WO2018133567A1

WO2018133567A1 - Neuron weight information processing method and system, neuron information processing method and system, and computer device

Info

Publication number: WO2018133567A1
Application number: PCT/CN2017/114659
Authority: WO
Inventors: 裴京; 邓磊; 施路平; 吴臻志; 田雷; 李国齐
Original assignee: 清华大学
Priority date: 2017-01-20
Filing date: 2017-12-05
Publication date: 2018-07-26

Abstract

A neuron weight information processing method, a neuron information processing method and system for input weight extension, and a computer device. The neuron weight information processing method comprises: receiving neuron output information about a front-end neuron and a weight index corresponding to the neuron output information (S100); reading, according to the weight index, a weight index information correlation to acquire weight information, wherein the weight index information correlation is a correlation between the weight index and the weight information (S200); and acquiring input information about the front-end neuron according to the weight information and the neuron output information (S300). Weight information is no longer directly transmitted between neurons, and an index of the weight information is transmitted, so that not only is the amount of information transmitted between networks saved on, but the setting of the weight information can also be flexibly changed.

Description

Neuron weight information processing method and system, neuron information processing method and system, and computer equipment

Related application

This application claims the application date of January 20, 2017, the application number is 201710042090.4, the name is "the input weight extension of the neuron information processing method and system", and the application number is 201710042087.2, the name is "neuron weight information processing method and system" The priority of the Chinese patent application is incorporated herein by reference in its entirety.

Technical field

The invention relates to the field of artificial neural network technology, in particular to a neuron weight information processing method and system, a neuron information processing method and system and a computer device with input weight extension.

Background technique

Most of today's artificial neural network research is still implemented in von Neumann computer software and with the high-performance GPGPU (General Purpose Graphic Processing Units) platform, the hardware overhead, energy consumption and information of the whole process. Processing speed is not optimistic. To this end, in recent years, the field of neuromorphic computing has developed rapidly, that is, the use of hardware circuits to directly construct neural networks to simulate the function of the brain, trying to achieve a massively parallel, low-energy, computing platform that can support complex mode learning.

However, in the traditional neuromorphic system, when the connection between the neurons in the neural network and the weight information are bound, the connection between each neuron needs to transmit the corresponding weight information, occupying more hardware. Processing resources, hardware constraints on network performance, and weight information reduces the processing power of neurons, limits the flexibility of neurons, and is not suitable for the hardware framework of general neural networks. Alternatively, the number of physical spaces of a single neuron weight memory is limited. When the number of input signals of a single neuron is greater than the number of physical spaces, the weight information corresponding to some input signals can only utilize the existing weights. Some neural networks that are sensitive to parameters greatly affect the application performance of neurons.

Summary of the invention

Based on this, it is necessary to provide a method and system for processing neuron weight information, a method and system for processing neuron information with input weight expansion, and a computer device.

A method and system for processing neuron weight information, the method comprising:

Receiving neuron output information of the front end neuron, and a weight index corresponding to the neuron output information;

Obtaining, according to the weight index, the correspondence information of the weight index information, and obtaining the weight information, where the correspondence relationship of the weight index information is a correspondence between the weight index and the weight information;

And inputting the input information of the front end neuron according to the weight information and the neuron output information.

In one embodiment, the front end neuron comprises an artificial neuron or a pulsed neuron.

In one embodiment, the weight index includes a storage address of weight information corresponding to the weight index.

In one embodiment, the receiving neuron output information of the front-end neuron, and the weight index corresponding to the neuron output information, include:

Receiving routing information output by the front end neuron, the routing information including neuron output information output by the front end neuron, and a weight index corresponding to the neuron output information;

Parsing the routing information, and acquiring the neuron output information and the weight index.

In one embodiment, the weight information includes:

The weight information calculated by the weight reduction algorithm is used according to the preset weight value range and the initial weight information.

In one embodiment, after the step of acquiring the input information of the front-end neuron according to the weight information and the neuron output information, the method further includes:

Calculating current neuron output information according to the neuron output algorithm according to the input information of the front end neuron and the read current neuron information;

Determining the destination information of the current neuron output information, and searching the destination index correspondence according to the destination information, and acquiring a weight index of the destination information, where the destination index correspondence includes a correspondence between the destination information and the weight index ;

The current neuron output information and the weight index are output.

In one embodiment, the output information of the front-end neuron received by the current neuron carries a weight index of the weight information between the front-end neuron and the current neuron, and the current neuron reads the weight according to the received weight index information. After the information, the corresponding arithmetic processing is performed according to different neuron models, and the input information of the front-end neurons is obtained. The weight information is no longer transmitted directly between the neurons, but the index of the weight information is transmitted, which not only saves the amount of information transmitted between the networks, but also can change the setting of the weight information of each neuron more flexibly, and improves the setting. Information processing capabilities of neural networks.

In one of the embodiments, the use of the weight index is applicable not only to the artificial neural network but also to the pulse neural network, which improves the information processing capability of the artificial neural network and the pulsed neural network.

In one embodiment, the storage address of the weight information is used as the index information, so that the neuron receiving the index information directly uses the storage address information to query the weight information, thereby improving the extraction efficiency of the weight information, thereby improving the entire neural network. Information processing efficiency.

In one of the embodiments, the weight index and the neuron output information are set in the routing information for transmission, and the existing routing data is fully utilized, thereby improving the information use efficiency between the neurons.

In one embodiment, the weight information calculated by the weight reduction algorithm is used according to a preset weight range. The weight value is within a preset value range and is a finite number of discrete values of the preset value range, which reduces the storage space of the hardware for storing the weight information, and maintains the precision of the weight information.

In one embodiment, after the current neuron calculates the output information of the current neuron, the current neuron output information and the weight index are output after the corresponding weight index is searched according to the target neuron of the output information. The current neuron sends the weight index to the neurons in the back end, and the weight index information is completely transmitted in the neural network, which improves the information processing capability of the neural network.

