Model learning method and apparatus, and network-side device

Abstract

This application discloses a model learning method, apparatus, and network-side device, belonging to the field of communication technology. The method in this application includes: a first network element receiving a first message from a second network element, the first message being used to indicate first information, the first information being used for online learning of a first model; the first network element performing online learning of the first model based on the first information.

Description

Technical Field

[0001] This application belongs to the field of communication technology, and in particular relates to a model learning method, apparatus and network-side equipment. Background Technology

[0002] In related communication technologies, the network element responsible for training the Artificial Intelligence (AI) model in the core network usually collects the model training data in advance, performs offline training of the model, and finally provides the trained model to the model inference network element for model inference, etc.

[0003] However, since the data in communication networks is dynamically changing or generated in real time, there may be a problem that pre-trained models cannot adapt to the characteristics of the data in communication networks. Summary of the Invention

[0004] The purpose of this application is to provide a model learning method, apparatus, and network-side device that enables the model to better conform to the data characteristics of the communication network.

[0005] In a first aspect, a model learning method is provided, comprising: a first network element receiving a first message from a second network element, the first message being used to indicate first information, the first information being used for online learning of a first model; and the first network element performing online learning of the first model based on the first information.

[0006] Secondly, a model learning method is provided, comprising: a second network element sending a first message to a first network element, the first message being used to indicate first information, the first information being used for online learning of a first model.

[0007] Thirdly, a model learning device is provided, comprising: a transmission module for receiving a first message from a second network element, the first message indicating first information, the first information being used for online learning of a first model; and a processing module for performing online learning of the first model based on the first information.

[0008] Fourthly, a model learning device is provided, comprising: a transmission module for sending a first message to a first network element, the first message being used to indicate first information, the first information being used for online learning of a first model.

[0009] Fifthly, embodiments of this application provide a network-side device, including: a memory, a processor, and computer-executable instructions stored in the memory and executable on the processor, wherein the computer-executable instructions, when executed by the processor, implement the steps of the method described in the first aspect or the second aspect. In a sixth aspect, embodiments of this application provide a computer-readable storage medium for storing computer-executable instructions that, when executed by a processor, implement the steps of the method described in the first or second aspect. In this embodiment of the application, the second network element provides the first network element with the first information, so that the first network element can realize the online learning of the first model based on the first information, thereby enabling the first model to better conform to the data characteristics of the communication network and ensuring the model inference performance. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic diagram of the communication network architecture provided in an exemplary embodiment of this application.

[0012] Figure 2 This is one of the flowcharts illustrating the model learning method provided in an exemplary embodiment of this application.

[0013] Figure 3 This is a second schematic flowchart of the model learning method provided in an exemplary embodiment of this application.

[0014] Figure 4 This is the third flowchart illustrating the model learning method provided in an exemplary embodiment of this application.

[0015] Figure 5a This is one of the interactive flow diagrams of the model learning method provided in an exemplary embodiment of this application.

[0016] Figure 5b This is the second schematic diagram of the interaction flow of the model learning method provided in an exemplary embodiment of this application.

[0017] Figure 5c This is the third schematic diagram of the interaction flow of the model learning method provided in an exemplary embodiment of this application.

[0018] Figure 6 This is the fourth flowchart illustrating the model learning method provided in an exemplary embodiment of this application.

[0019] Figure 7 This is one of the structural schematic diagrams of a model learning device provided in an exemplary embodiment of this application.

[0020] Figure 8 This is a second schematic diagram of the structure of the model learning device provided in an exemplary embodiment of this application.

[0021] Figure 9 This is a schematic diagram of the structure of a network-side device provided in an exemplary embodiment of this application. Detailed Implementation

[0022] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0023] The technical solutions provided in this application will be described in detail below with reference to the accompanying drawings and through some embodiments and application scenarios.

[0024] like Figure 1 The diagram illustrates a communication network architecture provided in an exemplary embodiment of this application. This communication network architecture may include user equipment (UE, also known as a terminal), radio access network (RAN), core network, etc.

[0025] The RAN manages radio resources, transmits user data received via N3 to the UE, and retrieves user data from the UE through the N3 interface. The RAN maps between Quality of Service (QoS) traffic in Dedicated Radio Bearer (DRB) and Protocol Data Unit (PDU) sessions.

[0026] The core network may include, but is not limited to, core network elements such as Access and Mobility Management Function (AMF), Session Management Function (SMF), Advanced User Plane Function (A-UPF), Policy Control Function (PCF), and Data Network (DN).

[0027] The AMF can be used to implement, but is not limited to, functions such as registration management, connection management, reachability management, and mobility management. Furthermore, the AMF can also be used to perform access authentication and access authorization. The AMF is a Non-Access Stratum (NAS) security endpoint that forwards SM NAS messages between the UE and the SMF.

[0028] SMF can be used to implement, but is not limited to, session establishment, modification and release, UE IP address allocation and management (including optional authorization functions), selection and control of User Plane (UP) functions, downlink data notification and other functions.

[0029] The SMF controls the UPF through N4 association. For example, the SMF can provide the UPF with Packet Detection Rules (PDR), Forwarding Action Rules (FAR), QoS Enforcement Rules (QER), and Usage Reporting Rules (URR). Among them, FAR is used to indicate how to detect user data traffic; QER and URR are used to instruct the UPF how to perform user data traffic forwarding, QoS processing, and usage reporting on user data traffic detected using PDR.

[0030] The UPF can be used to implement, but is not limited to, functions including, serving as an anchor point for intra / inter-Radio Access Technology (RAT) mobility, packet routing and forwarding, traffic usage reporting, user plane QoS processing, downlink packet buffering, and downlink data notification triggering. The GTP-U tunnel is used for the N3 interface between the RAN and the UPF. The GTP-U tunnel operates on a per-PDU session basis. For downlink traffic, the UPF binds the downlink traffic to the QoS traffic within the PDU session's GTP-U tunnel using the FAR received from the SMF. For uplink traffic, the RAN transmits user plane traffic to the QoS stream identified by the UE.

