Research map / Adversarial robustness and security / Certified and empirical robustness
Sensitivity Analysis of Deep Neural Networks
Hai Shu, Hongtu Zhu
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 the proposed FI measure instead of the Jacobian norm for simulated-outlier detection | MNIST classifiers · MNIST with simulated outliers | precision-recall curve performance | Improved |
| Use the proposed FI measure instead of the Jacobian norm for simulated-outlier detection | MNIST classifiers · MNIST with simulated outliers | ROC curve performance | No clear effect |
| Choose DenseNet121 rather than ResNet50 | CIFAR10 classifiers · CIFAR10 | FI: 2.87e-3 -> 1.10e-4 (75th percentile, training); 0.031 -> 1.83e-3 (75th percentile, test) | Improved |
| Choose DenseNet121 rather than ResNet50 | CIFAR10 classifiers · CIFAR10 | test accuracy: 88.70% -> 91.16% | Improved |
| Choose DenseNet121 rather than ResNet50 | MNIST classifiers · MNIST | FI | Mixed |
| Use multi-scale pixel neighborhoods rather than only a single pixel scale | FI-based vulnerable-region detection · CIFAR10 and MNIST test images | vulnerable-region coverage | Mixed |
Closest papers by meaning
- Contrasting Adversarial Perturbations: The Space of Harmless Perturbations (AAAI 2025 · similarity 0.88)
- Understanding Deep Networks via Extremal Perturbations and Smooth Masks (ICCV 2019 · similarity 0.87)
- Robustness of Classifiers to Universal Perturbations: A Geometric Perspective (ICLR 2018 · similarity 0.87)
- Learning perturbation sets for robust machine learning (ICLR 2021 · similarity 0.86)
- On the Minimal Adversarial Perturbation for Deep Neural Networks With Provable Estimation Error (TPAMI 2022 · similarity 0.86)
- Outlier Gradient Analysis: Efficiently Identifying Detrimental Training Samples for Deep Learning Models (ICML 2025 · similarity 0.84)
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.
