The Limitations of Deep Learning in Adversarial Settings
概
利用Jacobian矩阵构造adversarial samples,计算量比较大.
主要内容
目标:
\mathop{\arg \min} \limits_{\delta_X} \|\delta_X\|, \mathbf{s.t.} \: F(X+\delta_X)=Y^*.
\]
简而言之, 在原图像\(X\)上加一个扰动\(\delta_X\), 使得\(F\)关于\(X+\delta_X\)的预测为\(Y^*\)而非\(Y\).
若\(Y \in \mathbb{R}^M\)是一个\(M\)维的向量, 类别由下式确定
\]
\(F(X)=Y\)关于\(X\)的Jacobian矩阵为
\]
注意, 这里作者把\(X\)看成一个\(N\)维向量(只是为了便于理解).
因为我们的目的是添加扰动\(\delta_X\), 使得\(X+\delta_X\)的标签为我们指定的\(t\), 即我们希望
\]
作者希望改动部分元素, 即\(\|\delta_X\|_0\le \Upsilon\), 作者是构造了一个saliency_map来选择合适的\(i\), 并在其上进行改动, 具体算法如下:

saliency_map的构造之一是:
\begin{array}{ll}
0, & if \: \frac{\partial{F_t(X)}}{\partial X_i} <0 \:or \: \sum_{j \not= t} \frac{\partial F_j(X)}{\partial X_i} >0, \\
\frac{\partial{F_t(X)}}{\partial X_i} |\sum_{j \not= t} \frac{\partial F_j(X)}{\partial X_i}|, & otherwise.
\end{array}
\]
可以很直观的去理解, 改变标签, 自然希望\(F_t(X)\)增大, 其余部分减少, 故 \(\frac{\partial{F_t(X)}}{\partial X_i} <0 \:or \: \sum_{j \not= t} \frac{\partial F_j(X)}{\partial X_i} >0\)所对应的\(X_i\)自然是不重要的, 其余的是重要的, 其重要性用\(\frac{\partial{F_t(X)}}{\partial X_i} |\sum_{j \not= t} \frac{\partial F_j(X)}{\partial X_i}|\)来表示.
alg2, alg3
作者顺便提出了一个更加具体的算法, 应用于Mnist, max_iter 中的\(784\)即为图片的大小\(28 \times 28\), \(\Upsilon=50\), 相当于图片中\(50\%\)的像素发生了改变, 且这里采用了一种新的saliency_map, 其实质为寻找俩个指标\(p,q\)使得:

其实际的操作流程根据算法3. \(\theta\)是每次改变元素的量.


一些有趣的实验指标
Hardness measure


其中\(\epsilon(s,t,\tau)\)中, \(s\):图片标签, \(t\):目标标签, \(\tau\):成功率, \(\epsilon\)为改变像素点的比例. (12)是(11)的一个梯形估计, \(\tau_k\)由选取不同的\(\Upsilon_k\)来确定, \(H(s, t)\)越大说明将类别s改变为t的难度越大.
Adversarial distance

\(A(X,t)\)越大, 说明将图片\(X\)的标签变换至\(t\)的难度越大, 而一个模型的稳定性可以用下式衡量
R(F)=\min_{X,t} A(X,t).
\]
The Limitations of Deep Learning in Adversarial Settings的更多相关文章
- What are some good books/papers for learning deep learning?
What's the most effective way to get started with deep learning? 29 Answers Yoshua Bengio, ...
- Applied Deep Learning Resources
Applied Deep Learning Resources A collection of research articles, blog posts, slides and code snipp ...
- (转)Deep Learning Research Review Week 1: Generative Adversarial Nets
Adit Deshpande CS Undergrad at UCLA ('19) Blog About Resume Deep Learning Research Review Week 1: Ge ...
- 论文笔记之:UNSUPERVISED REPRESENTATION LEARNING WITH DEEP CONVOLUTIONAL GENERATIVE ADVERSARIAL NETWORKS
UNSUPERVISED REPRESENTATION LEARNING WITH DEEP CONVOLUTIONAL GENERATIVE ADVERSARIAL NETWORKS ICLR 2 ...
- Towards Deep Learning Models Resistant to Adversarial Attacks
目录 概 主要内容 Note Madry A, Makelov A, Schmidt L, et al. Towards Deep Learning Models Resistant to Adver ...
- (转) The major advancements in Deep Learning in 2016
The major advancements in Deep Learning in 2016 Pablo Tue, Dec 6, 2016 in MACHINE LEARNING DEEP LEAR ...
- 博弈论揭示了深度学习的未来(译自:Game Theory Reveals the Future of Deep Learning)
Game Theory Reveals the Future of Deep Learning Carlos E. Perez Deep Learning Patterns, Methodology ...
- [C3] Andrew Ng - Neural Networks and Deep Learning
About this Course If you want to break into cutting-edge AI, this course will help you do so. Deep l ...
- 0.读书笔记之The major advancements in Deep Learning in 2016
The major advancements in Deep Learning in 2016 地址:https://tryolabs.com/blog/2016/12/06/major-advanc ...
随机推荐
- 【Python】【Basic】MacOS上搭建Python开发环境
1. Python3 1.1. 下载地址:https://www.python.org/downloads/mac-osx/ 1.1.1. PKG包安装: 没啥可说的,点点点,下一步而已,不用手动配置 ...
- When should we write our own assignment operator in C++?
The answer is same as Copy Constructor. If a class doesn't contain pointers, then there is no need t ...
- 【Linux】【Services】【SaaS】Docker+kubernetes(7. 安装Docker私有镜像仓库)
1. 简介 1.1. 自己做个私有镜像,方便上传和下载,我也在docker官网注册了一个账号,做好的镜像可以传上去 1.2. Redhat自带私有镜像的功能,需要安装包,这是howto: https: ...
- 【Linux】【Services】【SaaS】Docker+kubernetes(2. 配置NTP服务chrony)
1. 简介 1.1. 这次使用另外一个轻量级的NTP服务,chrony.这是openstack推荐使用的ntp服务. 1.2. 官方网站:https://chrony.tuxfamily.org/ 2 ...
- 【Spring Framework】spring管理自己new的对象
使用AutowireCapableBeanFactory手动注入 使用.newInstance();创建对象的话,如果其他对象都使用Spring Autowired,还需要手动创建所有依赖的Bean: ...
- shiro免认证的路径配置
– ?:匹配一个字符,如/admin? 将匹配/admin1,但不匹配/admin 或/admin/:– *:匹配零个或多个字符串,如/admin 将匹配/admin./admin123,但不匹配/a ...
- 解决tensorflow和keras版本不相匹配的问题
查看安装版本 pip list https://docs.floydhub.com/guides/environments/ 查看对应版本 我感觉是我tensorflow版本装太高了,keras没有
- 【C/C++】PAT A1025 Ranking/算法笔记
题目意思大概是输入一堆人的学号,成绩,给出学号,总排名,考场号,考场内排名. 这是我第一次写的: #include <iostream> #include <algorithm> ...
- Redis集群断电恢复
再集群整体断点或关闭后,默认启动集群后,会成为孤立的单点,需要删除每个节点的pid文件,node.conf.并将RDB和AOF文件移动出来,再挨个启动每个节点,并用create创建集群脚本,重新创建集 ...
- gitlab 集成openldap
Setting up LDAP sign-in If you have an LDAP directory service such as Active Directory, you can conf ...