Data processing method and apparatus, and device and storage medium

Abstract

Provided in the present disclosure are a data processing method and apparatus, and a device and a storage medium. The method comprises: acquiring an AI service requirement by means of a first logic function plane; and by means of the first logic function plane and / or a second logic function plane, acquiring AI data corresponding to the AI service requirement, wherein the first logic function plane is a data plane, and the second logic function plane is an intelligence plane; or, the first logic function plane is the intelligence plane, and the second logic function plane is the data plane. Data collaboration and service collaboration can be accurately performed between the data plane and the intelligence plane, and the development of repeated functions can be reduced, thereby simplifying the system design and implementation of logic function planes, and reducing system development costs.

Description

Data processing methods, apparatus, equipment and storage media

[0001] This disclosure claims priority to Chinese Patent Application No. 202410500504.3, filed on April 24, 2024, entitled “Data Processing Method, Apparatus, Device and Storage Medium”, the entire contents of which are incorporated herein by reference. Technical Field

[0002] This disclosure relates to the field of communication technology, and to a data processing method, apparatus, device, and storage medium. Background Technology

[0003] In sixth-generation mobile communication technology, the network architecture can include logical functional planes for artificial intelligence (AI) services to improve the communication quality of sixth-generation mobile communication.

[0004] Currently, sixth-generation mobile communication technology only proposes that the network architecture can introduce a logical functional plane with AI services. However, how this logical functional plane can interact with the data plane (responsible for data services in mobile communication) is an urgent problem to be solved. Summary of the Invention

[0005] This disclosure relates to a data processing method, apparatus, device, and storage medium for solving the problem of information interaction between the logical functional plane and the data plane of AI services in related technologies.

[0006] In a first aspect, this disclosure provides a data processing method, which includes:

[0007] AI service requirements are obtained through the first logical functional plane;

[0008] AI data corresponding to AI service requirements is obtained through the first logical functional plane and / or the second logical functional plane.

[0009] Wherein, the first logical functional surface is the data surface, and the second logical functional surface is the intelligence surface; or, the first logical functional surface is the intelligence surface, and the second logical functional surface is the data surface.

[0010] In some embodiments, the AI ​​service requirements include at least one of the following:

[0011] Identification of the target AI model;

[0012] Information about the target AI model;

[0013] Identification of data related to the AI ​​model used;

[0014] The identifier of the training data for the AI ​​model to be trained.

[0015] In some embodiments, the AI ​​data includes at least one of the following:

[0016] Target AI model;

[0017] Data related to the AI ​​model used;

[0018] Training data for the AI ​​model to be trained.

[0019] In some embodiments, the information of the target AI model includes the type of the target AI model, the size of the parameter set of the target AI model, and the storage location of the target AI model;

[0020] In some embodiments, the data related to the AI ​​model used includes the input data and output data corresponding to the AI ​​model used.

[0021] In some embodiments, the first logical functional plane includes a resource pool; obtaining AI data corresponding to AI service requirements through the first logical functional plane includes:

[0022] In the resource pool of the first logical functional plane, obtain the AI ​​data corresponding to the AI ​​service requirement;

[0023] The AI ​​data refers to the AI ​​data collected by the second logical functional plane and stored in the resource pool of the first logical functional plane.

[0024] In some embodiments, the second logical functional plane includes a resource pool; obtaining AI data corresponding to AI service requirements through the second logical functional plane includes:

[0025] In the resource pool of the second logical functional plane, obtain the AI ​​data corresponding to the AI ​​service requirement;

[0026] The AI ​​data can be either AI data collected by the second logical functional surface and stored in the resource pool of the second logical functional surface, or AI data collected by the first logical functional surface and stored in the resource pool of the second logical functional surface.

[0027] In some embodiments, the first logical functional plane and the second logical functional plane include a shared resource pool; obtaining AI data corresponding to AI service requirements through the first logical functional plane and the second logical functional plane includes:

[0028] From the shared resource pool, obtain the AI ​​data corresponding to the AI ​​service request;

[0029] The AI ​​data refers to the AI ​​data collected by the second logical functional plane and stored in the shared resource pool.

[0030] In some embodiments, obtaining AI data corresponding to the AI ​​service request includes:

[0031] Determine the search criteria corresponding to the AI ​​service requirements;

[0032] The AI ​​data is retrieved from the resource pool based on the search criteria.

[0033] In some embodiments, the intelligent surface includes a first module and a second module. The first module is used to acquire the AI ​​data and process the AI ​​data. The second module is used to acquire AI service requests and provide feedback to the target user on the processing results corresponding to the AI ​​service requests.

[0034] In some embodiments, the data plane includes an acquisition module, a storage module, and a retrieval module, wherein the acquisition module is used to acquire AI data, the storage module is used to store AI data, and the retrieval module is used to query AI data that matches AI service requirements from the stored AI data.

[0035] In some embodiments, the first logical functional plane is an intelligence plane, the second logical functional plane is a data plane, the data plane includes a resource pool, and the AI ​​data is training data; obtaining AI data corresponding to AI service requirements through the first and second logical functional planes includes:

[0036] The AI ​​service request is sent to the first module of the intelligent surface through the second module of the intelligent surface;

[0037] The first module in the intelligent plane obtains the training data corresponding to the AI ​​service requirement and at least one AI node from the resource pool of the data plane, wherein the at least one AI node is a candidate node for AI model training.

[0038] In some embodiments, the first logical functional plane is a data plane, the second logical functional plane is an intelligence plane, the intelligence plane includes a resource pool, and the AI ​​data is the training data; obtaining AI data corresponding to AI service requirements through the first and second logical functional planes includes:

[0039] The AI ​​service request is sent from the second module in the intelligent plane to the first module in the intelligent plane. The AI ​​service request is the AI ​​service request obtained by the second module through the data plane.

[0040] The first module in the intelligent surface acquires training data and at least one AI node from the resource pool of the intelligent surface, wherein the at least one AI node is a candidate node for training the AI ​​model.

[0041] In some embodiments, the method further comprises:

[0042] The first module processes the at least one AI node to obtain an AI scheduling strategy, which includes the target AI node and the path information of the target AI node.

[0043] The first module sends the AI ​​scheduling strategy and the training data to the second module, and the second module sends the AI ​​scheduling strategy and the training data to the target user.

[0044] In some embodiments, the method further comprises:

[0045] The first module is used to pre-allocate AI resources and deploy AI services at the target AI node.

[0046] In some embodiments, the method further comprises:

[0047] The AI ​​data subscription information is sent to the shared resource pool through the first logical functional plane;

[0048] The subscription information is sent to the second logical functional plane through the shared resource pool;

[0049] The AI ​​data is updated in the shared resource pool using the second logical functional plane and the subscription information;

[0050] The updated AI data is sent to the first logical functional plane through the shared resource pool.

[0051] In some embodiments, the intelligent plane includes AI control functionality, and the data plane includes a data orchestrator (DO), at least one data agent (DA), and a data storage function (DRF), wherein...

[0052] The DO is used to receive AI service requests sent by the AICF and to acquire the data service capabilities of at least one DA;

[0053] The DO is also used to generate a logical network according to the AI ​​service requirements, and to obtain the data corresponding to the target DA in the DRF, wherein the logical network includes the target DA, and the data service capabilities of the target DA are matched with the AI ​​service requirements;

[0054] The DA is used to send the data service capabilities corresponding to the DA to the DO, and to send the data corresponding to the DA to the DRF;

[0055] The DA is also used to execute the data service capabilities corresponding to the DA.

[0056] Secondly, this disclosure provides a data processing apparatus, which includes a first acquisition module and a second acquisition module, wherein:

[0057] The first acquisition module is used to acquire AI service requirements through the first logical functional plane;

[0058] The second acquisition module is used to acquire AI data corresponding to AI service requirements through the first logical functional plane and / or the second logical functional plane;

[0059] Wherein, the first logical functional surface is the data surface, and the second logical functional surface is the intelligence surface; or, the first logical functional surface is the intelligence surface, and the second logical functional surface is the data surface.

[0060] Thirdly, this disclosure provides an electronic device, the network device including a memory, a transceiver, and a processor:

[0061] The memory is used to store computer programs;

[0062] The transceiver is used to send and receive data under the control of the processor;

[0063] The processor is configured to read the computer program from the memory and perform the following operations:

[0064] AI service requirements are obtained through the first logical functional plane;

[0065] AI data corresponding to AI service requirements is obtained through the first logical functional plane and / or the second logical functional plane.

[0066] Wherein, the first logical functional surface is the data surface, and the second logical functional surface is the intelligence surface; or, the first logical functional surface is the intelligence surface, and the second logical functional surface is the data surface.

[0067] Fourthly, this disclosure provides a processor-readable storage medium storing a computer program for causing a processor to perform the method described in the first aspect.

[0068] This disclosure relates to a data processing method, apparatus, device, and storage medium. An electronic device can obtain AI service requests through a first logical functional plane and acquire AI data corresponding to the AI ​​service requests through the first logical functional plane and / or a second logical functional plane. The first logical functional plane is a data plane, and the second logical functional plane is an intelligence plane; alternatively, the first logical functional plane is an intelligence plane, and the second logical functional plane is a data plane. In the above method, the data plane and the intelligence plane can collaboratively process AI service requests to obtain AI data corresponding to the AI ​​service requests, realizing data collaboration and data services between the data plane and the intelligence plane, and improving the accuracy and efficiency of information interaction between the data plane and the intelligence plane.

