Compilation and Verification of Quantum Software in the Noisy and Approximate Regime
Lead Research Organisation:
University of Oxford
Department Name: Computer Science
Abstract
Like classical computers, quantum computers need compilers, which are tools that translate code written by a human into something the machine can run. Unlike classical computers, today's quantum computers have extremely limited computational resources and are highly susceptible from noise from the environment. Furthermore, it is often impractical or impossible to implement a computation exactly, meaning quantum compilers need to make certain approximations, which can yield further errors.
In this project, we will develop the next generation of quantum compilers which are able to understand, and tame, these many sources of errors, and produce efficient software with independently-checkable guarantees of reliable performance.
In this project, we will develop the next generation of quantum compilers which are able to understand, and tame, these many sources of errors, and produce efficient software with independently-checkable guarantees of reliable performance.
Organisations
Publications
Kissinger A
(2024)
Scalable Spider Nests (...Or How to Graphically Grok Transversal Non-Clifford Gates)
in Electronic Proceedings in Theoretical Computer Science
| Description | Quantum circuits are a low-level language for programming quantum computers. Often, one is only interested in implementing a computation, via a quantum circuit, that is "approximately" correct, i.e. it approximates a target computation within a fixed tolerance. We have developed a new technique, called "phase squashing", for approximate optimisation of quantum circuits, which aims to reduce the time and space required to perform a computation, while guaranteeing the target computation is still approximated within a fixed parameter epsilon, which we call the "error budget". For this, we developed a new technique for bounding global errors in ZX diagrams (one of the main structures used in quantum circuit optimisation), wrote an implementation of the phase squashing technique based on the open-source software tool PyZX (https://github.com/zxcalc/pyzx), and benchmarked the results on several families of circuits, showing state of the art performance in some cases. This work is currently under peer review, and will be reported once published. |
| Exploitation Route | These techniques have already been implemented in open source tools used by the quantum computing researchers. They could easily be adapted by other quantum compiling tools used in industry, such as IBM's qiskit and Quantinuum's t|ket> compiler, in order to produce more efficient quantum circuits in instances where approximation up to a fixed tolerance is required. |
| Sectors | Digital/Communication/Information Technologies (including Software) |