[0031] The PCF is used to provide QoS policy rules to control plane functions for enforcement. The PCF translates Application Function (AF) requests into Policy Control and Charging (PCC) rules applicable to PDU sessions.

[0032] DN can include, but is not limited to, Unified Data Management (UDM). UDM performs tasks such as generating 3GPP AKA authentication credentials, granting access based on subscription data, managing UE service element registration (e.g., providing AMF storage for UEs, SMF storage for UE PDU sessions), and subscription management. UDM accesses the Unified Data Repository (UDR) to retrieve UE subscription data and store the UE context in the UDR. UDM and UDR can be deployed together.

[0033] In addition to the aforementioned core network elements, the core network may also include a Network Data Analytics Function (NWDAF). The NWDAF is a 5G core network (5GC) element located in the control plane, performing statistical data and machine learning-related tasks within the 5G system (5GS). The NWDAF can interact with different entities for various purposes. 1) The NWDAF can collect data based on event subscriptions provided by AMF, SMF, UPF, PCF, UDM, Network Slice Admission Control Function (NSACF), AF, and OAM. The NWDAF can interact directly with the AF or through the Network Exposure Function (NEF).

[0034] Optionally, NWDAF uses the Data Collection Coordination Function (DCCF) for analysis and data collection. 2) NWDAF retrieves information from the data repository, such as retrieving user-related UDRs via UDM or retrieving PFD information via NEF (or Packet Flow Description Function, PFDF) etc. 3) NWDAF collects location information data from the Location Services (LCS) system.

[0035] Optionally, NWDAF stores and retrieves information from the Analytic Data Repository Function (ADRF). Optionally, NWDAF analyzes and collects data from the Messaging Framework Adaptor Function (MFAF). 4) NWDAF retrieves information about network elements, for example, retrieving network element-related information from the Network Repository Function (NRF). 5) NWDAF provides analytics to consumers on demand.

[0036] 6) NWDAF provides batch data related to the analysis ID.

[0037] 7) NWDAF provides information on the accuracy of the analysis ID.

[0038] 8) NWDAF provides information on the accuracy of the AI ​​model or indications of model accuracy degradation. In this embodiment, the AI ​​model may also be referred to as a machine learning (ML) model, AI unit, neural network, etc., without limitation.

[0039] For the deployment of the aforementioned NWDAF, a single instance or multiple instances of the NWDAF can be deployed in a Public Land Mobile Network (PLMN). The NWDAF may include logical functions such as the Analytics Logical Function (AnLF) and the Model Training Logical Function (MTL·F). AnLF, as a logical function in NWDAF, is used to perform inference and derive analytical information, that is, to derive statistical data and / or predictions based on consumer requests for analysis, and to expose analytical services, namely Nnwdaf_AnalyticsSubscription or Nnwdaf_AnalyticsI Request.

[0040] MTLF is used to train AI models and expose new training services, such as providing pre-trained AI models.

[0041] An NWDAF can contain one MTLF or one AnLF, or both MTLF and AnLF; there is no restriction on this.

[0042] Based on the aforementioned communication network architecture, in related technologies, the network element responsible for Artificial Intelligence (AI) model training in the core network (such as NWDAF deploying MTLF) usually collects model training data in advance and performs offline training of the model. Finally, the trained model is provided to the model inference network element (such as NWDAF deploying AnLF) for model inference.

[0043] However, since data in communication networks is dynamically changing or generated in real time, pre-trained models may not be able to adapt to the characteristics of the communication network data. To address this, this application provides a model learning scheme that trains or updates the model through online learning, enabling the model to better conform to the data characteristics of the communication network and improving model inference performance. Optionally, the technical solutions provided in this embodiment can be applied to, but are not limited to, the aforementioned methods. Figure 1 The communication network architecture shown, such as the technical solution provided in this embodiment of the application, can also be applied to 4G communication network architecture, 6G communication network architecture, or other future communication network architectures, etc., without limitation.

[0044] Based on this, the technical solutions provided by the embodiments of this application will be described below with reference to the accompanying drawings and some examples.

[0045] Figure 2 This illustration shows a flowchart of a model learning method 200 provided in an embodiment of this application. This method 200 can be executed by a first network element. Figure 2 As shown, the method 200 may include the following steps.

[0046] S210, the first network element receives the first message from the second network element.

[0047] S220, the first network element performs online learning of the first model based on the first information.

[0048] The first network element is used to provide model training services, etc., that is, the first network element may include, but is not limited to, model training functions. For example, the first network element may be NWDAF, which deploys MTLF in the aforementioned communication network architecture.

[0049] The second network element is used to provide model inference services, etc. That is, the second network element may include, but is not limited to, model inference functions. For example, the second network element may be, but is not limited to, NWDAF deployed in the aforementioned communication network architecture.

[0050] Optionally, the first network element and the second network element can be independently deployed core network elements, access network elements, etc., that can exchange information through the core network interface. For example, the first network element can be NWDAF 1 deploying MTLF, and the second network element can be NWDAF 2 deploying AnLF. NWDAF 1 and NWDAF 2 are different, and this embodiment does not impose any restrictions on them.

[0051] The first message is used to indicate first information, which is used for online learning of the first model, such as training and updating the first model. In other words, compared to offline model training methods in related technologies, in this embodiment, the second network element provides the first network element with real-time data generated in the communication network via the first message. This assists the first network element in performing online learning of the first model based on the first information in the first message, enabling the first model to better conform to the data characteristics of the communication network and improving model inference performance.

[0052] Optionally, the aforementioned first model may be a model trained offline by the first network element and provided to the second network element, or it may be a model that the first network element is currently undergoing initial training, or it may be a model trained offline but provided to other network elements besides the second network element. No restrictions are placed here.

[0053] For example, in the case where the first model is a model trained offline by the first network element and provided to the second network element, the process of the first model providing the first model to the second network element may include, but is not limited to: the second network element sending a model request message to the first network element, the model request message being used to indicate the relevant information of the first model requested by the second network element; the first network element sending a model request response message to the second network element according to the model request message, the model request response message being used to indicate the first model to the second network element.

