Model training, data receiving method, device, equipment, medium and program product

Abstract

The application provides a model training and data receiving method and device, equipment, medium and program product, to solve the problem of low data generation efficiency in related technologies. The method comprises: receiving simulation data sent by a second device, the second device being a management service MnS producer; training a first model based on the simulation data to obtain a second model, the first model being a data generation model or a policy model. The application can improve the data generation efficiency.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a model training, data receiving method, apparatus, device, medium and program product. Background Technology

[0002] The 3GPP SA5 project's research on network digital twin management focuses on Network Digital Twin (NDT). Currently, the project has proposed nine key scenarios, one of which involves generating machine learning (ML) training data based on NDT. This utilizes NDT's high-fidelity network simulation capabilities to generate large amounts of high-quality data for training artificial intelligence (AI) / ML models. This addresses issues such as insufficient quantity, low quality, and imbalanced sample size in training data. However, because NDT generates data entirely from scratch, such as when generating a week's worth of traffic data, it is time-consuming, resulting in low data generation efficiency. Summary of the Invention

[0003] This application provides a model training, data receiving method, apparatus, device, medium, and program product to solve the problem of low data generation efficiency in related technologies.

[0004] To solve the above-mentioned technical problems, this application is implemented as follows:

[0005] In a first aspect, embodiments of this application provide a model training method applied to a first device, the first device being a machine learning (ML) training producer, the method comprising:

[0006] Receive simulation data sent by the second device, which is the management service MnS producer;

[0007] The first model is trained based on the simulation data to obtain the second model, wherein the first model is a data generation model or a strategy model.

[0008] Optionally, the second device has a network digital twin (NDT) function.

[0009] Optionally, the first model is a data generation model, and before receiving the simulation data sent by the second device, the method further includes:

[0010] Receive a first request sent by the MnS producer, the first request being used to request the generation of a data generation model;

[0011] The data generation model is generated based on the first request.

[0012] Optionally, training the first model based on the simulation data to obtain the second model includes:

[0013] The first data is generated using the data generation model;

[0014] The simulation data and the first data are compared to obtain a comparison result;

[0015] Based on the comparison results, the data generation model is trained, and the step of generating the first data using the data generation model is returned to be executed until the first preset condition is met, and the second model is obtained.

[0016] The method further includes:

[0017] The second model is sent to the MnS producer.

[0018] Optionally, the first model is a strategy model, and before receiving the simulation data sent by the second device, the method further includes:

[0019] Receive a second request sent by a third device, the second request being used to request the generation strategy model, the third device being an MnS consumer;

[0020] The strategy model is generated based on the second request.

[0021] Optionally, the method further includes:

[0022] Based on the second request, simulation information is determined, which includes at least one of simulation object, data type, and data requirements;

[0023] A third request is sent to the MnS producer, the third request carrying the simulation information, the third request being used to request the creation of an NDT instance;

[0024] Receive the second response message sent by the MnS producer.

[0025] Optionally, training the first model based on the simulation data to obtain the second model includes:

[0026] The reward function value is determined based on the simulation data;

[0027] The policy model is trained based on the reward function value, and the process returns to the step of determining the reward function value based on the simulation data until the second preset condition is met, thus obtaining the second model.

[0028] The method further includes:

[0029] The second model is sent to the MnS consumer.

[0030] Secondly, embodiments of this application provide a model training method applied to a second device, the second device being an MnS producer, the method comprising:

[0031] Create an NDT instance;

[0032] Simulation operations are performed based on the NDT instance to obtain simulation data, which is used to train a first model, which is a data generation model or a policy model.

[0033] The simulation data is sent to a first device, which is an ML training producer.

[0034] Optionally, creating an NDT instance includes:

[0035] The system receives a fourth request from a third device, which requests the creation of an NDT instance. The fourth request includes at least one of the following: simulation object, data type, and data requirements. The third device is an MnS consumer.

[0036] An NDT instance is created based on the fourth request;

[0037] Send a first response message to the MnS consumer.

[0038] Optionally, the first model is a data generation model, and before performing simulation operations based on the NDT instance to obtain simulation data, the method further includes:

[0039] Send a first request to the ML training producer, the first request being used to request the generation of a data generation model;

[0040] After sending the simulation data to the ML training producer, the method further includes:

[0041] Receive the second model sent by the ML training producer;

[0042] Execute the second model to obtain the target data;

[0043] The target data or the address of the target data is sent to the MnS consumer.

[0044] Optionally, creating an NDT instance includes:

[0045] Receive a third request sent by the ML training producer, the third request being used to request the creation of an NDT instance;

[0046] An NDT instance is created based on the third request;

[0047] Send a second response message to the ML training producer.

[0048] Thirdly, embodiments of this application provide a data receiving method applied to a third device, wherein the third device is an MnS consumer, the method comprising:

[0049] A fourth request is sent to the second device, the fourth request being used to request the creation of an NDT instance, the fourth request including at least one of simulation object, data type, and data requirements, the second device being an MnS producer;

[0050] Receive the first response message sent by the MnS producer;

[0051] Receive the target data or the address of the target data sent by the MnS producer;

[0052] or,

[0053] A second request is sent to the first device, the second request being used to request the generation of a policy model, the first device being an ML training producer;

[0054] Receive the second model sent by the ML training producer.

[0055] Fourthly, embodiments of this application provide a model training apparatus applied to a first device, the first device being a machine learning (ML) training producer, the apparatus comprising:

[0056] The first receiving module is used to receive simulation data sent by the second device, which is the management service MnS producer;

[0057] The training module is used to train the first model based on the simulation data to obtain the second model, wherein the first model is a data generation model or a strategy model.

[0058] Optionally, the second device has a network digital twin (NDT) function.

[0059] Optionally, the first model is a data generation model, and the apparatus further includes:

[0060] The second receiving module is used to receive a first request sent by the MnS producer, the first request being used to request the generation of a data generation model.

[0061] The first generation module is used to generate the data generation model based on the first request.

[0062] Optionally, the training module includes:

[0063] A generation unit is used to generate first data using the data generation model;

[0064] A comparison unit is used to compare the simulation data and the first data to obtain a comparison result;

[0065] The first training unit is used to train the data generation model based on the comparison results, and then return to the step of generating the first data using the data generation model until the first preset condition is met to obtain the second model.

