Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
October 2, 2025ComputersOpen Access

Do LLMs Offer a Robust Defense Mechanism Against Membership Inference Attacks on Graph Neural Networks?

View Full Paper
Ask AI
Bookmark
Share

Authors

AJAbdellah JnainiMKMohammed-Amine Koulali

Discussion

Loading...

Member takes

Overview

Proposed LLM-guided methods reduce membership inference attack accuracy in GNNs, implying protection for sensitive data.

Key Points

  • LLM-guided defense mechanisms effectively reduce membership inference attack accuracy and maintain node classification performance.
  • Three methods were introduced: posterior encoding with noise, knowledge distillation, and secure aggregation for GNNs.
  • Extensive experiments validated the effectiveness of these approaches on widely used graph neural network architectures.
  • The findings support the importance of balancing privacy measures with model performance in privacy-sensitive applications.

Cite This Study

Jnaini et al. (2025) studied this question.

synapsesocial.com/papers/68de68f183cbc991d0a217achttps://doi.org/10.3390/computers14100414
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. 1Subgraph Structure Membership Inference Attacks against Graph Neural Networks2024
  2. 2Large Language Models for Link Stealing Attacks Against Graph Neural Networks2024 · 1 citations
  3. 3Graph Transductive Defense: a Two-Stage Defense for Graph Membership Inference Attacks2024
  4. 4Can Large Language Models Improve the Adversarial Robustness of Graph Neural Networks?2024
  5. 5Poster: Enhancing GNN Robustness for Network Intrusion Detection via Agent-based Analysis2025