Model identifier configuration method and device and network equipment

Abstract

The embodiment of the invention provides a model identifier configuration method and device and network equipment, and the method comprises the steps: determining model identifier information corresponding to an AI model based on the generation position of the AI model for any AI model; generating a target model identifier corresponding to the AI model based on a definition mode corresponding to the AI model, and a mobile network identifier, a base station identifier, a cell identifier and model identifier information corresponding to the AI model; wherein the definition mode corresponding to the AI model comprises a basic mode and a non-basic mode; and generating model identification configuration information based on the indication information corresponding to the generation position and the target model identifier, and sending the model identification configuration information to the equipment corresponding to the generation position. By generating the target model identifier which uniquely identifies the AI model globally, identifier conflicts do not occur even if the AI model crosses domains, and the identifier requirements of massive network nodes in a 6G network are met; in addition, the global equipment adopts a unified identification mode, so that the model identification is convenient to manage.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of communication technology, and in particular to a model identification configuration method, device and network equipment. Background Art

[0002] Life Cycle Management (LCM) is an important concept proposed in the current 3rd Generation Partnership Project (3GPP) standardization research, covering artificial intelligence (AI) model data collection, model training, model reasoning, model management, model activation / deactivation, model switching, rollback, model monitoring, model update and other processes. AI model identification is an important process in life cycle management. Processes such as model reasoning and model management require a sound model identification mechanism as a basis.

[0003] In the related technology, each device informs the network side of the model identification (Identity ID) of the AI ​​model supported, and the network side completes AI model recognition based on the model ID.

[0004] However, the sixth generation mobile communication technology (6G) network involves the participation of massive network nodes. The use of a single model ID is prone to ID conflicts, making it difficult to achieve cross-domain recognition of AI models. Summary of the invention

[0005] Embodiments of the present application provide a model identification configuration method, apparatus, and network device to provide model identification for cross-domain identification.

[0006] In a first aspect, an embodiment of the present application provides a first model identification configuration method, the method comprising:

[0007] For any AI model, determine the model identification information corresponding to the AI ​​model based on the generation location of the AI ​​model;

[0008] Generate a target model identifier corresponding to the AI ​​model based on a definition mode corresponding to the AI ​​model, a mobile network identifier, a base station identifier, a cell identifier, and the model identifier information corresponding to the AI ​​model; wherein the definition mode corresponding to the AI ​​model includes a basic mode and a non-basic mode;

[0009] Model recognition configuration information is generated based on the indication information corresponding to the generation location and the target model identifier, and the model recognition configuration information is sent to the device corresponding to the generation location.

[0010] In some optional implementations, determining the model identification information corresponding to the AI ​​model based on the generation location of the AI ​​model includes:

[0011] If the generation location of the AI ​​model is the network side, determining that the model identification information corresponding to the AI ​​model is an identification of the AI ​​model in the corresponding cell; or

[0012] If the generation location of the AI ​​model is the terminal side, the model identification information corresponding to the AI ​​model is determined to be the terminal identification corresponding to the AI ​​model and the identification of the AI ​​model in the corresponding terminal.

[0013] In some optional implementations, generating a target model identifier corresponding to the AI ​​model based on a definition mode corresponding to the AI ​​model, a mobile network identifier, a base station identifier, a cell identifier, and the model identifier information corresponding to the AI ​​model includes:

[0014] If the definition mode corresponding to the AI ​​model is the basic mode, a preset calculation method is used to calculate the mobile network identifier, the base station identifier, the cell identifier and the model identifier information to obtain the target model identifier.

[0015] In some optional implementations, generating a target model identifier corresponding to the AI ​​model based on a definition mode corresponding to the AI ​​model, a mobile network identifier, a base station identifier, a cell identifier, and the model identifier information corresponding to the AI ​​model includes:

[0016] If the definition mode corresponding to the AI ​​model is a non-basic mode, the mobile network identifier, the base station identifier, the cell identifier, and the model identifier information are calculated using a preset calculation method to obtain a first model identifier corresponding to the AI ​​model;

[0017] The first model identifier is combined with target additional information to obtain the target model identifier; wherein the target additional information is determined based on the definition mode.

[0018] In some optional implementations, the non-basic mode includes association model definition, network slice definition, use case definition, function definition, and network element definition.

[0019] In some optional implementation manners, the target additional information is determined in the following manner:

[0020] If the definition mode is an associated model definition, the identifiers of other models associated with the AI ​​model are determined as the target additional information; or

[0021] If the definition mode is a network slice definition, an identifier of the network slice for generating the AI ​​model and an identifier of the AI ​​model in the network slice are determined as the target additional information; or

[0022] If the definition mode is a use case definition, an identifier of the use case corresponding to the AI ​​model and an identifier of the AI ​​model in the use case are determined as the target additional information; or

[0023] If the definition mode is function definition, an identifier of a function granularity corresponding to the function of the AI ​​model and an identifier of the function of the AI ​​model in the function granularity are determined as the target additional information; or

[0024] If the definition mode is network element definition, the identifier of the network element corresponding to the AI ​​model is determined as the target additional information.

[0025] In some optional implementations, combining the first model identifier with the target additional information to obtain the target model identifier includes:

[0026] The first model identifier and the target additional information are combined based on the position information corresponding to the definition mode to obtain the target model identifier.

[0027] In some optional implementations, the model identification configuration information also includes indication information corresponding to the definition mode.

[0028] In some optional implementations, generating model identification configuration information based on the indication information corresponding to the generation position and the target model identifier includes:

[0029] If the target model identifier does not exceed a preset number of digits, generating model identification configuration information including indication information corresponding to the generation position and the target model identifier; or

[0030] If the target model identifier exceeds the preset number of digits, the target model identifier is divided into a plurality of sub-identifiers; and for any sub-identifier, model recognition configuration information including indication information corresponding to the generation position and the sub-identifier is generated.

[0031] In a second aspect, an embodiment of the present application provides a first model identification configuration device, the device comprising:

[0032] An information determination module, configured to determine, for any AI model, model identification information corresponding to the AI ​​model based on a generation location of the AI ​​model;

[0033] An identification determination module, used to generate a target model identification corresponding to the AI ​​model based on a definition mode corresponding to the AI ​​model, a mobile network identification, a base station identification, a cell identification, and the model identification information corresponding to the AI ​​model; wherein the definition mode corresponding to the AI ​​model includes a basic mode and a non-basic mode;

[0034] The identification configuration module is used to generate model identification configuration information based on the indication information corresponding to the generation position and the target model identification, and send the model identification configuration information to the device corresponding to the generation position.

