CN108924910B - AI model updating method and related product - Google Patents
AI model updating method and related product Download PDFInfo
- Publication number
- CN108924910B CN108924910B CN201810828381.0A CN201810828381A CN108924910B CN 108924910 B CN108924910 B CN 108924910B CN 201810828381 A CN201810828381 A CN 201810828381A CN 108924910 B CN108924910 B CN 108924910B
- Authority
- CN
- China
- Prior art keywords
- model
- parameters
- updating
- parameter
- network
- 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.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W48/00—Access restriction; Network selection; Access point selection
- H04W48/16—Discovering, processing access restriction or access information
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0803—Configuration setting
- H04L41/0813—Configuration setting characterised by the conditions triggering a change of settings
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/06—Selective distribution of broadcast services, e.g. multimedia broadcast multicast service [MBMS]; Services to user groups; One-way selective calling services
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Security & Cryptography (AREA)
- Multimedia (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The embodiment of the application discloses an updating method of an AI model and a related product, wherein the method comprises the following steps: when the AI model needs to be updated, broadcasting perception information of a neighbor perception network NAN network; acquiring an access request of a neighboring device, wherein the access request comprises identification information of a neighboring network, and after the identification information is verified, allowing the neighboring device to access a NAN network; broadcasting an AI model update request in the NAN network, and receiving an AI model update response returned by the adjacent equipment through the NAN network; sending AI model parameters to the adjacent equipment through the NAN network, and receiving the AI model update parameters sent by the adjacent equipment; and updating the AI model by using the AI model updating parameters. The technical scheme provided by the application has the advantage of high user experience.
Description
Technical Field
The application relates to the technical field of electronic equipment, in particular to an updating method of an AI model and a related product.
Background
With the widespread use of electronic devices (such as mobile phones, tablet computers, and the like), the electronic devices have more and more applications and more powerful functions, and the electronic devices are developed towards diversification and personalization, and become indispensable electronic products in the life of users.
A Neighbor Awareness Network (NAN) is applied between electronic devices, where the NAN is a local area network in which NAN devices are directly or indirectly connected together, and information may directly circulate, and multiple devices are already connected to each other. The AI model, also called artificial intelligence model, processes information within the terminal based on artificial intelligence algorithms, such as performing face recognition algorithms, as well as AI assistants, etc.
The existing AI model can not be updated, so that the experience degree of a user is influenced when the AI model is low in calculation precision.
Disclosure of Invention
The embodiment of the application provides an updating method of an AI model and a related product, which can improve the experience of a user.
In a first aspect, an embodiment of the present application provides an AI model updating method, which is applied to an electronic device, and includes the following steps:
when the AI model needs to be updated, broadcasting perception information of a neighbor perception network NAN network;
acquiring an access request of a neighboring device, wherein the access request comprises identification information of a neighboring network, and after the identification information is verified, allowing the neighboring device to access a NAN network;
broadcasting an AI model update request in the NAN network, and receiving an AI model update response returned by the adjacent equipment through the NAN network;
sending an AI model parameter to a neighboring device through a NAN network, and receiving an AI model update parameter sent by the neighboring device, wherein the AI model update parameter is obtained after update calculation is completed according to the AI model parameter; and updating the AI model by using the AI model updating parameters.
In a second aspect, an electronic device is provided, the electronic device comprising:
the communication unit is used for broadcasting the perception information of the neighbor awareness network NAN network and acquiring the access request of the neighbor device when the AI model needs to be updated, wherein the access request comprises the identification information of the neighbor network;
the processing unit is used for allowing the adjacent equipment to access the NAN network after the identification information is verified;
the communication unit is used for broadcasting an AI model updating request in the NAN network, receiving an AI model updating response returned by the adjacent equipment through the NAN network, sending an AI model parameter to the adjacent equipment through the NAN network, and receiving an AI model updating parameter sent by the adjacent equipment, wherein the AI model updating parameter is a parameter obtained after updating calculation is completed according to the AI model parameter;
and the processing unit is also used for updating the AI model by using the AI model updating parameters.
In a third aspect, an embodiment of the present application provides an electronic device, including a processor, a memory, a communication interface, and one or more programs, where the one or more programs are stored in the memory and configured to be executed by the processor, and the program includes instructions for executing the steps in the first aspect of the embodiment of the present application.
In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program for electronic data exchange, where the computer program enables a computer to perform some or all of the steps described in the first aspect of the embodiment of the present application.
In a fifth aspect, embodiments of the present application provide a computer program product, where the computer program product includes a non-transitory computer-readable storage medium storing a computer program, where the computer program is operable to cause a computer to perform some or all of the steps as described in the first aspect of the embodiments of the present application. The computer program product may be a software installation package.
