IBOK

Carlini, Nicholas, and David Wagner. "Magnet and" efficient defenses against adversarial attacks" are not robust to adversarial examples." arXiv preprint arXiv:1711.08478 (2017). - 2017년 Carlini와 Wagner이 제안한 MagNet과 "Efficient Defenses~"의 반박 논문.- 1) Meng et al.의 MagNet (2017), 2) Zantedeschi et al.의 "Efficient Defenses Against Adversarial Attacks" (2017), 3) Shen et al.의 APE-GAN (2017) 3가지의 adve..

Cho, Seung Ju, et al. "DAPAS: Denoising Autoencoder to Prevent Adversarial attack in Semantic Segmentation." arXiv preprint arXiv:1908.05195 (2019). Abstract - Denoise autoencoder을 이용하여 semantic segmentation에 대한 adversarial defense 기법을 제안 Introduction - image를 pixel level에서 재구성하여 깨끗한 image를 만든다. - 1) Gaussian distribution, 2) Uniform distribution, 3) Bimodal distribution을 noise로 사용. - Dataset : ..

Meng, Dongyu, and Hao Chen. "Magnet: a two-pronged defense against adversarial examples." Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security. 2017. 이미지 출처 : https://www.youtube.com/watch?v=wZ-wIdAcWQE 2017년에 제안된, autoencoder을 이용한 adversarial defense 방법. 개요 - Adversarial attack에 대한 defense 방법. - Autoencoder을 이용하여 "sanitize" 한 input을 classifier에 넣는다. 장점 - target ..