Method, device, and readable storage medium for subscribing to model transmission state analysis in a network

Abstract

The present disclosure provides a method, apparatus, and readable storage medium for subscribing to an analysis of a model transmission state in a network, the method comprising: sending a first message to a network data analysis function (NWDAF) directly or through a network capability exposure function (NEF); and the AF receiving analysis information of an AI / ML model transmission state sent by the NWDAF directly or through the NEF, wherein the analysis information is determined by receiving data of an AI / ML model transmission state sent by another network function 5GC NF(s) received by the NWDAF; wherein the analysis information is intended to adjust network strategy parameters and / or information of an application layer model. The present disclosure enables effective analysis of an AI / ML model transmission state so that a network can effectively adjust a network transmission strategy based on the AI / ML model transmission state and a third party can obtain the analysis of the AI / ML model transmission state to adjust application layer information.

Description

Technology Field

[0001] The present disclosure relates to the field of communication technology, and in particular to a method, apparatus, and readable storage medium for subscribing to model transmission state analysis in a network.

[0002] The present disclosure claims priority to a Chinese patent application filed with the Chinese Patent Office on November 24, 2021, with application number 202111407051.2 and titled “Method, apparatus and readable storage medium for subscribing to model transmission state analysis in a network,” all of which are incorporated by reference into the present disclosure. Background Technology

[0003] Recently, due to technological innovations in artificial intelligence, its applications are becoming increasingly widespread. In the case of mobile devices, strict limitations on energy consumption, computation, and memory costs make it impossible to execute heavyweight AI / Machine Learning models (hereinafter referred to as AI / ML models). Consequently, current methods require transmitting the inferences from many AI / ML models from the mobile device to the cloud or other devices; in other words, AI / ML models must be transmitted to the cloud or other devices.

[0004] In addition, considering issues such as privacy protection for transmitted data and alleviation of the data transmission load on the network, the demand for the transmission of AI / ML models is also increasing. Here, 5G systems serve as channels for transmitting AI / ML models, and in order to enhance the intelligence capabilities of 5G networks and satisfy the requirements for the transmission of AI / ML models in 5G systems as agreed upon in SA1 #93e under TS 22.261, 5G systems must support the monitoring of AI-ML sessions and the exposure of status information to third parties.

[0005] However, since conventional technology cannot effectively analyze the transmission status of AI / ML models, the network cannot effectively adjust its transmission strategy based on the transmission status of AI / ML models, and a third party cannot obtain the analysis of the transmission status of AI / ML models to perform adjustments to application layer information. The problem to be solved

[0006] The present disclosure provides a method, apparatus, and readable storage medium for subscribing to an analysis of a model transmission state in a network, thereby solving the technical problem in which an AI / ML model transmission state cannot be effectively analyzed, furthermore, a network cannot effectively adjust a network transmission strategy based on the AI / ML model transmission state, and a third party cannot obtain an analysis of the AI / ML model transmission state to perform adjustment of application layer information. means of solving the problem

[0007] In a first aspect, the present disclosure provides a method for subscribing to a model transmission state analysis in a network, said method applied to an application function (AF), said method,

[0008] Sending a first message to a Network Data Analysis Function (NWDAF) directly or through a Network Capability Exposure Function (NEF); wherein the first message is intended to request a subscription to analysis information regarding the transmission status of an Artificial Intelligence / Machine Learning (AI / ML) model in the network;

[0009] The method includes the step of receiving analysis information on the AI / ML model transmission status sent from the NWDAF directly or through the NEF, wherein the analysis information is determined based on data on the AI / ML model transmission status sent from other network functions (5GC NF(s)) of the 5G core network received by the NWDAF.

[0010] Here, the above analysis information is intended to adjust network strategy parameters and / or application layer model information.

[0011] Optionally, the data of the AI / ML model transmission status is obtained by sending a second message to the 5GC NF(s) based on the parameters requested in the first message received by the NWDAF, and the second message is for collecting data to analyze the AI / ML model transmission status in the network.

[0012] Optionally, the parameters requested in the first message include at least one of a network data analysis identifier, an identifier of a user device (UE) or a group of UEs receiving the AI / ML model or an identifier of any UE satisfying the analysis condition, an identifier of an application using the AI / ML model, a region where the AI / ML model is transmitted, a network slice instruction of a protocol data unit (PDU) session of a quality of service flow transmitting the AI / ML model, a data network instruction of a PDU session of a quality of service flow transmitting the AI / ML model, a duration of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, a size of the AI / ML model transmission, a quality of service requirement instruction of a quality of service flow transmitting the AI / ML model, and / or a specific quality of service requirement instruction for transmitting the AI / ML model;

[0013] The second message comprises at least one of the following: the current location of the UE using the AI / ML model, an identifier of the application using the AI / ML model, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate and a downlink bit rate transmitting the AI / ML model, an uplink packet delay and a downlink packet delay of the AI / ML model, a quantity of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, a quantity of packets transmitted by the AI / ML model, a quantity of packet retransmissions of the AI / ML model, a data collection time, a time length of AI / ML model transmission, a start timestamp of AI / ML model transmission, an end timestamp of AI / ML model transmission, a size of the AI / ML transmission model, a network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, a data network of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and a service process for the AF;

[0014] The above analysis information includes at least one of: a network slice of a PDU session of a Quality of Service flow for transmitting an AI / ML model, an identifier of an application using the AI / ML model, area information using the AI / ML model, the validity period of the analysis result, a User Plane Function (UPF) providing AI / ML model transmission, a data network name of a PDU session of a Quality of Service flow for transmitting an AI / ML model, the size of the AI / ML transmission model, the time length of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate for transmitting the AI / ML model and a downlink bit rate for transmitting the AI / ML model, an uplink packet delay of the AI / ML model and a downlink packet delay of the AI / ML model, the number of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, the number of times the Quality of Service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission, the number of packet transmissions of the AI / ML model, and the number of packet retransmissions of the AI / ML model;

[0015] Herein, when an AI / ML model performs federated learning, the parameter requested in the first message further includes federated learning group information, and the federated learning group information includes at least one of an identifier for indicating the federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0016] Correspondingly, the second message further includes at least one of an identifier for indicating a federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0017] Correspondingly, the analysis information further includes at least one of an identifier for directing a federated learning group to be analyzed, an identifier of a UE or UE(s) participating in federated learning, and an identifier of each application providing an AI / ML model or participating in federated learning.

[0018] In an embodiment of the present disclosure, a request is sent to NWDAF to subscribe to analysis information on the AI / ML model transmission status in a network, and analysis information confirmed based on the collected 5GC NF(s) data sent by NWDAF is received, and network strategy parameters and / or application layer model information is adjusted based on the analysis information, thereby effectively analyzing the AI / ML model transmission status so that the network can effectively adjust the network transmission strategy based on the AI / ML model transmission status, and a third party obtains the analysis of the AI / ML model transmission status and performs adjustment of the application layer information.

[0019] Optionally, after receiving the analysis information, the method,

[0020] Based on the analysis information above, the method further includes the step of sending a first request to the strategy control function (PCF) directly or through the NEF;

[0021] Here, the first request is to request an update to network strategy parameters for AI / ML model transmission; and the network strategy parameters are to optimize the AI / ML model transmission state.

[0022] Optionally, the step of sending a first request to the strategy control function (PCF) directly or through the NEF based on the analysis information is:

[0023] Based on at least one of the analysis information above, the uplink direction bit rate and downlink direction bit rate for transmitting the AI / ML model, the uplink direction packet delay and downlink direction packet delay of the AI / ML model, the quantity of abnormal releases of the quality of service flow during the period of transmission of the AI / ML model, the quantity of packet transmissions of the AI / ML model, the quantity of packet retransmissions of the AI / ML model, and the number of times the quality of service flow reached a reporting threshold for abnormal releases during the period of transmission of the AI / ML model, a new quality of service parameter for transmitting the AI / ML model is determined; the new quality of service parameter includes at least one of a 5G quality of service identifier, reflective quality of service control, the uplink direction maximum bit rate for transmitting the AI / ML model, the downlink direction maximum bit rate for transmitting the AI / ML model, the uplink direction minimum bit rate for transmitting the AI / ML model, the downlink direction minimum bit rate for transmitting the AI / ML model, and the priority of the quality of service flow;

[0024] A step of determining the domain information and address information of the UE(s) transmitting the AI / ML model and each AF based on the identifier of the application using the AI / ML model, the domain information using the AI / ML model, the IP address information of the application service using the AI / ML model, the network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and the data network name of the PDU session of the Quality of Service flow for transmitting the AI / ML model among the above analysis information; or, if the AI / ML model performs federated learning, determining the domain information and address information of the UE(s) transmitting the AI / ML model and each AF based on the identifier indicating the federated learning group being analyzed among the above analysis information, the identifier of the UE or UE(s) participating in federated learning, and the identifier of each application providing the AI / ML model or participating in federated learning.

[0025] A step of determining a data network access identifier (DNAI) and the region information and address information of the UE(s) and the AF corresponding to the DNAI based on the region information and address information of the respective AF and the UE(s) transmitting the AI / ML model, and the region information and address information of the UE(s) and the AF corresponding to the DNAI is for providing a path that optimizes the AI / ML model transmission state;

[0026] It includes the step of sending the new service quality parameters, the DNAI, the UE(s) corresponding to the DNAI, and the area information and address information of each AF to the PCF directly or through the NEF as parameters requested in the first request.

[0027] Optionally, the step of determining a data network access identifier (DNAI) and the area information and address information of the UE(s) and the AF corresponding to the DNAI, based on the area information and address information of the UE(s) and the AF respectively transmitting the AI / ML model, is as follows:

[0028] A step of determining whether the current routing path is poor based on the area information and address information of the UE(s) transmitting the AI / ML model and each AF;

[0029] If the current routing path is poor, a step of determining the destination addresses of both parties during transmission of the AI / ML model based on the address information of the UE(s) transmitting the AI / ML model and each AF, and the area information of the UE(s);

[0030] A step of determining the nearest path based on the above destination address;

[0031] Based on the nearest path, the method includes the step of determining the region information and address information of the DNAI, the UE(s) corresponding to the DNAI, and each AF.

[0032] Optionally, the above first request specifically,

[0033] The PCF requests that the 5G Service Quality identifier, reflective Service Quality control, the maximum bit rate in the uplink direction for transmitting AI / ML models, the maximum bit rate in the downlink direction for transmitting AI / ML models, the minimum bit rate in the uplink direction for transmitting AI / ML models, the minimum bit rate in the downlink direction for transmitting AI / ML models, and the priority of the Service Quality flow among the PCC rules be adjusted based on the new Service Quality parameters, and instructs that the PCF provide feedback on the first update result after adjustment directly or through the NEF; wherein the first update result is determined by the result of the PCF adjusting the PCC rules based on the new Service Quality parameters;

[0034] Correspondingly, the above method is,

[0035] The method further includes the step of receiving the first update result sent from the PCF directly or via NEF, wherein the first update result includes whether the first request was accepted or rejected.

[0036] Optionally, the above first request specifically,

[0037] PCF requests that the Session Management Function (SMF) network element determine whether the session management strategy needs to be updated, and if it is determined that the session management strategy needs to be updated, PCF determines to send a second request to the SMF, wherein the parameters requested in the second request include at least one of DNAI, a traffic steering policy identifier, and traffic routing information; the second request is intended to determine the User Plane Function (UPF) selected by the SMF based on the new session management strategy and to provide the corresponding DNAI, traffic steering policy identifier, and traffic routing information;

[0038] Correspondingly, the above method is,

[0039] The method further includes the step of receiving a second update result sent from PCF directly or via NEF, wherein the second update result is determined by PCF based on whether a new session management strategy sent from SMF has updated the UPF path;

[0040] Here, the second update result includes whether the second request was accepted or rejected.

[0041] In an embodiment of the present disclosure, a new service quality parameter is determined based on analysis information, and the new service quality parameter is sent to the PCF so that the PCF adjusts the PCC rule accordingly based on the new service quality parameter or updates the SM strategy through the SMF and provides the domain information and address information of the DNAI and the UE(s) corresponding to the DNAI and each AF, and by receiving a notification to indicate that the first request sent from the PCF has been accepted or rejected, a request is implemented to adjust the network strategy based on the NWDAF analysis results to optimize the AI / ML model transmission state.

[0042] Optionally, after receiving the analysis information, the method,

[0043] Based on the analysis information above, the information of the application layer model is adjusted, wherein the information of the application layer model includes at least one of model compression, model size, model transmission period, model encoding, and decoding; and the information of the application layer model is for updating service quality parameters.

[0044] A step of determining new service quality parameters based on information of the application layer model after adjustment, wherein the new service quality parameters include a 5G service quality identifier, reflective service quality control, a maximum bit rate in the uplink direction for transmitting an AI / ML model, a maximum bit rate in the downlink direction for transmitting an AI / ML model, a minimum bit rate in the uplink direction for transmitting an AI / ML model, a minimum bit rate in the downlink direction for transmitting an AI / ML model, and a priority of the service quality flow;

[0045] The method further includes the step of sending a third request to the strategy control function (PCF) directly or through the NEF;

[0046] Here, the parameter requested in the third request includes the new service quality parameter, and the third request is intended to request an update to the service quality parameter.

[0047] Optionally, the above third request specifically,

[0048] It is intended to request that the PCF adjust the 5G Quality of Service identifier, reflective Quality of Service control, the maximum bit rate in the uplink direction for transmitting AI / ML models, the maximum bit rate in the downlink direction for transmitting AI / ML models, the minimum bit rate in the uplink direction for transmitting AI / ML models, the minimum bit rate in the downlink direction for transmitting AI / ML models, and the priority of Quality of Service flows among the PCC rules based on the above-mentioned new Quality of Service parameters;

[0049] Correspondingly, the above method is,

[0050] The method further includes the step of receiving a third update result sent from the PCF directly or through the NEF, wherein the third update result is determined based on the result of the PCF adjusting the PCC rule, and wherein the third update result includes whether the third request was accepted or rejected.

[0051] Optionally, after adjusting the information of the application layer model, the method,

[0052] The method further includes the step of directly sending model compression, model size, and model encoding and decoding among the information of the application layer model after adjustment to the PCF, and said information of the application layer model after adjustment is intended to support the PCF in adjusting the 5G Quality of Service identifier, reflective Quality of Service control, the maximum bit rate in the uplink direction for transmitting AI / ML models, the maximum bit rate in the downlink direction for transmitting AI / ML models, the minimum bit rate in the uplink direction for transmitting AI / ML models, the minimum bit rate in the downlink direction for transmitting AI / ML models, and the priority of the Quality of Service flow among the PCC rules;

[0053] Correspondingly, the above method is,

[0054] The method further includes the step of receiving a fourth update result sent by the PCF, wherein the fourth update result is determined based on the result of adjusting the PCC rule based on the information of the application layer model after adjustment by the PCF, and wherein the fourth update result includes the acceptance of the third request or the rejection of the third request.

[0055] Optionally, after adjusting the information of the application layer model, the method,

[0056] The method further includes the step of directly sending the model transmission period among the information of the application layer model after adjustment to the PCF, wherein the model transmission period among the information of the application layer model after adjustment is intended to support the PCF in adjusting the gate state parameters among the PCC rules; and wherein the gate state parameters are intended to support the SMF in updating the session management strategy based on the transmission start time and transmission end time among the gate states;

[0057] Correspondingly, the above method is,

[0058] Receiving a fifth update result sent by PCF, wherein the fifth update result is confirmed by PCF receiving the result of a new session management strategy sent by SMF, and further including a step in which the fifth update result includes that the third request was accepted or rejected.

[0059] In an embodiment of the present disclosure, based on analysis information, information of the application layer model is adjusted, and new quality of service parameters are determined based on the information of the application layer model after adjustment, and the new quality of service parameters are sent to the PCF so that the PCF adjusts the PCC rules accordingly based on the new quality of service parameters; or, information of the application layer model after direct adjustment is sent to the PCF so that the PCF adjusts the quality of service parameters among the PCC rules based on model compression, model size, model encoding and decoding, or, the PCF adjusts the gate state parameters among the PCC rules based on the model transmission period so that the SMF updates the session management strategy based on the transmission start time and transmission end time among the gate states, and by receiving a notification indicating whether a third request sent from the PCF has been accepted or rejected, it is possible to implement adjusting information of the application layer model based on the NWDAF analysis results and further updating QoS requirements to optimize the AI / ML model transmission state.

[0060] In a second aspect, the present disclosure provides a method for subscribing to a model transmission state analysis in a network, said method applied to a network data analysis function (NWDAF), said method,

[0061] Receiving a first message sent from an application function (AF) directly or through a network capability exposure function (NEF); wherein the first message is for requesting a subscription to analysis information of the transmission status of an artificial intelligence / machine learning (AI / ML) model in the network;

[0062] A step of sending a second message to another network function 5GC NF(s) of the 5G core network based on the parameters requested in the first message, wherein the second message is for collecting data to analyze the transmission status of an AI / ML model in the network;

[0063] The method includes the step of receiving data on the AI / ML model transmission status transmitted from other network functions 5GC NF(s) of the 5G core network, analyzing the data on the AI / ML model transmission status, and obtaining analysis information on the AI / ML model transmission status.

[0064] Here, the above analysis information is intended to adjust network strategy parameters and / or application layer model information through AF.

