{
  "id": 19058,
  "url": "https://arxiv.org/abs/2608.11495v1",
  "title": "Defending against Model Extraction for GNNs with Model Reprogramming",
  "summary": "Graph Neural Networks (GNNs) serve as the backbone for high-stakes applications in Machine-Learning-as-a-Service (MLaaS). Still, their black-box deployment exposes them to Model Extraction (ME) attacks, in which adversaries steal intellectual property by querying APIs. Existing defenses suffer from a critical ''Euclidean bias'': they transfer image-based strategies (e.g., random noise) to graphs, ignoring the complex topological dependencies between nodes, which often results in severe utility d",
  "authors": "Yan Wen, Zhenyi Wang, Heng Huang",
  "category": "research",
  "topics": "bias-fairness,copyright-ip",
  "orgs": null,
  "regions": null,
  "published_at": "2026-08-11T23:20:15.000Z",
  "fetched_at": "2026-08-13T05:10:37.786Z",
  "source_slug": "x-arxiv-cs-cr-ai-security",
  "source_name": "arXiv cs.CR (AI security)",
  "source_homepage": "https://arxiv.org/list/cs.CR/recent",
  "ethics_ai_record_url": "https://ethics.ai/record/19058",
  "original_url": "https://arxiv.org/abs/2608.11495v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}