Research map / Adversarial robustness and security / Attacks on text and graphs
Detection Based Defense Against Adversarial Examples From the Steganalysis Point of View
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 |
|---|---|---|---|
| Add MPM weighting to SRM features (ablation) | SRM detector · ImageNet-1000, FGSM ε=8 adversarial images | detection accuracy: 0.9903 -> 0.9938 | Helped |
| Add MPM weighting to SRM features (ablation) | SRM detector · ImageNet-1000, IGSM ε=8 adversarial images | detection accuracy: 0.9812 -> 0.9879 | Helped |
| Add MPM weighting to SRM features (ablation) | SRM detector · ImageNet-1000, Deepfool adversarial images | detection accuracy: 0.9491 -> 0.9527 | Helped |
| Add MPM weighting to SRM features (ablation) | SRM detector · ImageNet-1000, C&W adversarial images | detection accuracy: 0.9092 -> 0.9341 | Helped |
| Add MPM weighting to SPAM features (ablation) | SPAM detector · ImageNet-1000, FGSM ε=8 adversarial images | detection accuracy: 0.9709 -> 0.9806 | Helped |
| Add MPM weighting to SPAM features (ablation) | SPAM detector · ImageNet-1000, IGSM ε=8 adversarial images | detection accuracy: 0.9601 -> 0.9745 | Helped |
| Add MPM weighting to SPAM features (ablation) | SPAM detector · ImageNet-1000, Deepfool adversarial images | detection accuracy: 0.8481 -> 0.8629 | Helped |
| Add MPM weighting to SPAM features (ablation) | SPAM detector · ImageNet-1000, C&W adversarial images | detection accuracy: 0.6778 -> 0.8296 | Helped |
| Use SRM rather than SPAM features with MPM (ablation) | Enhanced steganalysis detector · ImageNet-1000, C&W adversarial images | detection accuracy: 0.8296 -> 0.9341 | Helped |
Closest papers by meaning
- Beating Attackers At Their Own Games: Adversarial Example Detection Using Adversarial Gradient Directions (AAAI 2021 · similarity 0.90)
- One Man's Trash Is Another Man's Treasure: Resisting Adversarial Examples by Adversarial Examples (CVPR 2020 · similarity 0.89)
- Are adversarial examples inevitable? (ICLR 2019 · similarity 0.88)
- A Direct Approach to Robust Deep Learning Using Adversarial Networks (ICLR 2019 · similarity 0.88)
- Detecting Adversarial Examples Through Image Transformation (AAAI 2018 · similarity 0.88)
- Evading Defenses to Transferable Adversarial Examples by Translation-Invariant Attacks (CVPR 2019 · 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.