Communication method and communication apparatus

Abstract

Provided in the present application are a communication method and a communication apparatus. The communication method comprises: a first network function receiving an analysis request message from a network consumer, wherein the analysis request message comprises a first analysis requirement for feeding back a first analysis result; and on the basis of the first message, the first network function requesting that a second network function determines a second model, wherein the second model is used for assisting a first model, which is deployed on the first network function, in acquiring the first analysis result, so as to meet the first analysis requirement. By means of the method in the embodiments of the present application, when a single local model in a first network function cannot meet the analysis requirement given by a network consumer, the first network function requests the acquisition of another model for collaborative inference, such that the analysis requirement of the network consumer for an analysis result is met, and the capability of guaranteeing a network service is enhanced.

Description

Communication method and communication device

[0001] This application claims priority to the Chinese patent application filed with the State Intellectual Property Office of China on January 2, 2024, with application number 202410013946.5 and application name “Communication Method and Communication Device”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the field of communication technology, and more specifically, to a communication method and a communication device. Background Art

[0003] The network data analytics function (NWDAF) has data collection, training, analysis, and reasoning functions. It can be used to collect relevant data from network elements, third-party servers, terminal devices, or network management systems, perform analysis and training based on the relevant data, and provide data analysis results to network elements, third-party servers, terminal devices, or network management systems. The network consumers served by the NWDAF (for example, network elements, third-party servers, terminal devices, or network management systems) expect the analysis results fed back by the NWDAF to meet given analysis requirements (for example, time requirements and / or accuracy requirements). When the NWDAF cannot meet the analysis requirements given by the network consumers, the NWDAF will request the network consumers to lower the analysis requirements for the feedback analysis results. This method can only be applied to situations where the network consumers do not have strict analysis requirements for the current analysis. When the network consumers have strict requirements for the current analysis, the current technology cannot meet the analysis requirements of the network consumers. Summary of the Invention

[0004] The present application provides a communication method and a communication device that can be implemented.

[0005] In a first aspect, a communication method is provided. The method can be performed by a first network function, or by a module (such as a chip or circuit) in the first network function, or by a logical node, logical module, or software that can implement all or part of the first network function. This application is not limited to this.

[0006] The method includes: a first network function receives a first message from a first device, the first message including a first analysis requirement for feeding back a first analysis result, and the first network function is deployed with a first model; the first network function sends a second message to a second network function based on the first message, the second message is used to determine a second model, and the second model is used to assist the first model in obtaining the first analysis result before the first time.

[0007] Exemplarily, the first device is a network consumer, such as a network element, a third-party server, a terminal device, or a network management system.

[0008] Exemplarily, the first network function and the second network function may be NWDAF.

[0009] Exemplarily, the first analysis requirement includes a first time requirement and / or a first accuracy requirement.

[0010] Through the above method, when a single local model in the first network function cannot meet the analysis requirements given by the network consumer, the first network function will request to obtain the other party's model for collaborative reasoning, thereby meeting the network consumer's analysis requirements for analysis results and enhancing the network service guarantee capability.

[0011] In combination with the first aspect, in some implementations of the first aspect, the second message includes first indication information, and the first indication information is used to indicate an execution subject of the second model.

[0012] Exemplarily, if the first network function has sufficient local resources to execute multiple models, the above-mentioned first indication information may indicate that the execution subject of the second model is the first network function; if the first network function does not have sufficient local resources to execute multiple models, the above-mentioned first indication information may indicate that the execution subject of the second model is other nodes or other devices, etc.

[0013] In conjunction with the first aspect, in certain implementations of the first aspect, the second message includes at least one of the following information:

[0014] The inference speed of the second model and the analysis type of the second model.

[0015] Through the above method, the first network function can determine the parameters of the other party's model based on the time requirement of the first device requesting feedback of the analysis results and the parameters of the local model, and feed back the parameters of the other party's model to the second network function. The second network function searches for the other party's model that meets the requirements, which can reduce the processing complexity of the second network function.

[0016] In conjunction with the first aspect, in certain implementations of the first aspect, the second message includes at least one of the following information:

[0017] The inference speed of the first model, the analysis type of the first model, and the acceleration ratio of the accelerated inference.

[0018] Through the above method, the first network function can directly feed back the parameters of the local model and the parameters of accelerated inference to the second network function. The second model determines the parameters of the other party's model and searches for the other party's model that meets the requirements, which can reduce the processing complexity of the first network function.

[0019] In combination with the first aspect, in some implementations of the first aspect, the above method also includes: the first network function receives a third message from the above second network function, where the third message is used to indicate the second model; the first network function uses the first model and the second model to perform reasoning to obtain a first analysis result.

[0020] Through the above method, when the first network function has sufficient local resources to execute multiple models, the first network function can execute multiple models, thereby reducing the delay of signaling interaction between multiple models.

[0021] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the first network function obtains the input data requirements of the first model and the output data requirements of the second model.

[0022] Specifically, the input data requirements of the above-mentioned first model include the input data of the above-mentioned first model, and the output data requirements of the above-mentioned second model include the number N of output values ​​of each inference of the second model during the inference process, where N is a positive integer greater than or greater than 1.

[0023] Exemplarily, the requirement for the first network function to obtain the input data of the first model may specifically include: the above-mentioned first network function may collect the input data of the first model according to the above-mentioned first message, or the DCCF may schedule the input data of the first model to the above-mentioned first network function. This application does not limit this.

[0024] Exemplarily, the requirement for the first network function to obtain the output data of the second model may specifically include: the requirement for the first network function to generate the output data of the second model, or the requirement for the second device to generate the output data of the second model and instruct it to the first network function, and the second device is the execution entity of the second model.

[0025] In combination with the first aspect, in certain implementations of the first aspect, the above method also includes: the first network function receives N first output values ​​from the second device; the first network function uses the first model to verify the N first output values, and determines to reject M first output values ​​of the N first output values; the first network function sends a fourth message to the second device, where the fourth message is used to indicate the M first output values.

[0026] Exemplarily, the fourth message may indicate the M first output values ​​by directly indicating the number of the M first output values, or the fourth message may indicate the M first output values ​​by indicating the first ratio.

[0027] Through the above method, the execution subject of the first model and the execution subject of the second model can align the input data requirements of the model and the output data requirements of the model, which can ensure that the negotiation reasoning process of multiple models proceeds smoothly.

[0028] In combination with the first aspect, in some implementations of the first aspect, the method further includes: the first network function obtains the requirements of the input data of the first model and the output data of the first model.

[0029] Specifically, the input data requirements of the above-mentioned first model include the input data of the above-mentioned first model, and the output data requirements of the above-mentioned first model include the number N of output values ​​of each inference of the first model during the inference process, where N is a positive integer greater than or greater than 1.

[0030] Exemplarily, the requirement for the first network function to obtain the input data of the first model may specifically include: the above-mentioned first network function may collect the input data of the first model according to the above-mentioned first message, or the DCCF may schedule the input data of the first model to the above-mentioned first network function. This application does not limit this.

[0031] Exemplarily, the requirement for the first network function to obtain the output data of the first model may specifically include: the requirement for the first network function to generate the output data of the first model, or the requirement for the second device to generate the output data of the first model and instruct it to the first network function, and the second device is the execution entity of the second model.

[0032] In combination with the first aspect, in certain implementations of the first aspect, the above method also includes: the first network function sends N first output values ​​to the second device; the first network function receives a fifth message from the second device, the fifth message being used to indicate M first output values, and the M first output values ​​are output values ​​rejected by the second device among the above N first output values.

[0033] Exemplarily, the fifth message may indicate the M first output values ​​by directly indicating the number of the M first output values, or the fifth message may indicate the M first output values ​​by indicating the first ratio.

[0034] Through the above method, the execution subject of the first model and the execution subject of the second model can align the input data requirements of the model and the output data requirements of the model, which can ensure that the negotiation reasoning process of multiple models proceeds smoothly.

[0035] In combination with the first aspect, in certain implementations of the first aspect, before the above-mentioned first network function receives the first message from the first device, the above-mentioned method also includes: the first network function receives a seventh message from the first device, and the seventh message includes a second analysis requirement for feedback of the first analysis result; when the first network function does not meet the first analysis requirement, the first network function sends an eighth message to the first device, and the eighth message includes a third analysis requirement for feedback of the first analysis result, and the third analysis requirement is lower than the above-mentioned second analysis requirement, and the third analysis requirement is lower than the above-mentioned first analysis requirement.

[0036] For example, when the first network function fails to meet the second time requirement included in the second analysis requirement using the first model for reasoning (i.e., the time taken by the first network function to reason using the first model exceeds the second time requirement), the third analysis requirement fed back by the first network function to the first device includes a third time requirement, where the third time requirement is the time requirement that can be met by the first network function using the first model for reasoning. In this case, the time required by the first time requirement included in the first analysis requirement fed back by the first device is later than the time required by the second time requirement but earlier than the time required by the third time requirement.

[0037] For example, when the first network function fails to meet the second accuracy requirement included in the second analysis requirement through reasoning using the first model, the third analysis requirement fed back by the first network function to the first device includes a third accuracy requirement, where the third accuracy requirement is the accuracy requirement that can be met by the first network function through reasoning using the first model. In this case, the first accuracy requirement included in the first analysis requirement fed back by the first device may be lower than or equal to the second accuracy requirement included in the second analysis requirement, but the first accuracy requirement included in the first analysis requirement fed back by the first device may be higher than the third accuracy requirement included in the third analysis requirement.