[0066] The device further includes:

[0067] The first sending module is used to send the second model to the MnS producer.

[0068] Optionally, the first model is a strategy model, and the apparatus further includes:

[0069] The third receiving module is used to receive a second request sent by a third device, the second request being used to request the generation strategy model, the third device being an MnS consumer;

[0070] The second generation module is used to generate the strategy model based on the second request.

[0071] Optionally, the device further includes:

[0072] The determination module is used to determine simulation information based on the second request, wherein the simulation information includes at least one of simulation object, data type, and data requirements;

[0073] The second sending module is used to send a third request to the MnS producer, the third request carrying the simulation information and being used to request the creation of an NDT instance;

[0074] The fourth receiving module is used to receive the second response message sent by the MnS producer.

[0075] Optionally, the training module includes:

[0076] The determining unit is used to determine the reward function value based on the simulation data;

[0077] The second training unit is used to train the policy model based on the reward function value, and then return to the step of determining the reward function value based on the simulation data until the second preset condition is met to obtain the second model.

[0078] The device further includes:

[0079] The third sending module is used to send the second model to the MnS consumer.

[0080] Fifthly, embodiments of this application provide a first device, which is a machine learning (ML) training producer. The first device includes a transceiver and a processor.

[0081] The transceiver is used to receive simulation data sent by the second device, which is a management service MnS producer;

[0082] The processor is used to train the first model based on the simulation data to obtain the second model, wherein the first model is a data generation model or a strategy model.

[0083] Optionally, the second device has a network digital twin (NDT) function.

[0084] Optionally, the first model is a data generation model, and the transceiver is further configured to receive a first request sent by the MnS producer, the first request being used to request the generation of a data generation model;

[0085] The processor is also configured to generate the data generation model based on the first request.

[0086] Optionally, the processor is specifically used for:

[0087] The first data is generated using the data generation model;

[0088] The simulation data and the first data are compared to obtain a comparison result;

[0089] Based on the comparison results, the data generation model is trained, and the step of generating the first data using the data generation model is returned to be executed until the first preset condition is met, and the second model is obtained.

[0090] The transceiver is also used to send the second model to the MnS producer.

[0091] Optionally, the first model is a policy model, and the transceiver is further configured to receive a second request sent by a third device, the second request being used to request the generation of a policy model, the third device being an MnS consumer;

[0092] The processor is also configured to generate the policy model based on the second request.

[0093] Optionally, the processor is further configured to determine simulation information based on the second request, the simulation information including at least one of simulation object, data type, and data requirements;

[0094] The transceiver is also used for:

[0095] A third request is sent to the MnS producer, the third request carrying the simulation information, the third request being used to request the creation of an NDT instance;

[0096] Receive the second response message sent by the MnS producer.

[0097] Optionally, the processor is specifically used for:

[0098] The reward function value is determined based on the simulation data;

[0099] The policy model is trained based on the reward function value, and the process returns to the step of determining the reward function value based on the simulation data until the second preset condition is met, thus obtaining the second model.

[0100] The transceiver is also used for:

[0101] The second model is sent to the MnS consumer.

[0102] Sixthly, embodiments of this application provide a model training apparatus applied to a second device, the second device being an MnS producer, the apparatus comprising:

[0103] Create a module for creating NDT instances;

[0104] The simulation module is used to perform simulation operations based on the NDT instance to obtain simulation data, which is used to train a first model, which is a data generation model or a policy model.

[0105] The fourth sending module is used to send the simulation data to the first device, which is an ML training producer.

[0106] Optionally, the creation module includes:

[0107] The first receiving unit is used to receive a fourth request sent by a third device. The fourth request is used to request the creation of an NDT instance. The fourth request includes at least one of the following: simulation object, data type, and data requirements. The third device is an MnS consumer.

[0108] The first creation unit is used to create an NDT instance based on the fourth request;

[0109] The first sending unit is used to send a first response message to the MnS consumer.

[0110] Optionally, the first model is a data generation model, and the apparatus further includes:

[0111] The fifth sending module is used to send a first request to the ML training producer, the first request being used to request the generation of a data generation model;

[0112] The device further includes:

[0113] The fifth receiving module is used to receive the second model sent by the ML training producer;

[0114] The execution module is used to execute the second model to obtain the target data;

[0115] The sixth sending module is used to send the target data or the address of the target data to the MnS consumer.

[0116] Optionally, the creation module includes:

[0117] The second receiving unit is used to receive the third request sent by the ML training producer, the third request being used to request the creation of an NDT instance;

[0118] The second creation unit is used to create an NDT instance based on the third request;

[0119] The second sending unit is used to send a second response message to the ML training producer.

[0120] In a seventh aspect, embodiments of this application provide a second device, the second device being an MnS producer, the second device including a transceiver and a processor, the processor being used for:

[0121] Create an NDT instance;

[0122] Simulation operations are performed based on the NDT instance to obtain simulation data, which is used to train a first model, which is a data generation model or a policy model.

[0123] The transceiver is used to send the simulation data to a first device, which is an ML training producer.

[0124] Optionally, the processor is specifically used for:

[0125] The system receives a fourth request from a third device, which requests the creation of an NDT instance. The fourth request includes at least one of the following: simulation object, data type, and data requirements. The third device is an MnS consumer.

[0126] An NDT instance is created based on the fourth request;

[0127] Send a first response message to the MnS consumer.

[0128] Optionally, the first model is a data generation model, and the transceiver is further used for:

[0129] Send a first request to the ML training producer, the first request being used to request the generation of a data generation model;

[0130] Receive the second model sent by the ML training producer;

[0131] The processor is also used for:

[0132] Execute the second model to obtain the target data;

[0133] The target data or the address of the target data is sent to the MnS consumer.

[0134] Optionally, the processor is specifically used for:

[0135] Receive a third request sent by the ML training producer, the third request being used to request the creation of an NDT instance;

[0136] An NDT instance is created based on the third request;

[0137] Send a second response message to the ML training producer.

