Applying Bayesian Techniques to the Optimization of Parameterized Quantum Circuits
Arthur G. Rattew · Open Scholarship Institutional Repository (Washington University in St. Louis) · 2020
In this work, we explore the challenges faced when creating quantum algorithms for near-term quantum computers. We examine the characteristics of problems that are amenable to such advantage, and the limitations of existing approaches. Additionally, we explore the importance of the classical optimizer in the Variational Quantum Eigensolver (VQE), and propose a Bayesian method to optimally configure a single parameter in a given quantum circuit. Experimental testing confirms that our method is significantly more tolerant to noise than the existing analytical approach and its variants (p < 0.008).