A Novel Intrusion Detection Scheme Using Variational Autoencoders

Security is a major challenge in Internet-of-Things (IoT) systems, and network intrusion detection systems (NIDS) play a key role in proposed solutions. In this paper, we propose VarSec which is a centralised unsupervised algorithm for network anomaly detection. It uses two variational autoencoders (VAEs) to process packet-level data, and combines their output to detect anomalous packets. We evaluate the performance of our proposed algorithm on a variety of attack datasets and compare it with existing solutions.