Hardware Trojan Detection Using Supervised Learning Techniques

Supervised Machine Learning (ML) technique is one of the many ways to detect the presence of Hardware Trojan (HT) in an IC. In this paper, four different classifier namely Logistic Regression, SVM, Random Forest and K-Nearest Neighbour have been used with ensemble learning technique AdaBoost. The used dataset in this experiment was collected from CAS lab, consisting of 18 types of ASIC benchmark circuits (14 synchronous and 4 asynchronous). In the result it has been seen that the Random Forest classifier is best with a classification accuracy of $\mathbf{9 9. 4 2 \%}$.