Conflict mitigation method for artificial intelligence modules, network device, and readable storage medium

Abstract

The application provides a conflict mitigation method of an artificial intelligence module, a network device and a readable storage medium, and relates to the technical field of communication. A method executed by a first network device comprises the following steps: when the first network device needs to deploy a first artificial intelligence (AI) decision module, acquiring performance change data corresponding to all AI decision modules from a second network device; determining first information according to the performance change data; the first information is used to indicate whether the second network device trains a target AI decision module; the target AI decision module is part of the all AI decision modules; sending a training request to the second network device; the training request at least carries a first number of the first network device, a second number of the target AI decision module and the first information. According to the scheme of the application, the number and time of joint training can be reduced, and the energy consumption and deployment time can be reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication, in particular to a conflict resolution method of an artificial intelligence module, a network device and a readable storage medium. BACKGROUND

[0002] The selection or configuration of an artificial intelligence (AI) algorithm target can cause conflicting actions, and conflict resolution is necessary. There are multiple joint training methods for conflict resolution: one is to train multiple agents together, that is, to merge the input state and the output action, and to use one neural network for training. When there are many agents, the state and action scale are too large, which can cause an increase in training time and a decline in effect. Another is multi-agent distributed training, which requires a large amount of interaction between agents, also requiring a large amount of training time and overhead. SUMMARY

[0003] The purpose of the present application is to provide a conflict resolution method of an artificial intelligence module, a network device and a readable storage medium, to solve the number and range of joint training or distributed training in related technologies, to reduce energy consumption and deployment time.

[0004] To achieve the above-mentioned purpose, an embodiment of the present application provides a conflict resolution method of an artificial intelligence module, executed by a first network device, comprising:

[0005] When the first network device needs to deploy a first artificial intelligence (AI) decision module, obtaining performance change data corresponding to all AI decision modules from a second network device;

[0006] According to the performance change data, determining first information; the first information is used to indicate whether the second network device trains a target AI decision module; the target AI decision module is part of the AI decision modules;

[0007] Sending a training request to the second network device; the training request at least carries a first number of the first network device, a second number of the target AI decision module and the first information.

[0008] Optionally, according to the performance change data, determining the first information comprises:

[0009] According to the performance change data, determining whether each performance parameter corresponding to the performance change data is less than a corresponding first preset range;

[0010] If each of the performance parameters is greater than or equal to the corresponding first preset range, it is determined that the first information carries a first parameter, and a first AI decision module is deployed; the first parameter is used to indicate that the second network device does not perform training.

[0011] If part of the performance parameters is less than the corresponding first preset range, it is determined that the first information carries a second parameter; the second parameter is used to indicate that the second network device performs hierarchical training on the target AI decision module; the hierarchical training is to divide all AI decision modules into a set, and to perform joint training on the set of target AI decision modules after the division.

[0012] Optionally, the method further includes:

[0013] When each of the performance change data is less than the corresponding first preset range, it is determined that the first information carries a third parameter; the third parameter is used to indicate that the attribute information corresponding to the first AI decision module is deleted in the first network device and / or the second network device.

[0014] Optionally, when the first information carries the second parameter, the method further includes:

[0015] Receiving performance result data sent by the second network device after performing hierarchical training on the target AI decision module;

[0016] According to the performance result data, determining a decision message of the first network device; the decision message carries the first number, the second number and second information; the second information is used to indicate whether to deploy the first AI decision module in the first network device;

[0017] Sending the decision message to the second network device.

[0018] Optionally, after sending the decision message to the second network device, the method further includes:

[0019] Receiving target parameter data corresponding to all AI decision modules after the second network device performs hierarchical training on the target AI decision module;

[0020] According to all of the target parameter data, modifying all AI decision modules of the first network device, and deploying the first AI decision module;

[0021] Issuing the first AI decision module to a decision execution body.

[0022] Optionally, when the first network device needs to deploy a first artificial intelligence (AI) decision module, performance change data corresponding to all AI decision modules is acquired from a second network device, including:

[0023] A verification message for verifying the first AI decision module is sent to the second network device, and the verification message carries at least the first number and a number corresponding to the first AI decision module;

[0024] A verification response message sent by the second network device is received, and the verification response message carries performance change data corresponding to all AI decision modules.

[0025] To achieve the above purpose, an embodiment of the present application further provides a conflict resolution method of an artificial intelligence module, executed by a second network device, including:

[0026] Performance change data of all artificial intelligence (AI) decision modules is sent to a first network device;

[0027] A training request sent by the first network device is received, and the training request carries at least a first number of the first network device, a second number of a target AI decision module, and first information; the first information is used to indicate whether the second network device trains the target AI decision module; the target AI decision module is part of the all AI decision modules;

[0028] According to the training request, an operation corresponding to the first information is performed.

[0029] Optionally, the method further includes:

[0030] A decision message sent by the first network device is received, and the decision message carries the first number, the second number, and second information; the second information is used to indicate whether a first AI decision module is deployed in the first network device;

[0031] When a fourth parameter is carried in the second information, attribute information corresponding to the first AI decision module is deleted in a running environment of the second network device; the fourth parameter is used to indicate that the first AI decision module is not deployed in the first network device;

[0032] When a fifth parameter is carried in the second information, the second network device performs hierarchical training on the target AI decision module; the fifth parameter is used to indicate that the first AI decision module is deployed in the first network device; the hierarchical training is to divide all AI decision modules into sets, and to perform joint training on the target AI decision module sets after the division;

[0033] Send the target parameter data corresponding to all AI decision modules after hierarchical training to the first network device.

[0034] Optionally, the method further includes:

[0035] Determine attribute information of the first AI decision module; the attribute information includes priority of the first AI decision module and performance threshold constraint parameter;

[0036] According to the priority of the plurality of AI decision modules, divide the plurality of AI decision modules into a first decision module set and a second decision module set; the first decision module set is used to represent all decision modules whose priority is not higher than the first AI decision module; the second decision module set is used to represent all decision modules whose priority is higher than the first AI decision module;

[0037] According to the performance threshold constraint parameter corresponding to each AI decision module in the second decision module set, jointly train the first decision module set, and determine a target parameter that causes a change in performance of the first decision module set during training;

[0038] When the target parameter meets a preset threshold range, complete hierarchical training of the target AI decision module.

[0039] To achieve the above object, embodiments of the present application also provide a network device, which is a first network device, comprising:

[0040] A first obtaining module is configured to, when the first network device needs to deploy a first artificial intelligence (AI) decision module, obtain performance change data corresponding to all AI decision modules from a second network device;

[0041] A first determining module is configured to determine first information according to the performance change data; the first information is used to indicate whether the second network device trains a target AI decision module; the target AI decision module is part of the all AI decision modules;

[0042] A first sending module is configured to send a training request to the second network device; the training request at least carries a first number of the first network device, a second number of the target AI decision module, and the first information.

[0043] To achieve the above object, embodiments of the present application also provide a network device, which is a second network device, comprising:

[0044] A second sending module is configured to send performance change data of all artificial intelligence (AI) decision modules to a first network device;

[0045] The first receiving module is configured to receive a training request sent by the first network device; the training request at least carries a first number of the first network device, a second number of a target AI decision module, and first information; the first information is used to indicate whether the second network device trains the target AI decision module; the target AI decision module is part of the AI decision modules.

