Adaptive learning rate BP neural network algorithm

A BP neural network and adaptive learning rate technology, applied in the field of artificial intelligence, can solve problems such as low algorithm complexity, long training time, easy to fall into local minimum, etc., to achieve shortened training time, simple algorithm program, and easy implementation Effect

Inactive Publication Date: 2018-09-14
NANJING UNIV OF SCI & TECH
View PDF0 Cites 12 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The technical problem solved by the present invention is to provide a BP neural network algorithm of adaptive learning rate, to solve the problem that the existing BP algorithm training time is long and easy to fall into ...

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
  • Adaptive learning rate BP neural network algorithm
  • Adaptive learning rate BP neural network algorithm
  • Adaptive learning rate BP neural network algorithm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0058] The simulation experiment based on the BP neural network algorithm adopts Matlab to carry out software programming, according to the improved BP algorithm proposed by the present invention, a set of irregular data to be corrected is imported into the program for training, and is corrected with the existing neural network algorithm. The results are compared.

[0059] Before the training starts, the structure of the neural network needs to be set. First, determine the number of hidden layers of the neural network and the number of neurons in each layer. The selection of hidden layers is related to training accuracy and training time. If too many hidden layers are selected, the gradient of the error surface will not be stable enough, and the network will have more local valleys, making it easier for the network to fall into local minima during the training process. And if too few hidden layers are selected, since the weight adjustment error of each hidden layer has a certa...

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 discloses an adaptive learning rate BP neural network algorithm. The back propagation process of the BP neural network algorithm is optimized, and different learning rates are dynamically adapted to the connection weight of different neurons so that optimization of each direction is ensured to search the optimal solution of the direction, the adjustment efficiency of the weight can be enhanced to the greatest extent to accelerate the convergence speed of the whole training. The algorithm comprises the step one: network initialization; step two: sample inputting; step three: forward propagation; step four: result judgment; and step five: back propagation. The algorithm is simple in program and easy to implement and great in improvement effect so that the training time of the BP neural network can be greatly shortened, the problem of being prone to the local minimum of the present algorithm can be effectively overcome, the universality is high and thus the algorithm has great practical application value.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence, in particular to a BP neural network algorithm with an adaptive learning rate. Background technique [0002] Artificial neural network is one of the research hotspots in the direction of artificial intelligence in modern science. It analyzes the network structure of the human brain nervous system and the system mechanism of human brain nerve work to establish a mathematical network model that imitates the work of the human brain, has real-time processing and storage functions. BP neural network is the most commonly used algorithm in artificial intelligence network, and it is one of the most mature network structures currently researched. Because of its super self-learning, self-organization, self-adaptation, associative memory and fault tolerance, its application is very extensive. It covers almost every aspect of life, including information processing, automatic control, economy...

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
IPC IPC(8): G06N3/08
CPCG06N3/08
Inventor 华雨王晓鸣
Owner NANJING UNIV OF SCI & TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products