[0069] It should be understood that the description in the foregoing summary section is not intended to limit the key or essential features of the embodiments of this disclosure, nor is it intended to restrict the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0070] To more clearly illustrate the technical solutions in this disclosure or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0071] Figure 1A is a schematic diagram of a network architecture provided in an embodiment of this disclosure;

[0072] Figure 1B is a schematic diagram of another network architecture provided in an embodiment of this disclosure;

[0073] Figure 1C is a schematic diagram of another network architecture provided in an embodiment of this disclosure;

[0074] Figure 2 is a flowchart of a data processing method provided in an embodiment of this disclosure;

[0075] Figure 3 is a schematic diagram of a process for acquiring AI data according to an embodiment of this disclosure;

[0076] Figure 4A is a schematic diagram of a process for acquiring AI data according to an embodiment of this disclosure;

[0077] Figure 4B is a schematic diagram of another process for acquiring AI data provided in an embodiment of this disclosure;

[0078] Figure 4C is a schematic diagram of another process for acquiring AI data provided in an embodiment of this disclosure;

[0079] Figure 4D is a schematic diagram of another process for acquiring AI data provided in an embodiment of this disclosure;

[0080] Figure 5A is a schematic diagram of a process for acquiring AI data according to an embodiment of this disclosure;

[0081] Figure 5B is a schematic diagram of another process for acquiring AI data provided in an embodiment of this disclosure;

[0082] Figure 6 is a schematic diagram of a process for acquiring AI data according to an embodiment of this disclosure;

[0083] Figure 7 is a schematic diagram of a method for acquiring AI data provided in an embodiment of this disclosure;

[0084] Figure 8 is a schematic diagram of a method for acquiring AI data provided in an embodiment of this disclosure;

[0085] Figure 9 is a schematic diagram of a method for sending AI scheduling strategies and training data according to an embodiment of this disclosure;

[0086] Figure 10A is a schematic diagram of the information interaction process between the intelligent plane and the data plane provided in an embodiment of this disclosure;

[0087] Figure 10B is a schematic diagram of the information interaction process between the intelligent plane and the data plane provided in an embodiment of this disclosure;

[0088] Figure 11 is a schematic diagram of information interaction between the intelligent plane and the data plane network functions provided in an embodiment of this disclosure;

[0089] Figure 12 is a schematic diagram of the structure of a data processing device provided in an embodiment of this disclosure;

[0090] Figure 13 is a schematic diagram of another data processing device provided in an embodiment of this disclosure;

[0091] Figure 14 is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0092] In this disclosure, the term "and / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0093] In this disclosure, the term "multiple" refers to two or more, and other quantifiers are similar.

[0094] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this disclosure.

[0095] This disclosure provides a data processing method, apparatus, device, and storage medium, wherein an electronic device can obtain AI service requirements through a first logical functional surface and obtain AI data corresponding to the AI ​​service requirements through the first logical functional surface and / or a second logical functional surface. In this way, the first logical functional surface and the second logical functional surface can perform AI data collaboration and AI service collaboration.

[0096] The method and apparatus are based on the same concept of the application. Since the methods and apparatus solve problems in similar ways, the implementation of the apparatus and methods can refer to each other, and the repeated parts will not be described again.

[0097] The technical solutions provided in this disclosure can be applied to a variety of systems. For example, applicable systems may include Long Term Evolution (LTE) systems, LTE Frequency Division Duplex (FDD) systems, LTE Time Division Duplex (TDD) systems, Long Term Evolution Advanced (LTE-A) systems, Universal Mobile Telecommunications System (UMTS), Worldwide Interoperability for Microwave Access (WiMAX) systems, 5G New Radio (NR) systems, and their evolved communication systems. These systems may include terminal equipment and network equipment. The systems may also include a core network component, such as Evolved Packet System (EPS) or 5G systems (5GS).

[0098] The terminal devices involved in the embodiments of this disclosure can be devices that provide voice and / or data connectivity to users, handheld devices with wireless connectivity, or other processing devices connected to a wireless modem. The names of the terminal devices may differ in different systems; for example, in a 5G system, a terminal device may be called User Equipment (UE). Wireless terminal devices can be USB storage devices, other personal computer memory devices, and dongles. They can also communicate with one or more core networks (CNs) via a Radio Access Network (RAN). Wireless terminal devices can be mobile terminal devices, such as mobile phones (or "cellular" phones) and computers with mobile terminal devices. For example, they can be portable, pocket-sized, handheld, computer-embedded, or vehicle-mounted mobile devices that exchange voice and / or data with the radio access network. Examples of such devices include Personal Communication Service (PCS) telephones, cordless phones, Session Initiated Protocol (SIP) phones, Wireless Local Loop (WLL) stations, Personal Digital Assistants (PDAs), personal computers, tablets, and Machine-type Communication (MTC) terminal devices. Wireless terminal devices can also be referred to as systems, subscriber units, subscriber stations, mobile stations, mobile devices, remote stations, access points, remote terminals, access terminals, user terminals, user agents, user devices, and wireless access devices and routers / modems that meet the limitations of this definition, but are not limited to these in the embodiments of this disclosure.

[0099] The network device involved in this disclosure can be a base station, which may include multiple cells providing services to terminals. Depending on the application, the base station may also be called an access point, or a device in the access network that communicates with wireless terminal devices through one or more sectors on the air interface, or other names. The network device can be used to exchange received air frames with Internet Protocol (IP) packets, acting as a router between the wireless terminal device and the rest of the access network, where the rest of the access network may include an Internet Protocol (IP) communication network. The network device can also coordinate the attribute management of the air interface. For example, the network device involved in this disclosure can be an evolved Node B (eNB or e-NodeB) in a long term evolution (LTE) system, a 5G base station (gNB) in a next generation system, or a Home evolved Node B (HeNB), relay node, femto, pico, network testing equipment, etc., and is not limited in this disclosure. In some network architectures, network devices may include centralized unit (CU) nodes and distributed unit (DU) nodes, which may also be geographically separated.

[0100] In related technologies, the network architecture of sixth-generation mobile communication (6G) technology can include a logical functional plane with AI services to improve the communication quality of 6G. For example, the logical functional plane with AI services can provide AI services to users. Currently, 6G technology only proposes that the network architecture can introduce a logical functional plane with AI services. However, the network architecture can also include other logical functional planes (such as the data plane). How the logical functional plane with AI services can interact with other logical functional planes is a technical problem that urgently needs to be solved.

[0101] To address the technical problems in related technologies, this disclosure provides a data processing method. The method involves acquiring AI service requirements through a first logical functional plane, determining search conditions corresponding to the AI ​​service requirements through the first and / or second logical functional planes, and acquiring AI data corresponding to the AI ​​service requirements from a resource pool based on the search conditions. The first logical functional plane is a data plane, and the second logical functional plane is an intelligence plane, or vice versa. This allows the first logical functional plane to accurately determine AI service requirements based on user business requests, and the first and / or second logical functional planes to acquire AI data corresponding to the AI ​​service requirements from the resource pool. This enables data and service collaboration between the first and second logical functional planes, improving the accuracy of information interaction between them.

[0102] The network architecture involved in the embodiments of this disclosure will now be described with reference to Figures 1A-1C.

[0103] Figure 1A is a schematic diagram of a network architecture provided by an embodiment of this disclosure. Referring to Figure 1A, the network architecture includes an intelligent plane and a data plane. The data plane and the intelligent plane can interact based on service-oriented interfaces, point-to-point interfaces, etc., and this disclosure does not limit this interaction. The data plane may include an AI data resource pool. The intelligent plane and the data plane can be logical functional planes.

[0104] Please refer to Figure 1A. The intelligent surface can provide AI services to users. For example, the intelligent surface can acquire the data needed to train AI models, provide AI models to users (UE, RAN, logical function plane, network function NF, and applications, etc.), process AI models (e.g., AI model distillation, pruning, compression, etc.), and manage data security elements such as cleaning and privacy removal of the acquired data. The intelligent surface supports AI models at the network element, network, and service levels.

[0105] Referring to Figure 1A, the data plane may include data service orchestration and control functions and data execution nodes, wherein the data execution nodes can provide data services to users. For example, data execution nodes can be used for data processing, data storage, data analysis, data retrieval, data acquisition, and other data services. The AI ​​data resource pool is used to store AI data. For example, the AI ​​data resource pool can store AI models, data used to train AI models, and the AI ​​model resource pool can also store input data of the AI ​​model when it is used, etc. This disclosure does not limit this aspect.

[0106] Figure 1B is a schematic diagram of another network architecture provided by an embodiment of this disclosure. Please refer to Figure 1B, which includes a network architecture. The network architecture may include an intelligent plane and a data plane, and the intelligent plane and data plane can interact with each other. In the embodiment shown in Figure 1B, the AI ​​data resource pool may be located in the intelligent plane; that is, the intelligent plane can store or retrieve AI data in the AI ​​data resource pool.

[0107] Figure 1C is a schematic diagram of another network architecture provided by an embodiment of this disclosure. Referring to Figure 1C, the network architecture includes an intelligent plane, a data plane, and an AI data resource pool. The intelligent plane and the data plane can interact with the AI ​​data resource pool. In the embodiment shown in Figure 1C, the AI ​​data resource pool can be a shared resource pool for the intelligent plane and the data plane, and both the intelligent plane and the data plane can store or retrieve AI data in the AI ​​data resource pool.