In the above embodiment, the output information of the front-end neuron received by the current neuron carries a weight index of the weight information between the front-end neuron and the current neuron, and the current neuron reads the weight information according to the received weight index information. According to different neuron models, corresponding operation processing is performed to obtain input information of the front-end neurons. The weight information is no longer transmitted directly between the neurons, but the index of the weight information is transmitted, which not only saves the amount of information transmitted between the networks, but also can change the setting of the weight information of each neuron more flexibly, and improves the setting. Information processing capability of neural networks

In an embodiment, a neuron information processing method for input weight expansion is further provided, and the method includes:

Determining a predetermined number of consecutive neurons as a synergistic group, determining a last neuron in the synergistic group as a valid neuron, and determining a neuron other than the effective neuron in the synergistic group as a coordinated nerve yuan;

The first coordinated neuron in the collaborative group obtains the horizontal accumulated intermediate information of the first coordinated neuron according to the received front-end neuron information;

The subsequent cooperative neurons in the collaborative group sequentially acquire the horizontal accumulation intermediate information of the cooperative neurons according to the received front-end neuron information and the lateral accumulation intermediate information of the front-end cooperative neurons, and the synergy The horizontal accumulation intermediate information of the last coordinated neuron in the group is determined as the horizontal accumulation information;

The effective neuron acquires collaborative output information according to the received front-end neuron information, the read current neuron information of the effective neuron, and the horizontal accumulation information;

The effective neurons output the coordinated output information.

In one embodiment, the determining the last neuron in the collaborative group as a valid neuron, and determining a neuron other than the effective neuron in the collaborative group as a cooperative neuron, includes:

Setting an issuance enablement identifier of a neuron in the collaborative group, the release enablement identifier includes allowing data to be issued or not allowing data to be issued, setting an issue enablement identifier of the effective neuron to allow data to be issued, and The issuance enablement identifier of the coordinated neuron is set to not allow the issuance of data.

In one embodiment, the front-end neuron information includes: front-end neuron output information, a connection weight index of the front-end neuron and the current neuron;

The first coordinated neuron in the collaborative group obtains the horizontal accumulated intermediate information of the first coordinated neuron according to the received front-end neuron information, including:

The first coordinated neuron in the collaborative group is read according to the connection weight index of the front-end neuron and the current neuron Taking the connection weight of the front-end neuron and the current neuron;

Obtaining horizontal accumulation intermediate information of the first coordinated neuron according to the connection weight of the front-end neuron and the current neuron, and the front-end neuron information;

The subsequent cooperative neurons in the collaborative group sequentially obtain the horizontal accumulation intermediate information of the cooperative neurons according to the received front-end neuron information and the horizontal accumulation intermediate information of the front-end cooperative neurons, including:

The subsequent cooperative neurons in the collaborative group sequentially index the connection weights of the front-end neurons and the current neurons according to the connection weight index of the front-end neurons and the current neurons;

And obtaining, according to the connection weight of the front-end neuron and the current neuron, the front-end neuron information, and the lateral accumulation intermediate information of the front-end cooperative neuron, the horizontal accumulation intermediate information of each coordinated neuron is obtained.

In one embodiment, the predetermined number of consecutive neurons are determined as a cooperative group, and the last neuron in the collaborative group is determined as a valid neuron, and the effective neural group is excluded from the collaborative group The neurons outside the element are identified as synergistic neurons, including:

Determining a predetermined number of consecutive artificial neurons as an artificial synergistic group, and determining a last artificial neuron in the artificial cooperative group as an artificial effective neuron, and excluding the artificial effective neurons in the artificial cooperative group Artificial neurons are identified as artificial synergistic neurons; or

Determining a predetermined number of consecutive pulse neurons as a pulse synergy group, and determining a last pulse neuron in the pulse synergy group as a pulse effective neuron, wherein the pulse synergy group is other than the pulse effective neuron The pulsed neurons are identified as pulse-synchronized neurons.

In one embodiment, when the collaborative group is a pulse cooperative group, the current neuron information includes historical membrane potential information;

After the step of outputting the collaborative output information by the effective neurons, the method further includes:

The historical membrane potential information of the pulsed effective neurons is updated.

In one embodiment, by determining a predetermined number of consecutive neurons as a cooperative group, only the information of the last neuron in the cooperative group is output, and the remaining neurons only accumulate the information and superimpose the information. The neurons to the back end are equivalent to all the neurons in the cooperative group, and the multiple inputs correspond to one valid output, which can fully utilize the weight information of the multiple inputs, breaking the existing neuron input. The shortcomings of limited weight types improve the information processing capabilities of neural networks.

In one embodiment, by setting a release enable identifier, the neuron is set to allow data to be issued or not allowed to be distributed, and a predetermined number of consecutive neurons are grouped into a coordinated group, and the cooperative group can be flexibly configured according to requirements. .

In one embodiment, the connection weight of the front-end neuron and the current neuron in the received front-end neuron information is indexed, and the connection weight of the front-end neuron and the current neuron is read, and then used to calculate the horizontal accumulation intermediate information. The weight information of each coordinated neuron in a collaborative group is fully utilized, and the cooperative output signal of the effective neuron output In the interest rate, the weight information of each cooperative neuron is embodied, which is equivalent to expanding the weight information of the effective neurons, thereby improving the information processing capability of the neural network.

In one embodiment, a predetermined number of consecutive artificial neurons are determined as an artificial cooperative group, or a predetermined number of consecutive pulse neurons are determined as a pulse cooperative group, in an artificial neural network or an impulse neural network, All can determine the synergy group, expand the input weight of a single neuron, and improve the information processing capability of the artificial neural network or the pulse neural network.

In one embodiment, in the pulse synergy group, after the pulse effective neuron outputs the cooperative output information, the historical membrane potential information of the pulsed effective neurons is updated, so that the entire cooperative group completes subsequent information processing, and the pulse The cooperative neurons do not update the historical membrane potential information, and in the subsequent information processing, the function of weight expansion is completed, and the information processing capability of the entire pulse neural network is improved by the pulse synergy group.

In the above embodiment, by determining a predetermined number of consecutive neurons as a cooperative group, only information of the last neuron in the cooperative group is output, and the remaining neurons only add information and then superimpose it. The neurons in the end group are equivalent to all the neurons in the cooperative group, and the plurality of input groups correspond to one effective output, and the weight information of the plurality of input groups can be fully utilized to break the existing single neuron input. The shortcomings of limited weight types improve the information processing capabilities of neural networks.

The present invention also provides a computer device, comprising: a memory, a processor, and a computer program stored on the memory and operable on the processor, the processor executing the computer program to implement any of the above embodiments The steps of the method.

