CN105046325B - A kind of circuit based on class MOS luminescent devices simulation biological neural network - Google Patents
A kind of circuit based on class MOS luminescent devices simulation biological neural network Download PDFInfo
- Publication number
- CN105046325B CN105046325B CN201510387991.8A CN201510387991A CN105046325B CN 105046325 B CN105046325 B CN 105046325B CN 201510387991 A CN201510387991 A CN 201510387991A CN 105046325 B CN105046325 B CN 105046325B
- Authority
- CN
- China
- Prior art keywords
- light
- neural network
- class
- luminescent devices
- optical signal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
A kind of circuit based on class MOS luminescent devices simulation biological neural network, belongs to semiconductor integrated circuit and biological neural network technical field.Transmission channel including more than two neural elementary cells and light, the receiving converter of light that the neural elementary cell includes being used to receive optical signal and convert optical signals to electric signal simulate the nerve synapse of biological neural network and simulate the neuron of biological neural network for the electric signal to be converted to the class MOS luminescent devices of optical signal；The transmission channel of the light is used for simulating the aixs cylinder of biological neural network；Neural elementary cell more than described two is connected by the transmission channel of light or/and Y-connection, so as to simulate biological neural network.The present invention simulation biological neural network circuit can it is compatible with CMOS technology and have the advantages that speed is fast, area is small, efficiency high, be easy to integrate, the research and development to artificial intelligence etc. brings positive meaning.
Description
Technical field
The present invention relates to semiconductor integrated circuit and biological neural network technical field, and in particular to one kind is based on class MOS
Luminescent device (i.e. the luminescent device of metalloid-insulator-semiconductor structure) simulates the circuit of biological neural network.
Background technology
Biological neural network is the nerve net that is made up of neuron, aixs cylinder, nerve synapse etc. using neuron as elementary cell
Network.Information transmission between neuron and neuron is associated by cynapse, the Nerve Terminals In The Human Skin effect of previous neuron
Cynapse is formed in cell space, dendron or aixs cylinder of next neuron etc.；One neuron can be acted on thousands of by aixs cylinder
Neuron up to ten thousand, can also be by dendron from thousands of neuron receive information.
In biological neural network and the at present method of simulation biological neural network it is attached and passes using electric signal
Pass, still, the transmission of electric signal has the problems such as noise is larger, speed is slower.In the electron-optical CMOS integrated circuits in the present age
In, class MOS luminescent devices can be counted as a preferable LMDS Light Coupled Device, and it is to be based on metalloid-insulator-semiconductor
A kind of luminescent device of structure, it includes class PMOS luminescent devices and class NMOS luminescent devices.(bibliography Kaikai Xu,
Student Member, IEEE, G.P.Li, Member, IEEE, " A Three-Terminal Silicon-PMOSFET-Like
Light-Emitting Device (LED) for Optical Intensity Modulation ", IEEE Photonics
Jounal.).Its operation principle lights similar with silicon PN junction：One class MOS luminescent device is considered as two same PN junctions two
Pole pipe is that S-B knots and D-B knots are in parallel, and the two diodes are controlled by same grid voltage, when the reversed bias voltage of two PN junctions reaches
During avalanche breakdown voltage, the source and drain of class MOS luminescent devices will light.The source and drain of class MOS luminescent devices is sent simultaneously
Photon intensity and device grid end voltage into positively related relation, i.e., with the rise of class MOS luminescent device grid voltages, it
The photon intensity that source and drain is sent is increased monotonically.Therefore, the present invention realizes biological neural network using optical signal instead of electric signal
In information transmission, it is proposed that it is a kind of based on class MOS luminescent devices simulation biological neural network circuit.
The content of the invention
The present invention proposes a kind of circuit based on class MOS luminescent devices simulation biological neural network, class MOS luminescent devices
It is a kind of luminescent device based on metalloid-insulator-semiconductor structure, it includes class PMOS luminescent devices and class NMOS
Luminescent device.The circuit of present invention simulation biological neural network can be compatible with CMOS technology and with speed is fast, area is small, efficiency
Height, it is easy to the advantages that integrated, the research and development to artificial intelligence etc. brings positive meaning.