[0108] The communication method provided in this disclosure will now be described in detail with reference to the embodiments.

[0109] Figure 2 is a flowchart of a communication method provided in an embodiment of this disclosure. Referring to Figure 2, the method includes:

[0110] S201. Obtain AI service requirements through the first logical functional plane.

[0111] AI service requirements may include at least one of the following:

[0112] Identification of the target AI model;

[0113] Information about the target AI model;

[0114] Identification of data related to the AI ​​model used;

[0115] The identifier of the training data for the AI ​​model to be trained.

[0116] In some embodiments, before obtaining AI service requests through the first logical functional plane, the electronic device can obtain AI service requests through the first logical functional plane. For example, the electronic device can receive AI service requests sent by a user through the first logical functional plane. These AI service requests may be related to the AI ​​service requested by the user (e.g., the user requests to execute services related to an AI model). The user may include a terminal device (UE), a base station (RAN), a logical functional plane, a network function (NF), and an application, etc., which are not limited in this embodiment. For example, the AI ​​service request sent by the user may be used to request prediction of cell communication quality in future periods, or to request prediction of communication failures in future periods, etc., which are not limited in this embodiment.

[0117] In some embodiments, the electronic device can process AI business requests through a first logical functional plane to obtain AI service requirements. For example, the electronic device can translate and parse AI business requests through the first logical functional plane to obtain AI service requirements.

[0118] In some embodiments, the target AI model can be an AI model that matches the AI ​​service request. For example, if a user sends an AI service request to the first logical functional plane to request a prediction of communication quality in a future period, the target AI model can be an AI model used to predict communication quality. For example, AI model 1 is used to predict communication quality, and AI model 2 is used to predict communication failures. If a user sends an AI service request to the first logical functional plane to request a prediction of communication quality in a future period, the electronic device can translate and parse the AI ​​service request to obtain an AI service requirement. The AI ​​service requirement may include an identifier of AI model 1 (the target AI model) (which may be a unique identifier pre-set by the electronic device, or an identifier determined by the electronic device according to any feasible implementation method; this embodiment does not limit this).

[0119] In some embodiments, the information of the target AI model may include any information related to the target AI model, such as the type of the target AI model (e.g., prediction and decision type, classification and recognition type, etc.), the size of the parameter set of the target AI model (e.g., the storage space occupied by the AI ​​model), and the storage location of the target AI model. This disclosure does not limit this information. Optionally, the electronic device may acquire the information of the target AI model according to any feasible implementation method, and this disclosure does not limit this method.

[0120] In some embodiments, the data related to the AI ​​model used may include input data and output data corresponding to the AI ​​model. For example, the AI ​​model used by the user can be used to predict communication quality in future time periods. Therefore, the AI ​​service request sent by the user to the first logical functional plane can be used to request the data (e.g., measured communication data) required to predict the communication quality of the cell. The first logical functional plane can translate and parse the AI ​​service request to obtain an AI service requirement. This AI service requirement may include the identifier of the input data of the AI ​​model used by the user (e.g., the cell identifier corresponding to the input data, the time period identifier corresponding to the input data, and the identifier of the type of input data). For example, if the AI ​​service request is used to request the data required to predict the communication quality of the cell in the next 10 minutes, the AI ​​service requirement may include the cell identifier, the identifier of the measurement data (data used to determine communication quality) in the last 10 minutes, etc.

[0121] In some embodiments, training data can be used to generate an AI model to be trained. For example, if a user determines that the AI ​​model to be trained can be an AI model for predicting the communication quality of a cell in a future period, the user can send an AI service request to a first logical functional plane. This AI service request can be used to request the training data required to train the AI ​​model. The first logical functional plane can translate and parse the AI ​​service request to obtain an AI service requirement, which may include the identifier of the training data. For example, for an AI model that predicts communication faults, the training data may include actual measurement data (data used to determine communication faults) and the communication faults corresponding to the measurement data.

[0122] In some embodiments, the network architecture may include a first logical functional plane and a second logical functional plane. The types of the first and second logical functional planes can be as follows:

[0123] Case 1: The first logical functional plane can be the data plane, and the second logical functional plane can be the intelligence plane.

[0124] In scenario 1, the first logical functional plane can be the data plane. The data plane can receive AI service requests sent by users and process these requests to obtain AI service requirements. For example, a user can send an AI service request to the data plane, which can be used to request a prediction of the cell's communication quality in a future time period. The data plane can translate and parse this AI service request to obtain AI service requirements, which may include the identifier of the AI ​​model for predicting communication quality and / or information about the AI ​​model for predicting communication quality.

[0125] Scenario 2: The first logical functional surface can be the intelligence surface, and the second logical functional surface can be the data surface.

[0126] In scenario 2, the first logical functional surface can be an intelligent surface. The intelligent surface can receive AI service requests sent by the user and process these requests to obtain AI service requirements. For example, a user can send an AI service request to the intelligent surface, which can be used to request data needed to predict communication quality. The intelligent surface can translate and parse this AI service request to obtain AI service requirements, which may include identifiers of the input data for the AI ​​model used to predict communication quality.

[0127] S202. Obtain AI data corresponding to AI service requirements through the first logical functional plane and / or the second logical functional plane.

[0128] AI data may include at least one of the following:

[0129] Target AI model;

[0130] Data related to the AI ​​model used;

[0131] Training data for the AI ​​model to be trained.

[0132] In some embodiments, AI data may be matched with AI service requirements. For example, if the AI ​​service requirement includes an identifier of an AI model, the AI ​​data obtained by the electronic device through the first logical functional surface and / or the second logical functional surface may include the AI ​​model; if the AI ​​service requirement includes an identifier of the training data of the AI ​​model to be trained, the AI ​​data obtained by the electronic device through the first logical functional surface and / or the second logical functional surface may include the training data of the AI ​​model.

[0133] In some embodiments, the electronic device obtains AI data corresponding to the AI ​​service request through a first logical functional plane and / or a second logical functional plane, in the following three cases:

[0134] The first scenario for obtaining AI data corresponding to AI service needs: The first logical functional plane includes a resource pool.

[0135] In some embodiments, the resource pool can be an AI data resource pool, and the electronic device can obtain the AI ​​data corresponding to the AI ​​service requirement according to the following feasible implementation: obtain the AI ​​data corresponding to the AI ​​service requirement from the resource pool of the first logical functional plane, wherein the AI ​​data can be AI data collected by the second logical functional plane and stored in the resource pool of the first logical functional plane.

[0136] It should be noted that the resource pool may include the first or second logical functional plane, and multiple AI data (such as multiple AI models).

[0137] In some embodiments, the second logical functional plane can collect AI data and perform de-identification processing on the collected AI data. The second logical functional plane can store the de-identified AI data in the resource pool of the first logical functional plane.

[0138] The following section uses AI data as an example to explain the process of collecting data from the second logical functional plane.

[0139] The AI ​​model can complete its initial registration in the second logical functional plane, enabling the second logical functional plane to perceive basic information about the AI ​​model's resource data. During AI model registration, the second logical functional plane can identify and verify the AI ​​model. If the AI ​​model passes verification, the second logical functional plane can assign a unique identifier to the AI ​​model. Furthermore, after the AI ​​model completes registration, it can periodically send its status data to the second logical functional plane.

[0140] It should be noted that the method for collecting AI data in the first logical functional plane is the same as the method for collecting AI data in the second logical functional plane, and will not be described again in this embodiment.

[0141] In this scenario, the resource pool is located on the first logical functional plane. After the second logical functional plane collects AI data, the electronic device can store the AI ​​data in the resource pool of the first logical functional plane through the second logical functional plane, thus enabling information interaction between the first and second logical functional planes. For example, the first logical functional plane can be a data plane, and the second logical functional plane can be an intelligence plane. The AI ​​model can register on the intelligence plane. After the intelligence plane obtains the AI ​​model data related to the AI ​​model, it can call the data storage service on the data plane to store the AI ​​model data, thereby forming an AI data resource pool on the data plane (or it can generate a resource pool on the data plane and store the AI ​​model data in the resource pool; this embodiment does not limit this).

[0142] In some embodiments, an electronic device may obtain AI data corresponding to an AI service request by: determining the search criteria corresponding to the AI ​​service request, and obtaining AI data from a resource pool based on the search criteria.

[0143] The search criteria may include AI service requirements. For example, search criteria may include the identifier of the AI ​​model, the type of the AI ​​model, the size of the parameter set of the AI ​​model, the location of the AI ​​model, etc., but this disclosure does not limit these criteria.

[0144] In some embodiments, electronic devices can retrieve AI data from a resource pool based on search criteria. For example, if the resource pool is located on a data surface, the data surface can retrieve AI data that meets the search criteria from the resource pool based on a data retrieval service. This allows the electronic device to improve the accuracy of the AI ​​data.

[0145] Below, referring to Figure 3, we will explain the process of acquiring AI data in the first case, taking the first logical functional surface as the intelligent surface and the second logical functional surface as the data surface.

[0146] Figure 3 is a schematic diagram illustrating a process for acquiring AI data according to an embodiment of this disclosure. Referring to Figure 3, it includes: an AI model, a data plane, and an intelligence plane. The data plane can serve as the data anchor point for the AI ​​model; that is, the AI ​​model can complete its registration within the data plane. The data plane can possess data acquisition and data processing functions. The data acquisition function can acquire AI model data, and the data processing function can perform anonymization processing on the AI ​​model data and store the AI ​​model data in the resource pool of the intelligence plane.