[0138] Eighthly, embodiments of this application provide a data receiving device applied to a third device, the third device being an MnS consumer, the device comprising:

[0139] The seventh sending module is used to send a fourth request to the second device. The fourth request is used to request the creation of an NDT instance. The fourth request includes at least one of the following: simulation object, data type, and data requirements. The second device is an MnS producer.

[0140] The sixth receiving module is used to receive the first response message sent by the MnS producer;

[0141] The seventh receiving module is used to receive the target data or the address of the target data sent by the MnS producer;

[0142] or,

[0143] The eighth sending module is used to send a second request to the first device, the second request being used to request the generation strategy model, the first device being an ML training producer;

[0144] The eighth receiving module is used to receive the second model sent by the ML training producer.

[0145] Ninthly, embodiments of this application provide a third device, the third device being an MnS consumer, the third device including a transceiver and a processor, the transceiver being used for:

[0146] A fourth request is sent to the second device, the fourth request being used to request the creation of an NDT instance, the fourth request including at least one of simulation object, data type, and data requirements, the second device being an MnS producer;

[0147] Receive the first response message sent by the MnS producer;

[0148] Receive the target data or the address of the target data sent by the MnS producer;

[0149] Alternatively, a second request is sent to the first device, which is a ML training producer, to request the generation of a policy model.

[0150] Receive the second model sent by the ML training producer.

[0151] In a tenth aspect, embodiments of this application provide an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, it implements the steps of the model training method as described in the first aspect above; or, when the program is executed by the processor, it implements the steps of the model training method as described in the second aspect above; or, when the program is executed by the processor, it implements the steps of the data receiving method as described in the third aspect above.

[0152] Eleventhly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the model training method as described in the first aspect above; or, when executed by a processor, implements the steps of the model training method as described in the second aspect above; or, when executed by a processor, implements the steps of the data receiving method as described in the third aspect above.

[0153] In this embodiment, the above-mentioned model training method is applied to an ML training producer. By receiving simulation data sent by an MnS producer, and training a first model based on the simulation data, a second model is obtained. The first model is a data generation model or a strategy model. This allows an AI-based data generation model to be obtained through the interaction between NDT and the ML training producer. The simulation data generated by NDT is used to train the data generation model, and then the data generation model is used to generate target data, thereby improving data generation efficiency. Attached Figure Description

[0154] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0155] Figure 1 This is one of the flowcharts of a model training method provided in the embodiments of this application;

[0156] Figure 2 This is one of the interactive schematic diagrams of a model training method provided in the embodiments of this application;

[0157] Figure 3 This is the second interactive schematic diagram of a model training method provided in the embodiments of this application;

[0158] Figure 4 This is the second flowchart of a model training method provided in the embodiments of this application;

[0159] Figure 5 This is one of the flowcharts of a data receiving method provided in the embodiments of this application;

[0160] Figure 6 This is a second flowchart of a data receiving method provided in an embodiment of this application;

[0161] Figure 7 This is one of the structural schematic diagrams of a model training device provided in the embodiments of this application;

[0162] Figure 8 This is a schematic diagram of the structure of a first device provided in an embodiment of this application;

[0163] Figure 9 This is a second schematic diagram of the structure of a model training device provided in an embodiment of this application;

[0164] Figure 10 This is a schematic diagram of the structure of a second device provided in an embodiment of this application;

[0165] Figure 11 This is one of the structural schematic diagrams of a data receiving device provided in the embodiments of this application;

[0166] Figure 12 This is a second schematic diagram of the structure of a data receiving device provided in an embodiment of this application;

[0167] Figure 13 This is a schematic diagram of the structure of a third device provided in an embodiment of this application. Detailed Implementation

[0168] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0169] In this application embodiment, a model training and data receiving method, apparatus, device, medium, and program product are proposed to solve the problem of low data generation efficiency in related technologies.

[0170] See Figure 1 , Figure 1 This is one of the flowcharts of a model training method provided in the embodiments of this application, applied to a first device, where the first device is a machine learning (ML) training producer, such as... Figure 1 As shown, the method includes the following steps:

[0171] Step 101: Receive simulation data sent by the second device, which is the management service MnS producer.

[0172] Specifically, the aforementioned ML training producer is responsible for generating or providing training data, setting up the training environment, and monitoring and managing the training process during machine learning. It can be a software platform or framework for creating, training, and optimizing machine learning models, or a hardware system providing efficient computing power for the machine learning models. The aforementioned simulation data can be data generated by a second device performing network simulation based on the Network Digital Twin (NDT) function, or it can be network state data obtained by running the initial model during network simulation. The aforementioned Management Service (MnS) producer can be a software platform or framework with NDT functionality, or it can be a hardware system.

[0173] Step 102: Train the first model based on the simulation data to obtain the second model. The first model is a data generation model or a strategy model.

[0174] Specifically, when the first model is a data generation model, training the first model based on the simulation data can be achieved by comparing the differences between the simulation data and the data generated by the first model to optimize the first model; when the first model is a policy model, training the first model based on the simulation data can be achieved by determining the accuracy of the first model based on the simulation data to update the first model.

[0175] In this embodiment, the above-mentioned model training method is applied to an ML training producer. By receiving simulation data sent by an MnS producer, and training a first model based on the simulation data, a second model is obtained. The first model is a data generation model or a strategy model. This allows an AI-based data generation model to be obtained through the interaction between NDT and the ML training producer. The simulation data generated by NDT is used to train the data generation model, and then the data generation model is used to generate target data, thereby improving data generation efficiency.

[0176] Optionally, the second device has a network digital twin (NDT) function.

[0177] Optionally, the first model is a data generation model, and before receiving the simulation data sent by the second device, the method further includes:

[0178] Receive a first request sent by the MnS producer, the first request being used to request the generation of a data generation model;

[0179] The data generation model is generated based on the first request.

[0180] Specifically, the first request may include a requirement for the data generation model, which may be a first model trained by an ML training producer according to the first request, used to generate the target data required by the MnS consumer, specifically an AI model or an ML model.

[0181] In this embodiment, the first model is a data generation model. The above-mentioned model training method receives a first request sent by the MnS producer and generates the data generation model based on the first request, so that an AI-based data generation model can be obtained through the interaction of NDT and ML training producers, thereby improving data generation efficiency.

[0182] Optionally, training the first model based on the simulation data to obtain the second model includes:

[0183] The first data is generated using the data generation model;

[0184] The simulation data and the first data are compared to obtain a comparison result;

[0185] Based on the comparison results, the data generation model is trained, and the step of generating the first data using the data generation model is returned to be executed until the first preset condition is met, and the second model is obtained.

[0186] The method further includes:

[0187] The second model is sent to the MnS producer.

[0188] Specifically, the first data and simulation data mentioned above can be simulation data corresponding to the same simulation object. The comparison result mentioned above can be the difference between the simulation data and the first data. The first preset condition mentioned above can be that the difference between the simulation data and the first data generated by the trained data generation model is lower than a preset value. Under the condition of satisfying the first preset condition, the second model satisfies the preset simulation requirements. The second model mentioned above can be the trained data generation model that meets the simulation requirements.

[0189] It is understandable that, if the first preset condition is not met, the steps of generating first data using the data generation model, comparing the simulation data with the first data, and training the data generation model based on the comparison result are repeated until the trained data generation model meets the preset simulation requirements.

[0190] In this embodiment, the above-mentioned model training method generates first data using the data generation model, compares the simulation data with the first data to obtain a comparison result, trains the data generation model based on the comparison result, and returns to execute the step of generating first data using the data generation model until the first preset condition is met to obtain a second model, thereby improving the realism of the data generated by the second model.

[0191] For example, Figure 2 This is one of the interactive schematic diagrams of a model training method provided in the embodiments of this application, such as... Figure 2 As shown, the specific steps of the method are as follows:

[0192] 1. MnS consumers send NDT creation / activation requests to MnS producers, including simulation objects (such as network elements, slices, etc.), data to be generated (traffic data, performance data, etc.), data volume, etc.

[0193] 2. MnS producers collect data based on the NDT function and complete the creation of NDT instances;

[0194] 3. The MnS producer notifies the MnS consumer that the required NDT function has been created;

[0195] 4. The MnS producer requests the ML training producer to send data to generate the model;

[0196] 5. ML training producers train data generation models based on demand;

[0197] 6. MnS producers generate data through simulation based on the NDT function;

[0198] 7. NDT sends the data to the ML training producer. The ML training producer compares the data generated by NDT with the data generated by the data generation model, and updates the data generation model. Step 7 is repeated until a data generation model that meets the preset requirements is trained.

[0199] 8. The MnS producer executes the updated data generation model to generate the target data;

[0200] 9. The MnS producer will send the target data or the address where the data is located back to the MnS consumer.

[0201] Optionally, the first model is a strategy model, and before receiving the simulation data sent by the second device, the method further includes:

[0202] Receive a second request sent by a third device, the second request being used to request the generation strategy model, the third device being an MnS consumer;

[0203] The strategy model is generated based on the second request.

[0204] Specifically, the second request may include requirements for the policy model, such as the type of model, training requirements for the model such as accuracy, training data, etc. The policy model may refer to a function that maps states to action probability distributions to determine which action to select in a given state.

[0205] Optionally, the method further includes:

[0206] Based on the second request, simulation information is determined, which includes at least one of simulation object, data type, and data requirements;

[0207] A third request is sent to the MnS producer, the third request carrying the simulation information, the third request being used to request the creation of an NDT instance;

[0208] Receive the second response message sent by the MnS producer.

[0209] Specifically, the aforementioned simulation object can be used to indicate the object studied or simulated during the simulation, such as network elements, slices, subnets, etc.; the aforementioned data type can be used to refer to the data category or format involved in the simulation process, such as traffic data, performance data, etc.; and the aforementioned data requirement can be used to refer to the amount of data required for the simulation process. The aforementioned third request can be a message sent by the ML training producer to request the MnS producer to create an NDT instance and execute network simulation. Specifically, it can include the policy model and the required network state data. The NDT instance is the simulation instance corresponding to the simulation information. The aforementioned second response information can be used to instruct the MnS producer to complete the creation of the NDT instance.

[0210] In this embodiment, the method determines simulation information based on the second request. The simulation information includes at least one of simulation object, data type, and data requirements. A third request is sent to the MnS producer, carrying the simulation information. The third request is used to request the creation of an NDT instance. The method also receives a second response message from the MnS producer. This allows the creation of an NDT instance required by the MnS consumer based on the interaction process between the ML training producer and the MnS producer. The NDT instance is then used to train the policy model for reinforcement learning without affecting the physical network.

[0211] Optionally, training the first model based on the simulation data to obtain the second model includes:

[0212] The reward function value is determined based on the simulation data;

[0213] The policy model is trained based on the reward function value, and the process returns to the step of determining the reward function value based on the simulation data until the second preset condition is met, thus obtaining the second model.

[0214] The method further includes:

[0215] The second model is sent to the MnS consumer.

[0216] Specifically, the simulation data mentioned above can be network state data obtained by the MnS producer performing network simulation and running the policy model. The reward function value mentioned above can be a numerical value fed back by the environment after the agent takes a certain action in reinforcement learning. This value can be used to measure the quality of the action and guide the agent to adjust its behavioral strategy to maximize future cumulative rewards. Specifically, it can be obtained by the ML training producer calculating the reward function based on the simulation data and updating the policy model based on the reward function value until a second preset condition is met. The second preset condition can refer to the reward function value corresponding to the trained policy model meeting a preset range; the second model is the trained policy model that meets the second preset condition.

[0217] In this embodiment, the above-mentioned model training method determines the reward function value based on the simulation data, trains the policy model based on the reward function value, and returns to execute the step of determining the reward function value based on the simulation data until the second preset condition is met, thereby obtaining the second model and improving the accuracy of the policy data generated by the second model.

[0218] For example, Figure 3 This is the second interactive schematic diagram of a model training method provided in the embodiments of this application, as shown below. Figure 3 As shown, the specific steps of the method are as follows:

[0219] 1. The MnS consumer sends a reinforcement learning training request, including the model type and the model's training requirements such as accuracy and the data required for training.