[0065] Optionally, the parameters requested in the first message include at least one of a network data analysis identifier, an identifier of a user device (UE) or a group of UEs receiving the AI / ML model or an identifier of any UE satisfying the analysis condition, an identifier of an application using the AI / ML model, a region where the AI / ML model is transmitted, a network slice instruction of a protocol data unit (PDU) session of a quality of service flow transmitting the AI / ML model, a data network instruction of a PDU session of a quality of service flow transmitting the AI / ML model, a duration of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, a size of the AI / ML model transmission, a quality of service requirement instruction of a quality of service flow transmitting the AI / ML model, and / or a specific quality of service requirement instruction for transmitting the AI / ML model;

[0066] The second message comprises at least one of the following: the current location of the UE using the AI / ML model, an identifier of the application using the AI / ML model, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate and a downlink bit rate transmitting the AI / ML model, an uplink packet delay and a downlink packet delay of the AI / ML model, a quantity of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, a quantity of packets transmitted by the AI / ML model, a quantity of packet retransmissions of the AI / ML model, a data collection time, a time length of AI / ML model transmission, a start timestamp of AI / ML model transmission, an end timestamp of AI / ML model transmission, a size of the AI / ML transmission model, a network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, a data network of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and a service process for the AF;

[0067] The above analysis information includes at least one of: a network slice of a PDU session of a Quality of Service flow for transmitting an AI / ML model, an identifier of an application using the AI / ML model, area information using the AI / ML model, the validity period of the analysis result, a User Plane Function (UPF) providing AI / ML model transmission, a data network name of a PDU session of a Quality of Service flow for transmitting an AI / ML model, the size of the AI / ML transmission model, the time length of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate for transmitting the AI / ML model and a downlink bit rate for transmitting the AI / ML model, an uplink packet delay of the AI / ML model and a downlink packet delay of the AI / ML model, the number of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, the number of times the Quality of Service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission, the number of packet transmissions of the AI / ML model, and the number of packet retransmissions of the AI / ML model;

[0068] Herein, when an AI / ML model performs federated learning, the parameter requested in the first message further includes federated learning group information, and the federated learning group information includes at least one of an identifier for indicating the federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0069] Correspondingly, the second message further includes at least one of an identifier for indicating a federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0070] Correspondingly, the analysis information further includes at least one of an identifier for directing a federated learning group to be analyzed, an identifier of a UE or UE(s) participating in federated learning, and an identifier of each application providing an AI / ML model or participating in federated learning.

[0071] In a third aspect, the present disclosure provides a device for subscribing to model transmission state analysis in a network, said device comprising a memory, a transceiver, and a processor;

[0072] The memory stores a computer program; the transceiver transmits and receives data under the control of the processor; and the processor reads the computer program in the memory,

[0073] A first message is sent to the Network Data Analysis Function (NWDAF) either directly or through the Network Capability Exposure Function (NEF); wherein the first message is intended to request a subscription to analysis information regarding the transmission status of an Artificial Intelligence / Machine Learning (AI / ML) model in the network;

[0074] It is for performing an operation in which analysis information of the AI / ML model transmission status sent from the NWDAF directly or through the NEF, said analysis information is determined based on data of the AI / ML model transmission status sent from other network functions 5GC NF(s) of the 5G core network received by the NWDAF, and

[0075] Here, the above analysis information is intended to adjust network strategy parameters and / or application layer model information.

[0076] In a fourth aspect, the present disclosure provides a device for subscribing to model transmission state analysis in a network, the device comprising a memory, a transceiver, and a processor;

[0077] The memory stores a computer program; the transceiver transmits and receives data under the control of the processor; and the processor reads the computer program in the memory,

[0078] Receives a first message sent from an application function (AF) directly or through a network capability exposure function (NEF); wherein the first message is intended to request a subscription to analysis information regarding the transmission status of an artificial intelligence / machine learning (AI / ML) model in the network;

[0079] Based on the parameters requested in the first message above, a second message is sent to other network functions 5GC NF(s) of the 5G core network, and the second message is intended to collect data for analyzing the transmission status of an AI / ML model in the network;

[0080] It is intended to perform an operation of receiving data on the AI / ML model transmission status transmitted from other network functions 5GC NF(s) of the 5G core network, analyzing the data on the AI / ML model transmission status, and obtaining analysis information on the AI / ML model transmission status.

[0081] Here, the above analysis information is intended to adjust network strategy parameters and / or application layer model information through AF.

[0082] In a fifth aspect, the present disclosure provides a device for subscribing to model transmission state analysis in a network, said device,

[0083] A sending unit that sends a first message to a network data analysis function (NWDAF) directly or through a network capability exposure function (NEF); wherein the first message is intended to request a subscription to analysis information regarding the transmission status of an artificial intelligence / machine learning (AI / ML) model in the network;

[0084] It includes an analysis unit that receives analysis information on the AI / ML model transmission status sent from the NWDAF directly or through the NEF, wherein the analysis information is determined based on data on the AI / ML model transmission status sent from other network functions (5GC NF(s)) of the 5G core network received by the NWDAF.

[0085] Here, the above analysis information is intended to adjust network strategy parameters and / or information on the application layer model.

[0086] In the sixth aspect, a device for subscribing to model transmission state analysis in a network is provided, said device,

[0087] A receiving unit that receives a first message sent from an application function (AF) directly or through a network capability exposure function (NEF); wherein the first message is intended to request a subscription to analysis information regarding the transmission status of an artificial intelligence / machine learning (AI / ML) model in a network;

[0088] A sending unit that sends a second message to another network function 5GC NF(s) of the 5G core network based on parameters requested in the first message, wherein the second message is for collecting data to analyze the transmission status of an AI / ML model in the network;

[0089] Includes an analysis unit that receives data on the AI / ML model transmission status transmitted from other network functions 5GC NF(s) of the 5G core network, analyzes the data on the AI / ML model transmission status, and obtains analysis information on the AI / ML model transmission status.

[0090] Here, the above analysis information is intended to adjust network strategy parameters and / or application layer model information through AF.

[0091] In a seventh aspect, the present disclosure provides a processor-readable storage medium, said processor-readable storage medium having a computer program stored therein, said computer program being for the processor to perform a method according to either the first aspect or the second aspect. Effects of the invention

[0092] The present disclosure provides a method, apparatus, and readable storage medium for subscribing to an analysis of a model transmission state in a network, and sending a first message to a network data analysis function (NWDAF) directly or through a network capability exposure function (NEF); wherein the first message is for requesting subscription to analysis information of an artificial intelligence / machine learning (AI / ML) model transmission state in a network; receiving the analysis information of an AI / ML model transmission state sent by the NWDAF directly or through the NEF, wherein the analysis information is determined based on data of an AI / ML model transmission state sent by another network function 5GC NF(s) of a 5G core network received by the NWDAF, and wherein the data of an AI / ML model transmission state is obtained by sending a second message to the 5GC NF(s) based on parameters requested in the first message received by the NWDAF, and the second message is for collecting data for analyzing an AI / ML model transmission state in a network; wherein the analysis information is for adjusting network strategy parameters and / or information of an application layer model. A request is sent to NWDAF to subscribe to analysis information on the transmission status of AI / ML models in the network, and analysis information confirmed based on the collected 5GC NF(s) data sent by NWDAF is received, and network strategy parameters and / or application layer model information is adjusted based on the analysis information, thereby effectively analyzing the transmission status of AI / ML models so that the network can effectively adjust the network transmission strategy based on the transmission status of AI / ML models, and a third party obtains the analysis of the transmission status of AI / ML models and performs adjustment of application layer information.

[0093] It should be understood that the description in the above-described section of the invention is not intended to limit the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure become easier to understand through the description below. Brief explanation of the drawing

[0094] In order to more clearly explain the technical methods according to the present disclosure or the prior art, the attached drawings to be used in describing the embodiments or prior art are briefly introduced below. The accompanying drawings described below are merely some embodiments of the present disclosure, and it is obvious that those skilled in the art can derive other accompanying drawings from these drawings without creative effort. FIG. 1 is a network configuration diagram of a method for subscribing to model transmission state analysis in a network provided in an embodiment of the present disclosure. FIG. 2 is a network configuration diagram of a 5GC supporting network data analysis provided in an embodiment of the present disclosure. FIG. 3 is a flowchart of a method for subscribing to model transmission state analysis in a network provided in Embodiment 1 of the present disclosure. FIG. 4 is a signaling flowchart of a method for subscribing to model transmission state analysis in a network provided in Embodiment 1 of the present disclosure. FIG. 5 is a signaling flowchart of a method for subscribing to model transmission state analysis in a network provided in Embodiment 2 of the present disclosure. FIG. 6 is a signaling flowchart of a method for subscribing to model transmission state analysis in a network provided in Embodiment 3 of the present disclosure. FIG. 7 is a flowchart of a method for subscribing to model transmission state analysis in a network provided in Embodiment 4 of the present disclosure. FIG. 8 is a structural diagram of a device for subscribing to model transmission state analysis in a network provided in an embodiment of the present disclosure. FIG. 9 is a structural diagram of a device for subscribing to model transmission state analysis in a network provided in another embodiment of the present disclosure. Fig. 10 is a structural diagram of a device for subscribing to model transmission state analysis in a network provided in yet another embodiment of the present disclosure. Fig. 11 is a structural diagram of a device for subscribing to model transmission state analysis in a network provided in yet another embodiment of the present disclosure. Specific details for implementing the invention

[0095] In this disclosure, the term “and / or” indicates a relationship between related objects, indicating that three relationships may exist. For example, A and / or B may represent three cases: A existing alone, A and B existing together, or B existing alone. The symbol “ / ” generally indicates that the related objects are in an “or” relationship.

[0096] Below, with reference to the accompanying drawings according to the embodiments of the present disclosure, technical solutions according to the embodiments of the present disclosure are described clearly and completely. It is obvious that the described embodiments are only partial embodiments of the present disclosure and not all embodiments. Based on the embodiments according to the present disclosure, other embodiments obtained by a person skilled in the art without creative labor are all within the scope of protection of the present disclosure.

[0097] To ensure a clear understanding of the technical solution of the present disclosure, the prior art solution is first introduced in detail. In the prior art, under the SA1 R18 requirements agreed upon in SA#93e, there are at least several cases requiring the transmission of AI / ML models, such as the following:

[0098] Scenario 1, deployment and sharing of AI / ML models. Due to changes in tasks or environments, the memory of the mobile terminal is limited, so all models cannot be loaded in advance. Therefore, the mobile terminal must download new AI / ML models in real time from the network via the 5G system.

[0099] Scenario 2, Federated learning algorithm using 5GS. When the cloud server trains a single global model, each terminal must integrate the model trained locally. Training iteration process: A terminal downloads a single global model from the cloud server and trains it using local data; the terminal reports intermediate training results to the cloud server; the cloud server integrates intermediate training results from all terminals, updates the global model, and then redistributes the global model to the terminals; the terminal performs the next iteration.

[0100] Scenario 3, Splitting of an AI / ML model between AI / ML endpoints. A single AI / ML model can be split into multiple parts based on the current task or environment. Parts that are computationally complex and energy-intensive are inferred by the network, while parts requiring privacy protection or sensitive to latency are inferred by the terminal. For example, the terminal downloads / loads a model, first infers a few specific layers / parts, and then sends the intermediate results to the network; the network then proceeds to execute the remaining layers / parts and feeds the inference results back to the terminal. Since this scenario involves transmitting partial models at the initial stage or in the middle, it may include the transmission of the model.

[0101] Therefore, as a channel for transmitting AI / ML models, the 5G system must support monitoring and exposing status information of AI / ML sessions to third parties in order to enhance the intelligence capabilities of the 5G network and satisfy the requirements for transmitting AI / ML models in the 5G system as agreed upon in SA1 #93e in TS 22.261. However, currently, there is no analysis of the AI / ML model transmission status, third parties cannot effectively coordinate their own behavior based on the AI / ML model transmission status, and the network cannot effectively coordinate its network state based on the AI / ML model transmission status.

[0102] As a result of further research, the inventor discovered that in order to effectively analyze the transmission state of an AI / ML model, interaction between the Application Function (abbreviated as 'AF'), the Network Exposure Function (abbreviated as 'NEF'), the Network Data Analytic Function (abbreviated as 'NWDAF'), and each Network Function (abbreviated as 'NF') is required. As illustrated in FIG. 1, AF may send a request to NWDAF to indicate subscribing to an analysis of the AI / ML model transmission status in the network, either directly or through NEF. NWDAF collects data from each network function (abbreviated as Network Function, 'NF') (i.e., NF(s)) within the 5G Core Network (abbreviated as 5G Core Network, '5GC') to analyze and provide feedback on the AI / ML model transmission status in the network. By implementing an effective analysis of the AI / ML model transmission status, NWDAF can effectively adjust the network status based on the AI / ML model transmission status in the network, and enable a third party to obtain the analysis of the AI / ML model transmission status and perform adjustment of its own behavioral data.

[0103] Accordingly, based on the aforementioned advanced research of the inventor, a method for subscribing to an analysis of a model transmission state in a network as presented in this disclosure is provided, wherein a first message is sent to a Network Data Analysis Function (NWDAF) either directly or through a Network Capability Exposure Function (NEF); wherein the first message is intended to request subscription to analysis information of an artificial intelligence / machine learning (AI / ML) model transmission state in a network; wherein the analysis information of an AI / ML model transmission state sent by the NWDAF is received either directly or through the NEF, and the analysis information is determined based on receiving data of an AI / ML model transmission state sent by another network function of a 5G core network (i.e., 5GC NF(s)) received by the NWDAF, and the data of the AI / ML model transmission state is obtained by sending a second message to the 5GC NF(s) based on parameters requested in the first message received by the NWDAF, and the second message is intended to collect data for analyzing the AI / ML model transmission state in a network; Here, the analysis information is intended to adjust network strategy parameters and / or application layer model information. A request is sent to NWDAF to subscribe to analysis information on the AI / ML model transmission status in the network, and analysis information confirmed based on collected 5GC NF(s) data sent by NWDAF is received. By adjusting network strategy parameters and / or application layer model information based on the analysis information, the AI / ML model transmission status is effectively analyzed, enabling the network to effectively adjust the network transmission strategy based on the AI / ML model transmission status, and allowing a third party to obtain the analysis of the AI / ML model transmission status and perform adjustment of application layer information.

[0104] FIG. 2 is a network configuration diagram of a 5GC that supports network data analysis provided in an embodiment of the present disclosure. As shown in FIG. 2, in an embodiment of the present disclosure, NWDAF is a network analysis function managed by an operator, and NWDAF can provide data analysis services for 5GC network functions, application functions (abbreviated as 'AF'), and operation administration and maintenance (abbreviated as 'OAM'). Here, the analysis results may be historical statistical information or predictive information. NWDAF can provide services to one or more network slices.

[0105] Here, 5GC includes various other functions. These are, respectively: User Plane Function (abbreviated as 'UPF'), Session Management Function (abbreviated as 'SMF'), Access and Mobility Management Function (abbreviated as 'AMF'), Unified Data Repository (abbreviated as 'UDR'), Network Exposure Function (abbreviated as 'NEF'), AF, Policy Control Function (abbreviated as 'PCF'), and Online Charging System (abbreviated as 'OCS'). Here, all of these other functions can be collectively referred to as NFs. NWDAF communicates with other function entities (5GC NF(s)) and OAMs within the 5G core network based on the Servification Interface.

[0106] Different NWDAF instances exist within 5GC that can provide different types of dedicated analytics. To enable consumer NFs to discover suitable NWDAF instances that can provide specific types of analytics, NWDAF instances must provide the Analytic ID they support when registering with the Network Repository Function (abbreviated as 'NRF'), and the Analytic ID represents the type of analytics (or analytics identifier). Thus, when consumer NFs query the NRF for NWDAF instances, they can provide the Analytic ID to indicate what type of analytics is required. 5GC network functions and OAMs determine how to use the data analytics provided by the Network Data Analytics Function (NWDAF) to improve network performance.

[0107] In an embodiment of the present disclosure, in one application scenario, AF requests that NWDAF provide an analysis of the AI / ML model delivery status, and the analysis result (or analysis information) includes an application identifier (i.e., Application ID) using the AI / ML model, area information using the AI / ML model, duration of delivery of the AI / ML model, size of the AI / ML delivery model, information regarding the Quality of Service (i.e., QoS) of delivery of the AI / ML model, network slice used for delivery of the AI / ML model, and data network name (abbreviated as 'DNN'); if federated learning is present, it further includes a group identifier (i.e., federated learning group ID), UE ID or UE group ID participating in federated learning, and address information of the application server providing the model or participating in federated learning. AF requests that the network strategy of the 5GS be adjusted or the application layer AI / ML model parameters be adjusted based on the data analysis provided by NWDAF to optimize the AI / ML model delivery status, and the 5GC NF(s) adjust the network strategy in accordance with the request or AF adjusts the application layer AI / ML model parameters based on the data analysis.

[0108] Here, when AF sends an analysis request to NWDAF, if AF is in a trusted area, AF can send the request directly to NWDAF; if AF is not in a trusted area, AF can send the request to NWDAF through NEF, that is, AF sends the request to NEF, and furthermore, NEF sends the request to NWDAF.

[0109] Accordingly, a request is sent to NWDAF to subscribe to analysis information on the AI / ML model transmission status in the network, either directly or through NEF, and analysis information confirmed based on the collected 5GC NF(s) data sent by NWDAF is received, and network strategy parameters and / or application layer model information is adjusted based on the analysis information, thereby effectively analyzing the AI / ML model transmission status so that the network can effectively adjust the network transmission strategy based on the AI / ML model transmission status, and a third party obtains the analysis of the AI / ML model transmission status and performs adjustment of application layer information.

[0110] Embodiments of the present disclosure are described below with reference to the attached drawings.

[0111] FIG. 3 is a flowchart of a method for subscribing to a model transmission state analysis in a network provided in Embodiment 1 of the present disclosure. As illustrated in FIG. 3, if the entity performing the method for subscribing to a model transmission state analysis in a network provided in the present embodiment is AF, the method for subscribing to a model transmission state analysis in a network provided in the embodiment of the present disclosure includes the following steps.

[0112] Step (101), AF sends a first message to the network data analysis function (NWDAF) either directly or through the network capability exposure function (NEF).

[0113] Here, the first message is intended to request a subscription to analysis information regarding the transmission status of an artificial intelligence / machine learning (AI / ML) model in a network.

[0114] In this embodiment, the parameters requested in the first message are a network data analysis identifier (i.e., Analytics ID), an identifier of a single user device (UE) or a group of UEs receiving the AI / ML model, or an identifier of any UE satisfying the analysis condition (i.e., Target of Analytics Reporting), an identifier of the application using the AI / ML model (i.e., Application ID), an area where the AI / ML model is transmitted (i.e., Area of ​​Interest (AoI)), a network slice instruction of the Protocol Data Unit (PDU) session of the Quality of Service flow transmitting the AI / ML model (i.e., S-NSSAI), a data network instruction of the PDU session of the Quality of Service flow transmitting the AI / ML model (i.e., DNN), a duration of the AI / ML model transmission (i.e., Model transmission duration), a start timestamp of the AI / ML model transmission (i.e., Model transmission start), a end timestamp of the AI / ML model transmission (i.e., Model transmission stop), a size of the AI / ML transmission model (i.e., Model size), a Quality of Service requirement instruction of the Quality of Service flow transmitting the AI / ML model (i.e., 5G QoS Identifier (5QI)), and / or It includes at least one of the specific service quality requirement directives (i.e., QoS Characteristics) for transmitting AI / ML models.