Technical scheme is as follows：
A kind of circuit based on class MOS luminescent devices simulation biological neural network, it is characterised in that including more than two
The transmission channel of neural elementary cell and light；
The neural elementary cell includes：For receiving optical signal and the optical signal being converted to connecing for the light of electric signal
Conversion equipment is received to simulate the nerve synapse of biological neural network；The electric signal that receiving converter for receiving light exports is simultaneously
The electric signal is converted into the class MOS luminescent devices of optical signal to simulate the neuron of biological neural network；
The transmission channel of the light is used for simulating the aixs cylinder of biological neural network；
Neural elementary cell more than described two is connected by the transmission channel of light or/and Y-connection, so as to simulate
Biological neural network.
Further, neural elementary cell more than described two by transmission channel (simulation aixs cylinder) series connection of light or/
And Y-connection, so as to realize a neuron by aixs cylinder by optical signal transmission to other neurons, by cynapse from other
Neuron receives the process of optical signal, realizes simulation biological neural network.
Further, the receiving converter of the light includes photodiode and current integration circuit, the photoelectricity two
Pole pipe input connects optical signal, and output end connects the input of current integration circuit；The input of the current integration circuit
The output end of photodiode is connected, output end connects the grid end of class MOS luminescent devices.
Further, the class MOS luminescent devices grid end (input) connection current integration circuit output end, source and
Drain terminal (output end) connects the transmission channel of light.
Further, the transmission channel of the light is transparent silica, and the transmission channel of light is by class MOS luminescent devices
The optical transport that source and drain terminal are sent to other light receiving converter.
Further, the neural elementary cell includes the receiving converter and class MOS luminescent devices of light, the light
The optical signal of receiving converter connection input, output end connect the grid end of class MOS luminescent devices, the class MOS luminescent devices
Source the transmission channel of light is connected with drain terminal.
Further, the class MOS luminescent devices are that one kind based on metalloid-insulator-semiconductor structure lights
Device, including class PMOS luminescent devices and class NMOS luminescent devices.
Beneficial effects of the present invention are：The present invention replaces what is transmitted in electrical signal simulation biological neural network using optical signal
Information, there is bigger bandwidth, faster transfer rate, low crosstalk noise, ripe technique and inexpensive；And silicon
Light emitting source and other modules can be integrated on same silicon by the realization of PN junction light emitting source, compatible with CMOS technology, be not required to volume
Outer technique is supported, is easy to implement.
Brief description of the drawings
Fig. 1 is a kind of structural representation of biological neural network based on the simulation of class MOS luminescent devices provided by the invention；
Fig. 2 is a kind of structural representation of the neural elementary cell based on class MOS luminescent devices provided by the invention；
Fig. 3 a are that a kind of three-dimensional structure of neural elementary cell based on class PMOS luminescent devices provided by the invention is illustrated
Figure；
Fig. 3 b are a kind of profile of the neural elementary cell based on class PMOS luminescent devices provided by the invention；
Fig. 4 a are a kind of neural elementary cell series connection based on two class PMOS luminescent devices provided in an embodiment of the present invention
Three dimensional structure diagram；
Fig. 4 b are a kind of neural elementary cell series connection based on two class PMOS luminescent devices provided in an embodiment of the present invention
The profile of structure；
Fig. 5 a are a kind of neural elementary cell Y-connection based on class PMOS luminescent devices provided in an embodiment of the present invention
Three dimensional structure diagram；
Fig. 5 b are a kind of neural elementary cell Y-connection based on class PMOS luminescent devices provided in an embodiment of the present invention
The profile of structure；
Fig. 6 is a kind of neural elementary cell cascaded structure schematic diagram based on class PMOS luminescent devices provided by the invention；
Fig. 7 is a kind of neural elementary cell Y-connection structural representation based on class PMOS luminescent devices provided by the invention
Figure.
Embodiment
As described in background, biological neural network is using neuron as base unit, by neuron, aixs cylinder and nerve
The neutral net of the compositions such as cynapse.Information transmission between neuron and neuron is associated by cynapse, previous god
Nerve Terminals In The Human Skin through member acts on cell space, dendron or aixs cylinder of next neuron etc. and forms cynapse；One neuron can be with
Thousands of neuron is acted on by aixs cylinder, can also be by dendron from thousands of neuron receive information.