[0038] Through the above method, the first network function does not need to pre-determine whether the analysis requirements can be met based on the first message. When the analysis requirements are not met, the first network function requests the network consumer to lower the analysis requirements. The network consumer may moderately lower the analysis requirements or may not lower the analysis requirements at all. Based on this, the first network device triggers the multi-model collaborative reasoning process. This method can reduce the processing complexity of the first network function.

[0039] In a second aspect, a communication method is provided. The method can be performed by a second network function, or by a module (such as a chip or circuit) in the second network function, or by a logical node, logical module or software that can implement all or part of the second network function. This application is not limited to this.

[0040] The method includes: a second network function receives a second message from a first network function; the second network function determines a second model based on the second message, the second model is used to assist the first model in obtaining a first analysis result to meet the first analysis requirement, and the first model is deployed on the first network function.

[0041] Exemplarily, the first network function and the second network function may be NWDAF.

[0042] Exemplarily, the first analysis requirement includes a first time requirement and / or a first accuracy requirement.

[0043] Through the above method, when the second network function can obtain the other party's model based on the request of the first network function to cooperate with the local model of the first network function for collaborative reasoning, it can meet the analysis requirements of network consumers for analysis results and enhance the network service guarantee capability.

[0044] In combination with the second aspect, in some implementations of the second aspect, the second message includes first indication information, where the first indication information is used to indicate an execution subject of the second model.

[0045] Exemplarily, if the first network function has sufficient local resources to execute multiple models, the above-mentioned first indication information may indicate that the execution subject of the second model is the first network function; if the first network function does not have sufficient local resources to execute multiple models, the above-mentioned first indication information may indicate that the execution subject of the second model is other nodes or other devices, etc.

[0046] In conjunction with the second aspect, in certain implementations of the second aspect, the second message includes at least one of the following information:

[0047] The inference speed of the second model and the analysis type of the second model.

[0048] Through the above method, the first network function can determine the parameters of the other party's model based on the time requirement of the first device requesting feedback of the analysis results and the parameters of the local model, and feed back the parameters of the other party's model to the second network function. The second network function searches for the other party's model that meets the requirements, which can reduce the processing complexity of the second network function.

[0049] In conjunction with the second aspect, in certain implementations of the second aspect, the second message includes at least one of the following information:

[0050] The inference speed of the first model, the analysis type of the first model, and the acceleration ratio of the accelerated inference.

[0051] Through the above method, the first network function can directly feed back the parameters of the local model and the parameters of accelerated inference to the second network function. The second model determines the parameters of the other party's model and searches for the other party's model that meets the requirements, which can reduce the processing complexity of the first network function.

[0052] In combination with the second aspect, in some implementations of the second aspect, the method further includes: the second network function sending a third message to the first network function, where the third message is used to indicate the second model.

[0053] Through the above method, when the first network function has sufficient local resources to execute multiple models, the first network function can execute multiple models, thereby reducing the delay of signaling interaction between multiple models.

[0054] In combination with the second aspect, in some implementations of the second aspect, the above method also includes: the second network function sends a sixth message to the second device, the sixth message is used to indicate the first network function and the above second model, and the second device is the execution entity of the second model.

[0055] Through the above method, a connection can be established between the execution subject of the first model and the execution subject of the second model, thereby ensuring that the negotiation and reasoning process of the multiple models proceeds smoothly.

[0056] On the third aspect, a communication method is provided, which can be executed by a second device, or by a module (such as a chip or circuit) in the second device, or by a logical node, logical module or software that can implement all or part of the second device. This application does not limit this.

[0057] The method includes: a second device receives a sixth message from a second network function, the sixth message is used to indicate a first network function and a second model, and the first network function is deployed with the first model; the second device uses the second model for reasoning, and the second model is used to assist the first model in obtaining a first analysis result to meet the first analysis requirement.

[0058] Exemplarily, the first analysis requirement includes a first time requirement and / or a first accuracy requirement.

[0059] Through the above method, a connection can be established between the execution subject of the first model and the execution subject of the second model, thereby ensuring that the negotiation and reasoning process of the multiple models proceeds smoothly.

[0060] In combination with the third aspect, in some implementations of the third aspect, the method further includes: the second device obtaining the requirements of the input data of the second model and the output data of the second model.

[0061] Specifically, the input data requirements of the above-mentioned second model include the input data of the above-mentioned second model, and the output data requirements of the above-mentioned second model include the number N of output values ​​of each inference of the second model during the inference process, where N is a positive integer greater than or greater than 1.

[0062] Exemplarily, the requirement for the second device to obtain the input data of the second model may specifically include: the above-mentioned first network function may indicate the collected input data of the model to the second device, or the DCCF may schedule the input data of the second model to the second device, which is not limited in this application.

[0063] Exemplarily, the requirement for the second device to obtain the output data of the second model may specifically include: the requirement for the second device to generate the output data of the second model, or the requirement for the first network function to generate the output data of the second model and instruct the second device.

[0064] In combination with the third aspect, in certain implementations of the third aspect, the above method also includes: the second device sends N first output values ​​to the first network function; the second device receives a fourth message from the first network function, the fourth message being used to indicate the M first output values, the M first output values ​​being output values ​​rejected by the first network function among the N first output values, where M is a positive integer less than or equal to N.

[0065] Exemplarily, the fourth message may indicate the M first output values ​​by directly indicating the number of the M first output values, or the fourth message may indicate the M first output values ​​by indicating the first ratio.

[0066] Through the above method, the execution subject of the first model and the execution subject of the second model can align the input data requirements of the model and the output data requirements of the model, which can ensure that the negotiation reasoning process of multiple models proceeds smoothly.

[0067] In combination with the third aspect, in some implementations of the third aspect, the method further includes: the second device obtaining the requirements of the input data of the second model and the output data of the first model.

[0068] Specifically, the input data requirements of the above-mentioned second model include the input data of the above-mentioned second model, and the output data requirements of the above-mentioned first model include the number N of output values ​​of each inference of the first model during the inference process, where N is a positive integer greater than or greater than 1.

[0069] Exemplarily, the requirement for the second device to obtain the input data of the second model may specifically include: the above-mentioned first network function may indicate the collected input data of the model to the second device, or the DCCF may schedule the input data of the second model to the second device, which is not limited in this application.

[0070] Exemplarily, the requirement for the second device to obtain the output data of the first model may specifically include: the requirement for the second device to generate the output data of the first model, or the requirement for the first network function to generate the output data of the first model and instruct the second device.

[0071] In combination with the third aspect, in certain implementations of the third aspect, the above method also includes: the second device receives N first output values ​​from the first network function; the second device uses the second model to verify the N first output values, and determines to reject M first output values ​​of the N first output values; the second device sends a fifth message to the first network function, where the fifth message is used to indicate the M first output values.

[0072] Exemplarily, the fifth message may indicate the M first output values ​​by directly indicating the number of the M first output values, or the fifth message may indicate the M first output values ​​by indicating the first ratio.

[0073] Through the above method, the execution subject of the first model and the execution subject of the second model can align the input data requirements of the model and the output data requirements of the model, which can ensure that the negotiation reasoning process of multiple models proceeds smoothly.

[0074] In a fourth aspect, a communication device is provided, which includes: a transceiver unit for receiving a first message from a first device, the first message including a first analysis requirement for feedback of a first analysis result, and the first network function is deployed with a first model; the transceiver unit is also for sending a second message to a second network function based on the above-mentioned first message, the second message is used to determine a second model, and the second model is used to assist the first model in obtaining the first analysis result to meet the first analysis requirement.

[0075] In combination with the fourth aspect, in some implementations of the fourth aspect, the second message includes first indication information, and the first indication information is used to indicate an execution subject of the second model.

[0076] In conjunction with the fourth aspect, in certain implementations of the fourth aspect, the second message includes at least one of the following information:

[0077] The inference speed of the second model and the analysis type of the second model.

[0078] In conjunction with the fourth aspect, in certain implementations of the fourth aspect, the second message includes at least one of the following information:

[0079] The inference speed of the first model, the analysis type of the first model, and the acceleration ratio of the accelerated inference.

[0080] In combination with the fourth aspect, in certain implementations of the fourth aspect, the above-mentioned transceiver unit is also used to receive a third message from the above-mentioned second network function, and the third message is used to indicate the second model; the above-mentioned communication device also includes: a processing unit, used to use the first model and the second model to perform inference to obtain a first analysis result.

[0081] In combination with the fourth aspect, in certain implementations of the fourth aspect, the transceiver unit is further configured to obtain input data requirements of the first model and output data requirements of the second model.

[0082] In combination with the fourth aspect, in certain implementations of the fourth aspect, the above-mentioned transceiver unit is also used to receive N first output values ​​from the second device; the above-mentioned processing unit is also used to use the first model to verify the N first output values, and determine to reject M first output values ​​among the N first output values; the above-mentioned transceiver unit is also used to send a fourth message to the second device, and the fourth message is used to indicate the M first output values.

[0083] In combination with the fourth aspect, in certain implementations of the fourth aspect, the transceiver unit is further configured to obtain input data requirements of the first model and output data requirements of the first model.

[0084] In combination with the fourth aspect, in certain implementations of the fourth aspect, the above-mentioned transceiver unit is also used to send N first output values ​​to the second device; the above-mentioned transceiver unit is also used to receive a fifth message from the second device, and the fifth message is used to indicate M first output values, and the M first output values ​​are the output values ​​rejected by the second device among the above-mentioned N first output values.