[0046] The first processing module is configured to perform an operation corresponding to the first information according to the training request. To achieve the above-mentioned purposes, embodiments of the present application further provide a network device, which comprises a transceiver, a processor, a memory, and a program or instructions stored in the memory and executable on the processor; the processor implements the conflict alleviation method of the artificial intelligence module according to any one of the above-mentioned embodiments when executing the program or instructions.

[0047] To achieve the above-mentioned purposes, embodiments of the present application further provide a readable storage medium having a program or instructions stored thereon, wherein the program or instructions are executable by a processor to implement the steps of the conflict alleviation method of the artificial intelligence module according to any one of the above-mentioned embodiments.

[0048] The above-mentioned technical solutions of the present application have the following advantages:

[0049] The scheme of the present application, when the first network device needs to deploy a first artificial intelligence (AI) decision module, obtains performance change data corresponding to all AI decision modules from a second network device, and determines whether there is a network conflict between the first AI decision module and multiple AI decision modules in the first network device through the performance change data; determines first information according to the performance change data, wherein the first information is used to indicate whether the second network device trains a target AI decision module, and the target AI decision module is part of the AI decision modules; sends a training request to the second network device; and the training request at least carries a first number of the first network device, a second number of the target AI decision module, and the first information. In the above-mentioned scheme, the first information is determined through the performance change data, so that when the deployment requirement is not met, the training is indicated through the first information to solve the conflict alleviation between the AI decision modules; the number and time of joint training are also reduced, and the energy consumption and deployment time are reduced. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 The structure diagram of the conflict alleviation architecture provided by the embodiments of the present application;

[0051] Figure 2 The flowchart of the conflict alleviation method of the artificial intelligence module applied to the first network device according to the embodiments of the present application;

[0052] Figure 3A flowchart of a conflict mitigation method for an artificial intelligence module of a second network device according to an embodiment of the present application;

[0053] Figure 4 A conflict mitigation flowchart between AI decisions based on hierarchical training according to an embodiment of the present application;

[0054] Figure 5 A structure diagram of a first network device according to an embodiment of the present application;

[0055] Figure 6 A structure diagram of a second network device according to an embodiment of the present application;

[0056] Figure 7 A structure diagram of a network device according to another embodiment of the present application. DETAILED DESCRIPTION

[0057] To make the technical problems, technical solutions and advantages of the present application clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0058] It should be understood that the term “one embodiment” or “an embodiment” mentioned throughout the specification means that the specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, “in one embodiment” or “in an embodiment” appearing throughout the specification does not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner.

[0059] In various embodiments of the present application, it should be understood that the size of the serial number of the following processes does not mean the order of execution, and the execution order of the processes should be determined by its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0060] It should be known that there can be multiple AI algorithms (AI decision modules) in an entity to optimize different goals. For example, in network intelligence, an AI algorithm is designed to solve each network optimization problem. The AI algorithm usually changes one or more parameters to optimize a specific metric. For example, in a network, the control goal of the radio resource management use case can be a cell, a user equipment (UE), or a bearer, etc. The radio resource management control content can include access control, bearer control, handover control, quality of service (QoS) control, resource allocation, etc. Because the selection or configuration of AI algorithm goals can lead to conflicting actions, conflict mitigation is necessary. Conflict mitigation is to solve the conflicting interactions between different AI algorithms. Network conflicts can be divided into three categories: direct conflict, indirect conflict and implicit conflict.

[0061] (1) Direct conflict: the conflict can be directly observed. For example, two or more AI algorithms request different settings for one or more parameters of the same configuration of the control target. (2) Indirect conflict: the conflict cannot be directly observed, but there is a certain dependency between the parameters and resources of the AI use case targets can be observed. (3) Implicit conflict: the conflict cannot be directly observed, and the dependency between the AI use cases is not obvious.

[0062] To solve the network conflict that may exist when multiple AI algorithms (AI decision modules) optimize different targets, with reference to Figure 1 The embodiment of the application provides a conflict mitigation architecture, which includes a management / control body 1, a digital twin body 2 and a decision execution body 3.

[0063] The management / control body 1 is a logical body that manages or controls the decision execution body 3, and can be deployed on entities such as network management devices, edge clouds, core network elements or base stations in the network. The management / control body 1 is composed of an AI decision module conflict mitigation decision unit 101, an AI decision module subscription unit 102 and an AI decision module repository 103. The AI decision module is a module that outputs intelligent decisions according to received data, and includes an AI model and related business logic functions, which can be regarded as an application program (APP). The AI decision modules are independent and can be deployed to cause conflicts. The AI decision module conflict mitigation decision unit 101 interacts with the digital twin body 2, and outputs conflict mitigation decisions based on the information fed back by the digital twin body 2. The AI decision module subscription unit 102 interacts with the decision execution body 3, and establishes the relationship between the AI decision module and the decision execution body 3, so that the AI decision module can control the decision execution body 3. The AI decision module repository 103 stores multiple AI decision modules, including attribute information of the AI decision modules and data of the AI decision modules.

[0064] The digital twin body 2 is a logical body of a digital decision execution body and a network environment, and is used for pre-validation of the AI decision module before actual deployment. It can be deployed on entities such as network management devices, core networks, edge clouds or base stations in the network. The digital twin body 2 is composed of an AI decision module management unit 201, an AI decision module data analysis unit 202 and an AI decision module running environment 203. The AI decision module management unit 201 has the ability to create, delete, decide to jointly train AI decision modules, and stores attribute information of the AI decision modules in the digital twin body 2. The AI decision module data analysis unit 202 collects running data of each AI decision module in the AI decision module running environment 203, and analyzes the performance results. The AI decision module running environment 203 is where the AI decision modules decided to be started / started by the management / control body 1 run / train in the twin environment.

[0065] The decision executor 3 is an executor of the AI decision module, which can be deployed in a core network, a base station, etc., such as a forwarding module of a router, a CU (centralized unit) / DU (distributed unit), etc., to execute decisions such as antenna adjustment, bandwidth allocation, etc.

[0066] It should be noted that the conflict mitigation architecture described above can be applied to the conflict mitigation method of the artificial intelligence module, for example, the first network device can be the management / control body 1 in Figure 1 , and the second network device can be the digital twin 2 in Figure 1 .

[0067] Referring to FIG. 1, Figure 2 , the embodiments of the present application provide a conflict mitigation method of an artificial intelligence module, which is executed by a first network device, comprising:

[0068] Step 21, when the first network device needs to deploy a first artificial intelligence AI decision module, obtaining performance change data corresponding to all AI decision modules from a second network device.

[0069] It should be noted that the first AI decision module is an AI decision module that needs to be deployed in the first network device, which can also be understood as an AI decision module to be deployed in the first network device, and the first AI decision module is preferably one AI decision module.

[0070] Correspondingly, after sending the related parameters of the first AI decision module to the second network device, the second network device needs to verify the first AI decision module in the verification environment, that is, the second network device runs all AI decision modules in the running environment, and compares the parameter data after running each AI decision module with the related parameters of the first AI decision module to determine the performance change data corresponding to all AI decision modules.