[0035] In some optional implementation manners, the information determination module is specifically used to:

[0036] If the generation location of the AI ​​model is the network side, determining that the model identification information corresponding to the AI ​​model is an identification of the AI ​​model in the corresponding cell; or

[0037] If the generation location of the AI ​​model is the terminal side, the model identification information corresponding to the AI ​​model is determined to be the terminal identification corresponding to the AI ​​model and the identification of the AI ​​model in the corresponding terminal.

[0038] In some optional implementation manners, the identification determination module is specifically used to:

[0039] If the definition mode corresponding to the AI ​​model is the basic mode, a preset calculation method is used to calculate the mobile network identifier, the base station identifier, the cell identifier and the model identifier information to obtain the target model identifier.

[0040] In some optional implementation manners, the identification determination module is specifically used to:

[0041] If the definition mode corresponding to the AI ​​model is a non-basic mode, the mobile network identifier, the base station identifier, the cell identifier, and the model identifier information are calculated using a preset calculation method to obtain a first model identifier corresponding to the AI ​​model;

[0042] The first model identifier is combined with target additional information to obtain the target model identifier; wherein the target additional information is determined based on the definition mode.

[0043] In some optional implementations, the non-basic mode includes association model definition, network slice definition, use case definition, function definition, and network element definition.

[0044] In some optional implementation manners, the identification determination module is further configured to determine the target additional information in the following manner:

[0045] If the definition mode is an associated model definition, the identifiers of other models associated with the AI ​​model are determined as the target additional information; or

[0046] If the definition mode is a network slice definition, an identifier of the network slice for generating the AI ​​model and an identifier of the AI ​​model in the network slice are determined as the target additional information; or

[0047] If the definition mode is a use case definition, an identifier of the use case corresponding to the AI ​​model and an identifier of the AI ​​model in the use case are determined as the target additional information; or

[0048] If the definition mode is function definition, an identifier of a function granularity corresponding to the function of the AI ​​model and an identifier of the function of the AI ​​model in the function granularity are determined as the target additional information; or

[0049] If the definition mode is network element definition, the identifier of the network element corresponding to the AI ​​model is determined as the target additional information.

[0050] In some optional implementation manners, the identification determination module is specifically used to:

[0051] The first model identifier and the target additional information are combined based on the position information corresponding to the definition mode to obtain the target model identifier.

[0052] In some optional implementations, the model identification configuration information also includes indication information corresponding to the definition mode.

[0053] In some optional implementations, the identification configuration module is used to:

[0054] If the target model identifier does not exceed a preset number of digits, generating model identification configuration information including indication information corresponding to the generation position and the target model identifier; or

[0055] If the target model identifier exceeds the preset number of digits, the target model identifier is divided into a plurality of sub-identifiers; and for any sub-identifier, model recognition configuration information including indication information corresponding to the generation position and the sub-identifier is generated.

[0056] In a third aspect, an embodiment of the present application provides a network device, comprising at least one processor and at least one memory, wherein the memory stores a computer program, and when the program is executed by the processor, the processor executes any model identification configuration method described in the first aspect above.

[0057] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium storing a computer program executable by a processor, wherein when the program runs on the processor, the processor executes any of the model identification configuration methods described in the first aspect.

[0058] In the embodiment of the present application, based on the definition mode corresponding to the AI ​​model, the mobile network identifier (used to identify the mobile network of the device where the AI ​​model is located), the base station identifier (used to identify the base station corresponding to the device where the AI ​​model is located), the cell identifier (used to identify the cell corresponding to the device where the AI ​​model is located) and the model identification information (the identifier within the location unit, related to the location), the first item of information realizes the multi-dimensional identification definition, and the last four items ensure the global uniqueness of the identifier, thereby generating a target model identifier that globally uniquely identifies the AI ​​model in a specific scenario, and there will be no identification conflict even across domains, meeting the identification requirements and multi-dimensional management requirements of massive network nodes in the 6G network; in addition, global devices adopt a unified identification method instead of each device adopting its own identification method, which facilitates the management of model identification; by sending model identification configuration information to the device corresponding to the generation location of the AI ​​model, the relevant network nodes are aware of the identification-related information. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0060] Figure 1 A flowchart of the first model identification configuration method provided in an embodiment of the present application;

[0061] Figure 2 A schematic diagram of a flow chart of a second model identification configuration method provided in an embodiment of the present application;

[0062] Figure 3 A schematic diagram of a flow chart of a third model identification configuration method provided in an embodiment of the present application;

[0063] Figure 4 A schematic diagram of target model identification of the first non-basic mode provided in an embodiment of the present application;

[0064] Figure 5 A schematic diagram of target model identification of the second non-basic mode provided in an embodiment of the present application;

[0065] Figure 6 A schematic diagram of target additional information provided in an embodiment of the present application;

[0066] Figure 7 A schematic diagram of a flow chart of a fourth model identification configuration method provided in an embodiment of the present application;

[0067] Figure 8 A schematic diagram of the structure of a model identification configuration device provided in an embodiment of the present application;

[0068] Fig. 9 A schematic diagram of the structure of a network device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0069] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0070] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0071] In the description of this application, it should be noted that, unless otherwise clearly specified and limited, the term "connection" should be understood in a broad sense, for example, it can be directly connected, or indirectly connected through an intermediate medium, or it can be the internal connection of two devices. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0072] Lifecycle management is an important concept proposed in the current 3GPP standardization research. The importance of AI model lifecycle management lies in that it ensures the quality and efficiency of AI models throughout their entire lifecycle and reduces potential usage costs. AI model lifecycle management covers processes such as AI model data collection, model training, model reasoning, model management, model activation / deactivation, model switching, rollback, model monitoring, and model update.

[0073] AI model identification is an important process in lifecycle management. Model reasoning and model management, for example, all require a sound model identification mechanism as a basis. Specific identification information is assigned to AI models within the network, and this identification information is notified to relevant nodes through broadcast or point-to-point transmission, thereby achieving automatic identification and unified management of the model on the network and terminal sides, facilitating subsequent model reasoning, activation, deactivation, rollback, switching and other operations.

[0074] In the related technology, each device informs the network side of the model ID of the AI ​​model supported, and the network side completes AI model identification based on the model ID.

[0075] However, the 6G network involves the participation of a large number of network nodes. The use of a single model ID is prone to ID conflicts, making it difficult to achieve cross-domain recognition of AI models. In addition, the model identification methods of various devices are not unified, making it difficult to manage model IDs.

[0076] In view of this, an embodiment of the present application proposes a model identification configuration method, device and network equipment, the method comprising: for any AI model, based on the generation location of the AI ​​model, determining the model identification information corresponding to the AI ​​model; based on the definition mode corresponding to the AI ​​model, the mobile network identification, base station identification, cell identification and the model identification information corresponding to the AI ​​model, generating a target model identification corresponding to the AI ​​model; wherein the definition mode corresponding to the AI ​​model includes a basic mode and a non-basic mode; based on the indication information corresponding to the generation location and the target model identification, generating model identification configuration information, and sending the model identification configuration information to the device corresponding to the generation location.