[0054] Optionally, the model request message may include, but is not limited to, at least one of the following: a model identifier or analysis identifier of the requested first model, an indication of whether the requested first model is a trained model, and an indication of whether the requested first model supports online learning.

[0055] Optionally, the model request response message may include, but is not limited to, at least one of the following: the model file corresponding to the first model, the model access address, an indication of whether the model has completed training, an indication of whether online learning is supported, a model identifier, and an analysis identifier.

[0056] In some embodiments, the first information used for online learning of the first model may include, but is not limited to, at least one of the following 10)-19).

[0057] 10) The model inference results of the first model.

[0058] The model inference results can also be referred to as the analysis information, prediction information, etc., obtained by the first model.

[0059] 11) Truth data, which corresponds to the model inference results of the first model.

[0060] The true data mentioned above can also be called calibration data or real data.

[0061] In this embodiment, by collecting the model inference results of the first model and the truth data for online learning of the first model, the first network element can use the truth data as a target to perform online learning of the first model, thereby improving the model inference performance, such as making the model inference results closer to the truth data.

[0062] 12) Model inference data, wherein the model inference result is obtained by inferring the model inference data through the first model.

[0063] 13) The first time is the time when the model inference data is collected.

[0064] 14) Second time, the second time is the time when the true value data is collected.

[0065] 15) The third time is the application time of the model inference results.

[0066] 16) Model inference feedback information, which is determined based on the model inference result.

[0067] Optionally, the model inference feedback information may include, but is not limited to, model inference quality information, model inference speed information, etc., for the first network element to determine whether to perform online learning on the first model.

[0068] For example, if the model inference quality information or the model inference speed does not meet the standard, it is determined that the first model needs to be trained online to improve the inference performance of the first model.

[0069] 17) First indication information, used to indicate the chronological relationship between the time of collection of the truth data and the time of application of the model inference results.

[0070] 18) First loss information, which is used to represent the loss determined based on the model inference result and the truth data.

[0071] The first loss information is used to calculate the gradient of the first model, and the gradient calculation result enables the online learning of the first model.

[0072] 19) Second loss information, which is used to represent the cumulative loss determined based on the model inference results and truth data over a predetermined time period.

[0073] The first loss information is used to calculate the gradient of the first model, and the gradient calculation result enables the online learning of the first model.

[0074] Optionally, if the first information does not include the first loss information and / or the second loss information, then when the first network element performs online learning of the first model based on the first information, it can perform the first loss information and / or the second loss information based on the model inference results, ground truth data, etc. included in the first information, for use in online learning of the first model, such as gradient calculation.

[0075] Optionally, the information in 10)-19) above can be collected by the second network element and provided to the first network element in real time or periodically for online learning of the first model.

[0076] In some embodiments, the first information may be determined by the second network element based on network data it generates, or it may be determined by the second network element based on network data provided by the consumer network element; no limitation is imposed here. The consumer network element may include, but is not limited to, PCF, Location Management Function (LMF), etc.

[0077] For example, if the first information is sent by the second network element based on network data provided by the consumer network element, then in this embodiment, the second network element can receive a sixth message from the consumer network element; wherein, the sixth message is used to indicate information related to the model inference result of the first model, and the first message is determined based on the sixth message.

[0078] Optionally, the sixth message (or information related to the model inference result) may include, but is not limited to, at least one of the following 21)-25).

[0079] 21) Truth data, which corresponds to the model inference results.

[0080] Optionally, the truth data in the first message may include part or all of the truth data in the sixth message, without limitation.

[0081] 22) The second time, the time of acquisition of the true value data.

[0082] 23) The third time is the application time of the model inference results.

[0083] 24) Model inference feedback information, which is determined based on the model inference result.

[0084] Optionally, the model inference feedback information may include, but is not limited to, model inference quality information, model inference speed information, etc., for the first network element to determine whether to perform online learning on the first model.

[0085] 25) First indication information, used to indicate the chronological relationship between the time of collection of the truth data and the time of application of the model inference results.

[0086] In some embodiments, the sixth message sent by the consumer network element to the second network element may be sent by the consumer network element on demand, or it may be sent according to a request from the second network element. For example, in the case where the message is sent according to a request from the second network element, the second network element may, when providing the consumer network element with the model inference result of the first model, instruct or request the consumer network element to provide relevant information about the model inference result, such as at least one of 21)-24) above, for the purpose of collecting real-time network data for the first model and performing online learning of the first model.

[0087] In some embodiments, the method embodiment 200 of this application may further include Figure 3 The S230 and / or S240 shown are used to determine the second network element participating in the online learning of the first model, ensuring the smooth execution of the online learning of the first model.

[0088] S230, the first network element determines the second network element that supports online learning of the model.

[0089] Optionally, the first network element can determine the second network element that supports online model learning in various ways. For example, the second network element can be determined by local query or by querying a third network element (such as NRF). No restrictions are imposed here.

[0090] S240, the first network element requests the second network element to join the online learning of the model.

[0091] Wherein, if the method embodiment 200 of this application includes Figure 3 S230 shown, but excluding S240, can be understood as follows: the first network element needs to independently determine the second network element that supports the online learning of the first model, so as to ensure that the first information provided by the second network element can be accurately used for the online learning of the first model.

[0092] If the method embodiment 200 of this application includes Figure 3 S240 shown, but excluding S230, can be understood as follows: through pre-configuration or protocol agreement, the first network element knows that the second network element supports online model learning, but needs to request the second network element to join the online model learning through the execution of S240 to ensure the smooth execution of the online learning of the first model.

[0093] If the method embodiment 200 of this application may further include Figure 3 S240 and S240 shown can be understood as follows: When the first network element determines that it needs to perform online learning of the first model, it needs to independently determine the second network element that supports the online learning of the first model, and request the second network element to join the online learning of the model to ensure the smooth execution of the online learning of the first model.

