Close

Presentation

GNN4IP: Graph Neural Network for Hardware Intellectual Property Piracy Detection
Time
Location
Event Type
Research Manuscript
Virtual Programs
Hosted in Virtual Platform
Keywords
Hardware Security: Primitives, Architecture, Design & Test
Topics
Security
DescriptionTime-to-market constraints and enormous circuit design costs have pushed the semiconductor industry toward reusable IP design. However, the IC supply chain's globalization exposes IP providers to theft and illegal redistribution. Current countermeasures come with additional hardware overhead and cannot guarantee IP security as advanced attacks are reported to bypass them. We propose a novel approach to detect IP piracy. We model the circuit as a graph and construct a graph neural network model to learn its behavior and reveal the piracy between designs. Our evaluation indicates 96% detection accuracy in our comprehensive dataset of RTL and gate-level netlist hardware designs.