[0115] Here, specific service quality requirements include, for example, transmission delay, packet error rate, etc. Please refer to the example table of parameters requested in the first message shown in Table 1 below.

[0116] Example table of parameters requested in the first message Parameter Name (i.e., parameter names) Parameter Value (i.e., parameter value) Parameter Description (i.e., parameter description) Analytics ID Define a new Analytics ID, e.g., AI / ML Model Transfer Status or AI / ML Model Transfer Performance; Network data analysis identifier Target of Analytics Reporting UE ID, UE group ID, or "any UE" Identifier of a single UE or a group of UEs receiving an AI / ML model, or identifier of any UE satisfying an analysis condition (e.g., within an AoI). Application ID Application ID of the application using the AI / ML model Identifier of an application using AI / ML models AoI (Area of ​​Interest) The area where AI / ML models are transmitted S-NSSAI Network slice instructions for the PDU session of the QoS flow transmitting AI / ML models DNN Data network instructions for PDU sessions of QoS flows transmitting AI / ML models Model transmission duration Period of AI / ML model transfer Model transmission start Start timestamp for AI / ML model transfer Model transmission stop AI / ML model transfer end timestamp Model size Size of AI / ML transfer models QoS requirements Include one or more of the parameters below: 5QI (5G QoS Identifier) The 5QI of the QoS flow transmitting the AI / ML model, and additional parameters applied (e.g., GFBR of the GBR 5QI (guaranteed flow bit rate), etc.) QoS request instructions for QoS flows transmitting AI / ML models QoS Characteristics Resource Type (GBR, non-GBR), Packet Delay Budget (PDB), and Packet Error Rate (PER) values ​​of the QoS flow Specify specific QoS requirements for transmitting AI / ML models, such as packet transmission delay, packet error rate, etc.

[0117] If learning exists, the parameters requested in the first message may further include Federated Learning (FL) group information; the Federated Learning (FL) group information includes at least one of an identifier for indicating the Federated Learning group to be analyzed (i.e., Federated Learning (FL) group ID), an identifier of a UE or UE(s) participating in the Federated Learning (i.e., Federated Learning (FL) UE ID or UE group ID), and an identifier of an application participating in the Federated Learning (i.e., Federated Learning (FL) Application ID). Please refer to the example table of parameters requested in the first message shown in Table 2 below.

[0118] Example table of parameters requested in the first message Parameter Name Parameter Value Parameter Description Analytics ID Define a new Analytics ID, e.g., AI / ML Model Transfer Status or AI / ML Model Transfer Performance; Network data analysis identifier Target of Analytics Reporting UE ID, UE group ID, or "any UE" Identifier of a single UE or a group of UEs receiving an AI / ML model, or identifier of any UE satisfying an analysis condition (e.g., within an AoI). Application ID Application ID of the application using the AI / ML model Identifier of an application using AI / ML models AoI (Area of ​​Interest) The area where AI / ML models are transmitted S-NSSAI Network slice instructions for the PDU session of the QoS flow transmitting AI / ML models DNN Data network instructions for PDU sessions of QoS flows transmitting AI / ML models Model transmission duration Period of AI / ML model transfer Model transmission start Start timestamp for AI / ML model transfer Model transmission stop AI / ML model transfer end timestamp Model size Size of AI / ML transfer models Federated Learning(FL) group information If available, federated learning group information >Federated Learning(FL) group ID If present, federated learning group identifier, federated learning group instructions to analyze >Federated Learning (FL) UE ID or UE group ID If so, UE(s) participating in federated learning >Federated Learning(FL) Application ID If present, the application identifier participating in federated learning QoS requirements Include one or more of the parameters below: 5QI (5G QoS Identifier) The 5QI of the QoS flow transmitting the AI / ML model, and additional parameters applied (e.g., GFBR of the GBR 5QI (guaranteed flow bit rate), etc.) QoS request instructions for QoS flows transmitting AI / ML models QoS Characteristics Resource Type (GBR, non-GBR), Packet Delay Budget (PDB), and Packet Error Rate (PER) values ​​of the QoS flow Specify specific QoS requirements for transmitting AI / ML models, such as packet transmission delay, packet error rate, etc.

[0119] In this embodiment, if the AF is untrusted (i.e., the AF is not in a trusted area), the AF sends a request to the NEF to subscribe to the AI / ML model exposure status, such as Nnef_AnalyticsExposure_Subscribe (i.e., subscribe to analytics exposure) or Nnef_AnalyticsExposure_Fetch (i.e., obtain analytics exposure). The NEF sends a first message to the NWDAF, which may be a request to subscribe to the AI / ML model exposure status Nnwdaf_AnalyticsSubscription_Subscribe (i.e., subscribe to analytics) or Nnwdaf_AnalyticsInfo_Request (i.e., analytics information), and which may request a subscription to analytics information of the AI / ML model exposure status on the network by carrying the parameters shown in the table. If the AF is trusted (i.e., the AF is in a trusted area), the AF sends the first message directly to the NWDAF.

[0120] Here, the subscription request for the AI / ML model transmission status can request a subscription for analysis information of the AI / ML model transmission status in the network by carrying parameters in Table 1 or Table 2.

[0121] Step (102), AF receives analysis information of the AI / ML model transmission status sent from the NWDAF directly or through the NEF.

[0122] Here, the analysis information is determined based on data of the AI / ML model transmission status sent from other network functions 5GC NF(s) of the 5G core network received by the NWDAF.

[0123] Optionally, the data of the AI / ML model transmission status is obtained by sending a second message to the 5GC NF(s) based on the parameters requested in the first message received by the NWDAF, and the second message is for collecting data to analyze the AI / ML model transmission status in the network.

[0124] In this embodiment, the second message includes the current location of the UE using the AI / ML model (i.e., UE location), the identifier of the application using the AI / ML model (i.e., Application ID, which may be a server identifier or an AF identifier), the Quality of Service flow identifier transmitting the AI / ML model (i.e., QFI), the uplink direction bit rate transmitting the AI / ML model (i.e., bit rate for UL direction) and the downlink direction bit rate transmitting the AI / ML model (i.e., bit rate for DL ​​direction), the uplink direction packet delay of the AI / ML model (i.e., Packet delay for UL direction) and the downlink direction packet delay of the AI / ML model (i.e., Packet delay for the DL direction), the quantity of abnormal releases of the Quality of Service flow (QoS Sustainability) during the duration of the AI / ML model transmission, the quantity of packet transmission of the AI / ML model, the quantity of packet retransmission of the AI / ML model (i.e., packet retransmission), the data collection time (i.e., Timestamp), the duration of the AI / ML model transmission (i.e., the duration of the AI / ML model transmission), and the start of the AI / ML model transmission. It includes at least one of a timestamp, an end timestamp of the AI / ML model transmission, the size of the AI / ML transmission model, a network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, a data network of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and a service process for the AF (i.e., IP filter information). Please refer to the second message example table shown in Table 3 below.

[0125] Second message example table Information (i.e., data) Source (i.e., source) Description (i.e., data explanation) UE location AMF Current location of UEs using AI / ML models Application ID SMF, AF Identifier of an application using AI / ML models QoS requirements >QFI SMF QoS flow identifier for transmitting AI / ML models QoS flow Bit Rate UPF Bit rate for the uplink direction for transmitting the AI / ML model (bit rate for UL direction); and bit rate for the downlink direction for transmitting the AI / ML model (bit rate for DL ​​direction). QoS flow Packet Delay UPF Packet delay for the UL direction of the AI / ML model; and packet delay for the DL direction of the AI / ML model. QoS Sustainability OAM TS 28.554

[10] Abnormal release quantity of QoS flow during AI / ML model transfer period Packet transmission UPF Packet transmission quantity of AI / ML models Packet retransmission UPF Packet retransmission quantity of AI / ML models Timestamp UPF, AF Data collection time Model transmission duration UPF, AF Time length of AI / ML model transfer Model transmission start UPF, AF Start timestamp for AI / ML model transfer Model transmission stop UPF, AF AI / ML model transfer end timestamp Model size UPF, AF Size of AI / ML transfer models S-NSSAI SMF Network slice of a PDU session of a QoS flow for transmitting AI / ML models DNN SMF DNN of PDU session of QoS flow for transmitting AI / ML models IP filter information AF Clarify the service flow used in the application

[0126] If federated learning exists, the second message further includes at least one of an identifier designation for specifying the federated learning group to be analyzed (i.e., Federated Learning (FL) group ID), a UE or UE(s) participating in federated learning (i.e., Federated Learning (FL) UE ID or UE group ID), and an application identifier participating in federated learning (i.e., Federated Learning (FL) Application ID). Please refer to the second message example table shown in Table 4 below.

[0127] Second message example table Information (i.e., data) Source (i.e., source) Description (i.e., data explanation) UE location AMF Current location of UEs using AI / ML models Application ID SMF, AF Identifier of an application using AI / ML models Federated Learning (FL) group ID SMF, AF If so, instructions for the federated learning group to analyze Federated Learning(FL) UE ID or UE group ID SMF, AF If so, UE(s) participating in federated learning Federated Learning (FL) Application ID SMF, AF If present, the application identifier participating in federated learning QoS requirements >QFI SMF QoS flow identifier for transmitting AI / ML models QoS flow Bit Rate UPF Uplink direction bit rate for transmitting AI / ML models (bit rate for UL direction); and downlink direction bit rate for transmitting AI / ML models (bit rate for DL ​​direction). QoS flow Packet Delay UPF Packet delay for the UL direction of the AI / ML model; and packet delay for the DL direction of the AI / ML model. QoS Sustainability OAM TS 28.554

[10] QoS flow abnormal release quantity during AI / ML model transfer period Packet transmission UPF Packet transmission quantity of AI / ML models Packet retransmission UPF Packet retransmission quantity of AI / ML models Timestamp UPF, AF Data collection time Model transmission duration UPF, AF Time length of AI / ML model transfer Model transmission start UPF, AF Start timestamp for AI / ML model transfer Model transmission stop UPF, AF AI / ML model transfer end timestamp Model size UPF, AF Size of AI / ML transfer models S-NSSAI SMF Network slice of a PDU session of a QoS flow for transmitting AI / ML models DNN SMF DNN of PDU session of QoS flow for transmitting AI / ML models IP filter information AF Clarify the service flow used in the application

[0128] In this embodiment, if the AF is unreliable, analysis information regarding the AI / ML model transmission status sent by the NEF is received, and said analysis information is sent by the NWDAF to the NEF. If the AF is reliable, analysis information regarding the AI / ML model transmission status sent directly by the NWDAF is received.

[0129] Herein, the analysis information includes at least one of a network slice of a PDU session of a Quality of Service flow for transmitting an AI / ML model, an identifier of an application using the AI / ML model, area information using the AI / ML model, a validity period of the analysis result (i.e., Validity period), a User Plane Function (UPF) providing AI / ML model transmission (i.e., UPF Info), a data network name of a PDU session of a Quality of Service flow for transmitting an AI / ML model, the size of the AI / ML transmission model, the duration of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, and Quality of Service requirements (i.e., QoS requirements); Quality of Service requirements include at least one of the following: a Quality of Service flow identifier (i.e., QFI) for transmitting the AI / ML model, an uplink bit rate and a downlink bit rate for transmitting the AI / ML model, an uplink packet delay and a downlink packet delay for the AI / ML model, the number of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, the number of times the reporting threshold for abnormal releases of the Quality of Service flow was reached during the period of AI / ML model transmission, the number of packet transmissions of the AI / ML model, and the number of packet retransmissions of the AI / ML model. Please refer to the example table of analysis information shown in Table 5 below.

[0130] Example table of analysis information Information Description List of analysis information on AI / ML model transmission status S-NSSAI Network slice of a PDU session of a QoS flow for transmitting AI / ML models Application ID Identifier of an application using AI / ML models Validity area Area information using AI / ML models Validity period Validity time of analysis results UPF Info UPF providing AI / ML model delivery DNN DNN of PDU session of QoS flow for transmitting AI / ML models Model size Size of AI / ML transfer models Model transmission duration Time length of AI / ML model transfer Model transmission start Start timestamp for AI / ML model transfer Model transmission stop AI / ML model transfer end timestamp QoS requirements QoS requirements >>QFI QoS flow identifier for transmitting AI / ML models >>QoS flow Bit Rate Uplink direction bit rate for transmitting AI / ML models (bit rate for UL direction); and downlink direction bit rate for transmitting AI / ML models (bit rate for DL ​​direction). >>QoS flow Packet Delay Packet delay for the UL direction of the AI / ML model; and packet delay for the DL direction of the AI / ML model. QoS Sustainability Abnormal release quantity of QoS flow during AI / ML model transfer period >>QoS Sustainability Reporting Threshold(s) Number of times the QoS flow reached the abnormal release reporting threshold during the AI / ML model transfer period >>Packet transmission Packet transmission quantity of AI / ML models >>Packet retransmission Packet retransmission quantity of AI / ML models

[0131] If federated learning exists, the analysis information further includes at least one of the following: an identifier for specifying the federated learning group to be analyzed, an identifier of a UE or UE(s) participating in the federated learning, and a reference to the identifier of each application providing the AI / ML model or participating in the federated learning (i.e., Application Server Instance Address). Please refer to the example analysis information table shown in Table 6 below.

[0132] Example table of analysis information Information Description List of analysis information on AI / ML model transmission status S-NSSAI Network slice of a PDU session of a QoS flow for transmitting AI / ML models Application ID Identifier of an application using AI / ML models Validity area Area information using AI / ML models Validity period Validity time of analysis results UPF Info UPF providing AI / ML model delivery DNN DNN of PDU session of QoS flow for transmitting AI / ML models Model size Size of AI / ML transfer models Model transmission duration Time length of AI / ML model transfer Model transmission start Start timestamp for AI / ML model transfer Model transmission stop AI / ML model transfer end timestamp >Federated Learning(FL) group ID If present, group ID instructions for federated learning >> List of FL UE(s) If present, UE ID or UE group ID participating in federated learning >>Application Server Instance Address If present, specify the Application Server Instance(s) (IP address of the Application Server) or FQDN of the Application Server that provide the model or participate in federated learning. QoS requirements QoS requirements >>QFI QoS flow identifier for transmitting AI / ML models >>QoS flow Bit Rate Uplink direction bit rate for transmitting AI / ML models (bit rate for UL direction); and downlink direction bit rate for transmitting AI / ML models (bit rate for DL ​​direction). >>QoS flow Packet Delay Packet delay for the UL direction of the AI / ML model; and packet delay for the DL direction of the AI / ML model. QoS Sustainability Quantity of abnormal QoS Flow releases during the AI / ML model transfer period >>QoS Sustainability Reporting Threshold(s) Number of times the QoS flow reached the abnormal release reporting threshold during the AI / ML model transfer period >>Packet transmission Packet transmission quantity of AI / ML models >>Packet retransmission Packet retransmission quantity of AI / ML models

[0133] In this embodiment, a request is sent to NWDAF to subscribe to analysis information on the AI / ML model transmission status in the network, and analysis information confirmed based on the collected 5GC NF(s) data sent by NWDAF is received, and network strategy parameters and / or application layer model information is adjusted based on the analysis information, thereby effectively analyzing the AI / ML model transmission status so that the network can effectively adjust the network transmission strategy based on the AI / ML model transmission status, and a third party obtains the analysis of the AI / ML model transmission status and performs adjustment of the application layer information.

[0134] For example, as illustrated in FIG. 4, FIG. 4 is a signaling flowchart of a method for subscribing to model transmission state analysis in a network provided in Example 1 of the present disclosure. FIG. 4 is a diagram of the signaling interactions between AF and NWDAF, and NEF and NF in the method for subscribing to model transmission state analysis in a network. The method for subscribing to model transmission state analysis in a network provided in the present embodiment includes the following steps (i.e., the signaling interaction process corresponding to Example 1: AF requests that NWDAF provide AI / ML model transmission state analysis): (wherein steps (4011) through (4016) are when AF is in an untrusted region, and steps (4021) through (4024) are when AF is in a trusted region.)

[0135] Step (4011), if AF is untrusted, AF sends a request to NEF to subscribe to the AI / ML model status Nnef_AnalyticsExposure_Subscribe (i.e., subscribe to analytics exposure) or Nnef_AnalyticsExposure_Fetch (i.e., obtain analytics exposure).

[0136] In this embodiment, the request may carry parameters in Table 1 or Table 2 and request a subscription to analysis information of the AI / ML model transmission status in the network.

[0137] Step (4012), NEF sends a request to NWDAF for an AI / ML model exposure transmission status subscription Nnwdaf_AnalyticsSubscription_Subscribe (i.e., analytics subscription) or Nnwdaf_AnalyticsInfo_Request (i.e. analytics info).

[0138] Here, the AI / ML model exposure transmission status subscription Nnwdaf_AnalyticsSubscription_Subscribe or Nnwdaf_AnalyticsInfo_Request request can be the first message.

[0139] In this embodiment, the request may carry parameters in Table 1 or Table 2 and request a subscription to analysis information of the AI / ML model transmission status in the network.

[0140] Step (4013), NWDAF calls Nnf_EventExposure_Subscribe (i.e., subscribe to event exposure) to collect data from 5GC NF(s).

[0141] Here, data collection is as shown in Table 3 or Table 4 and is intended to analyze the transmission status of AI / ML models in the network. The method by which NWDAF sends a second message to the 5GC NF(s) may be for NWDAF to call Nnf_EventExposure_Subscribe.

[0142] Step (4014), 5GC NF(s) call Nnf_EventExposure_Notify (i.e., event exposure notification) to feed back the desired data to NWDAF.

[0143] Step (4015), NWDAF calls Nnwdaf_AnalyticsSubscription_Notify (i.e., analytics subscription notification) or Nnwdaf_AnalyticsInfo_Request response (i.e., analytics information request response) to send analytics information of the AI / ML model transfer status to NEF.

[0144] Step (4016), NEF calls Nnef_AnalyticsExposure_Notify (i.e., Analytics Exposure Notification) or Nnef_AnalyticsExposure_Fetch response (i.e., Analytics Exposure Fetch Response) to send analytics information of the AI / ML model transmission status to AF.

[0145] Here, the analysis information is as shown in Table 5 or Table 6.

[0146] Step (4021), if AF is reliable, AF sends a request to subscribe to the AI / ML model transfer status directly to NWDAF, and the action performed is as described in Step (4012). Meanwhile, the consumer may be PCF or SMF.