[0094] In some embodiments, the process described in S240 of the first network element requesting the second network element to join the online learning of the model may include, but is not limited to, the following: Figure 4 The contents of S2401-S2402 shown are as follows.

[0095] S2401, the first network element sends a second message to the second network element.

[0096] The second message is used to request the second network element to join the online learning of the model. Optionally, the second message may include, but is not limited to, at least one of the following 31)-34).

[0097] 31) Model files that support online learning.

[0098] 32) The model access address for model files that support online model learning.

[0099] 33) Model identifiers for model files that support online model learning.

[0100] 34) Model analysis identifiers for model files that support online model learning.

[0101] In this embodiment, through the instructions in 31)-34) above, the second network element can clearly understand what model needs to be learned online, and then determine whether to agree to join the online model learning.

[0102] In some embodiments, if the first network element determines that it has provided the second network element with a first model that supports reinforcement learning or a first model that has not completed training before sending the second message, then the second message may include the model identifier or model analysis identifier, so as to indicate to the second network element which model to use for online model learning and reduce the instruction overhead.

[0103] S2402, the first network element receives a third message from the second network element.

[0104] The third message is used to indicate whether one agrees to join the online learning of the model. For example, a bit "1" can indicate agreement and "0" can indicate disagreement, and there are no restrictions on this.

[0105] In some embodiments, when the third message is used to indicate disagreement to join the online learning of the model, the third message may include reason information for disagreement to join the online learning of the model, so that the first network element can determine whether to re-request the second network element to join the online learning process of the first model based on the reason information.

[0106] Optionally, the reason information may include, but is not limited to: network element busy / overloaded, no model supporting online learning, lack of ability to support online learning, etc.

[0107] In some embodiments, the method embodiment 200 of this application may further include: the first network element sending a fourth message to a third network element (such as NRF); wherein the fourth message is used to register or update the relevant capabilities of online model learning with the third network element, so that other network elements with model training functions can query network elements that support online model learning from the third network element as needed.

[0108] In this embodiment, the fourth message may include, but is not limited to, at least one of the following 41)-43).

[0109] 41) Second instruction information, used to indicate whether it is supported as a member of online learning.

[0110] 42) The third instruction information is used to indicate whether online learning is supported.

[0111] 43) The fourth instruction information is used to indicate whether the collection of data required for online learning is supported.

[0112] Similarly, in some embodiments, the second network element may also send a seventh message to a third network element (such as NRF) for the first network element to query, on demand, network elements that support online model learning from the third network element. The seventh message may include, but is not limited to, at least one of the following 51)-53).

[0113] 51) The fifth instruction information is used to indicate whether it supports being a member of the model's online learning.

[0114] 52) The sixth instruction information is used to indicate whether online learning of the model is supported.

[0115] 53) The seventh instruction information is used to indicate whether the collection of data required for online model learning is supported.

[0116] In some embodiments, the method embodiment 200 of this application may further include: the first network element sending a fifth message to the second network element to indicate model-related information after online learning of the first model, so that the second network element can perform model inference based on the model updated through online learning, thereby ensuring model inference performance. The fifth message may include, but is not limited to, at least one of the following 61)-65).

[0117] 61) Second model.

[0118] The second model is obtained by updating the first model through online learning.

[0119] 62) At least some of the model parameters of the second model, such as model parameters in the second model that are different from those in the first model.

[0120] 63) Model identifier of the second model.

[0121] 64) The model access address of the second model, for the second network element to obtain the second model from the model access address.

[0122] 65) Model analysis identifier for the second model.

[0123] Based on the description of the aforementioned method embodiment 200, for ease of understanding, the model learning scheme provided in this application will be illustrated below with reference to Examples 1-3, as follows.

[0124] Example 1: Capability registration or update like Figure 5a As shown, the capability registration or update process of the first network element (such as an NWDAF including an MTLF) may include, but is not limited to, the following steps.

[0125] S511, the first network element sends a fourth message to the third network element (such as NRF), wherein the fourth message is used to register or update the relevant capabilities of the model's online learning with the third network element.

[0126] S512, the third network element stores the online learning capabilities of the first network element's model according to the fourth message.

[0127] S513, the third network element returns the registration or update results of the model's online learning capabilities to the first network element.

[0128] Similar to the capability registration or update process of the first network element, the second network element (such as NWDAF including AnLF) can also send a seventh message to the third network element to register or update the relevant capabilities of the model's online learning. This will not be elaborated further in this example 1.

[0129] The aforementioned implementation process provided in Example 1 can refer to the relevant description in the aforementioned method embodiment 200 and achieve the same or corresponding technical effects, which will not be repeated here. In addition, Example 1 may include, but is not limited to, the aforementioned S511-S513, and may include more or fewer steps than the aforementioned S511-S513, which will not be repeated here.

[0130] Example 2: Online Model Learning like Figure 5b As shown, the online learning process of the first model may include, but is not limited to, the following steps.

[0131] S521, the first network element (e.g., NWDAF including MTLF) discovers a second network element (e.g., NWDAF including AnLF) from the third network element (e.g., NRF) that supports online learning of the model.

[0132] S522, the first network element sends a second message to the second network element, the second message being used to request the second network element to join the online learning of the model.

[0133] S523, the second network element sends a third message to the first network element according to the second message, the third message being used to indicate whether it agrees to join the online learning of the model.

[0134] Optionally, if the third message is used to indicate disagreement with joining the online learning of the model, the third message may include information on the reason for disagreement with joining the online learning of the model.

[0135] S524, the consumer network element (such as PCF, LMF, etc.) sends an analysis request to the second network element, and the analysis request may include a model identifier (or analysis ID).

[0136] S525, the second network element 5GC collects model inference data, uses the first model that supports online learning to perform model inference (or analysis, prediction), obtains model inference results (or analysis / prediction results), and stores the model inference results, related inference data and corresponding time information (such as timestamps).

