Method and device for optimizing recommendation model

Abstract

The invention provides a recommendation model optimization method and device, and relates to the technical field of artificial intelligence. In the method, a second network element can send an expected target of a user for a recommendation model to a first network element through target information, and the first network element can determine a target parameter capable of influencing the implementation of the expected target and a parameter index corresponding to the target parameter according to the received target information and send the target parameter and the parameter index to the second network element. Wherein the parameter index is used for indicating the ability of the target parameter to realize the expected target. The second network element executes the target parameter according to the parameter index to obtain an execution result, and then the first network element and / or the second network element can determine whether to update the recommendation model or not according to whether the execution result reaches an expected target indicated by the target information or not. Therefore, the recommendation model can be updated according to the requirements of the user, the optimization efficiency of the recommendation model is improved, and the satisfaction degree of the user on the recommendation information generated by the recommendation model is improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence (AI) technology, and in particular to a method and apparatus for optimizing a recommendation model. Background Technology

[0002] Currently, emerging AI models have demonstrated powerful capabilities in knowledge extraction and reasoning. Consequently, exploration based on AI models is rapidly progressing in various fields such as computer vision, AI science, healthcare, and robotics, and numerous AI models with diverse functionalities are already in use in daily life. However, for a class of AI models used to generate recommendation strategies, the fact that user preferences are not static sometimes leads to recommendation strategies that fail to meet user needs. Summary of the Invention

[0003] This application provides a method and apparatus for optimizing a recommendation model. By analyzing the expected goals of the recommendation model and the corresponding execution results, it determines whether to update the recommendation model so that the model better meets the needs of users.

[0004] To achieve the above objectives, this application adopts the following technical solution:

[0005] Firstly, an optimization method for the recommendation model is provided, which can be executed by a first network element. Here, the first network element can refer to the first network element itself, or to the processor, circuit, module, logic node, chip, or chip system within the first network element that implements the method.

[0006] The method includes: receiving a first message, the first message including target information, the target information being used to indicate the expected target of the recommendation model; sending first recommendation information, the first recommendation information indicating target parameters and corresponding parameter indicators, the target parameters and parameter indicators being obtained based on the target information; the target parameters being used to achieve the expected target, the parameter indicators being used to indicate the ability of the target parameters to achieve the expected target; and receiving an execution result, the execution result being obtained based on the first recommendation information, the execution result being used to determine whether to update the recommendation model.

[0007] Based on the method provided in the first aspect above, the first network element determines the target parameters and corresponding parameter indicators that can be used to achieve the expected goal according to the received target information. The parameter indicators indicate which parameters among the target parameters, when executed, are more likely to achieve the expected goal indicated by the target information. The network element receiving the first recommendation information, such as the second network element, can execute the aforementioned target parameters according to the parameter indicators to obtain the execution result. The first network element can determine whether to update the aforementioned recommendation model based on whether the execution result meets the expected goal. In this way, the target information and execution results corresponding to the recommendation model can be analyzed to determine whether to update the recommendation model, making the execution results corresponding to the target parameters generated by the recommendation model increasingly closer to the target information and better meeting the user's needs.

[0008] As one possible implementation, the target information includes at least one of the following: experience quality information, network function attribute information, or network function service range information. Based on this, the first network element can determine target parameters and corresponding parameter indicators according to multiple target information. For the execution results of these multiple target parameters, it can analyze whether to update the model, thereby optimizing the recommendation model to meet the user's various expected goals and satisfying the user's needs from all aspects.

[0009] As one possible implementation, the first message also includes candidate parameters, which are related to the expected target and are used by the first network element to determine the target parameters. Based on this, the first network element can also refer to the candidate parameters when determining the target parameters according to the target information. For example, the candidate parameters can be used as the target parameters. Alternatively, one or more parameters can be added to the candidate parameters to obtain the target parameters. Referring to candidate parameters can reduce the overhead of the first network element in determining the target parameters and save time costs associated with optimizing the model.

[0010] As one possible implementation, the target parameters include service quality parameters, or at least one of the network element candidate list parameters; the parameter indicators include at least one of the expected goal achievement degree, expected goal relevance, or target parameter priority corresponding to the target parameters. Based on this, at least one of the expected goal achievement degree, expected goal relevance, and target parameter priority can be used to indicate that the service quality parameters can be used to indicate the ability to achieve the expected goal. Thus, the second network element is instructed to execute the target parameters according to the parameter indicators to obtain the execution result.

[0011] As one possible implementation, the method further includes receiving a second message, which indicates the receipt of execution results. Based on this, the first network element can receive the execution results obtained from executing the target parameters after receiving the indication of the second message. That is, after sending the target parameters and their corresponding parameter metrics, the first network element triggers the collection of execution results for the target parameters by receiving the second message, so as to determine whether to update the recommendation model based on the execution results.

[0012] As one possible implementation, the method further includes: receiving a third message, which indicates that at least one of the target information or candidate parameters should be updated. The third message includes one or more of the following: updated target information or updated candidate parameters. Based on this, the first network element can obtain new target information and candidate parameters according to the received updated target information or updated candidate parameters, and then redetermine the target parameters and the corresponding parameter indices based on the new target information and candidate parameters. In this way, even if the target information or candidate parameters change, partially updated target information or updated candidate parameters can be sent, thereby saving transmission resources and improving the efficiency of model training.

[0013] As one possible implementation, the method further includes: determining to update the recommendation model when the execution result and target information meet a first preset condition; wherein, meeting the first preset condition includes at least one of the following: the difference between the execution result and the target information is greater than or equal to a first preset threshold; or, the ratio of the execution result to the target information is less than or equal to a second preset threshold. Based on this, the first network element can analyze the execution result and target information, and update the recommendation model when the difference between the execution result and the target information is large. This makes the execution result corresponding to the recommendation information generated by the recommendation model closer to the target information and better meets the user's needs. When the difference between the execution result and the target information is small, the execution result can basically meet the user's needs, and the recommendation model is not updated.

[0014] As one possible implementation, the method further includes receiving a fourth message, which instructs the first network element to update the recommendation model used to generate the first recommendation information. Based on this, the first network element can determine to update the recommendation model after receiving the instruction of the fourth message. In other words, the fourth message can trigger an update to the recommendation model, making the updated recommendation model better meet user needs.

[0015] Secondly, an optimization method for the recommendation model is provided, which can be executed by a second network element. Here, the second network element can refer to the second network element itself, or to a processor, circuit, module, logic node, chip, or chip system within the second network element that implements the method.

[0016] The method includes: sending a first message, the first message including target information, the target information being used to indicate the expected target of the recommendation model; receiving first recommendation information, the first recommendation information indicating target parameters and corresponding parameter indicators, the target parameters and parameter indicators being obtained based on the target information, the target parameters being used to achieve the expected target, and the parameter indicators being used to indicate the ability of the target parameters to achieve the expected target; executing the target parameters based on the parameter indicators to obtain the corresponding execution result; and sending the execution result, the execution result being used to determine whether to update the recommendation model.

[0017] Based on the method provided in the second aspect above, the second network element sends target information corresponding to the expected target of the recommendation model. The second network element then executes the target parameters according to the received target parameters to obtain the corresponding execution result. Whether the execution result meets the expected target can be used to determine whether to update the recommendation model. In this way, by setting expected targets for the recommendation model and analyzing the target information and execution results corresponding to the expected targets, it is possible to determine whether to update the recommendation model. This ensures that the execution results corresponding to the target parameters generated by the recommendation model increasingly approximate the target information and better meet the user's needs.

[0018] As one possible implementation, the target information includes at least one of the following: experience quality information, network function attribute information, and network function service range information. Based on this, the second network element can send multiple target information based on various expected goals, and receive target parameters and corresponding parameter indicators determined according to the target information. Then, it executes the target parameters to obtain the execution result. This execution result can be used to determine whether to update the model, so as to optimize the recommendation model for various user expected goals and meet user needs from multiple perspectives.

[0019] As one possible implementation, the first message also includes candidate parameters, which are related to the expected target and are used to determine the target parameters. Based on this, the second network element can also send candidate parameters when sending target information to assist in determining the target parameters. The candidate parameters sent by the second network element reduce the overhead of determining the target parameters, thus saving time costs associated with optimizing the model.

[0020] As one possible implementation, the target parameters include at least one of the following: service quality parameters or network element candidate list parameters; the parameter indicators include at least one of the following: expected goal achievement degree, expected goal relevance, or target parameter priority. Based on this, at least one of the expected goal achievement degree, expected goal relevance, and target parameter priority can be used to indicate the ability of the service quality parameters and at least one of the network element candidate list parameters to achieve the expected goals. Thus, the second network element can adjust the values ​​of different target parameters and the execution order of different target parameters according to the indications of the parameter indicators, and execute the target parameters to obtain the execution results. This makes the execution results closer to the target information.

[0021] As one possible implementation, the method further includes sending a second message, which indicates that the execution result has been received. Based on this, the second network element can send the execution result obtained from executing the target parameters after sending the second message. In other words, the second network element triggers the collection of the execution result of the target parameters by sending the second message, so as to determine whether to update the recommendation model based on the execution result.

[0022] As one possible implementation, the method further includes: sending a third message, which indicates that at least one of the target information or candidate parameters should be updated. The third message includes one or more of the following: updated target information or updated candidate parameters. Based on this, the second network element can indicate the determination of new target information and candidate parameters by sending updated target information or updated candidate parameters, and then redetermine the target parameters and their corresponding parameter indices based on the new target information and candidate parameters. In this way, when the target information or candidate parameters change, only the partially changed updated target information or updated candidate parameters can be sent, thereby saving transmission resources and improving the efficiency of model training.

[0023] As one possible implementation, the method further includes: sending a fourth message when the execution result and target information meet a first preset condition. The fourth message instructs the update of the recommendation model. The first preset condition includes at least one of the following: the difference between the execution result and the target information is greater than or equal to a first preset threshold; or the ratio of the execution result to the target information is less than or equal to a second preset threshold. Based on this, the second network element can analyze the execution result and target information, and send a fourth message when the difference between the execution result and the target information is significant, instructing the update of the recommendation model. This makes the execution result corresponding to the recommendation information generated by the recommendation model closer to the target information and better meets the user's needs. When the difference between the execution result and the target information is small, it indicates that the execution result basically meets the user's needs, and therefore the fourth message is not sent.

[0024] Thirdly, a communication device is provided for implementing the method provided in the first aspect. The communication device can be the first network element in the first aspect. The communication device includes modules, units, or means corresponding to the above method, which can be implemented in hardware, software, or by hardware executing corresponding software. The hardware or software includes one or more modules or units corresponding to the above functions.

[0025] In one possible implementation, the communication device may include a processing module and an interface module. The processing module can be used to implement the processing functions described in the first aspect and any possible implementation thereof. The processing module may be, for example, a processor. The interface module, also referred to as an interface unit, is used to implement the sending and / or receiving functions described in the first aspect and any possible implementation thereof. The interface module may consist of an interface circuit, a transceiver, a transceiver unit, or a communication interface.

[0026] In one possible implementation, the processing module is configured to control the interface module to receive a first message, the first message including target information, which indicates the expected target of the recommendation model; the processing module is also configured to control the interface module to send first recommendation information, which indicates target parameters and corresponding parameter indicators, the target parameters and parameter indicators being obtained based on the target information; the target parameters are used to achieve the expected target, and the parameter indicators are used to indicate the ability of the target parameters to achieve the expected target; the processing module is also configured to control the interface module to receive an execution result, the execution result being obtained based on the first recommendation information, and the execution result being used to determine whether to update the recommendation model.

[0027] In one possible implementation, the target information includes at least one of the following: experience quality information, network function attribute information, and network function service scope information.

[0028] In one possible implementation, the first message also includes candidate parameters, which are related to the expected target and are used by the first network element to determine the target parameters.

[0029] In one possible implementation, the target parameters include service quality parameters and at least one of the network element candidate list parameters; the parameter indicators include at least one of the expected target achievement degree, expected target relevance, and target parameter priority corresponding to the target parameters.

[0030] In one possible implementation, the processing module is also used to control the interface module to receive a second message, which indicates the receipt of the execution result.

[0031] In one possible implementation, the processing module is further configured to control the interface module to receive a third message, the third message being used to indicate updating at least one of the target information or candidate parameters, the third message including one or more of the following: updated target information or updated candidate parameters.

[0032] In one possible implementation, the processing module is further configured to determine to update the recommendation model when the execution result and the target information meet a first preset condition; wherein, the execution result and the target information meeting the first preset condition includes at least one of the following: the difference between the execution result and the target information is greater than or equal to a first preset threshold; or, the ratio of the execution result to the target information is less than or equal to a second preset threshold.

[0033] In one possible implementation, the processing module is also used to control the interface module to receive a fourth message, which is used to instruct the recommendation model that generates the first recommendation information to be updated.

[0034] Fourthly, a communication device is provided for implementing the method provided in the second aspect. This communication device can be the second network element in the second aspect. The communication device includes modules, units, or means that implement the method described above. These modules, units, or means can be implemented in hardware, software, or by hardware executing corresponding software. The hardware or software includes one or more modules or units corresponding to the functions described above.

[0035] In one possible implementation, the communication device may include a processing module and an interface module. The processing module can be used to implement the processing functions in the second aspect described above and any possible implementation thereof. The processing module may be, for example, a processor. The interface module, also referred to as an interface unit, is used to implement the sending and / or receiving functions in the second aspect described above and any possible implementation thereof. The interface module may consist of an interface circuit, a transceiver, a transceiver unit, or a communication interface.

[0036] In one possible implementation, the interface module is used to send a first message, which includes target information indicating the expected goal of the recommendation model; the interface module is also used to receive first recommendation information, which indicates target parameters and corresponding parameter metrics, which are obtained based on the target information. The target parameters are used to achieve the expected goal, and the parameter metrics indicate the ability of the target parameters to achieve the expected goal; the processing module is used to execute the target parameters based on the parameter metrics to obtain the corresponding execution result; the interface module is also used to send the execution result, which is used to determine whether to update the recommendation model.

[0037] In one possible implementation, the target information includes at least one of the following: experience quality information, network function attribute information, and network function service scope information.

[0038] In one possible implementation, the first message also includes candidate parameters, which are related to the expected target and are used to determine the target parameters.

[0039] In one possible implementation, the target parameters include at least one of the following: service quality parameters or network element candidate list parameters; the parameter indicators include at least one of the following: expected target achievement degree, expected target relevance, or target parameter priority.

[0040] In one possible implementation, the interface module is also used to send a second message, which indicates that the execution result has been received.

[0041] In one possible implementation, the interface module is also used to send a third message, which indicates that at least one of the target information or candidate parameters should be updated. The third message includes one or more of the following: updated target information or updated candidate parameters.

[0042] In one possible implementation, the interface module is further configured to send a fourth message when the execution result and target information meet the first preset condition. The fourth message is used to indicate an update to the recommendation model. The execution result and target information meeting the first preset condition includes at least one of the following: the difference between the execution result and the target information is greater than or equal to the first preset threshold; or the ratio between the execution result and the target information is less than or equal to the second preset threshold.

[0043] Fifthly, a communication device is provided, comprising: a processor; configured to cause the communication device to perform the method described in any of the preceding aspects by executing a computer program (or computer-executable instructions) stored in a memory, and / or by means of logic circuitry. The communication device may be a first network element as described in the first aspect; or, the communication device may be a second network element as described in the second aspect. Optionally, the number of processors may be one or more.

[0044] In one possible implementation, the communication device also includes a memory.

[0045] In one possible implementation, the processor and memory are integrated together; or, the memory is independent of the processor.

[0046] In one possible implementation, the communication device further includes a communication interface for communicating with other devices, such as transmitting or receiving data and / or signals. Exemplarily, the communication interface may be a transceiver, circuit, bus, module, or other type of communication interface.

[0047] In one possible implementation, the processor and / or memory also include an AI module for implementing AI-related functions. The AI ​​module can implement the methods for optimizing the recommendation model described above through software, hardware, or a combination of both.

[0048] In one possible implementation, the communication device is a chip or a chip system. Optionally, when the communication device is a chip system, it can be composed of chips or may include chips and other discrete components.

[0049] In one possible implementation, the communication device is a chip or a chip system. Optionally, when the communication device is a chip system, it can be composed of chips or may include chips and other discrete components.

[0050] A sixth aspect provides a communication device, comprising: a processor and an interface circuit; the interface circuit being configured to receive a computer program or instructions and transmit them to the processor; the processor being configured to execute the computer program or instructions to cause the communication device to perform the method described in any of the preceding aspects. The communication device may be a first network element as described in the first aspect; or, the communication device may be a second network element as described in the second aspect. Optionally, the number of processors may be one or more.

[0051] In a seventh aspect, a computer-readable storage medium is provided that stores instructions which, when executed on a computer, cause the computer to perform the methods described in any of the preceding aspects.

[0052] Eighthly, a computer program product containing instructions is provided, which, when run on a computer, enables the computer to perform the methods described in any of the preceding aspects.

[0053] A ninth aspect provides a communication system comprising a first network element for performing the method described in the first aspect and a second network element for performing the method described in the second aspect.

[0054] The technical effects of any possible implementation of aspects three through nine can be found in the technical effects of any one of aspects one through two or different possible implementations of any one of aspects, and will not be repeated here.

[0055] Understandably, provided that the solutions do not contradict each other, the solutions in the above aspects can be combined. Attached Figure Description

[0056] Figure 1 This is a schematic diagram of the communication system architecture provided in the embodiments of this application;

[0057] Figure 2 Schematic diagram of the communication network architecture provided in the embodiments of this application Figure 1 ;

[0058] Figure 3 Schematic diagram of the communication network architecture provided in the embodiments of this application Figure 2 ;

[0059] Figure 4 This is a schematic diagram of the hardware structure of the communication device provided in the embodiments of this application;

[0060] Figure 5 A flowchart illustrating the optimization method for the recommendation model provided in this application embodiment. Figure 1 ;

[0061] Figure 6 Detailed schematic diagram of the optimization method for the recommendation model provided in the embodiments of this application Figure 2 ;

[0062] Figure 7 Detailed flowchart of the optimization method for the recommendation model provided in the embodiments of this application Figure 1 ;

[0063] Figure 8 Detailed flowchart of the optimization method for the recommendation model provided in the embodiments of this application Figure 2 ;

[0064] Figure 9 This is a schematic diagram of the structure of the communication device provided in the embodiments of this application. Detailed Implementation

[0065] To ensure that the recommendation strategies generated by the recommendation model meet user needs, this application provides a method and apparatus for optimizing the recommendation model. In this method, a first network element determines target parameters and parameter indicators based on received target information and sends them to a second network element. The second network element executes the target parameters according to the parameter indicators to obtain the execution result. The first network element and / or the second network element can determine whether to update the recommendation model based on the execution result and the target information. This enables updating the recommendation model according to user needs, improving the optimization efficiency of the recommendation model and increasing user satisfaction with the recommendation information generated by the model.

[0066] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0067] The method provided in this application can be used in various communication systems. For example, the communication system can be a Universal Mobile Telecommunications System (UMTS) system, an LTE system, a 5G communication system, a WiFi system, a 3GPP-related communication system, a future communication system, or a system integrating multiple systems, etc., without limitation. Among them, 5G can also be referred to as NR. The following uses... Figure 1 The method provided in this application will be described using the communication system 10 shown as an example. Figure 1 This is merely an illustrative diagram and does not constitute a limitation on the applicable scenarios of the technical solutions provided in this application.

[0068] like Figure 1 The diagram shown is a schematic diagram of the architecture of the communication system 10 provided in this application. Figure 1 In the invention, the communication system 10 may include network element 101 (corresponding to the first network element in the invention content) and network element 102 (corresponding to the second network element in the invention content).

[0069] In this application, network element 102 can send target information to network element 101. This target information is used to instruct the recommendation model on the expected goal of generating recommendation information. After receiving the target information, network element 101 can determine first recommendation information based on the target information, which includes target parameters and corresponding parameter indicators. Network element 101 can send the first recommendation information to network element 102. After receiving the first recommendation information, network element 102 can execute the target parameters according to the parameter indicators to obtain the execution result. Network element 102 or network element 101 can determine whether to update the recommendation model based on the execution result and the target information. If network element 102 determines to update the recommendation model, it can send a third message to network element 101 to instruct network element 101 to update the recommendation model.

[0070] In this application, network element 101 can be a network element with analytical logic function (AnLF), or it can be a network element with both AnLF and recommendation logic function (ReLF). AnLF and ReLF can be deployed together in the same core network element, or they can be deployed independently in different core network elements; there is no restriction. A network element deploying AnLF can be called an AnLF network element, and a network element deploying ReLF can be called a ReLF network element.

[0071] For example, network element 101 is an existing network element in the core network, such as an NWDAF network element. An NWDAF network element can have AnLF, model training logic function (MTLF), and ReLF; or, an NWDAF network element can have AnLF and MTLF; or, an NWDAF network element can have AnLF and ReLF. It should be understood that the above network element names are merely examples of network element 101's name. In specific applications, network element 101 can also be named using other methods without restriction.

[0072] In this application, network element 102 can be a terminal or a network function (NF). A terminal is a device with wireless transceiver capabilities. Terminals can be deployed on land, including indoors, outdoors, handheld, or vehicle-mounted; they can also be deployed on water (such as ships); and they can be deployed in the air (such as airplanes, balloons, and satellites). A terminal can also be called a terminal device, which can be user equipment (UE), mobile station (MS), mobile terminal (MT), or a device used to provide voice or data connectivity to users. UEs include handheld devices with wireless communication capabilities, vehicle-mounted devices (e.g., cars, bicycles, electric vehicles, airplanes, ships, trains, high-speed trains), wearable devices (e.g., smartwatches, smart bracelets, pedometers), or computing devices. For example, a UE can be a mobile phone, tablet computer, laptop computer, PDA, mobile internet device (MID), satellite terminal, or computer with wireless transceiver capabilities. UE can also be a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless modem, a point-of-sale (POS) machine, customer-premises equipment (CPE), a smart robot, a robotic arm, workshop equipment, smart home devices (e.g., refrigerators, televisions, air conditioners, electricity meters, etc.), a wireless terminal in industrial control, a wireless terminal in autonomous driving, a wireless terminal in telemedicine, a wireless terminal in a smart grid, a wireless terminal in transportation safety, a wireless terminal in a smart city, a wireless terminal in a smart home, an in-vehicle terminal, a roadside unit (RSU) with terminal functionality, or a flying device (e.g., a smart robot, a hot air balloon, a drone, an airplane), etc. A terminal can also be other devices with terminal functionality. For example, a terminal can also be a device that performs terminal functionality in device-to-device (D2D) communication.

[0073] In this application, user NFs include, but are not limited to, AMF network elements, SMF network elements, NRF network elements, and NWDAF network elements. The main functions of each network element are described below.

[0074] AMF: Used for access and mobility management, mainly for access management functions.

[0075] SMF: Used for session management, managing user session creation, deletion, etc., and maintaining session context and user plane forwarding pipeline information.

[0076] UPF: Responsible for processing user messages, such as forwarding, billing, and legitimate interception.

[0077] NWDAF: This function can be used to collect, analyze, and predict data. Data collection includes, but is not limited to, data from other NFs, such as AMF, SMF, PCF, etc., data collected from NEF or AF, or data collected from the OAM system. The logical function AnLF is primarily used to perform inference, analyze statistical and / or predictive information based on user NF requests, and expose the analysis service. The logical function MTLF trains the model based on the acquired data to obtain the trained model.

[0078] NRF (Network Element Provider): Used to provide network element discovery functionality, providing network element information corresponding to the network element type based on requests from other network elements. NRF also provides network element management services, such as network element registration, updates, deregistration, and network element status subscription and notification.

[0079] Optionally, the communication system 10 may also include at least one of the following: SMF network element, AMF network element, and NRF network element. Figure 1 (Not shown in the image).

[0080] Optionally, the above-mentioned communication system 10 can be applied to the 5G network currently under discussion, or to future communication networks, etc., and this application embodiment does not specifically limit it in this regard.

[0081] For example, communication system 10 is suitable for Figure 2The diagram shows a 5G network. This 5G network includes terminals, access network equipment, and core network equipment. Terminal equipment accesses the data network (DN) through the access network equipment and core network equipment. The core network equipment includes some or all of the following network functions: unified data management (UDM) network elements, network exposure function (NEF) network elements, application function (AF) network elements, policy control function (PCF) network elements, access and mobility management function (AMF) network elements, session management function (SMF) network elements, user plane function (UPF) network elements, network data analysis function (NWDAF) network elements, and network storage function (NRF) network elements.

[0082] exist Figure 2 In this configuration, the terminal accesses the 5G network through the RAN node, and communicates with the AMF network element through the N1 interface (N1). The RAN node communicates with the AMF network element through the N2 interface (N2) and with the UPF network element through the N3 interface (N3). The SMF network element communicates with the UPF network element through the N4 interface (N4), and the UPF network element accesses the DN through the N6 interface (N6). Furthermore, Figure 3 The network elements shown, such as AMF, SMF, UDM, NEF, PCF, NRF, NWDAF, or AF, can interact using service-oriented interfaces. For example, the service-oriented interface provided by the AMF network element is Namf; that of the SMF network element is Nsmf; that of the UDM network element is Nudm; that of the NEF network element is Nnef; that of the PCF network element is Npcf; that of the NRF network element is Nnrf; that of the AF network element is Naf; and that of the NWDAF network element is Nuwdaf.

[0083] Understandably, the device or entity corresponding to network element 101 in communication system 10 is... Figure 3The NWDAF network element shown is part of the 5G network. This NWDAF network element has AnLF, or has both AnLF and ReLF. The network element or entity corresponding to network element 102 in communication system 10 is... Figure 2 The terminals or other user NFs in the 5G network shown are not limited here.

[0084] For example, see Figure 3 Taking network element 101 as an example of an NWDAF network element, the NWDAF network element includes an analysis module, an acquisition module, and a recommendation model. The first network element is used to update the recommendation model. Specifically, the first network element may include an analysis module and an acquisition module. The acquisition module can receive target information from the second network element, and the analysis module can analyze the target information to obtain target parameters and corresponding parameter indicators, which are then sent to the second network element. The recommendation model is used to receive and respond to recommendation monitoring requests from the second network element.

[0085] Understandable. Figure 1 The communication system 10 shown is for illustrative purposes only and is not intended to limit the technical solutions of this application. Those skilled in the art should understand that in specific implementations, the communication system 10 may also include other network elements, and the number of each network element can be determined according to specific needs without limitation.

[0086] Optionally, this application Figure 1 Each network element or device in the application (such as network element 101, network element 102, etc.) can also be referred to as a communication device. It can be a general-purpose device or a special-purpose device. This application does not make any specific limitation in this regard.

[0087] Optionally, this application Figure 1 The functions of each network element or device (e.g., network element 101, network element 102, etc.) can be implemented by one device, multiple devices working together, or one or more functional modules within a single device. This application does not impose any specific limitations on these functions. It is understood that the aforementioned functions can be network elements within hardware devices, software functions running on dedicated hardware, a combination of hardware and software, or virtualization functions instantiated on a platform (e.g., a cloud platform).

[0088] In its specific implementation, this application Figure 1 Each network element or device (such as network element 101, network element 102, etc.) can adopt Figure 4 The shown composition structure, or including Figure 4 The components shown. Figure 4The diagram shows a hardware structure schematic of a communication device applicable to this application. It will be understood that the communication device 40 includes means of the necessary form, such as modules, units, elements, circuits, or interfaces, appropriately configured together to execute the solution provided in this application. For example, the communication device 40 includes one or more processors 401 for implementing the method provided in this application.

[0089] Processor 401 can be a general-purpose processor or a dedicated processor. For example, processor 401 can be a baseband processor or a central processing unit (CPU). The baseband processor can be used to process communication protocols and communication data, while the CPU can be used to control the communication device 40 (such as network element 101, network element 102, etc.), execute software programs, and process data from the software programs. Optionally, in one design, processor 401 may include program 405 (sometimes also referred to as code or instructions), which can be run on processor 401 to cause the communication device 40 to perform the methods described in the following embodiments. In yet another possible design, the communication device 40 includes circuitry (…). Figure 4 (Not shown), the circuit is used to implement the recommendation model optimization function in the following embodiments.

[0090] Optionally, the communication device 40 may include one or more memories 403. The memory 403 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM), cache, or other type of dynamic storage device capable of storing information and instructions. It may also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM), or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media, or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures that can be accessed by a computer, but is not limited thereto. The memory provided in this application may generally be non-volatile. Optionally, the memory 403 stores a program 407 (sometimes referred to as code or instructions), which can be run on the processor 401 to cause the communication device 40 to perform the methods described in the following method embodiments.

[0091] Optionally, the processor 401 may include an AI module 406, and / or the memory 403 may include an AI module 408. The aforementioned AI modules are used to implement AI-related functions. The AI ​​modules can implement recommendation model optimization methods through software, hardware, or a combination of both.

[0092] Optionally, data may also be stored in the processor 401 and / or the memory 403. The processor 401 and the memory 403 may be configured separately or integrated together.

[0093] Optionally, the communication device 40 may also include a transceiver 402 and / or an antenna 404. The processor 401, sometimes referred to as a processing unit, controls the communication device 40. The transceiver 402, sometimes referred to as a transceiver unit, transceiver, transceiver circuit, or transceiver, is used to realize the transmission and reception functions of the communication device 40 through the antenna 404.

[0094] Understandable. Figure 4 The structural composition shown does not constitute a limitation on the communication device, except... Figure 4 In addition to the components shown, the communication device may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0095] The method provided in this application will now be described with reference to the accompanying drawings. Each network element in the following embodiments may possess... Figure 4 The components shown are not described in detail.

[0096] It is understood that in this application, the first network element and the second network element can perform some or all of the steps in this application. These steps are merely examples, and this application can also perform other steps or variations thereof. Furthermore, the steps can be performed in different orders as presented in this application, and it is not necessary to perform all the steps in this application.

[0097] It is understood that the methods provided below in this application use the first network element and the second network element as examples to illustrate the interaction, but this application does not limit the execution subject of the interaction. For example, the first network element in the methods provided in the following embodiments of this application may also be a chip, chip system, or processor that supports the first network element in implementing the method, or it may be a logical node, logical module, or software that can implement all or part of the functions of the first network element; the second network element in the methods provided in the following embodiments of this application may also be a chip, chip system, or processor that supports the second network element in implementing the method, or it may be a logical node, logical module, or software that can implement all or part of the functions of the second network element.

[0098] The method provided in this application will now be described with reference to the accompanying drawings. Each network element or device in the following embodiments may possess... Figure 4 The components shown are not described in detail.

[0099] It is understood that the message names between network elements or the names of parameters in the messages in the following embodiments of this application are just examples, and other names may be used in the specific implementation. This application does not make any specific limitations on this.

[0100] It is understood that in this application, " / " can indicate that the objects before and after it are in an "or" relationship. For example, A / B can mean A or B; "and / or" can be used to describe three relationships between the related objects. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone, where A and B can be singular or plural. Furthermore, expressions like "at least one of A, B, and C" or "at least one of A, B, or C" are generally used to indicate any of the following: A exists alone; B exists alone; C exists alone; A and B exist simultaneously; A and C exist simultaneously; B and C exist simultaneously; A, B, and C exist simultaneously. The above examples using three elements (A, B, and C) illustrate the optional entries for this item. When the expression contains more elements, its meaning can be obtained according to the aforementioned rules.

[0101] To facilitate the description of the technical solutions of this application, the terms "first" and "second" may be used to distinguish technical features with the same or similar functions. The terms "first" and "second" do not limit the number or execution order, nor do they imply that they are necessarily different. In this application, the terms "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design scheme described as "exemplary" or "for example" should not be construed as being more preferred or advantageous than other embodiments or design schemes. The use of "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner for ease of understanding.

[0102] It is understood that the term "embodiment" used throughout the specification means that a specific feature, structure, or characteristic related to an embodiment is included in at least one embodiment of this application. Therefore, various embodiments throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It is understood that in the various embodiments of this application, the sequence number of each process does not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of this application.

[0103] It is understood that in this application, "when," "under the circumstances," "if," and "if" all refer to the corresponding processing that will be carried out under certain objective circumstances, and are not time-limited, nor do they require that there must be a judgment action when implemented, nor do they imply any other limitations.

[0104] It is understood that some optional features in this application can be implemented independently in certain scenarios without relying on other features, such as the current solution upon which they are based, to solve the corresponding technical problems and achieve the corresponding effects. Alternatively, they can be combined with other features as needed in certain scenarios. Correspondingly, the apparatus provided in this application can also implement these features or functions, which will not be elaborated here.

[0105] It is understood that the same step or step with the same function or technical feature in this application can be referenced and learned from each other in different embodiments.

[0106] In some embodiments, such as Figure 5 The image shows an optimization method for a recommendation model provided in this application, which may include the following steps:

[0107] S501: The second network element sends a first message to the first network element. Correspondingly, the first network element receives the first message from the second network element.

[0108] In this application, the first network element can be Figure 1 In the communication system shown, network element 101, the second network element can be... Figure 1 Network element 102 in the communication system shown. The second network element can determine the first message and send the first message to the first network element.

[0109] The first message includes target information, which indicates the expected goal of the recommendation model in the first network element. The second network element can use the target information to indicate the expected goal of the recommendation model to the first network element. In some examples, the target information can be determined by user requirements or by application layer requirements of the second network element.

[0110] Optionally, the target information includes at least one of the following: quality of experience information, network function attribute information, or network function service range information. Optionally, quality of experience information may include QoE, network function attribute information may include network element load, and network function service range may include the service range of the network element, etc.

[0111] For example, when a second network element requires the experience quality to reach a predetermined goal, it can send this requirement to the first network element through experience quality information. When a second network element requires the network function attributes to reach a predetermined goal, it can send this requirement to the first network element through network function attribute information. When a second network element requires the network function service range to reach a predetermined goal, it can send this requirement to the first network element through network function service range information. The target information may also include other information characterizing the expected goals of the second network element, without any limitations. It is understood that the first network element can obtain different types of performance requirements needed by the second network element by receiving different types of target information. In some embodiments, the first network element determines the expected target first recommendation information for implementing the recommendation model based on the received target information. The first recommendation information may indicate target parameters and corresponding parameter indicators. For example, the first recommendation information may include target parameters and the aforementioned parameter indicators, or the first recommendation information may include the identifier of the target parameter and the identifier of the aforementioned parameter indicator. The first recommendation information can be used to instruct the execution of the aforementioned target parameters. In this application, the number of target parameters can be one or more. When the number of target parameters is multiple, each target parameter may correspond to one parameter indicator, or multiple target parameters may correspond to one parameter indicator, without limitation.

[0112] Optionally, when sending target information, the second network element may also send candidate parameters corresponding to that target information. For example, the first message may also include candidate parameters. These candidate parameters are related to the expected target and can be used to determine the target parameters. Therefore, the candidate parameters can provide a reference for the first network element in determining the target parameters, reducing the burden on the first network element when determining the target parameters. This also combines the results of the second and first network elements in determining the target parameters, thereby improving user satisfaction with the target parameters.

[0113] In some embodiments, the first network element determines first recommendation information for achieving the expected target of the recommendation model based on the received target information and candidate parameters. The first recommendation information may include target parameters and parameter indices corresponding to the target parameters. The first recommendation information can be used to instruct the execution of the aforementioned target parameters.

[0114] In some examples, when determining the first recommendation information, the first network element can add parameters to the candidate parameters to obtain the target parameters and determine the parameter indicators corresponding to the target parameters; alternatively, it can use the candidate parameters as the target parameters and determine the parameter indicators corresponding to the target parameters.

[0115] S502: The first network element sends a first recommendation message to the second network element. Correspondingly, the second network element receives the first recommendation message from the first network element.

[0116] As described in S501, the first recommendation information indicates the target parameters and the corresponding parameter metrics. The parameter metrics indicate the ability of the target parameters to achieve the expected goals; that is, the influence of each target parameter on the recommendation model's achievement of the expected goals is different, and the parameter metrics characterize the information about the influence of the target parameters on the recommendation model's achievement of the expected goals.

[0117] Optionally, the target parameters include service quality parameters, or at least one of the parameters in the network element candidate list; the parameter indicators include at least one of the expected target achievement degree, expected target relevance, or target parameter priority corresponding to the target parameters. The service quality parameters can be QoS parameters, and the network element candidate list can be a list of instances of a certain type of network element.

[0118] Understandably, the service quality parameters and the network element candidate list parameters have different capabilities in achieving the expected goals of the recommendation model. In other words, the service quality parameters and the network element candidate list parameters have different influences on the recommendation model's achievement of its expected goals. The first network element can indicate the difference in capability or influence between the target parameters to the second network element through at least one of the following: expected goal achievement degree, expected goal relevance, or target parameter priority. For example, when the expected goal achievement degree of the service quality parameters is greater than that of the network element candidate list, the second network element will prioritize the execution of the service quality parameters. When the expected goal relevance of the network element candidate list is greater than that of the service quality parameters, the second network element will prioritize the execution of the network element candidate list.

[0119] Optionally, the first network element instructs the second network element to execute the target parameters and obtain the execution result by indicating the target parameters and corresponding parameter indicators. When executing the target parameters, the second network element can adjust the possible value range and combination of the target parameters according to the parameter indicators, and then execute the target parameters to obtain the corresponding execution result.

[0120] For example, the second network element can indicate the combination order of service quality parameters and network element candidate list parameters through target parameter priority, and can indicate the priority of executing service quality parameters and network element candidate list parameters through expected target achievement degree and / or expected target relevance. Based on the target achievement degree, expected target relevance, and target parameter priority, at least one of the service quality parameters and network element candidate list parameters is executed to obtain the corresponding execution result.

[0121] S503: The second network element sends the execution result obtained from the execution target parameters to the first network element. Correspondingly, the first network element receives the execution result obtained from the execution target parameters from the second network element.

[0122] In this application, the execution result can be used to determine whether to update the recommendation model.

[0123] One possible implementation is that the first network element can proactively trigger recommendation monitoring, obtain the execution result obtained by the second network element executing the target parameters, and determine whether to update the recommendation model based on the execution result.

[0124] Optionally, the first network element can receive the execution result from the second network element and determine whether to update the recommendation model based on the execution result. Furthermore, the first network element can also obtain other network data and network performance related to the execution result. This is used in conjunction with detection information to update the recommendation model, and no limitations are imposed here.

[0125] Another possible implementation is that the first network element can passively trigger recommendation monitoring based on the instruction of the second network element, obtain the execution result obtained by the second network element in executing the target parameters, and update the recommendation model based on the execution result.

[0126] Optional See also Figure 6 After executing S501, the method may further include:

[0127] S501a: The second network element sends a fourth message to the first network element. Correspondingly, the first network element receives the fourth message sent by the second network element.

[0128] The fourth message is used to instruct the target information or candidate parameters in the first message to be updated. When the first network element receives the fourth message, it updates the target information and candidate parameters in the first message and re-executes S502 and subsequent steps.

[0129] For example, the fourth message includes one or more of the following: updated target information or updated candidate parameters. The updated target information or updated candidate parameters are used to indicate the changed target information or candidate parameters. After receiving the fourth message, the first network element can determine the target parameters and corresponding parameter indices based on the updated target information and candidate parameters.

[0130] Optional, see Figure 6 Before executing S503, the method also includes:

[0131] S503a: The second network element sends a second message to the first network element. Correspondingly, the first network element receives the second message from the second network element.

[0132] The second message instructs the first network element to obtain the execution result. It can be understood that after receiving the instruction in the second message, the first network element can obtain the execution result from the second network element and update the recommendation model based on the execution result.

[0133] Optionally, the second message includes target parameters. The second network element can instruct the first network element to obtain the execution result corresponding to the target parameters through the second message, so as to train the recommendation model based on the execution result corresponding to the indicated target parameters.

[0134] Optional, see Figure 6 After executing S503, the method further includes:

[0135] S504: When the execution result and target information meet the first preset condition, the first network element updates the recommendation model.

[0136] One possible implementation is that the first network element can proactively trigger an update to the recommendation model based on the execution result. That is, after receiving the execution result, the first network element determines whether to update the recommendation model based on a comparison between the execution result and the target information. For example, if the difference between the recommended information and the target information is large, it indicates that the execution result has failed to achieve the expected goal of the recommendation model, and therefore the recommendation model can be updated. If the difference between the recommended information and the target information is small, it indicates that the execution result has basically achieved the expected goal of the recommendation model, and therefore the recommendation model is not updated.

[0137] For example, the first network element updates the recommendation model when the difference between the execution result and the target information is greater than or equal to a first preset threshold; or when the ratio of the execution result to the target information is less than or equal to a second preset threshold. It is understandable that when the execution result meets the first preset condition, the execution result does not meet the expected goal, therefore the recommendation model needs to be updated to optimize the user experience. When the execution result does not meet the first preset condition, it means that the execution result meets the user's expected goal, therefore the recommendation model does not need to be updated.

[0138] Another possible implementation is that the first network element can passively trigger the model update based on the instruction of the second network element. In other words, the second network element can determine whether to update the recommendation model based on the comparison between the execution result and the target information. In this case, the second network element may not need to send the execution result to the first network element.

[0139] For example, if there is a large difference between the recommended information and the target information, it means that the execution result has failed to achieve the expected goal of the recommendation model, so the recommendation model can be updated. If there is a small difference between the recommended information and the target information, it means that the execution result has basically achieved the expected goal of the recommendation model, so the recommendation model does not need to be updated.

[0140] Optional, see Figure 6 Before executing S504, the method also includes:

[0141] S504a: When the execution result and target information meet the first preset condition, the second network element sends a third message to the first network element. Correspondingly, the first network element receives the third message from the second network element.

[0142] The third message instructs the first network element to update the recommendation network model. The first network element updates the recommendation model upon receiving the third message.

[0143] For example, if the difference between the execution result and the target information is greater than a preset threshold, or if the ratio of the execution result to the target information is less than a preset threshold, the second network element sends a third message to the first network element. It is understandable that when the execution result meets the first preset condition, the execution result does not meet the expected goal, therefore a third message needs to be sent to instruct the first network element to update the recommendation model to optimize the user experience. When the execution result does not meet the first preset condition, it means that the execution result meets the user's expected goal, therefore the recommendation model does not need to be updated.

[0144] The above embodiments specifically describe an optimization method for a recommendation model, which can also be used in the following scenarios:

[0145] In some embodiments, see Figure 7 The following example, using ANLF as the first network element and user NF as the second network element, illustrates the application scenario of the above recommendation model optimization method. When ANLF and RELF are deployed together in the same core network element NWDAF, ANLF implements the function of the first network element in the aforementioned embodiment. Furthermore, this scenario also includes SMF, AMF, NRF, and MTLF assisting user NF and NWDAF in implementing a recommendation model optimization method. This optimization method may include:

[0146] S701: User NF sends a recommendation policy subscription or request to ANLF (i.e., the first message in the aforementioned embodiment).

[0147] The recommended strategy subscription or request includes target information. For example, target information may include one or more of the following: QoE (Quality of Experience) and expected NF (Network Function) attributes. For example, expected NF attributes may include NF type, service scope, etc.

[0148] Optionally, the recommended policy subscription or request may also include candidate parameters. For example, candidate parameters may include QoS parameters, a list of candidate network elements, flow control policies, etc.

[0149] In some examples, the recommendation strategy subscription or request also includes information such as the user's NF identifier and the monitoring time interval.

[0150] S702: AnLF calls the MTLF model, and MTLF subscribes to model monitoring from ANLF.

[0151] Optionally, MTLF is used to collect data from various data sources to update the recommendation model.

[0152] S703: AnLF analyzes and makes recommendations based on target information and candidate parameters, and outputs the first recommendation information.

[0153] The first recommendation information is used to indicate the target parameters and their corresponding metrics. In some examples, the metrics include the degree of achievement of the target parameters, the correlation between the target parameters and the expected goals, and the parameter priority or ranking.

[0154] Optionally, AnLF analyzes and makes recommendations based on the target information and candidate parameters, and outputs the first recommendation.

[0155] S704: AnLF sends a recommendation response to user NF, which includes the first recommendation information. Correspondingly, user NF receives the recommendation response from AnLF.

[0156] S705: User NF executes the target parameters indicated by the first recommendation information based on the received first recommendation information.

[0157] Optionally, before executing the target parameters indicated by the target parameters, the user NF can modify the target parameters accordingly based on the parameter indicators.

[0158] After the user executes the target parameters in NF, there are two ways to trigger NWDAF for recommendation monitoring, and the result will determine whether the recommendation model needs to be updated.

[0159] Method 1: Triggered via user NF

[0160] S706: User NF sends a monitoring request or subscription for the recommendation strategy to NWDAF (i.e., the second message in the aforementioned embodiment).

[0161] Optionally, the monitoring request or subscription for the recommended strategy may include the target parameters for the user's NF execution.

[0162] S707a: After AnLF receives a monitoring request or subscription for a recommendation strategy, it triggers AnLF to monitor the user's NF for recommendations, obtains the execution results, and analyzes the results to determine whether the recommendation model needs to be updated or optimized.

[0163] Method 2: Triggered via NWDAF

[0164] S707b: AnLF triggers NWDAF to collect and analyze data, including collecting the target parameters and execution results from the user's NF, and analyzing and determining whether the recommendation model needs to be updated or optimized.

[0165] This application offers two methods to determine whether the recommendation model needs to be updated or optimized.

[0166] Method 1: Determine whether to update the recommendation model using AnLF.

[0167] S708a: After AnLF receives the execution result, if the execution result of the target parameter does not meet expectations, that is, if the gap between the execution result and the target information is large, AnLF triggers the update of the recommendation model.

[0168] Method 2: Determine whether to update the recommendation model based on user NF.

[0169] S708b: User NF compares the execution result with the target information. If the execution result of the target parameters does not meet expectations, that is, if the difference between the execution result and the target information is large, then a model update request (i.e., the third message in the aforementioned embodiment) is sent to AnLF. After receiving the model update request, AnLF updates the recommendation model.

[0170] S709: If AnLF updates or optimizes the recommendation model, it will send a model update request to MTLF. The model update request includes the execution results of the user's NF and information such as target information and target parameters.

[0171] S710: MTLF collects data from various data sources to retrain and optimize recommendation models.

[0172] Optionally, MTLF will report the results of optimizing the recommendation model to AnLF.

[0173] In some embodiments, during the application of the recommendation model, steps S703-S710 are repeated to achieve simultaneous training of the recommendation model during its application.

[0174] S711: If user NF updates target information or candidate parameters, user NF sends a monitoring update request (i.e., the fourth message in the aforementioned embodiment) to AnLF, which includes the updated target information and candidate parameters.

[0175] S712: AnLF sends a monitoring update response to user NF.

[0176] In some embodiments, after user NF sends a monitoring update request to AnLF, user NF and AnLF repeat steps S702-S711 to re-optimize the recommendation model based on the updated target information, which will not be elaborated here.

[0177] S713: User NF sends a request to AnLF to pause the subscription recommendation strategy.

[0178] S714: User NF sends a request to AnLF to restore the subscription recommendation policy.

[0179] Optionally, both the request to pause the subscription recommendation strategy and the request to resume the subscription recommendation strategy can include monitoring instructions, target information, candidate parameters, and other information.

