Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
March 18, 2024Open Access

Problem space structural adversarial attacks for Network Intrusion Detection Systems based on Graph Neural Networks

View Full Paper
Ask AI
Bookmark
Share

Authors

AVAndrea VenturiDSDario StabiliMMMirco Marchetti

Discussion

Loading...

Member takes

Overview

Experimental evaluation demonstrates structural adversarial attacks compromise graph neural networks in network intrusion detection systems, highlighting critical architectural vulnerabilities.

Key Points

  • Graph neural networks exhibit high resilience against classical feature perturbations, but they remain highly susceptible to targeted structural attacks.
  • Experimental campaign against state-of-the-art graph neural networks evaluates feasible structural attacks under realistic problem space constraints.
  • Highlights the urgent requirement for robust topological defense mechanisms, as network intrusion detection systems face emerging structural threats.

Cite This Study

Venturi et al. (2024) studied this question.

synapsesocial.com/papers/68e73a7cb6db6435876b3a82https://doi.org/10.48550/arxiv.2403.11830
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Poster: Enhancing GNN Robustness for Network Intrusion Detection via Agent-based Analysis2025
  2. 2On adversarial attack detection in intrusion detection system with graph neural network2025
  3. 3Enhancing Multi-Class Attack Detection in Graph Neural Network through Feature Rearrangement2024 · 25 citations
  4. 4Topological Adversarial Attacks on Graph Neural Networks Via Projected Meta Learning2024 · 4 citations
  5. 5An optimized graph neural network approach for robust and explainable IoT intrusion detection against adversarial attacks2026 · 3 citations