Golden Reference-Free Hardware Trojan Localization Using Graph Convolutional Network

The globalization of the integrated circuit (IC) supply chain has moved most design, fabrication, and testing process from a single trusted entity to various untrusted third-party entities worldwide. risk using third-Party Intellectual Property (3PIP) is possibility for adversaries insert malicious modifications known as Hardware Trojans (HTs). These HTs can compromise integrity, deteriorate performance, deny service, alter functionality design. While numerous HT detection methods have been proposed in literature, crucial task localization overlooked. Moreover, few existing several weaknesses: reliance on golden reference, inability generalize all types HT, lack scalability, low resolution, manual feature engineering/property definition. To overcome their shortcomings, we propose novel, reference-free method at pre-silicon stage by leveraging graph convolutional network (GCN). In this work, convert design into its intrinsic data structure, graph, extract node attributes. Afterward, convolution performs automatic extraction nodes classify Trojan or benign. Our approach automated does not burden designer with code review. It locates signals 99.6% accuracy, 93.1% $F1$ -score, false-positive rate below 0.009%.