Q-RAIL: A Reliability-Aware Framework for Quantum Federated Learning on Heterogeneous Noisy Hardware
Q-RAIL addresses unreliable updates in Quantum Federated Learning (QFL) due to client hardware differences. It introduces a circuit- and calibration-aware method to calculate client-specific effective noise budgets from backend metadata and circuit statistics. This allows the framework to generate stabilized aggregation weights, leading to significantly improved performance over standard methods under hardware skew.