[0137] S526, the second network element provides the model inference results to the consumer network element and requests the consumer network element to provide feedback on relevant information about the model inference results, such as truth data and the time information for collecting the truth data.

[0138] S527, the consumer network element sends a sixth message to the second network element, wherein the sixth message is used to indicate information related to the model inference result of the first model.

[0139] S528, the second network element determines the first loss information or the second loss information based on the stored model inference results and the information related to the model inference results of the first model in the sixth message.

[0140] S529, the second network element sends a first message to the first network element. The first message is used to indicate first information, and the first information is used for online learning of the first model.

[0141] S530, the first network element uses the information included in the first message as model training data to perform online learning of the first model.

[0142] Optionally, in order to train or update the first model, the first network element may also calculate the gradient based on the loss information.

[0143] S531, the first network element sends a fifth message to the second network element to indicate relevant information about the second model obtained after online learning and updating the first model.

[0144] Optionally, in the aforementioned model learning process, S524-S529 can be executed repeatedly until the online learning of the first model ends.

[0145] In this Example 2, the first network element can obtain real-time data (such as first information) generated in the communication network through interaction with the second network element, so as to perform online learning of the first model, making the first model more in line with the data characteristics of the communication network and improving the model inference performance.

[0146] The aforementioned implementation process provided in Example 2 can refer to the relevant description in the aforementioned method embodiment 200 and achieve the same or corresponding technical effects, which will not be repeated here. In addition, Example 2 may include, but is not limited to, the aforementioned S521-S531, and may include more or fewer steps than the aforementioned S521-S431, which will not be repeated here.

[0147] Example 3: Online Model Learning like Figure 5c As shown, the difference between Example 3 and Example 2 is that in Example 2, after receiving the sixth message from the consumer network element, the second network element needs to determine the first loss information or the second loss information based on the stored model inference results and the information related to the model inference results of the first model in the sixth message, and then send the first message to the first network element based on the calculation results, i.e., S528; while in Example 3, after receiving the sixth message from the consumer network element, the second network element does not need to determine the first loss information or the second loss information, but directly sends the first message to the first network element, and then the first network element calculates the first loss information or the second loss information based on the received first message. That is to say, in Example 3, the first message may not include the first loss information and the second loss information.

[0148] Figure 6 This diagram illustrates a flowchart of a model learning method 600 provided in an embodiment of this application. This method 600 can be executed by a second network element. Figure 6 As shown, the method 600 may include the following steps.

[0149] S610, the second network element sends a first message to the first network element. The first message is used to indicate first information, and the first information is used for online learning of the first model.

[0150] In some embodiments, the first information includes at least one of the following: model inference data used for model inference of the first model; truth data corresponding to the model inference result; a first time, the first time being the time of acquisition of the model inference data; a second time, the second time being the time of acquisition of the truth data; a third time, the time of application of the model inference result; model inference feedback information determined based on the model inference result; a first indication information used to indicate the chronological relationship between the time of acquisition of the truth data and the time of application of the model inference result; a first loss information used to represent the loss determined according to the model inference result and the truth data; and a second loss information used to represent the cumulative loss determined according to the model inference result and the truth data within a predetermined time period.

[0151] In some embodiments, the method further includes: the second network element receiving a sixth message from a consumer network element; wherein the sixth message is used to indicate information related to the model inference result of the first model, and the first message is determined based on the sixth message.

[0152] In some embodiments, the sixth message includes at least one of the following: truth data, which corresponds to the model inference result; a second time, the time when the truth data was collected; a third time, the time when the model inference result was applied; model inference feedback information, which is determined based on the model inference result; and a first indication information, used to indicate the chronological relationship between the time when the truth data was collected and the time when the model inference result was applied.

[0153] In some embodiments, the model inference feedback information includes model inference quality information.

[0154] In some embodiments, the method further includes: the second network element receiving a second message from the first network element, the second message being used to request the second network element to join the online learning of the model; the second network element sending a third message to the first network element, the third message being used to indicate whether it agrees to join the online learning of the model.

[0155] In some embodiments, the second message includes at least one of the following: a model file that supports online learning; a model access address for the model file that supports online learning; a model identifier for the model file that supports online learning; and a model analysis identifier for the model file that supports online learning.

[0156] In some embodiments, the third message is used to indicate a disagreement to join the online learning of the model, and the third message includes information on the reasons for disagreeing to join the online learning of the model.

[0157] In some embodiments, the method further includes: the second network element sending a seventh message to the third network element; wherein the seventh message includes at least one of the following: a fifth indication message for indicating whether it supports being a member of online model learning; a sixth indication message for indicating whether it supports online model learning; and a seventh indication message for indicating whether it supports the collection of data required for online model learning.

[0158] In some embodiments, the method further includes: the second network element receiving a fifth message from the first network element; wherein the fifth message includes at least one of the following: a second model; at least some model parameters of the second model; a model identifier of the second model; a model access address of the second model; a model analysis identifier of the second model; wherein the second model is obtained by updating the first model through online learning.

[0159] In some embodiments, the first network element includes a model training function; the second network element includes a model inference function.

[0160] It is understood that each implementation in this method embodiment 600 has the same or corresponding technical features as each implementation in the aforementioned method embodiment 200. Therefore, the implementation process of each implementation in this method embodiment 600 can refer to the relevant description in the aforementioned method embodiment 200 and achieve the same or corresponding technical effects. To avoid repetition, it will not be described again here.

[0161] Figure 7 A schematic diagram of the model learning device 700 provided in this embodiment is shown. The device 700 can be a component of the first network element or applied to the first network element, or it can be the first network element itself; no limitation is made here. In this embodiment, the device 700 may include: a transmission module 710, configured to receive a first message from a second network element, the first message indicating first information used for online learning of a first model; and a processing module 720, configured to perform online learning of the first model based on the first information.