The invention also provides a neuron weight information processing system, comprising:

a front-end neuron output information receiving module, configured to receive neuron output information of the front-end neuron, and a weight index corresponding to the neuron output information;

The weight information obtaining module is configured to: according to the weight index, read the correspondence of the weight index information, and obtain the weight information, where the weight index information correspondence relationship is a correspondence between the weight index and the weight information;

The front-end neuron input information acquiring module is configured to acquire input information of the front-end neuron according to the weight information and the neuron output information.

In one embodiment, the front-end neuron outputs an information receiving module, including:

a routing information receiving unit, configured to receive routing information output by the front end neuron, where the routing information includes neuron output information output by the front end neuron, and a weight index corresponding to the neuron output information;

The routing information parsing unit is configured to parse the routing information, and obtain the neuron output information and the weight index.

The invention also provides a neuron information processing system with input weight expansion, comprising:

a collaborative group determining module, configured to determine a preset number of consecutive neurons as a collaborative group, and determine a last neuron in the collaborative group as a valid neuron, except the effective neurons in the collaborative group The neurons are identified as synergistic neurons;

a horizontally accumulating information acquiring module, configured to acquire, according to the received front-end neuron information and the horizontally accumulated intermediate information of the front-end cooperative neurons, the horizontal accumulation of the cooperative neurons Intermediate information, and determining lateral accumulation intermediate information of the last coordinated neuron in the collaborative group as lateral accumulation information;

a cooperative output information obtaining module, configured to acquire, by the valid neuron, collaborative output information according to the received front-end neuron information, the read current neuron information of the effective neuron, and the horizontal accumulation information;

And a cooperative output information output module, configured to output the collaborative output information by the effective neurons.

DRAWINGS

In order to clearly explain the technical solutions of the present application, the drawings to be used in the embodiments will be briefly described below. Obviously, the drawings in the following description are only some of the embodiments of the present application, and those skilled in the art can obtain other drawings according to the drawings without any creative work.

1 is a schematic flow chart of a method for processing a neuron weight information according to an embodiment;

2 is a schematic flow chart of a method for processing a neuron weight information according to another embodiment;

3 is a schematic structural diagram of a neuron weight information processing system according to an embodiment;

4 is a schematic structural diagram of a neuron weight information processing system of another embodiment;

FIG. 5 is a schematic flow chart of a neuron information processing method for input weight expansion according to an embodiment; FIG.

6 is a schematic flow chart of a neuron information processing method for input weight expansion in another embodiment;

7 is a schematic structural diagram of a neuron information processing system with an input weight extension according to an embodiment;

FIG. 8 is a schematic structural diagram of a neuron information processing system for input weight expansion of another embodiment.

detailed description

The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

FIG. 1 is a schematic flowchart of a method for processing a neuron weight information according to an embodiment. The method for processing a neuron weight information as shown in FIG. 1 includes:

Step S100: Receive neuron output information of the front end neuron, and a weight index corresponding to the neuron output information.

Specifically, the neuron output information of the front-end neuron is output information calculated by the front-end neuron; The weight index is index information for the current neuron to retrieve the weight information corresponding to the front-end neuron output information.

The weight indexing method can occupy a smaller information transmission space in the process of information transmission, which not only reduces the processing requirements of the hardware, but also needs to change the index information, so that the change of the weight information can be more flexibly and conveniently performed. The update makes it easier to update the weight information in the neural network.

Step S200: The weight index is read according to the weight index, and the weight index information is obtained, and the weight index information correspondence relationship is a correspondence between the weight index and the weight information.

Specifically, the weight index information correspondence relationship may be stored locally in the current neuron or may be stored in other locations in the neural network as long as the current neuron can be read.

Step S300, acquiring input information of the front-end neuron according to the weight information and the neuron output information.

Specifically, the read weight information is subjected to corresponding operation processing according to different neuron models to obtain input information of the front end neurons.

In this embodiment, the output information of the front-end neuron received by the current neuron carries a weight index of the weight information between the front-end neuron and the current neuron, and the current neuron reads the weight information according to the received weight index information. After that, according to different neuron models, corresponding operation processing is performed to obtain input information of the front-end neurons. The weight information is no longer transmitted directly between the neurons, but the index of the weight information is transmitted, which not only saves the amount of information transmitted between the networks, but also can change the setting of the weight information of each neuron more flexibly, and improves the setting. Information processing capabilities of neural networks.

In one embodiment, the front-end neuron, including artificial neurons or pulsed neurons, ie, front-end neurons and current neurons, may be artificial neurons or pulsed neurons, ie, artificial neural networks or pulsed nerves. The meta-network can use the weight indexing method in the transmission of the weight information.

Specifically, when the neuron is an artificial neuron, the neuron output information of the front-end neuron and the weight index corresponding to the neuron output information include: membrane potential information output by the front-end artificial neuron And the connection weight index of the front end artificial neuron and the current artificial neuron; the reading the weight index information corresponding relationship according to the weight index, and obtaining the weight information, including: according to the front end artificial neuron and the current artificial neuron Connect the weight index to read the connection weight between the front-end artificial neurons and the current artificial neurons.

When the neuron is a pulsed neuron, the neuron output information of the receiving front end neuron, and the weight index corresponding to the neuron output information, include: pulse tip information output by the front end pulse neuron, front end pulse The weighting index of the connection between the neuron and the current pulsed neuron; reading the weighted index information correspondence according to the weight index, and obtaining the weight information, including: reading according to the connection weight index of the front-end pulse neuron and the current pulse neuron The connection weight of the front-end pulse neuron and the current pulsed neuron is taken.

In this embodiment, the use of the weight index is applicable not only to the artificial neural network but also to the pulse neural network, and improves the information processing capability of the artificial neural network and the pulsed neural network.

Specifically, in addition to the storage address of the weight information, the weight information may be used as the weight index, or the index information of the weight information may be obtained by using other methods according to actual requirements, and then used as the index information of the weight information.

In this embodiment, the storage address of the weight information is used as the index information, so that the neuron receiving the index information directly uses the storage address information to query the weight information, thereby improving the extraction efficiency of the weight information, thereby improving the entire neural network. Information processing efficiency.

In one embodiment, the receiving neuron output information of the front-end neuron and the weight index corresponding to the neuron output information include: receiving routing information output by the front-end neuron, the routing information including the front-end neural network Meta-outputted neuron output information, and a weight index corresponding to the neuron output information; parsing the routing information, and acquiring the neuron output information and the weight index.