[0071] Here, the performance change data is the actual result generated after running in the running environment of the second network device, and the result data after comparing with the related parameters of the first AI decision module. For example, the actual result generated after running is A, and the related parameters of the first AI decision module are B, and the performance change data after comparison is the difference (A-B). For example, B is 90%, A is 85%, and the performance change data is -5%. Of course, -5% here is the performance change parameter of one AI decision module among all AI decision modules.

[0072] In the embodiments of the present application, whether the deployment of the first AI decision module will conflict with all AI decision modules in the first network device can be determined according to the obtained performance change data, and data support for subsequently determining whether all AI decision modules in the first network device need to be trained is provided.

[0073] In step 22, first information is determined according to the performance change data; the first information is used to indicate whether the second network device trains a target AI decision module; the target AI decision module is part of the AI decision modules in the first network device.

[0074] In the embodiments of the present application, whether the performance change data meets the deployment condition can be determined through the performance change data, that is, whether the current performance change data will conflict with the multiple AI decision modules in the first network device. If there is no conflict, it can be determined that the first information carries a parameter indicating that the first AI decision module can be directly deployed. If there is a conflict, it can be determined that the first information carries another parameter indicating that the target AI decision module needs to be trained. The first number and the second number are both unique numbers.

[0075] Optionally, the first number is the number of the first network device, which can be represented by M / C_ID. The second number is the number of the target AI decision module that needs to be trained, which can carry the unique identifier of each AI decision module that needs to be trained, which can be represented by AI_ID. The number AI_ID can be identified by numbers such as 0, 1, 2, 3, and the like.

[0076] In step 23, a training request is sent to the second network device; the training request at least carries the first number of the first network device, the second number of the target AI decision module, and the first information.

[0077] The present application can send a training request to the second network device, and the first information in the training request is used to indicate whether the second network device needs to be trained, which can reduce the number and range of AI decision models that need to be jointly trained, and reduce energy consumption and deployment time.

[0078] It should be noted that the second network device is a network device with a digital twin function.

[0079] Optionally, the above step 21 comprises:

[0080] The verification message for verifying the first AI decision module is sent to the second network device, and the verification message at least carries the first number and the number corresponding to the first AI decision module;

[0081] receive the verification response message sent by the second network device, wherein the verification response message carries the performance change data corresponding to all the AI decision modules.

[0082] In the embodiments of the present application, the first network device starts the AI decision module through monitoring network data or external triggering, and sends a verification message for verifying the first AI decision module to the second network device. The verification message can be represented by M / C-AI-TestRequest, and the data code of the first AI decision module is sent to the second network device. The M / C-AI-TestRequest includes the first number M / C_ID of the first network device, the attribute information AI_APP-Attribute of the first AI decision module, and the number corresponding to the first AI decision module.

[0083] Correspondingly, the second network device runs and collects the running data of all the AI decision modules in the running environment, compares each running data with the corresponding parameters of the first AI decision module, determines the performance result of each AI decision module, sets the performance result field of each AI decision module after verification, and determines the performance result field of all the AI decision modules as the performance change data corresponding to each AI decision module.

[0084] Here, the verification response message sent by the second network device is received, and the response message can be represented by M / C-AI-TestResponce. The verification response message carries the performance change data of all the AI decision modules, the first number and the second number.

[0085] It should be noted that the performance change data AI_threshold_Flag represents the performance change of all the AI decision modules, i.e., whether the corresponding function of each AI decision module decreases below a preset threshold, whether the corresponding amplitude value of each AI decision module exceeds a preset amplitude range, etc. Each performance change data at least includes one of the following: the number of AI decision modules whose performance decreases below the preset threshold, the percentage of performance decrease, etc.

[0086] Optionally, the verification message sent by the first network device to the second network device further carries the attribute information of the first AI decision module. The attribute information includes the priority and performance threshold constraint parameter of the first AI decision module.

[0087] Here, the priority of the first AI decision module can be represented by a priority AI_priority, which represents the importance of the decision output by the first AI decision module. The AI_priority can be comprehensively evaluated by factors such as duration, number of relevant decision executors, and importance of the business, and can be obtained by weighting the above-mentioned values. The AI_priority can be sequentially identified by numbers, and can be sequentially identified in descending order, in ascending order, or according to a certain preset rule.

[0088] The performance threshold constraint parameter can be represented by AI_threshold, which is a preset value of the execution result, representing the acceptable minimum value of the performance (i.e., optimization target) result of the model output decision under different environments, and can be considered as a constraint condition. The performance threshold constraint parameter can be represented in various ways. If the dimensions of the optimization target of the first AI decision module include multiple aspects, such as the need to optimize throughput and latency, the preset value of the performance threshold constraint parameter can be a vector, each optimization target of the vector having a preset value; or a number representing the weighted value of all optimization targets.

[0089] In combination Figure 1 As shown, the attribute information of each AI decision module and the data of each AI decision module of the first network device are stored in the AI decision module repository 13 of the management / control body 1, which can be as shown in Table 1. The AI decision module repository 13 will update the relevant data when a new AI decision module arrives or an existing AI decision module changes.

[0090] Table 1: AI decision module repository of management / control body A

[0091] AI_ID AI_priority AI_threshold AI_location 1 2 [2，6] C:\Users\cmcc\AImodel / 1 2 1 5 C:\Users\cmcc\AImodel / 2

[0092] Wherein, AI_location represents the storage location of each AI decision module code or data.

[0093] Optionally, the step 22 comprises:

[0094] According to the performance change data, it is determined whether each performance parameter corresponding to the performance change data is less than the corresponding first preset range;

[0095] If each performance parameter is greater than or equal to the corresponding first preset range, it is determined that the first information carries a first parameter, and the first AI decision module is deployed; the first parameter is used to indicate that the second network device does not perform training;

[0096] If the part of the performance parameters is less than the corresponding first preset range, it is determined that the first information carries a second parameter; the second parameter is used to indicate that the second network device performs hierarchical training on the target AI decision module; the hierarchical training is used to indicate that all AI decision modules are divided into a set, and the target AI decision module set after the division is jointly trained.

[0097] In the present application, the first preset range can represent whether each performance parameter meets the threshold range of deploying the first AI decision module. For example, if the performance change data includes three performance parameters, the first preset range corresponding to each performance parameter can be analyzed and compared respectively, the analysis result of each performance parameter is determined, and the multiple analysis results are determined as the performance change data.

[0098] In an implementation manner, if each performance parameter is greater than or equal to the corresponding first preset range, it indicates that each performance parameter meets the preset requirement, that is, the first network device meets the condition of deploying the first AI decision module, and it is determined that the first information carries a first parameter. The first information can be represented by a field, such as an AI_Train field. The AI_Train field is set with the first parameter False. The first parameter False indicates that the conflict between the first AI decision module and other AI decision modules is small, and the first AI decision module does not need to be trained. The first network device selects to deploy the first AI decision module.

[0099] In another implementation manner, if part of the performance parameters, but not all the performance parameters, are less than the corresponding first preset range, it indicates that network conflict may occur, that is, the performance change data shows that the performance has dropped to the threshold value, which indicates that the conflict between the first AI decision module and other AI decision modules is large, and the target AI decision module needs to be trained. The hierarchical training is used to indicate that all AI decision modules are divided into a set, and the target AI decision module set after the division is jointly trained. Here, it is determined that the first information carries a second parameter, that is, the AI_Train field is set with True. The second parameter is used to indicate that the second network device performs hierarchical training on the target AI decision module.

