Sampling Neural Network: A Novel Neural Network Based on Sampling Theorem

In this paper, a new neural network algorithm is proposed, which is based on the signal sampling theorem. A new activation function of neural network in Sa(t) form and a new neuron weight error correction method are proposed. The sampling neural network algorithm (SNN) has the following advantages:...

Full description

Saved in:
Bibliographic Details
Published in2021 6th International Symposium on Computer and Information Processing Technology (ISCIPT) pp. 717 - 720
Main Authors Cai, Gang, Yan Wu, Ling
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.06.2021
Subjects
activation function
Computer network reliability
Fitting
neuron weight error correction
Neurons
Reliability engineering
Reliability theory
Sampling Theorem
Signal sampling
SNN
Training
Online AccessGet full text

Cover

More Information
Summary:In this paper, a new neural network algorithm is proposed, which is based on the signal sampling theorem. A new activation function of neural network in Sa(t) form and a new neuron weight error correction method are proposed. The sampling neural network algorithm (SNN) has the following advantages: clarity in theory, simple in design, reliable in operation, convenient in hardware implementation, low in operations, low in input sample data requirements, and does not fall into local minimum. The specific implementation of the method fully considers the hardware implementation, so the neural network algorithm has certain application prospects and value.
DOI:10.1109/ISCIPT53667.2021.00151