期刊论文、会议论文

Wang, Jian

英文

Lecture Notes in Computer Science

0302-9743

2017

10637

131-139

24th International Conference on Neural Information Processing (ICONIP)

Lecture Notes in Computer Science

NOV 14-18, 2017

Guangzhou, PEOPLES R CHINA

[Zhang, Bingjie;Gao, Tao;Wang, Jian] China Univ Petr, Coll Sci, Qingdao 266580, Peoples R China.^[Li, Long] Hengyang Normal Univ, Coll Math & Stat, Hengyang 421008, Peoples R China.^[Sun, Zhanquan] Shandong Comp Sci Ctr, Shandong Prov Key Lab Comp Networks, Natl Supercomp Ctr Jinan, Jinan 250014, Shandong, Peoples R China.

Chinese Acad Sci, Inst Automat, Guangdong Univ Technol, S China Univ Technol, Springers Lecture Notes Comp Sci, IEEE CAA Journal Automatica Sinica, Asia Pacific Neural Network Soc

Liu, D Xie, S Li, Y Zhao, D ElAlfy, ESM

GEWERBESTRASSE 11, CHAM, CH-6330, SWITZERLAND

SPRINGER INTERNATIONAL PUBLISHING AG

978-3-319-70093-9; 978-3-319-70092-2

10.1007/978-3-319-70093-9_14

National Natural Science Foundation of ChinaNational Natural Science Foundation of China (NSFC) [61305075, 11401185]; China Postdoctoral Science FoundationChina Postdoctoral Science Foundation [2012M520624]; Natural Science Foundation of Shandong ProvinceNatural Science Foundation of Shandong Province [ZR2013FQ004, ZR2013DM015, ZR2015AL014]; Specialized Research Fund for the Doctoral Program of Higher Education of ChinaSpecialized Research Fund for the Doctoral Program of Higher Education (SRFDP) [20130133120014]; Fundamental Research Funds for the Central UniversitiesFundamental Research Funds for the Central Universities [13CX05016A, 14CX05042A, 15CX05053A, 15CX08011A, 15CX02064A]

本校为其他机构

In this paper, by constructing a generalized Armijo search method, a novel conjugate gradient (CG) model has been proposed to training a common three-layer backpropagation (BP) neural network. Compared with the classical gradient descent method, this algorithm efficiently accelerates the convergence speed due to the existence of the additional conjugate direction. Essentially, the optimal learning rate of each epoch is determined by the given inexact line search strategy. The presented model does not significantly increase the computational cost in dealing with real applications. Two benchmark...