[0147] Referring to Figure 3, after the data plane stores the anonymized AI model data in the intelligent plane's resource pool, the intelligent plane can manage and maintain this resource pool. When the intelligent plane receives a business request from a user (matching an AI model registered with the data plane), it can process the request to obtain the AI ​​service requirement. Based on this requirement, the intelligent plane can retrieve the AI ​​model stored in the data plane from the resource pool and send it to the user. This allows for information exchange between the data plane and the intelligent plane during the AI ​​model registration process, improving the accuracy of information interaction.

[0148] In some embodiments, the second case of obtaining AI data corresponding to AI service requirements is as follows: the second logical functional surface includes a resource pool.

[0149] Electronic devices can obtain AI data corresponding to AI service requirements in the following feasible ways: obtain AI data corresponding to AI service requirements from the resource pool of the second logical functional plane, wherein the AI ​​data can be AI data collected by the second logical functional plane and stored in the resource pool of the second logical functional plane, or the AI ​​data can be data collected by the first logical functional plane and stored in the resource pool of the second logical functional plane.

[0150] In this scenario, the resource pool is located on the second logical functional plane. In one scenario, the second logical functional plane can collect AI data and store it in its resource pool. When the first logical functional plane requests an AI service, the electronic device can retrieve the corresponding AI data from the resource pool of the second logical functional plane through the first logical functional plane, and then send the AI ​​data to the user through the first logical functional plane, thereby achieving information interaction between the first and second logical functional planes. In another scenario, the first logical functional plane can collect AI data and store it in the resource pool of the second logical functional plane. When the first logical functional plane requests an AI service, the electronic device can retrieve the corresponding AI data from the resource pool of the second logical functional plane through the first logical functional plane, and then send the AI ​​data to the user through the first logical functional plane. In this way, the first and second logical functional planes can accurately interact in both the acquisition and use of AI data.

[0151] Below, taking AI data as an example of an AI model, and referring to Figures 4A-4D, we will explain the process of obtaining AI data corresponding to AI service requirements in the second scenario.

[0152] Figure 4A is a schematic diagram illustrating a process for acquiring AI data according to an embodiment of this disclosure. Referring to Figure 4A, it includes an AI model, an intelligent surface, and a data surface. The AI ​​model can register on the intelligent surface, which can then send the registered AI model to the data surface. The data surface can store the AI ​​model in its resource pool based on its data storage function.

[0153] Referring to Figure 4A, when a user sends a business request to the intelligent surface, the intelligent surface can process the request, obtain the AI ​​service requirements, and query the data surface for AI models based on these requirements. The data surface, using its data proxy function and the AI ​​service requirements, can retrieve an AI model from its resource pool that meets the requirements and send that AI model to the intelligent surface. The intelligent surface can then send the AI ​​model to the user. In this way, the intelligent surface and the data surface can exchange information accurately during both the AI ​​model registration and AI model invocation processes.

[0154] Figure 4B is a schematic diagram illustrating another process for acquiring AI data according to an embodiment of this disclosure. Referring to Figure 4B, it includes an AI model, an intelligence plane, and a data plane. The data plane includes data retrieval, data processing, and data acquisition functions. The AI ​​model can complete registration on the data plane. The data acquisition function of the data plane can collect data from the AI ​​model and send the AI ​​model's data to the data processing function. The data processing function can perform anonymization processing on the AI ​​model's data and store the AI ​​model in a resource pool.

[0155] Referring to Figure 4B, when a user sends a business request to the intelligent surface, the intelligent surface can process the request, obtain the AI ​​service requirement, and query the data surface for an AI model based on the requirement. The data surface's data retrieval function can then retrieve a matching AI model from the resource pool and send it to the intelligent surface. The intelligent surface can then send the AI ​​model to the user. In this way, the data surface and the intelligent surface can accurately exchange information during the user's acquisition of the AI ​​model.

[0156] Figure 4C is a schematic diagram illustrating another process for acquiring AI data according to an embodiment of this disclosure. Referring to Figure 4C, it includes an AI model, a data plane, and an intelligence plane. The data plane includes data acquisition and data processing functions, and the intelligence plane includes a resource pool. The AI ​​model can register on the data plane. The data acquisition function of the data plane can collect data from the AI ​​model and send the data to the data processing function. The data processing function can de-identify the data of the AI ​​model and store the data in the resource pool of the intelligence plane.

[0157] Referring to Figure 4C, when a user sends a business request to the data plane, the data plane can process the request, obtain the AI ​​service requirements, and query the AI ​​model in the intelligent plane's resource pool based on these requirements. The intelligent plane can then respond with an AI model that meets the service requirements. After obtaining the AI ​​model, the data plane can send it to the user. In this way, the data plane and the intelligent plane can accurately exchange information during the AI ​​model registration and invocation processes.

[0158] Figure 4D is a schematic diagram illustrating another process for acquiring AI data provided in an embodiment of this disclosure. Referring to Figure 4D, it includes an AI model, an intelligent surface, and a data surface. The AI ​​model can be registered on the intelligent surface, which includes a resource pool. After registration, the resource pool stores the AI ​​model. When a user sends a business request to the data surface, the data surface processes the request to obtain AI service requirements. Based on its data analysis functions, the data surface retrieves an AI model matching the AI ​​service requirements from the intelligent surface's resource pool and sends the AI ​​model to the user. In this way, the data surface and the intelligent surface can exchange information accurately during the AI ​​model invocation process.

[0159] It should be noted that the method for electronic devices to obtain AI data corresponding to AI service requests is the same as the method for obtaining AI data corresponding to AI service requests in the first case, and will not be described again in this embodiment.

[0160] In some embodiments, a third scenario for obtaining AI data corresponding to AI service requests involves the first and second logical functional planes including a shared resource pool.

[0161] Electronic devices can acquire AI data corresponding to AI service requirements in the following feasible way: acquire AI data corresponding to AI service requirements from a shared resource pool, wherein the AI ​​data can be AI data collected and stored in the shared resource pool by the second logical functional plane.

[0162] In this scenario, the resource pool can be a shared resource pool for the first and second logical functional surfaces. Therefore, electronic devices can obtain AI data through the second logical functional surface and store the AI ​​data in the shared resource pool. Electronic devices can obtain AI service requirements through the first logical functional surface and, based on these requirements, obtain AI data from the shared resource pool that meets those requirements. In this way, the first logical functional surface can obtain AI data from the shared resource pool, and the second logical functional surface can store the AI ​​data in the shared resource pool. This allows for accurate information exchange between the first and second logical functional surfaces.

[0163] Below, taking AI data as an example of an AI model, and referring to Figures 5A and 5B, we will explain the process of obtaining AI data corresponding to AI service requirements in the third case.

[0164] Figure 5A is a schematic diagram illustrating a process for acquiring AI data according to an embodiment of this disclosure. Referring to Figure 5A, it includes: an AI model, an intelligent plane, a data plane, and a shared resource pool. The AI ​​model can be registered with the intelligent plane. After registration, the intelligent plane can store the AI ​​model in the shared storage resource pool. Users can send business requests to the data plane. The data plane processes the business requests to obtain AI service requirements, and based on these requirements, retrieves a matching AI model from the shared resource pool and sends the AI ​​model to the user. In this way, the intelligent plane can store AI data in the shared resource pool, and the data plane can retrieve the AI ​​data stored by the intelligent plane from the shared resource pool, enabling effective information exchange between the intelligent plane and the data plane.

[0165] In the embodiment shown in Figure 5A, the intelligent surface can serve as a data anchor point for the AI ​​model, and the intelligent surface can perform data maintenance processes such as writing, updating, and deleting in a shared resource pool.

[0166] Figure 5B is a schematic diagram illustrating another process for acquiring AI data provided in an embodiment of this disclosure. Referring to Figure 5B, it includes: an AI model, an intelligent plane, a data plane, and a shared resource pool. The AI ​​model can be registered on the data plane. After registration, the data plane can store the AI ​​model in the shared storage resource pool. Users can send business requests to the intelligent plane. The intelligent plane processes the business requests, obtains the AI ​​service requirements, and, based on these requirements, retrieves a matching AI model from the shared resource pool and sends the AI ​​model to the user. In this way, the data plane can store AI data in the shared resource pool, and the intelligent plane can retrieve the AI ​​data stored by the data plane from the shared resource pool, enabling effective information exchange between the intelligent plane and the data plane.

[0167] In the embodiment shown in Figure 5B, the data plane can serve as the data anchor point for the AI ​​model, and the data plane can be written, updated, deleted, and processed in a shared resource pool.

[0168] In some embodiments, in the third case, the above data processing method further includes: sending AI data subscription information to a shared resource pool through a first logical functional plane; sending subscription information to a second logical functional plane through the shared resource pool; updating the AI ​​data in the shared resource pool through the second logical functional plane and the subscription information; and sending the updated AI data to the first logical functional plane through the shared resource pool.

[0169] The subscription information is used to subscribe to corresponding AI data. For example, if a user subscribes to an AI model in a shared resource pool through Smart Face, Smart Face can periodically provide the user with data from that AI model.

[0170] In some embodiments, subscription information may include at least one of the following:

[0171] Identification of subscribed AI data;

[0172] Data from AI models;

[0173] AI data feedback cycle.