Specifically, the weight index is set in the routing information, and the weight index information is transmitted by using the information transmission data in the existing neural network. In the routing information, the index information may be stored using fixed length or variable length information bits.

In this embodiment, the weight index and the neuron output information are set and transmitted in the routing information, and the existing routing data is fully utilized, thereby improving the information use efficiency between the neurons.

In one embodiment, the weight information calculated by the weight reduction algorithm is used to limit the weight to some fixed value according to the preset weight value range and the initial weight information.

Specifically, the preset weight value range is a range formed by a maximum and a minimum weight required in the network; and the weight reduction algorithm may discretize the weight under the premise of ensuring the accuracy of the algorithm. For some fixed values, such as weight binarization, binarization algorithm.

In this embodiment, according to the preset weight range, the weight information calculated by the weight reduction algorithm, the fixed value of the weight value within a preset value range, is reduced for the storage weight. The storage space of the hardware of the information, and the accuracy of the weight information is maintained.

FIG. 2 is a schematic flowchart diagram of a method for processing a neuron weight information according to another embodiment. The method for processing a neuron weight information as shown in FIG. 2 includes:

Step S400, calculating current neuron output information according to the neuron output algorithm according to the input information of the front end neuron and the read current neuron information.

Specifically, when the neuron is an artificial neuron, the current artificial neuron information includes: a current artificial nerve Meta-bias information; then, according to the input information of the front-end neuron and the read current neuron information, calculating current neuron output information according to a neuron output algorithm, including: outputting according to the front-end artificial neuron The membrane potential information, the connection weight of the front end artificial neuron and the current artificial neuron, and the current artificial neuron bias information are used to calculate the current artificial neuron output information through a preset artificial neuron activation function.

When the neuron is a pulsed neuron, the current pulsed neuron information includes: historical membrane potential information and membrane potential leakage information; then the input information according to the front-end neuron and the read current neuron information Calculating, according to the neuron output algorithm, the current neuron output information, including: according to the front end pulse neuron input information, a connection weight of the front end pulse neuron and the current pulse neuron, the historical membrane potential information, the Membrane potential leakage information, calculated by pulse neuron calculation model, current pulse neuron output information

Step S500, determining destination information of the current neuron output information, and searching for a destination index correspondence relationship according to the destination information, and acquiring a weight index of the destination information, where the destination index correspondence relationship includes between the destination information and the weight index Correspondence.

Specifically, when the task of a neural network is determined, the connection relationship between the current neuron and the back-end neuron has been determined, and the destination information of the current neuron output information has also been determined. According to the destination information of the current neuron output information, the destination index correspondence relationship is searched, and the corresponding weight index can be used.

Step S600, outputting the current neuron output information and the weight index.

Specifically, outputting the current neuron output information and the weight index to the neurons of the back end can complete the entire transfer process of the weight index.

In this embodiment, after the current neuron calculates the output information of the current neuron for output, according to the target neuron of the output information, after searching for the corresponding weight index, the current neuron output information and the weight are output. index. The current neuron sends the weight index to the neurons in the back end, and the weight index information is completely transmitted in the neural network, which improves the information processing capability of the neural network.

FIG. 3 is a schematic structural diagram of a neuron weight information processing system according to an embodiment. The neuron weight information processing system shown in FIG. 3 includes:

The front-end neuron output information receiving module 100 is configured to receive neuron output information of the front-end neuron, and a weight index corresponding to the neuron output information; the front-end neuron includes an artificial neuron or a pulsed neuron. The weight index includes a storage address of the weight information corresponding to the weight index. The front-end neuron output information receiving module includes: a routing information receiving unit, configured to receive routing information output by the front-end neuron, where the routing information includes neuron output information output by the front-end neuron, and output with the neuron A weight index corresponding to the information; a routing information parsing unit, configured to parse the routing information, and obtain the neuron output information and the weight index.

The weight information obtaining module 200 is configured to read the weight index information corresponding relationship according to the weight index, and obtain the weight information, where the weight index information correspondence relationship is a correspondence between the weight index and the weight information; the weight information, The method includes: calculating, by using a weight reduction algorithm, the weight information according to a preset weight value range and initial weight information.

The front-end neuron input information obtaining module 300 is configured to acquire input information of the front-end neuron according to the weight information and the neuron output information.

In this embodiment, the output information of the front-end neuron received by the current neuron carries a weight index of the weight information between the front-end neuron and the current neuron, and the current neuron reads the weight information according to the received weight index information. After that, according to different neuron models, corresponding operation processing is performed to obtain input information of the front-end neurons. The weight information is no longer transmitted directly between the neurons, but the index of the weight information is transmitted, which not only saves the amount of information transmitted between the networks, but also can change the setting of the weight information of each neuron more flexibly, and improves the setting. Information processing capabilities of neural networks. The use of weight index is applicable not only to artificial neural networks, but also to pulsed neural networks, which improves the information processing capabilities of artificial neural networks and pulsed neural networks. The storage address of the weight information is used as the index information, so that the neuron receiving the index information directly uses the storage address information to query the weight information, thereby improving the extraction efficiency of the weight information, thereby improving the information processing efficiency of the entire neural network. The weight index and the neuron output information are set and transmitted in the routing information, and the existing routing data is fully utilized, thereby improving the information use efficiency between the neurons. According to the preset weight range, the weight information calculated by the weight reduction algorithm has a weight value within a preset value range, which reduces the storage space of the hardware for storing the weight information, and maintains the weight. The accuracy of the information.

4 is a schematic structural diagram of a neuron weight information processing system according to another embodiment, and the neuron weight information processing system shown in FIG. 4 includes:

The front-end neuron output information receiving module 100 is configured to receive neuron output information of the front-end neuron, and a weight index corresponding to the neuron output information.

The weight information obtaining module 200 is configured to read the weight index information corresponding relationship according to the weight index, and obtain weight information, where the weight index information correspondence relationship is a correspondence between the weight index and the weight information.

The current neuron output information obtaining module 400 is configured to calculate current neuron output information according to the neuron output algorithm according to the input information of the front end neuron and the read current neuron information.

The weight index determining module 500 is configured to determine destination information of the current neuron output information, and search for a destination index correspondence according to the destination information, and obtain a weight index of the destination information, where the destination index correspondence includes the destination information. The correspondence between the weight index and the weight index.