Based on this, the invention provides a kind of circuit based on class MOS luminescent devices to simulate biological neural network, with base
Exemplified by circuit described in biological neural network being simulated in the circuit of class PMOS luminescent devices, including two or more nerve elementary cell
With the transmission channel of light；The neural elementary cell includes the receiving converter and class PMOS luminescent devices of light, the light
Receiving converter is used for the nerve synapse for simulating biological neural network, receives optical signal and the optical signal is converted into telecommunications
Number；The class PMOS luminescent devices are used for the neuron for simulating biological neural network, receive the receiving converter output of light
The electric signal is simultaneously converted to optical signal by electric signal；The transmission channel of the light is used for the aixs cylinder for simulating biological neural network,
Transmit optical signal；Neural elementary cell more than described two by the transmission channel (simulation aixs cylinder) of light realize series connection or/and
Y-connection, so as to realize a neuron by aixs cylinder by optical signal transmission to other neurons, by cynapse from other god
The biological neural network transmittance process of optical signal is received through member, so as to simulate biological neural network.
Further, the neural elementary cell includes the receiving converter and class PMOS luminescent devices of light, for connecing
Receive optical signal and finally export the optical signal with optical signal after conversion, wherein, the receiving converter of light is used for mould
Intend nerve synapse, class PMOS luminescent devices are used for imictron；The neural elementary cell can pass through the transmission channel (mould of light
Intend aixs cylinder) series connection or/and Y-connection, obtain more complicated circuit, for simulating biological neural network system, realize a god
Through member by aixs cylinder by optical signal transmission to other neurons, the biological neural by cynapse from other neurons reception optical signal
Packet transmission course.In this manner it is achieved that a neuron can receive the information of other multiple neurons, while also can be by information
Multiple neurons are passed to, so as to realize the simulation of biological neural network.
Further, the receiving converter of the light includes photodiode and current integration circuit, the photoelectricity two
Pole pipe input connects optical signal, and output end connects the input of current integration circuit；The input of the current integration circuit
The output end of photodiode is connected, output end connects the grid end of class PMOS luminescent devices.The photodiode is to utilize light
Electrical effect is changed into optical signal the photoelectricity testing part of electric signal, the semiconductor devices that it is made up of a PN junction, has
One direction conductive characteristic, caused electric current is photoelectric current under the light irradiation of general illumination for it, and this current signal
With the change of light, respective change, the intensity of light are bigger, and photoelectric current is also bigger.Current integration circuit is by current signal
Integration produce the circuit of voltage signal, it a kind of resets integrator by what amplifier and integrating capacitor formed；Electric current is feeding back
Electric capacity upper integral, its gain size are determined by integrating capacitor, and when there is electric current to flow into input, sensing is produced in integrating capacitor
Electric charge, make output voltage slowly lifting, integral process starts, the stable integral voltage of final output.
Further, the class PMOS luminescent devices are the LMDS Light Coupled Devices for converting electrical signals to optical signal, its principle
Lighted with silicon PN junction similar：One class PMOS luminescent device is considered as two same PN junction diode i.e. S-B knots and D-B knots
Parallel connection, the two diodes are controlled by same grid voltage, when the reversed bias voltage of two PN junctions reaches avalanche breakdown voltage, class PMOS
The source and drain of luminescent device will light.The photon intensity and the grid of device that the source and drain of class PMOS luminescent devices is sent simultaneously
Terminal voltage is into positively related relation, i.e., with the rise of class PMOS luminescent device grid voltages, the photon intensity that source and drain is sent is dull
Increase.
Further, the transmission channel of light of the present invention is that photodiode and class PMOS luminescent devices surface cover
The silica of layer of transparent, the photodiode and class PMOS luminescent device common substrates.Silica is situated between as transmission
Matter enters the transmission of traveling optical signal, and according to the transmission characteristic of light, the transmission direction of light has arbitrariness, i.e. optical signal can be situated between in transmission
Propagated in all directions in matter, so as to realize reception converting means of the optical signal transmission of class PMOS luminescent devices to multiple light
Put, act on multiple neurons；It can also realize that the receiving converter of light receives optical signal from multiple class PMOS luminescent devices
And neuron is transferred to, simulate biological neural network.