[0174] In some embodiments, the data reporting cycle can be the cycle at which the first logical functional surface feeds back AI data to the user. For example, if a user subscribes to an AI model through the first logical functional surface, and the AI ​​data feedback cycle is 1 day, then the second logical functional surface retrieves the AI ​​model's data every 1 day and updates the AI ​​model in a shared resource pool. The first logical functional surface can then retrieve the AI ​​model from the shared resource pool and send it to the user.

[0175] In some embodiments, when a user has a need for AI data, the user can query AI data based on a first logical functional surface or a second logical functional surface. Furthermore, the user can subscribe to relevant AI data based on their needs. In this way, the first logical functional surface or the second logical functional surface can periodically provide AI data feedback to the user, thereby improving the accuracy of the AI ​​data.

[0176] The process of subscribing to AI data will be explained below with reference to Figure 6.

[0177] Figure 6 is a schematic diagram illustrating a process for acquiring AI data according to an embodiment of this disclosure. Referring to Figure 6, it includes: an AI model, an intelligent plane, a data plane, and a shared resource pool. The AI ​​model can be registered with the intelligent plane. After registration, the intelligent plane can store the AI ​​model in the shared storage resource pool. Users can send a service request to the data plane to subscribe to the AI ​​model stored in the intelligent plane. The data plane can determine the subscription information based on the service request and send the subscription information to the shared resource pool. The shared resource pool can forward the subscription information to the intelligent plane.

[0178] Referring to Figure 6, the intelligent surface can periodically perceive the AI ​​model based on the subscription information and store the perceived AI model in a shared resource pool. The shared resource pool can send the AI ​​model to the data surface, and the data surface can forward the AI ​​model to the user. In this way, after the user subscribes to the AI ​​model, the intelligent surface can periodically obtain the data of the AI ​​model (e.g., model updates), thereby improving the accuracy of the AI ​​model obtained by the user. Furthermore, the intelligent surface can perceive the AI ​​model data, and the data surface can forward the AI ​​model data, thus realizing information interaction between the intelligent surface and the data surface.

[0179] This disclosure provides a data processing method that obtains AI service requirements through a first logical functional surface, determines search conditions corresponding to the AI ​​service requirements through the first and / or second logical functional surfaces, and obtains AI data corresponding to the AI ​​service requirements from a resource pool based on the search conditions. The first logical functional surface is a data surface, and the second logical functional surface is an intelligence surface, or vice versa. In this way, the first logical functional surface can accurately determine the AI ​​service requirements, and the first and / or second logical functional surfaces can obtain the AI ​​data corresponding to the AI ​​service requirements from the resource pool, achieving data and service collaboration between the first and second logical functional surfaces and improving the accuracy of information interaction between them.

[0180] Based on the embodiment shown in Figure 2, the following describes, with reference to Figure 7, the method by which an electronic device obtains AI data corresponding to AI service requirements when the above-mentioned AI data is the training data of the AI ​​model to be trained.

[0181] Figure 7 is a schematic diagram of a method for acquiring AI data according to an embodiment of this disclosure. In the embodiment shown in Figure 7, the first logical functional plane is the intelligence plane, the second logical functional plane is the data plane, the data plane includes a resource pool, and the AI ​​data is training data. Referring to Figure 7, the method flow includes:

[0182] S701: Send an AI service request to the first module in the intelligent face through the second module in the intelligent face.

[0183] The intelligent surface may include a first module and a second module.

[0184] The first module is used to acquire and process AI data, while the second module is used to obtain the processing results corresponding to the AI ​​service requests sent by the target user. The data plane can include an acquisition module, a storage module, and a retrieval module. The acquisition module is used to collect AI data, the storage module is used to store AI data, and the retrieval module is used to query AI data that matches the AI ​​service request from the stored AI data.

[0185] In some embodiments, an electronic device can send an AI service request to a first module in the smart surface via a second module in the smart surface. For example, when a user sends an AI service request to the smart surface (e.g., the user invokes the function of the second module of the smart surface to make an AI service request, which may indicate requirements such as AI data and training), the second module in the smart surface can parse the service request, obtain the AI ​​service request, and send the AI ​​service request to the first module in the smart surface.

[0186] S702. Through the first module in the intelligent plane, obtain the training data corresponding to the AI ​​service requirements and at least one AI node from the resource pool of the data plane.

[0187] In some embodiments, at least one AI node can be a candidate node for training an AI model. For example, an AI node can be a UE, CN, or other nodes, and multiple AI nodes can be used for training the AI ​​model.

[0188] The training data can be AI data corresponding to AI service requests. For example, a user's business request to the intelligent surface can be used to request training of an AI model for predicting communication quality. The second module in the intelligent surface can determine the AI ​​service request corresponding to the business request, including the identifier of the training data. Based on the identifier of the training data, the second module of the intelligent surface can obtain the training data of the AI ​​model for predicting communication quality from the resource pool of the data surface. Furthermore, since the user needs to train the AI ​​model, the intelligent surface can also provide the user with at least one AI node. The user can train the AI ​​model for predicting communication quality based on at least one AI node and training data (the user can determine the node for training the AI ​​model among at least one AI node).

[0189] This disclosure provides a method for acquiring AI data. A second module in the intelligent plane sends an AI service request to a first module in the intelligent plane. The first module in the intelligent plane then retrieves the corresponding training data and at least one AI node from the resource pool of the data plane. In this way, the second module in the intelligent plane can determine the AI ​​service request, and the first module in the intelligent plane can retrieve the corresponding training data from the resource pool of the data plane based on the AI ​​service request, thereby achieving information interaction between the intelligent plane and the data plane.

[0190] Based on the embodiment shown in Figure 2, the following describes, with reference to Figure 8, the method by which an electronic device obtains AI data corresponding to AI service requirements when the above-mentioned AI data is the training data of the AI ​​model to be trained.

[0191] Figure 8 is a schematic diagram of a method for acquiring AI data according to an embodiment of this disclosure. In the embodiment shown in Figure 8, the first logical functional plane is the data plane, the second logical functional plane is the intelligence plane, the intelligence plane includes a resource pool, and the AI ​​data is the training data. Referring to Figure 8, the method flow includes:

[0192] S801. The AI ​​service request is sent from the second module in the intelligent surface to the first module in the intelligent surface.

[0193] Specifically, the AI ​​service requirements can be those obtained by the second module through the data plane. For example, if the first logical functional plane is the data plane, then the data plane can receive AI business requests sent by users, process these requests to obtain AI service requirements, and send these requirements to the second module of the intelligent plane. It should be noted that the data plane can also forward user-sent AI business requests to the second module of the intelligent plane, which can also process these requests to obtain AI service requirements.

[0194] In some embodiments, after the second module of the intelligent surface obtains the AI ​​service request, it can send the AI ​​service request to the first module.

[0195] S802. Through the first module in the intelligent surface, obtain training data and at least one AI node from the resource pool of the intelligent surface.

[0196] In some embodiments, the training data can be AI data corresponding to an AI service request. For example, an AI service request sent by a user to the data plane can be used to request training of an AI model for predicting communication quality. The data plane determines that the AI ​​service request may include an identifier for the training data. After receiving the AI ​​service request sent by the data plane, the second module of the intelligent plane can send the AI ​​service request to the first module. Since the resource pool is in the intelligent plane, the first module can obtain the training data for the AI ​​model for predicting communication quality from the resource pool of the data plane according to the identifier of the training data. Furthermore, since the user needs to train the AI ​​model, the intelligent plane can also obtain at least one AI node from the resource pool and provide at least one AI node to the user. The user can then train the AI ​​model for predicting communication quality based on at least one AI node and the training data.

[0197] This disclosure provides a method for acquiring AI data. An electronic device can obtain AI service requests through a data plane. After receiving the AI ​​service request, a second module in the intelligent plane can send it to a first module in the intelligent plane. The first module in the intelligent plane then retrieves the training data corresponding to the AI ​​service request and at least one AI node from its resource pool. In this way, the data plane can determine the AI ​​service request, and the intelligent plane can retrieve the corresponding training data from its resource pool based on that request, thereby enabling information interaction between the intelligent plane and the data plane.

[0198] Based on the embodiments shown in Figures 7 and 8, after the electronic device acquires training data and at least one AI node, the above data processing method further includes a method for sending AI scheduling strategies and training data to the target user. The method for sending AI scheduling strategies and training data to the target user will be described below with reference to Figure 9.

[0199] Figure 9 is a schematic diagram of a method for sending AI scheduling strategies and training data according to an embodiment of this disclosure. Referring to Figure 9, the method includes:

[0200] S901. The first module processes at least one AI node to obtain an AI scheduling strategy.

[0201] In some embodiments, the AI ​​scheduling strategy includes a target AI node and its path information. The target node can be a node used to train an AI model. For example, an electronic device, through a first module, determines AI node 1, AI node 2, and AI node 3 in a resource pool. The first module can select a target AI node from these three AI nodes as the node for training the AI ​​model. When selecting an AI node, the first module can determine the target AI node based on information about each AI node (e.g., computing power, data processing capabilities, etc.).

[0202] The path information can be the path a user takes to access a target AI node. For example, the resource pool can pre-store multiple AI nodes and the path information for each AI node. After the electronic device determines the target AI node, it can obtain the path information for that target AI node.

[0203] S902, The first module sends the AI ​​scheduling strategy and training data to the second module.

[0204] S903. Through the second module, send the AI ​​scheduling strategy and training data to the target user.

