Multiple-shot discrimination of quantum unitary channels and its implementation

Abstract

This thesis investigates multiple-shot discrimination of unitary quantum channels on NISQ hardware, comparing parallel, sequential, and hybrid strategies. On the theoretical side, we introduce a hybrid geometric solver for constructing optimal input states for perfect discrimination tasks. Experimentally, we observe that theoretically suboptimal circuits can outperform optimal ones on noisy quantum devices. Results obtained on 6- and 11-qubit systems with M3 error mitigation highlight the importance of hardware-aware transpilation, with high-depth, low-width circuits proving more robust than highly entangled parallel schemes. Finally, we report an unexpected anomaly on IBM hardware when evaluating purely parallel strategies.