Hardware-Trojan Detection at Gate-level Netlists using Gradient Boosting Decision Tree Models
Ryotaro Negishi, Tatsuki Kurihara, Nozomu Togawa
Abstract
Technological devices including consumer devices have become deeply embedded in people's lives, and their demand is growing every year. It has been indicated that outsourcing the design and manufacturing of ICs, which are essential for tech-nological devices, may lead to the insertion of hardware Trojans. This paper proposes a hardware-Trojan detection method at gate-level netlists based on the gradient boosting decision tree models. We firstly propose the optimal set of Trojan features among many feature candidates at a netlist level through thorough evaluations. Then, we evaluate various gradient boosting decision tree models and determine XGBoost is the best for hardware-Trojan detection. Finally, we construct an XGBoost-based hardware-Trojan detection method with its optimized hyperparameters. Evaluation experiments were conducted on the netlists from Trust-HUB benchmarks and showed the average F-measure of 0.842 using the proposed method. This value is 0.175 points higher than that of the existing best method.