Dueling Deep Q-Learning for Intrusion Detection

2026-08-11

AI Summary

This work addresses the limited generalization of traditional supervised learning methods against novel cyberattacks by proposing a novel intrusion detection approach based on Dueling Deep Q-Learning. By decoupling the value and advantage streams, the method enhances the stability and efficiency of reinforcement learning while integrating SHAP (SHapley Additive exPlanations) to provide interpretable decision-making. Experimental evaluation on the CIC-IDS2018 dataset demonstrates that the proposed framework achieves an average detection accuracy of 99.68% across multiple attack types, significantly improving adaptability to previously unseen attacks. The approach thus offers a compelling combination of high performance and model transparency, advancing the feasibility of reinforcement learning for real-world intrusion detection systems.

Abstract

Intrusion detection systems (IDS) and automated systems for detecting and reporting cyber threats, are commonly handled via supervised machine learning methods. Though effective, these models struggle to effectively adapt to new attack types. This study proposes a novel approach by employing a reward-based, dueling Q-learning model for IDS, achieving an average accuracy of 99.68% across multiple attack classes. The proposed model has a dueling network architecture which separates its predictions into value and advantage streams. This has the benefit of improving learning efficiency and stability. The model was trained on the CIC-IDS2018, a benchmark dataset based on real-world intrusion detection scenarios, having multiple attack classes such as DDoS, botnets, and brute-force attacks. Furthermore, Explainable AI (XAI), specifically SHAP (SHapley Additive exPlanations), was also integrated into the training and evaluation process to provide interpretability into the model's predictions.

Problem

• Intrusion Detection
• Cyber Threats
• Adaptation to New Attacks
• Supervised Machine Learning
• Network Security

Innovation

• Dueling Deep Q-Learning
• Intrusion Detection System
• Explainable AI
• SHAP
• Reinforcement Learning