[0100] Optionally, the method further includes:

[0101] When each performance parameter in the performance change data is less than the corresponding first preset range, it is determined that the first information carries a third parameter; the third parameter is used to indicate that the attribute information corresponding to the first AI decision module is deleted in the first network device and / or the second network device.

[0102] In the present application, the first preset range can represent whether each performance parameter meets the threshold range of deploying the first AI decision module. The first preset range corresponding to each performance parameter can be analyzed and compared respectively to determine the analysis result of each performance parameter, and the multiple analysis results are determined as performance change data. When each performance parameter in the performance change data is less than the corresponding first preset range, it indicates that each performance parameter in the performance change data does not meet the preset requirement, and the third parameter in the first information is determined to be False, indicating that the target AI decision module does not need to be trained, and the first AI decision module is not selected to be deployed. The attribute information of the first AI decision module in the first network device can be deleted, and the attribute information of the first AI decision module in the second network device is also deleted, and step 21 is re-executed to design new attribute information of the AI decision module again.

[0103] Optionally, when the second parameter is carried in the first information, the method further comprises:

[0104] receiving performance result data sent by the second network device after the target AI decision module is trained;

[0105] determining a decision message of the first network device according to the performance result data; the decision message carries the first number, the second number and second information; the second information is used to indicate whether the first AI decision module is deployed in the first network device;

[0106] sending the decision message to the second network device.

[0107] In the present application, when the second parameter is carried in the first information, it indicates that the second network device needs to train the target AI decision module, and the target AI decision module is part of the AI decision module. The performance result data sent by the second network device after the target AI decision module is trained can be a training success result or a training failure result. According to the training success result or the training failure result carried in the performance result data, the decision message of the first network device is determined, and the second information in the decision message indicates whether the first AI decision module is deployed or not.

[0108] Optionally, after the decision message is sent to the second network device, the above method further comprises:

[0109] receiving target parameter data corresponding to all AI decision modules after the target AI decision module is trained and sent by the second network device;

[0110] According to all the target parameter data, modify all the AI decision modules of the first network device, and deploy the first AI decision module;

[0111] issue the first AI decision module to the decision executor.

[0112] It should be noted that if the second information is represented by the field Decision, after sending the field Decision to the second network device, if the field Decision carries the False parameter, the second network device will delete the AI decision module from the running environment of the second network device, and will not receive all the target parameter data corresponding to the AI decision module sent by the second network device.

[0113] In the embodiment of the application, if the field Decision carries the True parameter, the AI decision module after hierarchical training is retained, the first network device can receive the target parameter data corresponding to all the AI decision modules after hierarchical training, modify the parameters corresponding to all the AI decision modules of the first network device through all the target parameter data, and after completing the modification, the first AI decision module can be deployed in the first network device. After the deployment of the first AI decision module is completed, the first network device subscribes to the decision executor to obtain the control management right of the decision executor, and issues the first AI decision module to the decision executor, which can ensure that there is no network conflict when the decision executor executes the corresponding action.

[0114] The method of the application can reduce the number and range of AI decision modules that need to be jointly trained, reduce energy consumption and deployment time.

[0115] Referring to Figure 3 The embodiment of the application also provides a conflict alleviation method of an artificial intelligence module, which is executed by a second network device and includes the following steps:

[0116] Step 31: Send the performance change data of all artificial intelligence (AI) decision modules to a first network device.

[0117] In the embodiment of the application, the first AI decision module is an AI decision module that needs to be deployed in the first network device. After receiving the related parameters of the first AI decision module, the second network device needs to verify the first AI decision module in a verification environment, that is, the second network device runs all the AI decision modules in the running environment, and compares the parameter data after the running of each AI decision module with the related parameters of the first AI decision module to determine the performance change data corresponding to all the AI decision modules.

[0118] For example, the actual result generated after running is A, the relevant parameter of the first AI decision module is B, and the performance change data after comparison is the difference value of (A-B). For another example, B is 90%, and A is 85%, and the performance change data is -5%. Of course, -5% here is the performance change parameter of one of all AI decision modules.

[0119] Of course, when there are multiple parameter difference results, a preset weight corresponding to the parameter difference result can be set to determine a numerical value by weighting. In this application, the performance change data of all AI decision modules is sent to the first network device, which is also used to determine whether the target AI decision module of the first network device needs subsequent hierarchical training.

[0120] Step 32, receiving the training request sent by the first network device; the training request at least carries the first number of the first network device, the second number of the target AI decision module, and the first information; the first information is used to indicate whether the second network device trains the target AI decision module; the target AI decision module is part of the AI decision module in the all AI decision modules.

[0121] Step 33, according to the training request, performing the operation corresponding to the first information.

[0122] In the embodiment of the application, the training request is received, and the training request carries the first number and the second number. The first number and the second number are unique numbers. Optionally, the first number is the number of the first network device, which can be represented by M / C_ID. The second number is the number of the target AI decision module that needs to be trained, which can carry the unique identifier of each AI decision module that needs to be trained, which can be represented by AI_ID. For example, the number AI_ID can be identified by numbers such as 0, 1, 2, 3... This application can indicate the second network device to execute the information of deleting the first AI decision module, or to execute the hierarchical training of the target AI decision module according to the parameter carried by the first information. In the case of hierarchical training, the hierarchical training is performed, and in the case of no training, the first AI decision module is directly deployed by the first network device, which reduces the energy consumption and deployment time.

[0123] Optionally, the step 31 comprises:

[0124] receiving a verification message for verifying an artificial intelligence (AI) decision module sent by the first network device; the verification message at least carries the first number and the number corresponding to the first AI decision module;

[0125] sending a verification response message to the first network device; the verification response message carries the performance change data corresponding to all the AI decision modules.

[0126] In the embodiment of the application, a verification message sent by the first network is received, and the verification message can be represented by M / C-AI-TestRequest. The AI decision module data code sent by the first network device is also received. The M / C-AI-TestRequest includes the first number M / C_ID of the first network device, the attribute information AI_APP-Attribute of the first AI decision module, and the number corresponding to the first AI decision module. The second network device runs all the AI decision modules in the running environment, collects the running data of all the AI decision modules in the running process, compares each running data with the parameter corresponding to the first AI decision module, determines the performance result of each AI decision module, sets the performance result field of each AI decision module after verification, and determines the performance result field of all the AI decision modules as the performance change data corresponding to each AI decision module.

[0127] The second network device sends a verification response message to the first network device, and the verification response message can be represented by M / C-AI-TestResponce. The M / C-AI-TestResponce needs to include the first number M / C_ID, the number corresponding to the first AI decision module, and the performance change data of all the AI decision modules. Here, the performance change data can be represented by AI_threshold_Flag.

[0128] It should be noted that the performance change data AI_threshold_Flag represents the performance change of all the AI decision modules, that is, whether the function corresponding to each AI decision module decreases below a preset threshold, whether the amplitude value corresponding to each AI decision module exceeds a preset amplitude range, and the like. Each performance change data at least includes one of the following: the number of the AI decision module whose performance decreases below the preset threshold, the percentage of performance decrease, and the like.

