Toscano-Medina3Hector Perez-Meana4Jose Portillo-Portillo5Jesus Olivares-Mercado6Luis Javier García Villalba7Instituto Politecnico Nacional, ESIME Culhuacan, Mexico City 04440, MexicoInstituto Politecnico Nacional, ESIME Culhuacan, Mexico City 04440, MexicoInstituto Politecnico Nacional, ESIME Culhuacan, Mexico City 04440, MexicoInstituto Politecnico Nacional, ESIME Culhuacan, Mexico City 04440, MexicoInstituto Politecnico Nacional, ESIME Culhuacan, Mexico City 04440, MexicoInstituto Politecnico Nacional, ESIME Culhuacan, Mexico City 04440, MexicoInstituto Politecnico Nacional, ESIME Culhuacan, Mexico City 04440, MexicoGroup of Analysis, Security and Systems (GASS), Department of Software Engineering and Artificial Intelligence (DISIA), Faculty of Computer Science and Engineering, Office 431, Universidad Complutense de Madrid (UCM), Calle Profesor José García Santesmases, 9, Ciudad Universitaria, 28040 Madrid, SpainConducting penetration testing (pentesting) in cybersecurity is a crucial turning point for identifying vulnerabilities within the framework of Information Technology (IT), where real malicious offensive behavior is simulated to identify potential weaknesses and strengthen preventive controls. Given the complexity of the tests, time constraints, and the specialized level of expertise required for pentesting, analysis and exploitation tools are commonly used. Although useful, these tools often introduce uncertainty in findings, resulting in high rates of false positives. To enhance the effectiveness of these tests, Machine Learning (ML) has been integrated, showing significant potential for identifying anomalies across various security areas through detailed detection of underlying malicious patterns. However, pentesting environments are unpredictable and intricate, requiring analysts to make extensive efforts to understand, explore, and exploit them. This study considers these challenges, proposing a recommendation system based on a context-rich, vocabulary-aware transformer capable of processing questions related to the target environment and offering responses based on necessary pentest batteries evaluated by a Reinforcement Learning (RL) estimator. This RL component assesses optimal attack strategies based on previously learned data and dynamically explores additional attack vectors. The system achieved an F1 score and an Exact Match rate over <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>97.0</mn></mrow></semantics></math></inline-formula>%, demonstrating its accuracy and effectiveness in selecting relevant pentesting strategies.https://www.mdpi.com/1424-8220/25/1/211penetration testingreinforcement learningrecommender systems
spellingShingle Ariadna Claudia Moreno
Aldo Hernandez-Suarez
Gabriel Sanchez-Perez
Linda Karina Toscano-Medina
Hector Perez-Meana
Jose Portillo-Portillo
Jesus Olivares-Mercado
Luis Javier García Villalba
Analysis of Autonomous Penetration Testing Through Reinforcement Learning and Recommender Systems
Sensors
penetration testing
reinforcement learning
recommender systems
title Analysis of Autonomous Penetration Testing Through Reinforcement Learning and Recommender Systems
title_full Analysis of Autonomous Penetration Testing Through Reinforcement Learning and Recommender Systems
title_fullStr Analysis of Autonomous Penetration Testing Through Reinforcement Learning and Recommender Systems
title_full_unstemmed Analysis of Autonomous Penetration Testing Through Reinforcement Learning and Recommender Systems
title_short Analysis of Autonomous Penetration Testing Through Reinforcement Learning and Recommender Systems
title_sort analysis of autonomous penetration testing through reinforcement learning and recommender systems
topic penetration testing
reinforcement learning
recommender systems
url https://www.mdpi.com/1424-8220/25/1/211
work_keys_str_mv AT ariadnaclaudiamoreno analysisofautonomouspenetrationtestingthroughreinforcementlearningandrecommendersystems
AT aldohernandezsuarez analysisofautonomouspenetrationtestingthroughreinforcementlearningandrecommendersystems
AT gabrielsanchezperez analysisofautonomouspenetrationtestingthroughreinforcementlearningandrecommendersystems
AT lindakarinatoscanomedina analysisofautonomouspenetrationtestingthroughreinforcementlearningandrecommendersystems
AT hectorperezmeana analysisofautonomouspenetrationtestingthroughreinforcementlearningandrecommendersystems
AT joseportilloportillo analysisofautonomouspenetrationtestingthroughreinforcementlearningandrecommendersystems
AT jesusolivaresmercado analysisofautonomouspenetrationtestingthroughreinforcementlearningandrecommendersystems
AT luisjaviergarciavillalba analysisofautonomouspenetrationtestingthroughreinforcementlearningandrecommendersystems