Communication method, communication node and storage medium

Abstract

Provided in the present application are a communication method, a communication node and a storage medium. The communication method comprises: a first communication device receiving a model update request sent by a second communication device, wherein the model update request is used for instructing the first communication device to update a stored model; and the first communication device updating the model on the basis of the model update request.

Description

Communication method, communication node and storage medium Technical Field

[0001] The present application relates to the field of wireless communication technologies, and in particular to a communication method, a communication node, and a storage medium. Background Art

[0002] 5GC network element devices can perform statistical data and machine learning related tasks in 5GS. The network element devices may include one or more logical functions of analysis logic functions and model training logic functions. Among them, the analysis logic function is used to perform reasoning, derive analysis information (that is, derive statistical data and / or predictions based on analysis consumer requests) and expose analysis services; the model training logic function is used to train machine learning (ML) models and expose new training services (such as providing trained ML models). The trained model needs to be updated after the reasoning effect is poor. Although the 23.288 protocol supports the storage and update of trained models, the update effect is poor and the efficiency is low.

[0003] Summary of the Invention

[0004] This application provides a communication method, a communication node, and a storage medium to solve the problem of poor model update effect.

[0005] An embodiment of the present application provides a communication method, applied to a first communication device, including:

[0006] Receive a model update request sent by a second communication device, where the model update request is used to instruct the first communication device to update a stored model; and perform model update according to the model update request.

[0007] This embodiment of the present application provides another communication method, applied to a second communication device, including:

[0008] A model update request is sent to the first communication device, so that the first communication device updates the model according to the model update request; wherein the model update request is used to instruct the first communication device to update the stored model.

[0009] This embodiment of the present application provides another communication method, which is applied to a first communication device, including:

[0010] Receive a model management storage request sent by a second communication device, wherein the model management storage request includes a storage processing identifier and model information; delete one or more local models according to the storage processing identifier in the model management storage request, and store the updated model file according to the model information.

[0011] This embodiment of the present application provides another communication method, applied to a second communication device, including:

[0012] A model management storage request is sent to a first communication device, so that the first communication device updates the model according to the model management storage request; wherein the model management storage request includes a storage processing identifier and model information.

[0013] An embodiment of the present application provides a communication node, comprising: a memory, a processor, a program stored on the memory and executable on the processor, and a data bus for implementing connection communication between the processor and the memory. When the program is executed by the processor, a communication method as described in any one of the embodiments of the present application is implemented.

[0014] An embodiment of the present application provides a storage medium for computer-readable storage, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the communication method described in any one of the embodiments of the present application.

[0015] The communication method, apparatus, communication node, and storage medium provided in the embodiments of the present application are as follows: a first communication device receives a model update request sent by a second communication device, and the model update request is used to instruct the first communication device to update a stored model; the first communication device can directly update the model according to the model update request without having to interact with the second communication device multiple times; the first communication device directly updates the model based on the instruction of the model update request, with high update efficiency and better effect; compared with related technologies that require multiple interactions to achieve model updates, the speed is faster, and the communication resources of the first communication device and the second communication device are effectively saved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] FIG1a is an architecture diagram of a core network provided by an embodiment;

[0017] FIG1b is a diagram illustrating an architecture of data collection for a network data analysis function provided by an embodiment;

[0018] FIG1c is a network data analysis exposure architecture provided by an embodiment;

[0019] FIG1d is a diagram illustrating an architecture of data collection coordination and functionality provided by an embodiment;

[0020] FIG1e is another network data analysis exposure architecture provided by an embodiment;

[0021] FIG1f is a diagram of a 5G system architecture including ADRF according to an embodiment;

[0022] FIG2 is a flow chart of a communication method provided by an embodiment;

[0023] FIG3 is a flow chart of another communication method provided by an embodiment;

[0024] FIG4 is a flowchart illustrating another communication process according to an embodiment;

[0025] FIG5 is a flowchart illustrating another communication process according to an embodiment;

[0026] FIG6 is a flowchart illustrating a model storage update process according to an embodiment;

[0027] FIG7 is a flowchart illustrating a model subscription update process according to an embodiment;

[0028] FIG8 is a flowchart illustrating a model storage update process according to an embodiment;

[0029] FIG9 is a schematic structural diagram of a communication device provided by an embodiment;

[0030] FIG10 is a schematic structural diagram of another communication device provided by an embodiment;

[0031] FIG11 is a schematic structural diagram of another communication device provided by an embodiment;

[0032] FIG12 is a schematic structural diagram of another communication device provided by an embodiment;

[0033] FIG13 is a schematic structural diagram of a communication node provided by an embodiment. DETAILED DESCRIPTION

[0034] In the absence of conflict, the embodiments and features in the embodiments of this application can be combined with each other in any manner.

[0035] Figure 1a is a diagram of a core network architecture provided by an embodiment, which is a diagram of a 5G core network architecture. This architecture has the following functions:

[0036] 1) UE, user equipment.

[0037] 2) RAN, Radio Access Network. The RAN manages radio resources, delivers user data received via N3 to the UE, and transmits user data from the UE over the N3 interface. The RAN maps QoS flows within DRBs (Dedicated Radio Bearers) and PDUs in a session.

[0038] 3) AMF, Access and Mobility Management Function. This function includes registration management, connection management, reachability management, and mobility management. This function also performs access authentication and access authorization. The AMF is the NAS security terminal and forwards SM NAS between the UE and SMF.

[0039] 4) SMF, Session Management Function, includes the following functions: session establishment, modification, and release; UE IP address allocation and management (including optional authorization functions); UP function selection and control; downlink data notification; etc. The SMF controls the UPF through N4 association. The SMF provides the UPF with PDR (Packet Detection Rules), instructing how to detect user data traffic; and provides FAR (Forwarding Action Rules), QER (QoS Enforcement Rules), and URR (Usage Reporting Rules), instructing the UPF on how to perform user data traffic forwarding, QoS processing, and usage reporting for user data traffic detected using the PDR.

[0040] 5) UPF, User Plane Function. This function includes the following: serving as an anchor point for intra- and inter-Radio Access Type (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 is per PDU session. For downlink traffic, the UPF uses the FAR received from the SMF to bind the downlink traffic to the QoS traffic within the PDU session GTP-U tunnel. For uplink traffic, the RAN transfers the user plane traffic to the QoS flow identified by the UE.

[0041] 6) PCF, Policy Control Function. PCF provides QoS policy rules to the control plane functions to enforce these rules. PCF converts AF requests into PCC rules applicable to the PDU session.

[0042] 7) UDM, unified data management. The UDM performs the generation of 3GPP AKA authentication credentials, authorization of access based on subscription data, UE registration with the serving NF (e.g., storing the service AMF for the UE and the PDU session storage SMF for the UE), and subscription management. The UDM accesses the UDR to retrieve UE subscription data and stores the UE context in the UDR. The UDM and UDR can be deployed together.

[0043] NWDAF (Network Data Analysis Function) is a 5GC NF located in the control plane that performs statistics and machine learning related tasks in 5GS. NWDAF can interact with different entities for different purposes:

[0044] Collect data based on event subscriptions provided by AMF, SMF, UPF, PCF, UDM, NSACF, AF (directly or through NEF) and OAM; in some embodiments, use DCCF (Data Collection Coordination Function) for analysis and data collection; retrieve information from data repositories (such as UDR that retrieves user-related information through UDM or PFD information through NEF (PFDF)); collect location information data from LCS systems; in some embodiments, store and retrieve information from ADRF (Analysis Data Storage Function); in some embodiments, analyze and collect data from MFAF (Information Transfer Framework Adapter Function); retrieve information about NF (for example, retrieve NF-related information from NRF); provide analysis to consumers on demand.

[0045] Provides batch data associated with an analysis ID.

[0046] Provides accuracy information for analysis IDs.

[0047] Provides ML model accuracy information or an indication of ML model accuracy degradation about an ML model.

[0048] A single instance or multiple instances of NWDAF can be deployed in a PLMN. NWDAF can include the following logical functions:

[0049] Analytics Logic Function (AnLF): A logic function in NWDAF that performs inference, derives analytical information (i.e., derives statistics and / or predictions based on analytics consumer requests) and exposes analytics services (i.e., Nnwdaf_AnalyticsSubscription or Nnwdaf_AnalyticsInfo).

[0050] Model Training Logical Function (MTLF): A logical function in NWDAF that is used to train machine learning (ML) models and expose new training services (e.g., serving trained ML models).

[0051] An NWDAF can contain a MTLF or an AnLF, or both logical functions.

