CN108764464A - Neuronal messages sending method, device and storage medium - Google Patents
Neuronal messages sending method, device and storage medium Download PDFInfo
- Publication number
- CN108764464A CN108764464A CN201810327257.6A CN201810327257A CN108764464A CN 108764464 A CN108764464 A CN 108764464A CN 201810327257 A CN201810327257 A CN 201810327257A CN 108764464 A CN108764464 A CN 108764464A
- Authority
- CN
- China
- Prior art keywords
- current neural
- pulse
- output information
- neuron
- information
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
- G06N3/063—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using electronic means
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Neurology (AREA)
- Electrotherapy Devices (AREA)
- Feedback Control In General (AREA)
Abstract
This application involves a kind of neuronal messages sending method, system and storage mediums.The method includes:Obtain front end neuron output information and Current neural member historical information;According to the front pulse neuron output information and the Current neural member historical information, Current neural metamessage is calculated;Determine that Current neural member output information, the Current neural member output information include at least two pulse spikes according to the Current neural metamessage;Export the Current neural member output information.Neural network compatibility artificial neuron and spiking neuron are enabled to using this method, and realize the mutual conversion that neuronal messages are sent in a manner of artificial neuron and sent in the form of spiking neuron.
Description
Technical field
This application involves nerual network technique fields, more particularly to a kind of neuronal messages sending method, device and deposit
Storage media.
Background technology
Flourishing for current era big data information network and Intelligent mobile equipment, produces the unstructured letter of magnanimity
Breath, sharp increase of the association to the high-effect process demand of these information.Traditional von neumann machine is above-mentioned in processing
Huge challenge of both being faced when problem.On the one hand it is that its processor and memory detach, due to use bus communication, together
Step, serial and concentration working method, when handling large complicated problem, not only high energy consumption, efficiency are low, but also towards numerical value meter
The characteristic of calculation keeps its software programming complexity when handling non-Formalization Problems high, or even cannot achieve.
Occurs reference human brain development class brain computing technique therewith.Neural network is made of a large amount of neurons, single nerve
Meta structure and behavior are fairly simple, and abundant network processes function can be but showed by definitely learning rules.But in tradition
Class brain computing technique in, the format of neuron output information is relatively simple, the development of neural network is produced certain
It restricts.
Invention content
Based on this, it is necessary to which, in traditional class brain computing technique, the format of neuron output information is relatively simple
The problem of, a kind of diversified method, apparatus of neuron output information format and storage medium are provided.
A kind of neuronal messages sending method, the method includes:
Obtain front end neuron output information and Current neural member historical information;
According to the front pulse neuron output information and the Current neural member historical information, Current neural member is calculated
Information;
Current neural member output information, the Current neural member output information packet are determined according to the Current neural metamessage
Include at least two pulse spikes;
Export the Current neural member output information.
It is described in one of the embodiments, that Current neural member output information is determined according to the Current neural metamessage,
Including:
According to the Current neural metamessage and information number correspondence, determine that number is provided in pulse;
The Current neural member output information is then exported, further includes:
The pulse spike for providing number equivalent number with the pulse by exporting, it is defeated to export the Current neural member
Go out information.
It is described in one of the embodiments, that Current neural member output information is determined according to the Current neural metamessage,
Including:
According to the Current neural metamessage and information frequency correspondence, determine that frequency is provided in pulse;
The Current neural member output information is then exported, further includes:
Rate-adaptive pacemaker pulse spike is provided by the pulse, exports the Current neural member output information.
The front end neuron output information in one of the embodiments, including:
Artificial neuron's output information or spiking neuron output information.
In one of the embodiments, after the step of exporting the Current neural member output information, the method is also
Including:
At least two pulse spike is sent one by one using impulsive neural networks routing packet, by the Current neural
First output information is sent to backend pulse neuron.
In one of the embodiments, after the step of exporting the Current neural member output information, the method is also
Including:
It route packet using artificial neural network and sends at least two pulse spike, the Current neural member is defeated
Go out information and is sent to rear end artificial neuron.
It is described in one of the embodiments, to send at least two pulse spike using artificial neural network routing packet
The Current neural member output information is sent to rear end artificial neuron by signal, including:
Obtain the pulse number and pulse frequency of the Current neural member output information;
According to the pulse number and the pulse frequency, artificial neuron's output information is determined;
It route packet using artificial neural network and sends artificial neuron's output information to rear end artificial neuron.
