ImmuneGAN: Bio-inspired Artificial Immune System to Secure IoT Ecosystem

The Internet of Things (IoT) has shown exponential growth in the flow of data in-between numerous low-energy based devices, this distributive and dynamic property of IoT network ecosystem possess a huge threat to data security. This paper proposes a layered artificial immune system approach, which is inspired by the natural immunity mechanism and adapts an architecture called ImmuneGAN, the proposed approach solves the low-data availability issue to train the machine learning model and adapts the implicit representation learning by using the min-max optimization process, the ImmuneGAN architecture tends to learn the underlying pattern distribution of the intrusion packets by simulating antibody population and perform memory acquisition task in order to provide enhanced response upon the subsequent interaction of attack with the environment, also the ImmuneGAN outperforms existing machine learning-based approaches and yields state of the art results to classify the attack packets with an accuracy of 99.87% and F1-score of 0.99874 on MQTTsetdataset.