杨洁

个人信息Personal Information

副教授

博士生导师

硕士生导师

性别:女

毕业院校:大连理工大学

学位:博士

所在单位:数学科学学院

学科:计算数学

办公地点:大连理工大学数学科学学院505

联系方式:0411-84708351-8205

电子邮箱:yangjiee@dlut.edu.cn

扫描关注

论文成果

当前位置: 中文主页 >> 科学研究 >> 论文成果

A New Conjugate Gradient Method with Smoothing L-1/2 Regularization Based on a Modified Secant Equation for Training Neural Networks

点击次数:

论文类型:期刊论文

发表时间:2018-10-01

发表刊物:NEURAL PROCESSING LETTERS

收录刊物:SCIE

卷号:48

期号:2,SI

页面范围:955-978

ISSN号:1370-4621

关键字:Feedforward neural networks; Conjugate gradient method; Modified secant equation; Regularization; Global convergence

摘要:Proposed in this paper is a new conjugate gradient method with smoothing L-1/2 regularization based on a modified secant equation for training neural networks, where a descent search direction is generated by selecting an adaptive learning rate based on the strong Wolfe conditions. Two adaptive parameters are introduced such that the new training method possesses both quasi-Newton property and sufficient descent property. As shown in the numerical experiments for five benchmark classification problems from UCI repository, compared with the other conjugate gradient training algorithms, the new training algorithm has roughly the same or even better learning capacity, but significantly better generalization capacity and network sparsity. Under mild assumptions, a global convergence result of the proposed training method is also proved.