[0052] The Data Collection Coordination and Function (DCCF) is also an NF on the 5GC control plane. The DCCF is responsible for coordinating the collection and distribution of data requested by NF consumers. It prevents data sources from processing multiple subscriptions to the same data and prevents multiple notifications containing the same information from being sent due to uncoordinated requests from data consumers.

[0053] The DCCF applies to:

[0054] NWDAF that requests data from a data source (e.g., for computational analysis).

[0055] NF consumers that request analytics from NWDAF data sources.

[0056] NF Consumer that requests data from ADRF data source.

[0057] ADRF that receives data from NF data source.

[0058] FIG1b is an architecture diagram of data collection of a network data analysis function provided by an embodiment. The network data analysis function NWDAF collects data of any network element NF through the network function Nnf to facilitate data analysis and model training by NWDAF.

[0059] FIG1c shows a network data analysis exposure architecture provided by an embodiment. NWDAF provides analysis, statistics, and predictions to other NFs for use by any NF consumer that subscribes to or requests analysis.

[0060] Figure 1d is an architectural diagram of a data collection coordination and function provided by an embodiment. The data collection coordination and function DCCF coordinates data collection and distribution between any network element NF and the network data analysis function NWDAF through the Ndccf service and the Nnf service; DCCF is registered through the network resource management function entity NRF, the unified data management UDM or the binding support function BSF; DCCF interacts with the messaging framework adaptor function (MFAF) through the data management configuration service Nmfaf to realize data collection and distribution. The messaging framework adapter function MFAF interacts with any network element NF through the Nnf service to realize data collection, and MFAF interacts with NWDAF through the Nmfaf service to realize data collection. By collecting data from any network element NF, NWDAF can perform data analysis and model training.

[0061] Figure 1e shows another network data analysis exposure architecture provided by an embodiment. DCCF helps send the analysis and prediction results provided by NWDAF. DCCF and NWDAF interact through the Nnwdaf service for use by any NF consumer who subscribes to or requests analysis.

[0062] Figure 1f shows a 5G system architecture diagram including ADRF, provided by one embodiment. The 5G system architecture allows ADRF to store and retrieve collected data and analyze it. The following options are supported:

[0063] ADRF exposes Nadrf services for other 5GC network functions (such as NWDAF) to store and retrieve data, which can be accessed through Nadrf services.

[0064] Based on the request of the network function or the configuration on the DCCF, the DCCF can determine the ADRF and interact with it directly or indirectly to request or store data. The interaction can be:

[0065] Direct: DCCF requests to store data in ADRF through Nadrf service or Ndccf_DataManagement_Notify (for example, when ADRF requests DCCF for data collection notification). In addition, DCCF retrieves data from ADRF through Nadrf service.

[0066] Indirect: DCCF requires the message framework to store data in ADRF through the Nadrf service or Nmfaf_3daDataManagement_Configure. The message framework may contain one or more adapters for converting between 3GPP defined protocols.

[0067] Note: The internal logic of the message framework is outside the scope of 3GPP. Only the MFAF and the interfaces between the MFAF and other 3GPP defined network functions are within the scope of 3GPP.

[0068] The consumer network function may specify in the request sent to the DCCF that the data provided by the data source needs to be stored in the ADRF.

[0069] The ADRF stores data received directly from the network function in a Nadrf_DataManagement_StorageRequest, or in a Ndccf_DataManagement_Notify, Nmfaf_3caDataManagement_Notify, or Nnwdaf_DataManagement_Notify sent from the DCCF, MFAF, or NWDAF.

[0070] ADRF checks whether the data consumer is authorized to access the ADRF service and provides the requested data according to the procedures specified in TS23.501[2] section 7.1.4.

[0071] Model storage and related signaling provided:

[0072] Nadrf_MLModelManagement_StorageRequest

[0073] Service operation name: Nadrf_MLModelManagement_StorageRequest

[0074] Description: This service operation is used by the consumer NF to request ADRF to store or update a machine learning model. The machine learning model stored in the NWDAF containing the MTLF or the machine learning model address is provided to ADRF in the request message.

[0075] Input, required: The NF ID of the NWDAF containing the MTLF, and the following set of contents:

[0076] A machine learning model; or a tuple of (a unique machine learning model identifier, the address of the model file (e.g., a URL or FQDN), and the required storage size for each machine learning model).

[0077] Note: ADRF can download the machine learning model file based on the machine learning model address provided by the NWDAF containing the MTLF. How ADRF downloads and locally stores the machine learning model is left to be implemented.

[0078] FIG2 is a flow chart of a communication method provided by an embodiment. As shown in FIG2 , the communication method according to the embodiment of the present application is applied to a first communication device. The method includes S110 to S120:

[0079] S110: Receive a model update request sent by the second communication device, where the model update request is used to instruct the first communication device to update a stored model.

[0080] A model update request can be understood as a communication request for instructing a communication device to update a model. The model update request may be a newly added service operation dedicated to model updates. A second communication device sends a model update request to a first communication device. Upon receiving the model update request, the first communication device determines that the model update request instructs the device to update a stored model.

[0081] S120: Update the model according to the model update request.

[0082] The first communications device updates the model in response to the model update request. The model update process includes deleting the previously stored model and storing the updated model. Under the instruction of the model update request, the first communications device directly determines the model to be updated, deletes it, and then saves the updated model. This allows the model update to be completed in a single interaction, i.e., upon receiving the model update request.

[0083] Related technologies can only update models through Nadrf_MLModelManagement_StorageRequest. However, when using Nadrf_MLModelManagement_StorageRequest to update a model, the model cannot be updated directly and multiple interactions are required to complete the update. For example, upon receiving a Nadrf_MLModelManagement_StorageRequest, the first communication device first deletes the stored model that needs to be updated, then sends feedback to the second communication device regarding the deleted model, and then stores the updated model after receiving instructions from the second communication device. This update method requires multiple interactions, is inefficient, and wastes communication resources of the communication devices.

[0084] In the communication method provided in the embodiment of the present application, a first communication device receives a model update request sent by a second communication device, and the model update request is used to instruct the first communication device to update a stored model; the first communication device can directly update the model according to the model update request without having to interact with the second communication device multiple times. The first communication device directly updates the model based on the instruction of the model update request, and the update efficiency is high and the effect is better. Compared with the related technology, the model update can be achieved through multiple interactions, the speed is faster, and the communication resources of the first communication device and the second communication device are effectively saved.

[0085] In some embodiments, the model update request is a newly added service operation, and the service operation name may be Nadrf_MLModelManagement_UpdateRequest. The second communication device may use this service operation to request the first communication device to update the stored machine learning model.

[0086] In some embodiments, the model update request includes first model update information; and updating the model according to the model update request includes:

[0087] One or more local models to be updated are determined according to the first model update information, the one or more local models to be updated are deleted, and one or more updated model files are stored.

[0088] The first model update information can be understood as information required for model update, for example, the model to be updated, the updated model file, and so on.

[0089] Parse the model update request to determine the information included in the model update request. When the model update request includes the first model update information, determine the local model that needs to be updated and the updated model file through the first model update information. The updated model file can be carried in the first model update information or downloaded through the first model update information. The local model is the model stored locally on the first communication device, and the number of local models is one or more. Delete one or more local models to be updated, and store the updated model file corresponding to each deleted local model to be updated accordingly, for example, in the storage space corresponding to the deleted local model, or release its storage space after deleting the local model, and store the updated model file in a new storage space, etc.

[0090] The protocol only supports a single update model, not a batch update model. This approach is very time-consuming and inefficient. The first communication device in this application can simultaneously perform batch updates on multiple local models when there are multiple local models. Compared with the related art that only supports a single update model, the batch update model can effectively improve the update efficiency and save time.

[0091] In some embodiments, the first model update information includes one or more of the following:

[0092] Update indication information; one or more updated model files; addresses of one or more updated model files; one or more model identifiers; one or more updated model storage space sizes; storage processing identifier.

[0093] The update indication information can be understood as information instructing the first communication device to perform an update, such as deletion or storage, of the model. Upon receiving the model update request, the first communication device parses the model update request to obtain first model update information. If the first model update information includes the update indication information, the first communication device can determine that the model is to be deleted or stored.

[0094] In the case where the first model update information includes an updated model file, the updated model file can be directly stored to implement the model update; in the case where the first model update information includes the address of the updated model file, the updated model file can be obtained through the address of the model file, and the updated model file can be stored to implement the model update. The local model to be updated is determined by the model identifier or the storage processing identifier. The model identifier can uniquely identify a model, and the storage processing identifier can identify a batch of models, that is, the models processed in the same batch can be identified by a storage processing identifier. The updated model storage space size can be used to indicate the storage space size of the updated model file, and the first communication device can allocate storage space for the updated model file according to the updated storage space size.

[0095] In some embodiments, the first model update information may include a storage processing identifier and one or more updated model files.

[0096] In some embodiments, the first model update information may include a storage process identifier and addresses of one or more updated model files.

[0097] In some embodiments, the first model update information may include at least one set of model identifications and an updated model file.

[0098] In some embodiments, the first model update information may include at least one set of model identification, an address of an updated model file, and an updated model storage space size.

[0099] In some embodiments, the address of the model file may be a URL or a FQDN.

[0100] In some embodiments, the model update request includes second model update information; and updating the model according to the model update request includes:

[0101] One or more local models to be updated are determined according to the second model update information; and the local models are updated according to the model parameters corresponding to each local model.

[0102] The second model update information is used to indicate that the parameters of the model are updated. The model parameters can be any type of parameters of the model, for example, a classification tree that is newly added or reduced in the model.

[0103] Parse the model update request and determine the information carried in the model update request. In the case where the model update request contains the second model update information, parse the second model update information to determine one or more local models to be updated, and determine the model parameters corresponding to each local model. The model update request can carry model parameters, or a model update parameter file containing model parameters, or the address of the model update parameter file, and obtain the corresponding model update parameter file through the address. For each local model, the corresponding parameters in the stored local model are replaced or adjusted through the model parameters to achieve the update of the local model. The present application updates the local model through the model parameters, and can only update the parameters that need to be updated, and maintain the original parameters unchanged for the parameters that do not need to be updated, thereby reducing the amount of data transmission and the amount of updated data, and improving the update efficiency.

[0104] In related art, if a second communication device trains a model for 10 rounds and, to ensure that model data is not lost, updates the new model to the first communication device after each round, the second communication device can only transfer the model file after each round of training to the first communication device, and the first communication device must delete the model file 10 times and then store the new model 10 times. If the trained model file is large (for example, a large model), frequently transferring large files not only easily leads to core network signaling congestion, but the first communication device's frequent deletion and storage of large files also places a heavy load on the CPU. The first communication device subscribes to the model from the second communication device. After the second communication device trains and updates the model, it can only transfer the entire new model file to the first communication device. The second communication device frequently uploads large model files, and the first communication device frequently downloads large model files, which places a heavy burden on both parties and the server. Updating the model parameters through the second model update information eliminates the need to transmit the entire model, effectively reducing the amount of data transmitted, improving update efficiency, and alleviating the load on the first communication device.

[0105] In some embodiments, the second model update information includes one or more of the following:

[0106] One or more model parameters; one or more model parameter types; one or more model update parameter files; one or more model identifiers; the addresses of one or more model update parameter files; and the estimated storage space size of one or more updated models.

[0107] The model update parameter file may include updated model parameters, and may also include parameter information for indicating how to update the model parameters, for example, indicating that the model parameters are increased or decreased, etc. In the case where the second model update information includes at least one of the model update parameter file and the model parameters, the parameters of the local model can be modified or adjusted directly through the model parameters and / or the model parameters in the model update parameter file to achieve the model update; in the case where the second model update information includes the address of the model update parameter file, the model update parameter file can be obtained through the address of the model update parameter file, and the local model can be updated according to the model update parameter file to achieve the model update. The model parameter type can be used to determine the parameter type that needs to be updated. The local model, model parameters, etc. that need to be updated can be determined according to the model parameter type. For example, the model containing this model parameter type can be used as the local model to be updated. The model parameter type can also be used together with the model identifier to determine the local model and model parameters to be updated. The local model to be updated is determined by the model identifier. The estimated size of the updated model storage space can be used to indicate the estimated storage space size of the updated model file. The second communication device can estimate the storage space size of the updated model file based on the size of the model update parameter file, the size of the model file to be updated, etc., to obtain the estimated size of the updated model storage space; the first communication device can allocate storage space for the updated model file based on the estimated size of the updated model storage space.

[0108] In some embodiments, the second model update information may include at least one set of model update parameter files and a model identifier.

[0109] In some embodiments, the second model update information may include the address of at least one set of model update parameter files, a model identifier, and an estimated size of the updated model storage space.

[0110] In some embodiments, the second model update information may include at least one set of model parameters and a model identifier.

[0111] In some embodiments, the second model update information may include at least one set of model parameters, a model parameter type, and a model identifier.

[0112] In some embodiments, the second model update information may include at least one set of model update parameter files, model parameter types, and model identifiers.

[0113] In some embodiments, the second model update information may include the address of at least one set of model update parameter files, model parameter types, model identifiers, and estimated size of the updated model storage space.

[0114] A model update parameter file can correspond to one or more model identifiers and update one or more models.

[0115] In some embodiments, the model update parameter file includes one or more model parameters.

[0116] The model update parameter file may contain one or more model parameters, and update one or more parameters of the model simultaneously.

[0117] In some embodiments, the model update request includes third model update information;

[0118] Performing a model update based on a model update request includes: determining one or more local models to be updated and model files to be merged and updated based on the third model update information; fusing the model files to be merged and updated with each local model to obtain and store the merged models. The third model update information can be understood as information required for the model update, such as the models to be merged and updated, the models required for the merged update, and so on.

[0119] The first communication device parses the model update request and determines the information carried in the model update request. When the first communication device determines that the model update request includes third model update information, the third model update information may include unique identification information corresponding to the model to be updated, such as a model identifier; the third model update information may also include information indicating the required model fusion, such as a model file; the third model update information is parsed to determine one or more local models to be updated, and the model files to be fused and updated are determined. After receiving the third model update information, the first communication device fuses the model files to be fused and updated with each local model, obtains a fused model, and stores the fused model. The fused model is a new model.

[0120] In some embodiments, the third model update information includes one or more of the following:

[0121] Model fusion update indication information; one or more model files to be fused and updated; one or more model identifiers; the addresses of one or more model files to be fused and updated; and one or more estimated storage sizes of updated models. The model fusion update indication information is used to instruct the communication device to fuse the models and generate a new model to implement a model update. Upon receiving the model fusion update indication information, the first communication device performs a model fusion update under the instruction of the model fusion update indication information.

[0122] In the case where the third model update information includes the model file to be merged and updated, the local model can be merged and updated directly through the model file to be merged and updated; in the case where the third model update information includes the address of the model file to be merged and updated, the model file to be merged and updated can be obtained through the address of the model file to be merged and updated, and the local model can be merged and updated according to the model file to be merged and updated. The local model to be updated is determined by the model identifier. The estimated size of the updated model storage space can be used to indicate the estimated storage space size of the updated model file. The second communication device can estimate the storage space size of the updated model file according to the size of the model file to be merged and updated, the size of the model file to be updated, etc., to obtain the estimated size of the updated model storage space; the first communication device can allocate storage space for the updated model file according to the estimated size of the updated model storage space.

[0123] During the model update process, after the first communication device receives the model update information, it is not sure what processing method to take to update the model. The present application instructs the first communication device to perform a model fusion update, or to delete or save the model through update indication information and model fusion update indication information, so as to facilitate the first communication device to distinguish different update types and methods.

[0124] In some embodiments, the third model update information may include at least one set of model files to be merged and updated and a model identifier.

[0125] In some embodiments, the third model update information may include at least one set of model identifiers, addresses of model files to be merged and updated, and estimated size of the updated model storage space.

[0126] In some embodiments, the third model update information may include at least one set of model files to be merged and updated and a model identifier.

[0127] In some embodiments, the method further comprises:

[0128] After the fusion is completed, a new model identifier is generated and sent to the second communication device.

[0129] After the model fusion is completed, a new model is generated, that is, the fused model, and a new model identifier is generated to identify the fused model. The method of generating the new model identifier can be implemented according to the original model identifier generation method, or a new identifier generation strategy can be defined to generate the identifier; the model identifier can be generated according to the type, size, generation order, etc. of the model.

[0130] In some embodiments, the method further comprises:

[0131] After the model update is completed, at least one of the following information is sent to the second communication device:

[0132] The first result indication is used to indicate whether one or more models have been successfully updated or failed to be updated; the model identifier of the updated model; the address of the updated model file; and the storage processing identifier. The first result indication can be understood as information indicating the result of the model update of the first communication device, and is used to indicate whether one or more models have been successfully updated or failed to be updated. In the case of multiple model updates, the first result indication can indicate a successful update when all models are successfully updated, otherwise the update fails; or, it can indicate the models that have been successfully updated, as well as the models that have failed to be updated, so that the second communication device can determine which models have been successfully updated and which models have failed to be updated. After updating multiple models, there are multiple model identifiers of the updated models and multiple addresses of the updated model files.

[0133] After the model update is completed, the first communication device may generate a first result indication based on the update result, indicating the success or failure of one or more model updates through the first result indication; after the model update is completed, the first communication device may send one or more of the following information to the second communication device: the first result indication, the model identifier of the updated model, the address of the updated model file, and a storage processing identifier. The address of the updated model file can be determined after the updated model file is stored; the storage processing identifier can be a new storage processing identifier generated during the current model update, which identifies the model updated during the current model update process. The address of the updated model file is usually generated after the model update is successful.

[0134] In some embodiments, the first communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF; in some embodiments, the second communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF.

[0135] The first communication device or the second communication device may be a network element that performs one or more of the following functions: a network data analysis function (NWDAF), an analysis database function (ADRF), and a location management function (LMF). The first communication device or the second communication device may also be a 5GC network element that performs other functions. For example, the first communication device may be an ADRF and the second communication device may be an NWDAF; or the first communication device may be an LMF and the second communication device may be an NWDAF; or the first communication device may be an ADRF and the second communication device may be an NWDAF that includes a MTLF; or the first communication device may be an NWDAF that includes an AnLF and the second communication device may be an NMDAF that includes an MTLF.

[0136] In the communication method provided in the embodiment of the present application, the second communication device instructs the first communication device to update the model by carrying different types of indication information, model update information, etc. in the model update request, and deletes and stores the model through the first model update information, thereby quickly implementing batch updates of the model and improving update efficiency; updates the parameters of the model through the second model update information, thereby reducing the amount of data transmission and alleviating the burden on the communication device; and performs model fusion update through the third model update information to quickly generate a new model.

[0137] FIG3 is a flow chart of another communication method provided by an embodiment. As shown in FIG3 , the communication method according to the embodiment of the present application is applied to a second communication device. The method includes S210:

[0138] S210: Send a model update request to the first communication device, so that the first communication device updates the model according to the model update request.

[0139] The model update request is used to instruct the first communication device to update a stored model.

[0140] After training the new model, the second communication device may send a model update request to the first communication device, instructing the first communication device to update the stored model. The first communication device performs model update according to the received model update request.

[0141] The second communication device can train a new model and then update the model when the model reasoning effect is not ideal. For example, the first communication device determines that the model reasoning effect is not ideal when performing model reasoning, and feeds back this result to the second communication device, and so on.

[0142] In the communication method provided in the embodiment of the present application, the second communication device sends a model update request to the first communication device, and the model update request is used to instruct the first communication device to update the stored model; the first communication device can directly update the model according to the model update request without having to interact with the second communication device multiple times. The first communication device directly updates the model based on the instruction of the model update request, and the update efficiency is high and the effect is better. Compared with the related technology, the model update can be achieved through multiple interactions, the speed is faster, and the communication resources of the first communication device and the second communication device are effectively saved.

[0143] In some embodiments, the model update request includes first model update information.

[0144] In some embodiments, the first model update information includes one or more of the following:

[0145] Update indication information; one or more updated model files; addresses of one or more updated model files; one or more model identifiers; one or more updated model storage space sizes; storage processing identifier.

[0146] In some embodiments, the model update request includes second model update information.

[0147] In some embodiments, the second model update information includes one or more of the following:

[0148] One or more model parameters; one or more model parameter types; one or more model update parameter files; one or more model identifiers; the addresses of one or more model update parameter files; and the estimated storage space size of one or more updated models.

[0149] In some embodiments, the model update parameter file includes one or more model parameters.

[0150] In some embodiments, the model update request includes third model update information.

[0151] In some embodiments, the third model update information includes one or more of the following:

[0152] Model fusion update indication information; one or more model files to be fused and updated; one or more model identifiers; the addresses of one or more model files to be fused and updated; and the estimated storage space size of one or more updated models.

[0153] In some embodiments, the method further comprises:

[0154] A new model identifier sent by the first communication device is received.

[0155] In some embodiments, the method further comprises:

[0156] Receive at least one of the following information sent by the first communication device:

[0157] The first result indication is used to indicate whether one or more models are updated successfully or failed; the model identifier of the updated model; the address of the updated model file; and the storage processing identifier.

[0158] After receiving the first result indication, the second communication device can determine the result of the model update based on the information indicating the success or failure of one or more model updates indicated by the first result indication. For models that failed to update, the second communication device can also determine whether to instruct the first communication device to re-update. After receiving the model identifier and model file address of the updated model, the second communication device can determine the new model identifier and model file address to facilitate subsequent operations such as deleting and updating the new model. After receiving the storage processing identifier, the second communication device can identify the batch of models processed this time to facilitate subsequent batch processing of the models based on the storage processing identifier.

[0159] In some embodiments, the first communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF; the second communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF.

[0160] The first communication device or the second communication device can be a network element with one or more functions including the network data analysis function NWDAF, the analysis database function ADRF and the positioning management function LMF. The first communication device or the second communication device can also be a 5GC network element with other functions.

[0161] In some embodiments, the method further includes: before generating the model update request, detecting whether a model update condition is satisfied.

[0162] The model update condition can be understood as a condition for determining whether to perform a model update, for example, whether the model accuracy meets the accuracy requirements, whether the time difference between the last model update and the current time exceeds a certain threshold, etc. The second communication device can determine whether the model accuracy meets the accuracy requirements by reasoning with the model, or it can receive reasoning results or model accuracy information sent by other communication devices (for example, the first communication device). The second communication device detects in real time or periodically whether the model update condition is met. After determining that the model update condition is met, it trains a new model, prepares for the model update, and generates a model update request to instruct the first communication device to perform a model update.

[0163] In the communication method provided in the embodiment of the present application, the second communication device instructs the first communication device to update the model by carrying different types of indication information, model update information, etc. in the model update request, and deletes and stores the model through the first model update information, thereby quickly implementing batch updates of the model and improving update efficiency; updates the parameters of the model through the second model update information, thereby reducing the amount of data transmission and alleviating the burden on the communication device; and performs model fusion update through the third model update information to quickly generate a new model.

[0164] FIG4 is a flow chart of another communication method provided by an embodiment. As shown in FIG4 , the communication method according to the embodiment of the present application is applied to a first communication device. The method includes S310 to S330:

[0165] S310: Receive a model management storage request sent by a second communication device, where the model management storage request includes a storage processing identifier.

[0166] A model management and storage request can be understood as a communication instruction for performing operations such as managing and storing a model. The model management and storage request carries a storage processing identifier, which can identify one or a batch of models. The first communication device receives the model management and storage request sent by the second communication device according to the agreed communication protocol.

[0167] S320: Update the model according to the model management storage request.

[0168] The second communication device parses the model management storage request, determines the information carried in the model management storage request, determines a batch of models that need to be updated this time according to the storage processing identifier carried in the model management storage request, and updates these models.

[0169] S330: Send a second result indication to the second communication device.

[0170] The second result indication is used to indicate whether the model update is successful or failed.

[0171] A second result indication is generated based on the update result, and the second result indication is used to indicate whether the model update is successful or failed, wherein a successful update may mean that all models are updated successfully, or that part of the model is updated successfully; a failed model update may mean that all updates fail, or that part of the model is updated failed; for example, when some models are updated successfully and some models fail to be updated, the second result indication indicates that the model update failed, or the second result indicates the specific models that have been updated successfully, and the models that have failed to be updated, so that the second communication device can determine which models have been updated successfully and which models have failed to be updated.

[0172] In the communication method provided in the embodiment of the present application, a first communication device receives a model management storage request sent by a second communication device, and the model management storage request includes a storage processing identifier. The first communication device can directly update the model in batches according to the model management storage request without having to interact with the second communication device multiple times. The first communication device directly updates the model based on the instruction of the model update request, with high update efficiency and better effect. Compared with related technologies that require multiple interactions to achieve model updates, the speed is faster, and batch updates can update multiple models at a time, effectively saving communication resources of the first communication device and the second communication device.

[0173] In some embodiments, the model management storage request is a Nadrf_MLModelManagement_StorageRequest.

[0174] In some embodiments, the model management storage request further includes model information; and updating the model according to the model management storage request includes:

[0175] One or more local models are deleted according to the storage processing identifier, and the updated model file is stored according to the model information.

[0176] Model information can be information describing the model, such as the model identifier and model file.

[0177] After parsing the model management storage request, the storage processing identifier and model information are determined. According to the storage processing identifier, one or more local models that need to be updated this time are determined. One or more local models are deleted. According to the model information, the updated model file is determined. The updated model file is stored to complete the batch update of the model.

[0178] In some embodiments, the method further comprises:

[0179] After the model update is completed, a new storage processing identifier is generated and sent to the second communication device.

[0180] After the model update is completed, a new storage processing identifier is generated to identify the model processed in this batch, and the new storage processing identifier is sent to the second communication device.

[0181] In some embodiments, the model information includes one or more of the following:

[0182] One or more updated model files; one or more addresses of updated model files; one or more model identifiers; one or more updated model storage space sizes.

[0183] When the model information includes one or more updated model files, the updated model files can be directly stored in batches to complete the model update; when the model information includes the address of the updated model file, the updated model file can be obtained through the address of the updated model file, and then stored in batches to complete the model update; one or more model identifiers can indicate the local model that needs to be updated; the first communication device can allocate storage space for the updated model file according to the size of the updated model storage space.

[0184] In some embodiments, the model information may include at least one updated model file.

[0185] In some embodiments, the model information may include at least one set of updated model files and a model identifier.

[0186] In some embodiments, the model information may include at least one set of model identification, an address of an updated model file, and an updated model storage space size.

[0187] In some embodiments, the first communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF; in some embodiments, the second communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF.

[0188] The first communication device or the second communication device can be a network element with one or more functions including the network data analysis function NWDAF, the analysis database function ADRF and the positioning management function LMF. The first communication device or the second communication device can also be a 5GC network element with other functions.

[0189] The first communication device can update the model in batches by storing the processing identifier and the model information, thereby saving model update time and improving model update efficiency.

[0190] FIG5 is a flow chart of another communication method provided by an embodiment. As shown in FIG5 , the communication method according to the embodiment of the present application is applied to a second communication device. The method includes S410 to S420:

[0191] S410: Send a model management storage request to the first communication device, so that the first communication device updates the model according to the model management storage request.

[0192] When a new model is trained or a model update is required, the second communication device sends a model management storage request to the first communication device, instructing the first communication device to update the model. The model management storage request includes a storage processing identifier, and the first communication device updates the model according to the model management storage request.

[0193] S420: Receive a second result indication sent by the first communication device.

[0194] The second result indication is used to indicate whether the model update is successful or failed.

[0195] After completing the model update, the first communication device generates a second result indication, indicating whether the model update is successful or failed. The first communication device feeds back the second result indication to the second communication device, and the second communication device can determine whether the model update is successful after receiving the second result indication. Among them, a successful update can mean that all models are updated successfully, or that part of the model is updated successfully; a failed model update can mean that all updates fail, or that part of the model fails; for example, when some models are updated successfully and some models fail to update, the second result indication indicates that the model update failed, or the second result indicates the specific models that have been updated successfully, and the models that have failed to update. The second communication device can determine which models have been updated successfully and which models have failed to update based on the second result indication.

[0196] In the communication method provided in the embodiment of the present application, the second communication device sends a model management storage request to the first communication device, and the model management storage request includes a storage processing identifier. The first communication device can directly update the model in batches according to the model management storage request without having to interact with the second communication device multiple times. The first communication device directly updates the model based on the instruction of the model update request, with high update efficiency and better effect. Compared with the related technology, multiple interactions are required to realize model update, and the speed is faster. In addition, batch update can realize the update of multiple models at a time, effectively saving the communication resources of the first communication device and the second communication device.

[0197] In some embodiments, the model management storage request also includes model information.

[0198] In some embodiments, a new storage processing identifier generated and sent by the first communication device after completing the model update is received.

[0199] In some embodiments, the model information includes one or more of the following:

[0200] One or more updated model files; one or more addresses of updated model files; one or more model identifiers; one or more updated model storage space sizes.

[0201] In some embodiments, the first communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF; in some embodiments, the second communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF.

[0202] The first communication device or the second communication device can be a network element with one or more functions including the network data analysis function NWDAF, the analysis database function ADRF and the positioning management function LMF. The first communication device or the second communication device can also be a 5GC network element with other functions.

[0203] The communication process is described through the following examples:

[0204] Example 1

[0205] A new service operation, Nadrf_MLModelManagement_UpdateRequest, has been added for model updates. A second device uses this request to request a first device to update a stored model, such as a machine learning model. The model file or the address of the model file stored on the second device is provided to the first device in the model update request message.

[0206] Solution 1 (traditional delete, save and update):

[0207] The model update request includes the NF ID of the second communication device and update indication information, and the update indication information is used to instruct the first communication device to delete or update the model.

[0208] The model update request also includes one or more of the following information:

[0209] The storage processing identifier received during storage; one or more sets of updated model files and corresponding model identifiers; one or more sets of model identifiers, the addresses of the updated model files (such as URLs or FQDNs) and the updated model storage space size.

[0210] Option 2 (model parameter update):

[0211] The model update request includes the NF ID of the second communication device; the model update request also includes one or more of the following information:

[0212] One or more sets of model parameters (such as newly added or reduced classification trees in the model), and the model identifier of the corresponding model to be updated; one or more sets of model parameters, model parameter types and the model identifier of the corresponding model to be updated; one or more sets of model parameter files and the model identifier of the corresponding model to be updated; one or more sets of model identifiers, the address of the model update parameter file (such as URL or FQDN), and the estimated storage space size of the updated model.

[0213] Solution 3 (model fusion update):

[0214] The model update request includes the NF ID of the second communication device and model fusion update indication information, and the model fusion update indication information is used to instruct the first communication device to perform model fusion update.

[0215] The model update request also includes one or more of the following information:

[0216] One or more model files to be merged and updated that need to be merged and updated on the original model and the model identifiers of one or more corresponding models to be updated; one or more groups of model identifiers, the addresses of the model files to be merged and updated (such as URL or FQDN), and the estimated size of the updated model storage space.

[0217] After completing the model update, the first communication device sends at least one of the following information to the second communication device:

[0218] The first result indication is used to indicate whether one or more models are updated successfully or failed; the model identifier of the updated model; the address of the updated model file; and the storage processing identifier.

[0219] After the model is successfully updated, the address (eg, URL or FQDN) of the updated model file may be generated.

[0220] If solution three is used to perform model fusion update, a new model identifier will be generated after the model update is successful.

[0221] After each round of model update, a new storage processing identifier may be output to identify the updated model.

[0222] Example 2

[0223] Figure 6 provides an example flow chart of a model storage update, taking the second communication device sending a model update request to the first communication device as an example to illustrate the process of model update. The second communication device may be NWDAF or other 5GC NF, and the first communication device may be ADRF.

[0224] 1. The second communication device stores one or more models in the first communication device.

[0225] 2. The first communication device stores one or more models.

[0226] 3. The first communication device returns a storage processing identifier to the second communication device.

[0227] Note: Operations 4, 5, and 6 can only be performed if the first communication device returns a storage processing identifier. Operations 1, 2, and 3 are prerequisites for operations 4, 5, and 6.

[0228] 4. The second communication device sends a model update request to the first communication device. The information carried can refer to the above embodiment.

[0229] 5. The first communication device updates the model locally.

[0230] If the information carried in the model update request adopts solution 1, the first communication device needs to take the following actions:

[0231] The first communication device deletes one or more local models according to the received one or more model identifiers, and stores the received one or more updated model files or the one or more updated model files downloaded from the address of the updated model files to complete the update.

[0232] Alternatively, the first communication device retrieves one or more local models to be updated through the received storage processing identifier, deletes the local model, and stores the received updated model file or one or more updated model files downloaded from the address of the updated model file to complete the update.

[0233] If the information carried in the model update request adopts solution 2, the first communication device needs to take the following actions:

[0234] The first communication device retrieves one or more local models to be updated based on the received one or more model identifiers, and updates the local model based on the received model parameter update (obtained directly from the second communication device or downloaded from a link provided by the second communication device) and the corresponding model identifiers.

[0235] If the input parameters adopt solution 3, the first communication device needs to take the following actions:

[0236] The first communication device retrieves one or more local models according to the received one or more model identifiers.

[0237] Obtain the model file to be merged and updated from the second communication device or download it from the provided link; perform a one-to-one fusion of the received model file to be merged and updated and the local model, and update the stored model; generate a new model identifier after fusion.

[0238] 6. The first communication device replies to the model update request, and the information carried is the same as that in the above embodiment.

[0239] If solution three is adopted, the first communication device transmits the newly generated model identifier after fusion back to the second communication device.

[0240] Example 3

[0241] FIG7 provides an example diagram of a process for updating a model subscription.

[0242] 1. The first communication device requests the trained model from the second communication device.

[0243] The first communication device is a consumer NF, which can be LMF, NWDAF with AnLF or other 5GC NF; the second communication device can be NWDAF with MTLF.

[0244] 2. The second communication device sends the trained model to the first communication device. This response carries one or more model identifiers.

[0245] 3. The first communication device uses the received model to perform inference

[0246] 4. The first communication device finds that the received model inference result does not meet expectations.

[0247] 5. The first communication device sends a message to the second communication device indicating that the model inference result did not meet expectations (optional)

[0248] 6. The second communication device prepares to update the model.

[0249] 7. The second communication device returns a model update request.

[0250] The information carried in the model update request in operation 7 can refer to the above embodiment:

[0251] 8. The first communication device updates the local model.

[0252] If solution 1 is adopted, the first communication device needs to take the following actions:

[0253] The first communication device deletes one or more local models according to the received one or more model identifiers, and stores the received one or more updated model files or the one or more updated model files downloaded from the address of the updated model file to complete the update.

[0254] When the first communication device is in ADRF, it can also be updated through the storage processing identifier. When the first communication device is LMF, NWDAF with AnLF or other 5GC NF, there is no storage processing identifier at this time, and it can be updated through the model identifier.

[0255] If solution 2 is adopted, the first communication device needs to take the following actions:

[0256] The first communication device retrieves one or more local models based on the received one or more model identifiers and updates the local models based on one or more sets of model parameter updates received (obtained directly from the second communication device or downloaded from a link provided by the second communication device) and the corresponding model identifiers.

[0257] If the input parameters adopt solution 3, the first communication device needs to take the following actions:

[0258] The first communication device retrieves one or more local models according to the received one or more model identifiers.

[0259] The model file to be merged and updated is obtained from the second communication device or downloaded from a provided link.

[0260] The received model file to be fused and updated is fused with the local model on a one-to-one basis to update the stored model.

[0261] After fusion, a new model identity is generated.

[0262] 9. The first communication device sends a first result indication to the second communication device, where the first result indication is used to indicate whether one or more models are updated successfully or failed.

[0263] After the model update is complete, determine the location of the updated model file.

[0264] If solution three is adopted, the first communication device generates one or more new model identifiers after completing the model update, and the first communication device sends the generated new model identifiers to the second communication device.

[0265] Model updates don't necessarily need to be initiated by the first communication device; they can also be initiated by the second communication device, for example, periodic model updates (every hour, once a day). If periodic model updates are used, the frequency of model updates needs to be negotiated between the first and second communication devices. Therefore, steps 4 and 5 in this embodiment are optional.

[0266] Example 4

[0267] Reuse operations and add some parameters to support batch update of models.

[0268] Service Operation Name: Nadrf_MLModelManagement_StorageRequest; This service operation is used by a second communication device to request a first communication device to store or update a model, such as a machine learning model. The address of the model file or model location stored on the second communication device is provided to the first communication device in the request message.

[0269] The model management storage request is Nadrf_MLModelManagement_StorageRequest.

[0270] The model management storage request includes the NF ID of the second communication device, a storage processing identifier, and the following set of contents:

[0271] Model file; or model identifier, address of updated model file (such as URL or FQDN), and updated model storage space size.

[0272] After the model update is completed, the first communication device generates a second result indication, where the second result indication is used to indicate whether the model update is successful or failed).

[0273] After the model is updated successfully, the first communication device may generate an address (eg, URL or FQDN) of the updated model file.

[0274] After the model update is completed, the first communication device may generate a new storage processing identifier, and output a new storage processing identifier after each round of model update to identify the model updated this time.

[0275] Figure 8 provides an example flow chart of a model storage update process, wherein the first communication device may be an ADRF, and the second communication device may be an NWDAF or other 5GC NF.

[0276] 1. The second communication device stores one or more models in the first communication device.

[0277] 2. The first communication device stores one or more models.

[0278] 3. The first communication device returns a storage processing identifier.

[0279] Note: Operations 4, 5, and 6 can only be performed if the first communication device returns a storage processing identifier.

[0280] 4. The second communication device needs to update the model in batches, sends the storage processing identifier in operation 3 to the first communication device, and sends the updated model file or tuple (the address of the updated model file and the model storage space size required for each updated model) to the first communication device.

[0281] 5. The first communication device retrieves one or more local models to be updated using the received storage processing identifier, deletes the local models, and stores the received updated model files or one or more updated model files downloaded from the updated model file address, completing the update.

[0282] 6. The first communication device issues a second result indication and a new storage processing identifier, where the second result indication is used to indicate whether the model update is successful or failed.

[0283] FIG9 is a schematic structural diagram of a communication apparatus provided by an embodiment. The apparatus is applied to a first communication device. As shown in FIG9 , the apparatus includes: a model update request receiving module 510 and a first model updating module 520 .

[0284] The model update request receiving module 510 is used to receive a model update request sent by the second communication device, wherein the model update request is used to instruct the first communication device to update the stored model; the first model updating module 520 is used to update the model according to the model update request.

[0285] In the communication method provided in the embodiment of the present application, a first communication device receives a model update request sent by a second communication device, and the model update request is used to instruct the first communication device to update a stored model; the first communication device can directly update the model according to the model update request without having to interact with the second communication device multiple times. The first communication device directly updates the model based on the instruction of the model update request, and the update efficiency is high and the effect is better. Compared with the related technology, the model update can be achieved through multiple interactions, the speed is faster, and the communication resources of the first communication device and the second communication device are effectively saved.

[0286] In some embodiments, the model update request includes first model update information; and updating the model according to the model update request includes:

[0287] Determine one or more local models to be updated according to the first model update information, delete the one or more local models to be updated, and store one or more updated model files.

[0288] In some embodiments, the first model update information includes one or more of the following:

[0289] Update indication information; one or more updated model files; addresses of one or more updated model files; one or more model identifiers; one or more updated model storage space sizes; storage processing identifier.

[0290] In some embodiments, the model update request includes second model update information; and performing a model update according to the model update request includes:

[0291] One or more local models to be updated are determined according to the second model update information; and the local models are updated according to the model parameters corresponding to each local model.

[0292] In some embodiments, the second model update information includes one or more of the following:

[0293] One or more model parameters; one or more model parameter types; one or more model update parameter files; one or more model identifiers; the addresses of one or more model update parameter files; and the estimated storage space size of one or more updated models.

[0294] In some embodiments, the model update parameter file includes one or more model parameters.

[0295] In some embodiments, the model update request includes third model update information; and updating the model according to the model update request includes:

[0296] Determine one or more local models to be updated and a model file to be fused and updated according to the third model update information; fuse the model file to be fused and updated with each local model respectively to obtain a fused model and store it.

[0297] In some embodiments, the third model update information includes one or more of the following:

[0298] Model fusion update indication information; one or more model files to be fused and updated; one or more model identifiers; the addresses of one or more model files to be fused and updated; and the estimated storage space size of one or more updated models.

[0299] In some embodiments, the apparatus further comprises:

[0300] The model identifier generation module is used to generate a new model identifier after the fusion is completed, and send the new model identifier to the second communication device.

[0301] In some embodiments, the device is further configured to:

[0302] After the model update is completed, at least one of the following information is sent to the second communication device:

[0303] A first result indication, wherein the first result indication is used to indicate that one or more models are updated successfully or failed; a model identifier of the updated model; an address of the updated model file; and a storage processing identifier.

[0304] In some embodiments, the first communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF; in some embodiments, the second communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF.

[0305] The communication device proposed in this embodiment and the communication method proposed in the above embodiment belong to the same concept. For technical details not fully described in this embodiment, please refer to any of the above embodiments, and this embodiment has the same effect as executing the communication method.

[0306] Figure 10 is a structural diagram of another communication device provided by an embodiment. The device is applied to a second communication device. As shown in Figure 10, the device includes: a model update request sending module 610; wherein, the model update request sending module 610 is used to send a model update request to a first communication device, so that the first communication device performs a model update according to the model update request; wherein, the model update request is used to instruct the first communication device to update the stored model.

[0307] In the communication method provided in the embodiment of the present application, the second communication device sends a model update request to the first communication device, and the model update request is used to instruct the first communication device to update the stored model; the first communication device can directly update the model according to the model update request without having to interact with the second communication device multiple times. The first communication device directly updates the model based on the instruction of the model update request, and the update efficiency is high and the effect is better. Compared with the related technology, the model update can be achieved through multiple interactions, the speed is faster, and the communication resources of the first communication device and the second communication device are effectively saved.

[0308] In some embodiments, the model update request includes first model update information.

[0309] In some embodiments, the first model update information includes one or more of the following:

[0310] Update indication information; one or more updated model files; addresses of one or more updated model files; one or more model identifiers; one or more updated model storage space sizes; storage processing identifier.

[0311] In some embodiments, the model update request includes second model update information.

[0312] In some embodiments, the second model update information includes one or more of the following:

[0313] One or more model parameters; one or more model parameter types; one or more model update parameter files; one or more model identifiers; the addresses of one or more model update parameter files; and the estimated storage space size of one or more updated models.

[0314] In some embodiments, the model update parameter file includes one or more model parameters.

[0315] In some embodiments, the model update request includes third model update information.

[0316] In some embodiments, the third model update information includes one or more of the following:

[0317] Model fusion update indication information; one or more model files to be fused and updated; one or more model identifiers; the addresses of one or more model files to be fused and updated; and the estimated storage space size of one or more updated models.

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

[0319] The model identification receiving module is used to receive a new model identification sent by the first communication device.

[0320] In some embodiments, the device is used to:

[0321] Receive at least one of the following information sent by the first communication device:

[0322] A first result indication, where the first result indication is used to indicate whether one or more models are updated successfully or failed; a model identifier of the updated model; an address of the updated model file; and a storage processing identifier.

[0323] In some embodiments, the first communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF; in some embodiments, the second communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF.

[0324] In some embodiments, the apparatus further comprises:

[0325] The update condition detection module is used to detect whether the model update condition is met before generating the model update request.

[0326] The communication device proposed in this embodiment and the communication method proposed in the above embodiment belong to the same concept. For technical details not fully described in this embodiment, please refer to any of the above embodiments, and this embodiment has the same effect as executing the communication method.

[0327] Figure 11 is a structural diagram of another communication device provided by an embodiment, which is applied to a first communication device. As shown in Figure 11, the device includes: a model management storage request receiving module 710, a second model updating module 720 and a second result indication sending module 730; wherein, the model management storage request receiving module 710 is used to receive a model management storage request sent by the second communication device, and the model management storage request includes a storage processing identifier; the second model updating module 720 is used to update the model according to the model management storage request; the second result indication sending module 730 is used to send a second result indication to the second communication device; wherein, the second result indication is used to indicate whether the model update is successful or failed.

[0328] In the communication method provided in the embodiment of the present application, a first communication device receives a model management storage request sent by a second communication device, and the model management storage request includes a storage processing identifier. The first communication device can directly update the model in batches according to the model management storage request without having to interact with the second communication device multiple times. The first communication device directly updates the model based on the instruction of the model update request, with high update efficiency and better effect. Compared with related technologies that require multiple interactions to achieve model updates, the speed is faster, and batch updates can update multiple models at a time, effectively saving communication resources of the first communication device and the second communication device.

[0329] In some embodiments, the model management storage request further includes model information, and updating the model according to the model management storage request includes:

[0330] One or more local models are deleted according to the storage processing identifier, and the updated model file is stored according to the model information.

[0331] In some embodiments, the apparatus further comprises:

[0332] The storage processing identifier sending module is used to generate a new storage processing identifier after completing the model update, and send the new storage processing identifier to the second communication device.

[0333] In some embodiments, the model information includes one or more of the following:

[0334] One or more updated model files; one or more addresses of updated model files; one or more model identifiers; one or more updated model storage space sizes.

[0335] In some embodiments, the first communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF; in some embodiments, the second communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF.

[0336] The communication device proposed in this embodiment and the communication method proposed in the above embodiment belong to the same concept. For technical details not fully described in this embodiment, please refer to any of the above embodiments, and this embodiment has the same effect as executing the communication method.

[0337] FIG12 is a schematic structural diagram of another communication apparatus provided by an embodiment. The apparatus is applied to a second communication device. As shown in FIG12 , the apparatus includes: a model management storage request sending module 810 and a second result indication receiving module 820 .

[0338] The model management storage request sending module 810 is used to send a model management storage request to the first communication device so that the first communication device updates the model according to the model management storage request; the second result indication receiving module 820 is used to receive the second result indication sent by the first communication device; wherein, the model management storage request includes a storage processing identifier, and the second result indication is used to indicate whether the model update is successful or failed.

[0339] In the communication method provided in the embodiment of the present application, the second communication device sends a model management storage request to the first communication device, and the model management storage request includes a storage processing identifier. The first communication device can directly update the model in batches according to the model management storage request without having to interact with the second communication device multiple times. The first communication device directly updates the model based on the instruction of the model update request, with high update efficiency and better effect. Compared with the related technology, multiple interactions are required to realize model update, and the speed is faster. In addition, batch update can realize the update of multiple models at a time, effectively saving the communication resources of the first communication device and the second communication device.

[0340] In some embodiments, the model management storage request further includes model information.

[0341] In some embodiments, the apparatus further comprises:

[0342] The storage processing identifier receiving module is used to receive a new storage processing identifier generated and sent by the first communication device after completing the model update.

[0343] In some embodiments, the model information includes one or more of the following:

[0344] One or more updated model files; one or more addresses of updated model files; one or more model identifiers; one or more updated model storage space sizes.

[0345] In some embodiments, the first communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF; in some embodiments, the second communication device is a network element with one or more of the following functions: network data analysis function NWDAF, analysis database function ADRF, positioning management function LMF.

[0346] The communication device proposed in this embodiment and the communication method proposed in the above embodiment belong to the same concept. For technical details not fully described in this embodiment, please refer to any of the above embodiments, and this embodiment has the same effect as executing the communication method.

[0347] An embodiment of the present application also provides a communication node. Figure 13 is a structural diagram of a communication node provided by an embodiment. As shown in Figure 13, the communication node provided by the present application includes a memory 920, a processor 910, and a computer program stored in the memory and executable on the processor. When the processor 910 executes the program, the above-mentioned communication method is implemented.

[0348] The communication node may also include a memory 920; the processor 910 in the communication node may be one or more, and Figure 13 takes one processor 910 as an example; the memory 920 is used to store one or more programs; the one or more programs are executed by the one or more processors 910, so that the one or more processors 910 implement the communication method as described in the embodiment of the present application.

[0349] The communication node further includes: a communication module 930 , an input device 940 and an output device 950 .

[0350] The processor 910 , memory 920 , communication module 930 , input device 940 and output device 950 in the communication node may be connected via a bus or other means. FIG13 takes the bus connection as an example.

[0351] The input device 940 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the communication node. The output device 950 may include a display device such as a display screen.

[0352] The communication module 930 may include a receiver and a transmitter. The communication module 930 is configured to perform information transmission and reception communication according to the control of the processor 910.

[0353] The memory 920, as a computer-readable storage medium, can be configured to store software programs, computer executable programs, and modules, such as program instructions / modules corresponding to the communication method described in the embodiments of the present application (for example, the model update request receiving module 510 and the first model update module 520 in the communication device, or the model update request sending module 610 in the communication device, or the model management storage request receiving module 710, the second model update module 720, and the second result indication sending module 730 in the communication device, or the model management storage request sending module 810 and the second result indication receiving module 820 in the communication device). The memory 920 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; and the data storage area may store data created according to the use of the communication node, etc. In addition, the memory 920 may include a high-speed random access memory and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 920 may further include a memory remotely located relative to the processor 910, and these remote memories may be connected to the communication node via a network. Examples of the aforementioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0354] An embodiment of the present application further provides a storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, it implements any communication method described in the embodiments of the present application.

[0355] Optionally, the communication method is applied to a first communication device, including: receiving a model update request sent by a second communication device, the model update request being used to instruct the first communication device to update a stored model; and performing a model update according to the model update request.

[0356] Optionally, the communication method, applied to the second communication device, includes: sending a model update request to the first communication device, so that the first communication device updates the model according to the model update request; wherein the model update request is used to instruct the first communication device to update the stored model.

[0357] Optionally, the communication method, applied to a first communication device, includes: receiving a model management storage request sent by a second communication device, the model management storage request including a storage processing identifier; updating the model according to the model management storage request; and sending a second result indication to the second communication device; wherein the second result indication is used to indicate whether the model update is successful or failed.

[0358] Optionally, the communication method, applied to the second communication device, includes: sending a model management storage request to the first communication device, so that the first communication device updates the model according to the model management storage request; receiving a second result indication sent by the first communication device; wherein the model management storage request includes a storage processing identifier, and the second result indication is used to indicate whether the model update is successful or failed.

[0359] The computer storage medium of the embodiment of the present application can adopt any combination of one or more computer-readable media.Computer-readable media can be computer-readable signal media or computer-readable storage media.Computer-readable storage media can be, for example, but not limited to: electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any combination of the above.More specific examples (non-exhaustive list) of computer-readable storage media include: electrical connections with one or more wires, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM), flash memories, optical fibers, portable CD-ROMs, optical storage devices, magnetic storage devices, or any suitable combination of the above.Computer-readable storage media can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0360] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0361] The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wire, optical cable, radio frequency (RF), etc., or any suitable combination of the foregoing.

