Research map / Adversarial robustness and security / Attacks on text and graphs
Privacy Leaks by Adversaries: Adversarial Iterations for Membership Inference Attack
Jing Xue, Zhishen Sun, Haishan Ye, Luo Luo, Xiangyu Chang, Guang Dai
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 |
|---|---|---|---|
| use number of SimBA iterations as membership score instead of confidence metrics | IMIA · CIFAR10 DenseNet score-based | AUROC / accuracy: AUROC 69.21 -> 75.20; accuracy 64.17 -> 68.97 | Improved |
| use iteration count instead of Modified Entropy | IMIA · CIFAR100 ResNet score-based | accuracy (%): 85.43 -> 83.62 | Hurt |
| use iteration count instead of Modified Entropy | IMIA · CIFAR10 ResNeXt score-based | accuracy (%): 67.95 -> 67.77 | No clear effect |
| use HSJA iteration count instead of boundary distance | IMIA · CIFAR10 VGG label-only | accuracy (%): 63.0 -> 65.10 | Improved |
| use HSJA iteration count instead of boundary distance | IMIA · STL10 label-only, all four models | AUROC and accuracy: DenseNet accuracy 75.29 -> 76.57 | Improved |
| use iteration count instead of boundary distance | IMIA · CIFAR100 label-only ResNet | accuracy (%): 81.59 -> 81.39 | No clear effect |
| use PGD iteration count instead of cross-entropy loss | IMIA · CIFAR100 white-box ResNet | accuracy (%): 85.68 -> 90.54 | Improved |
| use PGD iteration count instead of cross-entropy loss | IMIA · CIFAR10 white-box DenseNet | accuracy (%): 64.79 -> 70.15 | Improved |
| use PGD iteration count instead of loss | IMIA · STL10 white-box ResNeXt and VGG | accuracy (%): 66.39 -> 66.19 (ResNeXt) | No clear effect |
Closest papers by meaning
- Assessing Per-Sample Membership Inference Vulnerability without Retraining (NeurIPS 2026 · similarity 0.90)
- Causal Evaluation of Membership Inference Attacks (NeurIPS 2026 · similarity 0.89)
- Low-Cost High-Power Membership Inference Attacks (ICML 2024 · similarity 0.89)
- Chameleon: Increasing Label-Only Membership Leakage with Adaptive Poisoning (ICLR 2024 · similarity 0.88)
- Estimating Model-Level Membership Inference Vulnerability Without Reference Models (NeurIPS 2026 · similarity 0.88)
- Membership Inference Attacks against Fine-tuned Large Language Models via Self-prompt Calibration (NeurIPS 2024 · similarity 0.88)
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.