[0162] In some embodiments, the first information includes at least one of the following: model inference results of the first model; truth data, which corresponds to the model inference results of the first model; model inference data, which is obtained by inferring the model inference data through the first model; a first time, which is the time when the model inference data is collected; a second time, which is the time when the truth data is collected; a third time, which is the time when the model inference results are applied; model inference feedback information, which is determined based on the model inference results; a first indication information, which indicates the chronological relationship between the time when the truth data is collected and the time when the model inference results are applied; a first loss information, which represents the loss determined based on the model inference results and the truth data; and a second loss information, which represents the cumulative loss determined based on the model inference results and the truth data within a predetermined time period.

[0163] In some embodiments, the model inference feedback information includes model inference quality information.

[0164] In some embodiments, the processing module 720 is further configured to: determine the second network element that supports online model learning; and request the second network element to join the online model learning.

[0165] In some embodiments, requesting the second network element to join the online learning model includes: sending a second message to the second network element, the second message being used to request the second network element to join the online learning model; A third message is received from the second network element, the third message being used to indicate whether the user agrees to join the online learning of the model.

[0166] In some embodiments, the second message includes at least one of the following: a model file that supports online learning; a model access address for the model file that supports online learning; a model identifier for the model file that supports online learning; and a model analysis identifier for the model file that supports online learning.

[0167] In some embodiments, the third message is used to indicate a disagreement to join the online learning of the model, and the third message includes information on the reasons for disagreeing to join the online learning of the model.

[0168] In some embodiments, the transmission module 710 is further configured to send a fourth message to the third network element; wherein the fourth message includes at least one of the following: a second indication information for indicating whether it supports being a member of online model learning; a third indication information for indicating whether it supports online model learning; and a fourth indication information for indicating whether it supports the collection of data required for online model learning.

[0169] In some embodiments, the transmission module 710 is further configured to send a fifth message to the second network element; wherein the fifth message includes at least one of the following: a second model; at least some model parameters of the second model; a model identifier of the second model; a model access address of the second model; a model analysis identifier of the second model; wherein the second model is obtained by updating the first model through online learning.

[0170] The device 700 provided in this application embodiment can execute the methods described in the preceding method embodiments and achieve the functions and beneficial effects of the methods described in the preceding method embodiments, which will not be repeated here.

[0171] Figure 8 A schematic diagram of the model learning device 800 provided in an embodiment of this application is shown. The device 800 can be a component of the second network element or applied to the second network element, or it can be the second network element itself; no limitation is made here. In this embodiment, the device 800 may include: a transmission module 810, used to send a first message to a first network element, the first message indicating first information, and the first information being used for online learning of the first model.

[0172] In some embodiments, the first information includes at least one of the following: model inference data used for model inference of the first model; truth data corresponding to the model inference result; a first time, the first time being the time of acquisition of the model inference data; a second time, the second time being the time of acquisition of the truth data; a third time, the time of application of the model inference result; model inference feedback information determined based on the model inference result; a first indication information used to indicate the chronological relationship between the time of acquisition of the truth data and the time of application of the model inference result; a first loss information used to represent the loss determined according to the model inference result and the truth data; and a second loss information used to represent the cumulative loss determined according to the model inference result and the truth data within a predetermined time period.

[0173] In some embodiments, the transmission module 810 is further configured to receive a sixth message from a consumer network element; wherein the sixth message is configured to indicate information related to the model inference result of the first model, and the first message is determined based on the sixth message.

[0174] In some embodiments, the sixth message includes at least one of the following: truth data, which corresponds to the model inference result; a second time, the time when the truth data was collected; a third time, the time when the model inference result was applied; model inference feedback information, which is determined based on the model inference result; and a first indication information, used to indicate the chronological relationship between the time when the truth data was collected and the time when the model inference result was applied.

[0175] In some embodiments, the model inference feedback information includes model inference quality information.

[0176] In some embodiments, the transmission module 810 is further configured to receive a second message from the first network element, the second message being used to request the second network element to join the online learning of the model; the transmission module 810 is further configured to send a third message to the first network element, the third message being used to indicate whether it agrees to join the online learning of the model.

[0177] In some embodiments, the second message includes at least one of the following: a model file that supports online learning; a model access address for the model file that supports online learning; a model identifier for the model file that supports online learning; and a model analysis identifier for the model file that supports online learning.

[0178] In some embodiments, the third message is used to indicate a disagreement to join the online learning of the model, and the third message includes information on the reasons for disagreeing to join the online learning of the model.

[0179] In some embodiments, the transmission module 810 is further configured to send a seventh message to a third network element; wherein the seventh message includes at least one of the following: a fifth indication information for indicating whether it supports being a member of online model learning; a sixth indication information for indicating whether it supports online model learning; and a seventh indication information for indicating whether it supports the collection of data required for online model learning.

[0180] In some embodiments, the transmission module 810 is further configured to receive a fifth message from the first network element; wherein the fifth message includes at least one of the following: a second model; at least some model parameters of the second model; a model identifier of the second model; a model access address of the second model; a model analysis identifier of the second model; wherein the second model is obtained by updating the first model through online learning.

[0181] The device 800 provided in this application embodiment can execute the methods described in the preceding method embodiments and achieve the functions and beneficial effects of the methods described in the preceding method embodiments, which will not be repeated here.

[0182] Figure 9 This diagram illustrates the hardware structure of the network-side device implemented according to embodiments of this application. Referring to the diagram, at the hardware level, the network-side device includes a processor and optionally, an internal bus, a network interface, and memory. The memory may include RAM, such as high-speed random-access memory (RAM), or non-volatile memory, such as at least one disk storage device. Of course, the network-side device may also include other hardware required for other services.

[0183] The processor, network interface, and memory can be interconnected via an internal bus, which can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be categorized as an address bus, data bus, control bus, etc. For ease of illustration, only a single bidirectional arrow is used in this diagram, but this does not imply that there is only one bus or one type of bus.

[0184] Memory is used to store programs. Specifically, programs may include program code, which includes computer operation instructions. Memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0185] The processor reads the corresponding computer program from non-volatile memory into main memory and then executes it, forming a device at the logical level that is assigned to a specific user. The processor executes the program stored in memory and specifically performs the following: Figures 2-4 The methods disclosed in the embodiments shown achieve the functions and beneficial effects of the methods described in the preceding method embodiments, and will not be repeated here.

