QUFIG: GNN-Based Prediction of Quantum Fault Injection Vulnerabilities with Gate-Level Precision
The QuFIG framework uses a circuit-DAG-based Graph Neural Network (GNN) to predict the vulnerability score of every gate in a quantum circuit against different types of injected faults. It formulates fault analysis as a learning problem to rank gate-fault pairs by their predicted impact on circuit fidelity, allowing users to prioritize which gates need detailed inspection.