Research map / Adversarial robustness and security / Black-box and transfer attacks
Sensitive-Sample Fingerprinting of Deep Neural Networks
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Fingerprint the model with three Sensitive-Samples selected by MANC | Sensitive-Sample fingerprinting · Neural network trojan attack | missing rate: 5.93 -> 0.00 | Helped |
| Fingerprint the model with three Sensitive-Samples selected by MANC | Sensitive-Sample fingerprinting · Error-generic poisoning attack | missing rate: 12.26 -> 0.01 | Helped |
| Fingerprint the model with three Sensitive-Samples selected by MANC | Sensitive-Sample fingerprinting · Error-specific poisoning attack | missing rate: 2.20 -> 0.00 | Helped |
| Fingerprint the model with eight Sensitive-Samples selected by MANC | Sensitive-Sample fingerprinting · Model compression attack | missing rate: 48.93 -> 0.04 | Helped |
| Provide top-3 or top-5 labels rather than only top-1 (ablation) | Sensitive-Sample fingerprinting · Neural network trojan attack | missing rate: 5.93 -> 0.00 | Helped |
| Include class probabilities in the model output (ablation) | Sensitive-Sample fingerprinting · Neural network trojan attack | missing rate: 5.93 -> 0.21 | Helped |
| Fine-tune the compromised model on Sensitive-Samples (ablation) | Sensitive-Sample fingerprinting · Error-generic poisoning attack | missing rate: 12.26 -> 4.82 | Helped |
| Fine-tune the compromised model on Sensitive-Samples (ablation) | Sensitive-Sample fingerprinting · Error-specific poisoning attack | missing rate: 2.20 -> 0.02 | Helped |
| Use human-unnoticeable Sensitive-Samples instead of ordinary validation samples | Model integrity verification · Evaluated integrity attacks | missing rate | Helped |
| Add MANC sample selection rather than randomly selecting Sensitive-Samples (ablation) | Sensitive-Sample fingerprinting · Evaluated integrity attacks | missing rate | Helped |
Closest papers by meaning
- SDBF: Steep-Decision-Boundary Fingerprinting for Hard-Label Tampering Detection of DNN Models (CVPR 2025 · similarity 0.89)
- Intersecting-Boundary-Sensitive Fingerprinting for Tampering Detection of DNN Models (ICML 2024 · similarity 0.89)
- Queries, Representation & Detection: The Next 100 Model Fingerprinting Schemes (AAAI 2025 · similarity 0.86)
- United We Stand, Divided We Fall: Fingerprinting Deep Neural Networks via Adversarial Trajectories (NeurIPS 2024 · similarity 0.86)
- Assessing Per-Sample Membership Inference Vulnerability without Retraining (NeurIPS 2026 · similarity 0.85)
- Beyond Slow Signs in High-fidelity Model Extraction (NeurIPS 2024 · 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.