[0186] The above is as stated in this application. Figures 2-4 The methods disclosed in the illustrated embodiments can be applied to or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above methods can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0187] The network-side device can also execute the methods described in the preceding method embodiments and achieve the functions and beneficial effects of the methods described in the preceding method embodiments, which will not be repeated here.

[0188] Of course, in addition to software implementation, the network-side device of this application does not exclude other implementation methods, such as logic devices or a combination of hardware and software, etc. In other words, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0189] This application also proposes a computer-readable storage medium that stores one or more programs, which, when executed by a network-side device including multiple applications, cause the network-side device to perform... Figures 2-4 The methods disclosed in the embodiments shown achieve the functions and beneficial effects of the methods described in the preceding method embodiments, and will not be repeated here.

[0190] The computer-readable storage medium mentioned above includes read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, etc.

[0191] This application also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, which, when executed by a computer, implement the following process: Figures 2-4 The methods disclosed in the embodiments shown achieve the functions and beneficial effects of the methods described in the preceding method embodiments, and will not be repeated here.

[0192] Computer-readable storage media include both permanent and non-permanent, removable and non-removable media, which can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable storage media do not include transient media, such as modulated data signals and carrier waves.

[0193] In summary, the above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

[0194] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0195] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0196] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

Claims

1. A model learning method, comprising: The first network element receives a first message from the second network element. The first message is used to indicate first information, and the first information is used for the online learning of the first model. The first network element performs online learning of the first model based on the first information.

2. The method of claim 1, wherein, The first information includes at least one of the following: The model inference results of the first model; Truth data, which corresponds to the model inference results of the first model; Model inference data, wherein the model inference result is obtained by inferring the model inference data through the first model; The first time, where the first time is the time when the model inference data is collected; The second time, where the second time is the time when the truth data is collected; The third time is the application time of the model's inference results; Model inference feedback information, which is determined based on the model inference result; The first indication information is used to indicate the chronological relationship between the time of collection of the truth data and the time of application of the model inference results; First loss information, which represents the loss determined based on the model inference result and the truth data; The second loss information is used to represent the cumulative loss determined based on the model inference results and truth data over a predetermined time period.

3. The method of claim 1, wherein, The model inference feedback information includes model inference quality information.

4. The method of claim 1, wherein, The method further includes at least one of the following: The first network element determines the second network element that supports online model learning; The first network element requests the second network element to join the model's online learning.

5. The method of claim 4, wherein, The first network element requests the second network element to join the online learning of the model, including: The first network element sends a second message to the second network element, the second message being used to request the second network element to join the online learning of the model; The first network element receives a third message from the second network element, the third message being used to indicate whether it agrees to join the online learning of the model.

6. The method of claim 5, wherein, The second message includes at least one of the following: Model files that support online learning; The model access address for model files that support online model learning; Model identifiers for model files that support online learning; Model analysis identifiers for model files that support online model learning.

7. The method of claim 5, wherein, The third message is used to indicate if one does not agree to join the online learning of the model, and the third message includes information on the reason for not agreeing to join the online learning of the model.

8. The method of claim 1, wherein, The method further includes: The first network element sends a fourth message to the third network element; The fourth message includes at least one of the following: The second instruction information is used to indicate whether it is supported to be a member of online learning; The third instruction information is used to indicate whether online learning is supported; The fourth instruction is used to indicate whether the collection of data required for online learning is supported.

9. The method of claim 1, wherein, The method further includes: The first network element sends a fifth message to the second network element; The fifth message includes at least one of the following: Second model; At least some of the model parameters of the second model; The model identifier for the second model; The model access address for the second model; Model analysis identifiers for the second model; The second model is obtained by updating the first model through online learning.

10. The method according to any one of claims 1-9, wherein, The first network element includes a model training function; The second network element includes model inference functionality.

11. A model learning method, comprising: The second network element sends a first message to the first network element. The first message is used to indicate first information, and the first information is used for the online learning of the first model.

12. The method of claim 11, wherein, The first information includes at least one of the following: Model inference data, which is used for model inference of the first model; Truth data, which corresponds to the model inference results; The first time, where the first time is the time when the model inference data is collected; The second time, where the second time is the time when the truth data is collected; The third time is the application time of the model's inference results; Model inference feedback information, which is determined based on the model inference result; The first indication information is used to indicate the chronological relationship between the time of collection of the truth data and the time of application of the model inference results; First loss information, which represents the loss determined based on the model inference result and the truth data; The second loss information is used to represent the cumulative loss determined based on the model inference results and truth data over a predetermined time period.

13. The method of claim 11, wherein, The method further includes: The second network element receives the sixth message from the consumer network element; The sixth message is used to indicate information related to the model inference result of the first model, and the first message is determined based on the sixth message.

14. The method of claim 13, wherein, The sixth message includes at least one of the following: Truth data, which corresponds to the model inference results; The second time is the time when the truth data is collected. The third time is the application time of the model's inference results; Model inference feedback information, which is determined based on the model inference result; The first indication information is used to indicate the chronological relationship between the time of collection of the truth data and the time of application of the model inference results.

15. The method of claim 12 or 14, wherein, The model inference feedback information includes model inference quality information.

16. The method of claim 11, wherein, The method further includes: The second network element receives a second message from the first network element, the second message being used to request the second network element to join the online learning of the model; The second network element sends a third message to the first network element, the third message being used to indicate whether it agrees to join the online learning of the model.

17. The method of claim 16, wherein, The second message includes at least one of the following: Model files that support online learning; The model access address for model files that support online model learning; Model identifiers for model files that support online learning; Model analysis identifiers for model files that support online model learning.

18. The method of claim 16, wherein, The third message is used to indicate if one does not agree to join the online learning of the model, and the third message includes information on the reason for not agreeing to join the online learning of the model.