The weight index sending module 600 is configured to output the current neuron output information and the weight index.

In this embodiment, after the current neuron calculates the output information of the current neuron for output, according to the target neuron of the output information, after searching for the corresponding weight index, the current neuron output information and the weight are output. index. The current neuron sends the weight index to the neurons at the back end, and the weight index information is completely transmitted in the neural network. Improve the information processing capabilities of neural networks.

FIG. 5 is a schematic flowchart of a method for processing neuron information of an input weight extension according to an embodiment, and the method for processing neuron information of the input weight extension shown in FIG. 1 includes:

Step S100', determining a preset number of consecutive neurons as a cooperative group, determining a last neuron in the cooperative group as a valid neuron, and selecting a neuron other than the effective neuron in the cooperative group Determined as a synergistic neuron.

Specifically, the preset quantity may be flexibly set according to the demand for the expansion of the weight stock. After setting up a collaborative group, the entire collaborative group is equivalent to a valid node. Among the consecutive neurons in the collaborative group, only the last neuron can output information, so the last neuron is determined as a valid neuron. The remaining neurons are used for cooperative work, and the weight information is expanded without outputting the data, so it is determined as a cooperative neuron, and the last effective neuron is processed for information processing.

Step S200', the first coordinated neuron in the collaborative group acquires the horizontal accumulated intermediate information of the first coordinated neuron according to the received front-end neuron information.

Specifically, the first coordinated neuron in the collaborative group obtains the horizontal accumulated intermediate information for subsequent neuron accumulation according to the received front-end neuron information, and no longer reads its own neuron information, and Then calculate the output information.

Step S300', the subsequent cooperative neurons in the collaborative group sequentially acquire the horizontal accumulation intermediate information of the cooperative neurons according to the received front-end neuron information and the lateral accumulation intermediate information of the front-end cooperative neurons, and The lateral accumulation intermediate information of the last coordinated neuron in the collaborative group is determined as lateral accumulation information.

Specifically, the subsequent cooperative neuron, after processing the received front-end neuron information, together with the lateral accumulation intermediate information of the front-end cooperative neuron, acquires the horizontal accumulation intermediate information of the current coordinated neuron, that is, within the cooperative group The cooperative neuron only processes the front-end neuron information it receives, and further accumulates the intermediate information with the lateral accumulation of other cooperative neurons in the collaborative group until the last synergistic neuron, the last synergistic neuron. The acquired lateral accumulation intermediate information is determined as lateral accumulation information for subsequent calculation of effective neurons.

Step S400', the valid neuron acquires cooperative output information according to the received front-end neuron information, the read current neuron information of the effective neuron, and the horizontal accumulation information.

Specifically, the effective neurons, the received front-end neuron information, the current neuron information, and the horizontal accumulated information calculated by all the cooperative neurons of the front end are calculated, and finally obtained for output. Collaborative output information.

In step S500', the effective neurons output the coordinated output information.

In this embodiment, by determining a predetermined number of consecutive neurons as a cooperative group, only information of the last neuron in the cooperative group is output, and the remaining neurons only add information to the synergistic group. The neurons in the back end are equivalent to all the neurons in the cooperative group to one effective node, and the multiple input groups correspond to one effective output, which can be fully utilized. The weight information of the plurality of input groups breaks the shortcoming of the limited type of input weight of the existing single neuron, and improves the information processing capability of the neural network.

In one embodiment, an issue enable identifier of a neuron in the collaborative group is set, the release enable identifier includes allowing data to be issued or not allowed to be distributed, and setting an issue enable identifier of the effective neuron to The data is allowed to be issued, and the issuance enablement identifiers of all the cooperative neurons are set to not allow the data to be issued.

Specifically, the setting is issued with an enable identifier for setting a determined preset number of consecutive neurons as a cooperative group, and setting only the last neuron to output information.

In one embodiment, a predetermined number of consecutive artificial neurons are determined as an artificial cooperative group, and the last artificial neuron in the artificial cooperative group is determined as an artificial effective neuron, and the artificial cooperative group is divided. The artificial neurons outside the artificial effective neurons are determined as artificial cooperative neurons; or a preset number of consecutive pulse neurons are determined as a pulse cooperative group, and the last pulse neuron in the pulse cooperative group is determined as a pulse An effective neuron, wherein the pulsed neurons other than the pulsed effective neurons in the pulse synergy group are determined as pulse cooperative neurons.

In this embodiment, a predetermined number of consecutive artificial neurons are determined as an artificial cooperative group, or a preset number of consecutive pulse neurons are determined as a pulse cooperative group, in an artificial neural network or a pulsed neural network, The cooperative group can be determined to expand the input weight of a single neuron, and improve the information processing capability of the artificial neural network or the pulse neural network.

When the collaborative group is a pulse cooperative group, the current neuron information includes historical membrane potential information; after the step of outputting the cooperative output information by the effective neuron, the method further includes: updating the pulse effective The historical membrane potential information of neurons.

In this embodiment, in the pulse cooperation group, after the pulse effective neurons output the coordinated output information, the historical membrane potential information of the pulsed effective neurons is updated, so that the entire collaborative group completes subsequent information processing, and the pulse synergy The neurons do not update the historical membrane potential information, and in the subsequent information processing, the function of weight expansion is completed, and the information processing capability of the entire pulse neural network is improved through the pulse synergy group.

FIG. 6 is a schematic flowchart diagram of a method for processing neuron information of an input weight extension according to another embodiment, and the method for processing neuron information of the input weight extension shown in FIG. 2 includes:

The front-end neuron information includes: front-end neuron output information, a connection weight index of the front-end neuron and the current neuron.

For the artificial collaboration group, the front-end neuron information includes: membrane potential information outputted by the front-end artificial neurons, and a connection weight index of the front-end artificial neurons and the current artificial neurons.

For the pulse synergy group, the front-end neuron information includes: pulse tip information outputted by the front-end pulse neuron, and a connection weight index of the front-end pulse neuron and the current pulse neuron.

Step S100a', determining a preset number of consecutive neurons as a cooperative group, determining a last neuron in the cooperative group as a valid neuron, and selecting a neuron other than the effective neuron in the cooperative group Determined as a synergistic neuron.

Specifically, the same step S100'.