[0077] The above scheme is based on the definition mode corresponding to the AI ​​model, the mobile network identifier (used to identify the mobile network of the device where the AI ​​model is located), the base station identifier (used to identify the base station corresponding to the device where the AI ​​model is located), the cell identifier (used to identify the cell corresponding to the device where the AI ​​model is located) and the model identification information (the identifier within the location unit, related to the location). The first item of information realizes the multi-dimensional identification definition, and the last four items ensure the global uniqueness of the identifier, thereby generating a target model identifier that globally uniquely identifies the AI ​​model in a specific scenario. Even across domains, there will be no identification conflicts, which meets the identification requirements and multi-dimensional management requirements of massive network nodes in the 6G network; in addition, global devices use a unified identification method instead of each device using its own identification method, which facilitates the management of model identification; by sending model identification configuration information to the device corresponding to the generation location of the AI ​​model, the relevant network nodes are aware of the identification-related information.

[0078] The following will be combined with the accompanying drawings and specific embodiments to describe in detail the technical solution of the present application and how the technical solution of the present application solves the above technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0079] Figure 1A flow chart of a first model identification configuration method provided in an embodiment of the present application is applied to a network device, which may be an access and mobility management function (Access and Mobility Management Function, AMF); Figure 1 As shown, the following steps are included:

[0080] Step S101: for any AI model, based on the generation location of the AI ​​model, determine the model identification information corresponding to the AI ​​model.

[0081] In implementation, some location units (such as a community) will deploy more than one AI model, and internal identification (model identification information) is required to distinguish different AI models in the same location unit.

[0082] The method of determining the model identification information is related to the generation location of the AI ​​model. The network side only needs to consider the identification within the cell, while the terminal side needs to consider not only the identification of the terminal in the cell, but also the identification of the AI ​​model in the terminal.

[0083] Step S102: Generate a target model identifier corresponding to the AI ​​model based on a definition mode corresponding to the AI ​​model, a mobile network identifier, a base station identifier, a cell identifier corresponding to the AI ​​model, and the model identifier information.

[0084] Among them, the definition mode corresponding to the AI ​​model includes a basic mode and a non-basic mode.

[0085] The model identification information represents the identification within the location unit, and the same model identification information may appear in different location units; therefore, the identification outside the location unit must also be considered. The cell is used as the location unit, and the upper level is the base station. Different operators have different base stations.

[0086] Based on this, this embodiment integrates the mobile network identifier PLMN corresponding to the AI ​​model (used to identify the mobile network of the device where the AI ​​model is located, which is related to the operator), the base station identifier gNB ID (used to identify the base station corresponding to the device where the AI ​​model is located), and the cell identifier Cell ID (used to identify the cell corresponding to the device where the AI ​​model is located), as well as the model identification information within the cell, to generate a target model identifier that uniquely identifies the AI ​​model globally.

[0087] In addition, since 6G networks need to integrate AI technology to enable more intelligent application scenarios, different scenarios require different information;

[0088] Based on this, this embodiment sets up multiple definition modes to implement multi-dimensional identification definition and meet the multi-dimensional management requirements in 6G networks.

[0089] Step S103: Generate model recognition configuration information based on the indication information corresponding to the generation location and the target model identifier, and send the model recognition configuration information to the device corresponding to the generation location.

[0090] In implementation, after the network device generates a target model identifier that globally uniquely identifies the AI ​​model, it is also necessary to inform the relevant network nodes of the target model identifier based on the generation location;

[0091] Since the network device may need to inform the device corresponding to the generation location through other nodes, it is necessary to generate model recognition configuration information based on the indication information corresponding to the generation location and the target model identifier, and send the model recognition configuration information directly or indirectly to the device corresponding to the generation location of the AI ​​model.

[0092] Currently, 3GPP TR38.843 specifies three scenarios for model generation: terminal side, network side, and both terminal and network sides;

[0093] Exemplarily, for the network side, the model identification configuration information can be directly sent to the corresponding base station, such as sending the model identification configuration signaling to the base station through the NG interface (the interface between the wireless access network and the 5G core network); for the terminal side, the model identification configuration information needs to be sent to the corresponding terminal through the base station, such as the base station addresses the specified UE based on the user equipment (UserEquipment, UE) ID, and configures the global model identifier to the UE through the radio resource control (Radio Resource Control, RRC) signaling or downlink control information (Downlink Control Information, DCI), or through the packet data fusion protocol (Packet Data Convergence Protocol, PDCP) user plane to complete the model identification process; for both sides of the terminal network, the model identification configuration information on both sides needs to be sent to the corresponding base station, the base station retains its own corresponding model identification configuration information, and sends its corresponding model identification configuration information to the terminal.

[0094] This embodiment does not specifically limit the indication information corresponding to the above-mentioned generation position, and the generation position can be represented by different numbers of specific bits. When the generation position includes the terminal side, the network side, and both the terminal and the network, it can be indicated by two bits, such as 00 represents the terminal side, 01 represents the network side, and 10 represents both the terminal and the network.

[0095] The indication information corresponding to the above-mentioned generation position is only an exemplary description and can be set according to actual needs during implementation.

[0096] The above scheme is based on the definition mode corresponding to the AI ​​model, the mobile network identifier (used to identify the mobile network of the device where the AI ​​model is located), the base station identifier (used to identify the base station corresponding to the device where the AI ​​model is located), the cell identifier (used to identify the cell corresponding to the device where the AI ​​model is located) and the model identification information (the identifier within the location unit, related to the location). The first item of information realizes the multi-dimensional identification definition, and the last four items ensure the global uniqueness of the identifier, thereby generating a target model identifier that globally uniquely identifies the AI ​​model in a specific scenario. Even across domains, there will be no identification conflicts, which meets the identification requirements and multi-dimensional management requirements of massive network nodes in the 6G network; in addition, global devices use a unified identification method instead of each device using its own identification method, which facilitates the management of model identification; by sending model identification configuration information to the device corresponding to the generation location of the AI ​​model, the relevant network nodes are aware of the identification-related information.

[0097] In some optional implementations, for the terminal side or both the terminal and network sides, after receiving the model identification configuration information, the terminal replies with the model identification configuration completion signaling. The model identification configuration completion signaling is carried through the uplink DCI or the media access control layer control unit (MAC Control Element, MAC CE), or through the RRC reconfiguration completion (RRCReconfigurationComplete) signaling.

[0098] In some optional implementation manners, the above step S101 may be implemented by but not limited to the following methods:

[0099] If the generation location of the AI ​​model is the network side, determining that the model identification information corresponding to the AI ​​model is an identification of the AI ​​model in the corresponding cell; or

[0100] If the generation location of the AI ​​model is the terminal side, the model identification information corresponding to the AI ​​model is determined to be the terminal identification corresponding to the AI ​​model and the identification of the AI ​​model in the corresponding terminal.

[0101] The method of determining the model identification information as described above is related to the generation location of the AI ​​model. The network side only needs to consider the identification within the cell, while the terminal side needs to consider not only the identification of the terminal within the cell, but also the identification of the AI ​​model in the terminal.

[0102] Based on this, in this embodiment, for the AI ​​model on the network side, the corresponding model identification information is the identifier Cell Model ID of the AI ​​model in the corresponding cell;

[0103] For the AI ​​model on the terminal side, only the identification within the terminal is considered. The AI ​​identification in different terminals in the same cell may conflict. Therefore, the model identification information includes the terminal identification UE ID corresponding to the AI ​​model (the identification of the terminal deploying the AI ​​model in the cell) and the identification UE Model ID of the AI ​​model in the corresponding terminal.

[0104] Exemplarily, the UE ID can be the 5G-Temporary Mobile Subscriber Identity (TMSI) field of the 5G-Globally Unique Temporary Identifier (GUTI), a total of 32 bits. AMF identifies the terminal through the 5G-TMSI, and the TMSI is unique within the AMF. The UE Model ID is 4 bits. The Cell Model ID can be configured to 4 to 16 bits, which is associated with the number of AI models in the cell, the UE capabilities and AI features, and is also associated with other conditions such as configuration, scenarios, sites, etc.

[0105] Figure 2 A flow chart of the second model identification configuration method provided in the embodiment of the present application is as follows: Figure 2 As shown, the following steps are included:

[0106] Step S201: for any AI model, based on the generation location of the AI ​​model, determine the model identification information corresponding to the AI ​​model.

[0107] The specific implementation of step S201 may refer to other embodiments and will not be described in detail here.

[0108] Step S202: If the definition mode corresponding to the AI ​​model is the basic mode, a preset calculation method is used to calculate the mobile network identifier, the base station identifier, the cell identifier and the model identifier information to obtain the target model identifier.

[0109] During implementation, for the AI ​​model definition of the basic mode, only the above-mentioned information needs to be considered, and the mobile network identifier, base station identifier, cell identifier and model identifier information are calculated to obtain the target model identifier.

[0110] Taking the model identification information corresponding to the AI ​​model on the network side as the identification CellModel ID of the AI ​​model in the corresponding cell as an example, the target model identification GID = gNB ID*2 (36-length(gNB ID) +length(Cell Model ID)+CellID*2 length(Cell ModelID) +Cell Model ID+PLMN; where gNB ID is the base station identifier, Cell Model ID is the identifier of the AI ​​model in the corresponding cell, Cell ID is the cell identifier, and PLMN is the mobile network identifier.

[0111] Taking the model identification information corresponding to the AI ​​model on the terminal side including the terminal identification UE ID corresponding to the AI ​​model and the identification UE Model ID of the AI ​​model in the corresponding terminal as an example, the target model identification GID = gNB ID*2 (36-length(gNB ID) +Cell ID+UE ID*2 length(UE Model ID) +UE Model ID+PLMN; where gNB ID is the base station identifier, Cell ID is the cell identifier, UE ID is the terminal identifier, UE Model ID is the identifier of the AI ​​model in the corresponding terminal, and PLMN is the mobile network identifier.

[0112] Step S203: Generate model recognition configuration information based on the indication information corresponding to the generation location and the target model identifier, and send the model recognition configuration information to the device corresponding to the generation location.

[0113] The specific implementation of step S203 may refer to other embodiments and will not be described in detail here.

[0114] Figure 3 A flow chart of a third model identification configuration method provided in an embodiment of the present application is shown as follows: Figure 3 As shown, the following steps are included:

[0115] Step S301: for any AI model, based on the generation location of the AI ​​model, determine the model identification information corresponding to the AI ​​model.

[0116] The specific implementation of step S301 may refer to other embodiments and will not be described in detail here.

[0117] Step S302: If the definition mode corresponding to the AI ​​model is a non-basic mode, a preset calculation method is used to calculate the mobile network identifier, the base station identifier, the cell identifier and the model identifier information to obtain a first model identifier corresponding to the AI ​​model.

[0118] Since 6G networks need to integrate AI technology to enable more intelligent application scenarios, special information needs to be labeled in some scenarios to facilitate the management of AI models;

[0119] Based on this, for the AI ​​model definition of non-basic mode, this embodiment does not directly use the information calculated from the mobile network identifier, base station identifier, cell identifier and model identifier information as the target model identifier, but uses it as the first model identifier and combines it with some special target additional information.

[0120] Step S303: Combine the first model identifier with the target additional information to obtain the target model identifier; wherein the target additional information is determined based on the definition mode.

[0121] As described above, for the AI ​​model definition of the basic mode, it is necessary to combine the first model identifier with some special target additional information, where the target additional information is special information related to the defined mode.

[0122] Step S304: Generate model recognition configuration information based on the indication information corresponding to the generation location and the target model identifier, and send the model recognition configuration information to the device corresponding to the generation location.

[0123] The specific implementation of step S304 may refer to other embodiments and will not be described in detail here.

[0124] The above scheme, for the AI ​​model definition of non-basic mode, does not directly use the information calculated from the mobile network identifier, base station identifier, cell identifier and model identifier information as the target model identifier, but uses it as the first model identifier and combines it with the target additional information related to the defined mode to facilitate AI model management and realize multi-dimensional global model identifier definition.

[0125] In some optional implementations, the non-basic mode includes association model definition, network slice definition, use case definition, function definition, and network element definition.

[0126] Exemplarily, the above-mentioned association model definition means that the AI ​​model that needs to be identified is associated with other models; the above-mentioned network slice definition means that the AI ​​model that needs to be identified is generated based on network slice resources; the above-mentioned use case definition means that the AI ​​model that needs to be identified corresponds to a specific use case; the above-mentioned function definition means that the AI ​​model that needs to be identified corresponds to a specific function; the above-mentioned network element definition means that the AI ​​model that needs to be identified is generated and / or managed by a specific network element.

[0127] In some optional implementation manners, the target additional information is determined in the following manner:

[0128] If the definition mode is an associated model definition, the identifiers of other models associated with the AI ​​model are determined as the target additional information; or

[0129] If the definition mode is a network slice definition, an identifier of the network slice for generating the AI ​​model and an identifier of the AI ​​model in the network slice are determined as the target additional information; or

[0130] If the definition mode is a use case definition, an identifier of the use case corresponding to the AI ​​model and an identifier of the AI ​​model in the use case are determined as the target additional information; or

[0131] If the definition mode is function definition, an identifier of a function granularity corresponding to the function of the AI ​​model and an identifier of the function of the AI ​​model in the function granularity are determined as the target additional information; or

[0132] If the definition mode is network element definition, the identifier of the network element corresponding to the AI ​​model is determined as the target additional information.

