A Comprehensive Review on Adversarial Machine Learning and XAI for Network Intrusion Detection

View at publisher Summarise with VIRAT
Authors: Karthika Subbaraj, Ritujaa Balachandar, Tanushree Panneer Selvam
Cited by: 0
References: 27

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Machine-learning models are becoming more and more vital in cybersecurity systems, but they are vulnerable to adversarial attacks causing severe concerns for safety-critical applications. This paper presents a systematic literature review of recent works which combines adversarial machine learning with explainable AI (XAI) for cybersecurity systems. The reviewed papers cover generative adversarial training for robust intrusion detection, hybrid XAI-driven detection pipelines, and concept analyses exposing the systematic manipulation of latent representations. Across these works, XAI is applied to interpret model decisions as well as in detecting adversarial behavior. Several defense mechanisms such as adversarial training, deception mechanisms and rule-based reasoning have also been provided. The review shows new designs, trade-offs between robustness, and computational cost, and difficulties in building reliable AI-enabled security systems.

Fields of Study & Subjects

Adversarial Robustness in Machine Learning Network Security and Intrusion Detection Explainable Artificial Intelligence (XAI)

Publication details

Journal / source
International Conference on Computing, Communication, Control and Cyber-Physical Systems (I5CPS)
Pages
1–6
Publication date
26 February 2026
Language
EN