Step S200a', the first coordinated neuron in the cooperative group reads the connection weight of the front-end neuron and the current neuron according to the connection weight index of the front-end neuron and the current neuron; according to the front-end neuron and The connection weight of the current neuron, the front-end neuron information, and the horizontal accumulation intermediate information of the first coordinated neuron.

Specifically, the connection weight index of the front end artificial neuron and the current artificial neuron is an address information, and the current neuron is indexed according to the received connection weight of the front end artificial neuron and the current artificial neuron, and the current nerve is In the memory in the element, the connection weight of the front end artificial neuron and the current artificial neuron is read, and according to the connection weight information, the output information of the front end neuron can be used in the calculation process of participating in the current neuron output information. More accurately reflects the weight of the output information of the front-end neurons.

When the collaborative group is an artificial cooperative group, the front-end neuron information includes membrane potential information output by the front-end artificial neuron, based on the membrane potential information output by the front-end artificial neuron, and the read front-end neuron and current nerve The weight of the connection of the element, after multiplication, obtains the horizontal accumulation intermediate information of the first artificial coordinated neuron and puts it into the accumulator.

When the collaborative group is a pulse cooperative group, the front-end neuron information includes pulse tip information output by the front-end pulse neuron, according to pulse tip information output by the front-end pulse neuron, and the read front-end neuron and current nerve The weight of the connection of the element, after multiplication, obtains the horizontal accumulation intermediate information of the first pulse cooperative neuron and puts it into the accumulator.

Step S300a', the subsequent cooperative neuron in the collaborative group sequentially reads the connection weight of the front-end neuron and the current neuron according to the connection weight index of the front-end neuron and the current neuron; according to the front-end neuron Obtaining intermediate information of the connection weights of the current neurons, the front-end neuron information, and the lateral accumulation of the front-end cooperative neurons, acquiring the horizontal accumulation intermediate information of the cooperative neurons, and the last synergistic nerve in the synergistic group The horizontal accumulation intermediate information of the element is determined as the horizontal accumulation information.

Specifically, the subsequent cooperative neurons in the collaborative group separately calculate the received front-end neuron output information and the connected front-end neuron and the current neuron, and calculate according to the preset neuron mode, such as After multiplying, the horizontal accumulation intermediate information of the coordinated neurons of the front end connected thereto is accumulated to obtain the horizontal accumulation intermediate information of the current coordinated neurons. Until the last coordinated neuron obtains the horizontal accumulation intermediate information, it is confirmed as the horizontal accumulation information.

Step S400a', the valid neuron acquires cooperative output information according to the received front-end neuron information, the read current neuron information of the effective neuron, and the horizontal accumulation information.

Specifically, when the collaboration group is an artificial collaboration group, the current neuron information includes current artificial neuron offset information. The effective neuron obtains collaborative output information according to the received front-end neuron information, the read current neuron information of the effective neuron, and the horizontal accumulation information, including: a membrane output according to the front-end artificial neuron Potential information, The connection weight of the front-end neuron and the current neuron, the current artificial neuron bias information, and the cooperative output information of the artificial effective neuron are calculated by a preset artificial neuron activation function.

When the collaborative group is a pulsed neuron, the current neuron information includes historical membrane potential information and membrane potential leakage information. The effective neuron obtains cooperative output information according to the received front-end neuron information, the read current neuron information of the effective neuron, and the horizontal accumulation information, including: according to the pulse output by the front-end pulse neuron The cutting-edge information, the connection weight of the front-end neuron and the current neuron, the historical membrane potential information, and the membrane potential leakage information are calculated by a pulsed neuron calculation model to calculate cooperative output information of the pulsed effective neurons.

In step S500a', the effective neurons output the coordinated output information.

In this embodiment, after the connection weight of the front-end neuron and the current neuron in the received front-end neuron information is indexed, and the connection weight of the front-end neuron and the current neuron is read, it is used to calculate the horizontal accumulation intermediate information, and The weight information of each coordinated neuron in a collaborative group is fully utilized, and in the cooperative output information of the effective neuron output, the weight information of each coordinated neuron is embodied, which is equivalent to the weight information of the effective neurons. It has been expanded to improve the information processing capabilities of neural networks.

FIG. 7 is a schematic structural diagram of an input weight-expanded neuron information processing system according to an embodiment, and the input weight-expanded neuron information processing system shown in FIG. 3 includes:

The collaborative group determining module 100' is configured to determine a preset number of consecutive neurons as a cooperative group, and determine a last neuron in the collaborative group as a valid neuron, and save the effective neural group in the cooperative group The neurons outside the element are determined as cooperative neurons; and are used for setting an issue enablement identifier of the neurons in the collaborative group, the issue enablement identifier includes allowing data to be released or not allowing data to be released, and the effective neurons are The issuance enablement flag is set to allow the release of data, and the issue enablement identification of all of the coordinated neurons is set to not allow the release of data. The method includes: an artificial neuron determining unit, configured to determine a preset number of consecutive artificial neurons as an artificial cooperative group, and determine a last artificial neuron in the artificial cooperative group as an artificial effective neuron, and the artificial synergy The artificial neurons other than the artificial effective neurons in the group are determined as artificial cooperative neurons; or the pulsed neuron determining unit is configured to determine a preset number of consecutive pulse neurons as a pulse synergy group, and the pulse The last pulsed neuron in the cooperative group is determined as a pulsed effective neuron, and the pulsed neurons other than the pulsed effective neurons in the pulse cooperative group are determined as pulse cooperative neurons.

The horizontally accumulated information obtaining module 200 ′ is configured to: acquire, by the first coordinated neuron in the collaborative group, the horizontal accumulated intermediate information of the first coordinated neuron according to the received front-end neuron information; Each of the cooperative neurons sequentially acquires the horizontal accumulation intermediate information of the cooperative neurons according to the received front-end neuron information and the lateral accumulation intermediate information of the front-end cooperative neurons, and the last synergistic nerve in the synergistic group The horizontal accumulation intermediate information of the element is determined as horizontal accumulation information, and the horizontal accumulation intermediate information of the last coordinated neuron in the cooperation group is determined as horizontal accumulation information; the front end neuron information includes: front end neuron output information, front end a weighted index of the connection between the neuron and the current neuron; the horizontal accumulation information acquisition module 200' is used to coordinate the first coordinated neuron in the group, the root According to the connection weight index of the front-end neuron and the current neuron, the connection weight of the front-end neuron and the current neuron is read; according to the connection weight of the front-end neuron and the current neuron, and the front-end neuron information, The horizontally accumulating intermediate information of the first coordinated neuron; the subsequent cooperative neuron in the cooperative group sequentially reads the connection between the front-end neuron and the current neuron according to the connection weight index of the front-end neuron and the current neuron Weighting; obtaining, according to the connection weight of the front-end neuron and the current neuron, the front-end neuron information, and the lateral accumulation intermediate information of the front-end cooperative neuron, acquiring the horizontal accumulation intermediate information of each coordinated neuron. When the collaborative group is a pulse cooperative group, the current neuron information includes historical membrane potential information.

