QUFIG: GNN-Based Prediction of Quantum Fault Injection Vulnerabilities with Gate-Level Precision
QUFIG uses a circuit-DAG GNN to rank quantum gates by fault-injection impact under limited fidelity budgets.
QUFIG treats gate-level quantum fault analysis as learning-guided prioritization when fidelity inspection budgets are limited. A graph neural network on the circuit DAG predicts each gate's vulnerability to each injected fault type, measured by impact on circuit fidelity, then ranks gate-fault pairs. On QASMbench and HamLib MaxCut, it recovers high-impact pairs with fewer inspections than random and depth-based heuristics. It cuts gates requiring inspection by 2.9-19.8% while still identifying effective faults.
- Predicts per-gate vulnerability scores for each fault type.
- Ranks gate-fault pairs under a restricted fidelity budget.
- Outperforms random and depth heuristics on QASMbench and HamLib.
- Reduces gates inspected by 2.9-19.8 percent.
From the original article
1 min read at arxiv.orgThe growing scale and accessibility of quantum hardware exposed new reliability and security challenges in the quantum computing workflow, such as the run-time fault injection attacks in cloud-based quantum computing platforms. However, existing works fail to identify vulnerabilities with gate-level precision or adapt to run-time environments.
By Shihan Zhao · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.