CONVOLUTIONAL NEURAL NETWORK ON-CHIP LEARNING SYSTEM BASED ON NON-VOLATILE MEMORY

Title:

CONVOLUTIONAL NEURAL NETWORK ON-CHIP LEARNING SYSTEM BASED ON NON-VOLATILE MEMORY

Document Type and Number:

WIPO Patent Application WO/2020/052342

Kind Code:

A1

Abstract:

Disclosed by the present invention is a convolutional neural network on-chip learning system based on a non-volatile memory, comprising: an input module, a convolutional neural network module, an output module and a weight update module. The on-chip learning of the convolutional neural network module implements a synaptic function by using a characteristic which the conductance of a memristor changes according to an applied pulse, and the convolutional kernel value or synaptic weight value is stored in a memristor unit; the input module converts an input signal into a voltage signal required by the convolutional neural network module; the convolutional neural network module converts the input voltage signal level by level, and transmits the result to the output module to obtain an output of the network; and the weight update module adjusts the conductance value of the memristor in the convolutional neural network module according to the result of the output module to update a network convolutional kernel value or synaptic weight value. The present invention is designed to implement the on-chip learning of the convolutional neural network to implement the online processing of data, thereby achieving the requirements of high speed, low power consumption and low hardware costs on the basis of the high degree of parallelism of the network.

Inventors:

MIAO XIANGSHUI (CN)
LI YI (CN)
PAN WENQIAN (CN)

Application Number:

PCT/CN2019/095680

Publication Date:

March 19, 2020

Filing Date:

July 12, 2019

Export Citation:

Click for automatic bibliography generation Help

Assignee:

UNIV HUAZHONG SCIENCE TECH (CN)

International Classes:

G06N3/063

Foreign References:

CN109460817A	2019-03-12
CN107742153A	2018-02-27
CN108182471A	2018-06-19
US20180018559A1	2018-01-18

Other References:

HU, FEI ET AL.: "Circuit Design of Convolutional Neural Network Based on Memristor Crossbar Arrays", JOURNAL OF COMPUTER RESEARCH AND DEVELOPMENT, 31 May 2018 (2018-05-31), pages 1098 - 1104

Attorney, Agent or Firm:

HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY PATENT AGENCY CENTER (CN)

Download PDF:

View/Download PDF PDF Help

Previous Patent: LED STREET LIGHT CAPABLE OF AUTOMATICALLY ADJUSTING ANGLE

Next Patent: SHIFT REGISTER UNIT AND DRIVE METHOD THEREFOR, GATE DRIVER CIRCUIT, AND DISPLAY DEVICE