Therefore, according to the technical scheme, when the AI model needs to be updated, the NAN network is established, the AI model parameters are sent to the adjacent equipment after the adjacent equipment is allowed to access the NAN network, and the updated parameters are sent to the electronic equipment to execute the AI model after the adjacent equipment completes the updating of the AI model parameters.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1A is a schematic structural diagram of an electronic device according to an embodiment of the present disclosure;
fig. 1B is a network architecture diagram for implementing a neighbor awareness networking NAN access method disclosed in an embodiment of the present application;
fig. 2 is a schematic flowchart of an AI model updating method disclosed in an embodiment of the present application;
FIG. 3 is a schematic flow chart illustrating another AI model updating method disclosed in an embodiment of the present application;
fig. 4 is a schematic structural diagram of another electronic device disclosed in the embodiments of the present application;
fig. 5 is a schematic structural diagram of an electronic device of a neighbor awareness network NAN disclosed in an embodiment of the present application.
Detailed Description
In order to make the technical solutions of the present application better understood, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The terms "first," "second," and the like in the description and claims of the present application and in the above-described drawings are used for distinguishing between different objects and not for describing a particular order. Furthermore, the terms "include" and "have," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements listed, but may alternatively include other steps or elements not listed, or inherent to such process, method, article, or apparatus.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein can be combined with other embodiments.
In the method provided in the first aspect,
the AI model update request includes: a first type and a first number of layers of the AI model;
the AI model update response includes: a second type and a second number of layers of the AI model of the proximate device.
In the method provided in the first aspect,
the updating the AI model using the AI model update parameter specifically includes:
and replacing the parameters of the AI model with the updated parameters or taking the product obtained by performing multiplication operation on the parameters of the AI model and the updated parameters as the parameters of the updated AI model.
In the method provided in the first aspect,
the sending of the AI model parameter to the neighboring device through the NAN network specifically includes:
and m adjacent devices with the same first layer number and the same first type are obtained from the adjacent devices, and AI model parameters are sent to the m adjacent devices, wherein m is an integer greater than or equal to 1.
In the method provided in the first aspect,
if m is greater than or equal to 2, the sending the AI model parameters to the m neighboring devices specifically includes:
acquiring m load values of m adjacent devices, selecting the first adjacent device with the minimum load value from the m load values, and sending the AI model parameters to the first adjacent device.
In the method provided in the first aspect,
if m is greater than or equal to 2, the sending the AI model parameters to the m neighboring devices and receiving the AI model update parameters sent by the neighboring devices specifically include:
obtaining the parameter sizes of n layers of the AI model, if the parameter sizes of the n layers are three-dimensional data [ H ] [ W ] [ CI ], cutting each layer of parameters into m sub-parameters along the CI direction, distributing the m sub-parameters of each layer to m adjacent devices, receiving the m n sub-parameters returned by the m adjacent devices, splicing the m n sub-parameters along the CI direction to obtain an AI model updating parameter, wherein n is an integer larger than or equal to 2, H is the height value of the three-dimensional data, W is the width value of the three-dimensional data, and CI is the depth value of the three-dimensional data.
The electronic devices involved in the embodiments of the present application may include various handheld devices (e.g., smart phones), vehicle-mounted devices, wearable devices, computing devices or other processing devices connected to wireless modems with wireless communication functions, as well as various forms of User Equipment (UE), Mobile Stations (MS), terminal equipment (terminal device), and so on. For convenience of description, the above-mentioned devices are collectively referred to as electronic devices or terminals, and in addition, the neighboring devices also belong to one of the electronic devices or are electronic devices having a NAN function.
Referring to fig. 1A, fig. 1A is a schematic structural diagram of an electronic device according to an embodiment of the present disclosure, where the electronic device includes a control circuit and an input-output circuit, and the input-output circuit is connected to the control circuit.
The control circuitry may include, among other things, storage and processing circuitry. The storage circuit in the storage and processing circuit may be a memory, such as a hard disk drive memory, a non-volatile memory (e.g., a flash memory or other electronically programmable read only memory used to form a solid state drive, etc.), a volatile memory (e.g., a static or dynamic random access memory, etc.), etc., and the embodiments of the present application are not limited thereto. Processing circuitry in the storage and processing circuitry may be used to control the operation of the electronic device. The processing circuitry may be implemented based on one or more microprocessors, microcontrollers, digital signal processors, baseband processors, power management units, audio codec chips, application specific integrated circuits, display driver integrated circuits, and the like.
The storage and processing circuitry may be used to run software in the electronic device, such as play incoming call alert ringing application, play short message alert ringing application, play alarm alert ringing application, play media file application, Voice Over Internet Protocol (VOIP) phone call application, operating system functions, and so forth. The software may be used to perform some control operations, such as playing an incoming alert ring, playing a short message alert ring, playing an alarm alert ring, playing a media file, making a voice phone call, and performing other functions in the electronic device, and the embodiments of the present application are not limited.
The input-output circuit can be used for enabling the electronic device to input and output data, namely allowing the electronic device to receive data from the external device and allowing the electronic device to output data from the electronic device to the external device.