[0220] 2. The ML training producer determines the simulation object (such as network element, slice, subnet, etc.) and the data required for simulation (such as traffic data, performance data, etc.) based on the demand information in the reinforcement learning training request;

[0221] 3. The ML training producer sends an NDT creation / activation request to the MnS producer, including the simulation object and the data required for the simulation;

[0222] 4. MnS producers collect data based on the NDT function and complete the creation of NDT instances;

[0223] 5. The MnS producer notifies the ML training producer that the required NDT function has been created;

[0224] 6. ML training: Producers train an initial strategy model based on demand information.

[0225] 7. The ML training producer sends a simulation request to the MnS producer, which includes the policy model and the required network state data.

[0226] 8. The MnS producer performs network simulation based on the NDT function and runs the strategy model;

[0227] 9. The MnS producer feeds back the network state data after running the policy model to the ML training producer;

[0228] 10. The ML training producer calculates the reward function based on the network state data, updates the policy model based on the reward function, and repeats steps 7-10 until a policy model that meets the preset requirements is trained.

[0229] 11. The ML-trained producer feeds the trained policy model back to the MnS consumer.

[0230] See Figure 4 , Figure 4 This is a second flowchart of a model training method provided in an embodiment of this application, applied to a second device, which is an MnS producer, such as... Figure 4 As shown, the method includes the following steps:

[0231] Step 401: Create an NDT instance;

[0232] Step 402: Perform simulation operations based on the NDT instance to obtain simulation data. The simulation data is used to train the first model, which is a data generation model or a policy model.

[0233] Step 403: Send the simulation data to the first device, which is an ML training producer.

[0234] It should be noted that this embodiment is used as a reference for... Figure 1 The implementation method of the first device corresponding to the illustrated embodiment can be found in the following examples. Figure 1 To avoid repetition, the relevant descriptions in the embodiments shown will not be repeated in this embodiment.

[0235] Optionally, creating an NDT instance includes:

[0236] The system receives a fourth request from a third device, which requests the creation of an NDT instance. The fourth request includes at least one of the following: simulation object, data type, and data requirements. The third device is an MnS consumer.

[0237] An NDT instance is created based on the fourth request;

[0238] Send a first response message to the MnS consumer.

[0239] Optionally, the first model is a data generation model, and before performing simulation operations based on the NDT instance to obtain simulation data, the method further includes:

[0240] Send a first request to the ML training producer, the first request being used to request the generation of a data generation model;

[0241] After sending the simulation data to the ML training producer, the method further includes:

[0242] Receive the second model sent by the ML training producer;

[0243] Execute the second model to obtain the target data;

[0244] The target data or the address of the target data is sent to the MnS consumer.

[0245] Optionally, creating an NDT instance includes:

[0246] Receive a third request sent by the ML training producer, the third request being used to request the creation of an NDT instance;

[0247] An NDT instance is created based on the third request;

[0248] Send a second response message to the ML training producer.

[0249] The above optional implementation methods can be found in [reference]. Figure 1 To avoid repetition, the relevant descriptions in the embodiments shown will not be repeated in this embodiment.

[0250] See Figure 5 , Figure 5 This is a flowchart of a data receiving method provided in an embodiment of this application, applied to a third device, wherein the third device is an MnS consumer, such as... Figure 5 As shown, the method includes the following steps:

[0251] Step 501: Send a fourth request to the second device. The fourth request is used to request the creation of an NDT instance. The fourth request includes at least one of the following: simulation object, data type, and data requirements. The second device is an MnS producer.

[0252] Step 502: Receive the first response message sent by the MnS producer;

[0253] Step 503: Receive the target data or the address of the target data sent by the MnS producer;

[0254] Or, such as Figure 6 As shown, the method includes the following steps:

[0255] Step 601: Send a second request to the first device, the second request being used to request the generation of a strategy model, the first device being an ML training producer;

[0256] Step 602: Receive the second model sent by the ML training producer.

[0257] It should be noted that this embodiment is used as a reference for... Figure 1 The implementation method of the first device corresponding to the illustrated embodiment can be found in the following examples. Figure 1 To avoid repetition, the relevant descriptions in the embodiments shown will not be repeated in this embodiment.

[0258] See Figure 7 , Figure 7 This is one of the structural schematic diagrams of a model training device provided in an embodiment of this application, applied to a first device, which is a machine learning (ML) training producer, such as... Figure 7 As shown, the model training device 700 includes:

[0259] The first receiving module 701 is used to receive simulation data sent by the second device, which is a management service MnS producer;

[0260] The training module 702 is used to train the first model based on the simulation data to obtain the second model, wherein the first model is a data generation model or a strategy model.

[0261] Optionally, the second device has a network digital twin (NDT) function.

[0262] Optionally, the first model is a data generation model, and the apparatus further includes:

[0263] The second receiving module is used to receive a first request sent by the MnS producer, the first request being used to request the generation of a data generation model.

[0264] The first generation module is used to generate the data generation model based on the first request.

[0265] Optionally, the training module includes:

[0266] A generation unit is used to generate first data using the data generation model;

[0267] A comparison unit is used to compare the simulation data and the first data to obtain a comparison result;

[0268] The first training unit is used to train the data generation model based on the comparison results, and then return to the step of generating the first data using the data generation model until the first preset condition is met to obtain the second model.

[0269] The device further includes:

[0270] The first sending module is used to send the second model to the MnS producer.

[0271] Optionally, the first model is a strategy model, and the apparatus further includes:

[0272] The third receiving module is used to receive a second request sent by a third device, the second request being used to request the generation strategy model, the third device being an MnS consumer;

[0273] The second generation module is used to generate the strategy model based on the second request.

[0274] Optionally, the device further includes:

[0275] The determination module is used to determine simulation information based on the second request, wherein the simulation information includes at least one of simulation object, data type, and data requirements;

[0276] The second sending module is used to send a third request to the MnS producer, the third request carrying the simulation information and being used to request the creation of an NDT instance;

[0277] The fourth receiving module is used to receive the second response message sent by the MnS producer.

[0278] Optionally, the training module includes:

[0279] The determining unit is used to determine the reward function value based on the simulation data;

[0280] The second training unit is used to train the policy model based on the reward function value, and then return to the step of determining the reward function value based on the simulation data until the second preset condition is met to obtain the second model.