[0129] Optionally, the step 33 includes:

[0130] When the first information carries a first parameter, the second network device does not process; the first parameter is used to indicate that the second network device does not perform training.

[0131] When the first information carries a second parameter, the second network device performs hierarchical training on the target AI decision module; the second parameter is used to indicate that the second network device performs hierarchical training on the target AI decision module; the hierarchical training is to divide all the AI decision modules into sets, and to perform joint training on the set of target AI decision modules after the division.

[0132] In an implementation form of the application, the first information can be represented by a field, such as an AI_Train field. When the first information carries a first parameter, i.e., the AI_Train field is set with the first parameter False, the first parameter False indicates that the conflict of the first AI decision module with other AI decision modules is small, and no training is needed. The first network device can deploy the AI decision module that meets the first parameter, and thus the first parameter can be used to instruct the second network device not to perform training.

[0133] In another implementation form of the application, when the first information carries a second parameter, a network conflict can occur, i.e., the performance change data shows that the performance has dropped to a threshold, indicating that the conflict of the first AI decision module with other AI decision modules is large, and hierarchical training is needed, i.e., the AI_Train field is set with the second parameter True, and the second network device performs hierarchical training on the target AI decision module.

[0134] Optionally, the method further includes:

[0135] When the first information carries a third parameter, attribute information of the first AI decision module is deleted in the running environment of the second network device. The third parameter is used to instruct to delete the attribute information corresponding to the first AI decision module in the second network device.

[0136] In an implementation form of the application, the first information carries a third parameter, and the third parameter is set to False, indicating that no hierarchical training is needed for the target AI decision module, and the first network device selects not to deploy the first AI decision module. The attribute information of the first AI decision module in the second network device can be deleted, and new attribute information of the AI decision module is reissued by the first network device.

[0137] In the embodiment of the application, it is needed to verify whether there is a conflict with the started AI decision module in the second network device. After the second network device receives the M / C-AI-Request message of the first network device for deploying the first AI decision module, all AI decision modules in the running environment are run, and running data of all AI decision modules in the running process are collected. Each running data is compared with the parameter corresponding to the first AI decision module, and the performance change value of each AI decision module is calculated and fed back to the first network device. If the conflict of the first AI decision module with other AI decision modules is large, and the first network device still decides to deploy the first AI decision module, hierarchical training needs to be started to alleviate the conflict.

[0138] Optionally, the method further includes:

[0139] determine attribute information of the first AI decision module; the attribute information comprises: priority of the first AI decision module and performance threshold constraint parameter;

[0140] divide the plurality of AI decision modules into a first decision module set and a second decision module set according to the priority of the plurality of AI decision modules; the first decision module set is used to represent all decision modules whose priority is not higher than that of the first AI decision module; the second decision module set is used to represent all decision modules whose priority is higher than that of the first AI decision module;

[0141] jointly train the first decision module set according to the performance threshold constraint parameter corresponding to each AI decision module in the second decision module set, and determine a target parameter whose performance of the first decision module set changes during the training process;

[0142] when it is determined that the target parameter meets a preset threshold range, complete the hierarchical training of the target AI decision module.

[0143] It should be noted that the priority of the first AI decision module can be represented by priority AI_priority, which represents the importance of the AI decision module output decision, which can be comprehensively evaluated by factors such as duration, number of related decision executors, and importance of business, and can be obtained by weighting the above values. The AI_priority can be sequentially identified by numbers, and can be sequentially identified from large to small, from small to large, or according to a certain preset rule. The performance threshold constraint parameter can be represented by AI_threshold, which is a preset value of the execution result, representing the acceptable minimum value of the performance (i.e. optimization target) result of the model output decision under different environments, which can be considered as a constraint condition. The performance threshold constraint parameter can be represented in various ways. If the dimensions of the optimization target of the first AI decision module include multiple aspects, such as throughput and latency, the preset value of the performance threshold constraint parameter can be a vector, each optimization target of the vector having a preset value; or a number representing the weighted value of all optimization targets.

[0144] Here, the attribute information of all AI decision modules sent by the first network device can be stored in the second network device, and the storage form can be as shown in Table 2.

[0145] Table 2:

[0146] AI_ID M / C_ID AI_priority AI_threshold AI_location A-1 A 2 [2，6] C:\Users\cmcc\AImodel / A1 A-2 A 1 5 C:\Users\cmcc\AImodel / A2

[0147] In the present application, when the second network device performs hierarchical training, all AI decision modules in Table 2 can be divided into a first decision module set N and a second decision module set M according to the priority of the first AI decision module. The second decision module set M is all decision modules with a priority higher than the first AI decision module, and the first decision module set N is all decision modules with a priority not higher than the first AI decision module. The AI decision modules in the first decision module set N are jointly trained with the performance threshold of each AI decision module in the second decision module set M as a constraint to complete the hierarchical training of the target AI module.

[0148] For example, the plurality of AI decision modules are 10 AI decision modules, and the priorities are divided into a plurality of priorities, such as ten priorities, and the priorities of each AI decision module are different, or divided into four priorities, and some AI decision modules have the same priority, and the like. Here, an example with three priorities is given, such as priorities a, b, and c, with the priorities decreasing in turn, the priority of the first AI decision module is n, if the priorities of a and b are higher than n, and the priority of c is not higher than n, the AI decision modules corresponding to the priorities of a and b are divided into the second decision module set M, and the remaining AI decision modules are divided into the first decision module set N; if there are only three AI decision modules in the first decision module set N, the three AI decision modules need to be jointly trained, and the second decision module set M does not need to be trained. In this way, the AI decision modules are divided into sets, and the target AI decision module set after division (such as the first decision module set N with a priority not higher than the first AI decision module) is jointly trained, which is a hierarchical training method, and can solve the problem of conflict between AI decision modules.

[0149] It should be noted that the AI decision modules in set N are trained hierarchically, and when the output action in the training process causes the performance of the AI decision modules in set M to decrease to a threshold, a large penalty factor is fed back to ensure the performance of the AI decision modules in set M. The penalty factor can be understood as a correction parameter for ensuring the performance of the AI decision modules in set M.

[0150] It should be further noted that the state dimension of each AI decision module is S, and the action dimension is F. If the hierarchical training method is not used, the state dimension of the combined neural network is (|M|+|N|)*S, and the action dimension is (|M|+|N|)*F. In the present application, if the hierarchical training method is used, the state dimension is |N|*S, and the action dimension is |N|*F. It can be seen that the complexity of the neural network can be significantly reduced, and the training time and cost can be reduced.

[0151] Optionally, the first information carries a second parameter, and the second network device performs hierarchical training on the target AI decision module, including:

[0152] When the first information carries the second parameter, the second network device starts the hierarchical training.

[0153] The result of the hierarchical training is set as performance result data, and the performance result data is sent to the first network device.

[0154] In the embodiment of the application, the first information is represented by an AI_Train field, and the first information carries a second parameter, that is, the second parameter True is set in the AI_Train field. If the AI_Train field is set to True, the second network device starts the hierarchical training. The second network device sets the training result as performance result data after training and sends the performance result data to the first network device.

[0155] In contrast, the first network device decides whether to deploy the first AI decision module according to the result, sets a decision message MC-AI-Decision message, and sends the decision message to the second network device. The decision message needs to include the first number M / C_ID of the first network device, the second number AI_ID of the target AI decision module, and the field Decision of whether to deploy.