The input-output circuit may further include a sensor. The sensors may include ambient light sensors, optical and capacitive based infrared proximity sensors, ultrasonic sensors, touch sensors (e.g., optical based touch sensors and/or capacitive touch sensors, where the touch sensors may be part of a touch display screen or may be used independently as a touch sensor structure), acceleration sensors, gravity sensors, and other sensors, etc. The input-output circuit may further include audio components that may be used to provide audio input and output functionality for the electronic device. The audio components may also include a tone generator and other components for generating and detecting sound.
The input-output circuitry may also include one or more display screens. The display screen can comprise one or a combination of a liquid crystal display screen, an organic light emitting diode display screen, an electronic ink display screen, a plasma display screen and a display screen using other display technologies. The display screen may include an array of touch sensors (i.e., the display screen may be a touch display screen). The touch sensor may be a capacitive touch sensor formed by a transparent touch sensor electrode (e.g., an Indium Tin Oxide (ITO) electrode) array, or may be a touch sensor formed using other touch technologies, such as acoustic wave touch, pressure sensitive touch, resistive touch, optical touch, and the like, and the embodiments of the present application are not limited thereto.
The input-output circuitry may further include communications circuitry that may be used to provide the electronic device with the ability to communicate with external devices. The communication circuitry may include analog and digital input-output interface circuitry, and wireless communication circuitry based on radio frequency signals and/or optical signals. The wireless communication circuitry in the communication circuitry may include radio frequency transceiver circuitry, power amplifier circuitry, low noise amplifiers, switches, filters, and antennas. For example, the wireless communication circuitry in the communication circuitry may include circuitry to support Near Field Communication (NFC) by transmitting and receiving near field coupled electromagnetic signals. For example, the communication circuit may include a near field communication antenna and a near field communication transceiver. The communications circuitry may also include cellular telephone transceiver and antennas, wireless local area network transceiver circuitry and antennas, and so forth.
The input-output circuit may further include other input-output units. Input-output units may include buttons, joysticks, click wheels, scroll wheels, touch pads, keypads, keyboards, cameras, light emitting diodes and other status indicators, and the like.
The electronic device may further include a battery (not shown) for supplying power to the electronic device.
Referring to fig. 1B, fig. 1B is a network architecture diagram of a system for implementing a neighbor aware network NAN access method according to an embodiment of the present application, where the network architecture diagram includes an electronic device and at least one neighbor device, and specifically, the electronic device receives an access request sent by any neighbor device, where the access request carries unique identification information of the neighbor device, and the access request is used to request to access a target NAN network established by the electronic device; analyzing the access request to obtain the unique identification information; and when the unique identification information belongs to a preset identification information set, allowing the adjacent equipment to access the target NAN network. Therefore, when the electronic device establishes the NAN, the access request of the adjacent device can be received, the access request carries the unique identification information, and if the unique identification information passes the verification, the access to the NAN is allowed, so that the safe NAN can be obtained.
Based on the electronic device shown in fig. 1A and the system shown in fig. 1B, the following method for updating an AI model may be implemented, referring to fig. 2, and specifically includes the following steps:
101. and when the AI model is determined to need to be updated, broadcasting perception information of the NAN network.
The manner in which the AI model needs to be updated may be specifically that the calculation delay of the AI model exceeds the time threshold, but may also be other conditions, for example, that the use time of the AI model exceeds the set time.
In this embodiment of the present application, the sensing information may include: unique identification information of the NAN network, which may include at least one of: a telephone number, an Integrated Circuit Card Identifier (ICCID), an International Mobile Equipment Identifier (IMEI), and the like, without limitation. After receiving the broadcasted sensing information, the adjacent equipment analyzes the sensing information to obtain unique identification information, and then accesses to the NAN network according to the unique identification information. The specific access mode may be to determine whether the unique identification information belongs to preset identification information, where the preset identification information may be set by a user or default by a system. When the unique identification information belongs to the preset identification information set, the electronic device may allow the neighboring device to access the target NAN network, for example, the neighboring device may directly access the target NAN network, or the neighboring device needs to be registered, and when the registration is completed, the neighboring device may access the target NAN network.
102. And acquiring the access request of the adjacent equipment, wherein the access request comprises the identification information of the adjacent network, and after the identification information is verified, allowing the adjacent equipment to access the NAN network.
Optionally, the implementation method for verifying the identification information may specifically include:
a1, detecting a connection record between the proximity device and the electronic device;
a2, determining an intimacy degree value between the adjacent equipment and the electronic equipment according to the connection record;
a3, when the affinity value is larger than a preset threshold value, executing the step of allowing the adjacent device to access the target NAN network.
