A Blockchain Data Balance Using a Generative Adversarial Network Approach: Application to Smart House IDS

Abstract

The rapid development of information and communication technologies makes the Internet of Things (IoT) devices much more complex and heterogeneous. In this context, the massive end devices (IoTs) and the large volume of data raise security and privacy challenges. To tackle these issues, the joint use of the Bockchain (BC) and Machine Learning (ML) seems attractive to achieve decentralized, secure, intelligent and efficient management of networks. On the one hand, the BC can greatly facilitate the sharing of training data and ML models, the decentralization of intelligence, security, privacy and reliable ML decision-making. On the other hand, ML may have significant impacts on the development of BC in communications and networking systems, including energy and resource efficiency, scalability, security, privacy and smart contracting. An important aspect of security intends to detect unusual and potentially inappropriate activities according to traffic patterns. This paper focuses on the problem of imbalance data where the number of abnormal samples is significantly lower than that of the normal (secure) ones. In particular, this paper presents a new equilibrium model based on an exciting recent innovation in ML namely Generator Adverse Networks (GANs) to address the problem of class imbalance and data noise to Intrusion Detection System (IDS) performance. The proposed approach use is illustrated by a case study: a smart house system-based scenario.