A review of Quantum Neural Networks: Methods, Models, Dilemma

TL;DR

This review paper summarizes recent developments in Quantum Neural Networks, focusing on implementation methods, circuit models, and current challenges, highlighting the field's exploratory nature and potential for quantum advantage.

Contribution

It provides a comprehensive overview of QNN development over six years, detailing methods, models, and difficulties, serving as a valuable resource for future research.

Findings

QNN shows higher storage capacity and efficiency due to quantum properties.

Several quantum circuit models like QBM and QCVNN are introduced.

Main challenges include implementation difficulties and theoretical limitations.

Abstract

The rapid development of quantum computer hardware has laid the hardware foundation for the realization of QNN. Due to quantum properties, QNN shows higher storage capacity and computational efficiency compared to its classical counterparts. This article will review the development of QNN in the past six years from three parts: implementation methods, quantum circuit models, and difficulties faced. Among them, the first part, the implementation method, mainly refers to some underlying algorithms and theoretical frameworks for constructing QNN models, such as VQA. The second part introduces several quantum circuit models of QNN, including QBM, QCVNN and so on. The third part describes some of the main difficult problems currently encountered. In short, this field is still in the exploratory stage, full of magic and practical significance.