[0205] In some embodiments, the target user can be the user who sends the AI ​​service request. For example, the target user can send an AI service request to the first logical functional plane. This AI service request is used to request the training of an AI model. After the first logical functional plane obtains the training data for training the AI ​​model and the target AI node, it can send the training data and AI scheduling strategy (target AI node and path information of the target AI node) to the target user. The target user can determine the path to access the target AI node through the AI ​​scheduling strategy and train the AI ​​model using the target AI node and the training data.

[0206] S904. Through the first module, pre-allocate AI resources and deploy AI services in the target AI node.

[0207] In some embodiments, after the intelligent surface determines the target AI node, the electronic device can, through the first module of the intelligent surface, pre-allocate AI resources and deploy AI services on the target node, thereby facilitating the target user to train the AI ​​model on the target AI node. For example, if the target AI node is a CN, the intelligent surface can pre-allocate the resources required for training the AI ​​model and deploy AI services on that CN. It should be noted that the intelligent surface can pre-allocate AI resources and deploy AI services on the target node based on any feasible implementation method such as signaling or messaging, and this disclosure does not limit this.

[0208] The following section, in conjunction with Figures 10A and 10B, explains the information interaction process between the intelligent plane and the data plane when the user requirement is to train an AI model.

[0209] Figure 10A is a schematic diagram of the information interaction process between the intelligent plane and the data plane according to an embodiment of this disclosure. Referring to Figure 10A, it includes a user, an intelligent plane, and a data plane. The intelligent plane includes intelligent services (a second module) and AI service orchestration (a first module), while the data plane includes a resource pool and data retrieval functions. When a user sends an AI service request (to request training data for training an AI model) to the intelligent service of the intelligent plane, the intelligent service can process the AI ​​service request, obtain the AI ​​service requirement, and send the AI ​​service requirement to the AI ​​service orchestration.

[0210] Please refer to Figure 10A. AI service orchestration can retrieve the training data and AI nodes corresponding to the AI ​​service requirement from the resource pool of the data plane through the data retrieval function of the AI ​​service requirement, and send the training data and AI nodes to the AI ​​service orchestration. The AI ​​service orchestration can identify the target AI node among the AI ​​nodes, and based on the target AI node's identifier, path information, and training data, obtain a feedback strategy, and send the feedback strategy to the intelligent service.

[0211] Referring to Figure 10A, since AI service orchestration can identify target nodes, it can pre-allocate resources and deploy services within those nodes. Furthermore, intelligent services can send feedback strategies to users. Users can access the target AI node based on the path information in this feedback strategy and train the AI ​​model using training data within that node. Thus, when a user requests to train an AI model, the intelligent plane and data plane can exchange information to provide the user with training data and the node for training the AI ​​model. Moreover, through service calls and collaboration between the intelligent plane and data plane, the development of repetitive functions can be reduced, thereby simplifying the system design and implementation of the logical functionalities and lowering system development costs.

[0212] Figure 10B is a schematic diagram of the information interaction process between the intelligent plane and the data plane according to an embodiment of this disclosure. Referring to Figure 10B, it includes a user, an intelligent plane, and a data plane. The intelligent plane includes intelligent services (a second module), AI service orchestration (a first module), and a resource pool. When a user sends an AI service request (to request training data for training an AI model) to the data plane, the data plane can process the AI ​​service request, obtain the AI ​​service requirement, and send the AI ​​service requirement to the intelligent services in the intelligent plane. The intelligent services can then send the AI ​​service requirement to the AI ​​service orchestration.

[0213] Please refer to Figure 10B. AI service orchestration can obtain the corresponding training data and AI nodes from the resource pool of the intelligent surface based on the AI ​​service request. AI service orchestration can identify the target AI node among the AI ​​nodes, and based on the target AI node's identifier, path information, and training data, obtain a feedback strategy and send the feedback strategy to the intelligent service.

[0214] Referring to Figure 10B, since AI service orchestration can determine target nodes, it can pre-allocate resources and deploy services within those nodes. Furthermore, intelligent services can send feedback strategies to users. Users can access the target AI node based on the path information in this feedback strategy and train the AI ​​model using training data within that node. Thus, when a user requests to train an AI model, the intelligent plane and data plane can provide the user with training data and the node for training the AI ​​model through information exchange. Moreover, through service calls and collaboration between the intelligent plane and data plane, the development of repetitive functions can be reduced, thereby simplifying the system design and implementation of the logical functionalities and lowering system development costs.

[0215] It should be noted that in the embodiment shown in Figure 10B, when other users outside the data plane call AI services, since the resource pool is managed by the intelligent plane, the intelligent plane can perform AI data query and AI service orchestration internally, and the intelligent plane does not need to coordinate with the data plane.

[0216] This disclosure provides a method for sending AI scheduling strategies and training data. A first module processes at least one AI node to obtain an AI scheduling strategy. The first module then sends the AI ​​scheduling strategy and training data to a second module, which in turn sends the AI ​​scheduling strategy and training data to a target user. The first module pre-allocates AI resources and deploys AI services on the target AI node. Thus, when the target user needs to train an AI model, the electronic device accurately provides the target user with training data and training nodes for the AI ​​model through information interaction between the intelligent plane and the data plane. Furthermore, the service calls and collaboration between the intelligent plane and the data plane reduce the development of repetitive functions, thereby simplifying the system design and implementation of the logical functional plane and reducing system development costs.

[0217] Based on any of the above embodiments, the intelligent plane includes an AI control function AICF, and the data plane includes a data orchestrator DO, at least one data agent DA, and a data storage function DRF. The following, with reference to Figure 11, will describe in detail the information interaction process between the network functions related to the intelligent plane and the network functions related to the data plane.

[0218] Figure 11 is a schematic diagram of information interaction between network functions of the intelligent plane and the data plane provided in an embodiment of this disclosure. Referring to Figure 11, it includes AICF, DO, DA (UE), DA (RAN), DA (CN), DA (AF), and DRF. Each domain DA can register in the DO and send its data service capabilities (AICF can register and transmit data through a secure and trusted anchor point). Each domain DA can register in the DRF and send relevant DA data to the DRF in real time.

[0219] Please refer to Figure 11. The AICF can send network data-related service requests to the DO. The DO can translate the service requests (mapping them to data service requirements), assign task identifiers to the data service, select target DAs based on the data service capabilities sent by the DAs, and allocate data service functions to each target DA. The DO can orchestrate target DAs to form a logical network (the DO can send relevant data service function settings to each target DA, thus forming an overlay logical network; furthermore, if there are updates to data security and information protection technologies, as well as analysis tools, the DO will also send relevant updates to the target DAs).

[0220] Please refer to Figure 11. The DO (Domain Analyzer) retrieves data from the target DA (Data Analyzer) in the DRF (Data Request Framework) based on the selected target DA. The DO can send the address of the target DA that can directly interact with the AICF (AI Analyzer Cloud). Inter-domain target DAs can perform a series of actions such as data collection, preprocessing, storage, and analysis according to their assigned functions, forming a data service flow. Furthermore, inter-domain target DAs can be interconnected to form a cross-domain data collaboration architecture. If user data processing is involved during the process, the processing behavior needs to be recorded. The target DA can send AI data (e.g., requested data, analysis results, etc.) to the AICF.

[0221] In the embodiment shown in Figure 11, the DO can be used to receive AI service requests sent by the AICF and to acquire the data service capabilities of at least one DA. The DO can also be used to generate a logical network based on the AI ​​service requests and to acquire the data corresponding to the target DA in the DRF. The logical network can include the target DA, whose data service capabilities match the AI ​​service requests. The DA can be used to send its corresponding data service capabilities to the DO and to the DRF. The DA can also be used to execute its corresponding data service capabilities. In this way, the data plane and the intelligent plane can achieve information interaction between network functions.

[0222] Figure 12 is a schematic diagram of a data processing device provided in an embodiment of this disclosure. Referring to Figure 12, the data processing device 120 includes a first acquisition module 121 and a second acquisition module 122, wherein:

[0223] The first acquisition module 121 is used to acquire AI service requirements through the first logical functional plane;

[0224] The second acquisition module 122 is used to acquire AI data corresponding to AI service requirements through the first logical functional plane and / or the second logical functional plane;

[0225] Wherein, the first logical functional surface is the data surface, and the second logical functional surface is the intelligence surface; or, the first logical functional surface is the intelligence surface, and the second logical functional surface is the data surface.

[0226] In some embodiments, the AI ​​service requirements include at least one of the following:

[0227] Identification of the target AI model;

[0228] Information about the target AI model;

[0229] Identification of data related to the AI ​​model used;

[0230] The identifier of the training data for the AI ​​model to be trained.

[0231] In some embodiments, the AI ​​data includes at least one of the following:

[0232] Target AI model;

[0233] Data related to the AI ​​model used;

[0234] Training data for the AI ​​model to be trained.

[0235] In some embodiments, the information of the target AI model includes the type of the target AI model, the size of the parameter set of the target AI model, and the storage location of the target AI model;

[0236] In some embodiments, the data related to the AI ​​model used includes the input data and output data corresponding to the AI ​​model used.

[0237] In some embodiments, the second acquisition module 122 is used to:

[0238] In the resource pool of the first logical functional plane, obtain the AI ​​data corresponding to the AI ​​service requirement;

[0239] The AI ​​data refers to the AI ​​data collected by the second logical functional plane and stored in the resource pool of the first logical functional plane.

[0240] In some embodiments, the second acquisition module 122 is used to:

[0241] In the resource pool of the second logical functional plane, obtain the AI ​​data corresponding to the AI ​​service requirement;

[0242] The AI ​​data can be either AI data collected by the second logical functional surface and stored in the resource pool of the second logical functional surface, or AI data collected by the first logical functional surface and stored in the resource pool of the second logical functional surface.

