Structural and SCOAP Features Based Approach for Hardware Trojan Detection Using SHAP and Light Gradient Boosting Model

Hardware Trojan (HT) is the most critical threat due to outsourcing of Integrated circuit designing phases. Existing machine learning-based HT detection techniques at the pre-silicon IC phase use the structural or SCOAP gate-level netlist features for detection. However, these techniques either fail to detect the always-on-Trojans or low SCOAP Trojans, thus provides large false positives/negatives. Moreover, they fail to interpret the model prediction locally due to model-specific feature importances, identify the best feature subset in large retraining rounds, and also drop some relevant features. Therefore, to tackle these limitations, this paper proposes a new technique that utilizes structural and SCOAP features to detect HT from the gate-level netlist. The proposed technique utilizes the fastest model Light Gradient Boosting that uses gradient-based one-side sampling and exclusive feature bundling to reduce the computational time. Further, a model agnostic Shapley additive explanations (SHAP) is employed to identify each feature global and local impact on model prediction, thus making the prediction transparent. Moreover, a quartile-based feature selection method is proposed, which uses SHAP to identify the optimal feature set by keeping low retraining rounds. Experimental results show that the proposed technique accurately detects always-on-Trojans and HT nets from Trust-Hub, DeTrust, DeTest and MIMIC based Trojan benchmarks.