[0281] The device further includes:

[0282] The third sending module is used to send the second model to the MnS consumer.

[0283] It should be noted that the model training apparatus provided in this application embodiment is an apparatus capable of executing the above-described model training method. Therefore, all implementation methods in the above-described model training method embodiments are applicable to this apparatus and can achieve the same or similar beneficial effects. To avoid repetition, this embodiment will not elaborate further.

[0284] For details, see Figure 8 As shown in the figure, this application embodiment also provides a first device, which is a machine learning (ML) training producer, including a bus 801, a transceiver 802, an antenna 803, a bus interface 804, a processor 805, and a memory 806.

[0285] The transceiver 802 is used to receive simulation data sent by the second device, which is a management service MnS producer;

[0286] Furthermore, the processor 805 is used to train the first model based on the simulation data to obtain the second model, wherein the first model is a data generation model or a strategy model.

[0287] exist Figure 8In this document, a bus architecture (represented by bus 801) is used. Bus 801 can include any number of interconnected buses and bridges, linking various circuits including one or more processors represented by processor 805 and memory represented by memory 806. Bus 801 can also link 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. Bus interface 804 provides an interface between bus 801 and transceiver 802. Transceiver 802 can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 805 is transmitted over a wireless medium via antenna 803, which further receives data and transmits data to processor 805.

[0288] The processor 805 manages the bus 801 and handles general processing, and also provides various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. The memory 806 can be used to store data used by the processor 805 during operation.

[0289] Optionally, the processor 805 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a complex programmable logic device (CPLD).

[0290] Optionally, the second device has a network digital twin (NDT) function.

[0291] Optionally, the first model is a data generation model, and the transceiver 802 is further configured to receive a first request sent by the MnS producer, the first request being used to request the generation of a data generation model;

[0292] The processor 805 is also configured to generate the data generation model based on the first request.

[0293] Optionally, the processor 805 is specifically used for:

[0294] The first data is generated using the data generation model;

[0295] The simulation data and the first data are compared to obtain a comparison result;

[0296] Based on the comparison results, the data generation model is trained, and the step of generating the first data using the data generation model is returned to be executed until the first preset condition is met, and the second model is obtained.

[0297] The transceiver 802 is also used to send the second model to the MnS producer.

[0298] Optionally, the first model is a policy model, and the transceiver 802 is further configured to receive a second request sent by a third device, the second request being used to request the generation of a policy model, the third device being an MnS consumer;

[0299] The processor 805 is also configured to generate the policy model based on the second request.

[0300] Optionally, the processor 805 is further configured to determine simulation information based on the second request, the simulation information including at least one of simulation object, data type, and data requirements;

[0301] The transceiver 802 is also used for:

[0302] A third request is sent to the MnS producer, the third request carrying the simulation information, the third request being used to request the creation of an NDT instance;

[0303] Receive the second response message sent by the MnS producer.

[0304] Optionally, the processor 805 is specifically used for:

[0305] The reward function value is determined based on the simulation data;

[0306] The policy model is trained based on the reward function value, and the process returns to the step of determining the reward function value based on the simulation data until the second preset condition is met, thus obtaining the second model.

[0307] The transceiver 802 is also used for:

[0308] The second model is sent to the MnS consumer.

[0309] It should be noted that the electronic device provided in this application embodiment is a device capable of executing the above-described model training method. Therefore, all implementation methods in the above-described model training method embodiments are applicable to this electronic device and can achieve the same or similar beneficial effects. To avoid repetition, this embodiment will not elaborate further.

[0310] See Figure 9 , Figure 9This is a second schematic diagram of a model training device provided in an embodiment of this application, applied to a second device, which is an MnS producer, such as... Figure 9 As shown, the model training device 900 includes:

[0311] Create module 901 to create NDT instances;

[0312] Simulation module 902 is used to perform simulation operations based on the NDT instance to obtain simulation data, which is used to train a first model, the first model being a data generation model or a policy model.

[0313] The fourth sending module 903 is used to send the simulation data to the first device, which is an ML training producer.

[0314] Optionally, the creation module 901 includes:

[0315] The first receiving unit is used to receive a fourth request sent by a third device. The fourth request is used to request the creation of an NDT instance. The fourth request includes at least one of the following: simulation object, data type, and data requirements. The third device is an MnS consumer.

[0316] The first creation unit is used to create an NDT instance based on the fourth request;

[0317] The first sending unit is used to send a first response message to the MnS consumer.

[0318] Optionally, the first model is a data generation model, and the apparatus further includes:

[0319] The fifth sending module is used to send a first request to the ML training producer, the first request being used to request the generation of a data generation model;

[0320] The device further includes:

[0321] The fifth receiving module is used to receive the second model sent by the ML training producer;

[0322] The execution module is used to execute the second model to obtain the target data;

[0323] The sixth sending module is used to send the target data or the address of the target data to the MnS consumer.

[0324] Optionally, the creation module 901 includes:

[0325] The second receiving unit is used to receive the third request sent by the ML training producer, the third request being used to request the creation of an NDT instance;

[0326] The second creation unit is used to create an NDT instance based on the third request;

[0327] The second sending unit is used to send a second response message to the ML training producer.

[0328] It should be noted that the model training apparatus provided in this application embodiment is an apparatus capable of executing the above-described model training method. Therefore, all implementation methods in the above-described model training method embodiments are applicable to this apparatus and can achieve the same or similar beneficial effects. To avoid repetition, this embodiment will not elaborate further.

[0329] For details, see Figure 10 As shown in the embodiment of this application, a second device is also provided. The second device is an MnS producer, including a bus 1001, a transceiver 1002, an antenna 1003, a bus interface 1004, a processor 1005, and a memory 1006.

[0330] The processor 1005 is used for:

[0331] Create an NDT instance;

[0332] Simulation operations are performed based on the NDT instance to obtain simulation data, which is used to train a first model, which is a data generation model or a policy model.

[0333] The transceiver 1002 is used to send the simulation data to a first device, which is an ML training producer.