[0147] Step (4022), the action performed is as described in Step (4013). That is, Step (4013).

[0148] Step (4023), the action performed is as described in Step (4014). That is, Step (4014).

[0149] Step (4024), NWDAF directly sends analysis information of the AI / ML model transmission status to AF, and the operation performed is as described in Step (4016).

[0150] Example 2, after receiving the analysis information, the method

[0151] Based on the analysis information above, the method further includes the step of sending a first request to the strategy control function (PCF) directly or through the NEF.

[0152] Here, the first request is to request an update to network strategy parameters for AI / ML model transmission; and the network strategy parameters are to optimize the AI / ML model transmission state.

[0153] Specifically, based on the NWDAF analysis results, the AF requests that the network strategy be adjusted to optimize the AI / ML model delivery state. Specifically, if the AF is in a trusted region, the AF sends a first request directly to the PCF, requesting that the PCF update the network strategy parameters for AI / ML model delivery. If the AF is not in a trusted region, the AF sends a first request to the PCF via the NEF, requesting that the PCF update the network strategy parameters for AI / ML model delivery.

[0154] Optionally, sending a first request to the strategy control function (PCF) directly or through the NEF based on the analysis information above can be implemented through the following steps.

[0155] Step (a1), based on at least one of the analysis information, the uplink direction bit rate and the downlink direction bit rate for transmitting the AI / ML model, the uplink direction packet delay and the downlink direction packet delay of the AI / ML model, the quantity of abnormal releases of the quality of service flow during the period of transmission of the AI / ML model, the quantity of packet transmissions of the AI / ML model, the quantity of packet retransmissions of the AI / ML model, and the number of times the quality of service flow reached a reporting threshold of abnormal releases during the period of transmission of the AI / ML model, a new quality of service parameter for transmitting the AI / ML model is determined; the new quality of service parameter includes at least one of a 5G quality of service identifier, a reflective quality of service control, the uplink direction maximum bit rate for transmitting the AI / ML model, the downlink direction maximum bit rate for transmitting the AI / ML model, the uplink direction minimum bit rate for transmitting the AI / ML model, the downlink direction minimum bit rate for transmitting the AI / ML model, and the priority of the quality of service flow.

[0156] Step (a2), based on the identifier of the application using the AI / ML model, the area information using the AI / ML model, the IP address information of the application service using the AI / ML model, the network slice of the PDU session of the quality of service flow for transmitting the AI / ML model, and the data network name of the PDU session of the quality of service flow for transmitting the AI / ML model among the analysis information, the area information and address information of the UE(s) transmitting the AI / ML model and each AF are determined; or, if the AI / ML model performs federated learning, based on the identifier indicating the federated learning group being analyzed among the analysis information, the identifier of the UE(s) or UE(s) participating in federated learning, and the identifier of each application providing the AI / ML model or participating in federated learning, the area information and address information of the UE(s) transmitting the AI / ML model and each AF are determined.

[0157] Step (a3), based on the area information and address information of the UE(s) transmitting the AI / ML model and each AF, a data network access identifier (DNAI) and the area information and address information of the UE(s) and each AF corresponding to the DNAI are determined, and the area information and address information of the UE(s) and each AF corresponding to the DNAI are to provide a path that optimizes the AI / ML model transmission state.

[0158] Step (a4), the new service quality parameters, the DNAI and the UE(s) corresponding to the DNAI, and the area information and address information of each AF are sent to the PCF directly or through the NEF as parameters requested in the first request.

[0159] Specifically, AF determines new QoS parameters for transmitting AI / ML models based on analysis information regarding AI / ML model transmission obtained from NWDAF, such as: the uplink bit rate and downlink bit rate for transmitting AI / ML models, the uplink packet delay and downlink packet delay of AI / ML models, the number of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, the number of packet transmissions of the AI / ML model, the number of packet retransmissions of the AI / ML model, and the number of times the Quality of Service flow reached the reporting threshold for abnormal releases during the period of AI / ML model transmission, and provides the new QoS parameters to PCF.

[0160] Based on analysis information regarding AI / ML model transmission obtained from NWDAF, e.g., identifier of the application using the AI / ML model, area information using the AI / ML model, IP address information of the application service using the AI / ML model, network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and data network name of the PDU session of the Quality of Service flow for transmitting the AI / ML model, or based on analysis information regarding AI / ML model transmission obtained from NWDAF, e.g., identifier indicating the federated learning group being analyzed, identifier of the UE or UE(s) participating in federated learning, and identifier of each application providing the AI / ML model or participating in federated learning, the AF determines the area information and address information of the UE(s) transmitting the AI / ML model and each AF, and further determines the Data Network Access Identifier (DNAI) and the UE(s) corresponding to the DNAI and each AF's area information and address information. AF sends a first request to PCF, either directly or through the NEF, carrying new QoS parameters, DNAI, UE(s) corresponding to the DNAI, and region information and address information of each AF, and requests PCF to optimize the AI / ML model transmission state by updating relevant strategy parameters for AI / ML model transmission based on the parameters of the request carried in the first request.

[0161] Optionally, step (a3) ​​can be implemented through the following steps.

[0162] Step (a31), based on the area information and address information of the UE(s) transmitting the AI / ML model and each AF, determine whether the current routing path is poor;

[0163] Step (a32), if the current routing path is not good, determine the destination addresses of both parties during transmission of the AI / ML model based on the address information of the UE(s) transmitting the AI / ML model and each AF and the area information of the UE(s);

[0164] Step (a33), based on the destination address, determine the nearest path;

[0165] Step (a34), based on the nearest path, the region information and address information of the DNAI and the UE(s) corresponding to the DNAI and each AF are determined.

[0166] Specifically, if the current routing path is determined to be poor, a DNAI capable of providing a superior service experience or performance is selected, and the DNAI that transmits AI / ML models to the PCF, along with the area information and address information of the corresponding UE(s) and AF(s), is provided.

[0167] Optionally, the above first request specifically,

[0168] The PCF requests that the 5G Service Quality identifier, reflective Service Quality control, the maximum bit rate in the uplink direction for transmitting AI / ML models, the maximum bit rate in the downlink direction for transmitting AI / ML models, the minimum bit rate in the uplink direction for transmitting AI / ML models, the minimum bit rate in the downlink direction for transmitting AI / ML models, and the priority of the Service Quality flow among the PCC rules be adjusted based on the new Service Quality parameters, and instructs that the PCF provide feedback on the first update result after adjustment directly or through the NEF; wherein the first update result is determined by the result of the PCF adjusting the PCC rules based on the new Service Quality parameters;

[0169] Correspondingly, the above method is,

[0170] The method further includes the step of receiving the first update result sent from the PCF directly or via NEF, wherein the first update result includes whether the first request was accepted or rejected.

[0171] Specifically, AF requests that, based on the new QoS parameters provided by PCF, it correspondingly adjust the 5G Quality of Service identifier, reflective Quality of Service control, the maximum bit rate in the uplink direction for transmitting AI / ML models, the maximum bit rate in the downlink direction for transmitting AI / ML models, the minimum bit rate in the uplink direction for transmitting AI / ML models, the minimum bit rate in the downlink direction for transmitting AI / ML models, and the priority of the Quality of Service flow among the PCC rules, and notify the result of the first update directly or via NEF, that is, notify AF whether the corresponding request has been accepted or rejected. Optionally, the first request specifically,

[0172] PCF requests that the Session Management Function (SMF) network element determine whether the session management strategy needs to be updated, and if it is determined that the SMF needs to update the session management strategy, PCF determines that it will send a second request to the SMF, wherein the parameters requested in the second request include at least one of DNAI, a traffic steering policy identifier, and traffic routing information; the second request is intended to determine the User Plane Function (UPF) selected by the SMF based on the new session management strategy and to provide the corresponding DNAI, traffic steering policy identifier, and traffic routing information;

[0173] Correspondingly, the above method is,

[0174] Receives a second update result sent from PCF directly or via NEF, said second update result is determined by PCF based on whether the new session management strategy sent from SMF has updated the UPF path;

[0175] Here, the second update result includes whether the second request was accepted or rejected.

[0176] Specifically, the AF requests the PCF to determine whether the Session Management Function (SMF) network element needs to update the session management strategy. If the PCF determines that the SMF needs to update the strategy information, the PCF initiates a second request to the SMF containing request parameters such as DNAI, traffic steering policy identifiers, and traffic routing information. The SMF then determines the selected User Plane Function (UPF) based on the new session management strategy and updates the session management strategy—that is, updates the SM strategy—by providing the corresponding DNAI, traffic steering policy identifiers, and traffic routing information. Finally, the result of the second update is notified directly or via the NEF, that is, the AF is notified whether the request was accepted or rejected.

[0177] For example, referring to FIG. 5, FIG. 5 is a signaling flowchart of a method for subscribing to a model transmission state analysis in a network provided in Example 2 of the present disclosure. FIG. 5 is a diagram of the signaling interaction between AF, NEF, and PCF in the method for subscribing to a model transmission state analysis in a network. The method for subscribing to a model transmission state analysis in a network provided in the present embodiment includes the following steps (i.e., a signaling interaction process corresponding to Example 2: based on the NWDAF analysis results, AF requests that the network strategy be adjusted to optimize the AI / ML model transmission state): (wherein steps (5011) through (5014) are when AF is not in a trusted area, and steps (5021) through (5022) are when AF is in a trusted area.)

[0178] Step (5010), AF signs from NWDAF and retrieves the AI / ML model transmission status analysis.

[0179] In this embodiment, AF obtains an AI / ML model transmission status analysis through NEF or by directly subscribing to NWDAF, as described in steps (4011) to (4024) above.

[0180] Step (5011), AF sends a Nnef_AFsessionWithQoS_Update request (i.e., Quality of Service based AF session update) to NEF.

[0181] In this embodiment, if the AF is unreliable, for an AF session that has already been established and has QoS requirements (i.e., the AF session), the AF can optimize the AI / ML model delivery status by sending a Nnef_AFsessionWithQoS_Update (i.e., Quality of Service-based AF session update) request to the NEF to update the relevant strategy parameters for AI / ML model delivery.

[0182] Specifically, the AF determines new QoS parameters for transmitting the AI / ML model based on analysis information regarding AI / ML model transmission obtained from the NWDAF, namely QoS flow Bit Rate (i.e., Quality of Service flow bit rate, e.g., uplink direction bit rate and downlink direction bit rate for transmitting the AI / ML model), QoS flow Packet Delay (i.e., Quality of Service flow packet delay, e.g., uplink direction packet delay and downlink direction packet delay of the AI / ML model), QoS Sustainability, Packet transmission, Packet retransmission, and QoS Sustainability Reporting Threshold(s); and determines the new QoS parameters using the PCF, namely the 5G QoS Identifier (5QI, i.e., 5G Quality of Service Identifier), Reflective QoS Control (i.e., Reflective Quality of Service Control), UL-maximum bitrate (i.e., uplink direction maximum bit rate for transmitting the AI / ML model), and DL-maximum bitrate (i.e., downlink direction It provides a maximum bit rate), UL-guaranteed bit rate (i.e., the lowest bit rate in the uplink direction for transmitting AI / ML models), DL-guaranteed bit rate (i.e., the lowest bit rate in the downlink direction for transmitting AI / ML models), and Priority Level (i.e., priority level), which increases or decreases according to the increase or decrease in prediction results so that 5GS satisfies the QoS requirements for model transmission.

[0183] Meanwhile, AF determines the area and address information of the UE(s) transmitting the AI / ML model and AF(s) based on analysis information regarding AI / ML model transmission obtained from NWDAF, such as the AF ID transmitting the AI / ML model, UE area information using the AI / ML model, IP address information of the application service using the AI / ML model, network slice used for AI / ML model transmission, and DNN information. Alternatively, the analysis information may include Federated Learning (FL) group ID, Federated Learning (FL) UE ID or UE group ID, and Federated Learning (FL) Application ID regarding federated learning group information. AF determines the area and address information of the UE(s) transmitting the AI / ML model and AF(s); if the current routing path is determined to be poor, it selects a DNAI capable of providing a superior service experience or performance, and provides the DNAI transmitting the AI / ML model and the corresponding area and address information of the UE(s) and AF(s) to the PCF.

[0184] Here, the process of determining whether the current routing path is poor can be as follows: The AF determines the current routing path based on the AF's IP address information and the UE(s)'s zone information (which can correspond to IDs such as AMF, SMF, UPF, etc., and the UPF's N6 interface is connected to the DN). For example, if some UEs / servers join / leave the (Federated Cluster Group) before the next transmission, some paths are poor, or the number of hops is too high, the path is poor.

[0185] The process of selecting a DNAI capable of providing a superior service experience or performance can be as follows: Based on received IP address information and UE(s) area information, the destination addresses of both parties to transmission can be determined, and accordingly, a superior (closest) routing path can be determined, which corresponds to DNAI.

[0186] Step (5012), NEF sends an Npcf_PolicyAuthorization_Update request (i.e., a strategy approval update request) to PCF.

[0187] In this embodiment, the NEF sends the above-described information to the PCF via an Npcf_PolicyAuthorization_Update request (i.e., a strategy authorization update request) to update the relevant strategy parameters used for the transmission of the AI / ML model, thereby optimizing the AI / ML model transmission status.

[0188] Here, the Npcf_PolicyAuthorization_Update request can be the second request.

[0189] Step (5013), PCF notifies NEF of the result (i.e., PCF sends an Npcf_PolicyAuthorization_Update response (strategy approval update response) to NEF).

[0190] In this embodiment, based on the information provided by the NEF, specifically, based on the new QoS parameters provided by the PCF, the PCF adjusts the 5G QoS Identifier (5QI), Reflective QoS Control, UL-maximum bitrate, DL-maximum bitrate, UL-guaranteed bitrate, DL-guaranteed bitrate, and Priority Level among the PCC rules accordingly, and notifies the NEF of the result by sending an Npcf_PolicyAuthorization_Update response (strategy authorization update response).

[0191] If it is determined by the PCF that the SMF needs to update strategy information, the PCF initiates an Npcf_SMPolicyControl_UpdateNotify request (i.e., a session strategy control update notification request) (DNAI, Per DNAI: Traffic steering policy identifier, Per DNAI: N6 traffic routing information) to the SMF to update the SM strategy, and the SMF selects a UPF based on the strategy and provides DNAI (Per DNAI: Traffic steering policy identifier, Per DNAI: N6 traffic routing information). Here, the Npcf_SMPolicyControl_UpdateNotify request may be a second request.

[0192] Step (5014), NEF sends the Nnef_AFsessionWithQoS_Update response to AF.

[0193] In this embodiment, the NEF sends an Nnef_AFsessionWithQoS_Update response (i.e., an AF session update response based on quality of service) to notify the AF whether the request has been accepted or rejected.

[0194] Step (5021), AF sends an Npcf_PolicyAuthorization_Update request directly to PCF.

[0195] In this embodiment, if the AF is reliable, the AF directly sends an Npcf_PolicyAuthorization_Update request to the PCF to update the relevant strategy parameters for AI / ML model transmission and optimizes the AI / ML model transmission status, and the operation performed is as described in step (5012).

[0196] Step (5022), PCF notifies AF of the result directly (i.e., PCF sends the Npcf_PolicyAuthorization_Update response (i.e., strategy authorization update response) directly to AF).

[0197] In this embodiment, first, the operation performed by the PCF based on the information provided by the AF is as described in step (5012). The PCF directly notifies the AF whether the request has been accepted or rejected.

[0198] Example 3, after receiving the analysis information, the method can also be implemented through the following steps.

[0199] Step (b1), based on the analysis information, adjusts the information of the application layer model, wherein the information of the application layer model includes at least one of model compression, model size, model transmission period, model encoding, and decoding; and the information of the application layer model is for updating service quality parameters;

[0200] Step (b2), based on the information of the application layer model after adjustment, the new service quality parameters are determined, and the new service quality parameters include a 5G service quality identifier, reflective service quality control, a maximum bit rate in the uplink direction for transmitting an AI / ML model, a maximum bit rate in the downlink direction for transmitting an AI / ML model, a minimum bit rate in the uplink direction for transmitting an AI / ML model, a minimum bit rate in the downlink direction for transmitting an AI / ML model, and a priority of the service quality flow;

[0201] Step (b3), send a third request to the strategy control function (PCF) directly or through the NEF;

[0202] Here, the parameter requested in the third request includes the new service quality parameter, and the third request is intended to request an update to the service quality parameter.

[0203] Specifically, AF updates the quality of service parameters by adjusting the information of the application layer model based on the QoS request information provided by NWDAF, such as adjusting model compression, model size, model transmission duration, model encoding, and decoding. Based on the information of the application layer model after adjustment, AF finalizes the new quality of service parameters and sends a third request carrying the new quality of service parameters to PCF, either directly or via NEF.

[0204] Optionally, the above third request specifically,

[0205] It is intended to request that the PCF adjust the 5G Quality of Service identifier, reflective Quality of Service control, the maximum bit rate in the uplink direction for transmitting AI / ML models, the maximum bit rate in the downlink direction for transmitting AI / ML models, the minimum bit rate in the uplink direction for transmitting AI / ML models, the minimum bit rate in the downlink direction for transmitting AI / ML models, and the priority of Quality of Service flows among the PCC rules based on the above-mentioned new Quality of Service parameters;

[0206] Correspondingly, the above method is,

[0207] The method further includes the step of receiving a third update result sent from the PCF directly or through the NEF, wherein the third update result is determined based on the result of the PCF adjusting the PCC rule, and wherein the third update result includes whether the third request was accepted or rejected.

[0208] Specifically, AF requests that, based on the new QoS parameters provided by PCF, PCC (i.e., strategy and charging control) rules be adjusted accordingly, including the 5G Quality of Service identifier, reflective Quality of Service control, maximum bit rate in the uplink direction for transmitting AI / ML models, maximum bit rate in the downlink direction for transmitting AI / ML models, minimum bit rate in the uplink direction for transmitting AI / ML models, minimum bit rate in the downlink direction for transmitting AI / ML models, and priority of Quality of Service flows, and PCF adjusts the PCC rules and notifies AF, either directly or through NEF, whether the request has been accepted or rejected.

[0209] Optionally, after adjusting the information of the application layer model, the method can also be implemented through the following steps.

[0210] Among the information of the application layer model after adjustment, model compression, model size, and model encoding and decoding are directly sent to the PCF, and said information of the application layer model after adjustment is intended to support the PCF in adjusting the 5G Quality of Service identifier, reflective Quality of Service control, the maximum bit rate in the uplink direction for transmitting AI / ML models, the maximum bit rate in the downlink direction for transmitting AI / ML models, the minimum bit rate in the uplink direction for transmitting AI / ML models, the minimum bit rate in the downlink direction for transmitting AI / ML models, and the priority of Quality of Service flows among PCC rules;

[0211] Correspondingly, the above method is,

[0212] The method further includes the step of receiving a fourth update result sent by the PCF, wherein the fourth update result is determined based on the result of adjusting the PCC rule based on the information of the application layer model after adjustment by the PCF, and wherein the fourth update result includes the acceptance of the third request or the rejection of the third request.

[0213] Specifically, the PCF adjusts the aforementioned QoS parameters among the PCC rules based on model compression, model size, model encoding, and decoding. Examples include: 5G Quality of Service identifier, reflective Quality of Service control, maximum bit rate in the uplink direction for transmitting AI / ML models, maximum bit rate in the downlink direction for transmitting AI / ML models, minimum bit rate in the uplink direction for transmitting AI / ML models, minimum bit rate in the downlink direction for transmitting AI / ML models, and priority of Quality of Service flows. The PCF adjusts the PCC rules and notifies the AF, either directly or via the NEF, whether the corresponding request has been accepted or rejected.

[0214] Optionally, after adjusting the information of the application layer model, the method,

[0215] The model transmission period among the information of the application layer model after adjustment is directly sent to the PCF, and the model transmission period among the information of the application layer model after adjustment is intended to support the PCF in adjusting the gate state parameters among the PCC rules. The gate state parameters are intended to support the SMF in updating the session management strategy based on the transmission start time and transmission end time during the gate state;

[0216] Correspondingly, the above method is,

[0217] Receiving a fifth update result sent by PCF, wherein the fifth update result is confirmed by PCF receiving the result of a new session management strategy sent by SMF, and further including a step in which the fifth update result includes that the third request was accepted or rejected.

[0218] Specifically, PCF adjusts the Gate status (i.e., Gate status parameter) among PCC rules based on the model transmission period and updates the SM strategy, SMF feeds back to PCF based on the transmission start and end times affecting the flow, and PCF notifies AF whether the request has been accepted or rejected, either directly or through NEF.

[0219] For example, as illustrated in FIG. 6, FIG. 6 is a signaling flowchart of a method for subscribing to a model transmission state analysis in a network provided in Example 3 of the present disclosure, FIG. 6 is a diagram of the signaling interaction between AF, NEF, and PCF in the method for subscribing to a model transmission state analysis in a network. The method for subscribing to a model transmission state analysis in a network provided in the present embodiment includes the following steps (i.e., a signaling interaction process corresponding to Example 3: AF adjusts information of the application layer model based on analysis information provided by NWDAF, for example, adjusting model compression, model size, model transmission duration, model encoding and decoding, etc., and further updates QoS requirements, similar to the step process of Example 2): (wherein steps (6012) to (6014) are when AF is in an untrusted region, and steps (6021) to (6022) are when AF is in a trusted region.)

[0220] Step (6010), AF retrieves the AI / ML model transfer status analysis signed from NWDAF.

[0221] Specifically, AF obtains an AI / ML model transmission status analysis through NEF or by directly subscribing to NWDAF, as described in steps (4011) to (4024) above.

[0222] Step (6011), AF coordinates application layer behavior based on analysis information.

[0223] In this embodiment, AF adjusts information of the application layer model based on QoS request information provided by NWDAF, such as model compression, model size, model transmission duration, model encoding, and decoding.

[0224] Step (6012), AF sends a Nnef_AFsessionWithQoS_Update request to NEF.

[0225] Step (6013), NEF sends an Npcf_PolicyAuthorization_Update request to PCF.

[0226] Here, the Npcf_PolicyAuthorization_Update request can be a third-party request.

[0227] Step (6014), PCF notifies NEF of the result (i.e., PCF sends an Npcf_PolicyAuthorization_Update response to NEF).

[0228] Step (6015), NEF sends the Nnef_AFsessionWithQoS_Update response to AF.

[0229] Step (6021), AF sends an Npcf_PolicyAuthorization_Update request directly to PCF.

[0230] Step (6022), PCF notifies AF of the result directly (i.e., PCF sends the Npcf_PolicyAuthorization_Update response (i.e., strategy authorization update response) directly to AF).

[0231] Specifically, AF requests that the session perform a QoS update. Specifically, AF determines new QoS parameters for transmitting the AI / ML model based on adjusted model compression, model size, model transmission duration, model encoding and decoding, etc., and provides the new QoS parameters, 5G QoS Identifier (5QI), Reflective QoS Control, UL-maximum bitrate, DL-maximum bitrate, UL-guaranteed bitrate, DL-guaranteed bitrate, and Priority Level, to PCF. Alternatively, AF directly sends information regarding model compression, model size, model transmission duration, model encoding and decoding adjustments, etc., to PCF.

[0232] Step (6013), PCF adjusts the 5G QoS Identifier (5QI), Reflective QoS Control, UL-maximum bitrate, DL-maximum bitrate, UL-guaranteed bitrate, DL-guaranteed bitrate, and Priority Level among PCC rules based on the new QoS parameters provided; or PCF adjusts the above-mentioned QoS parameters among PCC rules based on model compression, model size, model encoding and decoding; or PCF adjusts the Gate status among PCC rules based on the model transmission period, updates the SM strategy, and SMF feeds back to PCF based on the transmission start and end times affecting the flow; finally, notifies AF whether the request has been accepted or rejected through the method of Step (5014) or Step (5022) of Example 2.

[0233] In the present disclosure, NWDAF receives an AI / ML model transmission status analysis request sent by AF, and parameters included in the request. Input data collected by NWDAF from 5GC NF(s) is for analyzing the AI / ML model transmission status in the network; NWDAF performs the analysis and sends AI / ML model transmission status analysis information to AF; AF requests a QoS update of the AF session based on the AI / ML model transmission status analysis information and adjusts the strategies of network elements, such as the corresponding Strategy Control Function (PCF) and Session Management Function (SMF) network elements; AF adjusts relevant parameters of application layer model information based on the AI / ML model transmission status analysis information and further adjusts the QoS to optimize the AI / ML model transmission status. A third party can obtain the status of AI / ML model transmission, and based on the analysis results of the model transmission, the network can adjust its own behavior to satisfy the demand for AI / ML model transmission. Additionally, based on the analysis results of the model transmission, the third party can adjust the behavior of the model application layer to implement highly efficient transmission of the AI / ML model and secure the service experience and service performance of the AI / ML model transmission.

[0234] FIG. 7 is a flowchart of a method for subscribing to a model transmission state analysis in a network provided in Embodiment 4 of the present disclosure. As illustrated in FIG. 7, if the entity performing the method for subscribing to a model transmission state analysis in a network provided in the present embodiment is NWDAF, the method for subscribing to a model transmission state analysis in a network provided in the embodiment of the present disclosure includes the following steps.

[0235] Step (701), NWDAF receives a first message sent from an application function (AF) directly or through a network capability exposure function (NEF); wherein the first message is to request a subscription to analysis information of the transmission status of an artificial intelligence / machine learning (AI / ML) model in the network;

[0236] Step (702), NWDAF sends a second message to other network functions 5GC NF(s) of the 5G core network based on the parameters requested in the first message, and the second message is for collecting data to analyze the transmission status of AI / ML models in the network.

[0237] Step (703), NWDAF receives data of the AI / ML model transmission status sent from other network functions 5GC NF(s) of the 5G core network, analyzes the data of the AI / ML model transmission status, and obtains analysis information of the AI / ML model transmission status.

[0238] Here, the above analysis information is intended to adjust network strategy parameters and / or application layer model information through AF.

[0239] Optionally, the parameters requested in the first message include at least one of a network data analysis identifier, an identifier of a user device (UE) or a group of UEs receiving the AI / ML model or an identifier of any UE satisfying the analysis condition, an identifier of an application using the AI / ML model, a region where the AI / ML model is transmitted, a network slice instruction of a protocol data unit (PDU) session of a quality of service flow transmitting the AI / ML model, a data network instruction of a PDU session of a quality of service flow transmitting the AI / ML model, a duration of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, a size of the AI / ML model transmission, a quality of service requirement instruction of a quality of service flow transmitting the AI / ML model, and / or a specific quality of service requirement instruction for transmitting the AI / ML model;

[0240] The second message comprises at least one of the following: the current location of the UE using the AI / ML model, an identifier of the application using the AI / ML model, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate and a downlink bit rate transmitting the AI / ML model, an uplink packet delay and a downlink packet delay of the AI / ML model, a quantity of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, a quantity of packets transmitted by the AI / ML model, a quantity of packet retransmissions of the AI / ML model, a data collection time, a time length of AI / ML model transmission, a start timestamp of AI / ML model transmission, an end timestamp of AI / ML model transmission, a size of the AI / ML transmission model, a network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, a data network of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and a service process for the AF;

[0241] The above analysis information includes at least one of: a network slice of a PDU session of a Quality of Service flow for transmitting an AI / ML model, an identifier of an application using the AI / ML model, area information using the AI / ML model, the validity period of the analysis result, a User Plane Function (UPF) providing AI / ML model transmission, a data network name of a PDU session of a Quality of Service flow for transmitting an AI / ML model, the size of the AI / ML transmission model, the time length of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate for transmitting the AI / ML model and a downlink bit rate for transmitting the AI / ML model, an uplink packet delay of the AI / ML model and a downlink packet delay of the AI / ML model, the number of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, the number of times the Quality of Service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission, the number of packet transmissions of the AI / ML model, and the number of packet retransmissions of the AI / ML model;

[0242] Herein, if the AI / ML model performs federated learning, the parameter requested in the first message further includes federated learning group information, and the federated learning group information includes at least one of an identifier for indicating the federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0243] Correspondingly, the second message further includes at least one of an identifier for specifying a federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0244] Correspondingly, the analysis information further includes at least one of an identifier for specifying the federated learning group to be analyzed, an identifier of a UE or UE(s) participating in federated learning, and an indication of each application identifier that provides an AI / ML model or participates in federated learning.

[0245] In this embodiment, a request to subscribe to analysis information of the AI / ML model transmission status in a network sent by AF is received, and based on the parameters requested in the first message, data is collected from other network functions 5GC NF(s) of the 5G core network, data of the AI / ML model transmission status sent by other network functions 5GC NF(s) of the 5G core network is received, and the data of the AI / ML model transmission status is analyzed to obtain analysis information of the AI / ML model transmission status, thereby effectively analyzing the AI / ML model transmission status and allowing AF to adjust network strategy parameters and / or information of application layer models based on the analysis information, so that the network can effectively adjust the network transmission strategy based on the AI / ML model transmission status, and a third party obtains the analysis of the AI / ML model transmission status to perform adjustment of application information.

[0246] It should be explained here that the method for subscribing to model transmission state analysis in a network provided in the embodiment of the present disclosure can implement all method steps implemented in the method embodiment illustrated in FIG. 4 and can achieve the same technical effects, and here, specific descriptions of parts identical to the method embodiment and advantageous effects in this embodiment are not repeated.

[0247] FIG. 8 is a structural diagram of a device for subscribing to a model transmission state analysis in a network provided in an embodiment of the present disclosure. As shown in FIG. 8, the device for subscribing to a model transmission state analysis in a network provided in this embodiment is applied to an AF. At this time, the device for subscribing to a model transmission state analysis in a network provided in this embodiment includes a transceiver (800) that receives and sends data under the control of a processor (810).

[0248] Here, in FIG. 8, the bus architecture may specifically include any number of interconnected buses and bridges that connect together various circuits of one or more processors, represented by the processor (810), and memory, represented by the memory (820). The bus architecture may also connect together various other circuits, such as peripheral devices, voltage stabilizers, and power management circuits, all of which are known in the art and are therefore not described further in the text. The bus interface provides an interface. The transceiver (800) may be a plurality of elements, namely a transmitter and a receiver, and provides a unit for communicating with various other devices over a transmission medium, such as a wireless channel, a wired channel, an optical cable, etc. The processor (810) is responsible for managing the bus architecture and general processing, and the memory (820) may store data used when the processor (810) performs operations.

[0249] The processor (810) may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a complex programmable logic device (CPLD), and the processor (810) may use a multi-core architecture.

[0250] In this embodiment, the memory (820) is for storing a computer program; the transceiver (800) is for transmitting and receiving data under the control of the processor (810); and the processor (810) reads the computer program in the memory (820),

[0251] A first message is sent to the Network Data Analysis Function (NWDAF) either directly or through the Network Capability Exposure Function (NEF); wherein the first message is intended to request a subscription to analysis information regarding the transmission status of an Artificial Intelligence / Machine Learning (AI / ML) model in the network;

[0252] It is to perform an operation in which analysis information of the AI / ML model transmission status sent by the NWDAF is received directly or through the NEF, said analysis information is determined based on data of the AI / ML model transmission status sent by other network functions (5GC NF(s)) of the 5G core network received by the NWDAF, said AI / ML model transmission status data is obtained by sending a second message to the 5GC NF(s) based on parameters requested in the first message received by the NWDAF, and said second message is for collecting data to analyze the AI / ML model transmission status in the network;

[0253] Here, the above analysis information is intended to adjust network strategy parameters and / or application layer model information.

[0254] Optionally, the parameters requested in the first message include at least one of a network data analysis identifier, an identifier of a user device (UE) or a group of UEs receiving the AI / ML model or an identifier of any UE satisfying the analysis condition, an identifier of an application using the AI / ML model, a region where the AI / ML model is transmitted, a network slice instruction of a protocol data unit (PDU) session of a quality of service flow transmitting the AI / ML model, a data network instruction of a PDU session of a quality of service flow transmitting the AI / ML model, a duration of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, a size of the AI / ML model transmission, a quality of service requirement instruction of a quality of service flow transmitting the AI / ML model, and / or a specific quality of service requirement instruction for transmitting the AI / ML model;

[0255] The second message comprises at least one of the following: the current location of the UE using the AI / ML model, an identifier of the application using the AI / ML model, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate and a downlink bit rate transmitting the AI / ML model, an uplink packet delay and a downlink packet delay of the AI / ML model, a quantity of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, a quantity of packets transmitted by the AI / ML model, a quantity of packet retransmissions of the AI / ML model, a data collection time, a time length of AI / ML model transmission, a start timestamp of AI / ML model transmission, an end timestamp of AI / ML model transmission, a size of the AI / ML transmission model, a network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, a data network of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and a service process for the AF;

[0256] The above analysis information includes at least one of: a network slice of a PDU session of a Quality of Service flow for transmitting an AI / ML model, an identifier of an application using the AI / ML model, area information using the AI / ML model, the validity period of the analysis result, a User Plane Function (UPF) providing AI / ML model transmission, a data network name of a PDU session of a Quality of Service flow for transmitting an AI / ML model, the size of the AI / ML transmission model, the time length of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate for transmitting the AI / ML model and a downlink bit rate for transmitting the AI / ML model, an uplink packet delay of the AI / ML model and a downlink packet delay of the AI / ML model, the number of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, the number of times the Quality of Service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission, the number of packet transmissions of the AI / ML model, and the number of packet retransmissions of the AI / ML model;

[0257] Herein, if the AI / ML model performs federated learning, the parameter requested in the first message further includes federated learning group information, and the federated learning group information includes at least one of an identifier for indicating the federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0258] Correspondingly, the second message further includes at least one of an identifier for indicating a federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0259] Correspondingly, the analysis information further includes at least one of an identifier for specifying the federated learning group to be analyzed, an identifier of a UE or UE(s) participating in federated learning, and an indication of each application identifier that provides an AI / ML model or participates in federated learning.

[0260] Optionally, the processor (810) also,

[0261] After receiving the above analysis information, it is intended to send a first request to the strategy control function (PCF) directly or through the above NEF based on the above analysis information;

[0262] Here, the first request is to request an update to network strategy parameters for AI / ML model transmission; and the network strategy parameters are to optimize the AI / ML model transmission state.

[0263] Optionally, when the processor (810) sends a first request to the strategy control function (PCF) directly or through the NEF based on the analysis information, specifically,

[0264] Based on at least one of the analysis information above, the uplink direction bit rate and downlink direction bit rate for transmitting the AI / ML model, the uplink direction packet delay and downlink direction packet delay of the AI / ML model, the quantity of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, the quantity of packet transmissions of the AI / ML model, the quantity of packet retransmissions of the AI / ML model, and the number of times the Quality of Service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission, a new Quality of Service parameter for transmitting the AI / ML model is determined; the new Quality of Service parameter includes at least one of a 5G Quality of Service identifier, reflective Quality of Service control, the uplink direction maximum bit rate for transmitting the AI / ML model, the downlink direction maximum bit rate for transmitting the AI / ML model, the uplink direction minimum bit rate for transmitting the AI / ML model, the downlink direction minimum bit rate for transmitting the AI / ML model, and the priority of the Quality of Service flow;

[0265] Based on the identifier of the application using the AI / ML model, the region information using the AI / ML model, the IP address information of the application service using the AI / ML model, the network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and the data network name of the PDU session of the Quality of Service flow for transmitting the AI / ML model among the above analysis information, the region information and address information of the UE(s) transmitting the AI / ML model and each AF; or, if the AI / ML model performs federated learning, based on the identifier indicating the federated learning group being analyzed among the above analysis information, the identifier of the UE or UE(s) participating in federated learning, and the identifier of each application providing the AI / ML model or participating in federated learning, the region information and address information of the UE(s) transmitting the AI / ML model and each AF are determined;

[0266] Based on the domain information and address information of the UE(s) transmitting the AI / ML model and each AF, a data network access identifier (DNAI) and the domain information and address information of the UE(s) and each AF corresponding to the DNAI are determined, and the domain information and address information of the UE(s) and each AF corresponding to the DNAI are intended to provide a path that optimizes the AI / ML model transmission state;

[0267] The above new service quality parameters, the above DNAI, the UE(s) corresponding to the above DNAI, and the area information and address information of each AF are sent to the PCF directly or through the above NEF as parameters requested in the above first request.