[0133] The following are the descriptions of the above-mentioned non-basic definition modes:

[0134] 1. Association model definition

[0135] As mentioned above, the above-mentioned associated model definition means that the AI ​​model that needs to be identified and defined is associated with other models.

[0136] In practice, with the rise of large models, the combination of network models has become a research hotspot. By utilizing the powerful generalization capabilities of large models and the customized service capabilities of small and medium models, terminals can obtain significantly improved service quality. Therefore, it is necessary to study the model identification definition in the model combination scenario. The large model is a regional large model generated based on a base station or a central cloud server. It is obtained by using a large amount of concentrated computing resources to execute the model training mechanism, and has a wide coverage; the small model is associated with the large model, has a smaller coverage, and is more customized. It can be obtained by simplifying the model structure parameters of the large model.

[0137] In this embodiment, with respect to the above-mentioned association model definition, the identifiers of other models associated with the AI ​​model are determined as target additional information.

[0138] Since the AI ​​model may be associated with multiple other models during implementation, and may be associated with a large model, a small model, or a combination of large and small models, the target additional information may include the identifiers of multiple associated other models, and different identifiers may be separated by special symbols.

[0139] For example, the identifier length of the large model is 4 bits, which can be generated in the core network and needs to be associated with the corresponding core network element, or generated by the operator based on the provincial and municipal base station network management system; the identifier length of the small model is 4 to 16 bits, which can be generated by the municipal and county or lower-level base station network management system, or obtained by splitting some modules from the large model, and used to perform network management, service quality optimization and other functions. The specific length can be set according to the actual application scenario.

[0140] 2. Network Slicing Definition

[0141] As mentioned above, the above network slice definition means that the AI ​​model that needs to be identified and defined is generated based on network slice resources.

[0142] During implementation, the network slicing service provided by operators can support flexible cutting and combination of network resources to meet the diverse needs of different industries and businesses.

[0143] In this embodiment, with respect to the above-mentioned network slice definition, the network slice identifier S-NSSAI of the AI ​​model and the identifier Slice Model ID of the AI ​​model in the network slice are determined as target additional information; S-NSSAI includes 8 bits of slice service type (Slice Service Type, SST) information and 24 bits of slice differentiator (SliceDifferentiator, SD) information.

[0144] For example, there may be more than one network slice in the network, and different slices are distinguished by the network slice identifier; the same network slice may generate multiple AI models, and different AI models in the same network slice are distinguished by the AI ​​model identifier in the network slice.

[0145] 3. Use case definition

[0146] As mentioned above, the above use case definition means that the AI ​​model that performs identification definition corresponds to a specific use case.

[0147] During implementation, since 6G wireless air interface scenarios support a large variety of use cases, different use cases have significantly different requirements for artificial intelligence model performance. Therefore, in some scenarios, it is necessary to define the AI ​​smart model ID based on specific use cases.

[0148] In this embodiment, with respect to the above use case definition, the identifier of the use case corresponding to the AI ​​model, the Use Case ID, and the identifier of the AI ​​model in the use case, the Use Case Model ID, are determined as target additional information.

[0149] Exemplarily, the Use Case ID is represented by 4 bits of information, and the Use Case Model ID is represented by 4 to 16 bits of information. The specific length is related to the number of use cases in the cell and user needs, and is also related to other conditions such as configuration, scenarios, sites, etc. This embodiment does not make specific limitations on this.

[0150] For example, there may be more than one use case in the network, and different use cases are distinguished by the identifier of the use case; the same use case may correspond to multiple AI models, and different AI models in the same use case are distinguished by the identifier of the AI ​​model in the use case.

[0151] In addition, the implementation levels of use cases include base station level and cell level. Base station level use cases represent regional multi-cell collaborative AI use cases, such as mobility management, switching control, etc. Cell level use cases represent single-cell AI use cases, such as beam management, channel state information (CSI) compression feedback, positioning enhancement, etc. Therefore, the target additional information may also include the implementation level of the use case to facilitate the management of the AI ​​model corresponding to the use case.

[0152] 4. Functional definition

[0153] As mentioned above, the above function definition means that the AI ​​model that performs identification definition corresponds to a specific function.

[0154] In the function-based lifecycle management mechanism defined by 3GPP, specific AI models are used to perform one or several AI functions, and these functions are associated with AI features.

[0155] In this embodiment, with respect to the above-mentioned function definition, FunctionalityID#, an identifier of the function granularity corresponding to the AI ​​model, and Functionality Model ID, an identifier of the function of the AI ​​model in the function granularity, are determined as target additional information.

[0156] Exemplarily, Functionality ID# is represented by 4 bits of information, and Functionality Model ID is represented by 4 to 8 bits of information. The specific length is related to the number of functions supported in the cell, UE capabilities and AI features, and is also related to other conditions such as configuration, scenarios, sites, etc. This embodiment does not make specific limitations on this.

[0157] In implementation, three functional granularities can be set: network-level Functionality ID#1, base station-level Functionality ID#2, and cell-level Functionality ID#3. Different functional granularities are used to adapt to various application scenarios. Network-level functions may include supervised learning, semi-supervised learning, and unsupervised learning; base station-level functions may include classification, regression, clustering, etc.; cell-level functions may include federated learning, reinforcement learning, etc.

[0158] 5. Network element definition

[0159] As described above, the above network element definition means that the AI ​​model for identification definition is generated and / or managed by a specific network element.

[0160] During implementation, since the lifecycle management of AI models involves collaboration between multiple network nodes, the correspondence between network elements and AI models needs to be clarified in some scenarios. The same network element may be responsible for the generation and management of different AI models, and different network elements may also participate in the generation and management of the same artificial intelligence model. In order to map network elements to artificial intelligence models, it is necessary to define a global model identifier based on network elements. The above-mentioned network elements can be core network elements, such as AMF / User Plane Function (UPF); or they can be access network elements, such as centralized units (CU) / distributed units (DU), which are not specifically limited in this embodiment.

[0161] In this embodiment, with respect to the above-mentioned network element definition, the identifier NF ID of the network element corresponding to the AI ​​model is determined as the target additional information.

[0162] Exemplarily, for core network elements, the NF ID includes one or more of the following: NFRegion ID that identifies the area covered by the element, which can be represented by 8 bits of information; NF Set ID that uniquely identifies the set of elements within the NF region, which can be represented by 10 bits of information; NF Pointer that identifies one or more elements in the set of elements, which can be represented by 6 bits of information. The combination of NF Set ID and NF Pointer ID is unique within a NF Region. For access network elements, NF ID can be identified using 4 bits of information.

