Hardware Trojan Detection using Graph Neural Networks

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 around world. 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 functionality intended design. Various HT detection methods have been proposed in literature; however, many fall short due their reliance on golden reference circuit, limited scope, need manual code review, or inability scale with large modern designs. We propose novel reference-free method both Register Transfer Level (RTL) gate-level netlists by leveraging Graph Neural Networks (GNNs) learn behavior circuit through Data Flow (DFG) representation hardware evaluate our model custom dataset expanding Trusthub benchmarks trusthub1. results demonstrate that approach detects unknown 97% recall (true positive rate) very fast 21.1ms RTL 84% 13.42s Gate-Level Netlist.