Unlock instant, AI-driven research and patent intelligence for your innovation.

Brain-like Computing System

A computing system and brain-like technology, applied in the field of semiconductor chips, can solve the problems of rising power consumption, large area occupied by synapses, and difficulty in effectively expanding the scale of integrated chips, and achieve the effect of reducing power consumption and area.

Active Publication Date: 2020-01-03
SHANGHAI CIYU INFORMATION TECH CO LTD
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0014] The problem to be solved by the present invention is that the existing brain-inspired computing system occupies a large area of ​​synapses, it is difficult to effectively expand the scale of the integrated chip, and the power consumption increases due to leakage

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
  • Brain-like Computing System
  • Brain-like Computing System
  • Brain-like Computing System

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0046] can refer to image 3 Embodiment 1 of the present invention provides a brain-like computing system, including: more than one neuron and more than one MTJ synapse, and neurons are connected through the MTJ synapse. Specifically, more than one neuron includes neurons that input charges and neurons that output charges, each neuron that inputs charges is connected to more than one MTJ synapse, and each neuron that outputs charges is also connected to more than one MTJ synapse MTJ synapses.

[0047] The neuron structure in this embodiment is as image 3 As shown, it can contain a charge integrator and a pulse generator. The charge integrator collects the current input from the input point of the external or other neurons, and accumulates the input current. When the charge accumulation reaches the set value, The trigger pulse generator fires a pulse to each output synapse while the charge is zeroed. The structure of neurons in this embodiment will be further described late...

Embodiment 2

[0088] The brain-inspired computing system provided in Embodiment 1 can realize the function of simulating the pulse neural network. However, at present, a large number of application calculations are generally carried out using traditional neural networks. Therefore, the market needs brain-inspired systems that can adapt to application calculations under traditional neural networks. chip.

[0089] Therefore, on the basis of the first embodiment, the second embodiment also provides a dual-mode brain-inspired computing system capable of realizing both the function of simulating the pulse neural network and the function of the traditional neural network.

[0090] Such as Figure 8 As shown, the dual-mode brain-like computing system includes: more than one neuron and more than one MTJ synapse containing MTJ, the MTJ synapse includes a memory MTJ and a reference MTJ; the output of the MTJ synapse is connected to the The neurons that input charges in the brain-like computing syste...

Embodiment 3

[0107] The brain-inspired computing system provided in Embodiment 1 can realize the function of simulating a pulse neural network; while the dual-mode brain-inspired computing system provided in Embodiment 2 can realize both the function of simulating a pulse neural network and the function of a traditional neural network ; The brain-inspired computing system provided in this embodiment can realize traditional neural network functions.

[0108] The structure of the brain-inspired computing system provided in this embodiment is the same as that of the brain-inspired computing system in Embodiment 1. The difference is that in this embodiment, the normal working mode of the brain-inspired computing system is traditional neural network model.

[0109] Therefore, in this embodiment, the neuron includes a charge integrator and a pulse generator; the charge integrator is suitable for collecting charges input from the outside or input points of other neurons, and accumulating the inpu...

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 a brain-like computing system, which comprises more than one neuron and more than one MTJ (Magnetic Tunneling Junction) synapse, wherein the MTJ synapse comprises a memory MTJand a reference MTJ; an output end of the MTJ synapse is connected to neurons inputting the charges in the brain-like computing system, and an input end of the MTJ synapse is connected to neurons outputting the charges in the brain-like computing system; the output end of the MTJ synapse is set at a reference potential, the memory MTJ is suitable for receiving a first pulse from the input end ofthe MTJ synapse, the reference MTJ is suitable for receiving a second pulse from the input end of the MTJ synapse, the first pulse and the second pulse are transmitted simultaneously and have the sameshape and opposite signs; the neurons are connected through the MTJ synapses; the reference MTJs contained by all MTJ synapses are at the same prescribed state; and the neuron comprises a charge integrator, a pulse generator and a mode controller connected with the charge integrator and the pulse generator. The technical scheme of the invention can effectively reduce the area of the synapse, expand the scale of an integrated chip in the brain-like computing system and reduce the power consumption.

Description

technical field [0001] The invention relates to the field of semiconductor chips, in particular to a brain-like computing system. Background technique [0002] About Magnetic Tunnel Junction: [0003] Magnetic Tunneling Junction (MTJ, Magnetic Tunneling Junction) is a structure composed of two layers of ferromagnetic materials sandwiching a very thin layer of non-ferromagnetic insulating material, such as figure 1 shown. Its main current application is in magnetic random access memory (MRAM, Magnetic Random Access Memory) memory chips. [0004] refer to figure 1 , the lower layer of ferromagnetic material is a reference layer with a fixed magnetization direction, and the upper ferromagnetic material is a memory layer with a variable magnetization direction, and its magnetization direction can be parallel or antiparallel to the fixed magnetization layer. Due to the effect of quantum physics, the current can pass through the intermediate tunnel barrier layer, but the resis...

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): G06N3/063
CPCG06N3/063
Inventor 戴瑾
Owner SHANGHAI CIYU INFORMATION TECH CO LTD