As shown in figure 1, it is a kind of biological neural net of the breadboardin based on class PMOS luminescent devices provided by the invention
The structural representation of network, wherein, the circuit based on class PMOS luminescent devices simulation biological neural network includes：For receiving light letter
Number and by optical signal after a series of conversions finally with optical signal export neural elementary cell 10；For nerve is substantially single
The transmission for the light that the optical signal transmission of member receives optical signal to multiple neural elementary cells and to multiple neural elementary cells is led to
The aixs cylinder of biological neural network is simulated on road 20.The neural elementary cell based on class PMOS luminescent devices includes the reception of light
Conversion equipment 11 and class PMOS luminescent devices 12, for receiving optical signal and finally believing the optical signal with light after conversion
Number output.Wherein, the receiving converter analog neuron cynapse of light, class PMOS luminescent device imictrons；The neural base
This unit can obtain more complicated circuit, for simulating by transmission channel (simulation aixs cylinder) series connection or/and Y-connection of light
Biological neural network system, realize neuron by aixs cylinder by optical signal transmission to other neurons, by cynapse from other god
The biological neural network transmittance process of optical signal is received through member.
As shown in Fig. 2 the neural elementary cell based on class PMOS luminescent devices includes the receiving converter 11 and class of light
PMOS luminescent devices 12；The receiving converter of the light includes photodiode 111 and current integration circuit 112.The class
Source and the drain terminal ground connection of PMOS luminescent devices, substrate connect high voltage, and the voltage of the receiving converter output of light connects grid end, made
Its source and drain terminal light.
The course of work of circuit provided by the invention based on class PMOS luminescent devices simulation biological neural network is as follows：It is first
First, the receiving converter of light is photodiode and current integration circuit cascaded structure, and the optical signal of input is converted into pair
The voltage signal answered；Then corresponding voltage signal is inputted to the grid end of class PMOS luminescent devices, lighted as input class PMOS
The gate voltage signal of device, class PMOS luminescent devices optical signal corresponding to generation in the presence of gate voltage signal, as next base
In the input signal of the neural elementary cell of class PMOS luminescent devices, make next nerve based on class PMOS luminescent devices basic
Unit produces optical signal.So, the letter for connecting multiple neural elementary cells is used as by the optical signal propagated in silica
Number, complete the simulation to biological neural network.
Fig. 3 a, 3b are that a kind of three-dimensional structure of the neural elementary cell based on class PMOS luminescent devices provided by the invention is shown
Intention and profile；As shown in Fig. 3 a, 3b, the neural elementary cell based on class PMOS luminescent devices includes the pole of photoelectricity two
Pipe, current integration circuit and class PMOS luminescent devices.When extraneous optical signal transmission to photodiode input, the pole of photoelectricity two
Pipe converts optical signal into corresponding current signal, and current signal enters current integration circuit by the output end of photodiode
Input, current integration circuit integrate to current signal, are converted to corresponding voltage signal；Voltage signal is by current integration
After circuit output end output, as class PMOS luminescent device gate voltage signals, the source and drain of class PMOS luminescent devices is set to send necessarily
The optical signal of intensity.
Fig. 4 a, 4b are the three of a kind of neural elementary cell series connection based on two class PMOS luminescent devices provided by the invention
Tie up structural representation and profile.As shown in Fig. 4 a, 4b, extraneous optical signal inputs the photoelectricity two of first neural elementary cell
Pole pipe, photodiode convert optical signal into corresponding current signal, the electricity that current integration circuit exports photodiode
Stream signal is converted to voltage signal, grid voltage of this voltage signal as the class PMOS luminescent devices of first neural elementary cell
Input signal；For class PMOS luminescent devices after gate voltage signal is received, its source and drain sends the optical signal of some strength.First
The optical signal that individual class PMOS luminescent device source and drains are sent is transferred to second after the transmission channel L1 (silica) of light
The receiving converter of the light of individual neural elementary cell, converts optical signal into corresponding voltage signal, and this voltage signal is made
For the grid voltage input signal of the class PMOS luminescent devices of second neural elementary cell, send second class PMOS luminescent device
The optical signal of some strength.So, two neural elementary cells based on class PMOS luminescent devices are completed by optical signal
Series connection.
Fig. 5 a, 5b are the three of a kind of neural elementary cell Y-connection based on class PMOS luminescent devices provided by the invention
Tie up structural representation and profile.Pass through Y-connection structure in the biological neural network simulated based on class PMOS luminescent devices
Simulation neuron in biological nervous system can act on thousands of neuron by aixs cylinder.One nerve is basic
The neuron of unit be class PMOS luminescent devices when receiving grid end input voltage V caused by the receiving converter of light, class
PMOS luminescent devices source and drain terminal can send the optical signal of some strength.Because the direction of propagation of light has arbitrariness, class
Optical signal caused by PMOS luminescent devices can be transferred to the light of multiple neural elementary cells of others by the transmission channel of light
Receiving converter, the connection of a neural elementary cell and multiple neural elementary cells is completed by optical signal, is constituted
The structure of the Y-connection of neural elementary cell based on class PMOS luminescent devices.
In Fig. 5 a, 5b, there are the transmission channel L1 and L2 of light in the class PMOS luminescent devices both sides of a neural elementary cell,
The transmission channel L1 and L2 of light connect a neural elementary cell respectively.Class PMOS luminescent devices are changed in the reception for receiving light
After grid end input voltage V caused by device, class PMOS luminescent devices source and drain terminal send the optical signal of some strength；Optical signal
The receiving converter of the light of two other neural elementary cell is transferred to by transmission channel L1, L2 of light respectively, light is believed
The voltage signal of corresponding size number is changed into, the two voltage signals are respectively as two nerves based on class PMOS luminescent devices
The grid voltage input signal of class PMOS luminescent devices in elementary cell, make the class PMOS luminescent devices difference of two neural elementary cells
Send the optical signal of some strength.The Y-connection of neural elementary cell is so completed by optical signal.
Fig. 6 is a kind of series connection of multiple neural elementary cells based on class PMOS luminescent devices provided in an embodiment of the present invention
Structural representation.Wherein, the optical signal of each neural elementary cell output can be as the input of next neural elementary cell
Optical signal.The certain ambient light of receiving converter to the light in a neural elementary cell based on class PMOS luminescent devices
Signal, then the receiving converter of the light of the neural elementary cell can convert optical signals into voltage signal, the voltage signal make
For the grid end input voltage of the class PMOS luminescent devices in this neural elementary cell based on class PMOS luminescent devices, make class
PMOS luminescent devices source and drain terminal send optical signal；The optical signal is basic as next nerve after the transmission channel of light
The input optical signal of unit, it is transferred to the reception converting means of the light of second neural elementary cell based on class PMOS luminescent devices
Put；Optical signal caused by the class PMOS luminescent devices of second neural elementary cell based on class PMOS luminescent devices is by light
But also as the 3rd neural elementary cell input optical signal based on class PMOS luminescent devices after transmission channel.Each neural base
This unit is sequentially connected by the transmission channel of light, so completes the string of multiple neural elementary cells by the connection of optical signal
Connection.
Fig. 7 is that a kind of multiple neural elementary cell stars based on class PMOS luminescent devices provided in an embodiment of the present invention connect
The structural representation connect.Certain outer of receiving converter to the light in the neural elementary cell based on class PMOS luminescent devices
Boundary's optical signal, then the receiving converter of light can convert optical signals into voltage signal, the voltage signal is as this neural base
The grid end input voltage of class PMOS luminescent devices in this unit, makes class PMOS luminescent devices source and drain terminal send optical signal；
The reception that the optical signal can be transferred to the light of multiple neural elementary cells of others by the transmission channel of multiple different light turns
In changing device, the input optical signal as multiple neural elementary cells of others；The light of multiple neural elementary cells of others
Receiving converter can convert optical signals into voltage signal, respectively as the class PMOS in corresponding neural elementary cell
The grid end input voltage of luminescent device, multiple class PMOS luminescent devices sources and drain terminal is set to send the light letter of varying strength respectively
Number.The optical signal that such a neural elementary cell is sent can pass through light as the input signal of multiple neural elementary cells
The connection of signal completes the Y-connection of multiple neural elementary cells.
Claims (3)
1. a kind of circuit based on class MOS luminescent devices simulation biological neural network, it is characterised in that including more than two god
Transmission channel through elementary cell and light；
The neural elementary cell includes：Reception for receiving optical signal and be converted to the optical signal light of electric signal turns
Changing device simulates the nerve synapse of biological neural network；Electric signal that receiving converter for receiving light exports and by institute
State electric signal and be converted to the class MOS luminescent devices of optical signal to simulate the neuron of biological neural network；The reception of the light turns
Changing device includes photodiode and current integration circuit, and the photodiode input connects optical signal, output end connection
The input of current integration circuit, the output end of the input connection photodiode of the current integration circuit, output end connect
Connect the grid end of class MOS luminescent devices；The output end of the grid end connection current integration circuit of the class MOS luminescent devices, source and leakage
The transmission channel of end connection light；
The transmission channel of the light is used for simulating the aixs cylinder of biological neural network；
Neural elementary cell more than described two is connected by the transmission channel of light or/and Y-connection, so as to simulate biology
Neutral net.
2. the circuit according to claim 1 based on class MOS luminescent devices simulation biological neural network, it is characterised in that
Neural elementary cell more than described two is connected by the transmission channel of light or/and Y-connection, so as to realize a nerve
By optical signal transmission to other neurons, the process of optical signal, realization are received by cynapse by aixs cylinder from other neurons for member
Simulate biological neural network.
3. the circuit according to claim 1 based on class MOS luminescent devices simulation biological neural network, it is characterised in that
The transmission channel of the light is transparent silica, the light that the transmission channel of light sends class MOS luminescent devices source and drain terminal
Transmit to the receiving converter of other light.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510387991.8A CN105046325B (en) | 2015-07-06 | 2015-07-06 | A kind of circuit based on class MOS luminescent devices simulation biological neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510387991.8A CN105046325B (en) | 2015-07-06 | 2015-07-06 | A kind of circuit based on class MOS luminescent devices simulation biological neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105046325A CN105046325A (en) | 2015-11-11 |
| CN105046325B true CN105046325B (en) | 2017-12-15 |
Family
ID=54452854
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510387991.8A Expired - Fee Related CN105046325B (en) | 2015-07-06 | 2015-07-06 | A kind of circuit based on class MOS luminescent devices simulation biological neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105046325B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11853871B2 (en) | 2018-06-05 | 2023-12-26 | Lightelligence PTE. Ltd. | Optoelectronic computing systems |
| US12001946B2 (en) | 2020-04-20 | 2024-06-04 | Lightelligence PTE. Ltd. | Optoelectronic computing systems |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017210550A1 (en) | 2016-06-02 | 2017-12-07 | Massachusetts Institute Of Technology | Apparatus and methods for optical neural network |
| CN107808192B (en) * | 2017-11-28 | 2024-04-26 | 西南大学 | Transmission structure for controllable spike signal inhibition between two mutually-coupled VCSELs photonic nerves |
| JP7426099B2 (en) * | 2018-03-27 | 2024-02-01 | バー‐イラン、ユニバーシティー | Optical neural network unit and optical neural network configuration |
| SG11202011824PA (en) | 2018-06-05 | 2020-12-30 | Lightelligence Inc | Optoelectronic computing systems |
| CN109211122B (en) * | 2018-10-30 | 2020-05-15 | 清华大学 | Ultra-precise displacement measurement system and method based on optical neural network |
| CN109784486B (en) * | 2018-12-26 | 2021-04-23 | 中国科学院计算技术研究所 | Optical neural network processor and training method thereof |
| US11734556B2 (en) | 2019-01-14 | 2023-08-22 | Lightelligence PTE. Ltd. | Optoelectronic computing systems |
| CN110276440B (en) * | 2019-05-19 | 2023-03-24 | 南京惟心光电***有限公司 | Convolution operation accelerator based on photoelectric calculation array and method thereof |
| CN110991628B (en) * | 2019-11-02 | 2023-04-18 | 复旦大学 | Neuron circuit based on charge pump |
| TWI806042B (en) | 2020-04-29 | 2023-06-21 | 新加坡商光子智能私人有限公司 | Optoelectronic processing apparatus, system and method |
