Research map / Adversarial robustness and security / Certified and empirical robustness
Fingerprinting Deep Neural Networks for Ownership Protection: An Analytical Approach
Guang Yang, Ziye Geng, Yihang Chen, Changqing Luo
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 |
|---|---|---|---|
| place fingerprints at a stretch factor chosen inside the theoretical admissible interval, rather than fixed bounds or margins (ablation) | AnaFP · CIFAR-10 CNN | AUC | Improved |
| discard infeasible stretch factors (vs C-lower/C-upper fixed to bounds) (ablation) | AnaFP · CIFAR-10 CNN | AUC | Improved |
| use high-confidence anchors (margin >= 5.0) instead of mid-confidence (ablation) | AnaFP · CIFAR-10 CNN | AUC: 0.791 -> 0.958 | Improved |
| use high-confidence anchors (margin >= 5.0) instead of mid-confidence (ablation) | AnaFP · MNIST MLP | AUC: 0.755 -> 0.961 | Improved |
| moderately relax quantile thresholds from 0.3/0.7 to 0.5/0.5 (ablation) | AnaFP · CIFAR-10 CNN | AUC: 0.848 -> 0.957 | Improved |
| excessively relax quantile thresholds to 0.9/0.1 (ablation) | AnaFP · CIFAR-10 CNN | AUC: 0.959 -> 0.896 | Hurt |
| increase surrogate pool diversity from low to medium/high (ablation) | AnaFP · CIFAR-10 CNN | AUC | No clear effect |
| increase surrogate pool size beyond 6 models (ablation) | AnaFP · CIFAR-10 CNN | AUC | No clear effect |
| use AnaFP instead of baselines under KD attack | fingerprinting · CIFAR-100 CNN, KD | AUC: 0.596 vs 0.598 (UAP) | No clear effect |
Closest papers by meaning
- SDBF: Steep-Decision-Boundary Fingerprinting for Hard-Label Tampering Detection of DNN Models (CVPR 2025 · similarity 0.90)
- United We Stand, Divided We Fall: Fingerprinting Deep Neural Networks via Adversarial Trajectories (NeurIPS 2024 · similarity 0.89)
- Intersecting-Boundary-Sensitive Fingerprinting for Tampering Detection of DNN Models (ICML 2024 · similarity 0.89)
- SIF: Semantically In-Distribution Fingerprints for Large Vision-Language Models (CVPR 2026 · similarity 0.86)
- Fingerprinting Denoising Diffusion Probabilistic Models (CVPR 2025 · similarity 0.86)
- Scalable Fingerprinting of Large Language Models (NeurIPS 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.