[0163] In some optional implementations, combining the first model identifier with the target additional information to obtain the target model identifier includes:

[0164] The first model identifier and the target additional information are combined based on the position information corresponding to the definition mode to obtain the target model identifier.

[0165] In implementation, for non-basic modes, the first model identifier is the information that actually globally uniquely identifies the AI ​​model, and the target additional information is only for the convenience of management. Directly integrating the first model identifier with the target additional information will not only affect the global unique identification, but also make it difficult to reflect the target additional information required for management;

[0166] Based on this, this embodiment sets specific location information corresponding to each defined mode for non-basic modes, and combines the first model identifier and the target additional information based on the specific location information. The first model identifier and the target additional information correspond to different positions respectively. Subsequently, the first model identifier and the target additional information can be found from the target model identifier based on the specific location information.

[0167] See also Figure 4 As shown, the first model identifier and the target additional information can be directly combined. Figure 4 Taking the example of the target additional information being in front and the first model identifier being in the back, the location information can be set according to actual applications.

[0168] See also Figure 5 As shown, in order to more efficiently identify the first model identifier and the target additional information, the first model identifier and the target additional information may be separated by a first preset character. Figure 5 Taking the first preset character △ as an example, other symbols may be used as the first preset character in implementation.

[0169] Since the target additional information may contain multiple additional information, such as the identification of multiple other models associated with the AI ​​model in the associated model definition; the identification of the network slice of the AI ​​model and the identification of the AI ​​model in the network slice in the network slice definition, refer to Figure 6 As shown, multiple pieces of additional information in the target additional information may be separated by a second preset character. Figure 6 Taking the second preset character σ as an example, other symbols may be used as the second preset character in implementation.

[0170] In some optional implementations, the model identification configuration information also includes indication information corresponding to the definition mode.

[0171] Since the target model identifiers of the above-mentioned multiple definition modes are generated in different ways, it is necessary to make the relevant network nodes aware of not only the target model identifiers but also the definition modes;

[0172] Based on this, the model identification configuration information also includes indication information corresponding to the definition mode.

[0173] This embodiment does not specifically limit the indication information corresponding to the above definition mode, and the definition mode can be represented by different numbers of specific bits. When the definition mode includes the basic mode, association model definition, network slice definition, use case definition, function definition and network element definition, at least three bits are required for indication, and additional bits are required when there are other definition modes. As shown in Table 1, there are 6 bits in the model identification configuration information as indication information:

[0174] Table 1

[0175]

[0176]

[0177] Table 1 above is an exemplary description, and other methods may be used in implementation to indicate the generation location and define the mode.

[0178] Figure 7 A flowchart of a fourth model identification configuration method provided in an embodiment of the present application is shown as follows: Figure 7 As shown, the following steps are included:

[0179] Step S701: For any AI model, based on the generation location of the AI ​​model, determine the model identification information corresponding to the AI ​​model.

[0180] Step S702: Generate a target model identifier corresponding to the AI ​​model based on the definition mode corresponding to the AI ​​model, the mobile network identifier, the base station identifier, the cell identifier corresponding to the AI ​​model, and the model identifier information.

[0181] The specific implementation of steps S701 to S702 may refer to other embodiments and will not be described in detail here.

[0182] Step S703: Determine whether the target model identifier exceeds a preset number of digits.

[0183] In the implementation, since multiple pieces of information need to be added to the model recognition configuration information, taking the terminal side as an example, refer to Table 2:

[0184] Table 2

[0185] Fields Number of bits DCI Identifier 1 UE ID Not fixed Instructions It is related to the indication method. Table 1 takes 6 bits as an example. Target model identification Not fixed Trigger mode 1 Use cycle 4 Reserved bits Not fixed

[0186] Among them, the downlink control information (DCI) field, 1 represents downlink DCI;

[0187] UE ID is the identifier of the device corresponding to the generated position, and the corresponding bit depends on the definition method of UE ID;

[0188] The indication information includes the above-mentioned generation position and indication information corresponding to the definition mode, and the specific bits can be set according to the actual application scenario;

[0189] The target model identifier is the identifier of the above-mentioned globally unique AI model;

[0190] The trigger mode indicates whether the target model identification is stored based on a trigger, such as 1 for event triggering (UE starts storing the global model ID after receiving a confirmation message from the access base station); 0 for direct storage without event triggering.

[0191] The usage period indicates the number of time slots after which the model identification needs to be re-executed.

[0192] Since the model identification configuration information contains multiple contents, the number of bits left for the target model identifier is limited. If the target model identifier is too long, the model identification configuration information may not be able to carry the complete target model identifier;

[0193] Based on this, this embodiment sets a preset number of bits, which is determined based on the reserved position in the model identification configuration information; by judging whether the target model identifier exceeds the preset number of bits, it is determined whether the model identification configuration information can carry a complete target model identifier.

[0194] Step S704: if the target model identifier does not exceed a preset number of digits, generating model recognition configuration information including indication information corresponding to the generation position and the target model identifier.

[0195] In implementation, if the target model identifier does not exceed the preset number of digits, it means that the model identification configuration information can carry the complete target model identifier;

[0196] Based on this, model recognition configuration information including indication information corresponding to the generation position and the target model identifier is directly generated.

[0197] Step S705: if the target model identifier exceeds the preset number of digits, the target model identifier is divided into a plurality of sub-identifiers; for any sub-identifier, the model recognition configuration information including the indication information corresponding to the generation position and the sub-identifier is generated.

[0198] In implementation, if the target model identifier exceeds the preset number of digits, it means that the model identification configuration information is difficult to carry the complete target model identifier;

[0199] Based on this, the target model identifier is divided into multiple sub-identifiers, each of which does not exceed the preset number of digits; and model identification configuration information including indication information corresponding to the generation position and the sub-identifier is generated. In other words, the target model identifier is carried by multiple model identification configuration information.

[0200] Step S706: Send the model recognition configuration information to the device corresponding to the generation location

[0201] The specific implementation of step S706 may refer to other embodiments and will not be described in detail here.

[0202] like Figure 8 As shown, the embodiment of the present application provides a model identification configuration device 800, which includes:

[0203] An information determination module 801 is used to determine, for any AI model, model identification information corresponding to the AI ​​model based on the generation location of the AI ​​model;

[0204] An identification determination module 802 is used to generate a target model identification corresponding to the AI ​​model based on a definition mode corresponding to the AI ​​model, a mobile network identification, a base station identification, a cell identification and the model identification information corresponding to the AI ​​model; wherein the definition mode corresponding to the AI ​​model includes a basic mode and a non-basic mode;

[0205] The identification configuration module 803 is used to generate model identification configuration information based on the indication information corresponding to the generation position and the target model identifier, and send the model identification configuration information to the device corresponding to the generation position.