[0085] In combination with the fourth aspect, in certain implementations of the fourth aspect, before the above-mentioned transceiver unit is used to receive the first message from the first device, the above-mentioned transceiver unit is also used to receive a seventh message from the first device, and the seventh message includes a second analysis requirement for feedback of the first analysis result; when the above-mentioned communication device does not meet the first analysis requirement, the above-mentioned transceiver unit is also used to send an eighth message to the first device, and the eighth message includes a third analysis requirement for feedback of the first analysis result, and the third analysis requirement is lower than the above-mentioned second analysis requirement, and the third analysis requirement is lower than the above-mentioned first analysis requirement.

[0086] In a fifth aspect, a communication device is provided, which includes: a transceiver unit for receiving a second message from a first network function; the communication device also includes: a processing unit for determining a second model based on the second message, the second model being used to assist the first model in obtaining a first analysis result to meet the first analysis requirement, and the first model being deployed on the first network function.

[0087] In combination with the fifth aspect, in certain implementations of the fifth aspect, the second message includes first indication information, and the first indication information is used to indicate an execution subject of the second model.

[0088] In conjunction with the fifth aspect, in certain implementations of the fifth aspect, the second message includes at least one of the following information:

[0089] The inference speed of the second model and the analysis type of the second model.

[0090] In conjunction with the fifth aspect, in certain implementations of the fifth aspect, the second message includes at least one of the following information:

[0091] The inference speed of the first model, the analysis type of the first model, and the acceleration ratio of the accelerated inference.

[0092] In combination with the fifth aspect, in some implementations of the fifth aspect, the above-mentioned transceiver unit is further used to send a third message to the first network function, where the third message is used to indicate the second model.

[0093] In combination with the fifth aspect, in certain implementations of the fifth aspect, the above-mentioned transceiver unit is also used to send a sixth message to the second device, where the sixth message is used to indicate the first network function and the above-mentioned second model, and the second device is the execution entity of the second model.

[0094] In a sixth aspect, a communication device is provided, which further includes: a transceiver unit for receiving a sixth message from a second network function, the sixth message being used to indicate a first network function and a second model, the first network function being deployed with a first model; the communication device further includes: a processing unit for reasoning using the second model, the second model being used to assist the first model in obtaining a first analysis result to meet the first analysis requirement.

[0095] In combination with the sixth aspect, in certain implementations of the sixth aspect, the transceiver unit is further used to obtain input data requirements of the second model and output data requirements of the second model.

[0096] In combination with the sixth aspect, in certain implementations of the sixth aspect, the above-mentioned transceiver unit is further used to send N first output values ​​to the first network function; the transceiver unit is also used to receive a fourth message from the first network function, the fourth message being used to indicate the M first output values, the M first output values ​​being the output values ​​rejected by the first network function among the N first output values, where M is a positive integer less than or equal to N.

[0097] In combination with the sixth aspect, in certain implementations of the sixth aspect, the transceiver unit is further used to obtain input data requirements of the second model and output data requirements of the first model.

[0098] In combination with the sixth aspect, in certain implementations of the sixth aspect, the above-mentioned transceiver unit is further used to receive N first output values ​​from the first network function; the above-mentioned processing unit is further used to use the second model to verify the N first output values, and determine to reject M first output values ​​among the N first output values; the above-mentioned transceiver unit is also used to send a fifth message to the first network function, and the fifth message is used to indicate the M first output values.

[0099] In the seventh aspect, a communication device is provided, comprising a processor, wherein the processor is used to cause the communication device to perform the method described in the first aspect and any possible embodiment of the first aspect, or to cause the communication device to perform the method described in the second aspect and any possible embodiment of the second aspect, or to cause the communication device to perform the method described in the third aspect and any possible embodiment of the third aspect, by executing a computer program or instruction or through a logic circuit.

[0100] In a possible implementation, the communication device further includes a memory for storing the computer program or instruction.

[0101] In a possible implementation, the communication device further includes a communication interface, which is used to input and / or output signals.

[0102] In an eighth aspect, a communication device is provided, comprising a logic circuit and an input / output interface, the input / output interface being used to input and / or output signals, the logic circuit being used to execute the method described in the first aspect and any possibility of the first aspect, or to execute the method described in the second aspect and any possibility of the second aspect, or to execute the method described in the third aspect and any possibility of the third aspect.

[0103] In the ninth aspect, a computer-readable storage medium is provided, on which a computer program or instruction is stored. When the computer program or the instruction is run on a computer, the method described in the first aspect and any possibility of the first aspect is executed, or the method described in the second aspect and any possibility of the second aspect is executed, or the method described in the third aspect and any possibility of the third aspect is executed.

[0104] In the tenth aspect, a computer program product is provided, comprising instructions, which, when executed on a computer, cause the method described in the first aspect and any possible method of the first aspect to be executed, or cause the method described in the second aspect and any possible method of the second aspect to be executed, or cause the method described in the third aspect and any possible method of the third aspect to be executed.

[0105] In the eleventh aspect, a communication system is provided, which includes the above-mentioned first network function and / or the above-mentioned second network function and / or the above-mentioned second device, the first network function is used to execute the method described in the above-mentioned first aspect and any possibility of the first aspect, the second network function is used to execute the method described in the above-mentioned second aspect and any possibility of the second aspect, and the second device is used to execute the method described in the above-mentioned third aspect and any possibility of the third aspect.

[0106] For the relevant explanations and descriptions of the beneficial effects of the fourth to eleventh aspects, please refer to the descriptions of the first to third aspects. BRIEF DESCRIPTION OF THE DRAWINGS

[0107] FIG1 is a schematic diagram of a network architecture 100 to which the technical solution of the present application can be applied.

[0108] FIG2 is a schematic flow chart of a communication method 200 provided in an embodiment of the present application.

[0109] FIG3 is a schematic flowchart of a communication method 300 provided in an embodiment of the present application.

[0110] FIG4 is a schematic flowchart of a communication method 400 provided in an embodiment of the present application.

[0111] FIG5 is a schematic flowchart of a communication method 500 provided in an embodiment of the present application.

[0112] FIG6 is a schematic flowchart of a communication method 600 provided in an embodiment of the present application.

[0113] FIG7 is a schematic flowchart of a communication method 700 provided in an embodiment of the present application.

[0114] FIG8 is a schematic block diagram of a communication device 800 applicable to an embodiment of the present application.

[0115] FIG9 is a schematic block diagram of a communication device 900 applicable to an embodiment of the present application.

[0116] FIG10 is a schematic block diagram of a communication device 1000 applicable to an embodiment of the present application. DETAILED DESCRIPTION

[0117] The technical solutions in the embodiments of the present application will be described below with reference to the accompanying drawings.

[0118] To facilitate understanding of the embodiments of the present application, the following points are explained before introducing the embodiments of the present application.

[0119] In this application, "used to indicate" or "indicates" can include direct indication and indirect indication, or "used to indicate" or "indicates" can indicate explicitly and / or implicitly. For example, when describing that a certain information is used to indicate information I, it can include that the information directly indicates I or indirectly indicates I, but it does not necessarily mean that the information contains I.

[0120] In the embodiments shown below, the first, second, third, fourth and various numbers are only used for the convenience of description and are not intended to limit the scope of the embodiments of the present application. For example, different messages are distinguished.

[0121] In the embodiments of this application, words such as "exemplary," "for example," "illustratively," and "as another example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as an "exemplary" in this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner.

[0122] The terms "include", "comprising", "having" and variations thereof mean "including but not limited to", unless specifically emphasized otherwise.

[0123] In the embodiments of the present application, the descriptions involving A sending a message, information or data to B, and B receiving a message, information or data from A are intended to illustrate to which object the message, information or data is to be sent, and do not limit whether they are sent directly or indirectly via other nodes.

[0124] The technical solutions provided in this application can be applied to various communication systems. For example, the fifth generation (5G) or NR system, LTE system, LTE frequency division duplex (FDD) system, LTE time division duplex (TDD) system, etc. The technical solutions provided in this application can also be applied to non-terrestrial network (NTN) communication systems such as satellite communication systems. The technical solutions provided in this application can also be applied to device to device (D2D) communication, vehicle-to-everything (V2X) communication, machine to machine (M2M) communication, machine type communication (MTC), and Internet of Things (IoT) communication systems or other communication systems. The technical solutions provided in this application can also be applied to future communication systems, such as the sixth generation (6G) mobile communication system.

[0125] As an example, FIG1 shows a schematic diagram of a network architecture 100 .

[0126] As shown in Figure 1, the network architecture takes the 5G system (5GS) as an example. The network architecture may include user equipment (UE), (radio) access network (R)AN) equipment, user plane function (UPF), unified data management (UDM), operations, administration and management (OAM), access and mobility management function (AMF), session management function (SMF), network exposure function (NEF), network repository function (NRF), network data analytics function (NWDAF), application function (AF), policy control function (PCF), unified data repository (UDR), and data collection coordination function (DCCF).

[0127] The following briefly describes the various parts involved in the network architecture in Figure 1.

[0128] 1.UE

[0129] The UE in this application may be any type of mobile terminal, fixed terminal or portable terminal. The UE in this application includes but is not limited to: a user unit, a user station, a mobile station, a mobile station, a remote station, a remote terminal device, a mobile terminal device, a user terminal device, a wireless communication device, a user agent, a user device, a cellular phone, a cordless phone, a session initiation protocol (SIP) phone, a wireless local loop (WLL) station, a personal digital assistant (PDA), a handheld device with wireless communication function, a computing device or other processing device connected to a wireless modem, an in-vehicle device, a wearable device, a terminal device in an Internet of Things (IoT) system, a home appliance, a virtual reality device, a user device in a 2G / 3G / 4G / 5G / 6G network or a user device in a future evolved public land mobile network (PLMN) or a user device in a future vehicle network, etc., and this application is not limited thereto.

