Research on deep neural network model construction and overfitting

Tong-han Li; Hui Zhang; Linchang Fan; Hao Wang; Qian Liu

doi:10.1117/12.2639137

15 July 2022 Research on deep neural network model construction and overfitting

Tong-han Li, Hui Zhang, Linchang Fan, Hao Wang, Qian Liu

Proceedings Volume 12258, International Conference on Neural Networks, Information, and Communication Engineering (NNICE 2022); 122580G (2022) https://doi.org/10.1117/12.2639137
Event: International Conference on Neural Networks, Information, and Communication Engineering (NNICE 2022), 2022, Qingdao, China

Abstract

In recent years, deep neural network has been widely used in image recognition, natural language processing, computer vision and other fields, but it is prone to overfitting during network training. To solve this problem, this paper uses TensorFlow2.0 framework to construct multilayer perceptron deep network for Fashion-MNIST dataset, and uses dropout algorithm to solve the overfitting problem in the process of network training. The research results show that the dropout algorithm is applied to the deep neural network, which can make the deep neural network model have strong generalization ability and can effectively solve the overfitting problem of the training network. The research on overfitting problem has important practical significance for reducing the identification error of deep network.

Citation Download Citation

Tong-han Li, Hui Zhang, Linchang Fan, Hao Wang, and Qian Liu "Research on deep neural network model construction and overfitting", Proc. SPIE 12258, International Conference on Neural Networks, Information, and Communication Engineering (NNICE 2022), 122580G (15 July 2022); https://doi.org/10.1117/12.2639137

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available