Application of ZX-calculus to Quantum Architecture Search

• DOI: 10.48550/arxiv.2406.01095
• Publication Date: 2024-06-03
• Authors: Tom Ewen; Ivica Turkalj; Patrick Holzer; Mark-Oliver Wolf
• Keywords: Quantum Physics; FOS: Physical sciences; Quantum Physics (quant-ph)

This paper presents a novel approach to quantum architecture search by integrating the techniques of ZX-calculus with Genetic Programming (GP) optimize structure parameterized circuits employed in Quantum Machine Learning (QML). Recognizing challenges designing efficient for QML, we propose GP framework that utilizes mutations defined via ZX-calculus, graphical language can simplify visualizing and working circuits. Our methodology focuses on evolving aim enhancing their capability approximate functions relevant various machine learning tasks. We introduce several mutation operators inspired transformation rules investigate impact efficiency accuracy The empirical analysis involves comparative study where these are applied diverse set regression problems, measuring performance metrics such as percentage valid after mutation, improvement objective, well circuit depth width. results indicate certain ZX-calculus-based perform significantly better than others Architecture Search (QAS) all considered. They suggest ZX-diagram based QAS shallower more uniformly allocated gates crude genetic optimization model.