DGA GAN——GAN在安全中的应用

DGA的模型：https://github.com/Yuren-Zhong/DeepDGA CNN、LSTM、双向LSTM 论文可以看https://openreview.net/pdf?id=BJLmN8xRW

DGA GAN的论文：https://arxiv.org/pdf/1610.01969.pdf 

国内的一些实验demo http://www.itboth.com/d/fuE77fIBNvAn 使用的是LSTM，但是效果不是很好

有时间自己可以写代码实验下。

 

DGA的数据：

Dataset

Training on the dataset built by https://github.com/andrewaeva/DGA, including 1,000,000 legit domains and 801,667 DGA generated domains. While calling dataset.load_data, if specify filter=True then legal domains that end with different suffixes than DGA generated domains are not loaded, and also suffixes of all loaded domains are removed.

Evaluating on the dataset provided by http://data.netlab.360.com/feeds/dga/dga.txt, https://github.com/philarkwright/DGA-Detection, https://github.com/nickwallen/botnet-dga-classifier and https://github.com/ClickSecurity/data_hacking .