Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Structure of deep memory convolution neural network and construction method of structure

A convolutional neural network and network structure technology, applied in neural learning methods, biological neural network models, neural architectures, etc.

Active Publication Date: 2017-11-21
南方电网互联网服务有限公司
View PDF3 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a structure of deep memory convolutional neural network and its construction method, to a certain extent overcome the shortcomings of the representation ability of complex functions in the case of limited samples and computing units, and improve the existing convolutional neural network. computing efficiency

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Structure of deep memory convolution neural network and construction method of structure
  • Structure of deep memory convolution neural network and construction method of structure
  • Structure of deep memory convolution neural network and construction method of structure

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0089] The invention improves the network operation efficiency by improving the structure of the convolutional neural network and adding memory into the convolutional network.

[0090] The present invention will be described in detail below in conjunction with the accompanying drawings and examples.

[0091] 1. Network structure

[0092] Part 1: Convolutional Neural Network Structure with Clustering Dimensionality Reduction with Five Convolutional Layers

[0093] 1) The first convolution layer selects 96 convolution operators, each convolution operator is a 16×16 grayscale image block, and the image block contains 72 different shapes of straight lines and 8 different sizes of discs and 16 different shapes of rings;

[0094] 2) The expression of the convolution process of the first convolutional layer is:

[0095]

[0096] in for image P 0 The gray value at pixel [2i-1+x,2j-1+y], Represents the weight of the convolution operator at position [x,y], is the convolved ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention relates to the field of learning-based neural networks, in particular to a structure of a deep memory convolution neural network and a construction method of the structure. The structure comprises a five-convolution layer clustering and dimension reduction containing convolution neural network structure, a deep memory neural network structure and a BP network structure. The invention also relates to the construction method of the network structure. According to the invention, the disadvantage of the expression ability of a complex function under the conditions of limited samples and limited calculation units can be overcome to a certain degree and efficiency of the current convolution neural network is improved.

Description

technical field [0001] The invention relates to the field of learning-based neural networks, in particular to a structure of a deep memory convolutional neural network and a construction method thereof. Background technique [0002] The concept of deep learning originated from the research of artificial neural networks, and one of the earliest deep learning structures is the multi-layer perceptron MLP. Deep learning discovers distributed features of data by combining low-level features to form more abstract high-level representations. As a typical algorithm for traditional multi-layer network training, BP network has defects in large-scale input problems. The ubiquitous local minima in the non-convex objective cost function of deep architectures are the main source of training difficulties. References (Hinton G E, Osindero S, Teh Y W.A fast learning algorithm for deep belief nets[J].Neuralcomputation,2006,18(7):1527-1554.) proposed a deep belief network (DBN) based on deep...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06N3/04G06N3/08
CPCG06N3/084G06N3/045
Inventor 鄢化彪黄绿娥尹宝勇刘祚时
Owner 南方电网互联网服务有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com