An automated geometric space curve approach for designing dynamically corrected gates

Abstract

The noisy nature of quantum hardware necessitates the implementation of high-fidelity quantum gates in a noise-insensitive manner. While there exist many powerful methods for designing dynamically corrected gates, they typically involve an exploration across a large-dimensional landscape filled with solutions that are only locally optimal, making it challenging to find globally optimal ones. Moreover, these methods often use a single cost function to try to accomplish the two disparate goals of achieving a target gate and suppressing noise, and this can lead to unnecessary tradeoffs between the two and, consequently, lower fidelities. Here, we present a method for designing dynamically corrected gates called B\'ezier Ansatz for Robust Quantum (BARQ) control to address these challenges. Rather than numerically optimizing the controls directly, BARQ instead makes use of the Space Curve Quantum Control formalism in which the quantum evolution is mapped to a geometric space curve. In the formulation used by BARQ, the boundary conditions of the space curve determine the target gate, while its shape determines the noise robustness of the corresponding gate. This allows the target gate to be fixed upfront, so that numerical optimization is only needed to achieve noise-robustness, and this is performed efficiently using a control-point parameterization of the space curve. In this way, BARQ eliminates the gate-fixing and noise-robustness tradeoff while also providing a global perspective into the control landscape, and allows for ample freedom to design experimentally friendly and robust control pulses. The pulse design is facilitated through the developed software package qurveros.