The preset threshold value can be set by the user or defaulted by the system. If the proximity device and the electronic device have been previously connected, there may be a connection record between the proximity device and the electronic device, for example, a bluetooth connection has been previously made between the electronic device and the proximity device, or alternatively, the electronic device initiates a hotspot, the proximity device accesses the hotspot, and so on. Furthermore, the electronic device may detect a connection record between the proximity device and the electronic device, and further determine an affinity value between the proximity device and the electronic device according to the connection record, where the affinity value is used to represent a previous association between the proximity device and the electronic device, and the connection record may include at least one of the following information: connection duration, connection times, connection frequency, connection mode, data transmission amount, etc., which are not limited herein. The connection duration is duration or total duration of each connection between the neighboring device and the electronic device, the connection frequency is the connection frequency between the neighboring device and the electronic device within a period of time, and the data transmission amount may include a size of a data memory in which the electronic device sends data to the neighboring device, or a size of a data memory in which the neighboring device sends data to the electronic device. Thus, the intimacy degree between the proximity device and the electronic device can be determined according to the connection record, for example, according to the number of connections, such as: the affinity value is 5 when the connection is made 5 times, and the affinity value between the neighboring device and the electronic device is determined, for example, according to the connection time length, wherein the longer the connection time length, the larger the affinity value. When the affinity value is greater than the preset threshold value, the electronic device may allow the neighboring device to access the target NAN network.
103. And broadcasting the AI model update request in the NAN network, and receiving an AI model update response returned by the adjacent equipment through the NAN network.
The AI model update request may include: a first type of AI model, a first number of layers.
The purpose of carrying these two parameters is to inform the neighboring devices of the AI model that needs to be updated and the corresponding number of layers, because the AI model update is actually a training process for the AI model, which is described below, where the AI model takes a neural network model as an example and the number of layers may be 10, but in practical applications, the number of layers may be other.
The neural network model training is divided into two stages of forward operation and reverse operation, and specifically comprises the following steps:
the layer 1 of the neural network model receives training samples, the training samples are used as input data of the layer and are calculated with weight data of the layer to obtain output results of the layer 1, the output results of the layer are used as input data of the next layer to execute calculation of the next layer, forward operation of the layer 10 is sequentially executed to obtain forward operation results, the forward operation results are used as reverse input data of the layer 10 of reverse operation, the reverse input data of the layer 10 and the weight data of the layer 10 are executed to obtain weight parameters of the layer 10, the reverse input data of the layer 10 and the forward input data of the layer 10 are calculated to obtain reverse output results of the layer 10, for the reverse operation, the output results of the layer 10 are reverse input data of the previous layer, so that the weight parameters of each layer can be obtained by executing the reverse operation of the layer 10, and the weight parameters of each layer are multiplied with the weight of the layer to obtain updated weight, and performing multiple iterations to complete the updating of the neural network model. From the above description, it can be obtained that if the number of layers of the neural network is different, the updated weight parameters are also different, and the input data are also different, so it is necessary to ensure that the number of layers of the AI model is the same.
104. Sending an AI model parameter to a neighboring device through a NAN network, and receiving an AI model update parameter sent by the neighboring device, wherein the AI model update parameter is obtained after update calculation is completed according to the AI model parameter;
105. the AI model is updated using the AI model update parameters.
The specific manner of updating the AI model by using the AI model update parameter may include:
and replacing the parameters of the AI model with the updated parameters or taking the product obtained by performing product operation on the parameters of the AI model and the updated parameters as the parameters of the updated AI model.
The technical scheme has two modes for updating the parameters of the AI model, wherein the first mode is a replacement mode, the first mode has the advantages of being direct, the calculation amount of the electronic equipment is small, but the data volume transmitted by the updating parameters of the replacement mode is large, so that the updating efficiency is influenced, the other mode is that the updating parameters are weighted values of original parameters, the product obtained by multiplying the original parameters and the weighted values is the updated parameters, the mode can reduce the data volume transmitted by the updating parameters, because the AI model has a plurality of layers, generally more than 10 layers, each layer has different parameters, if all the parameters are transmitted, the data volume is large, although the data transmission through the NAN network does not need the network flow, but the large data volume needs a long time to receive, therefore, the time delay is relatively large, and in addition, since the electronic device itself has the existing parameters of the AI model, most of the parameters of the AI model do not need to be updated by updating the AI model, that is, the weighted value of the parameter is 1, and it is not necessary to transmit the same parameter value to the electronic device at this time.
According to the technical scheme, when the AI model needs to be updated, the NAN network is established, the AI model parameters are sent to the adjacent equipment after the adjacent equipment is allowed to access the NAN network, and the updated parameters are sent to the electronic equipment to execute the AI model after the adjacent equipment completes the updating of the AI model parameters.
Optionally, as the AI model update response, it includes: a second type and a second number of layers of the AI model of the proximate device; the method for sending the AI model parameters to the neighboring device through the NAN network specifically includes:
and acquiring m adjacent devices with the same first layer number and the same first type, and sending the AI model parameters to the m adjacent devices.
