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

Hopfield neural network system based on electromagnetic radiation effect, processor chip and terminal

A technology of electromagnetic radiation and neural network, applied in the field of neural dynamics, which can solve the problems of insufficient security, insufficient dynamic diversification, insufficient instantaneous and intermittent chaos, etc.

Active Publication Date: 2021-09-10
CHANGSHA UNIVERSITY OF SCIENCE AND TECHNOLOGY
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The applicant found in the research that the system constructed by the currently proposed Hopfield neural network model still has the defects of insufficient dynamic diversification a

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
  • Hopfield neural network system based on electromagnetic radiation effect, processor chip and terminal
  • Hopfield neural network system based on electromagnetic radiation effect, processor chip and terminal
  • Hopfield neural network system based on electromagnetic radiation effect, processor chip and terminal

Examples

Experimental program
Comparison scheme
Effect test

no. 1 example ；

[0037] The first embodiment of the present invention provides a kind of Hopfield neural network system based on the effect of electromagnetic radiation. In this system, the Hopfield neural network based on the effect of electromagnetic radiation is constructed, wherein the second neuron of the Hopfield neural network based on the effect of electromagnetic radiation With self-feedback synaptic weight and flux-controlled memristive electromagnetic radiation with secondary internal state, this embodiment provides a Hopfield neural network based on the electromagnetic radiation effect as:

[0038]

[0039] Among them, x, y, z represent state variables; tanh(x), tanh(y), tanh(z) represent nonlinear neuron activation functions; α, β, μ, η represent memristive parameters; φ represents memristor The magnetic flux of the memristor; v represents the input voltage of the memristor; i represents the output current of the memristor; ρ represents the intensity of electromagnetic radiation...

no. 2 example ；

[0084] The second embodiment of the present invention provides a processor chip equipped with the Hopfield neural network system based on electromagnetic radiation described in the first embodiment.

[0085] In this embodiment, a digital circuit is designed on the vivado 2018.3 simulation platform to further verify various types of attractors generated by the system model provided in the first embodiment. First, implementing neuron activation functions, which consist of infinite exponential progressions, are challenging to implement directly numerically. Here, a second-order piecewise function is introduced, which can well approximate the hyperbolic tangent function, which can be defined as:

[0086]

[0087]

[0088] Where L=2, β=1 and θ=0.25, β and θ can determine H s (z) Slope and gain between -L≤z≤L.

[0089] Secondly, for the continuous Hopfield neural network system, the high-precision fourth-order Runge-Kutta method is used to discretize it, as shown in formula ...

no. 3 example ；

[0094] With the continuous development and popularization of communication technology, the frequency of data information being stolen, illegally copied and disseminated is getting higher and higher. As the main carrier of information dissemination, the security problems of images are becoming more and more serious. Chaotic systems with initial sensitivity and unpredictability are widely used in the field of information security. The most important applications include pseudo-random number generators and image encryption. Multi-rolling chaotic attractors have more complex topology, better randomness and Larger key space undoubtedly improves the security performance of encrypted data. Compared with ordinary chaotic systems, the Hopfield neural network model with chaotic behavior has better nonlinear and associative memory effects, and can produce chaotic matrices with good diffusion effects. Therefore, the system provided by the first embodiment of the present invention has bro...

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 a Hopfield neural network system based on an electromagnetic radiation effect, a processor chip and a terminal, the system constructs a Hopfield neural network based on the electromagnetic radiation effect, and a second neuron of the Hopfield neural network based on the electromagnetic radiation effect has a self-feedback synaptic weight and has magnetic flux of a secondary internal state to control memristive electromagnetic radiation. Compared with a system constructed by an existing Hopfield neural network model, the method has the advantages that the number of balance points in the Hopfield neural network is changed through the electromagnetic radiation intensity, so that the system generates richer dynamic behaviors, the system has high randomness and excellent safety, and the system is suitable for chaos engineering application.

Description

technical field [0001] The invention relates to the technical field of neural dynamics, in particular to a Hopfield neural network system based on electromagnetic radiation, a processor chip and a terminal. Background technique [0002] Neural dynamics has achieved great success in application fields such as pattern recognition, automatic control, signal processing, medicine, artificial intelligence, transportation, and economy. So far, hundreds of artificial neurons and neural network models have been developed and utilized, which has greatly promoted the rapid development of neural dynamics. Among them, J.J. Hopfield proposed a neural network stored in a way very similar to the brain, That is the famous Hopfield neural network. The Hopfield neural network composed of nonlinear units belongs to a more complex chaotic system. It is a neural network model composed of a large number of nonlinear processing units. It has rich dynamic characteristics and can reflect many basic ...

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): G06N3/06G06N3/063G06F17/11
CPCG06N3/061G06N3/063G06F17/11
Inventor 余飞沈辉张梓楠黄园媛蔡烁
Owner CHANGSHA UNIVERSITY OF SCIENCE AND TECHNOLOGY
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