| US11537866B2 (en) * | 2020-05-21 | 2022-12-27 | Globalfoundries U.S. Inc. | Optical neuro-mimetic devices |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5422982A (en) * | 1991-05-02 | 1995-06-06 | Dow Corning Corporation | Neural networks containing variable resistors as synapses |
| CN101997538A (en) * | 2009-08-19 | 2011-03-30 | 中国科学院半导体研究所 | Pulse coupling based silicon-nanowire complementary metal oxide semiconductors (CMOS) neuronal circuit |
| CN102456157A (en) * | 2010-10-20 | 2012-05-16 | 北京大学 | Nerve cell apparatus and nerve network |
| CN103078054A (en) * | 2013-01-04 | 2013-05-01 | 华中科技大学 | Unit, device and method for simulating biological neuron and neuronal synapsis |
| CN103530690A (en) * | 2013-10-31 | 2014-01-22 | 中国科学院上海微***与信息技术研究所 | Nerve cell element and neural network |
-
2015
- 2015-07-06 CN CN201510387991.8A patent/CN105046325B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5422982A (en) * | 1991-05-02 | 1995-06-06 | Dow Corning Corporation | Neural networks containing variable resistors as synapses |
| CN101997538A (en) * | 2009-08-19 | 2011-03-30 | 中国科学院半导体研究所 | Pulse coupling based silicon-nanowire complementary metal oxide semiconductors (CMOS) neuronal circuit |
| CN102456157A (en) * | 2010-10-20 | 2012-05-16 | 北京大学 | Nerve cell apparatus and nerve network |
| CN103078054A (en) * | 2013-01-04 | 2013-05-01 | 华中科技大学 | Unit, device and method for simulating biological neuron and neuronal synapsis |
| CN103530690A (en) * | 2013-10-31 | 2014-01-22 | 中国科学院上海微***与信息技术研究所 | Nerve cell element and neural network |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11853871B2 (en) | 2018-06-05 | 2023-12-26 | Lightelligence PTE. Ltd. | Optoelectronic computing systems |
| US12001946B2 (en) | 2020-04-20 | 2024-06-04 | Lightelligence PTE. Ltd. | Optoelectronic computing systems |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105046325A (en) | 2015-11-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105046325B (en) | A kind of circuit based on class MOS luminescent devices simulation biological neural network | |
| CN106788442B (en) | D/A converting circuit and method, source electrode driver and display device | |
| Posch et al. | Retinomorphic event-based vision sensors: bioinspired cameras with spiking output | |
| CN102820880B (en) | Level shifter | |
| CN105099553B (en) | A kind of visible light communication method of reseptance and its system based on neuroid | |
| CN107949911A (en) | Solid-state image pickup and camera device | |
| CN104272395B (en) | Sampling apparatuses using the optics timing of time interleaving | |
| CN102769059B (en) | Bridge joint solar cell and solar power system | |
| CN106921470A (en) | A kind of visible light communication method of reseptance and its reception system based on neuroid | |
| CN106990642A (en) | The optical analog to digital conversion device demultiplexed based on modulator multichannel | |
| CN103346835B (en) | High speed visible light multiple-input multiple-output system and communication means thereof | |
| CN110263296A (en) | A kind of matrix-vector multiplier and its operation method based on photoelectricity computing array | |
| CN106449668A (en) | Photoelectric conversion element, photoelectric conversion apparatus using the same, distance detection sensor, and information processing system | |
| CN107342814A (en) | A kind of neural net equalizer based on visible light communication | |
| CN107393505A (en) | Photosensitive circuit and photosensitive circuit driving method, display device | |
| Pour et al. | Energy harvesting using substrate photodiodes | |
| Kim et al. | Photo-responsible synapse using Ge synaptic transistors and GaAs photodetectors | |
| CN110009102A (en) | A kind of accelerated method of the depth residual error network based on photoelectricity computing array | |
| CN108599771A (en) | D/A converting circuit, method and display device | |
| CN110276440A (en) | A kind of convolution algorithm accelerator and its method based on photoelectricity computing array | |
| CN110244817A (en) | A kind of Solving Partial Differential Equations device and its method based on photoelectricity computing array | |
| CN103178901A (en) | Optical fiber simulated Raman scattering effect based optical neuron and establishment method thereof | |
| CN109993283A (en) | The accelerated method of depth convolution production confrontation network based on photoelectricity computing array | |
| CN110245324A (en) | A kind of de-convolution operation accelerator and its method based on photoelectricity computing array | |
| CN204595666U (en) | Current source and array, sensing circuit and amplifying circuit |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20171215 Termination date: 20200706 |