[0130] 2. (R)AN equipment

[0131] The (R)AN device of the present application can manage radio resources, provide access services for UE, and further complete the forwarding of control signals and UE data between UE and the core network.

[0132] For example, the (R)AN can be a node in a radio access network. The (R)AN can be a base station, an evolved NodeB (eNodeB), a transmission reception point (TRP), a home base station (e.g., a home evolved NodeB, or home NodeB, HNB), a Wi-Fi access point (AP), a remote radio unit (RRU), a mobile switching center, a next-generation NodeB (gNB) in a 5G mobile communication system, a next-generation base station in a 6G mobile communication system, or a base station in a future mobile communication system. The (R)AN device can also be a module or unit that performs some of the functions of a base station, such as a centralized unit (CU) or a distributed unit (DU). The (R)AN can also be a device that performs base station functions in a D2D communication system, a V2X communication system, an M2M communication system, or an IoT communication system. The (R)AN can also be a network device in an NTN, that is, the (R)AN can be deployed on a high-altitude platform or a satellite. (R)AN can be a macro base station, a micro base station, an indoor station, a relay node, a donor node, etc.

[0133] The embodiments of the present application do not limit the specific technology, device form and name adopted by the (R)AN. For the convenience of description, the (R)AN will be collectively referred to as the access network device below.

[0134] 3. UPF

[0135] The UPF is used for packet routing and forwarding, as well as quality of service (QoS) processing for user plane data. User data can be accessed to the data network (DN) through the UPF.

[0136] 4. DN

[0137] DN is an operator network mainly used to provide data services to terminals, such as the Internet, a third-party service network, or an IP Multimedia Service (IMS) network.

[0138] 5. OAM

[0139] OAM, short for network management, is primarily used for routine network and service analysis, forecasting, planning, and configuration, as well as network and service testing and fault management. OAM interacts with the RAN to obtain information such as radio channel conditions and radio resource utilization on the RAN side.

[0140] 6. AMF

[0141] The main functions of AMF include managing user registration, reachability detection, SMF node selection, access authorization and authentication, mobility management, and mobile state transition management.

[0142] 7. SMF

[0143] SMF is mainly used for session management, UE Internet Protocol (IP) address allocation and management, selection and management of UPF, policy control and charging function interface termination points, and downlink data notification.

[0144] 8. NEF

[0145] NEF is mainly used to securely open up services and capabilities provided by the 3rd Generation Partnership Project (3GPP) network functions to the outside world, and supports secure interaction between 3GPP networks and third-party applications.

[0146] 9. UDM

[0147] Used for unified data management, UE subscription data management, UE identity storage and management, UE access authentication, registration or mobility management, etc.

[0148] 10. NRF

[0149] NRF is mainly responsible for providing network capabilities and opening up events to the outside world, as well as receiving relevant external information.

[0150] 11. PCF

[0151] PCF is a unified policy framework used to guide network behavior and provide policy rule information to network functions (such as AMF, SMF, etc.) or UE.

[0152] 12. UDR

[0153] UDR is mainly responsible for providing storage capabilities for contract data, policy data, and capability exposure-related data.

[0154] 13. AF

[0155] The AF primarily supports the delivery of application-side requirements to the network, such as Quality of Service (QoS) requirements or user status event subscriptions. The AF can be deployed within the operator's network or a third-party AF.

[0156] 14. NWDAF

[0157] NWDAF can have at least one of the following functions:

[0158] Data collection, model training, model feedback, analysis result inference, analysis result feedback, etc. Among them, the data collection function refers to collecting data from network elements, third-party servers, terminal devices or network management systems; the model training function refers to analyzing and training the model based on relevant input data; the model feedback function refers to sending the trained artificial intelligence (AI) model / machine learning (ML) model to the network element that supports the inference function; the analysis result inference function determines the data analysis results based on the trained AI model / ML model and the inference data; the analysis result feedback function can provide data analysis results to network elements, third-party servers, terminal devices or network management systems.

[0159] Based on different functions, NWDAF can be further divided into the analytics logical function (AnLF) that supports analytical reasoning and the model training logical function (MTLF) that supports model training. AnLF can request model information from MTLF and perform reasoning based on the analysis results fed back by MTLF.

[0160] The AnLF and MTLF can each be a separate functional entity, or they can be co-located with other functional entities. For example, the AnLF can be co-located with the AMF or the SMF.

[0161] 15. DCCF

[0162] DCCF is responsible for connecting various network functions (NFs) of data sources and NWDAF. DCCF can collect data from various NFs and dispatch the collected data to NWDAF.

[0163] It should be understood that the network functions included in the communication architecture 100 listed above are merely exemplary and the present application is not limited thereto.

[0164] In the network architecture 100 shown in Figure 1 , the N2 interface is the interface between the RAN device and the AMF, used to transmit radio parameters and non-access stratum (NAS) signaling. The N3 interface is the interface between the RAN device and the UPF, used to transmit user plane data. The N4 interface is the interface between the SMF and the UPF, used to transmit service policies, tunnel identification information for N3 connections, data cache indication information, and downlink data notification messages. The N6 interface is the interface between the DN and the UPF, used to transmit user plane data.

[0165] It should be understood that in the above network architecture 100, different network functions can exchange information through service interfaces. For example, NWDAF can collect data generated by UE on other network functions (such as AMF, SMF, etc.) through service interfaces (such as Namf, Nsmf, etc.) provided by these network functions, and provide data analysis results (Analytics), models, data (data), etc. to other network functions (such as AMF, PCF, etc.) through the Nnwdaf interface.

[0166] It should be understood that the network architecture 100 applied to the embodiment of the present application is only an example of a network architecture from the perspective of a traditional point-to-point architecture and a service-oriented architecture. The network architecture applicable to the embodiment of the present application is not limited thereto. Any network architecture that can realize the functions of each of the above-mentioned network elements is applicable to the embodiment of the present application. In addition, other network architectures applicable to the embodiment of the present application may not include all the network functions shown in the above-mentioned network architecture 100, or other network architectures applicable to the embodiment of the present application may also include network functions other than the network functions shown in the above-mentioned network architecture 100, and this application does not limit this.

[0167] It should be noted that the names of the various network functions and interfaces in this application are merely examples. This application does not exclude the possibility that the network functions may be renamed in the future, or that the functions of the network functions may be merged. As technology evolves, any device or network element capable of implementing the aforementioned network functions falls within the scope of protection of this application. Furthermore, the aforementioned network functions may also be referred to as instances, entities, devices, apparatuses, or modules, and this application does not specifically limit these terms.

[0168] The network consumers served by the NWDAF (e.g., network elements, third-party servers, terminal devices, or network management systems) expect the NWDAF's feedback analysis results to meet given analysis requirements (e.g., time requirements and / or accuracy requirements). If the NWDAF cannot meet the analysis requirements given by the network consumer, the NWDAF will request the network consumer to lower the analysis requirements for the feedback analysis results. This method is only applicable when the network consumer has relaxed analysis requirements for the current analysis. If the network consumer has strict analysis requirements for the current analysis, current technologies cannot meet the network consumer's analysis requirements.

[0169] Based on the above technical problems, the present application provides a communication method 200. When a single local model in the NWDAF cannot meet the analysis requirements given by the network consumer, another model is found to assist the local model in inferring and analyzing the results to meet the analysis requirements given by the network consumer, thereby enhancing the service guarantee capability of the NWDAF in the network.

[0170] Figure 2 is a schematic flow chart of a communication method 200 provided in an embodiment of the present application. In this embodiment, the method is illustrated by taking the first device, the first network function, and the second network function as the execution subjects of the interactive illustration as examples, but the present application does not limit the execution subjects of the interactive illustration. For example, the first device in Figure 2 can also be a chip, a chip system, or a processor that supports the method that can be implemented by the first device, or a logic module or software that can implement all or part of the first device; the first network function can also be a chip, a chip system, or a processor that supports the method that can be implemented by the first network function, or a logic module or software that can implement all or part of the first network function; the second network function can also be a chip, a chip system, or a processor that supports the method that can be implemented by the second network function, or a logic module or software that can implement all or part of the second network function.

[0171] The communication method 200 may include the following steps:

[0172] In step S210, the first device sends a first message to the first network function, where the first message includes a first analysis request for feeding back a first analysis result. Accordingly, the first network function receives the first message from the first device.

[0173] Illustratively, the first analysis requirement may include a time requirement and / or an accuracy requirement.

[0174] Exemplarily, the first device may be the aforementioned network consumer.

[0175] In step S212, when the first model fails to meet the first analysis requirement, the first network function sends a second message to the second network function, requesting the second network function to determine a second model. Accordingly, the second network function receives the second message from the first network function.

[0176] Among them, the above-mentioned first model is a local model deployed on the first network function.

[0177] The second message may include first indication information, where the first indication information is used to indicate an execution subject of the second model. For example, the execution subject of the second model may be the first network function or other entity, which is not limited in this application.

[0178] Exemplarily, the first network function requesting the second network function to determine the second model can be divided into the following five cases:

[0179] The first scenario: the first analysis requirement includes a time requirement, and the first network function determines that the local first model cannot meet the time requirement.

[0180] The first network function requests the second model from the second network function to accelerate the reasoning of the first analysis result. The second message can carry information in the following two ways:

[0181] Method 1: The second message may include at least one of the following information:

[0182] The inference speed of the second model, the analysis type of the second model, etc.

[0183] In the first method, the first network function determines the conditions that the second model needs to meet based on the information of the local model and the acceleration ratio of the accelerated reasoning and feeds back to the second network function. The second network function then searches for a second model that meets the conditions.

[0184] Method 2: The second message may include at least one of the following information:

[0185] The inference speed of the above-mentioned first model, the analysis type of the above-mentioned first model, the acceleration ratio of the accelerated inference, etc.

[0186] The second method mentioned above is that the first network function provides the second network function with information about the local model and the acceleration ratio of accelerated reasoning. The second network function is required to determine the conditions that the second model needs to meet and find a second model that meets the requirements.

[0187] Second scenario: the first analysis requirement includes an accuracy requirement and the first network function determines that the local first model cannot meet the accuracy requirement.

[0188] The first network function requests a second model that meets the accuracy requirement from the second network function to assist the first model in obtaining the first analysis result to meet the accuracy requirement. The more parameters a model has, the more accurate the analysis result inferred by the model. In this case, the second model has more parameters than the first model, or the second model is a large model and the first model is a small model.

[0189] A third scenario: the first analysis requirement includes a time requirement and an accuracy requirement, and the first network function determines that the local first model can meet the time requirement but cannot meet the accuracy requirement.

[0190] The first network function requests the second network function to obtain a second model that meets the accuracy requirement to assist the first model in obtaining the first analysis result to meet the accuracy requirement. In this case, the second model has more parameters than the first model, or the second model is a large model and the first model is a small model.

[0191] A fourth situation: the first analysis requirement includes a time requirement and an accuracy requirement, and the first network function determines that the local first model can meet the accuracy requirement but cannot meet the time requirement.

[0192] The first network function requests the second model from the second network function to accelerate the reasoning of the first analysis result. At this time, the second message can carry information in the above-mentioned method 1 or the above-mentioned method 2.

[0193] A fifth situation: the first analysis requirement includes a time requirement and an accuracy requirement, and the first network function determines that the local first model cannot meet the time requirement and the accuracy requirement.

[0194] The first network function requests the second network function to use a second model that meets the accuracy requirement to accelerate the inference of the first analysis result. In this case, the second model has more parameters than the first model, or the second model is a large model and the first model is a small model.

[0195] Step S214: The second network function determines a second model based on the second message, where the second model is used to assist the first model in obtaining the first analysis result to meet the first analysis requirement.

[0196] Exemplarily, the first network function may be NWDAF supporting AnLF, and the second network function may be NEDAF supporting MTLF.

[0197] Through the above communication method 200, when a single local model in the NWDAF cannot meet the analysis requirement of a network consumer requesting to obtain analysis results, another model is found to assist the local model in obtaining analysis results to meet the analysis requirement, thereby enhancing the service guarantee capability of the NWDAF in the network.

[0198] Specifically, the following embodiment takes the first network function as AnLF1 and the second network function as MTLF as an example to describe in detail the technical solution provided by the embodiment of the present application, as shown in the schematic flowchart of the communication method 300 in FIG3 .

[0199] The communication method 300 may include the following steps:

[0200] In step S310, the first device sends a first message to AnLF1. In response, AnLF1 receives the first message from the first device. The first message includes a first analysis request for feeding back a first analysis result.

[0201] The communication method 300 is described in detail using the example of satisfying the time requirement included in the first analysis requirement. This corresponds to the first or fourth case in which the first network function requests the second network function to determine the second model. The first analysis requirement may include a first time requirement. For example, the first time requirement is to feedback the first analysis result before a first time, or the first time requirement is to feedback the first analysis result within a first time period.

[0202] Exemplarily, the first message may be a subscription message (Nnwdaf_AnalyticsSubscription_Subscribe or Nnwdaf-_AnalyticsInfo_Request), which further includes at least one of the following information:

[0203] Analysis ID, accuracy requirements for the first analysis result, etc.

[0204] Step S312: AnLF1 collects input data of the model according to the first message.

[0205] For example, AnLF1 may collect input data of the model from corresponding network functions, devices, nodes, etc. according to the content included in the first message. The input data of the model may also be referred to as inference data of the model.

[0206] Step S314 : AnLF1 determines, based on the first message, the collected model input data, the time for collecting the model input data, and the inference speed of the first model, that the first time requirement cannot be met by using the first model.

[0207] For example, the first time requirement is to feedback the first analysis result within 500 seconds (s). AnLF1 uses 100s to collect input data for 10,000 models, and the remaining time is 400s. The inference speed of the first model is 10 / second (s). It takes 1000s for AnLF1 to complete the inference of 10,000 inference data using the first model, which far exceeds the remaining 400s. Therefore, AnLF1 can determine that the first time requirement cannot be met using the first model. At this time, AnLF1 triggers multi-model collaborative accelerated inference.

[0208] Step S316 , AnLF1 determines the inference speed of the second model based on the first message, the collected model input data, the time of collecting the model input, and the inference speed of the first model.

[0209] For example, the first time requirement is to feedback the first analysis result within 500 seconds (s). AnLF1 used 100s to collect input data for 10,000 models, and the remaining time is 400s. The current inference speed of the first model is 10 / second (s). If the first model can complete the inference of 10,000 inference data in the remaining 400s, the inference speed of the first model needs to be accelerated to 25 / s. Therefore, the acceleration ratio of the accelerated inference required is 25 / 10=2.5. For example, in order to achieve the effect of 2.5 times accelerated inference, the inference speed of the second model is approximately 100 / s.

[0210] In step S318, AnLF1 sends a second message to MTLF. Correspondingly, MTLF receives the second message from AnLF1.

[0211] Specifically, the second message includes at least one of the following information:

[0212] The inference speed of the second model, the analysis type of the second model, the first indication information, the analysis identifier, the number of second models, the filtering information of the second model, etc.

[0213] Illustratively, when the number of second models included in the second message is greater than 1, AnLF1 may request MTLF to determine multiple second models through the second message, which is not limited in this application.

[0214] For example, the analysis type of the second model included in the second message may include parameters such as the model structure of the second model and the word segmentation type of the second model. The word segmentation type of the second model indicates the word segmentation algorithm used by the second model. The analysis type of the second model should be consistent with the analysis type of the first model.

[0215] Exemplarily, the first indication information included in the second message is used to indicate the execution subject of the second model. Exemplarily, if AnLF1 has sufficient local resources to execute multiple models, AnLF1 can indicate through the first indication information that the execution subject of the second model is AnLF1, so MTLF needs to feedback the second model to AnLF1 after determining the second model; if AnLF1 does not have sufficient local resources to execute multiple models, AnLF1 can indicate through the first indication information that the execution subject of the second model is the second device below, so MTLF needs to feedback the second model to the second device after determining the second model. Optionally, AnLF1 may not carry the above-mentioned first indication information in the second message. AnLF1 not carrying the first indication information in the second message may be a default that MTLF feeds back the second model to AnLF1 after determining the second model, or a default that MTLF does not feed back the second model to AnLF1 after determining the second model. This application does not limit this.

[0216] Exemplarily, the filtering information of the second model included in the second message is used to indicate the conditions met for training the second model. When the MTLF does not find a second model that meets the requirements, the MTLF can train a second model that meets the requirements based on the filtering information of the second model.

[0217] Step S320: The MTLF searches for a second model that meets the requirements based on the second message.

[0218] After MTLF finds the second model that meets the requirements, it needs to feed back the second model to the execution entity of the second model.

[0219] 1. If the first indication information indicates that the execution subject of the second model is AnLF1, or if the default MTLF feeds back the second model to AnLF1 after determining the second model, the communication method 300 further includes the following steps S322 to S324:

[0220] In step S322, MTLF sends a third message to AnLF1, where the third message is used to indicate the second model. Accordingly, AnLF1 receives the third message from MTLF.

[0221] Exemplarily, the third message includes at least one of the following information:

[0222] Analysis identifier, identifier of the second model (Model ID), file address of the second model, etc.

[0223] Exemplarily, the third message may be a Nnwdaf_MLModelProvision_Subscribe message.

[0224] In step S324 , AnLF1 performs inference based on the local first model and the local second model to obtain the first analysis result.

[0225] 2. If the first indication information indicates that the execution subject of the second model is the second device, or if it is defaulted that the MTLF does not feed back the second model to the AnLF1 after determining the second model, the communication method 300 further includes the following steps S326 to S346:

[0226] Optionally, in step S326 , the MTLF determines that the execution entity of the second model is AnLF2 (an example of the second device).

[0227] It should be noted that when the first indication information indicates that the execution subject of the second model is the second device, the MTLF does not need to determine the execution subject of the second model. When the second message does not include the first indication information and it is assumed that the MTLF does not feedback the second model to AnLF1 after determining the second model, the MTLF needs to determine the execution subject of the second model.

[0228] For example, the MTLF may determine AnLF2 from the connected multiple AnLFs as the execution subject of the second model. For example, the current computing resources, algorithm resources, etc. of the AnLF2 may meet the resource requirements for executing the second model.

[0229] Optionally, in step S328, MTLF sends a message #1 to AnLF1, where the message #1 is used to instruct AnLF2. Correspondingly, AnLF1 receives the message #1 from MTLF.

[0230] It should be noted that when the first indication information indicates that the execution subject of the second model is the second device, the MTLF does not need to send message #1 to AnLF1. When the second message does not include the first indication information and it is assumed that the MTLF does not feedback the second model to AnLF1 after determining the second model, the MTLF needs to determine the execution subject of the second model and send message #1 to AnLF1.

[0231] Exemplarily, the message #1 may include the identifier of AnLF2, the address of AnLF2, and the like.

[0232] In step S330, the MTLF sends a sixth message to AnLF2, where the sixth message is used to indicate the second model and AnLF1. Accordingly, AnLF2 receives the sixth message from the MTLF.

[0233] Exemplarily, the sixth message may include an analysis identifier, an identifier of the second model (Model ID), a file address of the second model, an identifier of AnLF1, an address of AnLF1, and the like.

[0234] The above steps S326 to S330 can establish a connection between the execution subject of the first model and the execution subject of the second model.

[0235] After the execution entity of the first model and the execution entity of the second model establish a connection, they need to negotiate the input data requirements and output data requirements of the models. There are two specific ways:

[0236] Method 1: Assume that the first model is a large model and the second model is a small model.

[0237] In step S332 , AnLF1 and AnLF2 obtain the input data requirements of the model and the output data requirements of the second model.

[0238] For example, the input data requirements of the model may include the input data of the first model and the input data of the second model. AnLF1 may indicate the collected input data of the model to AnLF2. For example, AnLF1 may send a tag of the collected input data set of the model or the data set itself to AnLF2.

[0239] Illustratively, the output data requirement of the second model may include the number N of output values ​​(tokens) of each inference of the second model during the inference process, where N is a positive integer greater than or equal to 1. The output data requirement of the second model may be generated by AnLF1 and indicated to AnLF2, or the output data requirement of the second model may be generated by AnLF2 and indicated to AnLF1, which is not limited in this application.

[0240] In step S334, AnLF2 performs an inference using the second model and generates N first output values ​​(tokens), and sends the N first output values ​​to AnLF1. Accordingly, AnLF1 receives the N first output values ​​from AnLF2.

[0241] Step S336 : AnLF1 verifies the N first output values ​​using the first model, and determines to reject M first output values ​​among the N first output values.

[0242] Optionally, the step S336 may also be: AnLF1 verifies the N first output values ​​using the first model, and determines to reject a first proportion of the N first output values. The first proportion may also be called a rejection proportion.

[0243] In step S338, AnLF1 sends a fourth message to AnLF2, where the fourth message is used to indicate the M first output values. Correspondingly, AnLF2 receives the fourth message from AnLF1.

[0244] Optionally, the fourth message may include the first ratio or the number of the M first output values, etc., which is not limited in this application.

[0245] For example, the N first output values ​​are 7 tokens. After verifying the 7 tokens, AnLF1 determines to accept the first 5 tokens and reject the last 2 tokens. Then, AnLF1 instructs AnLF2 to reject the last 2 tokens of the 7 tokens, and AnLF2 can re-infer the last 2 tokens accordingly.

[0246] Or, for another example, the N first output values ​​are 7 tokens. After verifying the 7 tokens, AnLF1 determines to accept the first 5 tokens and reject the last 2 tokens. Then, AnLF1 instructs AnLF2 to reject 2 / 7 of the 7 tokens, and AnLF2 can re-infer the 2 / 7 tokens based on this.

[0247] The above steps S334 to S338 are used to perform multiple inferences until the first analysis result is inferred.

[0248] Method 2: Assume that the first model is a small model and the second model is a large model.

[0249] In step S340 , AnLF1 and AnLF2 obtain the input data requirements of the model and the output data requirements of the first model.

[0250] For example, the input data requirements of the model may include the input data of the first model and the input data of the second model. AnLF1 may indicate the collected input data of the model to AnLF2. For example, AnLF1 may send a tag of the collected input data set of the model or the data set itself to AnLF2.

[0251] Illustratively, the output data requirement of the first model may include the number N of output values ​​(tokens) of each inference of the first model during the inference process, where N is a positive integer greater than or equal to 1. The output data requirement of the first model may be generated by AnLF1 and indicated to AnLF2, or the output data requirement of the first model may be generated by AnLF2 and indicated to AnLF1, and this application is not limited to this.

[0252] In step S342, AnLF1 performs an inference using the first model and generates N first output values ​​(tokens), and sends the N first output values ​​to AnLF2. Accordingly, AnLF2 receives the N first output values ​​from AnLF1.

[0253] In step S344 , AnLF2 verifies the N first output values ​​using the second model, and determines to reject M first output values ​​among the N first output values.

[0254] Optionally, the step S344 may also be: AnLF2 verifies the N first output values ​​using the second model, and determines to reject a first proportion of the N first output values. The first proportion may also be called a rejection proportion.

[0255] In step S346, AnLF2 sends a fifth message to AnLF1, where the fifth message is used to indicate the M first output values. Accordingly, AnLF1 receives the fifth message from AnLF2.

[0256] Optionally, the fifth message may include the first ratio or the number of the M first output values, etc., which is not limited in this application.

[0257] For example, the N first output values ​​are 7 tokens. After verifying the 7 tokens, AnLF2 determines to accept the first 5 tokens and reject the last 2 tokens. Then, AnLF2 instructs AnLF1 to reject the last 2 tokens of the 7 tokens, so that AnLF1 can re-infer the last 2 tokens.

[0258] Or, for another example, the N first output values ​​are 7 tokens. After verifying the 7 tokens, AnLF2 determines to accept the first 5 tokens and reject the last 2 tokens. Then, AnLF2 instructs AnLF1 to reject 2 / 7 of the 7 tokens, and AnLF1 can re-infer the 2 / 7 tokens based on this.

[0259] The above steps S340 to S346 are used to perform multiple inferences until the first analysis result is inferred.

[0260] Step S348: AnLF1 feeds back the first analysis result to the first device.

[0261] Specifically, the AnLF feeding back the first analysis result to the first device can meet the above-mentioned first time requirement.

[0262] The technical solution of the above method 300 describes in detail how to find another model to assist the local model in accelerating the process of obtaining analysis results, promotes the collaboration between different network functions of NWDAF, and thus enhances the service assurance capability of NWDAF in the network.

[0263] In the above-mentioned communication method 300, AnLF1 determines the relevant parameters of the second model, and MTLF searches for a second model that meets the requirements. The present application may also provide a communication method 400, in which MTLF determines the relevant parameters of the second model and searches for a second model that meets the requirements, as shown in FIG4 , which is a schematic flow chart of the communication method 400.

[0264] The communication method 400 may include the following steps:

[0265] In step S410, the first device sends a first message to AnLF1. In response, AnLF1 receives the first message from the first device. The first message includes a first analysis request for feeding back a first analysis result.

[0266] Communication method 400 is described in detail using the example of satisfying the time requirement included in the first analysis requirement. This corresponds to the first or fourth case in which the first network function requests the second network function to determine the second model. The first analysis requirement may include a first time requirement. Exemplarily, the first time requirement is to feedback the first analysis result before a first time, or the first time requirement is to feedback the first analysis result within a first time period.

[0267] Exemplarily, the first message may be a subscription message (Nnwdaf_AnalyticsSubscription_Subscribe or Nnwdaf-_AnalyticsInfo_Request), which further includes at least one of the following information:

[0268] Analysis ID, accuracy requirements for the first analysis result, etc.

[0269] Step S412: AnLF1 collects input data of the model according to the first message.

[0270] For example, AnLF1 may collect input data of the model from corresponding network functions, devices, nodes, etc. according to the content included in the first message. The input data of the model may also be referred to as inference data of the model.

[0271] Step S414 : AnLF1 determines, based on the first message, the collected model input data, the time for collecting the model input data, and the inference speed of the first model, that the first time requirement cannot be met by using the first model.

[0272] For example, the first time requirement is to feedback the first analysis result within 500 seconds (s). AnLF1 uses 100s to collect input data for 10,000 models, and the remaining time is 400s. The inference speed of the first model is 10 / second (s). It takes 1000s for AnLF1 to complete the inference of 10,000 inference data using the first model, which far exceeds the remaining 400s. Therefore, AnLF1 can determine that the first time requirement cannot be met using the first model. At this time, AnLF1 triggers multi-model collaborative accelerated inference.

[0273] In step S416, AnLF1 sends a second message to MTLF, where the second message is used to request MTLF to determine a second model. Correspondingly, MTLF receives the second message from AnLF1.

[0274] Specifically, the second message includes at least one of the following information:

[0275] The reasoning speed of the first model, the analysis type of the first model, the acceleration ratio of the accelerated reasoning, the first indication information, the analysis identifier, the number of the second models, the filtering information of the second model, etc.

[0276] Exemplarily, the analysis type of the first model included in the second message may include parameters such as the model structure of the first model and the word segmentation type of the first model, where the word segmentation type of the first model indicates the word segmentation algorithm used by the first model.

[0277] For example, the acceleration ratio of the accelerated reasoning included in the second message can be calculated as follows: Assume that the first time requirement is to feedback the first analysis result within 500 seconds (s). AnLF1 uses 100s to collect input data for 10,000 models, and the remaining time is 400s. The current reasoning speed of the first model is 10 / second (s). If the first model can complete the reasoning of 10,000 reasoning data in the remaining 400s, the reasoning speed of the first model needs to be accelerated to 25 / s. Therefore, the required acceleration ratio of the accelerated reasoning is 25 / 10=2.5.

[0278] Illustratively, when the number of second models included in the second message is greater than 1, AnLF1 may request MTLF to determine multiple second models through the second message, which is not limited in this application.

[0279] Exemplarily, the first indication information included in the second message is used to indicate the execution subject of the second model. Exemplarily, if AnLF1 has sufficient local resources to execute multiple models, AnLF1 can indicate through the first indication information that the execution subject of the second model is AnLF1, so MTLF needs to feedback the second model to AnLF1 after determining the second model; if AnLF1 does not have sufficient local resources to execute multiple models, AnLF1 can indicate through the first indication information that the execution subject of the second model is the second device below, so MTLF needs to feedback the second model to the second device after determining the second model. Optionally, AnLF1 may not carry the above-mentioned first indication information in the second message. AnLF1 not carrying the first indication information in the second message may be a default that MTLF feeds back the second model to AnLF1 after determining the second model, or a default that MTLF does not feed back the second model to AnLF1 after determining the second model. This application does not limit this.

[0280] Exemplarily, the filtering information of the second model included in the second message is used to indicate the conditions met for training the second model. When the MTLF does not find a second model that meets the requirements, the MTLF can train a second model that meets the requirements based on the filtering information of the second model.

[0281] Step S418: The MTLF determines the inference speed of the second model, the analysis type of the second model, etc. according to the second message.

[0282] Assuming that the current inference speed of the first model included in the second message is 10 / s and the acceleration factor of the accelerated inference is 2.5, it can be determined that the inference speed of the second model is approximately 100 / s.

[0283] The analysis type of the second model should be consistent with the analysis type of the first model included in the second message.

[0284] In step S420 , the MTLF searches for a second model that meets the requirements based on the determined inference speed of the second model, the analysis type of the second model, and the like.

[0285] After MTLF finds the second model that meets the requirements, it needs to feed back the second model to the execution entity of the second model.

[0286] Steps S422 to S448 may refer to the above-mentioned steps S322 to S348 and will not be repeated here.

[0287] The technical solution of the above method 400 describes in detail how to find another model to assist the local model in accelerating the process of obtaining analysis results, promotes the collaboration between different network functions of NWDAF, and thus enhances the service assurance capability of NWDAF in the network.

[0288] In the above communication method 300 and the above communication method 400, the execution subject of the first model first collects data, then pre-judges whether it is necessary to trigger multi-model collaborative accelerated reasoning, then determines the execution subject of the second model, and finally the execution subject of the first model instructs the collected data to the execution subject of the second model or the execution subject of the second model collects data by itself. This process may require two data collection processes in succession. If the number of models for collaborative accelerated reasoning is greater, this process requires more data collection processes. Based on this, the present application can also provide a communication method 500, which can reduce the number of times data is collected from the data source during the above multi-model collaborative accelerated reasoning process, thereby reducing the time for data collection.

[0289] As shown in FIG5 , which is a schematic flow chart of a communication method 500, the communication method 500 may include the following steps:

[0290] In step S510, the first device sends a first message to AnLF1. In response, AnLF1 receives the first message from the first device. The first message includes a first analysis request for feeding back a first analysis result.

[0291] Communication method 500 is described in detail using the example of satisfying the time requirement included in the first analysis requirement. This corresponds to the first or fourth case in which the first network function requests the second network function to determine the second model. The first analysis requirement may include a first time requirement. Exemplarily, the first time requirement is to feedback the first analysis result before a first time, or the first time requirement is to feedback the first analysis result within a first time period.

[0292] Exemplarily, the first message may be a subscription message (Nnwdaf_AnalyticsSubscription_Subscribe or Nnwdaf-_AnalyticsInfo_Request), which further includes at least one of the following information:

[0293] Analysis ID, accuracy requirements for the first analysis result, etc.

[0294] In step S512, AnLF1 estimates the amount of input data required to be collected for the model, the time required to collect the input data for the model, etc., and determines that the first time requirement cannot be met by using the first model based on the first message, the estimated amount of input data for the model, the estimated time required to collect the input data for the model, and the inference speed of the first model.

[0295] For example, the first time requirement is to feedback the first analysis result within 500 seconds (s). AnLF1 estimates that it needs to collect input data for 10,000 models, and estimates that the time to collect the input data for these 10,000 models is about 100s, and the remaining time is 400s. The inference speed of the first model is 10 / s. It takes 1000s for AnLF1 to complete the inference of 10,000 inference data using the first model, which far exceeds the remaining 400s. Therefore, AnLF1 can determine that the first time requirement cannot be met using the first model. At this time, AnLF1 triggers multi-model collaborative accelerated inference.

[0296] In step S514, AnLF1 determines the inference speed of the second model based on the first message, the estimated input data volume of the model, the estimated time for collecting the input data of the model, and the inference speed of the first model.

[0297] For example, the first time requirement is to feedback the first analysis result within 500 seconds (s). AnLF1 needs to use 100s to collect the input data of 10,000 models, and the remaining reasoning time is 400s. The current reasoning speed of the first model is 10 / s. If the first model can complete the reasoning of 10,000 reasoning data in the remaining 400s, the reasoning speed of the first model needs to be accelerated to 25 / s. Therefore, the acceleration ratio of the accelerated reasoning required is 25 / 10=2.5. For example, in order to achieve the effect of 2.5 times accelerated reasoning, the reasoning speed of the second model is approximately 100 / s.

[0298] Step S516: AnLF1 sends a second message to MTLF. Correspondingly, MTLF receives the second message from AnLF1.

[0299] Specifically, the second message includes at least one of the following information:

[0300] The inference speed of the second model, the analysis type of the second model, the first indication information, the analysis identifier, the number of second models, the filtering information of the second model, etc.

[0301] Illustratively, when the number of second models included in the second message is greater than 1, AnLF1 may request MTLF to determine multiple second models through the second message, which is not limited in this application.

[0302] For example, the analysis type of the second model included in the second message may include parameters such as the model structure of the second model and the word segmentation type of the second model. The word segmentation type of the second model indicates the word segmentation algorithm used by the second model. The analysis type of the second model should be consistent with the analysis type of the first model.

[0303] Exemplarily, the first indication information included in the second message is used to indicate the execution subject of the second model. Exemplarily, if AnLF1 has sufficient local resources to execute multiple models, AnLF1 can indicate through the first indication information that the execution subject of the second model is AnLF1, so MTLF needs to feedback the second model to AnLF1 after determining the second model; if AnLF1 does not have sufficient local resources to execute multiple models, AnLF1 can indicate through the first indication information that the execution subject of the second model is the second device below, so MTLF needs to feedback the second model to the second device after determining the second model. Optionally, AnLF1 may not carry the above-mentioned first indication information in the second message. AnLF1 not carrying the first indication information in the second message may be a default that MTLF feeds back the second model to AnLF1 after determining the second model, or a default that MTLF does not feed back the second model to AnLF1 after determining the second model. This application does not limit this.

[0304] Exemplarily, the filtering information of the second model included in the second message is used to indicate the conditions met for training the second model. When the MTLF does not find a second model that meets the requirements, the MTLF can train a second model that meets the requirements based on the filtering information of the second model.

[0305] Step S320: The MTLF searches for a second model that meets the requirements based on the second message.

[0306] After MTLF finds the second model that meets the requirements, it needs to feed back the second model to the execution entity of the second model.

[0307] 1. If the first indication information indicates that the execution subject of the second model is AnLF1, or if the default MTLF feeds back the second model to AnLF1 after determining the second model, the communication method 500 further includes the following steps S520 to S524:

[0308] In step S322, MTLF sends a third message to AnLF1, where the third message is used to indicate the second model. Accordingly, AnLF1 receives the third message from MTLF.

[0309] Exemplarily, the third message includes at least one of the following information:

[0310] Analysis identifier, identifier of the second model (Model ID), file address of the second model, etc.

[0311] Exemplarily, the third message may be a Nnwdaf_MLModelProvision_Subscribe message.

[0312] Step S522: AnLF1 collects input data of the model according to the first message.

[0313] For example, AnLF1 may collect input data of the model from corresponding network functions, devices, nodes, etc. according to the content included in the first message. The input data of the model may also be referred to as inference data of the model.

[0314] In step S524 , AnLF1 performs inference based on the local first model and the local second model to obtain the first analysis result.

[0315] 2. If the first indication information indicates that the execution subject of the second model is the second device, or if it is defaulted that the MTLF does not feed back the second model to the AnLF1 after determining the second model, the communication method 500 further includes the following steps S526 to S548:

[0316] Optionally, in step S526 , the MTLF determines that the execution entity of the second model is AnLF2 (an example of the second device).

[0317] It should be noted that when the first indication information indicates that the execution subject of the second model is the second device, the MTLF does not need to determine the execution subject of the second model. When the second message does not include the first indication information and it is assumed that the MTLF does not feedback the second model to AnLF1 after determining the second model, the MTLF needs to determine the execution subject of the second model.

[0318] For example, the MTLF may determine AnLF2 from the connected multiple AnLFs as the execution subject of the second model. For example, the current computing resources, algorithm resources, etc. of the AnLF2 may meet the resource requirements for executing the second model.

[0319] Optionally, in step S528, MTLF sends a message #1 to AnLF1, where the message #1 is used to instruct AnLF2. Correspondingly, AnLF1 receives the message #1 from MTLF.

[0320] It should be noted that when the first indication information indicates that the execution subject of the second model is the second device, the MTLF does not need to send message #1 to AnLF1. When the second message does not include the first indication information and it is assumed that the MTLF does not feedback the second model to AnLF1 after determining the second model, the MTLF needs to determine the execution subject of the second model and send message #1 to AnLF1.

[0321] Exemplarily, the message #1 may include the identifier of AnLF2, the address of AnLF2, and the like.

[0322] In step S530, the MTLF sends a sixth message to AnLF2, where the sixth message is used to indicate the second model and AnLF1. Accordingly, AnLF2 receives the sixth message from the MTLF.