[0243] In some embodiments, the second acquisition module 122 is used to:

[0244] From the shared resource pool, obtain the AI ​​data corresponding to the AI ​​service request;

[0245] The AI ​​data refers to the AI ​​data collected by the second logical functional plane and stored in the shared resource pool.

[0246] In some embodiments, the second acquisition module 122 is used to:

[0247] Determine the search criteria corresponding to the AI ​​service requirements;

[0248] The AI ​​data is retrieved from the resource pool based on the search criteria.

[0249] In some embodiments, the intelligent surface includes a first module and a second module. The first module is used to acquire the AI ​​data and process the AI ​​data. The second module is used to acquire AI service requests and provide feedback to the target user on the processing results corresponding to the AI ​​service requests.

[0250] In some embodiments, the data plane includes an acquisition module, a storage module, and a retrieval module, wherein the acquisition module is used to acquire AI data, the storage module is used to store AI data, and the retrieval module is used to query AI data that matches AI service requirements from the stored AI data.

[0251] In some embodiments, the second acquisition module 122 is used to:

[0252] The AI ​​service request is sent to the first module of the intelligent surface through the second module of the intelligent surface;

[0253] The first module in the intelligent plane obtains the training data corresponding to the AI ​​service requirement and at least one AI node from the resource pool of the data plane, wherein the at least one AI node is a candidate node for AI model training.

[0254] In some embodiments, the second acquisition module 122 is used to:

[0255] The AI ​​service request is sent from the second module in the intelligent plane to the first module in the intelligent plane. The AI ​​service request is the AI ​​service request obtained by the second module through the data plane.

[0256] The first module in the intelligent surface acquires training data and at least one AI node from the resource pool of the intelligent surface, wherein the at least one AI node is a candidate node for training the AI ​​model.

[0257] Figure 13 is a schematic diagram of another data processing apparatus provided in an embodiment of this disclosure. Based on the embodiment shown in Figure 12, and referring to Figure 13, the data processing apparatus 120 further includes a processing module 123, wherein the processing module 123 is used for:

[0258] The first module processes the at least one AI node to obtain an AI scheduling strategy, which includes the target AI node and the path information of the target AI node.

[0259] The first module sends the AI ​​scheduling strategy and the training data to the second module, and the second module sends the AI ​​scheduling strategy and the training data to the target user.

[0260] In some embodiments, the processing module 123 is configured to:

[0261] The first module is used to pre-allocate AI resources and deploy AI services at the target AI node.

[0262] In some embodiments, the second acquisition module 122 is used to:

[0263] The AI ​​data subscription information is sent to the shared resource pool through the first logical functional plane;

[0264] The subscription information is sent to the second logical functional plane through the shared resource pool;

[0265] The AI ​​data is updated in the shared resource pool using the second logical functional plane and the subscription information;

[0266] The updated AI data is sent to the first logical functional plane through the shared resource pool.

[0267] In some embodiments, the intelligent plane includes AI control functionality, and the data plane includes a data orchestrator (DO), at least one data agent (DA), and a data storage function (DRF), wherein...

[0268] The DO is used to receive AI service requests sent by the AICF and to acquire the data service capabilities of at least one DA;

[0269] The DO is also used to generate a logical network according to the AI ​​service requirements, and to obtain the data corresponding to the target DA in the DRF, wherein the logical network includes the target DA, and the data service capabilities of the target DA are matched with the AI ​​service requirements;

[0270] The DA is used to send the data service capabilities corresponding to the DA to the DO, and to send the data corresponding to the DA to the DRF;

[0271] The DA is also used to execute the data service capabilities corresponding to the DA.

[0272] It should be noted that the division of units in the embodiments of this disclosure is illustrative and only represents one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units described above can be implemented in hardware or as software functional units.

[0273] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a processor-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to related technologies, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0274] It should be noted that the apparatus provided in this disclosure can implement all the method steps implemented in the above method embodiments and can achieve the same technical effect. Therefore, the parts and beneficial effects that are the same as those in the method embodiments will not be described again here.

[0275] Figure 14 is a schematic diagram of an electronic device provided in an embodiment of this disclosure. Referring to Figure 14, the electronic device includes a memory 1410, a transceiver 1420, and a processor 1430.

[0276] The memory 1410 is used to store computer programs;

[0277] The transceiver 1420 is used to send and receive data under the control of the processor;

[0278] The processor 1430 is configured to read the computer program in the memory and perform the following operations:

[0279] AI service requirements are obtained through the first logical functional plane;

[0280] AI data corresponding to AI service requirements is obtained through the first logical functional plane and / or the second logical functional plane.

[0281] Wherein, the first logical functional surface is the data surface, and the second logical functional surface is the intelligence surface; or, the first logical functional surface is the intelligence surface, and the second logical functional surface is the data surface.

[0282] In some embodiments, the AI ​​service requirements include at least one of the following:

[0283] Identification of the target AI model;

[0284] Information about the target AI model;

[0285] Identification of data related to the AI ​​model used;

[0286] The identifier of the training data for the AI ​​model to be trained.

[0287] In some embodiments, the AI ​​data includes at least one of the following:

[0288] Target AI model;

[0289] Data related to the AI ​​model used;

[0290] Training data for the AI ​​model to be trained.

[0291] In some embodiments, the information of the target AI model includes the type of the target AI model, the size of the parameter set of the target AI model, and the storage location of the target AI model;

[0292] In some embodiments, the data related to the AI ​​model used includes the input data and output data corresponding to the AI ​​model used.

[0293] In some embodiments, the first logical functional plane includes a resource pool; obtaining AI data corresponding to AI service requirements through the first logical functional plane includes:

[0294] In the resource pool of the first logical functional plane, obtain the AI ​​data corresponding to the AI ​​service requirement;

[0295] The AI ​​data refers to the AI ​​data collected by the second logical functional plane and stored in the resource pool of the first logical functional plane.

[0296] In some embodiments, the second logical functional plane includes a resource pool; obtaining AI data corresponding to AI service requirements through the second logical functional plane includes:

[0297] In the resource pool of the second logical functional plane, obtain the AI ​​data corresponding to the AI ​​service requirement;

[0298] The AI ​​data can be either AI data collected by the second logical functional surface and stored in the resource pool of the second logical functional surface, or AI data collected by the first logical functional surface and stored in the resource pool of the second logical functional surface.

[0299] In some embodiments, the first logical functional plane and the second logical functional plane include a shared resource pool; obtaining AI data corresponding to AI service requirements through the first logical functional plane and the second logical functional plane includes:

[0300] From the shared resource pool, obtain the AI ​​data corresponding to the AI ​​service request;

[0301] The AI ​​data refers to the AI ​​data collected by the second logical functional plane and stored in the shared resource pool.

[0302] In some embodiments, obtaining AI data corresponding to the AI ​​service request includes:

[0303] Determine the search criteria corresponding to the AI ​​service requirements;

[0304] The AI ​​data is retrieved from the resource pool based on the search criteria.

[0305] In some embodiments, the intelligent surface includes a first module and a second module. The first module is used to acquire the AI ​​data and process the AI ​​data. The second module is used to acquire AI service requests and provide feedback to the target user on the processing results corresponding to the AI ​​service requests.

[0306] In some embodiments, the data plane includes an acquisition module, a storage module, and a retrieval module, wherein the acquisition module is used to acquire AI data, the storage module is used to store AI data, and the retrieval module is used to query AI data that matches AI service requirements from the stored AI data.

[0307] In some embodiments, the first logical functional plane is an intelligence plane, the second logical functional plane is a data plane, the data plane includes a resource pool, and the AI ​​data is training data; obtaining AI data corresponding to AI service requirements through the first and second logical functional planes includes:

[0308] The AI ​​service request is sent to the first module of the intelligent surface through the second module of the intelligent surface;

[0309] The first module in the intelligent plane obtains the training data corresponding to the AI ​​service requirement and at least one AI node from the resource pool of the data plane, wherein the at least one AI node is a candidate node for AI model training.

[0310] In some embodiments, the first logical functional plane is a data plane, the second logical functional plane is an intelligence plane, the intelligence plane includes a resource pool, and the AI ​​data is the training data; obtaining AI data corresponding to AI service requirements through the first and second logical functional planes includes:

[0311] The AI ​​service request is sent from the second module in the intelligent plane to the first module in the intelligent plane. The AI ​​service request is the AI ​​service request obtained by the second module through the data plane.

[0312] The first module in the intelligent surface acquires training data and at least one AI node from the resource pool of the intelligent surface, wherein the at least one AI node is a candidate node for training the AI ​​model.

[0313] In some embodiments, the method further comprises:

[0314] The first module processes the at least one AI node to obtain an AI scheduling strategy, which includes the target AI node and the path information of the target AI node.

[0315] The first module sends the AI ​​scheduling strategy and the training data to the second module, and the second module sends the AI ​​scheduling strategy and the training data to the target user.

[0316] In some embodiments, the method further comprises:

[0317] The first module is used to pre-allocate AI resources and deploy AI services at the target AI node.

[0318] In some embodiments, the method further comprises:

[0319] The AI ​​data subscription information is sent to the shared resource pool through the first logical functional plane;

[0320] The subscription information is sent to the second logical functional plane through the shared resource pool;

[0321] The AI ​​data is updated in the shared resource pool using the second logical functional plane and the subscription information;

[0322] The updated AI data is sent to the first logical functional plane through the shared resource pool.