[0180] In other embodiments, see Figure 8 The following example, using ANLF and ReLF as the first network element and user NF as the second network element, illustrates the application scenario of the above recommendation model optimization method. When ANLF and ReLF are independently deployed in different core network elements, ReLF, in combination with ANLF, implements the function of the first network element in the aforementioned embodiment. Furthermore, this scenario also includes SMF, AMF, NRF, and MTLF assisting user NF and NWDAF in implementing a recommendation model optimization method. This optimization method may include:

[0181] S801: User NF sends a recommendation policy subscription or request to ReLF (i.e., the first message in the aforementioned embodiment).

[0182] The recommended strategy subscription or request includes target information. For example, target information may include one or more of the following: QoE, expected NF attributes, such as NF type, service scope, etc.

[0183] Optionally, the recommended policy subscription or request may also include candidate parameters. For example, candidate parameters may include QoS parameters, a list of candidate network elements, etc. Flow control policy

[0184] In some examples, the recommendation strategy subscription or request also includes information such as the user's NF identifier and the monitoring time interval.

[0185] S802: ReLF sends an analytics function subscription message to AnLF. Correspondingly, AnLF receives the subscription message from ReLF.

[0186] The subscription messages carry the aforementioned target information.

[0187] S803: AnLF analyzes and makes recommendations based on target information and candidate parameters, and outputs the first recommendation information.

[0188] The first recommendation information is used to indicate the target parameters and their corresponding metrics. In some examples, the metrics include the degree of achievement of the target parameters, the correlation between the target parameters and the expected goals, and the parameter priority or ranking.

[0189] Optionally, AnLF analyzes and makes recommendations based on the target information and candidate parameters, and outputs the first recommendation.

[0190] S804: AnLF sends an analysis response to ReLF. Relatively, ReLF receives the analysis response from AnLF.

[0191] S805: ReLF sends a recommendation response to user NF, which includes the first recommendation information. Correspondingly, user NF receives a recommendation response from AnLF.

[0192] S806: User NF executes the target parameters indicated by the target parameters based on the first recommendation information received, in order to meet the expected goal.

[0193] Optionally, before executing the target parameters indicated by the target parameters, the user NF can modify the target parameters accordingly based on the parameter indicators.

[0194] After the user executes the target parameters in NF, there are two ways to trigger NWDAF for recommendation monitoring, and the result will determine whether the recommendation model needs to be updated.

[0195] Method 1: Triggered via user NF

[0196] S807: User NF sends a monitoring request or subscription for the recommendation strategy to NWDAF (i.e., the second message in the aforementioned embodiment).

[0197] Optionally, the monitoring request or subscription for the recommended strategy may include the target parameters for the user's NF execution.

[0198] S808a: After receiving a monitoring request or subscription for the recommendation strategy, ReLF is triggered to perform recommendation monitoring and obtain the execution results. Based on the analysis of the execution results, it is determined whether the recommendation model needs to be updated or optimized.

[0199] Method 2: Triggered via NWDAF

[0200] S808b: ReLF triggers NWDAF to collect data for analysis, including collecting the target parameters and execution results from the user's NF, and analyzing and determining whether the recommendation model needs to be updated or optimized.

[0201] After ReLF receives the execution result, there are two ways to determine whether the recommendation model needs to be updated or optimized.

[0202] Method 1: Determine whether to update the recommendation model using ReLF.

[0203] S809a: After ReLF receives the execution result, if the execution result of the target parameter does not meet expectations, ReLF triggers an update of the recommendation model.

[0204] Optionally, if the execution result differs significantly from the target information, ReLF will trigger an update of the recommendation model.

[0205] Method 2: Determine whether to update the recommendation model based on user NF.

[0206] S809b: User NF compares the execution result with the target information. If the execution result of the target parameters does not meet expectations, that is, if the difference between the execution result and the target information is large, then a model update request (i.e., the third message in the aforementioned embodiment) is sent to ReLF. After receiving the model update request, ReLF updates the recommendation model.

[0207] S810: If ReLF updates or optimizes the recommendation model, it will send a model update request to MTLF. The model update request includes the execution results of the user's NF and information such as target information and target parameters.

[0208] S811: MTLF collects data from various data sources to retrain and optimize the recommendation model.

[0209] Optionally, MTLF will report the results of optimizing the recommendation model to ReLF.

[0210] In some embodiments, during the application of the recommendation model, steps S803-S810 are repeated to achieve simultaneous training of the recommendation model during its application.

[0211] S812: If user NF updates target information or candidate parameters, user NF sends a monitoring update request (i.e., the fourth message in the aforementioned embodiment) to ReLF, which includes the updated target information and candidate parameters.

[0212] S813: ReLF sends a monitoring update response to the user NF.

[0213] In some embodiments, after the user NF sends a monitoring update request to the ReLF, the user NF and the ReLF repeat steps S802-S811 to realize the process of re-optimizing the recommendation model based on the updated target information, which will not be elaborated here.

[0214] S814: User NF sends a request to ReLF to pause the subscription recommendation strategy.

[0215] S815: User NF sends a request to ReLF to restore the subscription recommendation policy.

[0216] Optionally, both the request to pause the subscription recommendation strategy and the request to resume the subscription recommendation strategy can include monitoring instructions, target information, candidate parameters, and other information.

[0217] The above mainly describes the solution provided in this application from the perspective of interaction between various network elements. Correspondingly, this application also provides a communication device, which can be the first network element in the above method embodiments, or a device containing the first network element, or a component usable in the first network element; the communication device can also be the second network element in the above method embodiments, or a device containing the second network element, or a component usable in the second network element. It is understood that the aforementioned first network element, etc., includes hardware structures and / or software modules corresponding to the execution of each function in order to achieve the above functions. Those skilled in the art should readily recognize that, based on the unit and algorithm operations of the various examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0218] This application can divide the first network element or the second network element into functional modules based on the above method example. For example, each function can be divided into its own functional modules, or two or more functions can be integrated into one processing module. The integrated modules can be implemented in hardware or as software functional modules. It is understood that the module division in this application is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0219] For example, when dividing the functional modules in an integrated manner, Figure 9 A schematic diagram of a communication device 90 is shown. The communication device 90 includes an interface module 901 and a processing module 902. The interface module 901, also called an interface unit, is used to perform transmit and receive operations. For example, it can be an interface circuit, a transceiver, a transceiver unit, or a communication interface. The processing module 902, also called a processing unit, is used to perform operations other than transmit and receive operations. For example, it can be a processing circuit or a processor.

[0220] In some embodiments, the communication device 90 may further include a storage module. Figure 9 (Not shown in the image) is used to store program instructions and data.

[0221] In one example, the communication device is a first network element, which can be used to implement any of the methods executed by the first network element in the foregoing embodiments.

[0222] For example, in one possible implementation, interface module 901 is used to receive a first message, the first message including target information, the target information being used to indicate the expected target of the recommendation model; interface module 901 is also used to send first recommendation information, the first recommendation information indicating target parameters and corresponding parameter indicators, the target parameters and parameter indicators being obtained based on the target information; the target parameters are used to achieve the expected target, and the parameter indicators are used to indicate the ability of the target parameters to achieve the expected target; processing module 902 is used to receive execution results, the execution results being obtained based on the first recommendation information, and the execution results being used to determine whether to update the recommendation model.

[0223] In one example, the communication device is a second network element, which can be used to implement any of the methods executed by the second network element in the foregoing embodiments.

[0224] For example, interface module 901 is used to send a first message, which includes target information and is used to indicate the expected goal of the recommendation model; interface module 901 is also used to receive first recommendation information, which indicates target parameters and corresponding parameter indicators. The target parameters and parameter indicators are obtained based on the target information. The target parameters are used to achieve the expected goal, and the parameter indicators are used to indicate the ability of the target parameters to achieve the expected goal; processing module 902 is used to execute the target parameters based on the parameter indicators to obtain the corresponding execution result; interface module 901 is also used to send the execution result, which is used to determine whether to update the recommendation model.

[0225] When the communication device is used to implement the functions of the first network element or the second network element, other functions that the communication device 90 can implement can be found in [reference needed]. Figure 5 The relevant descriptions of the embodiments shown will not be elaborated upon further.

[0226] In a simplified embodiment, those skilled in the art will recognize that the communication device 90 can employ... Figure 4 The form shown. For example, Figure 4 The processor 401 can call computer execution instructions stored in the memory 403 to cause the communication device 90 to execute the method described in the above method embodiment.

[0227] For example, Figure 9 The functions / implementation process of the processing module 902 and the interface module 901 can be achieved through... Figure 4 The processor 401 in the memory calls computer execution instructions stored in the memory 403 to implement the function. Alternatively, Figure 9 The function / implementation process of the processing module 902 can be achieved through... Figure 4 The processor 401 in the memory calls computer execution instructions stored in the memory 403 to implement this. Figure 9The function / implementation process of interface module 901 can be accessed through... Figure 4 This is achieved using transceiver 402.

[0228] It is understood that one or more of the above modules or units can be implemented by software, hardware, or a combination of both. When any of the above modules or units are implemented by software, the software exists as computer program instructions and is stored in memory. The processor can be used to execute the program instructions and implement the above method flow. The processor can be built into a SoC (System-on-Chip) or ASIC, or it can be a separate semiconductor chip. In addition to the core that executes software instructions for computation or processing, the processor may further include necessary hardware accelerators, such as field-programmable gate arrays (FPGAs), PLDs (Programmable Logic Devices), or logic circuits that implement dedicated logic operations.

[0229] When the above modules or units are implemented in hardware, the hardware can be any one or any combination of a CPU, microprocessor, digital signal processing (DSP) chip, microcontroller unit (MCU), artificial intelligence processor, ASIC, SoC, FPGA, PLD, application-specific digital circuit, hardware accelerator, or non-integrated discrete device, which can run the necessary software or perform the above method flow independently of software.

[0230] Optionally, this application also provides a chip system, including: at least one processor and an interface, wherein the at least one processor is coupled to a memory via the interface, and when the at least one processor executes a computer program or instructions in the memory, the method in any of the above method embodiments is executed. In one possible implementation, the chip system further includes a memory. Optionally, the chip system may be composed of chips or may include chips and other discrete devices; this application does not specifically limit this.

[0231] Optionally, this application also provides a computer-readable storage medium. All or part of the processes in the above method embodiments can be executed by a computer program instructing related hardware. This program can be stored in the aforementioned computer-readable storage medium. When executed, the program can include the processes described in the above method embodiments. The computer-readable storage medium can be an internal storage unit of the communication device in any of the foregoing embodiments, such as a hard disk or memory of the communication device. The aforementioned computer-readable storage medium can also be an external storage device of the communication device, such as a plug-in hard disk, smart memory card (SMC), secure digital card (SD), flash card, etc., equipped on the communication device. Further, the aforementioned computer-readable storage medium can include both internal storage units and external storage devices of the communication device. The aforementioned computer-readable storage medium is used to store the aforementioned computer program and other programs and data required by the communication device. The aforementioned computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0232] Optionally, this application also provides a computer program product. All or part of the processes in the above method embodiments can be executed by a computer program instructing related hardware. This program can be stored in the above computer program product, and when executed, it can include the processes described in the above method embodiments.

[0233] Optionally, this application also provides computer instructions. All or part of the processes in the above method embodiments can be executed by computer instructions instructing related hardware (such as a computer, processor, first network element, or second network element). The program can be stored in the aforementioned computer-readable storage medium or the aforementioned computer program product.

[0234] Optionally, this application also provides a communication system, including: Figure 5 The first network element and the second network element in the illustrated embodiment.

[0235] Optionally, this application also provides a communication system, including: Figure 6 The first network element and the second network element in the illustrated embodiment.

[0236] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0237] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of modules or units is merely a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be electrical, mechanical, or other forms.

[0238] The units described as separate components may or may not be physically separate. A component shown as a unit can be one or more physical units; that is, it can be located in one place or distributed in multiple different locations. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0239] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0240] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An optimization method for a recommendation model, characterized in that, The method comprises: receiving a first message, the first message comprising target information, the target information being used to indicate an expected target of a recommendation model; sending first recommendation information, the first recommendation information indicating a target parameter and a parameter index corresponding to the target parameter, the target parameter and the parameter index being obtained according to the target information, the target parameter being used to achieve the expected target, and the parameter index being used to indicate an ability of the target parameter to achieve the expected target; receiving an execution result, the execution result being obtained based on the first recommendation information, and the execution result being used to determine whether to update the recommendation model.

2. The method of claim 1, wherein, The target information comprises at least one of quality of experience information, network function attribute information, or network function service range information.

3. The method according to claim 1 or 2, characterized in that, The first message further comprises a candidate parameter, the candidate parameter being related to the expected target, and the candidate parameter being used to determine the target parameter.

4. The method according to any one of claims 1 to 3, characterized in that, The target parameter comprises at least one of a quality of service parameter or a network element candidate list parameter; and the parameter index comprises at least one of an expected target achievement degree corresponding to the target parameter, an expected target relevance, or a target parameter priority.

5. The method according to any one of claims 1 to 4, characterized in that, The method further comprises: receiving a second message, the second message being used to indicate the execution result.

6. The method of claim 3, wherein, The method further comprises: receiving a third message, the third message being used to indicate at least one of updated target information or updated candidate parameter, the third message comprising one or more of the updated target information or the updated candidate parameter.

7. The method according to any one of claims 1 to 6, characterized in that, The method further comprises: determining to update the recommendation model when the execution result and the target information satisfy a first preset condition; The execution result and the target information satisfying the first preset condition comprises at least one of: a difference between the execution result and the target information being greater than or equal to a first preset threshold; or a ratio of the execution result to the target information being less than or equal to a second preset threshold.

8. An optimization method for a recommendation model, characterized in that, The method comprises: sending a first message, the first message comprising target information, the target information being used to indicate an expected target of a recommendation model; receiving first recommendation information, the first recommendation information indicating a target parameter and a parameter index corresponding to the target parameter, the target parameter and the parameter index being obtained according to the target information, the target parameter being used to achieve the expected target, and the parameter index being used to indicate an ability of the target parameter to achieve the expected target; based on the parameter index, executing the target parameter to obtain a corresponding execution result; sending the execution result, the execution result being used to determine whether to update the recommendation model.

9. The method of claim 8, wherein, The target information comprises at least one of quality of experience information, network function attribute information, or network function service range information.

10. The method according to claim 8 or 9, characterized in that, The first message further comprises a candidate parameter, the candidate parameter being related to the expected target, and the candidate parameter being used to determine the target parameter.

11. The method according to any one of claims 8-10, characterized in that, The target parameter comprises at least one of a quality of service parameter or a network element candidate list parameter; and the parameter index comprises at least one of an expected target achievement degree corresponding to the target parameter, an expected target relevance, or a target parameter priority.

12. The method according to any one of claims 8-11, characterized in that, The method further includes: sending a second message, the second message being used to indicate the execution result.

13. The method of claim 10, wherein, The method further includes: sending a third message, the third message being used to indicate updating at least one of the target information or the candidate parameter, the third message including one or more of the following: updated target information or updated candidate parameter.

14. The method according to any one of claims 8 to 13, characterized in that, The method further includes: when the execution result and the target information satisfy a first preset condition, sending a fourth message, the fourth message being used to indicate updating the recommendation model; wherein the execution result and the target information satisfy the first preset condition, including at least one of the following: a difference between the execution result and the target information is greater than or equal to a first preset threshold; or a ratio of the execution result to the target information is less than or equal to a second preset threshold.

15. A communications device, characterized by The communication device includes units or modules for performing the method of any one of claims 1-7, or units or modules for performing the method of any one of claims 8-14.

16. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer program instructions, which when executed, implement the method of any one of claims 1-7, or implement the method of any one of claims 8-14.

17. A computer program product comprising instructions, characterized in that, When the computer program product is running on a computer, it causes the method of any one of claims 1-7 to be implemented, or causes the method of any one of claims 8-14 to be implemented.

18. A communications device, characterized by including: a processor coupled to a memory, the memory being used to store programs or instructions, when the programs or instructions are executed by the processor, causing the device to perform the method of any one of claims 1-7, or perform the method of any one of claims 8-14.

19. A communication system, characterized by including: a device for performing the method of any one of claims 1-7, and a device for performing the method of any one of claims 8-14.

Images