[0334] exist Figure 10 In this document, a bus architecture (represented by bus 1001) is used. Bus 1001 may include any number of interconnected buses and bridges, linking various circuits including one or more processors represented by processor 1005 and memory represented by memory 1006. Bus 1001 may also link 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. Bus interface 1004 provides an interface between bus 1001 and transceiver 1002. Transceiver 1002 may be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 1005 is transmitted over a wireless medium via antenna 1003, which further receives data and transmits it to processor 1005.

[0335] Processor 1005 is responsible for managing bus 1001 and general processing, and can also provide various functions, including timing, peripheral interface, voltage regulation, power management, and other control functions. Memory 1006 can be used to store data used by processor 1005 during operation.

[0336] Optionally, the processor 1005 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a complex programmable logic device (CPLD).

[0337] Optionally, the processor 1005 is specifically used for:

[0338] The system receives a fourth request from a third device, which requests the creation of an NDT instance. The fourth request includes at least one of the following: simulation object, data type, and data requirements. The third device is an MnS consumer.

[0339] An NDT instance is created based on the fourth request;

[0340] Send a first response message to the MnS consumer.

[0341] Optionally, the first model is a data generation model, and the transceiver 1002 is further used for:

[0342] Send a first request to the ML training producer, the first request being used to request the generation of a data generation model;

[0343] Receive the second model sent by the ML training producer;

[0344] The processor 1005 is further configured to: execute the second model to obtain target data;

[0345] The target data or the address of the target data is sent to the MnS consumer.

[0346] Optionally, the processor 1005 is specifically used for:

[0347] Receive a third request sent by the ML training producer, the third request being used to request the creation of an NDT instance;

[0348] An NDT instance is created based on the third request;

[0349] Send a second response message to the ML training producer.

[0350] It should be noted that the electronic device provided in this application embodiment is a device capable of executing the above-described model training method. Therefore, all implementation methods in the above-described model training method embodiments are applicable to this electronic device and can achieve the same or similar beneficial effects. To avoid repetition, this embodiment will not elaborate further.

[0351] See Figure 11 , Figure 11 This is a schematic diagram of the structure of a data receiving device provided in an embodiment of this application, applied to a third device, wherein the third device is an MnS consumer, such as... Figure 11 As shown, the data receiving device 1100 includes:

[0352] The seventh sending module 1101 is used to send a fourth request to the second device. The fourth request is used to request the creation of an NDT instance. The fourth request includes at least one of the following: simulation object, data type, and data requirements. The second device is an MnS producer.

[0353] The sixth receiving module 1102 is used to receive the first response message sent by the MnS producer;

[0354] The seventh receiving module 1103 is used to receive the target data or the address of the target data sent by the MnS producer;

[0355] Or, such as Figure 12 As shown, the data receiving device 1200 includes:

[0356] The eighth sending module 1201 is used to send a second request to the first device, the second request being used to request the generation strategy model, the first device being an ML training producer;

[0357] The eighth receiving module 1202 is used to receive the second model sent by the ML training producer.

[0358] It should be noted that the data receiving device provided in this application embodiment is a device capable of executing the above-described data receiving method. Therefore, all implementation methods in the above-described model training method embodiments are applicable to this device and can achieve the same or similar beneficial effects. To avoid repetition, this embodiment will not elaborate further.

[0359] For details, see Figure 13 As shown in the figure, this application embodiment also provides a third device, which is an MnS consumer, including a bus 1301, a transceiver 1302, an antenna 1303, a bus interface 1304, a processor 1305, and a memory 1306.

[0360] The transceiver 1302 is used for:

[0361] A fourth request is sent to the second device, the fourth request being used to request the creation of an NDT instance, the fourth request including at least one of simulation object, data type, and data requirements, the second device being an MnS producer;

[0362] Receive the first response message sent by the MnS producer;

[0363] Receive the target data or the address of the target data sent by the MnS producer;

[0364] Alternatively, a second request is sent to the first device, which is a ML training producer, to request the generation of a policy model.

[0365] Receive the second model sent by the ML training producer.

[0366] exist Figure 13 In this document, a bus architecture (represented by bus 1301) is used. Bus 1301 can include any number of interconnected buses and bridges, linking various circuits including one or more processors represented by processor 1305 and memory represented by memory 1306. Bus 1301 can also link 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. Bus interface 1304 provides an interface between bus 1301 and transceiver 1302. Transceiver 1302 can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 1305 is transmitted over a wireless medium via antenna 1303, which further receives data and transmits it to processor 1305.

[0367] Processor 1305 manages bus 1301 and general processing, and also provides various functions, including timing, peripheral interface, voltage regulation, power management, and other control functions. Memory 1306 can be used to store data used by processor 1305 during operation.

[0368] Optionally, the processor 1305 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a complex programmable logic device (CPLD).

[0369] It should be noted that the electronic device provided in this application embodiment is a device capable of executing the above-described data receiving method. Therefore, all implementation methods in the above-described data receiving method embodiments are applicable to this electronic device and can achieve the same or similar beneficial effects. To avoid repetition, this embodiment will not elaborate further.

[0370] This application also provides an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor. When the program is executed by the processor, it implements the various processes of the above-described model training method or data receiving method embodiments and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0371] This application also provides a computer-readable storage medium storing a computer program. When executed by a processor, this computer program implements the various processes of the above-described model training method or data receiving method embodiments, achieving the same technical effects. To avoid repetition, further details are omitted here. The computer-readable storage medium may include, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0372] This application also provides a computer program product, including computer instructions. When executed by a processor, these computer instructions implement the various processes of the above-described model training method or data receiving method embodiments and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0373] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0374] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0375] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A model training method, characterized in that, Applied to a first device, which is a machine learning (ML) training producer, the method includes: Receive simulation data sent by the second device, which is the management service MnS producer; The first model is trained based on the simulation data to obtain the second model, wherein the first model is a data generation model or a strategy model.

2. The method according to claim 1, characterized in that, The second device has Network Digital Twin (NDT) functionality.

3. The method according to claim 1, characterized in that, The first model is a data generation model. Before receiving the simulation data sent by the second device, the method further includes: Receive a first request sent by the MnS producer, the first request being used to request the generation of a data generation model; The data generation model is generated based on the first request.