[0323] In some embodiments, the intelligent plane includes AI control functionality, and the data plane includes a data orchestrator (DO), at least one data agent (DA), and a data storage function (DRF), wherein...

[0324] The DO is used to receive AI service requests sent by the AICF and to acquire the data service capabilities of at least one DA;

[0325] The DO is also used to generate a logical network according to the AI ​​service requirements, and to obtain the data corresponding to the target DA in the DRF, wherein the logical network includes the target DA, and the data service capabilities of the target DA are matched with the AI ​​service requirements;

[0326] The DA is used to send the data service capabilities corresponding to the DA to the DO, and to send the data corresponding to the DA to the DRF;

[0327] The DA is also used to execute the data service capabilities corresponding to the DA.

[0328] In Figure 14, the bus architecture may include any number of interconnected buses and bridges, linking various circuits together, including one or more processors represented by processor 1430 and memory represented by memory 1410. 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. Transceiver 1420 may be multiple elements, including transmitters and receivers, providing units for communicating with various other devices over transmission media, including wireless channels, wired channels, optical fibers, etc. Processor 1430 is responsible for managing the bus architecture and general processing, and memory 1410 may store data used by processor 1430 during operation.

[0329] In some embodiments, the processor 1430 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), and the processor may also adopt a multi-core architecture.

[0330] It should be noted that the physical device provided in this disclosure can implement all the method steps implemented by the physical device in the above method embodiments and can achieve the same technical effect. Therefore, the parts and beneficial effects that are the same as those in the method embodiments will not be described again here.

[0331] This disclosure also provides a processor-readable storage medium storing a computer program for causing a processor to perform the method described in any of the above method embodiments.

[0332] Processor-readable storage media can be any available medium or data storage device that a computer can access, including but not limited to magnetic storage (such as floppy disks, hard disks, magnetic tapes, magneto-optical disks (MOs), etc.), optical storage (such as CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (such as ROMs, EPROMs, EEPROMs, non-volatile memory (NAND flash), solid-state drives (SSDs)).

[0333] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the method described in any of the above method embodiments.

[0334] Those skilled in the art will understand that embodiments of this disclosure can be provided as methods, systems, or computer program products. Therefore, this disclosure can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this disclosure can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0335] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-executable instructions. These computer-executable instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more flowchart illustrations and / or one or more block diagrams.

[0336] These processor-executable instructions may also be stored in a processor-readable memory that can instruct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the processor-readable memory produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0337] These processor-executable instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

[0338] Obviously, those skilled in the art can make various modifications and variations to this disclosure without departing from its spirit and scope. Therefore, if such modifications and variations fall within the scope of the claims of this disclosure and their equivalents, this disclosure is also intended to include such modifications and variations.

Claims

1. A data processing method, wherein, include: AI service requirements are obtained through the first logical functional plane; AI data corresponding to AI service requirements is obtained through the first logical functional plane and / or the second logical functional plane. Wherein, the first logical functional surface is the data surface, and the second logical functional surface is the intelligence surface; or, the first logical functional surface is the intelligence surface, and the second logical functional surface is the data surface.

2. The method according to claim 1, wherein, The AI ​​service requirements include at least one of the following: Identification of the target AI model; Information about the target AI model; Identification of data related to the AI ​​model used; The identifier of the training data for the AI ​​model to be trained.

3. The method according to claim 1, wherein, The AI ​​data includes at least one of the following: Target AI model; Data related to the AI ​​model used; Training data for the AI ​​model to be trained.

4. The method according to claim 2, wherein, The information of the target AI model includes the type of the target AI model, the size of the parameter set of the target AI model, and the storage location of the target AI model; The data related to the AI ​​model used includes the input data and output data corresponding to the AI ​​model used.

5. The method according to claim 1, wherein, The first logical functional plane includes a resource pool; obtaining AI data corresponding to AI service requirements through the first logical functional plane includes: In the resource pool of the first logical functional plane, obtain the AI ​​data corresponding to the AI ​​service requirement; The AI ​​data refers to the AI ​​data collected by the second logical functional plane and stored in the resource pool of the first logical functional plane.

6. The method according to claim 1, wherein, The second logical functional plane includes a resource pool; it obtains AI data corresponding to AI service requirements through the second logical functional plane, including: In the resource pool of the second logical functional plane, obtain the AI ​​data corresponding to the AI ​​service requirement; The AI ​​data can be either AI data collected by the second logical functional surface and stored in the resource pool of the second logical functional surface, or AI data collected by the first logical functional surface and stored in the resource pool of the second logical functional surface.

7. The method according to claim 1, wherein, The first and second logical functional planes include a shared resource pool; AI data corresponding to AI service requirements is obtained through the first and second logical functional planes, including: From the shared resource pool, obtain the AI ​​data corresponding to the AI ​​service request; The AI ​​data refers to the AI ​​data collected by the second logical functional plane and stored in the shared resource pool.

8. The method according to any one of claims 5-7, wherein, Obtaining the AI ​​data corresponding to the AI ​​service request includes: Determine the search criteria corresponding to the AI ​​service requirements; The AI ​​data is retrieved from the resource pool based on the search criteria.

9. The method according to any one of claims 1-7, wherein, The intelligent interface includes a first module and a second module. The first module is used to acquire the AI ​​data and process the AI ​​data. The second module is used to acquire AI service requests and provide feedback to the target user on the processing results corresponding to the AI ​​service requests.

10. The method according to any one of claims 1-7, wherein, The data plane includes a collection module, a storage module, and a retrieval module. The collection module is used to collect AI data, the storage module is used to store AI data, and the retrieval module is used to query AI data that matches the AI ​​service requirements from the stored AI data.

11. The method according to claim 9, wherein, The first logical functional plane is the intelligence plane, and the second logical functional plane is the data plane. The data plane includes a resource pool, and the AI ​​data is training data. AI data corresponding to AI service requirements is obtained through the first and second logical functional planes, including: The AI ​​service request is sent to the first module of the intelligent surface through the second module of the intelligent surface; The first module in the intelligent plane obtains the training data corresponding to the AI ​​service requirement and at least one AI node from the resource pool of the data plane, wherein the at least one AI node is a candidate node for AI model training.

12. The method according to claim 9, wherein, The first logical functional plane is the data plane, and the second logical functional plane is the intelligence plane. The intelligence plane includes a resource pool, and the AI ​​data is training data. AI data corresponding to AI service requirements is obtained through the first and second logical functional planes, including: The AI ​​service request is sent from the second module in the intelligent plane to the first module in the intelligent plane, wherein the AI ​​service request is the AI ​​service request obtained by the second module through the data plane; The first module in the intelligent surface acquires training data and at least one AI node from the resource pool of the intelligent surface, wherein the at least one AI node is a candidate node for training the AI ​​model.

13. The method according to claim 11 or 12, wherein, The method further includes: The first module processes the at least one AI node to obtain an AI scheduling strategy, which includes the target AI node and the path information of the target AI node. The first module sends the AI ​​scheduling strategy and the training data to the second module, and the second module sends the AI ​​scheduling strategy and the training data to the target user.

14. The method according to claim 13, wherein, The method further includes: The first module is used to pre-allocate AI resources and deploy AI services at the target AI node.

15. The method according to claim 7, wherein, The method further includes: The AI ​​data subscription information is sent to the shared resource pool through the first logical functional plane; The subscription information is sent to the second logical functional plane through the shared resource pool; The AI ​​data is updated in the shared resource pool using the second logical functional plane and the subscription information; The updated AI data is sent to the first logical functional plane through the shared resource pool.

16. The method according to claim 1, wherein, The intelligent plane includes AI control functions, and the data plane includes a data orchestrator (DO), at least one data agent (DA), and a data storage function (DRF). The DO is used to receive AI service requests sent by the AICF and to acquire the data service capabilities of at least one DA; The DO is also used to generate a logical network according to the AI ​​service requirements, and to obtain the data corresponding to the target DA in the DRF, wherein the logical network includes the target DA, and the data service capabilities of the target DA are matched with the AI ​​service requirements; The DA is used to send the data service capabilities corresponding to the DA to the DO, and to send the data corresponding to the DA to the DRF; The DA is also used to execute the data service capabilities corresponding to the DA.

17. A data processing apparatus, wherein, It includes a first acquisition module and a second acquisition module, wherein: The first acquisition module is used to acquire AI service requirements through the first logical functional plane; The second acquisition module is used to acquire AI data corresponding to AI service requirements through the first logical functional plane and / or the second logical functional plane; Wherein, the first logical functional surface is the data surface, and the second logical functional surface is the intelligence surface; or, the first logical functional surface is the intelligence surface, and the second logical functional surface is the data surface.

18. An electronic device, wherein, Includes memory, transceiver, and processor: The memory is used to store computer programs; The transceiver is used to send and receive data under the control of the processor; The processor is configured to read the computer program from the memory and perform the following operations: AI service requirements are obtained through the first logical functional plane; AI data corresponding to AI service requirements is obtained through the first logical functional plane and / or the second logical functional plane. Wherein, the first logical functional surface is the data surface, and the second logical functional surface is the intelligence surface; or, the first logical functional surface is the intelligence surface, and the second logical functional surface is the data surface.

19. A processor-readable storage medium, wherein, The processor-readable storage medium stores a computer program that causes the processor to perform the method according to any one of claims 1 to 15.

Images