Specifically, the horizontal accumulation information acquisition module 200 ′, when implementing the hardware circuit by using specific components, the horizontal accumulation intermediate information generated by each coordinated neuron in the cooperation group is transmitted to the next through the shared register. Synergistic neurons or effective neurons are used for membrane potential accumulation, and this way of feedback addition can be achieved with an accumulator. More specifically, the cooperative neuron obtains the horizontal accumulation intermediate information of the front-end cooperative neurons by reading the shared register. After the valid neuron outputs the information, the shared register needs to be cleared to wait for the next or next collaborative group to work properly. In the design of the neural network circuit, in order to simplify the circuit structure, the input circuit circuits of the cooperative neurons in the cooperative group and the last effective neurons may be the same, that is, the same as the effective neurons, and the synergistic neurons are also The input circuit has the function of reading the current neuron information, and the current neuron input information of each cooperative neuron is set to 0 by using the software design method.

The collaborative output information obtaining module 300' is configured to obtain the collaborative output information according to the received front-end neuron information, the read current neuron information of the effective neuron, and the horizontal accumulated information.

The collaborative output information output module 400' is configured to output the coordinated output information by the valid neurons.

In this embodiment, by determining a predetermined number of consecutive neurons as a cooperative group, only information of the last neuron in the cooperative group is output, and the remaining neurons only add information to the synergistic group. The neurons in the back end are equivalent to all the neurons in the cooperative group, and the multiple inputs correspond to one effective output. The weight information of the multiple inputs can be fully utilized, and the existing neuron input weights are broken. The shortcomings of the limited type improve the information processing capability of the neural network. By setting the release enable identifier, the neurons are set to allow data to be released or not allowed to be distributed, and a predetermined number of consecutive neurons are grouped into a coordinated group, and the cooperative group can be flexibly configured according to requirements. After receiving the weight of the connection between the front-end neuron and the current neuron in the received front-end neuron information, reading the connection weight of the front-end neuron and the current neuron, and calculating the horizontal accumulation intermediate information, each of the cooperative groups The weight information of the cooperative neurons is fully utilized, and in the cooperative output information of the effective neuron output, the weight information of each coordinated neuron is embodied, which is equivalent to expanding the weight information of the effective neurons, thereby improving The information processing capability of the neural network. Determining a preset number of consecutive artificial neurons as an artificial cooperative group, or determining a preset number of consecutive pulse neurons as a pulse synergistic group, and determining a cooperative group in an artificial neural network or a pulsed neural network The expansion of input weights of individual neurons improves the information processing capability of artificial neural networks or pulsed neural networks.

FIG. 8 is a schematic structural diagram of a neuron information processing system for input weight expansion according to another embodiment. The input weight-expanded neuron information processing system shown in FIG. 4 includes:

The collaborative group determining module 100' is configured to determine a preset number of consecutive neurons as a cooperative group, and determine a last neuron in the collaborative group as a valid neuron, and save the effective neural group in the cooperative group Neurons outside the element are identified as synergistic neurons.

The horizontally accumulated information obtaining module 200 ′ is configured to: acquire, by the first coordinated neuron in the collaborative group, the horizontal accumulated intermediate information of the first coordinated neuron according to the received front-end neuron information; Each of the cooperative neurons sequentially acquires the horizontal accumulation intermediate information of the cooperative neurons according to the received front-end neuron information and the lateral accumulation intermediate information of the front-end cooperative neurons, and the last synergistic nerve in the synergistic group The horizontal accumulation intermediate information of the element is determined as the horizontal accumulation information.

The historical membrane potential update module 500' is configured to update the historical membrane potential information of the pulsed effective neurons.

Based on the same inventive concept, an embodiment of the present invention further provides a computer device including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer The steps of the method mentioned in the above embodiments are implemented at the time of the program.

A person skilled in the art can understand that all or part of the process of implementing the foregoing embodiment can be completed by a computer program or instruction related hardware, and the program can be stored in a computer readable storage medium. When executed, the flow of an embodiment of the methods as described above may be included. Any reference to a memory, storage, database or other medium used in the various embodiments provided herein may include non-volatile and/or volatile memory. Non-volatile memory can include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of formats, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronization chain. Synchlink DRAM (SLDRAM), Memory Bus (Rambus) Direct RAM (RDRAM), Direct Memory Bus Dynamic RAM (DRDRAM), and Memory Bus Dynamic RAM (RDRAM).

The technical features of the above-described embodiments may be arbitrarily combined. For the sake of brevity of description, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction between the combinations of these technical features, All should be considered as the scope of this manual.

The above-mentioned embodiments are merely illustrative of several embodiments of the present invention, and the description thereof is more specific and detailed, but is not to be construed as limiting the scope of the invention. It should be noted that a number of variations and modifications may be made by those skilled in the art without departing from the spirit and scope of the invention. Therefore, the scope of the invention should be determined by the appended claims.