4. The method according to claim 3, characterized in that, The step of training the first model based on the simulation data to obtain the second model includes: The first data is generated using the data generation model; The simulation data and the first data are compared to obtain a comparison result; Based on the comparison results, the data generation model is trained, and the step of generating the first data using the data generation model is returned to be executed until the first preset condition is met, and the second model is obtained. The method further includes: The second model is sent to the MnS producer.

5. The method according to claim 1, characterized in that, The first model is a strategy model. Before receiving simulation data sent by the second device, the method further includes: Receive a second request sent by a third device, the second request being used to request the generation strategy model, the third device being an MnS consumer; The strategy model is generated based on the second request.

6. The method according to claim 5, characterized in that, The method further includes: Based on the second request, simulation information is determined, which includes at least one of simulation object, data type, and data requirements; A third request is sent to the MnS producer, the third request carrying the simulation information, the third request being used to request the creation of an NDT instance; Receive the second response message sent by the MnS producer.

7. The method according to claim 5, characterized in that, The step of training the first model based on the simulation data to obtain the second model includes: The reward function value is determined based on the simulation data; The policy model is trained based on the reward function value, and the process returns to the step of determining the reward function value based on the simulation data until the second preset condition is met, thus obtaining the second model. The method further includes: The second model is sent to the MnS consumer.

8. A model training method, characterized in that, Applied to a second device, which is a MnS producer, the method includes: Create an NDT instance; Simulation operations are performed based on the NDT instance to obtain simulation data, which is used to train a first model, which is a data generation model or a policy model. The simulation data is sent to a first device, which is an ML training producer.

9. The method according to claim 8, characterized in that, The creation of an NDT instance includes: Receive a fourth request sent by a third device, the fourth request being used to request the creation of an NDT instance, the fourth request including at least one of simulation object, data type and data requirements, the third device being an MnS consumer; An NDT instance is created based on the fourth request; Send a first response message to the MnS consumer.

10. The method according to claim 9, characterized in that, The first model is a data generation model. Before performing simulation operations based on the NDT instance to obtain simulation data, the method further includes: Send a first request to the ML training producer, the first request being used to request the generation of a data generation model; After sending the simulation data to the ML training producer, the method further includes: Receive the second model sent by the ML training producer; Execute the second model to obtain the second data; Send the second data or the address of the second data to the MnS consumer.

11. The method according to claim 8, characterized in that, The creation of an NDT instance includes: Receive a third request sent by the ML training producer, the third request being used to request the creation of an NDT instance; An NDT instance is created based on the third request; Send a second response message to the ML training producer.

12. A data receiving method, characterized in that, Applied to a third device, wherein the third device is an MnS consumer, the method includes: A fourth request is sent to the second device, the fourth request being used to request the creation of an NDT instance, the fourth request including at least one of simulation object, data type and data requirements, the second device being an MnS producer; Receive the first response message sent by the MnS producer; Receive the second data or the address of the second data sent by the MnS producer; or, A second request is sent to the first device, the second request being used to request the generation of a policy model, the first device being an ML training producer; Receive the second model sent by the ML training producer.

13. A model training device, characterized in that, Applied to a first device, the first device being a machine learning (ML) training producer, the apparatus includes: The first receiving module is used to receive simulation data sent by the second device, which is the management service MnS producer; The training module is used to train the first model based on the simulation data to obtain the second model, wherein the first model is a data generation model or a strategy model.

14. A first device, characterized in that, The first device is a machine learning (ML) training producer, and the first device includes a transceiver and a processor. The transceiver is used to receive simulation data sent by the second device, which is a management service MnS producer; The processor is used to train the first model based on the simulation data to obtain the second model, wherein the first model is a data generation model or a strategy model.

15. A model training device, characterized in that, Applied to a second device, which is a MnS producer, the apparatus includes: Create a module for creating NDT instances; The simulation module is used to perform simulation operations based on the NDT instance to obtain simulation data, which is used to train a first model, which is a data generation model or a policy model. The fourth sending module is used to send the simulation data to the first device, which is an ML training producer.

16. A second device, characterized in that, The second device is an MnS producer, and the second device includes a transceiver and a processor, the processor being used for: Create an NDT instance; Simulation operations are performed based on the NDT instance to obtain simulation data, which is used to train a first model, which is a data generation model or a policy model. The transceiver is used to send the simulation data to a first device, which is an ML training producer.

17. A data receiving device, characterized in that, Applied to a third device, said third device being an MnS consumer, said device includes: The seventh sending module is used to send a fourth request to the second device. The fourth request is used to request the creation of an NDT instance. The fourth request includes at least one of the following: simulation object, data type, and data requirements. The second device is an MnS producer. The sixth receiving module is used to receive the first response message sent by the MnS producer; The seventh receiving module is used to receive the second data or the address of the second data sent by the MnS producer; or, The eighth sending module is used to send a second request to the first device, the second request being used to request the generation strategy model, the first device being an ML training producer; The eighth receiving module is used to receive the second model sent by the ML training producer.

18. A third device, characterized in that, The third device is an MnS consumer, and the third device includes a transceiver and a processor. The transceiver is used for: A fourth request is sent to the second device, the fourth request being used to request the creation of an NDT instance, the fourth request including at least one of simulation object, data type, and data requirements, the second device being an MnS producer; Receive the first response message sent by the MnS producer; Receive the target data or the address of the target data sent by the MnS producer; Alternatively, a second request is sent to the first device, which is a ML training producer, to request the generation of a policy model. Receive the second model sent by the ML training producer.

19. An electronic device, characterized in that, include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the model training method as described in any one of claims 1 to 7; or, the program, when executed by the processor, implements the steps of the model training method as described in any one of claims 8 to 11; or, the program, when executed by the processor, implements the steps of the model training method as described in claim 12.

20. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the model training method as described in any one of claims 1 to 7; or, when executed by a processor, the computer program implements the steps of the model training method as described in any one of claims 8 to 11; or, when executed by a processor, the computer program implements the steps of the model training method as described in claim 12.

21. A computer program product, characterized in that, The method includes computer instructions that, when executed by a processor, implement the steps of the model training method as described in any one of claims 1 to 7; or, when executed by a processor, the computer instructions implement the steps of the model training method as described in any one of claims 8 to 11; or, when executed by a processor, the computer instructions implement the steps of the model training method as described in claim 12.

Images