ECCV2020—对抗防御Paper List

1. Regional Homogeneity: Towards Learning Transferable Universal Adversarial Perturbations Against Defenses
2. Adversarial Ranking Attack and Defense
3. Segmentations-Leak: Membership Inference Attacks and Defenses in Semantic Image Segmentation
4. Anti-Bandit Neural Architecture Search for Model Defense
5. Open-set Adversarial Defense
6. Defense Against Adversarial Attacks via Controlling Gradient Leaking on Embedded Manifolds
7. Manifold Projection for Adversarial Defense on Face Recognition
8. 如有遗漏，请及时提出…