It is described according to the pulse number and the pulse frequency in one of the embodiments, determine artificial neuron
Output information, including:
Spiking neuron changing value is calculated according to the pulse number and/or the pulse frequency;
According to the spiking neuron changing value, the pulse number and the pulse frequency, artificial transmission coefficient is determined;
According to the artificial transmission coefficient, the pulse number and the pulse frequency determine that the artificial neuron is defeated
Go out information.
A kind of neuronal messages sending device, including:
Data obtaining module, for obtaining front end neuron output information and Current neural member historical information;
Current information computing module, for being gone through according to the front pulse neuron output information and the Current neural member
History information calculates Current neural metamessage;
Output information computing module, for determining Current neural member output information, institute according to the Current neural metamessage
It includes at least two pulse spikes to state Current neural member output information;
Message output module, for exporting the Current neural member output information.
A kind of computer equipment, including memory and processor, the memory are stored with computer program, the processing
The step of device realizes any of the above embodiment institute's providing method when executing the computer program.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
The step of any of the above embodiment institute's providing method is realized when row:
Above-mentioned neuronal messages sending method, device and storage medium.Pass through the output information of front end neuron, current god
Historical information through member and Current neural metamessage determine the output information of Current neural member, and Current neural member is defeated
It includes at least two pulse spikes to go out information.The matching of Current neural member output information and rear end neuron is realized, it is real
The diversity for having showed neuron output information format enhances the flexibility of neural network.
Description of the drawings
Fig. 1 is the flow diagram of neuronal messages sending method in one embodiment;
Fig. 2 is the module diagram of neuronal messages sending device in one embodiment;
Fig. 3 is the internal structure chart of one embodiment Computer equipment.
Specific implementation mode
It is with reference to the accompanying drawings and embodiments, right in order to make the object, technical solution and advantage of the application be more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not
For limiting the application.
In one embodiment, as shown in Figure 1, providing a kind of neuronal messages sending method, include the following steps:
Step S100 obtains front end neuron output information and Current neural member historical information.
Specifically, the front end neuron output information is the output for the front end neuron being connect with Current neural member
Information.The Current neural member historical information is included in the historical information that the front end neuron stored in Current neural member is sent.
It is appreciated that the front end neuron can be impulsive neural networks neuron (Spiking Neural Network, SNN),
It can also be the neuron (Artificial Neural Network, ANN) of artificial neural network.The impulsive neural networks
Neuron uses the algorithm based on STDP (spike-time-dependent plasticity) to build complete synaptic neural net
Network.Connection between the neuron of impulsive neural networks is realized using 1 bit quantity, and the neuron of the impulsive neural networks
The refractory period of a period of time can be entered after the activation.It should be understood that neuron is no longer sent out in refractory period, to input signal
Raw response.Within the regular hour, the frequency and pattern of pulse granting represent different information.The artificial neural network
Connection between the neuron of network is realized using more bit quantities (such as 8 bits), and does not have refractory period.
Specifically, Current neural member includes the neuron based on biological neuron Burst phenomenons.
Step S102 is calculated according to the front pulse neuron output information and the Current neural member historical information
Current neural metamessage.
Specifically, the Current neural metamessage refers to the output information according to front end neuron as Current neural member
Input information believes the output of input information, that is, front end neuron of Current neural member according to the Current neural member historical information
Breath carries out neural network computing, obtains Current neural metamessage.
Step S104 determines Current neural member output information, the Current neural member according to the Current neural metamessage
Output information includes at least two pulse spikes.
Specifically, refer to Current neural member to after being connect with Current neural member according to the Current neural member output information
The output information of terminal nerve member transmission.
Step S106 exports the Current neural member output information.
Above-mentioned neuronal messages sending method passes through the output information of front end neuron, the historical information of Current neural member
And Current neural metamessage, determine the output information of Current neural member, and the output information of Current neural member includes at least
Two pulse spikes, realize the matching of Current neural member output information and rear end neuron, realize neuron output
The diversity of information format enhances the flexibility of neural network.