[0362] The computer program code for performing the operations of the present application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and also conventional procedural programming languages ​​such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., using an Internet service provider to connect via the Internet).

[0363] The above description is merely an exemplary embodiment of the present application and is not intended to limit the scope of protection of the present application.

[0364] It will be appreciated by those skilled in the art that the term user terminal covers any suitable type of wireless user equipment, such as a mobile phone, a portable data processing device, a portable web browser or a vehicle-mounted mobile station.

[0365] In general, various embodiments of the present application may be implemented in hardware or dedicated circuits, software, logic, or any combination thereof. For example, some aspects may be implemented in hardware, while other aspects may be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device, although the present application is not limited thereto.

[0366] Embodiments of the present application may be implemented by executing computer program instructions by a data processor of a mobile device, for example, in a processor entity, or by hardware, or by a combination of software and hardware. The computer program instructions may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages.

[0367] The block diagram of any logical flow in the drawings of this application may represent program operations, or may represent interconnected logical circuits, modules and functions, or may represent a combination of program operations and logical circuits, modules and functions. A computer program may be stored on a memory. The memory may be of any type suitable for the local technical environment and may be implemented using any suitable data storage technology, such as but not limited to read-only memory (ROM), random access memory (RAM), optical storage devices and systems (digital versatile discs (DVD) or compact disks (CD), etc.). Computer-readable media may include non-transitory storage media. The data processor may be of any type suitable for the local technical environment, such as but not limited to a general-purpose computer, a special-purpose computer, a microprocessor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a programmable logic device (FPGA) and a processor based on a multi-core processor architecture.

Claims

1. A communication method, applied to a first communication device, comprising: receiving a model update request sent by a second communication device, where the model update request is used to instruct the first communication device to update a stored model; Perform a model update according to the model update request.

2. The communication method according to claim 1, wherein: The model update request includes first model update information; Updating the model according to the model update request includes: At least one local model to be updated is determined according to the first model update information, the at least one local model to be updated is deleted, and at least one updated model file is stored.

3. The communication method according to claim 2, wherein: The first model update information includes at least one of the following: Update instructions; At least one updated model file; The address of at least one updated model file; At least one model identifier; At least one updated model storage size; Storage processing identifier. The communication method according to claim 1 , wherein: The model update request includes second model update information; Updating the model according to the model update request includes: determining at least one local model to be updated according to the second model update information; The local models are updated according to the model parameters corresponding to each local model. The communication method according to claim 4 , wherein: The second model update information includes at least one of the following: at least one model parameter; At least one model parameter type; At least one model update parameter file; At least one model identifier; The address of at least one model update parameter file; At least one updated estimated model storage size. The communication method according to claim 5 , wherein: The model update parameter file contains at least one model parameter.

7. The communication method according to claim 1, wherein: The model update request includes third model update information; Updating the model according to the model update request includes: Determining at least one local model to be updated and a model file to be merged and updated according to the third model update information; The model file to be fused and updated is fused with each local model respectively to obtain a fused model and store it.

8. The communication method according to claim 7, wherein: The third model update information includes at least one of the following: Model fusion update indication information; At least one model file to be merged and updated; At least one model identifier; The address of at least one model file to be merged and updated; At least one updated estimated model storage size.

9. The communication method according to claim 7, further comprising: After the fusion is completed, a new model identifier is generated and sent to the second communication device.

10. The communication method according to any one of claims 1 to 9, further comprising: After the model update is completed, at least one of the following information is sent to the second communication device: a first result indication, the first result indication being used to indicate that at least one model update is successful or failed; The model ID of the updated model; The address of the updated model file; Storage processing identifier.

11. The communication method according to any one of claims 1 to 9, wherein: The first communication device is a network element having at least one of the following functions: a network data analysis function NWDAF, an analysis database function ADRF, and a location management function LMF; The second communication device is a network element having at least one of the following functions: a network data analysis function NWDAF, an analysis database function ADRF, and a positioning management function LMF.

12. A communication method, applied to a second communication device, comprising: Sending a model update request to the first communication device, so that the first communication device updates the model according to the model update request; The model update request is used to instruct the first communication device to update the stored model.

13. The communication method according to claim 12, wherein: The model update request includes first model update information.

14. The communication method according to claim 13, wherein: The first model update information includes at least one of the following: Update instructions; At least one updated model file; The address of at least one updated model file; At least one model identifier; At least one updated model storage size; Storage processing identifier.

15. The communication method according to claim 12, wherein: The model update request includes second model update information. The communication method according to claim 15 , wherein: The second model update information includes at least one of the following: at least one model parameter; At least one model parameter type; At least one model update parameter file; At least one model identifier; The address of at least one model update parameter file; At least one updated estimated model storage size.

17. The communication method according to claim 16, wherein: The model update parameter file contains at least one model parameter.

18. The communication method according to claim 12, wherein: The model update request includes third model update information.

19. The communication method according to claim 18, wherein: The third model update information includes at least one of the following: Model fusion update indication information; At least one model file to be merged and updated; At least one or more model identifiers; The address of at least one model file to be merged and updated; At least one updated estimated model storage size.

20. The communication method according to claim 18, further comprising: Receive a new model identifier sent by the first communication device.

21. The communication method according to any one of claims 12 to 20, further comprising: Receive at least one of the following information sent by the first communication device: a first result indication, the first result indication being used to indicate that at least one model update is successful or failed; The model ID of the updated model; The address of the updated model file; Storage processing identifier.

22. The communication method according to any one of claims 12 to 20, wherein: The first communication device is a network element having at least one of the following functions: a network data analysis function NWDAF, an analysis database function ADRF, and a location management function LMF; The second communication device is a network element having at least one of the following functions: a network data analysis function NWDAF, an analysis database function ADRF, and a positioning management function LMF.

23. The communication method according to any one of claims 12 to 20, wherein: Before sending the model update request to the first communication device, the method further includes: Check whether the model update conditions are met.

24. A communication method, applied to a first communication device, comprising: receiving a model management storage request sent by a second communication device, wherein the model management storage request includes a storage processing identifier; Performing model update according to the model management storage request; sending a second result indication to the second communication device; The second result indication is used to indicate whether the model update is successful or failed.

25. The communication method according to claim 24, wherein: The model management storage request further includes model information, and updating the model according to the model management storage request includes: At least one local model is deleted according to the storage processing identifier, and the updated model file is stored according to the model information.

26. The communication method according to claim 25, wherein: The model information includes at least one of the following: At least one or more updated model files; The address of at least one or more updated model files; At least one or more model identifiers; At least one updated model storage size.

27. The communication method according to claim 24, further comprising: After the model update is completed, a new storage processing identifier is generated and sent to the second communication device.

28. The communication method according to any one of claims 24 to 27, wherein: The first communication device is a network element having at least one of the following functions: a network data analysis function NWDAF, an analysis database function ADRF, and a location management function LMF; The second communication device is a network element having at least one of the following functions: a network data analysis function NWDAF, an analysis database function ADRF, and a positioning management function LMF.

29. A communication method, applied to a second communication device, comprising: sending a model management storage request to the first communication device, so that the first communication device updates the model according to the model management storage request; receiving a second result indication sent by the first communication device; The model management storage request includes a storage processing identifier, and the second result indication is used to indicate whether the model update is successful or failed.

30. The communication method according to claim 29, wherein: The model management storage request also includes model information.

31. The communication method according to claim 30, wherein: The model information includes at least one of the following: At least one updated model file; The address of at least one updated model file; At least one or more model identifiers; At least one updated model storage size.

32. The communication method according to claim 29, further comprising: Receive a new storage processing identifier generated and sent by the first communication device after completing the model update.

33. The communication method according to any one of claims 29 to 32, wherein: The first communication device is a network element having at least one of the following functions: a network data analysis function NWDAF, an analysis database function ADRF, and a location management function LMF; The second communication device is a network element having at least one of the following functions: a network data analysis function NWDAF, an analysis database function ADRF, and a positioning management function LMF.

34. A communication node comprising: A memory, a processor, a program stored in the memory and executable on the processor, and a data bus configured to implement connection communication between the processor and the memory, wherein the program, when executed by the processor, implements the communication method according to any one of claims 1 to 33.

35. A storage medium configured as a computer-readable storage medium, wherein the storage medium stores at least one program, and the at least one program can be executed by at least one processor to implement the communication method according to any one of claims 1 to 33.

Images