SpikeLM: A Second-Order Supervised Learning Algorithm for Training Spiking Neural Networks

摘要：

For networks of spiking neurons which encode information in individual spike firing times, a second-order supervised learning algorithm, SpikeLM, is derived akin to the traditional Levenberg-Marquardt algorithm. The backpropagation computation procedure is obtained by introducing specific matrices. The classical XOR problem and a function approximation experiment have been applied to validate the improved algorithm. The experiment results roughly show the potential performance of this new algorithm in the machine learning field.

作者: Yongji Wang Jian Huang

作者单位: Department of Control Science and Engineering Huazhong University of Science and Technology Wuhan City, Hubei Province, P.R. China, 430074

会议类型: 国际会议

会议名称: 第三届国际脉冲动力系统及应用学术会议

会议地点: 青岛

会议语种:英文

在线出版日期: 2006-07-21（万方平台首次上网日期，不代表论文的发表时间）