IoT Security: A Comparative Analysis of Intrusion Detection Systems Based on Machine Learning, Deep Learning and Transfer Learning Techniques

Internet of Things (IoT) regroups a huge number of smart devices connected and ex-changed data between them and with other objects and systems via the internet. The explosive growth of internet-connected IoT devices engenders large amounts of data and then the attack surface is increased, thus it is important to implement suitable security techniques to assure the privacy and reliability and protect IoT data from security threats. To do this, there are several solutions in the literature; however, we have focused our research on solutions based on artificial intelligence and more precisely Machine and Deep learning. Among these solutions, we present Intrusion Detection Systems (IDS). This survey presents IoT security challenges. It highlights available datasets for IoT and provides a taxonomy of IDSs and gives comprehensive analysis of re-cent proposed IDS in the literature based on different techniques and designed specifically for IoT system.