mlsanityData to 7 Oct 2026

Deep hierarchical reinforcement agents for automated penetration testing

arXiv 2109.06449 ↗First posted 14 Sep 2021cs.AI cs.CR cs.LGcited by 19

Not on the map. Why?

Abstract (arXiv)

Penetration testing the organised attack of a computer system in order to test existing defences has been used extensively to evaluate network security. This is a time consuming process and requires in-depth knowledge for the establishment of a strategy that resembles a real cyber-attack. This paper presents a novel deep reinforcement learning architecture with hierarchically structured agents called HA-DRL, which employs an algebraic action decomposition strategy to address the large discrete action space of an autonomous penetration testing simulator where the number of actions is exponentially increased with the complexity of the designed cybersecurity network. The proposed architecture is shown to find the optimal attacking policy faster and more stably than a conventional deep Q-learning agent which is commonly used as a method to apply artificial intelligence in automatic penetration testing.

Key facts. Cited by 19 papers in the map (data through 7 Oct 2026). Builds on 13 of its 23 references that are in the map. First posted on arXiv 14 Sep 2021.

Builds on

13 of 23 references.

References · 13

  1. 2015Continuous control with deep reinforcement learningcited by 3,279
  2. 2015Dueling Network Architectures for Deep Reinforcement Learningcited by 769
  3. 2018QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learningcited by 697
  4. 2015Deep Reinforcement Learning in Large Discrete Action Spacescited by 158
  5. 2017Action Branching Architectures for Deep Reinforcement Learningcited by 90
  6. 2018Learn What Not to Learn: Action Elimination with Deep Reinforcement Learningcited by 77
  7. 2019Learning Action Representations for Reinforcement Learningcited by 75
  8. 2015Reinforcement Learning with Parameterized Actionscited by 61
  9. 2019Hybrid Actor-Critic Reinforcement Learning in Parameterized Action Spacecited by 53
  10. 2019Autonomous Penetration Testing using Reinforcement Learningcited by 33
  11. 2020Q-Learning in enormous action spaces via amortized approximate maximizationcited by 30
  12. 2019Growing Action Spacescited by 29
1 more
  1. 2020CybORG: An Autonomous Cyber Operations Research Gymcited by 15

Built on by

Cited by year

Citing papers per year: 2021: 1; 2022: 6; 2023: 5; 2024: 4; 2025: 3.

19 papers.

Sort:

By area:

  1. 2022Beyond CAGE: Investigating Generalization of Learned Autonomous Network Defense Policiescited by 22
  2. 2023Automated Cyber Defence: A Reviewcited by 19
  3. 2021Intrusion Prevention through Optimal Stoppingcited by 18
  4. 2023Inroads into Autonomous Network Defence using Explained Reinforcement Learningcited by 17
  5. 2023Learning Near-Optimal Intrusion Responses Against Dynamic Attackerscited by 13
  6. 2023A Multiagent CyberBattleSim for RL Cyber Operation Agentscited by 8
  7. 2022Behaviour-Diverse Automatic Penetration Testing: A Curiosity-Driven Multi-Objective Deep Reinforcement Learning Approachcited by 7
  8. 2023Deep Reinforcement Learning for Autonomous Cyber Defence: A Surveycited by 6
  9. 2025AutoPentester: An LLM Agent-based Framework for Automated Pentestingcited by 6
  10. 2022Learning Security Strategies through Game Play and Optimal Stoppingcited by 5
  11. 2024SoK: On the Offensive Potential of AIcited by 5
  12. 2022Exposing Surveillance Detection Routes via Reinforcement Learning, Attack Graphs, and Cyber Terraincited by 3
7 more
  1. 2024Knowledge-Informed Auto-Penetration Testing Based on Reinforcement Learning with Reward Machinecited by 3
  2. 2024Evaluation of Reinforcement Learning for Autonomous Penetration Testing using A3C, Q-learning and DQNcited by 3
  3. 2025xOffense: An Autonomous Multi-Agent Framework for Penetration Testing with Domain-Adapted Large Language Modelscited by 2
  4. 2025Training RL Agents for Multi-Objective Network Defense Taskscited by 1
  5. 2022GAIL-PT: A Generic Intelligent Penetration Testing Framework with Generative Adversarial Imitation Learningcited by 0
  6. 2022Tutorial on Course-of-Action (COA) Attack Search Methods in Computer Networkscited by 0
  7. 2024Discovering Command and Control (C2) Channels on Tor and Public Networks Using Reinforcement Learningcited by 0