Introduction
Accurately characterizing quantum operations is a persistent bottleneck in building reliable quantum computers. Traditional Quantum Process Tomography, while thorough, scales exponentially with the number of qubits and gates, becoming impractical for modern devices. Researchers from Tohoku University, Nara Institute of Science and Technology, and the University of Information Technology (Vietnam National University, Ho Chi Minh City), led by Dr. Le Bin Ho, propose Compilation-Based Quantum Process Tomography or CQPT as a more efficient alternative.
Overcoming Quantum Measurement Obstacles
QPT reconstructs an unknown quantum process by preparing many input states and performing tomographically complete measurements. The resource cost explodes as systems grow: more qubits and deeper circuits demand exponentially more measurement settings and data. This makes routine characterization and frequent calibration infeasible for near-term and future quantum hardware.
How CQPT Works: A Smarter Approach
CQPT reframes characterization as a compilation problem. Instead of fully reconstructing the process, a trainable model attempts to revert the device’s noisy output back to the known input state. The optimization targets a single measurement outcome per input state, drastically reducing measurement overhead. Two formulations address different noise models: a Kraus-based version suited to general noisy channels and a Choi-based version that leverages process-state isomorphism when applicable. Both use classical optimization to tune a parameterized recovery that approximates the inverse of the unknown process.
Paving the Way for Reliable Quantum Hardware
By slashing measurement requirements and keeping classical overhead manageable, CQPT tackles a key scaling challenge in quantum hardware development. Its practical benefits include faster device calibration, targeted error identification, and streamlined inputs for quantum error correction protocols. So far, the team has validated CQPT with numerical simulations demonstrating feasibility across realistic noise models. The next steps are experimental demonstrations on physical quantum processors and integration into calibration workflows to accelerate the path toward robust, fault tolerant machines.
Institutions and lead researcher: Tohoku University, Nara Institute of Science and Technology (NAIST), University of Information Technology (Vietnam National University, Ho Chi Minh City), and Dr. Le Bin Ho.
