Research map / Adversarial robustness and security / Attacks on text and graphs
Adversarial Example Detection Using Latent Neighborhood Graph
Ahmed Abusnaina, Yuhang Wu, Sunpreet Arora, Yizhen Wang, Fei Wang, Hao Yang, David Mohaisen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add adversarial examples to the reference set (ablation) | LNG · CIFAR-10 PGD | AUC: 85.62 -> 91.39 | Improved |
| add adversarial examples to the reference set (ablation) | k-NNG · CIFAR-10 PGD | AUC: 54.29 -> 85.78 | Improved |
| add adversarial examples to the reference set (ablation) | k-NNG · ImageNet CW | AUC: 50.49 -> 79.03 | Improved |
| add adversarial examples to the reference set (ablation) | LNG · CIFAR-10 CWwb | AUC: 80.77 -> 84.38 | Improved |
| add adversarial examples to the reference set (ablation) | LNG · ImageNet CW | AUC: 81.24 -> 86.05 | Improved |
| add adversarial examples to the reference set (ablation) | LNG · CIFAR-10 CW | AUC: 89.91 -> 89.74 | No clear effect |
| estimate edges between neighborhood nodes (LNG instead of k-NNG) (ablation) | graph discriminator · CIFAR-10 PGD | AUC: 85.78 -> 91.39 | Improved |
| estimate edges between neighborhood nodes (LNG instead of k-NNG) (ablation) | graph discriminator · ImageNet CW | AUC: 79.03 -> 86.05 | Improved |
| estimate edges between neighborhood nodes (LNG instead of NC) (ablation) | graph discriminator · CIFAR-10 CWwb | AUC: 81.60 -> 84.38 | Improved |
| connect all nodes (AC) instead of LNG (ablation) | graph discriminator · CIFAR-10 CWwb | AUC: 84.38 -> 80.43 | Hurt |
| LNG instead of center-connected graph (CC) (ablation) | graph discriminator · CIFAR-10 CW | AUC: 91.51 -> 89.74 | Hurt |
| increase neighbors k from 2 to 5 (ablation) | k-NNG · CIFAR-10 FGSM | AUC: 96.39 -> 99.54 | Improved |
Closest papers by meaning
- Beating Attackers At Their Own Games: Adversarial Example Detection Using Adversarial Gradient Directions (AAAI 2021 · similarity 0.88)
- Adversarial Training for Graph Neural Networks: Pitfalls, Solutions, and New Directions (NeurIPS 2023 · similarity 0.88)
- Be Your Own Neighborhood: Detecting Adversarial Examples by the Neighborhood Relations Built on Self-Supervised Learning (ICML 2024 · similarity 0.88)
- ML-LOO: Detecting Adversarial Examples with Feature Attribution (AAAI 2020 · similarity 0.86)
- GOLD: Graph Out-of-Distribution Detection via Implicit Adversarial Latent Generation (ICLR 2025 · similarity 0.85)
- Adversarial Robust Generalization of Graph Neural Networks (ICML 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