[0206] In some optional implementation manners, the information determination module 801 is specifically configured to:

[0207] If the generation location of the AI ​​model is the network side, determining that the model identification information corresponding to the AI ​​model is an identification of the AI ​​model in the corresponding cell; or

[0208] If the generation location of the AI ​​model is the terminal side, the model identification information corresponding to the AI ​​model is determined to be the terminal identification corresponding to the AI ​​model and the identification of the AI ​​model in the corresponding terminal.

[0209] In some optional implementation manners, the identification determination module 802 is specifically configured to:

[0210] If the definition mode corresponding to the AI ​​model is the basic mode, a preset calculation method is used to calculate the mobile network identifier, the base station identifier, the cell identifier and the model identifier information to obtain the target model identifier.

[0211] In some optional implementation manners, the identification determination module 802 is specifically configured to:

[0212] If the definition mode corresponding to the AI ​​model is a non-basic mode, the mobile network identifier, the base station identifier, the cell identifier, and the model identifier information are calculated using a preset calculation method to obtain a first model identifier corresponding to the AI ​​model;

[0213] The first model identifier is combined with target additional information to obtain the target model identifier; wherein the target additional information is determined based on the definition mode.

[0214] In some optional implementations, the non-basic mode includes association model definition, network slice definition, use case definition, function definition, and network element definition.

[0215] In some optional implementation manners, the identification determination module 802 is further configured to determine the target additional information in the following manner:

[0216] If the definition mode is an associated model definition, the identifiers of other models associated with the AI ​​model are determined as the target additional information; or

[0217] If the definition mode is a network slice definition, an identifier of the network slice for generating the AI ​​model and an identifier of the AI ​​model in the network slice are determined as the target additional information; or

[0218] If the definition mode is a use case definition, an identifier of the use case corresponding to the AI ​​model and an identifier of the AI ​​model in the use case are determined as the target additional information; or

[0219] If the definition mode is function definition, an identifier of a function granularity corresponding to the function of the AI ​​model and an identifier of the function of the AI ​​model in the function granularity are determined as the target additional information; or

[0220] If the definition mode is network element definition, the identifier of the network element corresponding to the AI ​​model is determined as the target additional information.

[0221] In some optional implementation manners, the identification determination module 802 is specifically configured to:

[0222] The first model identifier and the target additional information are combined based on the position information corresponding to the definition mode to obtain the target model identifier.

[0223] In some optional implementations, the model identification configuration information also includes indication information corresponding to the definition mode.

[0224] In some optional implementations, the identification configuration module 803 is used to:

[0225] If the target model identifier does not exceed a preset number of digits, generating model identification configuration information including indication information corresponding to the generation position and the target model identifier; or

[0226] If the target model identifier exceeds the preset number of digits, the target model identifier is divided into a plurality of sub-identifiers; and for any sub-identifier, model recognition configuration information including indication information corresponding to the generation position and the sub-identifier is generated.

[0227] Since the device is the device in the method in the embodiment of the present application, and the principle of solving the problem by the device is similar to that of the method, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0228] Based on the same technical concept, the embodiment of the present application also provides a network device 900, such as Fig. 9 As shown, it includes at least one processor 901 and a memory 902 connected to the at least one processor. The specific connection medium between the processor 901 and the memory 902 is not limited in the embodiment of the present application. Fig. 9 For example, the processor 901 and the memory 902 are connected via a bus 903. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig. 9 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0229] Among them, the processor 901 is the control center of the network device, which can use various interfaces and lines to connect various parts of the network device, and realize data processing by running or executing instructions stored in the memory 902 and calling data stored in the memory 902. Optionally, the processor 901 may include one or more processing units, and the processor 901 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application program, etc., and the modem processor mainly processes the issuance of instructions. It is understandable that the above-mentioned modem processor may not be integrated into the processor 901. In some embodiments, the processor 901 and the memory 902 may be implemented on the same chip, and in some embodiments, they may also be implemented separately on independent chips.

[0230] Processor 901 can be a general-purpose processor, such as a CPU, a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, and can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application. A general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the method disclosed in the embodiment of the model identification configuration method can be directly embodied as a hardware processor execution, or a combination of hardware and software modules in the processor.

[0231] The memory 902 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules. The memory 902 may include at least one type of storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory, a random access memory (Random Access Memory, RAM), a static random access memory (Static Random Access Memory, SRAM), a programmable read-only memory (Programmable Read Only Memory, PROM), a read-only memory (Read Only Memory, ROM), an electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), a magnetic memory, a disk, an optical disk, etc. The memory 902 is any other medium that can be used to carry or store a desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 902 in the embodiment of the present application can also be a circuit or any other device that can realize a storage function, for storing program instructions and / or data.

[0232] In the embodiment of the present application, the memory 902 stores a computer program. When the program is executed by the processor 901, the processor 901 executes:

[0233] For any AI model, determine the model identification information corresponding to the AI ​​model based on the generation location of the AI ​​model;

[0234] Generate a target model identifier corresponding to the AI ​​model based on a definition mode corresponding to the AI ​​model, a mobile network identifier, a base station identifier, a cell identifier, and the model identifier information corresponding to the AI ​​model; wherein the definition mode corresponding to the AI ​​model includes a basic mode and a non-basic mode;

[0235] Model recognition configuration information is generated based on the indication information corresponding to the generation location and the target model identifier, and the model recognition configuration information is sent to the device corresponding to the generation location.

[0236] In some optional implementations, the processor 901 specifically performs:

[0237] If the generation location of the AI ​​model is the network side, determining that the model identification information corresponding to the AI ​​model is an identification of the AI ​​model in the corresponding cell; or

[0238] If the generation location of the AI ​​model is the terminal side, the model identification information corresponding to the AI ​​model is determined to be the terminal identification corresponding to the AI ​​model and the identification of the AI ​​model in the corresponding terminal.

[0239] In some optional implementations, the processor 901 specifically performs:

[0240] If the definition mode corresponding to the AI ​​model is the basic mode, a preset calculation method is used to calculate the mobile network identifier, the base station identifier, the cell identifier and the model identifier information to obtain the target model identifier.

[0241] In some optional implementations, the processor 901 specifically performs:

[0242] If the definition mode corresponding to the AI ​​model is a non-basic mode, the mobile network identifier, the base station identifier, the cell identifier, and the model identifier information are calculated using a preset calculation method to obtain a first model identifier corresponding to the AI ​​model;

[0243] The first model identifier is combined with target additional information to obtain the target model identifier; wherein the target additional information is determined based on the definition mode.

[0244] In some optional implementations, the non-basic mode includes association model definition, network slice definition, use case definition, function definition, and network element definition.