The reason is that if the AI model of the neighboring devices is different from the above types, such as the types of the AI models of the neighboring devices or the number of layers, the neighboring devices cannot update the AI model, and therefore, the neighboring devices with the same type and the same number of layers need to be selected to perform the AI model update.
Optionally, the sending the AI model parameters to the m neighboring devices by the method may further include:
acquiring m load values of m adjacent devices, selecting the first adjacent device with the minimum load value from the m load values, and sending the AI model parameters to the first adjacent device.
The selection of the neighboring device with the smallest load value according to the load value can improve the updating efficiency because the computation amount is very large for the AI model updating, and the product operation number of one forward operation may reach 10 in the convolutional neural network as an example6Next, the calculation amount of the backward operation is larger than that of the forward operation, and in addition, the forward operation and the backward operation are trained to be performed many times, so that the load increase for the neighboring devices is large, so that the neighboring devices with smaller load values need to be selected to perform the AI model update.
Optionally, the sending the AI model parameters to the m neighboring devices by the method, and receiving the AI model update parameters sent by the neighboring devices may further include:
and obtaining the parameter sizes of n layers of the AI model, if the parameter sizes of the n layers are all three-dimensional data (H) (W) (CI), cutting each layer of parameters into m sub-parameters along the CI direction, distributing the m sub-parameters of each layer to m adjacent devices, receiving the m n layers of sub-parameters returned by the m adjacent devices, and splicing the m n layers of sub-parameters along the CI direction to obtain the updated parameters of the AI model. N is an integer of 2 or more, and m is an integer of 2 or more. Where H is the height value of the three-dimensional data, W is the width value of the three-dimensional data, and CI is the depth value of the three-dimensional data, it should be noted that the specific value of H, W, CI may be different for the n-layer parameters.
Referring to fig. 3, fig. 3 provides an AI model updating method, which is performed between an electronic device and a neighboring device, for convenience of description, the electronic device may be the electronic device shown in fig. 1B, the neighboring device may be the neighboring device shown in fig. 1B, and the method shown in fig. 3 may be implemented under the framework of the system shown in fig. 1B, and referring to fig. 3, the NAN-based AI assistant interaction method provided in fig. 3 includes the following steps:
s301, when the electronic equipment determines that the AI model needs to be updated, the perception information of the neighbor awareness network NAN network is broadcasted;
step S302, after receiving the perception information, the adjacent equipment sends an access request to the electronic equipment, wherein the access request comprises identification information of the adjacent network;
step S303, the electronic device acquires the access request of the adjacent device, and after the identification information is verified, the adjacent device is allowed to access the NAN network.
Step S304, broadcasting an AI model update request in the NAN network, and receiving an AI model update response returned by the neighboring device through the NAN network.
Step S305, transmitting AI model parameters to adjacent equipment through a NAN network;
step S306, the adjacent equipment trains the AI model according to the AI model parameters to obtain AI model update parameters, and sends the AI model update parameters to the electronic equipment;
step S307, the electronic device updates the AI model using the AI model update parameter.
According to the technical scheme, when the AI model needs to be updated, the NAN network is established, the AI model parameters are sent to the adjacent equipment after the adjacent equipment is allowed to access the NAN network, and the updated parameters are sent to the electronic equipment to execute the AI model after the adjacent equipment completes the updating of the AI model parameters.
Referring to fig. 4, fig. 4 is a schematic structural diagram of another electronic device disclosed in the embodiment of the present application, and as shown in the drawing, the electronic device includes a processor, a memory, a communication interface, and one or more programs, where the one or more programs are stored in the memory and configured to be executed by the processor, and the program includes instructions for performing the following steps:
when the AI model needs to be updated, broadcasting perception information of a neighbor perception network NAN network;
acquiring an access request of a neighboring device, wherein the access request comprises identification information of a neighboring network, and after the identification information is verified, allowing the neighboring device to access a NAN network;
broadcasting an AI model update request in the NAN network, and receiving an AI model update response returned by the adjacent equipment through the NAN network;
sending an AI model parameter to a neighboring device through a NAN network, and receiving an AI model update parameter sent by the neighboring device, wherein the AI model update parameter is obtained after update calculation is completed according to the AI model parameter; and updating the AI model by using the AI model updating parameters.
Optionally, the AI model update request includes: a first type and a first number of layers of the AI model;
the AI model update response includes: a second type and a second number of layers of the AI model of the proximate device.
Optionally, the updating the AI model by using the AI model update parameter specifically includes:
and replacing the parameters of the AI model with the updated parameters or taking the product obtained by performing multiplication operation on the parameters of the AI model and the updated parameters as the parameters of the updated AI model.
Optionally, the sending the AI model parameter to the neighboring device through the NAN network specifically includes:
and m adjacent devices with the same first layer number and the same first type are obtained from the adjacent devices, and AI model parameters are sent to the m adjacent devices, wherein m is an integer greater than or equal to 1.