It should be understood that although each step in the flow chart of Fig. 1 is shown successively according to the instruction of arrow, this
A little steps are not that the inevitable sequence indicated according to arrow executes successively.Unless expressly state otherwise herein, these steps
It executes there is no the limitation of stringent sequence, these steps can execute in other order.Moreover, at least part in Fig. 1
Step may include that either these sub-steps of multiple stages or stage are executed in synchronization to multiple sub-steps
It completes, but can execute at different times, the execution sequence in these sub-steps or stage is also not necessarily to be carried out successively,
But it can either the sub-step of other steps or at least part in stage execute in turn or alternately with other steps.
In one embodiment, Current neural member output information is determined according to the Current neural metamessage, including:According to
Current neural metamessage and information number correspondence determine that number is provided in pulse.Specifically, pulse is provided contains in number
The output information of Current neural member.The output Current neural member output information in one of the embodiments, is also wrapped
It includes:The pulse spike for providing number equivalent number with pulse by exporting, exports the Current neural member output information.Tool
Body, the output information that pulse is provided to the Current neural member for including in number are converted in pulse spike number and include
The output information of Current neural member.It is sent out it should be understood that realizing different neurons by the conversion of above- mentioned information bearing mode
Put the compatibility of mode.
In one embodiment, Current neural member output information is determined according to the Current neural metamessage, including:According to
The Current neural metamessage and information frequency correspondence determine that frequency is provided in pulse.Specifically, pulse is provided wraps in frequency
The output information of Current neural member is contained.Current neural member output information is exported in one of the embodiments, further includes:It is logical
Extra pulse provides rate-adaptive pacemaker pulse spike, exports Current neural member output information.Specifically, pulse is provided in frequency
Including Current neural member output information be converted to pulse spike transmission frequency in include Current neural member it is defeated
Go out information.It should be understood that realizing the compatibility of different neuron releasing modes by the conversion of above- mentioned information bearing mode.
In one embodiment, neuron output information in front end includes that artificial neuron's output information or spiking neuron are defeated
Go out information.In one embodiment, after the step of exporting the Current neural member output information, the method further includes:
At least two pulse spike is sent one by one using impulsive neural networks routing packet, and the Current neural member is exported and is believed
Breath is sent to backend pulse neuron.Specifically, the basic model of impulsive neural networks neuron is:For a pulse nerve
Member receives the cell space film potential raising of spiking neuron after the output information from front end neuron, when film potential is increased to
When on one threshold value, which is issued by pulse as output, shields all connect with the neuron at once later
Front end cynapse no longer receive pulse, at this time film potential fall back to resetting voltage and maintain a period of time, just receive again later
Pulse input.
In one embodiment, after the step of exporting the Current neural member output information, the method further includes:
It route packet using artificial neural network and send at least two pulse spike, the Current neural member output information is sent out
It send to rear end artificial neuron.
Specifically, the output information of front end neuron is sent to Current neural member, if rear end neuron is artificial neuron,
Then Current neural member sends the output information of Current neural member in the form of artificial neuron;If rear end neuron is pulse nerve
Member, then Current neural member the output information of Current neural member is sent in the form of spiking neuron.Specifically, artificial neural network
Human brain neuroid is abstracted from information processing angle, establishes naive model, is formed by different connection types different
Network.Neural network is a kind of operational model, by being interconnected to constitute between a large amount of neuron.Each node on behalf is a kind of
Specific output function, referred to as excitation function.
In one embodiment, described to send at least two pulse spikes letter using artificial neural network routing packet
Number, the Current neural member output information is sent to rear end artificial neuron, including:Obtain the Current neural member output letter
The pulse number and pulse frequency of breath;According to the pulse number and the pulse frequency, artificial neuron's output information is determined;
It route packet using artificial neural network and sends artificial neuron's output information to rear end artificial neuron.It should be understood that
By above-mentioned neuron sending method, the spiking neuron information of reception can be sent to artificial god by Current neural member
Through member.
In one embodiment, described according to the pulse number and the pulse frequency, determine that artificial neuron exports
Information, including:Spiking neuron changing value is calculated according to the pulse number and/or the pulse frequency;According to the pulse
Neuron changing value, the pulse number and the pulse frequency determine artificial transmission coefficient;According to the artificial transmission system
Number, the pulse number and the pulse frequency, determine artificial neuron's output information.Specifically, Current neural member is logical
The artificial transmission coefficient of adjustment is crossed, the accuracy for sending information is improved.
In one embodiment, front end neuron is spiking neuron, Current neural member receiving front-end spiking neuron hair
The information sent.Specifically, the information of spiking neuron is mutually level pulse, and different pulses can be defined as 0 or 1.Arteries and veins
The pulse granting frequency for rushing neuron will produce Current neural member different influences.Readily comprehensible, different pulse frequency
Different information is contained in time-domain.The film potential accumulation mode of Current neural member is LIT in one of the embodiments,
Model.Specifically, in LIT models, the accumulation mode of film potential is indicated with following formula:
Wherein, Vi(t) it is film potentials of the neuron i in moment t,For activation primitive, VLFor the threshold of activation primitive
The threshold value of value and membrane potential of neurons is more than the thresholded neuron and provides a spike, WijFor front end neuron j and currently
The connection weight of neuron i, TWFor time window width, δjThe time of spike is provided in current time window for front end neuron j
Point, (Δ t) is an attenuation function to K, is reduced rapidly, V as Δ increasesleakFor the film potential leakage rate of unit time.
Specifically, when the cell space film potential accumulation of Current neural member reaches threshold value, multiple pulse spikes are generated, then
Into refractory period.
After its first cell space film potential of Current neural reaches threshold value in one of the embodiments, the reading of Current neural member is worked as
The granting frequency n and granting frequency f of preceding neuron output information, and impulsive neural networks routing packet is generated with frequency f.Specifically
Ground often generates an impulsive neural networks routing packet, provides number counter and add 1, when granting number counter reaches n,
Stop generating routing packet.Current neural member is total at this time produces n impulsive neural networks routing and wraps, subsequent Current neural member into
Enter the refractory period of certain time, the film potential of Current neural member resets.
In one embodiment, front end neuron is artificial neuron, and the cell space film potential of Current neural member reaches threshold value
Afterwards, it reads the granting frequency n of its Current neural member output information and provides frequency f, and generate an artificial neuron and route packet.
The numerical value during a variable k is wrapped as routing is set, the numerical value k in the routing packet is the function of n and f:
K=n+ α f
Wherein α>0, it is meant that, if Current neural member calculates the routing packet for providing n frequency with SNN patterns as f,
The variation delta V of rear end membrane potential of neuronsSNNIt is represented by:
ΔVSNN=n (w+Vleak)
Wherein, w is the connection weight of two neurons.The change of rear end ANN membrane potential of neurons is provided in the form of ANN at this time
Change amount Δ VANN:
ΔVANN=kw=nw+ α fw
By choosing α appropriate so that Δ VANN≈ΔVSNN, ensure sending the shadow to rear end neuron in the form of ANN
Ring Approximate Equivalent.It can be seen that due to VleakPresence, provide that frequency f is bigger, and total leakage rate is smaller, and k is bigger, i.e. f and k is
Positive correlation.
In one embodiment, as shown in Fig. 2, providing a kind of neuronal messages sending device, including:Acquisition of information mould
Block 200, for obtaining front end neuron output information and Current neural member historical information;
Current information computing module 202, for according to the front pulse neuron output information and the Current neural
First historical information calculates Current neural metamessage;
Output information computing module 204, for determining Current neural member output information according to the Current neural metamessage,
The Current neural member output information includes at least two pulse spikes;
Message output module 206, for exporting the Current neural member output information.
Specific restriction about neuronal messages sending device may refer to above for neuronal messages sending method
Restriction, details are not described herein.Modules in above-mentioned neuronal messages sending device can be fully or partially through software, hard
Part and combinations thereof is realized.Above-mentioned each module can be embedded in or in the form of hardware independently of in the processor in computer equipment,
It can also in a software form be stored in the memory in computer equipment, the above modules are executed in order to which processor calls
Corresponding operation.
In one embodiment, a kind of computer equipment is provided, which can be server, internal junction
Composition can be as shown in Figure 3.The computer equipment includes processor, memory, the network interface connected by system bus.Its
In, the processor of the computer equipment is for providing calculating and control ability.The memory of the computer equipment includes non-volatile
Property storage medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program.The built-in storage is
The operation of operating system and computer program in non-volatile memory medium provides environment.The network interface of the computer equipment
For being communicated by network connection with external terminal.To realize that a kind of neuron is believed when the computer program is executed by processor
Cease sending method.
It will be understood by those skilled in the art that structure shown in Fig. 3, is only tied with the relevant part of application scheme
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme, specifically, computer equipment
May include either combining certain components than more or fewer components as shown in the figure or being arranged with different components.
In one embodiment, a kind of computer equipment, including memory and processor are provided, is stored in memory
The step of computer program, which realizes any of the above embodiment institute's providing method when executing computer program.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
The step of any of the above embodiment institute's providing method is realized when machine program is executed by processor.
One of ordinary skill in the art will appreciate that realizing all or part of flow in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the flow of the embodiment of above-mentioned each method.Wherein,
Any reference to memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
Each technical characteristic of above example can be combined arbitrarily, to keep description succinct, not to above-described embodiment
In each technical characteristic it is all possible combination be all described, as long as however, the combination of these technical characteristics be not present lance
Shield is all considered to be the range of this specification record.
The several embodiments of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, under the premise of not departing from the application design, various modifications and improvements can be made, these belong to the protection of the application
Range.Therefore, the protection domain of the application patent should be determined by the appended claims.
Claims (10)
1. a kind of neuronal messages sending method, which is characterized in that the method includes:
Obtain front end neuron output information and Current neural member historical information;
According to the front pulse neuron output information and the Current neural member historical information, Current neural member letter is calculated
Breath;
Determine that Current neural member output information, the Current neural member output information include extremely according to the Current neural metamessage
Few two pulse spikes;
Export the Current neural member output information.
2. neuronal messages sending method according to claim 1, which is characterized in that described according to Current neural member
Information determines Current neural member output information, including:
According to the Current neural metamessage and information number correspondence, determine that number is provided in pulse;
The Current neural member output information is then exported, further includes:
The pulse spike for providing number equivalent number with the pulse by exporting exports the Current neural member output letter
Breath.
3. neuronal messages sending method according to claim 1, which is characterized in that described according to Current neural member
Information determines Current neural member output information, including:
According to the Current neural metamessage and information frequency correspondence, determine that frequency is provided in pulse;
The Current neural member output information is then exported, further includes:
Rate-adaptive pacemaker pulse spike is provided by the pulse, exports the Current neural member output information.
4. neuronal messages sending method according to claim 1, which is characterized in that the front end neuron output letter
Breath, including:
Artificial neuron's output information or spiking neuron output information.
5. neuronal messages sending method according to claim 1, which is characterized in that exporting, the Current neural member is defeated
After the step of going out information, the method further includes:
At least two pulse spike is sent one by one using impulsive neural networks routing packet, the Current neural member is defeated
Go out information and is sent to backend pulse neuron.
6. neuronal messages sending method according to claim 1, which is characterized in that exporting, the Current neural member is defeated
After the step of going out information, the method further includes:
It route packet using artificial neural network and send at least two pulse spike, the Current neural member is exported and is believed
Breath is sent to rear end artificial neuron.
7. neuronal messages sending method according to claim 6, which is characterized in that described to utilize artificial neural network road
At least two pulse spike is sent by packet, the Current neural member output information is sent to rear end artificial neuron
Member, including:
Obtain the pulse number and pulse frequency of the Current neural member output information;
According to the pulse number and the pulse frequency, artificial neuron's output information is determined;
It route packet using artificial neural network and sends artificial neuron's output information to rear end artificial neuron.
8. neuronal messages sending method according to claim 7, which is characterized in that it is described according to the pulse number and
The pulse frequency determines artificial neuron's output information, including:
Spiking neuron changing value is calculated according to the pulse number and/or the pulse frequency;
According to the spiking neuron changing value, the pulse number and the pulse frequency, artificial transmission coefficient is determined;
According to the artificial transmission coefficient, the pulse number and the pulse frequency determine that the artificial neuron exports letter
Breath.