Claims

A method for processing a neuron weight information, the method comprising:

Receiving neuron output information of the front end neuron, and a weight index corresponding to the neuron output information;

Obtaining, according to the weight index, the correspondence information of the weight index information, and obtaining the weight information, where the correspondence relationship of the weight index information is a correspondence between the weight index and the weight information;

And inputting the input information of the front end neuron according to the weight information and the neuron output information.
The method for processing neuron weight information according to claim 1, wherein:

The front end neurons include artificial neurons or pulsed neurons.
The method for processing neuron weight information according to claim 1, wherein:

The weight index includes a storage address of the weight information corresponding to the weight index.
The method for processing a neuron weight information according to claim 1, wherein the receiving the neuron output information of the front-end neuron and the weight index corresponding to the neuron output information comprises:

Receiving routing information output by the front end neuron, the routing information including neuron output information output by the front end neuron, and a weight index corresponding to the neuron output information;

Parsing the routing information, and acquiring the neuron output information and the weight index.
The method for processing a neuron weight information according to claim 1, wherein the weight information comprises:

The weight information calculated by the weight reduction algorithm is used according to the preset weight value range and the initial weight information.
The method for processing a neuron weight information according to claim 1, wherein after the step of acquiring input information of the front-end neuron based on the weight information and the neuron output information, the method further includes :

Calculating current neuron output information according to the neuron output algorithm according to the input information of the front end neuron and the read current neuron information;

Determining the destination information of the current neuron output information, and searching the destination index correspondence according to the destination information, and acquiring a weight index of the destination information, where the destination index correspondence includes a correspondence between the destination information and the weight index ;

The current neuron output information and the weight index are output.
A neuron information processing method for input weight expansion, characterized in that the method comprises:

Determining a predetermined number of consecutive neurons as a synergistic group, determining a last neuron in the synergistic group as a valid neuron, and determining a neuron other than the effective neuron in the synergistic group as a coordinated nerve yuan;

The first coordinated neuron in the collaborative group obtains the horizontal accumulated intermediate information of the first coordinated neuron according to the received front-end neuron information;

The subsequent cooperative neurons in the collaborative group sequentially acquire the horizontal accumulation intermediate information of the cooperative neurons according to the received front-end neuron information and the lateral accumulation intermediate information of the front-end cooperative neurons, and the synergy The horizontal accumulation intermediate information of the last coordinated neuron in the group is determined as the horizontal accumulation information;

The valid neuron according to the received front-end neuron information, the current neuron information of the read valid neuron And the horizontally accumulating information, obtaining collaborative output information;

The effective neurons output the coordinated output information.
The method for processing a neuron information of the input weight extension according to claim 7, wherein the last neuron in the cooperative group is determined as a valid neuron, and the effective neural group is excluded from the cooperative group The neurons outside the element are identified as synergistic neurons, including:

Setting an issuance enablement identifier of a neuron in the collaborative group, the release enablement identifier includes allowing data to be issued or not allowing data to be issued, setting an issue enablement identifier of the effective neuron to allow data to be issued, and The issuance enablement identifier of the coordinated neuron is set to not allow the issuance of data.
The input weight extension neuron information processing method according to claim 7, wherein:

The front-end neuron information includes: front-end neuron output information, a connection weight index of the front-end neuron and the current neuron;

The first coordinated neuron in the collaborative group obtains the horizontal accumulated intermediate information of the first coordinated neuron according to the received front-end neuron information, including:

The first cooperative neuron in the cooperative group reads the connection weight of the front-end neuron and the current neuron according to the connection weight index of the front-end neuron and the current neuron;

Obtaining horizontal accumulation intermediate information of the first coordinated neuron according to the connection weight of the front-end neuron and the current neuron, and the front-end neuron information;

The subsequent cooperative neurons in the collaborative group sequentially obtain the horizontal accumulation intermediate information of the cooperative neurons according to the received front-end neuron information and the horizontal accumulation intermediate information of the front-end cooperative neurons, including:

The subsequent cooperative neurons in the collaborative group sequentially index the connection weights of the front-end neurons and the current neurons according to the connection weight index of the front-end neurons and the current neurons;

And obtaining, according to the connection weight of the front-end neuron and the current neuron, the front-end neuron information, and the lateral accumulation intermediate information of the front-end cooperative neuron, the horizontal accumulation intermediate information of each coordinated neuron is obtained.
The input weight extension neuron information processing method according to claim 7, wherein the predetermined number of consecutive neurons are determined as a cooperative group, and the last neuron in the cooperative group is determined to be effective. a neuron, wherein the neuron other than the effective neuron in the synergistic group is determined as a synergistic neuron, including:

Determining a predetermined number of consecutive artificial neurons as an artificial synergistic group, and determining a last artificial neuron in the artificial cooperative group as an artificial effective neuron, and excluding the artificial effective neurons in the artificial cooperative group Artificial neurons are identified as artificial synergistic neurons; or

Determining a predetermined number of consecutive pulse neurons as a pulse synergy group, and determining a last pulse neuron in the pulse synergy group as a pulse effective neuron, wherein the pulse synergy group is other than the pulse effective neuron The pulsed neurons are identified as pulse-synchronized neurons.
The input weight extension neuron information processing method according to claim 10, characterized in that:

When the collaborative group is a pulse cooperative group, the current neuron information includes historical membrane potential information;

After the step of outputting the collaborative output information by the effective neurons, the method further includes:

The historical membrane potential information of the pulsed effective neurons is updated.
A computer device, comprising: a memory, a processor, and a computer program stored on the memory and operable on the processor, the processor executing the computer program to implement any one of claims 1-11 The steps of the method.
A neuron weight information processing system, comprising:

a front-end neuron output information receiving module, configured to receive neuron output information of the front-end neuron, and a weight index corresponding to the neuron output information;

The weight information obtaining module is configured to: according to the weight index, read the correspondence of the weight index information, and obtain the weight information, where the weight index information correspondence relationship is a correspondence between the weight index and the weight information;

The front-end neuron input information acquiring module is configured to acquire input information of the front-end neuron according to the weight information and the neuron output information.
A neuron information processing system for input weight expansion, comprising:

a collaborative group determining module, configured to determine a preset number of consecutive neurons as a collaborative group, and determine a last neuron in the collaborative group as a valid neuron, except the effective neurons in the collaborative group The neurons are identified as synergistic neurons;

a horizontal accumulation information obtaining module, configured, by the first coordinated neuron in the collaborative group, acquiring, according to the received front-end neuron information, horizontal accumulated intermediate information of the first coordinated neuron; Synchronizing neurons, according to the received front-end neuron information, and the lateral accumulation intermediate information of the front-end cooperative neurons, acquiring the lateral accumulation intermediate information of the cooperative neurons, and the last synergistic neuron in the synergistic group The horizontal accumulation intermediate information is determined as the horizontal accumulation information;

a cooperative output information obtaining module, configured to acquire, by the valid neuron, collaborative output information according to the received front-end neuron information, the read current neuron information of the effective neuron, and the horizontal accumulation information;

And a cooperative output information output module, configured to output the collaborative output information by the effective neurons.