19. The method of claim 11, wherein, The method further includes: The second network element sends a seventh message to the third network element; The seventh message includes at least one of the following: The fifth instruction is used to indicate whether support is provided as a member for online model learning; The sixth instruction indicates whether online learning of the model is supported; The seventh instruction indicates whether the collection of data required for online model learning is supported.

20. The method of claim 11, wherein, The method further includes: The second network element receives the fifth message from the first network element; The fifth message includes at least one of the following: Second model; At least some of the model parameters of the second model; The model identifier for the second model; The model access address for the second model; Model analysis identifiers for the second model; The second model is obtained by updating the first model through online learning.

21. The method according to any one of claims 11-20, wherein, The first network element includes a model training function; The second network element includes model inference functionality.

22. A model learning device, comprising: The transmission module is used to receive a first message from the second network element, the first message being used to indicate first information, and the first information being used for online learning of the first model; The processing module is used to perform online learning of the first model based on the first information.

23. The apparatus of claim 22, wherein, The first information includes at least one of the following: The model inference results of the first model; Truth data, which corresponds to the model inference results of the first model; Model inference data, wherein the model inference result is obtained by inferring the model inference data through the first model; The first time, where the first time is the time when the model inference data is collected; The second time, where the second time is the time when the truth data is collected; The third time is the application time of the model's inference results; Model inference feedback information, which is determined based on the model inference result; The first indication information is used to indicate the chronological relationship between the time of collection of the truth data and the time of application of the model inference results; First loss information, which represents the loss determined based on the model inference result and the truth data; The second loss information is used to represent the cumulative loss determined based on the model inference results and truth data over a predetermined time period.

24. The apparatus of claim 23, wherein, The model inference feedback information includes model inference quality information.

25. The apparatus of claim 22, wherein, The processing module is also used for at least one of the following: Identify the second network element that supports online learning of the model; Request the second network element to join the model's online learning.

26. The apparatus of claim 25, wherein, The request for the second network element to join the model's online learning includes: Send a second message to the second network element, the second message being used to request the second network element to join the model's online learning; A third message is received from the second network element, the third message being used to indicate whether the user agrees to join the online learning of the model.

27. The apparatus of claim 26, wherein, The second message includes at least one of the following: Model files that support online learning; The model access address for model files that support online model learning; Model identifiers for model files that support online learning; Model analysis identifiers for model files that support online model learning.

28. The apparatus of claim 26, wherein, The third message is used to indicate if one does not agree to join the online learning of the model, and the third message includes information on the reason for not agreeing to join the online learning of the model.

29. The apparatus of claim 22, wherein, The transmission module is also used to send a fourth message to the third network element; The fourth message includes at least one of the following: The second instruction information is used to indicate whether it supports being a member of the model's online learning; The third indication information is used to indicate whether online learning of the model is supported; The fourth instruction indicates whether the collection of data required for online model learning is supported.

30. The apparatus of claim 22, wherein, The transmission module is also used to send a fifth message to the second network element; The fifth message includes at least one of the following: Second model; At least some of the model parameters of the second model; The model identifier for the second model; The model access address for the second model; Model analysis identifiers for the second model; The second model is obtained by updating the first model through online learning.

31. A model learning device, comprising: The transmission module is used to send a first message to the first network element. The first message is used to indicate first information, and the first information is used for online learning of the first model.

32. The apparatus of claim 31, wherein, The first information includes at least one of the following: Model inference data, which is used for model inference of the first model; Truth data, which corresponds to the model inference results; The first time, where the first time is the time when the model inference data is collected; The second time, where the second time is the time when the truth data is collected; The third time is the application time of the model's inference results; Model inference feedback information, which is determined based on the model inference result; The first indication information is used to indicate the chronological relationship between the time of collection of the truth data and the time of application of the model inference results; First loss information, which represents the loss determined based on the model inference result and the truth data; The second loss information is used to represent the cumulative loss determined based on the model inference results and truth data over a predetermined time period.

33. The apparatus of claim 31, wherein, The transmission module is also used to receive a sixth message from the consumer network element; The sixth message is used to indicate information related to the model inference result of the first model, and the first message is determined based on the sixth message.

34. The apparatus of claim 33, wherein, The sixth message includes at least one of the following: Truth data, which corresponds to the model inference results; The second time is the time when the truth data is collected. The third time is the application time of the model's inference results; Model inference feedback information, which is determined based on the model inference result; The first indication information is used to indicate the chronological relationship between the time of collection of the truth data and the time of application of the model inference results.

35. The apparatus of claim 32 or 34, wherein, The model inference feedback information includes model inference quality information.

36. The apparatus of claim 31, wherein, The transmission module is further configured to receive a second message from the first network element, the second message being used to request the second network element to join the online learning of the model; The transmission module is also used to send a third message to the first network element, the third message being used to indicate whether or not one agrees to join the online learning of the model.

37. The apparatus of claim 36, wherein, The second message includes at least one of the following: Model files that support online learning; The model access address for model files that support online model learning; Model identifiers for model files that support online learning; Model analysis identifiers for model files that support online model learning.

38. The apparatus of claim 36, wherein, The third message is used to indicate if one does not agree to join the online learning of the model, and the third message includes information on the reason for not agreeing to join the online learning of the model.

39. The apparatus of claim 31, wherein, The transmission module is also used to send a seventh message to a third network element; The seventh message includes at least one of the following: The fifth instruction is used to indicate whether support is provided as a member for online model learning; The sixth instruction indicates whether online learning of the model is supported; The seventh instruction indicates whether the collection of data required for online model learning is supported.

40. The apparatus of claim 31, wherein, The transmission module is also used to receive a fifth message from the first network element; The fifth message includes at least one of the following: Second model; At least some of the model parameters of the second model; The model identifier for the second model; The model access address for the second model; Model analysis identifiers for the second model; The second model is obtained by updating the first model through online learning.

41. A network-side device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the method as described in any one of claims 1-21.

42. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method as described in any one of claims 1-21.

Images