9. a kind of neuronal messages sending device, which is characterized in that including:
Data obtaining module, for obtaining front end neuron output information and Current neural member historical information;
Current information computing module, for being believed according to the front pulse neuron output information and the Current neural member history
Breath calculates Current neural metamessage;
Output information computing module, it is described to work as determining Current neural member output information according to the Current neural metamessage
Preceding neuron output information includes at least two pulse spikes;
Message output module, for exporting the Current neural member output information.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is by processor
The step of method described in any one of claim 1-8 can be realized when execution.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810327257.6A CN108764464B (en) | 2018-04-12 | 2018-04-12 | Neuron information sending method, device and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810327257.6A CN108764464B (en) | 2018-04-12 | 2018-04-12 | Neuron information sending method, device and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108764464A true CN108764464A (en) | 2018-11-06 |
CN108764464B CN108764464B (en) | 2020-10-16 |
Family
ID=63981776
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810327257.6A Active CN108764464B (en) | 2018-04-12 | 2018-04-12 | Neuron information sending method, device and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108764464B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113627599A (en) * | 2021-06-22 | 2021-11-09 | 中国科学院深圳先进技术研究院 | Processing method and processing device for neuron signals and readable storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120084241A1 (en) * | 2010-09-30 | 2012-04-05 | International Business Machines Corporation | Producing spike-timing dependent plasticity in a neuromorphic network utilizing phase change synaptic devices |
CN106845633A (en) * | 2017-01-25 | 2017-06-13 | 清华大学 | Neutral net information conversion method and system |
CN106875004A (en) * | 2017-01-20 | 2017-06-20 | 清华大学 | Composite mode neuronal messages processing method and system |
-
2018
- 2018-04-12 CN CN201810327257.6A patent/CN108764464B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120084241A1 (en) * | 2010-09-30 | 2012-04-05 | International Business Machines Corporation | Producing spike-timing dependent plasticity in a neuromorphic network utilizing phase change synaptic devices |
CN106875004A (en) * | 2017-01-20 | 2017-06-20 | 清华大学 | Composite mode neuronal messages processing method and system |
CN106845633A (en) * | 2017-01-25 | 2017-06-13 | 清华大学 | Neutral net information conversion method and system |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113627599A (en) * | 2021-06-22 | 2021-11-09 | 中国科学院深圳先进技术研究院 | Processing method and processing device for neuron signals and readable storage medium |
CN113627599B (en) * | 2021-06-22 | 2023-11-24 | 深圳微灵医疗科技有限公司 | Neuron signal processing method, processing device, and readable storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN108764464B (en) | 2020-10-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106875004B (en) | Composite mode neuronal messages processing method and system | |
KR101793011B1 (en) | Efficient hardware implementation of spiking networks | |
CN106875005B (en) | Adaptive threshold neuronal messages processing method and system | |
KR20160076520A (en) | Causal saliency time inference | |
KR20160076533A (en) | Methods and apparatus for tagging classes using supervised learning | |
CN111612125B (en) | Novel HTM time pool method and system for online learning | |
KR20160125967A (en) | Method and apparatus for efficient implementation of common neuron models | |
KR20160084401A (en) | Implementing synaptic learning using replay in spiking neural networks | |
WO2015112643A1 (en) | Monitoring neural networks with shadow networks | |
CN106845633A (en) | Neutral net information conversion method and system | |
CN106845632A (en) | Impulsive neural networks information is converted to the method and system of artificial neural network information | |
CN108401254A (en) | A kind of wireless network resource distribution method based on intensified learning | |
CN112329327B (en) | Hardware-aware liquid state machine network generation method and system | |
CN113673688A (en) | Weight generation method, data processing method and device, electronic device and medium | |
Xiao et al. | Network security situation prediction method based on MEA-BP | |
KR101825937B1 (en) | Plastic synapse management | |
Vogelstein et al. | Spike timing-dependent plasticity in the address domain | |
CN112308221A (en) | Working memory hardware implementation method based on reserve pool calculation | |
KR101782760B1 (en) | Dynamically assigning and examining synaptic delay | |
CN108764464A (en) | Neuronal messages sending method, device and storage medium | |
CN110378476A (en) | Approximate realization method, system and medium for maximum pooling layer of pulse convolution neural network | |
CN107798384B (en) | Iris florida classification method and device based on evolvable pulse neural network | |
CN113869505A (en) | Pulse neural network optimization method and system based on brain glial cells | |
CN113269113A (en) | Human behavior recognition method, electronic device, and computer-readable medium | |
CN109697511B (en) | Data reasoning method and device and computer equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |