The SC Solutions team has won a Small Business Technology Transfer (STTR) Phase I funding from U.S. Army Research Laboratories to develop the necessary software to help commercialize a scalable software tool for validation and verification of quantum systems, such as quantum computers.
Quantum information processors are rapidly transitioning from small-scale, research-grade experiments to full-stack, commercial platforms. IBM, Rigetti, and others have each launched cloud-accessible quantum computers with up to 20 qubits, and Google has announced development of a 72-qubit device. As the hardware has matured, intricate experimental controls have been replaced by drag-and-drop web front ends, and the community of potential users for these new quantum computers has grown from domain experts to anyone with an internet connection. But quantum information processors remain extremely complicated devices, and they can fail in myriad interesting and unexpected ways. The R&D community urgently needs simple tools that can identify and characterize faults in these intermediate-scale devices and report that information in an intuitive, easily-digestible form. Unfortunately, while quantum systems scale ever larger and their interfaces grow more easily accessible, the tools to benchmark their performance continues to lag.
Modern Quantum Computing, Verification and Validation (QCVV) techniques have proven themselves well-suited to the one- and two-qubit gates that form the key building blocks of larger-scale devices. But dealing with more qubits is different, and large-scale system performance generally falls short of expectations derived from such one- and two-qubit characterizations. Crosstalk and correlated decoherence effects play increasingly important roles as systems scale up, but many small-scale QCVV techniques are completely blind to their effects. Furthermore, the software programs that implement these protocols are often cumbersome, research-grade tools written by, and effectively for, QCVV experts. As quantum devices mature, the software tools that characterize and benchmark their performance must adapt to the new realities of both larger hardware and a broader user base.
The SC team will develop a user-friendly software package containing state-of-the-art tools for performing QCVV on the full spectrum of circuit-model quantum computing hardware platforms. This package will be designed from the start with many-qubit processors in mind. Our collaborative effort will heavily leverage the SNL-developed, open-source software package pyGSTi (www.pygsti.info). This well-known toolset implements experiment design and data analysis for the most commonly used QCVV protocols, including gate set tomography (GST), randomized benchmarking (RB), and others. This tool has been under continuous development at SNL for several years, and in steady use by quantum computing hardware experimentalists at SNL and at large. To broaden the existing expert-dominated user-base, SC will apply its user-interface (UI) design capabilities acquired over the past twenty-five years in developing and licensing real-time control software to the semiconductor and LED manufacturing and aerospace industries. The resulting software package will combine flexible and scalable software tools for QCVV with a user-friendly interface and comprehensive documentation. We expect this tool will prove to be indispensable to the broad spectrum of the quantum computing community.