Machine learning in CyberSecurity

Projects for ML in CyberSecurity course Saarland University WS 2018/19

Repository Contains the following items:

1. Project_1
2. Project_2
3. Project_3 with report

Project_1:

A digit classifier is implemented based on the popular MNIST dataset. The dataset is based on a seminal paper, which immensely popularized (convolutional) neural networks. This is a great starting point for ML research and this dataset/model has been a stepping stone for numerous other tasks such as GANs, Adversarial Perturbations

Project_2:

Contains an implementation of Evasion attack (craft adversarial examples) and also defence of the model

Proejct_3:

The goal of this project (and the competition) is: given a dataset based on varied properties of different Windows machines, to be able to predict a particular machine's probability of getting infected by malware.