[0268] Optionally, when the processor (810) determines the data network access identifier (DNAI) and the region information and address information of the UE(s) and the AF corresponding to the DNAI based on the region information and address information of the respective AF and the UE(s) transmitting the AI / ML model, specifically,

[0269] Determining whether the current routing path is poor based on the area and address information of the UE(s) transmitting the AI / ML model and each AF;

[0270] If the current routing path is poor, determining the destination addresses of both parties during the transmission of the AI / ML model based on the address information of the UE(s) transmitting the AI / ML model and their respective AFs, and the area information of the UE(s);

[0271] Determining the nearest path based on the above destination address;

[0272] Based on the nearest path, it includes determining the region information and address information of the DNAI, the UE(s) corresponding to the DNAI, and each AF.

[0273] Optionally, the above first request specifically,

[0274] The PCF requests that the 5G Service Quality identifier, reflective Service Quality control, the maximum bit rate in the uplink direction for transmitting AI / ML models, the maximum bit rate in the downlink direction for transmitting AI / ML models, the minimum bit rate in the uplink direction for transmitting AI / ML models, the minimum bit rate in the downlink direction for transmitting AI / ML models, and the priority of the Service Quality flow among the PCC rules be adjusted based on the new Service Quality parameters, and instructs that the PCF provide feedback on the first update result after adjustment directly or through the NEF; wherein the first update result is determined by the result of the PCF adjusting the PCC rules based on the new Service Quality parameters;

[0275] Correspondingly, the processor (810) also specifically,

[0276] Receives the first update result sent from the PCF directly or via NEF, and the first update result includes that the first request has been accepted or that the first request has been rejected.

[0277] Optionally, the above first request specifically,

[0278] PCF requests that the Session Management Function (SMF) network element determine whether the session management strategy needs to be updated, and if it is determined that the session management strategy needs to be updated, PCF determines to send a second request to the SMF, wherein the parameters requested in the second request include at least one of DNAI, a traffic steering policy identifier, and traffic routing information; the second request is intended to determine the User Plane Function (UPF) selected by the SMF based on the new session management strategy and to provide the corresponding DNAI, traffic steering policy identifier, and traffic routing information;

[0279] Correspondingly, the processor (810) also specifically,

[0280] Receives a second update result sent from PCF directly or via NEF, said second update result is determined by PCF based on whether the new session management strategy sent from SMF has updated the UPF path;

[0281] Here, the second update result includes whether the second request was accepted or rejected.

[0282] Optionally, the processor (810) also,

[0283] After receiving the above analysis information, the information of the application layer model is adjusted based on the above analysis information, and the information of the application layer model includes at least one of model compression, model size, model transmission period, model encoding, and decoding; and the information of the application layer model is for updating service quality parameters;

[0284] Based on the information of the application layer model after adjustment, the new service quality parameters are determined, and the new service quality parameters include a 5G service quality identifier, reflective service quality control, a maximum bit rate in the uplink direction for transmitting an AI / ML model, a maximum bit rate in the downlink direction for transmitting an AI / ML model, a minimum bit rate in the uplink direction for transmitting an AI / ML model, a minimum bit rate in the downlink direction for transmitting an AI / ML model, and a priority of the service quality flow;

[0285] Sending a third request to the Strategy Control Function (PCF) directly or through the above NEF;

[0286] Here, the parameter requested in the third request includes the new service quality parameter, and the third request is intended to request an update to the service quality parameter.

[0287] Optionally, the above third request specifically,

[0288] It is intended to request that the PCF adjust the 5G Quality of Service identifier, reflective Quality of Service control, the maximum bit rate in the uplink direction for transmitting AI / ML models, the maximum bit rate in the downlink direction for transmitting AI / ML models, the minimum bit rate in the uplink direction for transmitting AI / ML models, the minimum bit rate in the downlink direction for transmitting AI / ML models, and the priority of Quality of Service flows among the PCC rules based on the above-mentioned new Quality of Service parameters;

[0289] Correspondingly, the processor (810) also,

[0290] Receive a third update result sent from the PCF directly or through the NEF, the third update result is confirmed by the PCF based on the result of adjusting the PCC rule, and the third update result includes whether the third request was accepted or rejected.

[0291] Optionally, the processor (810) also,

[0292] After adjusting the information of the above application layer model, the model compression, model size, and model encoding and decoding among the information of the adjusted application layer model are directly sent to the PCF, and the information of the adjusted application layer model is intended to support the PCF in adjusting the 5G Quality of Service identifier, reflective Quality of Service control, the maximum bit rate in the uplink direction for transmitting AI / ML models, the maximum bit rate in the downlink direction for transmitting AI / ML models, the minimum bit rate in the uplink direction for transmitting AI / ML models, the minimum bit rate in the downlink direction for transmitting AI / ML models, and the priority of the Quality of Service flow among the PCC rules;

[0293] Correspondingly, the processor (810) also,

[0294] Receives a fourth update result sent by PCF, and the fourth update result is determined based on the result of PCF adjusting PCC rules based on information of the application layer model after adjustment, and the fourth update result includes the acceptance of the third request or the rejection of the third request.

[0295] Optionally, the processor (810) also,

[0296] After adjusting the information of the above application layer model, the model transmission period among the information of the adjusted application layer model is directly sent to the PCF, and the model transmission period among the information of the adjusted application layer model is intended to support the PCF in adjusting the gate state parameters among the PCC rules; and the gate state parameters are intended to support the SMF in updating the session management strategy based on the transmission start time and transmission end time during the gate state;

[0297] Correspondingly, the processor (810) also,

[0298] Receives the 5th update result sent by the PCF, the 5th update result is confirmed by the PCF receiving the result of the new session management strategy sent by the SMF, and the 5th update result includes that the 3rd request was accepted or that the 3rd request was rejected.

[0299] It should be explained here that a device subscribing to the model transmission state analysis in a network provided in this disclosure can implement all method steps implemented in the method embodiments illustrated in FIGS. 3 to 6 and can achieve the same technical effects, and here, specific descriptions of parts identical to the method embodiments and advantageous effects in this embodiment are not repeated.

[0300] FIG. 9 is a structural diagram of a device for subscribing to a model transmission state analysis in a network provided in another embodiment of the present disclosure. As shown in FIG. 9, the device for subscribing to a model transmission state analysis in a network provided in this embodiment is applied to AF, and the device (900) for subscribing to a model transmission state analysis in a network provided in this embodiment is,

[0301] A first message is sent to a network data analysis function (NWDAF) either directly or through a network capability exposure function (NEF); wherein the first message is a sending unit (901) intended to request a subscription to analysis information of the transmission status of an artificial intelligence / machine learning (AI / ML) model in the network;

[0302] It includes an analysis unit (902) that receives analysis information of the AI / ML model transmission status sent from the NWDAF directly or through the NEF, and the second message is for collecting data to analyze the AI / ML model transmission status in the network;

[0303] Here, the above analysis information is intended to adjust network strategy parameters and / or application layer model information.

[0304] Optionally, the parameters requested in the first message include at least one of a network data analysis identifier, an identifier of a user device (UE) or a group of UEs receiving the AI / ML model or an identifier of any UE satisfying the analysis condition, an identifier of an application using the AI / ML model, a region where the AI / ML model is transmitted, a network slice instruction of a protocol data unit (PDU) session of a quality of service flow transmitting the AI / ML model, a data network instruction of a PDU session of a quality of service flow transmitting the AI / ML model, a duration of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, a size of the AI / ML model transmission, a quality of service requirement instruction of a quality of service flow transmitting the AI / ML model, and / or a specific quality of service requirement instruction for transmitting the AI / ML model;

[0305] The second message comprises at least one of the following: the current location of the UE using the AI / ML model, an identifier of the application using the AI / ML model, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate and a downlink bit rate transmitting the AI / ML model, an uplink packet delay and a downlink packet delay of the AI / ML model, a quantity of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, a quantity of packets transmitted by the AI / ML model, a quantity of packet retransmissions of the AI / ML model, a data collection time, a time length of AI / ML model transmission, a start timestamp of AI / ML model transmission, an end timestamp of AI / ML model transmission, a size of the AI / ML transmission model, a network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, a data network of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and a service process for the AF;

[0306] The above analysis information includes at least one of: a network slice of a PDU session of a Quality of Service flow for transmitting an AI / ML model, an identifier of an application using the AI / ML model, area information using the AI / ML model, the validity period of the analysis result, a User Plane Function (UPF) providing AI / ML model transmission, a data network name of a PDU session of a Quality of Service flow for transmitting an AI / ML model, the size of the AI / ML transmission model, the time length of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, a Quality of Service request, a Quality of Service flow identifier transmitting the AI / ML model, an uplink bit rate transmitting the AI / ML model and a downlink bit rate transmitting the AI / ML model, an uplink packet delay of the AI / ML model and a downlink packet delay of the AI / ML model, the number of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, the number of times the Quality of Service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission, the number of packet transmissions of the AI / ML model, and the number of packet retransmissions of the AI / ML model;

[0307] Herein, if the AI / ML model performs federated learning, the parameter requested in the first message further includes federated learning group information, and the federated learning group information includes at least one of an identifier for indicating the federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0308] Correspondingly, the second message further includes at least one of an identifier for specifying a federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0309] Correspondingly, the analysis information further includes at least one of an identifier for specifying the federated learning group to be analyzed, an identifier of a UE or UE(s) participating in federated learning, and an indication of each application identifier that provides an AI / ML model or participates in federated learning.

[0310] Optionally, the dispatch unit also,

[0311] After receiving the above analysis information, it is intended to send a first request to the strategy control function (PCF) directly or through the above NEF based on the above analysis information;

[0312] Here, the first request is to request an update to network strategy parameters for AI / ML model transmission; and the network strategy parameters are to optimize the AI / ML model transmission state.

[0313] Optionally, the transmitting unit specifically determines a new quality of service parameter for transmitting an AI / ML model based on at least one of the analysis information, specifically the uplink direction bit rate and downlink direction bit rate for transmitting the AI / ML model, the uplink direction packet delay and the downlink direction packet delay of the AI / ML model, the number of abnormal releases of the quality of service flow during the period of transmission of the AI / ML model, the number of packet transmissions of the AI / ML model, the number of packet retransmissions of the AI / ML model, and the number of times the quality of service flow reached a reporting threshold for abnormal releases during the period of transmission of the AI / ML model; said new quality of service parameter includes at least one of a 5G quality of service identifier, reflective quality of service control, the uplink direction maximum bit rate for transmitting the AI / ML model, the downlink direction maximum bit rate for transmitting the AI / ML model, the uplink direction minimum bit rate for transmitting the AI / ML model, the downlink direction minimum bit rate for transmitting the AI / ML model, and the priority of the quality of service flow;

[0314] Based on the identifier of the application using the AI / ML model, the region information using the AI / ML model, the IP address information of the application service using the AI / ML model, the network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and the data network name of the PDU session of the Quality of Service flow for transmitting the AI / ML model among the above analysis information, the region information and address information of the UE(s) transmitting the AI / ML model and each AF; or, if the AI / ML model performs federated learning, based on the identifier indicating the federated learning group being analyzed among the above analysis information, the identifier of the UE or UE(s) participating in federated learning, and the identifier of each application providing the AI / ML model or participating in federated learning, the region information and address information of the UE(s) transmitting the AI / ML model and each AF are determined;

[0315] Based on the domain information and address information of the UE(s) transmitting the AI / ML model and their respective AFs, a data network access identifier (DNAI) and the domain information and address information of the UE(s) and their respective AFs corresponding to the DNAI are determined, and the domain information and address information of the UE(s) and their respective AFs corresponding to the DNAI are intended to provide a path that optimizes the AI / ML model transmission state.

[0316] The above new service quality parameters, the above DNAI, the UE(s) corresponding to the above DNAI, and the area information and address information of each AF are sent to the PCF directly or through the above NEF as parameters requested in the above first request.

[0317] Optionally, the sending unit also specifically,

[0318] Based on the area information and address information of the UE(s) transmitting the AI / ML model and each AF, determine whether the current routing path is poor;

[0319] If the current routing path is poor, determine the destination addresses of both parties during the transmission of the AI / ML model based on the address information of the UE(s) transmitting the AI / ML model and their respective AFs, and the area information of the UE(s);

[0320] Based on the above destination address, determine the nearest path;

[0321] Based on the nearest path, the region information and address information of the DNAI, the UE(s) corresponding to the DNAI, and each AF are determined.

[0322] Optionally, the above first request specifically,

[0323] The PCF requests that the 5G Service Quality identifier, reflective Service Quality control, the maximum bit rate in the uplink direction for transmitting AI / ML models, the maximum bit rate in the downlink direction for transmitting AI / ML models, the minimum bit rate in the uplink direction for transmitting AI / ML models, the minimum bit rate in the downlink direction for transmitting AI / ML models, and the priority of the Service Quality flow among the PCC rules be adjusted based on the new Service Quality parameters, and instructs that the PCF provide feedback on the first update result after adjustment directly or through the NEF; wherein the first update result is determined by the result of the PCF adjusting the PCC rules based on the new Service Quality parameters;

[0324] Correspondingly, the dispatch unit also,

[0325] Receives the first update result sent from the PCF directly or via NEF, and the first update result includes that the first request has been accepted or that the first request has been rejected.

[0326] Optionally, the above first request specifically,

[0327] PCF requests that the Session Management Function (SMF) network element determine whether the session management strategy needs to be updated, and if it is determined that the session management strategy needs to be updated, PCF determines to send a second request to the SMF, wherein the parameters requested in the second request include at least one of DNAI, a traffic steering policy identifier, and traffic routing information; the second request is intended to determine the User Plane Function (UPF) selected by the SMF based on the new session management strategy and to provide the corresponding DNAI, traffic steering policy identifier, and traffic routing information;

[0328] Correspondingly, the receiving unit also,

[0329] Receives a second update result sent from PCF directly or via NEF, said second update result is determined by PCF based on whether the new session management strategy sent from SMF has updated the UPF path;

[0330] Here, the second update result includes whether the second request was accepted or rejected.

[0331] Optionally, the device further comprises a fixed unit; the fixed unit is,

[0332] After receiving the above analysis information, the information of the application layer model is adjusted based on the above analysis information, and the information of the application layer model includes at least one of model compression, model size, model transmission period, model encoding, and decoding; and the information of the application layer model is for updating service quality parameters;

[0333] Based on the information of the application layer model after adjustment, the new service quality parameters are determined, and the new service quality parameters include a 5G service quality identifier, reflective service quality control, a maximum bit rate in the uplink direction for transmitting an AI / ML model, a maximum bit rate in the downlink direction for transmitting an AI / ML model, a minimum bit rate in the uplink direction for transmitting an AI / ML model, a minimum bit rate in the downlink direction for transmitting an AI / ML model, and a priority of the service quality flow;

[0334] Sending a third request to the Strategy Control Function (PCF) directly or through the above NEF;

[0335] Here, the parameter requested in the third request includes the new service quality parameter, and the third request is intended to request an update to the service quality parameter.

[0336] Optionally, the above third request specifically,

[0337] It is intended to request that the PCF adjust the 5G Quality of Service identifier, reflective Quality of Service control, the maximum bit rate in the uplink direction for transmitting AI / ML models, the maximum bit rate in the downlink direction for transmitting AI / ML models, the minimum bit rate in the uplink direction for transmitting AI / ML models, the minimum bit rate in the downlink direction for transmitting AI / ML models, and the priority of Quality of Service flows among the PCC rules based on the above-mentioned new Quality of Service parameters;

[0338] Correspondingly, the receiving unit also,

[0339] Receive a third update result sent from the PCF directly or through the NEF, the third update result is confirmed by the PCF based on the result of adjusting the PCC rule, and the third update result includes whether the third request was accepted or rejected.

[0340] Optionally, the dispatch unit also,

[0341] After adjusting the information of the above application layer model, the model compression, model size, and model encoding and decoding among the information of the adjusted application layer model are directly sent to the PCF, and the information of the adjusted application layer model is intended to support the PCF in adjusting the 5G Quality of Service identifier, reflective Quality of Service control, the maximum bit rate in the uplink direction for transmitting AI / ML models, the maximum bit rate in the downlink direction for transmitting AI / ML models, the minimum bit rate in the uplink direction for transmitting AI / ML models, the minimum bit rate in the downlink direction for transmitting AI / ML models, and the priority of the Quality of Service flow among the PCC rules;

[0342] Correspondingly, the receiving unit also,

[0343] Receives a fourth update result sent by PCF, and the fourth update result is determined based on the result of PCF adjusting PCC rules based on information of the application layer model after adjustment, and the fourth update result includes the acceptance of the third request or the rejection of the third request.

[0344] Optionally, the dispatch unit also,

[0345] After adjusting the information of the above application layer model, the model transmission period among the information of the adjusted application layer model is directly sent to the PCF, and the model transmission period among the information of the adjusted application layer model is intended to support the PCF in adjusting the gate state parameters among the PCC rules; and the gate state parameters are intended to support the SMF in updating the session management strategy based on the transmission start time and transmission end time during the gate state;

[0346] Correspondingly, the receiving unit also,

[0347] Receives the 5th update result sent by the PCF, the 5th update result is confirmed by the PCF receiving the result of the new session management strategy sent by the SMF, and the 5th update result includes that the 3rd request was accepted or that the 3rd request was rejected.

[0348] It should be explained here that the device subscribing to the model transmission state analysis in the network provided in this disclosure can implement all method steps implemented in the method embodiments of FIGS. 3 to 6 and can achieve the same technical effects, and here, specific descriptions of parts identical to the method embodiments and advantageous effects in this embodiment are not repeated.

[0349] FIG. 10 is a structural diagram of a device for subscribing to a model transmission state analysis in a network provided in another embodiment of the present disclosure. As shown in FIG. 10, the device for subscribing to a model transmission state analysis in a network provided in this embodiment is applied to an NWDAF. At this time, the device for subscribing to a model transmission state analysis in a network provided in this embodiment includes a transceiver (1000) that receives and sends data under the control of a processor (1010).

[0350] Here, in FIG. 10, the bus architecture may specifically include any number of interconnected buses and bridges that connect together various circuits of one or more processors, represented by the processor (1010), and memory, represented by the memory (1020). The bus architecture may also connect together various other circuits, such as peripheral devices, voltage stabilizers, and power management circuits, all of which are known in the art and are therefore not described further in the text. The bus interface provides an interface. The transceiver (1000) may be a plurality of elements, namely a transmitter and a receiver, and provides a unit for communicating with various other devices over a transmission medium, such as a wireless channel, a wired channel, an optical cable, etc. The processor (1010) is responsible for managing the bus architecture and general processing, and the memory (1020) may store data used when the processor (1010) performs operations.

[0351] The processor (1010) may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a complex programmable logic device (CPLD), and the processor (1010) may use a multi-core architecture.

