论文阅读笔记:Threat of Adversarial Attacks on Deep Learning in Computer Vision: A Survey
2018-03-14 08:32
2531 查看
论文阅读笔记:Threat of Adversarial Attacks on Deep Learning in Computer Vision: A Survey
要点
(universal adversarial perturbation)[1]3D物理对抗样本
对抗样本生成
Box-constrained L-BFGSminρminρ
Fast Gradient Sign Method (FGSM)
ρ=ϵsign(∇J(θ,Ic,l))ρ=ϵsign(∇J(θ,Ic,l))
one-shot 生成方式要求ϵϵ不能太小
Basic Iterative Method (BIM)
Ik+1=clip(Ik+αsign(∇J(θ,Ik,l))Ik+1=clip(Ik+αsign(∇J(θ,Ik,l))
Iterative Least-likely Class Method (ILCM)
选取测试概率最低的那个类作为target class进行BIM targeted adversarial example generation
Jacobian-based Saliency Map Attack (JSMA)
One Pixel Attack
Carlini and Wagner Attacks (C&W)
DeepFool
Universal Adversarial Perturbations
对几乎所有输入均有效、不依赖于输入的对抗扰动
UPSET(Universal Perturbations for Steering to Exact Targets)
ANGRI(Antagonistic Network for Generating Rogue Images)
对抗训练
参考文献
[1]相关文章推荐
- 【论文阅读笔记】Deep Learning in Medical Imaging: Overview and Future Promise of an Exciting New Technique
- 论文笔记 Ensemble of Deep Convolutional Neural Networks for Learning to Detect Retinal Vessels in Fundus
- 1705.Person Re-Identification by Deep Joint Learning of Multi-Loss Classification 论文阅读笔记
- 1705.Person Re-Identification by Deep Joint Learning of Multi-Loss Classification 论文阅读笔记
- [深度学习论文笔记][Weight Initialization] Exact solutions to the nonlinear dynamics of learning in deep lin
- 【论文阅读笔记】CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning
- 【论文阅读笔记】Deep Learning based Recommender System: A Survey and New Perspectives
- The development and prosperous of deep learning theory applying in computer vision(Image part)
- 关于The Limitations of Deep Learning in Adversarial Settings的理解
- 论文阅读:End-to-End Learning of Deformable Mixture of Parts and Deep Convolutional Neural Networks for H
- 生成对抗网络学习笔记3----论文unsupervised representation learning with deep convolutional generative adversarial
- 笔记 NEURAL NETS FOR VISION CVPR 2012 Tutorial on Deep Learning
- 论文笔记:A survey on vision-based human action recognition.
- [论文阅读笔记] DeepNeuron: An Open Deep Learning Toolbox for Neuron Tracing
- Feed Forward and Backward Run in Deep Convolution Neural Network 论文阅读笔记
- 论文笔记 A Large Contextual Dataset for Classification,Detection and Counting of Cars with Deep Learning
- 论文笔记:A Review on Deep Learning Techniques Applied to Semantic Segmentation
- Deep Learning in Computer Vision
- cuDNN: efficient Primitives for Deep Learning 论文阅读笔记
- 李菲菲课程笔记:Deep Learning for Computer Vision – Introduction to Convolution Neural Networks