GNN4IP: Graph Neural Network for Hardware Intellectual Property Piracy Detection
This addresses IP theft in the semiconductor industry, offering a detection method with high accuracy, though it is incremental as it builds on existing graph-based and neural network approaches.
The authors tackled the problem of hardware intellectual property piracy detection by proposing GNN4IP, a graph neural network method that models circuits as graphs, achieving 96% accuracy in detecting piracy and 100% accuracy in recognizing original IP in obfuscated versions.
Aggressive time-to-market constraints and enormous hardware design and fabrication costs have pushed the semiconductor industry toward hardware Intellectual Properties (IP) core design. However, the globalization of the integrated circuits (IC) supply chain exposes IP providers to theft and illegal redistribution of IPs. Watermarking and fingerprinting are proposed to detect IP piracy. Nevertheless, they come with additional hardware overhead and cannot guarantee IP security as advanced attacks are reported to remove the watermark, forge, or bypass it. In this work, we propose a novel methodology, GNN4IP, to assess similarities between circuits and detect IP piracy. We model the hardware design as a graph and construct a graph neural network model to learn its behavior using the comprehensive dataset of register transfer level codes and gate-level netlists that we have gathered. GNN4IP detects IP piracy with 96% accuracy in our dataset and recognizes the original IP in its obfuscated version with 100% accuracy.