Optionally, if m is greater than or equal to 2, the sending the AI model parameters to the m neighboring devices specifically includes:
acquiring m load values of m adjacent devices, selecting the first adjacent device with the minimum load value from the m load values, and sending the AI model parameters to the first adjacent device.
Optionally, if m is greater than or equal to 2, the sending the AI model parameters to the m neighboring devices and receiving the AI model update parameters sent by the neighboring devices specifically include:
obtaining the parameter sizes of n layers of the AI model, if the parameter sizes of the n layers are three-dimensional data [ H ] [ W ] [ CI ], cutting each layer of parameters into m sub-parameters along the CI direction, distributing the m sub-parameters of each layer to m adjacent devices, receiving the m n sub-parameters returned by the m adjacent devices, splicing the m n sub-parameters along the CI direction to obtain an AI model updating parameter, wherein n is an integer larger than or equal to 2, H is the height value of the three-dimensional data, W is the width value of the three-dimensional data, and CI is the depth value of the three-dimensional data.
The above description has introduced the solution of the embodiment of the present application mainly from the perspective of the method-side implementation process. It is understood that the electronic device comprises corresponding hardware structures and/or software modules for performing the respective functions in order to realize the above-mentioned functions. Those of skill in the art will readily appreciate that the present application is capable of hardware or a combination of hardware and computer software implementing the various illustrative elements and algorithm steps described in connection with the embodiments provided herein. Whether a function is performed as hardware or computer software drives hardware depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
In the embodiment of the present application, the electronic device may be divided into the functional units according to the method example, for example, each functional unit may be divided corresponding to each function, or two or more functions may be integrated into one processing unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit. It should be noted that the division of the unit in the embodiment of the present application is schematic, and is only a logic function division, and there may be another division manner in actual implementation.
Referring to fig. 5, fig. 5 is a schematic structural diagram of a neighbor awareness network NAN electronic device disclosed in an embodiment of the present application, which is applied to the electronic device shown in fig. 1A, and the electronic device may specifically include: a communication unit 501 and a processing unit 502,
a communication unit 501, configured to broadcast awareness information of a neighbor awareness network NAN network when it is determined that an AI model needs to be updated, and acquire an access request of a neighbor device, where the access request includes identification information of the neighbor network;
a processing unit 502, configured to allow the neighboring device to access a NAN network after verifying the identification information;
the communication unit 501 is configured to broadcast an AI model update request in a NAN network, receive an AI model update response returned by the neighboring device through the NAN network, send an AI model parameter to the neighboring device through the NAN network, and receive an AI model update parameter sent by the neighboring device, where the AI model update parameter is a parameter obtained after update calculation is completed according to the AI model parameter;
the processing unit 502 is further configured to update the AI model using the AI model update parameter.
According to the technical scheme, when the AI model needs to be updated, the NAN network is established, the AI model parameters are sent to the adjacent equipment after the adjacent equipment is allowed to access the NAN network, and the updated parameters are sent to the electronic equipment to execute the AI model after the adjacent equipment completes the updating of the AI model parameters.
Optionally, the perception information includes: the unique identification information, the accessing to the NAN network according to the awareness information specifically includes:
the processing unit 503 is specifically configured to determine whether the unique identifier information belongs to a preset identifier information set, and if it is determined that the unique identifier information belongs to the preset identifier information set, allow access to the NAN network.
Embodiments of the present application also provide a computer storage medium, wherein the computer storage medium stores a computer program for electronic data exchange, and the computer program enables a computer to execute part or all of the steps of any one of the AI model updating methods described in the above method embodiments.
Embodiments of the present application also provide a computer program product including a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform part or all of the steps of any one of the AI model updating methods as set forth in the above method embodiments.
It should be noted that, for simplicity of description, the above-mentioned method embodiments are described as a series of acts or combination of acts, but those skilled in the art will recognize that the present application is not limited by the order of acts described, as some steps may occur in other orders or concurrently depending on the application. Further, those skilled in the art should also appreciate that the embodiments described in the specification are preferred embodiments and that the acts and modules referred to are not necessarily required in this application.
In the foregoing embodiments, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus may be implemented in other manners. For example, the above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one type of division of logical functions, and there may be other divisions when actually implementing, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not implemented. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of some interfaces, devices or units, and may be an electric or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit may be implemented in the form of hardware, or may be implemented in the form of a software program module.
The integrated units, if implemented in the form of software program modules and sold or used as stand-alone products, may be stored in a computer readable memory. Based on such understanding, the technical solution of the present application may be substantially implemented or a part of or all or part of the technical solution contributing to the prior art may be embodied in the form of a software product stored in a memory, and including several instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method described in the embodiments of the present application. And the aforementioned memory comprises: various media capable of storing program codes, such as a usb disk, a read-only memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and the like.
Those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by associated hardware instructed by a program, which may be stored in a computer-readable memory, which may include: flash disk, ROM, RAM, magnetic or optical disk, and the like.