[0156] Optionally, the method further includes:

[0157] receiving the decision message sent by the first network device; the decision message carries the first number, the second number, and second information; the second information is used to indicate whether to deploy the first AI decision module in the first network device;

[0158] When the second information carries a fourth parameter, attribute information corresponding to the first AI decision module is deleted in the running environment of the second network device; the fourth parameter is used to indicate that the first AI decision module is not deployed in the first network device;

[0159] When the second information carries a fifth parameter, the second network device performs hierarchical training on the target AI decision module; the fifth parameter is used to indicate that the first AI decision module is deployed in the first network device;

[0160] The target parameter data corresponding to all AI decision modules after hierarchical training is sent to the first network device.

[0161] In an implementation form of the application, the decision message sent by the first network device is received, such as represented by an MC-AI-Decision message; the MC-AI-Decision message carries a first number M / C_ID, a second number AI_ID and second information; the second information can be represented by Decision, which is used to indicate whether the first AI decision module is deployed in the first network device; the fourth parameter is carried in the second information Decision, which is used to indicate that the first AI decision module is not deployed in the first network device, and the attribute information of the first AI decision module in the second network device is deleted, i.e., when the fourth parameter is represented as False, the attribute information of the first AI decision module is deleted in the running environment of the second network device.

[0162] In another implementation form of the application, if the second information Decision carries a fifth parameter, and the fifth parameter is represented as True, the second network device retains the first AI decision module, and the second network device performs hierarchical training on the target AI decision module, sends the parameter data corresponding to all AI decision module codes that have been modified after hierarchical training of all AI decision modules to the first network device, so that the first network device can modify the parameters and then deploy the first AI decision module, which can avoid network conflicts caused by deployment of the first network device.

[0163] The application also provides an AI decision conflict mitigation process based on hierarchical training, as shown in Figure 1 to Figure 4 , wherein the first network device is the management / control body 1 in Figure 1 , and the second network device is the digital twin 2, and the interaction process among the three is as follows:

[0164] Step 1: The management / control body decision starts the AI decision module through monitoring network data or external triggering, etc.

[0165] Step 2: The AI decision module conflict mitigation solution function of the management / control body sends an AI decision module verification request M / C-AI-TestRequest to the AI decision module management function of the digital twin, and sends the data code of the AI decision module from the AI decision module repository to the AI decision module management function.

[0166] Step 3: The AI decision module management function of the digital twin deploys the AI decision module into the AI decision module running environment and starts running.

[0167] Step 4: The AI decision module data analysis function of the digital twin collects the running data of each AI decision module in the AI decision module running environment, analyzes the performance results, and sets the performance result data after verification.

[0168] Step 5, the digital twin sends a verification response M / C-AI-TestResponce to the AI decision module conflict buffer solution function of the management / control body, which needs to include the number of the management / control body M / C_ID, the number of the AI decision module AI_ID and the performance result data AI_threshold_Flag after verification.

[0169] The performance result data AI_threshold_Flag after verification indicates the performance change of all AI decision modules in the running environment after new deployment, that is, whether it falls below the threshold, the magnitude of exceeding. This field at least contains one of the following: the number of AI decision module whose performance falls below the threshold, the percentage of performance decline, etc.

[0170] Step 6, the AI decision module conflict buffer solution function of the management / control body decides whether joint training is needed, and sends the MC-AI-TrainRequest message to the AI decision module management function of the digital twin, which needs to include the number of the management / control body M / C_ID, the number of the AI decision module AI_ID and the field AI_Train whether to train, AI_Train field indicates whether the management / control body requires the digital twin to train the AI decision module.

[0171] The basis for judging whether to train jointly can be as follows:

[0172] 1) If the AI_threshold_Flag field received by the management / control body shows that the performance does not fall below the threshold, it means that the conflict of the AI decision module with other AI decision modules is small, and there is no need to train, and the AI_Train field is set to False;

[0173] 2) If the AI_threshold_Flag field received by the management / control body shows that the performance falls below the threshold, it means that the conflict of the AI decision module with other AI decision modules is large, and the module can be selected not to be deployed (deleted), and the AI_Train field is set to False;

[0174] 3) If the AI_threshold_Flag field received by the management / control body shows that the performance falls below the threshold, it means that the conflict of the AI decision module with other AI decision modules is large, and joint training is needed, and the AI_Train field is set to True;

[0175] In a specific embodiment, if the AI_Train field is set to False, the management / control body deletes / starts the AI decision module; if started (the conflict of the AI decision module with other AI decision modules is small), the AI decision module of the management / control body subscribes to the function to send a subscription request to the decision execution body, obtains the management / control right of the decision execution body, so that the decision output by the AI decision module can be issued to the decision execution body.

[0176] In another specific embodiment, if the AI_Train field is set to True, the digital twin starts hierarchical training; the digital twin sets the training result as the performance result data after training and sends it to the management / control body. The management / control body decides whether to deploy the AI decision module according to the result and sets the MC-AI-Decision message, which needs to include the number M / C_ID of the management / control body, the number AI_ID of the AI decision module and the field Decision of whether to deploy, and sends it to the digital twin.

[0177] If Decision is False, the digital twin deletes the AI decision module from the AI decision module running environment.

[0178] If Decision is True, the digital twin retains the AI decision module and sends all modified AI decision module codes to the management / control body; the management / control body deploys the modified AI decision module. The AI decision module of the management / control body subscribes to the function to send a subscription request to the decision execution body, obtains the management / control right of the decision execution body, so that the decision output by the AI decision module can be issued to the decision execution body.

[0179] To sum up, the scheme of the present application utilizes digital twins to realize conflict mitigation in a digital network, can intuitively find various conflicts including implicit conflicts, provides reliable basis for conflict mitigation, and reduces trial and error risks; further sets priorities and thresholds for AI decisions, as much as possible to reduce the number and range of models that need to be jointly trained to make AI decisions, and reduces energy consumption and deployment time.

[0180] Referring to Figure 5 The embodiment of the present application also provides a network device, which is a first network device and comprises:

[0181] The first acquisition module 51 is configured to acquire, from a second network device, performance change data corresponding to all AI decision modules when the first network device needs to deploy a first artificial intelligence (AI) decision module.

[0182] The first determining module 52 is configured to determine first information according to the performance change data, wherein the first information is used to indicate whether the second network device trains a target AI decision module, and the target AI decision module is part of the AI decision modules.

[0183] The first sending module 53 is configured to send a training request to the second network device, wherein the training request at least carries the first number of the first network device, a second number of the target AI decision module, and the first information.

[0184] Optionally, the first determining module 52 includes:

[0185] The first determining unit is configured to determine whether each performance parameter corresponding to the performance change data is less than a corresponding first preset range according to the performance change data.

[0186] The second determining unit is configured to determine that the first information carries a first parameter and deploy a first AI decision module if each performance parameter is greater than or equal to the corresponding first preset range, wherein the first parameter is used to indicate that the second network device does not train.

[0187] The third determining unit is configured to determine that the first information carries a second parameter if part of the performance parameters are less than the corresponding first preset range, wherein the second parameter is used to indicate that the second network device performs hierarchical training on the target AI decision module, and the hierarchical training is to divide all AI decision modules into sets and perform joint training on the target AI decision module sets after the division.