[0245] In some optional implementations, the processor 901 further executes:

[0246] If the definition mode is an associated model definition, the identifiers of other models associated with the AI ​​model are determined as the target additional information; or

[0247] If the definition mode is a network slice definition, an identifier of the network slice for generating the AI ​​model and an identifier of the AI ​​model in the network slice are determined as the target additional information; or

[0248] If the definition mode is a use case definition, an identifier of the use case corresponding to the AI ​​model and an identifier of the AI ​​model in the use case are determined as the target additional information; or

[0249] If the definition mode is function definition, an identifier of a function granularity corresponding to the function of the AI ​​model and an identifier of the function of the AI ​​model in the function granularity are determined as the target additional information; or

[0250] If the definition mode is network element definition, the identifier of the network element corresponding to the AI ​​model is determined as the target additional information.

[0251] In some optional implementations, the processor 901 specifically performs:

[0252] The first model identifier and the target additional information are combined based on the position information corresponding to the definition mode to obtain the target model identifier.

[0253] In some optional implementations, the model identification configuration information also includes indication information corresponding to the definition mode.

[0254] In some optional implementations, the processor 901 specifically performs:

[0255] If the target model identifier does not exceed a preset number of digits, generating model identification configuration information including indication information corresponding to the generation position and the target model identifier; or

[0256] If the target model identifier exceeds the preset number of digits, the target model identifier is divided into a plurality of sub-identifiers; and for any sub-identifier, model recognition configuration information including indication information corresponding to the generation position and the sub-identifier is generated.

[0257] Since the network device is the network device in the method in the embodiment of the present application, and the principle of solving the problem by the network device is similar to that of the method, the implementation of the network device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0258] Based on the same technical concept, an embodiment of the present application also provides a computer-readable storage medium, which stores a computer program executable by a processor. When the program runs on the processor, the processor executes the steps of the above-mentioned model identification configuration method.

[0259] In some optional embodiments, various aspects of the model identification configuration method provided in the present application can also be implemented in the form of a program product, which contains computer-executable instructions. When the program product is run on a computer device, the computer-executable instructions are used to enable the computer device to execute the steps of the model identification configuration method according to various exemplary embodiments of the present application described above in this specification.

[0260] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0261] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0262] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0263] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0264] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0265] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A model identification configuration method, characterized in that: The method includes: For any AI model, determine the model identification information corresponding to the AI ​​model based on the generation location of the AI ​​model; Generate a target model identifier corresponding to the AI ​​model based on a definition mode corresponding to the AI ​​model, a mobile network identifier, a base station identifier, a cell identifier, and the model identifier information corresponding to the AI ​​model; wherein the definition mode corresponding to the AI ​​model includes a basic mode and a non-basic mode; Model recognition configuration information is generated based on the indication information corresponding to the generation location and the target model identifier, and the model recognition configuration information is sent to the device corresponding to the generation location.

2. The method according to claim 1, characterized in that Based on the generation location of the AI ​​model, determining model identification information corresponding to the AI ​​model includes: If the generation location of the AI ​​model is the network side, determining that the model identification information corresponding to the AI ​​model is an identification of the AI ​​model in the corresponding cell; or If the generation location of the AI ​​model is the terminal side, the model identification information corresponding to the AI ​​model is determined to be the terminal identification corresponding to the AI ​​model and the identification of the AI ​​model in the corresponding terminal.

3. The method according to claim 1, characterized in that Based on the definition mode corresponding to the AI ​​model, the mobile network identifier, the base station identifier, the cell identifier and the model identifier information corresponding to the AI ​​model, generating a target model identifier corresponding to the AI ​​model, including: If the definition mode corresponding to the AI ​​model is the basic mode, a preset calculation method is used to calculate the mobile network identifier, the base station identifier, the cell identifier and the model identifier information to obtain the target model identifier.

4. The method according to claim 1, characterized in that Based on the definition mode corresponding to the AI ​​model, the mobile network identifier, the base station identifier, the cell identifier and the model identifier information corresponding to the AI ​​model, generating a target model identifier corresponding to the AI ​​model, including: If the definition mode corresponding to the AI ​​model is a non-basic mode, the mobile network identifier, the base station identifier, the cell identifier, and the model identifier information are calculated using a preset calculation method to obtain a first model identifier corresponding to the AI ​​model; The first model identifier is combined with target additional information to obtain the target model identifier; wherein the target additional information is determined based on the definition mode.

5. The method according to claim 4, characterized in that The non-basic mode includes association model definition, network slice definition, use case definition, function definition and network element definition.

6. The method according to claim 5, characterized in that The target additional information is determined by: If the definition mode is an associated model definition, the identifiers of other models associated with the AI ​​model are determined as the target additional information; or If the definition mode is a network slice definition, an identifier of the network slice for generating the AI ​​model and an identifier of the AI ​​model in the network slice are determined as the target additional information; or If the definition mode is a use case definition, an identifier of the use case corresponding to the AI ​​model and an identifier of the AI ​​model in the use case are determined as the target additional information; or If the definition mode is function definition, an identifier of a function granularity corresponding to the function of the AI ​​model and an identifier of the function of the AI ​​model in the function granularity are determined as the target additional information; or If the definition mode is network element definition, the identifier of the network element corresponding to the AI ​​model is determined as the target additional information.

7. The method according to claim 4, characterized in that Combining the first model identifier with the target additional information to obtain the target model identifier includes: The first model identifier and the target additional information are combined based on the position information corresponding to the definition mode to obtain the target model identifier.

8. The method according to any one of claims 3 to 7, characterized in that: The model identification configuration information also includes indication information corresponding to the definition mode.

9. The method according to claim 1, characterized in that Generate model recognition configuration information based on the indication information corresponding to the generation position and the target model identifier, including: If the target model identifier does not exceed a preset number of digits, generating model identification configuration information including indication information corresponding to the generation position and the target model identifier; or If the target model identifier exceeds the preset number of digits, the target model identifier is divided into a plurality of sub-identifiers; and for any sub-identifier, model recognition configuration information including indication information corresponding to the generation position and the sub-identifier is generated.

10. A model identification configuration device, characterized in that: The device includes: An information determination module, configured to determine, for any AI model, model identification information corresponding to the AI ​​model based on a generation location of the AI ​​model; An identification determination module, used to generate a target model identification corresponding to the AI ​​model based on a definition mode corresponding to the AI ​​model, a mobile network identification, a base station identification, a cell identification, and the model identification information corresponding to the AI ​​model; wherein the definition mode corresponding to the AI ​​model includes a basic mode and a non-basic mode; The identification configuration module is used to generate model identification configuration information based on the indication information corresponding to the generation position and the target model identification, and send the model identification configuration information to the device corresponding to the generation position.

11. A network device, characterized in that: The method comprises at least one processor and at least one memory, wherein the memory stores a computer program, and when the program is executed by the processor, the processor executes the method according to any one of claims 1 to 9.

Images