一种基于LM的量子神经网络训练算法

计算机科学 ›› 2013, Vol. 40 ›› Issue (9): 221-224.

一种基于LM的量子神经网络训练算法

张翼鹏,陈亮,郝欢

解放军理工大学通信工程学院南京210007;解放军理工大学通信工程学院南京210007;解放军理工大学通信工程学院南京210007

出版日期:2018-11-16 发布日期:2018-11-16
基金资助:
本文受国家自然科学基金项目(61072042)资助

LM Based Training Algorithm for Quantum Neural Networks

ZHANG Yi-peng,CHENG Liang and HAO Huan

Online:2018-11-16 Published:2018-11-16

摘要/Abstract

摘要： 针对量子神经网络的训练结果易陷入局部极小值的问题,将Levenberg-Marquardt(LM)算法引入到原训练算法中,从而提高网络收敛速度与训练效果。并通过改进原训练算法的迭代步骤,解决训练过程中网络权值与量子间隔不同的目标函数相互冲突引起的输出均方误差和波动的问题。实验结果表明,相比原训练算法,引入LM后的训练算法可以大幅减少迭代次数,显著降低网络收敛值,提高量子神经网络的分类效果。

关键词: 量子神经网络,Levenberg-Marquardt算法,最速下降,量子间隔中图法分类号TP183文献标识码A

Abstract: Aiming at the question that the result trained by quantum neural networks is easy to fall into the local least value,the Levenberg-Marquardt algorithm was introduced into the original training algorithm to increase the training speed and improve the performance of the network．In addition,the conflict between different objective functions used for the training synaptic weights and quantum intervals can cause mean square error fluctuates．In order to solve this problem,the iteration order of the original training algorithm was adjusted．The experimental results show that,compared with the original training algorithm,the algorithm using LM can significantly reduce the number of iterations,significantly decrease the mean square error when the network convergences,and improve the classification results of quantum neural network.

Key words: Quantum neural networks,Levenberg-marquardt algorithm,Steepest descent,Quantum intervals

张翼鹏,陈亮,郝欢. 一种基于LM的量子神经网络训练算法[J]. 计算机科学, 2013, 40(9): 221-224. https://doi.org/

ZHANG Yi-peng,CHENG Liang and HAO Huan. LM Based Training Algorithm for Quantum Neural Networks[J]. Computer Science, 2013, 40(9): 221-224. https://doi.org/

参考文献

[1] Hagan M T,Demuth H B,Beale M H．Neural Network Design [M]．Boston:PWS Pub Co.,1995
[2] Purushothaman G,Karayiannis N B．Quantum neural networks(QNNs):inherently fuzzy feedforward neural networks[J]．Neural Networks,IEEE Transactions on,1997,8(3):679-693
[3] Karayiannis N B,Xiong Y．Training Reformulated Radial Basis Function Neural Networks Capable of Identifying Uncertainty in Data Classification[J]．IEEE Transactions on Neural Networks,2006,17(5):1222-1234
[4] 肖婧,谭阳红．基于新特征提取法和量子神经网络的手写数字识别[J]．电子测量技术,2009(06):84-87
[5] 盖怀存,张小锋,江泽涛．基于量子神经网络的人脸识别技术研究[J]．计算机工程与应用,2010,46(8):187-189
[6] 王金明,王耿,郑国宏,等．一种量子神经网络说话人识别方法[J]．解放军理工大学学报:自然科学版,2012,13(3):242-246
[7] 孙健,张雄伟,孙新建．一种新的量子神经网络训练算法[J]．信号处理,2011,27(9):1306-1312
[8] 张鸿燕,耿征．Levenberg-Marquardt算法的一种新解释[J]．计算机工程与应用,2009,45(19):5-8
[9] Indrajit M,Srikanta R．Comparing the performance of neuralnetworks developed by using Levenberg-Marquardt and Qusai-Newton with the gradient descent algorithm for modeling a multiple response grinding process[J]．Expert Systems with Applications,2012,39(3):2397-2407
[10] Wilamowski B M,Hao Yu．Improved Computation for Levenberg-Marquardt Training[J]．IEEE Transactions on Neural Networks,2010,21(6):930-937

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed