Project for the course Research Methodologies and Scientific Writing course at @KTH (EIT Data Science Master Programme)
The aim of the project was to develop an intrusion detection system using Deep Learning Algorithms. The main implementation is done using Tensorflow following the feature work of another paper.
The proposed solution is a Deep Neural Network composed by a Stacked Denoising Autoencoder in the first layers to extract the most meaningful features followed by a Simple Softmax Regression for classification purposes.
The repository contains the presentation and the paper written for the project regarding the implementation and the results.
For training and see the hold out results
$ python3 main.py
For training and see the results on the testset
$ python3 main.py --test=True