[0188] Optionally, the network device further includes:

[0189] The second determining module is configured to determine that the first information carries a third parameter if each performance parameter in the performance change data is less than the corresponding first preset range, wherein the third parameter is used to indicate that attribute information corresponding to a first AI decision module is deleted in the first network device and / or the second network device.

[0190] Optionally, the network device further includes:

[0191] The third receiving module is configured to receive performance result data sent by the second network device after the hierarchical training on the target AI decision module.

[0192] a third determining module, configured to determine a decision message of the first network device according to the performance result data; the decision message carries the first number, the second number and second information; the second information is used to indicate whether to deploy the first AI decision module in the first network device;

[0193] a third sending module, configured to send the decision message to the second network device.

[0194] Optionally, the network device described above further includes:

[0195] a fourth receiving module, configured to receive target parameter data corresponding to all AI decision modules after hierarchical training of the target AI decision module sent by the second network device;

[0196] a second processing module, configured to modify all AI decision modules of the first network device after modification according to all the target parameter data, and deploy the first AI decision module;

[0197] a third processing module, configured to issue the first AI decision module to a decision executor.

[0198] Optionally, the first obtaining module 51 described above includes:

[0199] a first sending unit, configured to send a verification message used to verify the first AI decision module to the second network device; the verification message at least carries the first number and a number corresponding to the first AI decision module;

[0200] a first receiving unit, configured to receive a verification response message sent by the second network device; the verification response message carries performance change data corresponding to all the AI decision modules.

[0201] The implementation embodiments of the conflict resolution method of the artificial intelligence module are applicable to the embodiments of the network device, and the same technical effects can be achieved.

[0202] Referring to Figure 6 The embodiments of the present application further provide a network device, which is a second network device and includes:

[0203] a second sending module 61, configured to send performance change data of all artificial intelligence (AI) decision modules to a first network device;

[0204] The first receiving module 62 is configured to receive a training request sent by the first network device; the training request at least carries a first number of the first network device, a second number of a target AI decision module, and first information; the first information is used to indicate whether the second network device trains the target AI decision module; the target AI decision module is part of the AI decision modules;

[0205] The first processing module 63 is configured to perform an operation corresponding to the first information according to the training request.

[0206] Optionally, the network device described above further includes:

[0207] The fifth receiving module is configured to receive a decision message sent by the first network device; the decision message carries the first number, the second number, and second information; the second information is used to indicate whether a first AI decision module is deployed in the first network device;

[0208] The fourth processing module is configured to delete attribute information corresponding to the first AI decision module in the running environment of the second network device when the second information carries a fourth parameter; the fourth parameter is used to indicate that the first AI decision module is not deployed in the first network device;

[0209] The fifth processing module is configured to perform hierarchical training on the target AI decision module by the second network device when the second information carries a fifth parameter; the fifth parameter is used to indicate that the first AI decision module is deployed in the first network device; the hierarchical training is to divide all AI decision modules into sets and perform joint training on the target AI decision module sets after the division;

[0210] The sixth processing module is configured to send target parameter data corresponding to all AI decision modules after the hierarchical training to the first network device.

[0211] Optionally, the network device described above further includes:

[0212] The fourth determining module is configured to determine attribute information of the first AI decision module; the attribute information includes a priority of the first AI decision module and a performance threshold constraint parameter;

[0213] The seventh processing module is configured to divide a plurality of AI decision modules into a first decision module set and a second decision module set according to the priorities of the plurality of AI decision modules; the first decision module set is used to represent all decision modules whose priorities are not higher than that of the first AI decision module; the second decision module set is used to represent all decision modules whose priorities are higher than that of the first AI decision module.

[0214] The eighth processing module is used to perform joint training on the first decision module set based on the performance threshold constraint parameters corresponding to each AI decision module in the second decision module set, and to determine the target parameters that cause the performance of the first decision module set to change during the training process.

[0215] The ninth processing module is used to complete the hierarchical training of the target AI decision module when the target parameters meet the preset threshold range.

[0216] The aforementioned implementation embodiments of the conflict mitigation method for the artificial intelligence module are all applicable to the embodiments of this network device and can achieve the same technical effect.

[0217] Another embodiment of the network device in this application, such as Figure 7 As shown, it includes a transceiver 710, a processor 700, a memory 720, and a program or instructions stored in the memory 720 and executable on the processor 700; when the processor 700 executes the program or instructions, it implements the conflict mitigation method applied to the artificial intelligence module described above.

[0218] The transceiver 710 is used to receive and send data under the control of the processor 700.

[0219] Among them, Figure 7 In this context, the bus architecture may include any number of interconnected buses and bridges, specifically linking various circuits together, represented by one or more processors (processor 700) and memory (memory 720). The bus architecture may also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. The bus interface provides an interface. The transceiver 710 may be multiple elements, including transmitters and receivers, providing a unit for communicating with various other devices over a transmission medium. The processor 700 is responsible for managing the bus architecture and general processing, and the memory 720 may store data used by the processor 700 during operation.

[0220] This application provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the steps in the conflict mitigation method of the artificial intelligence module described above and achieve the same technical effect. To avoid repetition, this will not be repeated here.

[0221] The processor mentioned above is the processor in the conflict mitigation method of the artificial intelligence module described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0222] The exemplary embodiments described above are with reference to the accompanying drawings. Many different forms and embodiments are feasible without departing from the spirit and teachings of this application. Therefore, this application should not be construed as limiting the exemplary embodiments set forth herein. Rather, these exemplary embodiments are provided to make this application complete and convey the scope of this application to those skilled in the art. In these drawings, component dimensions and relative dimensions may be exaggerated for clarity. The terminology used herein is for the purpose of describing particular exemplary embodiments only and is not intended to be limiting. As used herein, unless clearly indicated otherwise, the singular forms “a,” “an,” and “the” are intended to include all such forms. It will be further understood that the terms “comprising” and / or “including”, when used in this specification, indicate the presence of the stated features, integers, steps, operations, components, and / or elements, but do not exclude the presence or addition of one or more other features, integers, steps, operations, components, and / or groups thereof. Unless otherwise indicated, when stated, a range of values ​​includes the upper and lower limits of the range and any subranges in between.

[0223] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principles described in this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A conflict mitigation method for an artificial intelligence module, characterized in that, The method is performed by a first network device, comprising: When the first network device needs to deploy a first artificial intelligence (AI) decision module, obtaining, from a second network device, performance change data corresponding to all AI decision modules; the performance change data is actual results generated after running in a running environment of the second network device, and result data obtained after comparing with related parameters of the first AI decision module; According to the performance change data, determining first information; the first information is used to indicate whether the second network device trains a target AI decision module; the target AI decision module is part of the AI decision modules; Sending a training request to the second network device; the training request at least carries a first number of the first network device, a second number of the target AI decision module, and the first information; According to the performance change data, determining first information, comprising: according to the performance change data, determining whether each performance parameter corresponding to the performance change data is less than a corresponding first preset range; if each performance parameter is greater than or equal to the corresponding first preset range, determining that the first information carries a first parameter, and deploying the first AI decision module; the first parameter is used to indicate that the second network device does not train; if part of the performance parameters are less than the corresponding first preset range, it is determined that the first information carries a second parameter; the second parameter is used to indicate that the second network device performs hierarchical training on the target AI decision module; the hierarchical training is to divide all AI decision modules into sets, and to jointly train the target AI decision module sets after division.

2. The method of claim 1, wherein, The method further comprises: When each performance parameter in the performance change data is less than the corresponding first preset range, it is determined that the first information carries a third parameter; the third parameter is used to indicate that the attribute information corresponding to the first AI decision module is deleted in the first network device and / or the second network device.

3. The method of claim 1, wherein, When the first information carries the second parameter, the method further comprises: Receiving performance result data sent by the second network device after performing hierarchical training on the target AI decision module; According to the performance result data, determining a decision message of the first network device; the decision message carries the first number, the second number, and second information; the second information is used to indicate whether to deploy the first AI decision module in the first network device; Sending the decision message to the second network device.

4. The method of claim 3, wherein, After sending the decision message to the second network device, the method further comprises: Receiving target parameter data corresponding to all AI decision modules after hierarchical training of the target AI decision module sent by the second network device; According to all the target parameter data, modifying all AI decision modules of the first network device, and then deploying the first AI decision module; Downlinking the first AI decision module to a decision execution body.

5. The method of claim 1, wherein, When a first network device needs to deploy a first artificial intelligence (AI) decision module, performance change data corresponding to all AI decision modules is obtained from a second network device, including: A verification message for verifying the first AI decision module is sent to the second network device, and the verification message at least carries the first number and a number corresponding to the first AI decision module; A verification response message sent by the second network device is received, and the verification response message carries the performance change data corresponding to all AI decision modules.

6. A conflict mitigation method for an artificial intelligence module, characterized in that, The method is performed by a second network device, including: Performance change data of all artificial intelligence (AI) decision modules is sent to a first network device; the performance change data is actual results generated after running in the running environment of the second network device, and result data after comparison with related parameters of a first AI decision module; A training request sent by the first network device is received; the training request at least carries a first number of the first network device, a second number of a target AI decision module, and first information; the first information is used to indicate whether the second network device trains the target AI decision module; the target AI decision module is part of the all AI decision modules; wherein the first network device determines whether each performance parameter corresponding to the performance change data is less than a corresponding first preset range according to the performance change data; if each performance parameter is greater than or equal to the corresponding first preset range, it is determined that a first parameter is carried in the first information, and the first AI decision module is deployed; the first parameter is used to indicate that the second network device does not train; if part of the performance parameters are less than the corresponding first preset range, it is determined that a second parameter is carried in the first information; the second parameter is used to indicate that the second network device performs hierarchical training on the target AI decision module; the hierarchical training is to divide all AI decision modules into sets, and to jointly train the target AI decision module sets after division; According to the training request, the operation corresponding to the first information is performed.

7. The method of claim 6, wherein, The method further includes: A decision message sent by the first network device is received; the decision message carries the first number, the second number, and second information; the second information is used to indicate whether the first AI decision module is deployed in the first network device; When a fourth parameter is carried in the second information, attribute information corresponding to the first AI decision module is deleted in the running environment of the second network device; the fourth parameter is used to indicate that the first AI decision module is not deployed in the first network device; When a fifth parameter is carried in the second information, the second network device performs hierarchical training on the target AI decision module; the fifth parameter is used to indicate that the first AI decision module is deployed in the first network device; the hierarchical training is to divide all AI decision modules into sets, and to jointly train the target AI decision module sets after division; Send the target parameter data corresponding to all AI decision modules after hierarchical training to the first network device.

8. The method of claim 6, wherein, The method further comprises: determining attribute information of the first AI decision module; the attribute information comprises priority and performance threshold constraint parameters of the first AI decision module; According to the priority of the plurality of AI decision modules, the plurality of AI decision modules are divided into a first decision module set and a second decision module set; the first decision module set is used to represent all decision modules whose priority is not higher than the first AI decision module; the second decision module set is used to represent all decision modules whose priority is higher than the first AI decision module; According to the performance threshold constraint parameters corresponding to each AI decision module in the second decision module set, the first decision module set is jointly trained, and the target parameters of the first decision module set whose performance changes during training are determined; When the target parameters meet the preset threshold range, the hierarchical training of the target AI decision module is completed.

9. A network device, the network device being a first network device, characterized in that, Comprise: The first acquisition module is used for acquiring the performance change data corresponding to all AI decision modules from the second network device when the first network device needs to deploy the first artificial intelligence AI decision module; the performance change data is the actual result generated after running in the running environment of the second network device, and the result data after comparing the related parameters of the first AI decision module; The first determination module is used for determining the first information according to the performance change data; the first information is used to indicate whether the second network device trains the target AI decision module; The target AI decision module is part of the AI decision modules in the all AI decision modules; The first determination module comprises: a first determination unit configured to determine whether each performance parameter corresponding to the performance change data is less than a corresponding first preset range according to the performance change data; a second determination unit configured to determine that the first information carries a first parameter and deploy the first AI decision module if each performance parameter is greater than or equal to the corresponding first preset range; the first parameter is used to indicate that the second network device does not train; a third determination unit configured to determine that the first information carries a second parameter if part of the performance parameters are less than the corresponding first preset range; the second parameter is used to indicate that the second network device performs hierarchical training on the target AI decision module; the hierarchical training is to divide all AI decision modules into sets and jointly train the target AI decision module set after division; The first sending module is used for sending a training request to the second network device; the training request carries at least the first number of the first network device, the second number of the target AI decision module and the first information.

10. A network device, the network device being a second network device, characterized in that Comprise: The second sending module is used for sending the performance change data of all artificial intelligence AI decision modules to the first network device; The performance change data is actual result generated after running in the running environment of the second network device, and result data after comparison with relevant parameters of the first AI decision module; The first receiving module is configured to receive a training request sent by the first network device; the training request at least carries a first number of the first network device, a second number of a target AI decision module, and first information; the first information is used to indicate whether the second network device trains the target AI decision module; the target AI decision module is part of the AI decision modules; wherein the first network device determines whether each performance parameter corresponding to the performance change data is less than a corresponding first preset range according to the performance change data; if each performance parameter is greater than or equal to the corresponding first preset range, it is determined that the first information carries a first parameter, and the first AI decision module is deployed; the first parameter is used to indicate that the second network device does not train; if part of the performance parameters are less than the corresponding first preset range, it is determined that the first information carries a second parameter; the second parameter is used to indicate that the second network device performs hierarchical training on the target AI decision module; the hierarchical training is to divide all AI decision modules into sets, and to jointly train the target AI decision module sets after division; The first processing module is configured to perform an operation corresponding to the first information according to the training request.

11. A network device comprising: A transceiver, a processor, a memory, and a program or instructions stored on the memory and executable on the processor; characterized in that the processor executes the program or instructions to implement the conflict resolution method of the artificial intelligence module according to any one of claims 1 to 5, or executes the program or instructions to implement the conflict resolution method of the artificial intelligence module according to any one of claims 6 to 8.

12. A readable storage medium, on which a program or instructions are stored, characterized in that, The program or instructions are executed by the processor to implement the conflict resolution method of the artificial intelligence module according to any one of claims 1 to 5, or the program or instructions are executed by the processor to implement the conflict resolution method of the artificial intelligence module according to any one of claims 6 to 8.

Images