Adaptive Circuit Learning of Born Machine: Towards Realization of Amplitude Embedding and Data Loading (2311.17798v1)

Abstract: With the progress in the quantum algorithm in recent years, much of the existing literature claims the exponential quantum advantage against their classical counterpart. However, many of these successes hinge on the assumption that arbitrary states can be efficiently prepared in quantum circuits. In reality, crafting a circuit to prepare a generic $n$-qubit quantum state demands an operation count on the order of $\mathcal{O}(2^n)$, which is prohibitively demanding for the quantum algorithm to demonstrate its advantage against the classical one. To tackle this data-loading problem, numerous strategies have been put forward. Nonetheless, most of these approaches only consider a very simple and easy-to-implement circuit structure, which has been shown to suffer from serious optimization issues. In this study, we harness quantum circuits as Born machines to generate probability distributions. Drawing inspiration from methods used to investigate electronic structures in quantum chemistry and condensed matter physics, we present a novel algorithm "Adaptive Circuit Learning of Born Machine" (ACLBM) that dynamically expands the ansatz circuit. Our algorithm is tailored to selectively integrate two-qubit entangled gates that best capture the complex entanglement present within the target state. Empirical results underscore the proficiency of our approach in encoding real-world data through amplitude embedding, demonstrating not only compliance with but also enhancement over the performance benchmarks set by previous research.