A Programmable Threshold Circuit

A circuit and threshold technology, applied in the field of integrated circuits and neural networks, can solve the problems of a large number of tubes, inconvenient integration and high cost, and achieve the effects of low cost, simple structure and low power consumption

Active Publication Date: 2017-02-01
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The technical problem to be solved by the present invention is: aiming at the defects of large number of tubes, high cost, high power consumption, and inconvenient integration of the circuits used in the neural network in the traditional technology, a simple structure, low cost, and low power consumption are proposed. , Programmable threshold circuit for easy integration, can be applied in artificial neural network

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
  • A Programmable Threshold Circuit
  • A Programmable Threshold Circuit
  • A Programmable Threshold Circuit

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0034] Such as Figure 4 As shown, in this example, the neuron field effect transistor is an N-type neuron field effect transistor, its drain terminal is connected to the voltage source VDD, and the source terminal is grounded through the resistor R6; each weight adjustment circuit unit is composed of a memristor The top electrode of the memristor is connected to the input signal, and the bottom electrode is connected to the control grid of the neuron field effect transistor to control the voltage of the control grid, and the bottom electrode is grounded through the resistor; the output signal of the whole circuit is controlled by the neuron The source terminal voltage of the field effect tube is drawn out.

[0035] From Figure 4 It can be seen that the weight adjustment circuit unit composed of memristor M1 and resistor R1 is connected to the first control gate of the neuron field effect transistor, and M1 is connected to the input signal V1; the circuit unit composed of me...

Embodiment 2

[0053] Such as Figure 9 As shown, the programmable threshold circuit structure in this example is the same as Figure 4 The structures in are similar, and the weight adjustment circuit units are composed of memristors and resistors. The difference is that the positions of the memristors and resistors in this example are equivalent to those in Figure 4 The neuron FET is still an N-type neuron FET, its drain terminal is connected to the voltage source VDD, and the source terminal is grounded through the resistor R6; each weight adjustment circuit unit is composed of a memristor One end of the resistance is connected to the input signal, the other end is connected to the control grid, the top electrode of the memristor is connected to the control grid, and the bottom electrode is grounded.

[0054] From Figure 9 It can be seen that the weight adjustment circuit unit composed of memristor M1 and resistor R1 is connected to the first control gate of the neuron field effect tra...

Embodiment 3

[0056] Such as Figure 10 As shown, the programmable threshold circuit structure in this example is the same as Figure 4 The structure in is similar, the difference is that the Figure 4 The resistors in the weight adjustment circuit unit are all replaced by memristors, that is, the weight adjustment circuit unit in this example is composed of the first memristor and the second memristor; in this example, the neuron field effect transistor It is still an N-type neuron field effect transistor, its drain terminal is connected to a voltage source VDD, and its source terminal is grounded through a resistor R1.

[0057] From Figure 10 It can be seen that the weight adjustment circuit unit composed of memristor M1 and memristor M6 is connected to the first control gate of the neuron field effect transistor, and M1 is connected to the input signal V1; the memristor M2 and the memristor The weight adjustment circuit unit composed of M7 is connected to the second control grid of the...

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 integrated circuits and neural networks. Aiming at the defects of large number of tubes, high cost, high power consumption, and inconvenient integration of circuits used in neural networks in traditional technologies, a simple structure, low cost, and functional A programmable threshold circuit with low power consumption and easy integration; the circuit includes a neuron field effect transistor, and the neuron field effect transistor includes a plurality of control gates, and also includes a number equal to the control gates of the neuron field effect transistor A weight adjustment circuit unit, the weight adjustment circuit unit is connected to the control grid in one-to-one correspondence, the weight adjustment circuit unit is used to adjust the synaptic weight of the neuron field effect transistor, and the neuron The FET is an N-type neuron FET or a P-type neuron FET. The invention is suitable for artificial neural network.

Description

technical field [0001] The present invention relates to the field of integrated circuits and neural networks, and in particular to a programmable threshold circuit, which is used to simulate the weight adjustment function of time-series plasticity of neurons, and is also used for the two working states of neurons (learning state, calculation state) simulation. Background technique [0002] As digital computers encounter insoluble difficulties in fuzzy pattern recognition, associative memory and self-learning, neural network computing has regained people's attention, and artificial neural networks have emerged as the times require. An artificial neural network refers to a computing system that operates in a manner similar to a biological brain, based on an electronic system. Artificial neural networks create connections between processing units that are essentially equivalent in function to the neurons of a biological brain. Therefore, the foundation of the neural network i...

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 Patents(China)
IPC IPC(8): G06N3/06
Inventor 刘洋胡绍刚徐艳飞董华吴霜毅于奇
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA
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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap