节点文献
神经网络不同类型、功能及其特性的演化
Evolution of Different Types, Function and Characteristics of Neural Networks
【Author】 Cong Shuang;Wang Da-Xin;Department of Automation, University of Science and Technology of China;
【机构】 中国科学技术大学自动化系;
【摘要】 深度神经网络正在被应用到各个领域中。本文分别对浅和深度不同类型神经网络的功能及其特性进行了演化分析,揭示了各种具有不同结构及其算法的网络功能及应用方面的优缺点,人们可以利用这些特点,根据应用的需要,选择合适的网络来获得更好的应用效果。
【Abstract】 Deep neural network is being applied to various fields. In this paper, the functions and characteristics of different types of shallow and deep neural networks are analyzed respectively, and the advantages and disadvantages of various network functions and applications with different structures and algorithms are revealed. People can use these characteristics to select appropriate networks according to the needs of applications to obtain better application effects.Finally, several applications of deep neural networks are given.
【Key words】 BP Network; Convolutional Neural Network; Boltzmann Machine Network,Deep Belief Network; Auto-Encoders; Generate Adversarial Network;
- 【会议录名称】 第23届中国系统仿真技术及其应用学术年会(CCSSTA23rd 2022)会议论文集
- 【会议时间】2022-10
- 【分类号】TP183