Research map / Adversarial robustness and security / Certified and empirical robustness
Characterizing Adversarial Subspaces Using Local Intrinsic Dimensionality
Read the paperCheck 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 |
|---|---|---|---|
| Use LID features instead of KD+BU features | Adversarial-example detector · MNIST Opt | AUC (%): 95.35 -> 99.24 | Helped |
| Use LID features instead of KD+BU features | Adversarial-example detector · CIFAR-10 Opt | AUC (%): 93.77 -> 98.94 | Helped |
| Use LID features instead of KD+BU features | Adversarial-example detector · SVHN Opt | AUC (%): 90.66 -> 97.60 | Helped |
| Train the LID detector on FGM examples and test on other attack strategies | LID detector · CIFAR-10 BIM-a, BIM-b, JSMA, and Opt | AUC (%): 82.30; 91.61; 89.93; 93.32 | Helped |
| Increase LID-estimation minibatch size from 100 to 1000 | LID-based detector · Opt attacks on MNIST, CIFAR-10, and SVHN | AUC | Helped |
| Add a LID term to the adaptive Opt attack objective | LID-based detector · MNIST adaptive Opt, Scenario 1 | Attack failure rate (%): 100 | No clear effect |
| Add a LID term to the adaptive Opt attack objective | LID-based detector · CIFAR-10 adaptive Opt, Scenario 2 | Attack failure rate (%): 95.7 | No clear effect |
Closest papers by meaning
- Decision Boundary Analysis of Adversarial Examples (ICLR 2018 · similarity 0.89)
- Explaining the role of Intrinsic Dimensionality in Adversarial Training (ICML 2025 · similarity 0.87)
- Understanding Adversarial Examples From the Mutual Influence of Images and Perturbations (CVPR 2020 · similarity 0.87)
- Be Your Own Neighborhood: Detecting Adversarial Examples by the Neighborhood Relations Built on Self-Supervised Learning (ICML 2024 · similarity 0.87)
- Contrasting Adversarial Perturbations: The Space of Harmless Perturbations (AAAI 2025 · similarity 0.86)
- Adversarial Training Can Provably Improve Robustness: Theoretical Analysis of Feature Learning Process Under Structured Data (ICLR 2025 · similarity 0.86)
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.