| 深度学习神经网络的新型自适应激活函数 |
| |
| 引用本文： | 刘宇晴,王天昊,徐旭.深度学习神经网络的新型自适应激活函数[J].吉林大学学报(理学版),2002,57(4):857-859. |
| |
| 作者姓名： | 刘宇晴 王天昊 徐旭 |
| |
| 作者单位： | 吉林大学 数学学院, 长春 130012 |
| |
| 摘 要： | 构造一个含有参数的光滑激活函数用于深度学习神经网络, 通过基于误差反向传播算法建立参数的在线修正公式, 避免了梯度弥散、 不光滑及过
拟合等问题. 与一些常用的激活函数进行对比实验结果表明, 新的激活函数在多个数据集上效果均较好.
|
| 关 键 词： | 激活函数 卷积神经网络 机器学习 |
| 收稿时间： | 2019-01-17 |
New Adaptive Activation Function for Deep Learning Neural Networks |
| |
| LIU Yuqing,WANG Tianhao,XU Xu.New Adaptive Activation Function for Deep Learning Neural Networks[J].Journal of Jilin University: Sci Ed,2002,57(4):857-859. |
| |
| Authors: | LIU Yuqing WANG Tianhao XU Xu |
| |
| Institution: | College of Mathematics, Jilin University, Changchun 130012, China |
| |
| Abstract: | A smooth activation function with a parameter was constructed for the deep learning neural networks. The online correction formula for this parameter was established based on the error back propagation algorithm, which avoided the problemsof gradients losing, non smooth and over fitting. Compared with some popular activation functions, the results show that the new activation function works well on many data sets. |
| |
| Keywords: | activation function convolutional neural network machine learning |
|
| 点击此处可从《吉林大学学报(理学版)》浏览原始摘要信息 |
| 点击此处可从《吉林大学学报(理学版)》下载免费的PDF全文 |