The foregoing detailed description of the embodiments of the present application has been presented to illustrate the principles and implementations of the present application, and the above description of the embodiments is only provided to help understand the method and the core concept of the present application; meanwhile, for a person skilled in the art, according to the idea of the present application, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present application.
Claims (10)
1. An AI model updating method, characterized in that the method is applied to an electronic device, and the method comprises the following steps:
when the AI model needs to be updated, broadcasting perception information of a neighbor perception network NAN network;
the method comprises the steps of obtaining an access request of adjacent equipment, wherein the access request comprises identification information of the adjacent equipment, and allowing the adjacent equipment to access a NAN network after the identification information is verified;
broadcasting an AI model update request in the NAN network, and receiving an AI model update response returned by the adjacent equipment through the NAN network;
sending an AI model parameter to a neighboring device through a NAN network, and receiving an AI model update parameter sent by the neighboring device, wherein the AI model update parameter is obtained after update calculation is completed according to the AI model parameter; and updating the AI model by using the AI model updating parameters.
2. The method of claim 1,
the AI model update request includes: a first type and a first number of layers of the AI model;
the AI model update response includes: a second type and a second number of layers of the AI model of the proximate device.
3. The method according to claim 1, wherein the updating the AI model using the AI model update parameters specifically comprises:
and replacing the parameters of the AI model with the updated parameters or taking the product obtained by performing multiplication operation on the parameters of the AI model and the updated parameters as the parameters of the updated AI model.
4. The method of claim 2, wherein the sending the AI model parameters to the neighboring device via the NAN network specifically comprises:
and m adjacent devices with the same first layer number and the same first type are obtained from the adjacent devices, and AI model parameters are sent to the m adjacent devices, wherein m is an integer greater than or equal to 1.
5. The method according to claim 4, wherein if m is greater than or equal to 2, the sending AI model parameters to the m neighboring devices specifically includes:
acquiring m load values of m adjacent devices, selecting the first adjacent device with the minimum load value from the m load values, and sending the AI model parameters to the first adjacent device.
6. The method according to claim 4, wherein if m is greater than or equal to 2, the sending the AI model parameters to the m neighboring devices, and receiving the AI model update parameters sent by the neighboring devices specifically includes:
obtaining the parameter sizes of n layers of the AI model, if the parameter sizes of the n layers are three-dimensional data [ H ] [ W ] [ CI ], cutting each layer of parameters into m sub-parameters along the CI direction, distributing the m sub-parameters of each layer to m adjacent devices, receiving the m n sub-parameters returned by the m adjacent devices, splicing the m n sub-parameters along the CI direction to obtain an AI model updating parameter, wherein n is an integer larger than or equal to 2, H is the height value of the three-dimensional data, W is the width value of the three-dimensional data, and CI is the depth value of the three-dimensional data.
7. An electronic device, characterized in that the electronic device comprises:
the communication unit is used for broadcasting the perception information of the neighbor awareness network NAN when the AI model needs to be updated, and acquiring an access request of a neighbor device, wherein the access request comprises the identification information of the neighbor device;
the processing unit is used for allowing the adjacent equipment to access the NAN network after the identification information is verified;
the communication unit is used for broadcasting an AI model updating request in the NAN network, receiving an AI model updating response returned by the adjacent equipment through the NAN network, sending an AI model parameter to the adjacent equipment through the NAN network, and receiving an AI model updating parameter sent by the adjacent equipment, wherein the AI model updating parameter is a parameter obtained after updating calculation is completed according to the AI model parameter;
and the processing unit is also used for updating the AI model by using the AI model updating parameters.
8. The electronic device of claim 7,
the AI model update request includes: a first type and a first number of layers of the AI model;
the AI model update response includes: a second type and a second number of layers of the AI model of the proximate device.
9. An electronic device comprising a processor, a memory, a communication interface, and one or more programs stored in the memory and configured to be executed by the processor, the programs comprising instructions for performing the steps in the method of any of claims 1-6.