[0352] In this embodiment, the memory (1020) is for storing a computer program; the transceiver (1000) is for transmitting and receiving data under the control of the processor; and the processor (1010) reads the computer program in the memory (1020),

[0353] Receives a first message sent from an application function (AF) directly or through a network capability exposure function (NEF); wherein the first message is intended to request a subscription to analysis information regarding the transmission status of an artificial intelligence / machine learning (AI / ML) model in the network;

[0354] Based on the parameters requested in the first message above, a second message is sent to other network functions 5GC NF(s) of the 5G core network, and the second message is intended to collect data for analyzing the transmission status of an AI / ML model in the network;

[0355] It is intended to perform an operation of receiving data on the AI / ML model transmission status transmitted from other network functions (5GC NF(s)) of the 5G core network, analyzing the data on the AI / ML model transmission status, and obtaining analysis information on the AI / ML model transmission status;

[0356] Here, the above analysis information is intended to adjust network strategy parameters and / or application layer model information through AF.

[0357] Optionally, the parameters requested in the first message include at least one of a network data analysis identifier, an identifier of a user device (UE) or a group of UEs receiving the AI / ML model or an identifier of any UE satisfying the analysis condition, an identifier of an application using the AI / ML model, a region where the AI / ML model is transmitted, a network slice instruction of a protocol data unit (PDU) session of a quality of service flow transmitting the AI / ML model, a data network instruction of a PDU session of a quality of service flow transmitting the AI / ML model, a duration of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, a size of the AI / ML model transmission, a quality of service requirement instruction of a quality of service flow transmitting the AI / ML model, and / or a specific quality of service requirement instruction for transmitting the AI / ML model;

[0358] The second message comprises at least one of the following: the current location of the UE using the AI / ML model, an identifier of the application using the AI / ML model, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate and a downlink bit rate transmitting the AI / ML model, an uplink packet delay and a downlink packet delay of the AI / ML model, a quantity of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, a quantity of packets transmitted by the AI / ML model, a quantity of packet retransmissions of the AI / ML model, a data collection time, a time length of AI / ML model transmission, a start timestamp of AI / ML model transmission, an end timestamp of AI / ML model transmission, a size of the AI / ML transmission model, a network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, a data network of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and a service process for the AF;

[0359] The above analysis information includes at least one of: a network slice of a PDU session of a Quality of Service flow for transmitting an AI / ML model, an identifier of an application using the AI / ML model, area information using the AI / ML model, the validity period of the analysis result, a User Plane Function (UPF) providing AI / ML model transmission, a data network name of a PDU session of a Quality of Service flow for transmitting an AI / ML model, the size of the AI / ML transmission model, the time length of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, a Quality of Service request, a Quality of Service flow identifier transmitting the AI / ML model, an uplink bit rate transmitting the AI / ML model and a downlink bit rate transmitting the AI / ML model, an uplink packet delay of the AI / ML model and a downlink packet delay of the AI / ML model, the number of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, the number of times the Quality of Service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission, the number of packet transmissions of the AI / ML model, and the number of packet retransmissions of the AI / ML model;

[0360] Herein, if the AI / ML model performs federated learning, the parameter requested in the first message further includes federated learning group information, and the federated learning group information includes at least one of an identifier for indicating the federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0361] Correspondingly, the second message further includes at least one of an identifier for specifying a federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0362] Correspondingly, the analysis information further includes at least one of an identifier for specifying the federated learning group to be analyzed, an identifier of a UE or UE(s) participating in federated learning, and an indication of each application identifier that provides an AI / ML model or participates in federated learning.

[0363] It should be explained here that a device subscribing to the model transmission state analysis in a network provided in this disclosure can implement all method steps implemented in the method embodiments illustrated in FIG. 4 and FIG. 7 and can achieve the same technical effects, and here, specific descriptions of parts identical to the method embodiments and advantageous effects in this embodiment are not repeated.

[0364] FIG. 11 is a structural diagram of a device for subscribing to a model transmission state analysis in a network provided in another embodiment of the present disclosure. As shown in FIG. 11, the device for subscribing to a model transmission state analysis in a network provided in an embodiment of the present disclosure is applied to an NWDAF, wherein the device (1100) for subscribing to a model transmission state analysis in a network provided in this embodiment is,

[0365] A receiving unit (1101) that receives a first message sent from an application function (AF) directly or through a network capability exposure function (NEF); wherein the first message is intended to request a subscription to analysis information of the transmission status of an artificial intelligence / machine learning (AI / ML) model in the network;

[0366] Based on the parameters requested in the first message above, a second message is sent to another network function 5GC NF(s) of the 5G core network, and the second message is for collecting data to analyze the transmission status of an AI / ML model in the network; comprising a sending unit (1102).

[0367] The analysis unit (1103) is also for receiving data of the AI / ML model transmission status sent from other network functions 5GC NF(s) of the 5G core network, and for analyzing the data of the AI / ML model transmission status to obtain analysis information of the AI / ML model transmission status;

[0368] Here, the above analysis information is intended to adjust network strategy parameters and / or application layer model information through AF.

[0369] Optionally, the parameters requested in the first message include at least one of a network data analysis identifier, an identifier of a user device (UE) or a group of UEs receiving the AI / ML model or an identifier of any UE satisfying the analysis condition, an identifier of an application using the AI / ML model, a region where the AI / ML model is transmitted, a network slice instruction of a protocol data unit (PDU) session of a quality of service flow transmitting the AI / ML model, a data network instruction of a PDU session of a quality of service flow transmitting the AI / ML model, a duration of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, a size of the AI / ML model transmission, a quality of service requirement instruction of a quality of service flow transmitting the AI / ML model, and / or a specific quality of service requirement instruction for transmitting the AI / ML model;

[0370] The second message comprises at least one of the following: the current location of the UE using the AI / ML model, an identifier of the application using the AI / ML model, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate and a downlink bit rate transmitting the AI / ML model, an uplink packet delay and a downlink packet delay of the AI / ML model, a quantity of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, a quantity of packets transmitted by the AI / ML model, a quantity of packet retransmissions of the AI / ML model, a data collection time, a time length of AI / ML model transmission, a start timestamp of AI / ML model transmission, an end timestamp of AI / ML model transmission, a size of the AI / ML transmission model, a network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, a data network of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and a service process for the AF;

[0371] The above analysis information includes at least one of: a network slice of a PDU session of a Quality of Service flow for transmitting an AI / ML model, an identifier of an application using the AI / ML model, area information using the AI / ML model, the validity period of the analysis result, a User Plane Function (UPF) providing AI / ML model transmission, a data network name of a PDU session of a Quality of Service flow for transmitting an AI / ML model, the size of the AI / ML transmission model, the time length of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate for transmitting the AI / ML model and a downlink bit rate for transmitting the AI / ML model, an uplink packet delay of the AI / ML model and a downlink packet delay of the AI / ML model, the number of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, the number of times the Quality of Service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission, the number of packet transmissions of the AI / ML model, and the number of packet retransmissions of the AI / ML model;

[0372] Herein, if the AI / ML model performs federated learning, the parameter requested in the first message further includes federated learning group information, and the federated learning group information includes at least one of an identifier for indicating the federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0373] Correspondingly, the second message further includes at least one of an identifier for indicating a federated learning group to be analyzed, a UE identifier or UE(s) identifier participating in federated learning, and an application identifier participating in federated learning;

[0374] Correspondingly, the analysis information further includes at least one of an identifier for specifying the federated learning group to be analyzed, an identifier of a UE or UE(s) participating in federated learning, and an indication of each application identifier that provides an AI / ML model or participates in federated learning.

[0375] It should be explained here that the device subscribing to the model transmission state analysis in the network provided in this disclosure can implement all method steps implemented in the method embodiment of FIG. 4 and FIG. 7 and can achieve the same technical effects, and here, specific descriptions of parts identical to the method embodiment and advantageous effects in this embodiment are not repeated.

[0376] It should be noted that the division of units in the embodiments of the present disclosure is exemplary and merely a division of a type of logical function, and other division methods may exist when actually implemented. Furthermore, in each embodiment of the present disclosure, each functional unit may be integrated into a single processing unit, each unit may exist physically independently, or two or more units may be integrated into a single unit. The aforementioned integrated unit may be implemented in the form of hardware or in the form of a software functional unit.

[0377] If an integrated unit is implemented in the form of a software functional unit and sold or used as a separate product, it may be stored on a single processor-readable storage medium. Based on this understanding, the essence of the technical solution of the present disclosure, the part contributing to the prior art, or all or part of said technical solution may be implemented in the form of a software product, said computer software product may be stored on a single storage medium and may include some instructions so that a single computer device (which may be a personal computer, a server, or a network device, etc.) or a processor performs all or part of the steps of the method of each embodiment of the present disclosure. However, the storage medium described above includes various media capable of storing program code, such as a USB, a mobile hard disk, Read-Only Memory (ROM), Random Access Memory (RAM), a magnetic disk, or an optical disk.

[0378] Embodiments of the present disclosure further provide a processor-readable storage medium, wherein a computer program is stored in the processor-readable storage medium, and the computer program is intended to enable a processor to perform any one of the method embodiments described above.

[0379] Here, processor-readable storage media may be any available medium or data storage device that a processor can access, and include, but is not limited to, magnetic memory (e.g., floppy disk, hard disk, magnetic tape, magneto-optical disk (MO), etc.), optical memory (e.g., optical disk (CD), digital universal disk (DVD), Blu-ray disk (BD), holographic universal disk (HVD), etc.), and semiconductor memory (e.g., ROM, erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), non-volatile memory (NAND FLASH), solid-state disk (SSD)).

[0380] Those skilled in the art should understand that embodiments of the present disclosure may be provided as methods, systems, or computer program products. Accordingly, the present disclosure may be used in the form of complete hardware embodiments, complete software embodiments, or embodiments combining software and hardware aspects. Additionally, the present disclosure may be used in the form of computer program products implemented on one or more computer-usable storage media (including, but not limited to, magnetic disk memory and optical memory, etc.) containing computer-usable program code.

[0381] The present disclosure has been described with reference to flowcharts and / or block diagrams of the methods, apparatus (systems), and computer program products of the embodiments of the present disclosure. It should be understood that each process and / or block within the flowchart and / or block diagram, and combinations of processes and / or blocks within the flowchart and / or block diagram, can be implemented by computer-executable instructions. By providing such computer-executable instructions to a processor of a general-purpose computer, a dedicated computer, an embedded processor, or other programmable data processing device to create an apparatus, the instructions executed by the processor of the computer or other programmable data processing device may be used to create an apparatus for implementing a function specified in one or more processes of the flowchart and / or one or more boxes of the block diagram.

[0382] These processor-executable instructions are stored in processor-readable memory that can induce a computer or other programmable data processing device to operate in a specific manner, and the instructions stored in said processor-readable memory may produce a manufactured product including an instruction unit, said instruction unit implements a function specified in one or more processes of a flowchart and / or one or more blocks of a block diagram.

[0383] These processor-executable instructions may be loaded into a computer or other programmable data processing device to perform a series of operation steps on the computer or other programmable processing device to generate processing implemented by the computer, and accordingly, the instructions executed on the computer or other programmable processing device provide steps for implementing a function specified in one or more processes of a flowchart and / or one or more blocks of a block diagram.

[0384] It is obvious to those skilled in the art that various changes and modifications can be made to the present disclosure without departing from the spirit and scope of the present disclosure. Thus, if such modifications and modifications to the present disclosure fall within the scope of the claims of the present disclosure and the scope of the equivalent art, the present disclosure is intended to include such modifications and modifications.

