Improving Convolutional Neural Networks with Competitive Activation Function | |
Ying, Yao1; Zhang, Nengbo2; He, Ping3,4; Peng, Silong5 | |
刊名 | SECURITY AND COMMUNICATION NETWORKS |
2021-05-13 | |
卷号 | 2021页码:9 |
ISSN号 | 1939-0114 |
DOI | 10.1155/2021/1933490 |
通讯作者 | He, Ping(pinghecn@qq.com) |
英文摘要 | The activation function is the basic component of the convolutional neural network (CNN), which provides the nonlinear transformation capability required by the network. Many activation functions make the original input compete with different linear or nonlinear mapping terms to obtain different nonlinear transformation capabilities. Until recently, the original input of funnel activation (FReLU) competed with the spatial conditions, so FReLU not only has the ability of nonlinear transformation but also has the ability of pixelwise modeling. We summarize the competition mechanism in the activation function and then propose a novel activation function design template: competitive activation function (CAF), which promotes competition among different elements. CAF generalizes all activation functions that use competition mechanisms. According to CAF, we propose a parametric funnel rectified exponential unit (PFREU). PFREU promotes competition among linear mapping, nonlinear mapping, and spatial conditions. We conduct experiments on four datasets of different sizes, and the experimental results of three classical convolutional neural networks proved the superiority of our method. |
资助项目 | National Natural Science Foundation of China[11705122] ; Science and Technology Program of Sichuan[2020YFH0124] ; Guangdong Basic and Applied Basic Research Foundation[2021A1515011342] ; Zigong Key Science and Technology Project of China[2020YGJC01] |
WOS研究方向 | Computer Science ; Telecommunications |
语种 | 英语 |
出版者 | WILEY-HINDAWI |
WOS记录号 | WOS:000668944400001 |
资助机构 | National Natural Science Foundation of China ; Science and Technology Program of Sichuan ; Guangdong Basic and Applied Basic Research Foundation ; Zigong Key Science and Technology Project of China |
内容类型 | 期刊论文 |
源URL | [http://ir.ia.ac.cn/handle/173211/45255] |
专题 | 自动化研究所_智能制造技术与系统研究中心_多维数据分析团队 |
通讯作者 | He, Ping |
作者单位 | 1.Northeastern Univ, Coll Informat Sci & Engn, Shenyang 110819, Peoples R China 2.Shenzhen Univ, Coll Comp Sci & Software Engn, Shenzhen 518060, Peoples R China 3.Jinan Univ, Sch Intelligent Syst Sci & Engn, Guangzhou 519070, Guangdong, Peoples R China 4.Sichuan Univ Sci & Engn, Artificial Intelligence Key Lab Sichuan Prov, Zigong 643000, Sichuan, Peoples R China 5.Chinese Acad Sci, Inst Automat, Beijing 100190, Peoples R China |
推荐引用方式 GB/T 7714 | Ying, Yao,Zhang, Nengbo,He, Ping,et al. Improving Convolutional Neural Networks with Competitive Activation Function[J]. SECURITY AND COMMUNICATION NETWORKS,2021,2021:9. |
APA | Ying, Yao,Zhang, Nengbo,He, Ping,&Peng, Silong.(2021).Improving Convolutional Neural Networks with Competitive Activation Function.SECURITY AND COMMUNICATION NETWORKS,2021,9. |
MLA | Ying, Yao,et al."Improving Convolutional Neural Networks with Competitive Activation Function".SECURITY AND COMMUNICATION NETWORKS 2021(2021):9. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论