10. A computer-readable storage medium, characterized in that a computer program for electronic data exchange is stored, wherein the computer program causes a computer to perform the method according to any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810828381.0A CN108924910B (en) | 2018-07-25 | 2018-07-25 | AI model updating method and related product |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810828381.0A CN108924910B (en) | 2018-07-25 | 2018-07-25 | AI model updating method and related product |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108924910A CN108924910A (en) | 2018-11-30 |
| CN108924910B true CN108924910B (en) | 2021-03-09 |
Family
ID=64418327
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810828381.0A Active CN108924910B (en) | 2018-07-25 | 2018-07-25 | AI model updating method and related product |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108924910B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110659047A (en) * | 2019-09-16 | 2020-01-07 | Oppo广东移动通信有限公司 | Algorithm model updating method and device, electronic equipment and computer readable storage medium |
| CN112651510A (en) * | 2019-10-12 | 2021-04-13 | 华为技术有限公司 | Model updating method, working node and model updating system |
| CN113377931A (en) * | 2020-03-09 | 2021-09-10 | 香港理工大学深圳研究院 | Language model collaborative learning method, system and terminal of interactive robot |
| EP4135442A4 (en) * | 2020-04-21 | 2023-04-26 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Communication method and related device |
| CN113747462A (en) * | 2020-05-30 | 2021-12-03 | 华为技术有限公司 | Information processing method and related equipment |
| WO2021248423A1 (en) * | 2020-06-12 | 2021-12-16 | 华为技术有限公司 | Artificial intelligence resource scheduling method and apparatus, storage medium, and chip |
| CN112527344B (en) * | 2020-11-11 | 2022-09-23 | 联想(北京)有限公司 | Collaborative updating method and device for distributed AI model and program |
| CN114765771A (en) * | 2021-01-08 | 2022-07-19 | 展讯通信（上海）有限公司 | Model updating method and device, storage medium, terminal and network side equipment |
| WO2022257068A1 (en) * | 2021-06-10 | 2022-12-15 | Oppo广东移动通信有限公司 | Control method, terminal device, and network device |
| CN116419322A (en) * | 2021-12-31 | 2023-07-11 | 维沃移动通信有限公司 | AI network information transmission method and device and communication equipment |
| CN117440449A (en) * | 2022-07-12 | 2024-01-23 | 维沃移动通信有限公司 | AI model information transmission method, device and equipment |
| CN117834427A (en) * | 2022-09-26 | 2024-04-05 | 维沃移动通信有限公司 | Method and device for updating AI model parameters and communication equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106126827A (en) * | 2016-06-28 | 2016-11-16 | 华中科技大学 | A kind of monitoring method of numerical control equipment health index |
| CN106230849A (en) * | 2016-08-22 | 2016-12-14 | 中国科学院信息工程研究所 | A kind of smart machine machine learning safety monitoring system based on user behavior |
| CN108093030A (en) * | 2017-11-29 | 2018-05-29 | 杭州古北电子科技有限公司 | A kind of artificial intelligence model dispositions method based on Cloud Server |
| CN108243216A (en) * | 2016-12-26 | 2018-07-03 | 华为技术有限公司 | Method, end side equipment, cloud side apparatus and the end cloud cooperative system of data processing |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014179529A1 (en) * | 2013-05-02 | 2014-11-06 | Consortium P, Inc. | Scalable real-time location detection based on overlapping neural networks |
-
2018
- 2018-07-25 CN CN201810828381.0A patent/CN108924910B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106126827A (en) * | 2016-06-28 | 2016-11-16 | 华中科技大学 | A kind of monitoring method of numerical control equipment health index |
| CN106230849A (en) * | 2016-08-22 | 2016-12-14 | 中国科学院信息工程研究所 | A kind of smart machine machine learning safety monitoring system based on user behavior |
| CN108243216A (en) * | 2016-12-26 | 2018-07-03 | 华为技术有限公司 | Method, end side equipment, cloud side apparatus and the end cloud cooperative system of data processing |
| CN108093030A (en) * | 2017-11-29 | 2018-05-29 | 杭州古北电子科技有限公司 | A kind of artificial intelligence model dispositions method based on Cloud Server |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108924910A (en) | 2018-11-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108924910B (en) | AI model updating method and related product | |
| CN108683444B (en) | Data transmission method and related product | |
| CN109120683B (en) | Data transmission method and related product | |
| CN108834132B (en) | Data transmission method and equipment and related medium product | |
| CN105024984A (en) | Permission setting method, apparatus and system | |
| CN110210605B (en) | Hardware operator matching method and related product | |
| CN105208458A (en) | Virtual frame display method and device | |
| CN105681161A (en) | Method, device and system for adding friend account | |
| CN110475020B (en) | Equipment control method and related product | |
| CN107317918B (en) | Parameter setting method and related product | |
| CN104954159A (en) | Network information statistics method and device | |
| CN103763112A (en) | User identity protection method and apparatus | |
| CN110876184B (en) | Carrier aggregation configuration method and related product | |
| CN109151726B (en) | Data processing method of neighbor awareness network NAN and related product | |
| CN103546632B (en) | A kind of quarter-bell ringing method for mobile terminal, device and mobile terminal | |
| CN103687063A (en) | Network connecting method, network connecting device and terminal equipment | |
| CN103491240A (en) | Alarm clock sounding method, device and mobile terminal | |
| CN104426747A (en) | Instant messaging method, terminal and system | |
| CN103200077A (en) | Method, device and system for data interaction during voice communication | |
| CN105721282A (en) | Messages display method and device | |
| CN108566311B (en) | Carrier aggregation configuration method and related equipment | |
| CN108900385B (en) | Network connection method and related product | |
| CN108966290B (en) | Network connection method and related product | |
| CN104918130A (en) | Methods for transmitting and playing multimedia information, devices and system | |
| CN106792434A (en) | A kind of focus method for building up and terminal device |
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 |