Claims

Claim 1 In a method for subscribing to a model transmission state analysis in a network, the method for subscribing to a model transmission state analysis in a network is applied to an application function (AF), and the method for subscribing to a model transmission state analysis in a network sends a first message to a network data analysis function (NWDAF) directly or through a network capability exposure function (NEF); Herein, the first message is for requesting subscription to analysis information regarding the transmission status of an artificial intelligence / machine learning (AI / ML) model in a network; receiving analysis information regarding the transmission status of an AI / ML model sent by the NWDAF directly or through the NEF, wherein the analysis information is determined based on data regarding the transmission status of an AI / ML model sent by other network functions (5GC NF(s)) of the 5G core network received by the NWDAF, wherein the analysis information is for adjusting at least one of network strategy parameters and information on an application layer model; and after receiving the analysis information, the method for subscribing to the analysis of the model transmission status in the network further includes the step of sending a first request to a strategy control function (PCF) directly or through the NEF based on the analysis information; wherein the first request is for requesting an update to network strategy parameters for AI / ML model transmission; and the network strategy parameters are for optimizing the AI / ML model transmission status.Herein, the step of sending a first request to a strategy control function (PCF) directly or through the NEF based on the analysis information comprises determining a new quality of service parameter for transmitting an AI / ML model based on at least one of the following: the uplink direction bit rate and the downlink direction bit rate for transmitting the AI / ML model among the analysis information; the uplink direction packet delay and the downlink direction packet delay of the AI / ML model; the quantity of abnormal releases of the quality of service flow during the period of AI / ML model transmission; the quantity of packet transmissions of the AI / ML model; the quantity of packet retransmissions of the AI / ML model; and the number of times the quality of service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission; The above new service quality parameter comprises at least one of a 5G service quality identifier, reflective service quality control, a maximum bit rate in the uplink direction for transmitting an AI / ML model, a maximum bit rate in the downlink direction for transmitting an AI / ML model, a minimum bit rate in the uplink direction for transmitting an AI / ML model, a minimum bit rate in the downlink direction for transmitting an AI / ML model, and a priority of a service quality flow; based on the identifier of an application using an AI / ML model, area information using an AI / ML model, IP address information of an application service using an AI / ML model, a network slice of a protocol data unit (PDU) session of a service quality flow for transmitting an AI / ML model, and a data network name of a PDU session of a service quality flow for transmitting an AI / ML model among the analysis information, the area information and address information of the UE(s) transmitting the AI / ML model and their respective AFs are determined;Alternatively, if the AI / ML model performs federated learning, a step of determining the domain information and address information of the UE(s) transmitting the AI / ML model and each AF based on the indications of the federated learning group being analyzed among the analysis information, the UE identifier or UE(s) identifier participating in the federated learning, and each application identifier providing the AI / ML model or participating in the federated learning; a step of determining the data network access identifier (DNAI) and the domain information and address information of the UE(s) and each AF corresponding to the DNAI based on the domain information and address information of the UE(s) transmitting the AI / ML model and each AF, wherein the domain information and address information of the UE(s) and each AF corresponding to the DNAI is for providing a path that optimizes the AI / ML model transmission state; and a step of sending the new quality of service parameter, the DNAI, and the domain information and address information of the UE(s) and each AF corresponding to the DNAI to the PCF directly or through the NEF as parameters requested in the first request; characterized by comprising: a method for subscribing to a model transmission state analysis in a network. Claim 2 In claim 1, the data of the AI / ML model transmission status is obtained by sending a second message to the 5GC NF(s) based on the parameters requested in the first message received by the NWDAF, and the second message is for collecting data to analyze the AI / ML model transmission status in the network; wherein the parameters requested in the first message include at least one of a network data analysis identifier, an identifier of a single user device (UE) or a group of UEs receiving the AI / ML model or an identifier of any U satisfying the analysis conditions, an identifier of an application using the AI / ML model, an area where the AI / ML model is transmitted, a network slice instruction of the PDU session of the Quality of Service flow transmitting the AI / ML model, a data network instruction of the PDU session of the Quality of Service flow transmitting the AI / ML model, a duration of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, a size of the AI / ML transmission model, a Quality of Service requirement instruction of the Quality of Service flow transmitting the AI / ML model, and a specific Quality of Service requirement instruction for transmitting the AI / ML model;The second message comprises at least one of the following: the current location of the UE using the AI / ML model, an identifier of the application using the AI / ML model, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate and a downlink bit rate transmitting the AI / ML model, an uplink packet delay and a downlink packet delay of the AI / ML model, a quantity of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, a quantity of packets transmitted by the AI / ML model, a quantity of packet retransmissions of the AI / ML model, a data collection time, a time length of AI / ML model transmission, a start timestamp of AI / ML model transmission, an end timestamp of AI / ML model transmission, a size of the AI / ML transmission model, a network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, a data network of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and a service process for the AF;The above analysis information includes at least one of a network slice of a PDU session of a Quality of Service flow for transmitting an AI / ML model, an identifier of an application using the AI / ML model, area information using the AI / ML model, the validity period of the analysis result, a User Plane Function (UPF) providing AI / ML model transmission, a data network name of a PDU session of a Quality of Service flow for transmitting the AI / ML model, the size of the AI / ML transmission model, the time length of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate for transmitting the AI / ML model and a downlink bit rate for transmitting the AI / ML model, an uplink packet delay of the AI / ML model and a downlink packet delay of the AI / ML model, the number of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, the number of times the Quality of Service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission, the number of packet transmissions of the AI / ML model, and the number of packet retransmissions of the AI / ML model; wherein the AI / ML model is federated When learning is performed, the parameter requested in the first message further includes federated learning group information, and the federated learning group information includes at least one of an identifier for indicating the federated learning group to be analyzed, an identifier for a UE or UE(s) participating in federated learning, and an identifier for an application participating in federated learning; correspondingly, the second message further includes at least one of an identifier for indicating the federated learning group to be analyzed, an identifier for a UE or UE(s) participating in federated learning, and an identifier for an application participating in federated learning;Correspondingly, a method for subscribing to model transmission state analysis in a network, characterized in that the analysis information further comprises at least one of an identifier for indicating a federated learning group to be analyzed, an identifier of a UE or UE(s) participating in federated learning, and an indication of each application identifier that provides an AI / ML model or participates in federated learning. Claim 3 delete Claim 4 A method for subscribing to a model transmission state analysis in a network, characterized in that, in claim 1, the step of determining a data network access identifier (DNAI) and the area information and address information of the UE(s) and the AF corresponding to the DNAI based on the area information and address information of the UE(s) and the AF corresponding to the DNAI, comprises: a step of determining whether the current routing path is poor based on the area information and address information of the UE(s) and the AF corresponding to the AI / ML model; if the current routing path is poor, a step of determining the destination addresses of both parties during transmission based on the address information of the UE(s) and the AF corresponding to the AI / ML model and the area information of the UE(s); a step of determining the nearest path based on the destination addresses; and a step of determining the DNAI and the area information and address information of the UE(s) and the AF corresponding to the DNAI based on the nearest path. Claim 5 In claim 1, the first request specifically requests that the PCF adjust the 5G service quality identifier, reflective service quality control, the uplink direction maximum bit rate for transmitting AI / ML models, the downlink direction maximum bit rate for transmitting AI / ML models, the uplink direction minimum bit rate for transmitting AI / ML models, the downlink direction minimum bit rate for transmitting AI / ML models, and the priority of the service quality flow among the PCC rules based on the new service quality parameters, and instructs the PCF to feed back the first update result after adjustment directly or through the NEF; wherein the first update result is determined by the result of the PCF adjusting the PCC rules based on the new service quality parameters; correspondingly, the method for subscribing to a model transmission status analysis in the network further comprises the step of receiving the first update result sent by the PCF directly or through the NEF, wherein the first update result includes the acceptance of the first request or the rejection of the first request. Claim 6 In paragraph 1, the first request specifically requests the PCF to determine whether a session management function (SMF) network element needs to update a session management strategy, and if it is determined that the SMF needs to update the session management strategy, the PCF determines to send a second request to the SMF, and the parameters requested in the second request include at least one of a DNAI, a traffic steering policy identifier, and traffic routing information; The second request is to confirm the User Plane Function (UPF) selected by the SMF based on a new session management strategy and to provide the corresponding DNAI, traffic steering policy identifier, and traffic routing information; correspondingly, a method for subscribing to a model transmission state analysis in the network further comprises the step of receiving a second update result sent from the PCF directly or through the NEF, wherein the second update result is confirmed by the PCF based on whether the new session management strategy sent from the SMF has updated the UPF path; wherein the second update result includes whether the second request was accepted or rejected. Claim 7 In claim 1 or 2, the method of subscribing to an analysis of the model transmission status in the network after receiving the analysis information adjusts the information of the application layer model based on the analysis information, and the information of the application layer model includes at least one of model compression, model size, model transmission period, model encoding, and decoding; A method for subscribing to a model transmission status analysis in a network, characterized by comprising: a step of providing information on the application layer model for updating service quality parameters; a step of determining new service quality parameters based on the information on the application layer model after adjustment, wherein the new service quality parameters include a 5G service quality identifier, reflective service quality control, a maximum bit rate in the uplink direction for transmitting an AI / ML model, a maximum bit rate in the downlink direction for transmitting an AI / ML model, a minimum bit rate in the uplink direction for transmitting an AI / ML model, a minimum bit rate in the downlink direction for transmitting an AI / ML model, and a priority of the service quality flow; and further comprising a step of sending a third request to the PCF directly or through the NEF, wherein the parameters requested in the third request include the new service quality parameters, and the third request is for requesting an update of the service quality parameters. Claim 8 In claim 7, the third request is specifically to request that the PCF adjust the 5G service quality identifier, reflective service quality control, the uplink direction maximum bit rate for transmitting AI / ML models, the downlink direction maximum bit rate for transmitting AI / ML models, the uplink direction minimum bit rate for transmitting AI / ML models, the downlink direction minimum bit rate for transmitting AI / ML models, and the priority of the service quality flow among the PCC rules based on the new service quality parameters; correspondingly, the method for subscribing to a model transmission status analysis in the network further comprises the step of receiving a third update result sent by the PCF directly or through the NEF, wherein the third update result is determined based on the result of the PCF adjusting the PCC rules, and the third update result includes whether the third request was accepted or the third request was rejected. Claim 9 In claim 7, the method for subscribing to a model transmission status analysis in the network after adjusting the information of the application layer model further comprises the step of directly sending the model compression, model size, and model encoding and decoding among the information of the application layer model after adjustment to the PCF, and the information of the application layer model after adjustment is intended to support the PCF in adjusting the 5G Quality of Service identifier, reflective Quality of Service control, the maximum bit rate in the uplink direction for transmitting the AI / ML model, the maximum bit rate in the downlink direction for transmitting the AI / ML model, the minimum bit rate in the uplink direction for transmitting the AI / ML model, the minimum bit rate in the downlink direction for transmitting the AI / ML model, and the priority of the Quality of Service flow among the PCC rules; correspondingly, the method for subscribing to a model transmission status analysis in the network further comprises the step of receiving a fourth update result sent from the PCF, wherein the fourth update result is determined based on the result of adjusting the PCC rules based on the information of the application layer model after adjustment, and the fourth update result includes the step of the third request being accepted or the third request being rejected. Further including; or, a method for subscribing to a model transmission state analysis in the network after adjusting the information of the application layer model further includes the step of directly sending the model transmission period among the information of the application layer model after adjustment to the PCF, wherein the model transmission period among the information of the application layer model after adjustment is to support the PCF in adjusting the gate state parameter among the PCC rules; and wherein the gate state parameter is to support the SMF in updating a session management strategy based on the transmission start time and transmission end time among the gate states;Correspondingly, a method for subscribing to a model transmission status analysis in the network further comprises the step of receiving a fifth update result sent by the PCF, wherein the fifth update result is confirmed by the PCF receiving the result of a new session management strategy sent by the SMF, and wherein the fifth update result includes the acceptance of the third request or the rejection of the third request. Claim 10 In a method for subscribing to a model transmission state analysis in a network, the method for subscribing to a model transmission state analysis in a network is applied to a network data analysis function (NWDAF), and the method for subscribing to a model transmission state analysis in a network receives a first message sent from an application function (AF) directly or through a network capability exposure function (NEF); Herein, the first message is for requesting a subscription to analysis information regarding the transmission status of an artificial intelligence / machine learning (AI / ML) model in a network; the second message is sent to another network function 5GC NF(s) of the 5G core network based on the parameters requested in the first message, and the second message is for collecting data to analyze the transmission status of the AI / ML model in the network; the method includes receiving data regarding the transmission status of the AI / ML model sent from another network function 5GC NF(s) of the 5G core network, analyzing the data regarding the transmission status of the AI / ML model, and obtaining analysis information regarding the transmission status of the AI / ML model, wherein the analysis information is for adjusting at least one of network strategy parameters and information of an application layer model through the AF; the analysis information is also used for the AF to send a first request to a strategy control function (PCF), either directly or through the NEF, based on the analysis information; wherein the first request is for requesting an update to network strategy parameters for AI / ML model transmission; The above network strategy parameters are intended to optimize the AI / ML model transmission state;Herein, the analysis information specifically determines a new quality of service parameter for transmitting an AI / ML model based on at least one of the following: the uplink bit rate and downlink bit rate for transmitting the AI / ML model among the analysis information; the uplink packet delay and downlink packet delay of the AI / ML model; the quantity of abnormal releases of the quality of service flow during the period of AI / ML model transmission; the quantity of packet transmissions of the AI / ML model; the quantity of packet retransmissions of the AI / ML model; and the number of times the quality of service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission; The above new service quality parameter comprises at least one of a 5G service quality identifier, reflective service quality control, a maximum bit rate in the uplink direction for transmitting an AI / ML model, a maximum bit rate in the downlink direction for transmitting an AI / ML model, a minimum bit rate in the uplink direction for transmitting an AI / ML model, a minimum bit rate in the downlink direction for transmitting an AI / ML model, and a priority of a service quality flow; based on the identifier of an application using an AI / ML model, area information using an AI / ML model, IP address information of an application service using an AI / ML model, a network slice of a protocol data unit (PDU) session of a service quality flow for transmitting an AI / ML model, and a data network name of a PDU session of a service quality flow for transmitting an AI / ML model among the analysis information, the area information and address information of the UE(s) transmitting the AI / ML model and their respective AFs are determined;Alternatively, if the AI / ML model performs federated learning, a step of determining the domain information and address information of the UE(s) transmitting the AI / ML model and each AF based on the indications of an identifier indicating the federated learning group analyzed by the AF among the analysis information, a UE identifier or UE(s) identifier participating in the federated learning, and each application identifier providing the AI / ML model or participating in the federated learning; a step of determining the Data Network Access Identifier (DNAI) and the domain information and address information of the UE(s) and each AF corresponding to the DNAI based on the domain information and address information of the UE(s) transmitting the AI / ML model and each AF, wherein the domain information and address information of the UE(s) and each AF corresponding to the DNAI is for providing a path to optimize the AI / ML model transmission state; and a step of sending the new quality of service parameter, the DNAI, and the domain information and address information of the UE(s) and each AF corresponding to the DNAI to the PCF directly or through the NEF as parameters requested in the first request; characterized in that the method of subscribing to a model transmission state analysis in a network is used to perform the above steps. Claim 11 In paragraph 10, the parameters requested in the first message include a network data analysis identifier, an identifier of a single user device (UE) or a group of UEs receiving the AI / ML model or an identifier of any U satisfying the analysis condition, an identifier of the application using the AI / ML model, a region where the AI / ML model is transmitted, a network slice instruction of the PDU session of the Quality of Service flow transmitting the AI / ML model, a data network instruction of the PDU session of the Quality of Service flow transmitting the AI / ML model, a duration of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, a size of the AI / ML model transmitted, a Quality of Service requirement instruction of the Quality of Service flow transmitting the AI / ML model, and a specific Quality of Service requirement instruction for transmitting the AI / ML model; and the second message includes the current location of the UE using the AI / ML model, an identifier of the application using the AI / ML model, a Quality of Service flow identifier transmitting the AI / ML model, an uplink bit rate transmitting the AI / ML model and a downlink bit rate transmitting the AI / ML model, an uplink packet delay of the AI / ML model, and the AI / ML model Downlink direction packet delay, quantity of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, quantity of packet transmissions of the AI / ML model, quantity of packet retransmissions of the AI / ML model, data collection time, time length of AI / ML model transmission, start timestamp of AI / ML model transmission, end timestamp of AI / ML model transmission, size of the AI / ML transmission model, network slice of the PDU session of the Quality of Service flow for transmitting the AI / ML model, data network of the PDU session of the Quality of Service flow for transmitting the AI / ML model, and at least one of the service processes for the AF;The above analysis information includes at least one of a network slice of a PDU session of a Quality of Service flow for transmitting an AI / ML model, an identifier of an application using the AI / ML model, area information using the AI / ML model, the validity period of the analysis result, a User Plane Function (UPF) providing AI / ML model transmission, a data network name of a PDU session of a Quality of Service flow for transmitting the AI / ML model, the size of the AI / ML transmission model, the time length of the AI / ML model transmission, a start timestamp of the AI / ML model transmission, an end timestamp of the AI / ML model transmission, an identifier of the Quality of Service flow transmitting the AI / ML model, an uplink bit rate for transmitting the AI / ML model and a downlink bit rate for transmitting the AI / ML model, an uplink packet delay of the AI / ML model and a downlink packet delay of the AI / ML model, the number of abnormal releases of the Quality of Service flow during the period of AI / ML model transmission, the number of times the Quality of Service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission, the number of packet transmissions of the AI / ML model, and the number of packet retransmissions of the AI / ML model; wherein the AI / ML model is federated When learning is performed, the parameter requested in the first message further includes federated learning group information, and the federated learning group information includes at least one of an identifier for indicating the federated learning group to be analyzed, an identifier for a UE or UE(s) participating in federated learning, and an identifier for an application participating in federated learning; correspondingly, the second message further includes at least one of an identifier for indicating the federated learning group to be analyzed, an identifier for a UE or UE(s) participating in federated learning, and an identifier for an application participating in federated learning;Correspondingly, a method for subscribing to model transmission state analysis in a network, characterized in that the analysis information further comprises at least one of an identifier for indicating a federated learning group to be analyzed, an identifier of a UE or UE(s) participating in federated learning, and an indication of each application identifier that provides an AI / ML model or participates in federated learning. Claim 12 In a device subscribing to model transmission state analysis in a network, the device sends a first message to a network data analysis function (NWDAF) directly or through a network capability exposure function (NEF); Herein, the first message comprises a sending unit for requesting a subscription to analysis information regarding the transmission status of an artificial intelligence / machine learning (AI / ML) model in a network; an analysis unit that receives analysis information regarding the transmission status of an AI / ML model sent by the NWDAF directly or through the NEF, wherein the analysis information is determined based on data regarding the transmission status of an AI / ML model sent by other network functions (5GC NF(s)) of the 5G core network received by the NWDAF; wherein the analysis information is for adjusting at least one of network strategy parameters and information of an application layer model; and the sending unit also, after receiving the analysis information, sends a first request to a strategy control function (PCF) directly or through the NEF based on the analysis information; wherein the first request is for requesting an update to network strategy parameters for AI / ML model transmission; The above network strategy parameter is intended to optimize the AI / ML model transmission state; wherein, specifically, the sending unit determines a new quality of service parameter for transmitting the AI / ML model based on at least one of the following: the uplink direction bit rate and the downlink direction bit rate for transmitting the AI / ML model among the analysis information; the uplink direction packet delay and the downlink direction packet delay of the AI / ML model; the quantity of abnormal releases of the quality of service flow during the period of AI / ML model transmission; the quantity of packet transmissions of the AI / ML model; the quantity of packet retransmissions of the AI / ML model; and the number of times the quality of service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission;The above new service quality parameter includes at least one of a 5G service quality identifier, reflective service quality control, a maximum bit rate in the uplink direction for transmitting an AI / ML model, a maximum bit rate in the downlink direction for transmitting an AI / ML model, a minimum bit rate in the uplink direction for transmitting an AI / ML model, a minimum bit rate in the downlink direction for transmitting an AI / ML model, and a priority of a service quality flow; and based on the identifier of the application using the AI / ML model, area information using the AI / ML model, IP address information of the application service using the AI / ML model, a network slice of the Protocol Data Unit (PDU) session of the service quality flow for transmitting the AI / ML model, and the data network name of the PDU session of the service quality flow for transmitting the AI / ML model among the analysis information, the area information and address information of the UE(s) transmitting the AI / ML model and their respective AFs are determined; Alternatively, if the AI / ML model performs federated learning, the domain information and address information of the UE(s) transmitting the AI / ML model and each AF are determined based on the indications of the identifier indicating the federated learning group analyzed by the AF among the analysis information, the UE identifier or UE(s) identifier participating in the federated learning, and each application identifier providing the AI / ML model or participating in the federated learning; and the domain information and address information of the UE(s) transmitting the AI / ML model and each AF are determined based on the domain information and address information of the UE(s) corresponding to the DNAI and each AF, and the domain information and address information of the UE(s) corresponding to the DNAI and each AF are intended to provide a path that optimizes the AI / ML model transmission state;A device for subscribing to a model transmission state analysis in a network, characterized by sending the above new service quality parameters, the DNAI, the area information and address information of the UE(s) corresponding to the DNAI and each AF, as parameters requested in the first request, directly or through the NEF to the PCF. Claim 13 An apparatus characterized in that, in claim 12, the data of the AI / ML model transmission status is obtained by sending a second message to the 5GC NF(s) based on the parameters requested in the first message received by the NWDAF, and the second message is for collecting data to analyze the AI / ML model transmission status in the network. Claim 14 A device for subscribing to model transmission state analysis in a network, wherein the device receives a first message sent from an application function (AF) directly or through a network capability exposure function (NEF); Herein, the first message is for requesting a subscription to analysis information regarding the transmission status of an artificial intelligence / machine learning (AI / ML) model in a network, comprising: a receiving unit that sends a second message to another network function 5GC NF(s) of a 5G core network based on parameters requested in the first message, wherein the second message is for collecting data to analyze the transmission status of an AI / ML model in a network, comprising: an analysis unit that receives data regarding the transmission status of an AI / ML model sent from another network function 5GC NF(s) of a 5G core network, analyzes the data regarding the transmission status of an AI / ML model, and obtains analysis information regarding the transmission status of an AI / ML model; wherein the analysis information is for adjusting at least one of network strategy parameters and information of an application layer model through the AF; and the analysis information is also used for the AF to send a first request to a strategy control function (PCF), either directly or through the NEF, based on the analysis information; wherein the first request is for requesting an update to network strategy parameters for AI / ML model transmission. The above network strategy parameters are intended to optimize the AI / ML model transmission state;Herein, the analysis information specifically determines a new quality of service parameter for transmitting an AI / ML model based on at least one of the following: the uplink bit rate and downlink bit rate for transmitting the AI / ML model among the analysis information; the uplink packet delay and downlink packet delay of the AI / ML model; the quantity of abnormal releases of the quality of service flow during the period of AI / ML model transmission; the quantity of packet transmissions of the AI / ML model; the quantity of packet retransmissions of the AI / ML model; and the number of times the quality of service flow reached a reporting threshold for abnormal releases during the period of AI / ML model transmission; The above-mentioned new service quality parameter comprises at least one of a 5G service quality identifier, reflective service quality control, a maximum bit rate in the uplink direction for transmitting an AI / ML model, a maximum bit rate in the downlink direction for transmitting an AI / ML model, a minimum bit rate in the uplink direction for transmitting an AI / ML model, a minimum bit rate in the downlink direction for transmitting an AI / ML model, and a priority of a service quality flow; based on the identifier of an application using an AI / ML model, area information using an AI / ML model, IP address information of an application service using an AI / ML model, a network slice of a protocol data unit (PDU) session of a service quality flow for transmitting an AI / ML model, and a data network name of a PDU session of a service quality flow for transmitting an AI / ML model among the analysis information, the area information and address information of the UE(s) transmitting the AI / ML model and their respective AFs are determined;Alternatively, if the AI / ML model performs federated learning, the device is characterized by being used to perform the following operations: determining the area information and address information of the UE(s) transmitting the AI / ML model and each AF based on the indications of the federated learning group being analyzed among the analysis information, the UE identifier or UE(s) identifier participating in the federated learning, and each application identifier providing the AI / ML model or participating in the federated learning; determining the area information and address information of the UE(s) transmitting the AI / ML model and each AF based on the area information and address information of each AF and each AF, wherein the area information and address information of the UE(s) and each AF corresponding to the DNAI are for providing a path that optimizes the AI / ML model transmission state; and sending the new quality of service parameter, the DNAI, and the area information and address information of the UE(s) and each AF corresponding to the DNAI to the PCF directly or through the NEF as parameters requested in the first request. Claim 15 A processor-readable storage medium, wherein a computer program is stored in the processor-readable storage medium, and the computer program is configured to enable the processor to perform a method of subscribing to a model transmission state analysis in a network according to any one of claims 1, 2, 4 through 6, 10, and 11. Claim 16 delete Claim 17 delete Claim 18 delete Claim 19 delete Claim 20 delete Claim 21 delete Claim 22 delete Claim 23 delete

Images