Variational Quantum Gate Optimization

We propose a gate optimization method, which we call variational quantum gate optimization (VQGO). VQGO is a method to construct a target multi-qubit gate by optimizing a parametrized quantum circuit which consists of tunable single-qubit gates with high fidelities and fixed multi-qubit gates with limited controlabilities. As an example, we apply the proposed scheme to the models relevant to superconducting qubit systems. We show in numerical simulations that the high-fidelity CNOT gate can be constructed with VQGO using cross-resonance gates with finite crosstalk. We also demonstrate that fast and a high-fidelity four-qubit syndrome extraction can be implemented with simultaneous cross-resonance drives even in the presence of non-commutative crosstalk. VQGO gives a pathway for designing efficient gate operations for quantum computers.