INTRUSION DETECTION USING DEEP NEURAL NETWORK

Abstract

Intrusion detection is very attractive topic for both system administrators and security researchers. The problem of intrusion detection can be tackled by machine learning models, based on statistical algorithms or artificial neural networks, to identify abnormal behaviours from those of users accessing systems. The recent development of machine learning techniques and increasing computational power of graphical processing units contribute significantly to the wide spread of the deep learning technique. This report investigates the application of deep neural network to the problem of intrusion detection and compares with typical machine
learning techniques based on NSL-KDD dataset.