基于量子门组单元的神经网络及其应用  

A Neural Network ModelBased-on Quantum Gates Cell and Its Applications

作　　者：解光军[1] 周典[2] 范海秋[1] 操礼程[1] 

机构地区：[1]合肥工业大学理学院,安徽合肥230009 [2]合肥工业大学管理学院,安徽合肥230009

出　　处：《系统工程理论与实践》2005年第5期113-117,共5页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(60302014);安徽省自然科学基金(03042202);中国博士后科学基金(20040350578)

摘　　要：　以通用量子门组(即相移门和受控非门)作为基本的计算单元,构造出全新的量子神经元模型,并由此组成前馈型结构网络.仿真结果表明,就文中算例而言,该量子神经网络的计算性能优于传统的神经网络.Abstract Quantum computation is well known for its particular computational performance. In this paper, a novel quantum neuron model is constructed based on the universal quantum gates unit (i.e. phase-shift gate and controlled-NOT gate), which acts as the basic computational component, with this type of neuron a feed-forward network structure is also built. The simulation results show that the quantum neuron network is superior to classical BP and RBF network for two financial data analysis examples.

关 键 词：量子计算 量子神经网络 量子门 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术] TP18