[0323] Exemplarily, the sixth message may include an analysis identifier, an identifier of the second model (Model ID), a file address of the second model, an identifier of AnLF1, an address of AnLF1, and the like.

[0324] The above steps S526 to S530 can establish a connection between the execution subject of the first model and the execution subject of the second model.

[0325] After the execution subject of the first model and the execution subject of the second model establish a connection, they need to negotiate the requirements for the input data and output data of the models.

[0326] Specifically, the input data requirements of the model negotiated by the execution subject of the first model and the execution subject of the second model can specifically be the input data of the negotiated model, including step S532, DCCF can collect the input data of the model from the corresponding network functions, devices, nodes, etc. according to the content included in the above-mentioned first message, and dispatch the collected input data of the model to the execution subject AnLF1 of the first model and the execution subject AnLF2 of the second model, so that AnLF1 obtains the input data of the first model and AnLF2 obtains the input data of the second model.

[0327] Exemplarily, the DCCF may indicate the collected input data of the model to AnLF1 and AnLF2. For example, the DCCF may send tags of the collected input data sets of the model or the data sets themselves to AnLF1 and AnLF2.

[0328] Specifically, the execution subject of the first model and the execution subject of the second model negotiate the output data requirements of the model in the following two ways:

[0329] Method 1: Assume that the first model is a large model and the second model is a small model.

[0330] In step S534 , AnLF1 and AnLF2 obtain the output data requirements of the second model.

[0331] Illustratively, the output data requirement of the second model may include the number N of output values ​​(tokens) of each inference of the second model during the inference process, where N is a positive integer greater than or equal to 1. The output data requirement of the second model may be generated by AnLF1 and indicated to AnLF2, or the output data requirement of the second model may be generated by AnLF2 and indicated to AnLF1, which is not limited in this application.

[0332] In step S536, AnLF2 performs an inference using the second model and generates N first output values ​​(tokens), and sends the N first output values ​​to AnLF1. Accordingly, AnLF1 receives the N first output values ​​from AnLF2.

[0333] Step S538 : AnLF1 verifies the N first output values ​​using the first model, and determines to reject M first output values ​​among the N first output values.

[0334] Optionally, the step S538 may also be: AnLF1 verifies the N first output values ​​using the first model, and determines to reject a first proportion of the N first output values. The first proportion may also be called a rejection proportion.

[0335] In step S540, AnLF1 sends a fourth message to AnLF2, where the fourth message is used to indicate the M first output values. Correspondingly, AnLF2 receives the fourth message from AnLF1.

[0336] Optionally, the fourth message may include the first ratio or the number of the M first output values, etc., which is not limited in this application.

[0337] For example, the N first output values ​​are 7 tokens. After verifying the 7 tokens, AnLF1 determines to accept the first 5 tokens and reject the last 2 tokens. Then, AnLF1 instructs AnLF2 to reject the last 2 tokens of the 7 tokens, and AnLF2 can re-infer the last 2 tokens accordingly.

[0338] Or, for another example, the N first output values ​​are 7 tokens. After verifying the 7 tokens, AnLF1 determines to accept the first 5 tokens and reject the last 2 tokens. Then, AnLF1 instructs AnLF2 to reject 2 / 7 of the 7 tokens, and AnLF2 can re-infer the 2 / 7 tokens based on this.

[0339] The above steps S536 to S540 are used to perform multiple inferences until the first analysis result is inferred.

[0340] Method 2: Assume that the first model is a small model and the second model is a large model.

[0341] In step S542 , AnLF1 and AnLF2 obtain the output data requirements of the first model.

[0342] Illustratively, the output data requirement of the first model may include the number N of output values ​​(tokens) of each inference of the first model during the inference process, where N is a positive integer greater than or equal to 1. The output data requirement of the first model may be generated by AnLF1 and indicated to AnLF2, or the output data requirement of the first model may be generated by AnLF2 and indicated to AnLF1, and this application is not limited to this.

[0343] In step S544, AnLF1 performs an inference using the first model and generates N first output values ​​(tokens), and sends the N first output values ​​to AnLF2. Accordingly, AnLF2 receives the N first output values ​​from AnLF1.

[0344] Step S546 : AnLF2 verifies the N first output values ​​using the second model, and determines to reject M first output values ​​among the N first output values.

[0345] Optionally, the step S546 may also be: AnLF2 verifies the N first output values ​​using the second model, and determines to reject a first proportion of the N first output values. The first proportion may also be called a rejection proportion.

[0346] In step S548, AnLF2 sends a fifth message to AnLF1, where the fifth message is used to indicate the M first output values. Accordingly, AnLF1 receives the fifth message from AnLF2.

[0347] Optionally, the fifth message may include the first ratio or the number of the M first output values, etc., which is not limited in this application.

[0348] For example, the N first output values ​​are 7 tokens. After verifying the 7 tokens, AnLF2 determines to accept the first 5 tokens and reject the last 2 tokens. Then, AnLF2 instructs AnLF1 to reject the last 2 tokens of the 7 tokens, so that AnLF1 can re-infer the last 2 tokens.

[0349] Or, for another example, the N first output values ​​are 7 tokens. After verifying the 7 tokens, AnLF2 determines to accept the first 5 tokens and reject the last 2 tokens. Then, AnLF2 instructs AnLF1 to reject 2 / 7 of the 7 tokens, and AnLF1 can re-infer the 2 / 7 tokens based on this.

[0350] The above steps S544 to S548 are used to perform multiple inferences until the first analysis result is inferred.

[0351] Step S550 : AnLF1 feeds back the first analysis result to the first device.

[0352] Specifically, the AnLF feeding back the first analysis result to the first device can meet the above-mentioned first time requirement.

[0353] The technical solution of method 500 describes in detail how to find another model to assist the local model in accelerating the acquisition of analysis results, promoting collaboration between different network functions of the NWDAF, thereby enhancing the service assurance capabilities of the NWDAF in the network. Furthermore, communication method 500 can use the DCCF to collect data from data sources and dispatch it to the corresponding execution entities of multiple models. This can reduce the number of data collection attempts from data sources during the aforementioned multi-model collaborative accelerated inference process, thereby reducing data collection time.

[0354] In the above-mentioned communication method 500, AnLF1 determines the relevant parameters of the second model, and MTLF searches for a second model that meets the requirements. The present application may also provide a communication method 600, in which MTLF determines the relevant parameters of the second model and searches for a second model that meets the requirements, as shown in FIG6 , which is a schematic flow chart of the communication method 600.

[0355] The communication method 600 may include the following steps:

[0356] In step S610, the first device sends a first message to AnLF1. In response, AnLF1 receives the first message from the first device. The first message includes a first analysis request for feeding back a first analysis result.

[0357] Communication method 600 is described in detail using the example of satisfying the time requirement included in the first analysis requirement. This corresponds to the first or fourth case in which the first network function requests the second network function to determine the second model. The first analysis requirement may include a first time requirement. For example, the first time requirement is to feedback the first analysis result before a first time, or the first time requirement is to feedback the first analysis result within a first time period.

[0358] Exemplarily, the first message may be a subscription message (Nnwdaf_AnalyticsSubscription_Subscribe or Nnwdaf-_AnalyticsInfo_Request), which further includes at least one of the following information:

[0359] Analysis ID, accuracy requirements for the first analysis result, etc.

[0360] In step S612, AnLF1 estimates the amount of input data required to be collected for the model, the time required to collect the input data for the model, etc., and determines that the first time requirement cannot be met by using the first model based on the first message, the estimated amount of input data for the model, the estimated time required to collect the input data for the model, and the inference speed of the first model.

[0361] For example, the first time requirement is to feedback the first analysis result within 500 seconds (s). AnLF1 estimates that it needs to collect input data for 10,000 models, and estimates that the time to collect the input data for these 10,000 models is about 100s, and the remaining time is 400s. The inference speed of the first model is 10 / s. It takes 1000s for AnLF1 to complete the inference of 10,000 inference data using the first model, which far exceeds the remaining 400s. Therefore, AnLF1 can determine that the first time requirement cannot be met using the first model. At this time, AnLF1 triggers multi-model collaborative accelerated inference.

[0362] In step S614, AnLF1 sends a second message to MTLF, where the second message is used to request MTLF to determine a second model. Correspondingly, MTLF receives the second message from AnLF1.

[0363] Specifically, the second message includes at least one of the following information:

[0364] The reasoning speed of the first model, the analysis type of the first model, the acceleration ratio of the accelerated reasoning, the first indication information, the analysis identifier, the number of the second models, the filtering information of the second model, etc.

[0365] Exemplarily, the analysis type of the first model included in the second message may include parameters such as the model structure of the first model and the word segmentation type of the first model, where the word segmentation type of the first model indicates the word segmentation algorithm used by the first model.

[0366] Exemplarily, the acceleration ratio of the accelerated reasoning included in the second message can be calculated as follows: Assume that the first time requirement is to feedback the first analysis result within 500 seconds (s). AnLF1 estimates that it needs to collect input data for 10,000 models, and estimates that the time to collect the input data of these 10,000 models is approximately 100s, and the remaining time is 400s. The current reasoning speed of the first model is 10 / second (s). If the first model can complete the reasoning of 10,000 reasoning data in the remaining 400s, the reasoning speed of the first model needs to be accelerated to 25 / s. Therefore, the required acceleration ratio of accelerated reasoning is 25 / 10=2.5.

[0367] Illustratively, when the number of second models included in the second message is greater than 1, AnLF1 may request MTLF to determine multiple second models through the second message, which is not limited in this application.

[0368] Exemplarily, the first indication information included in the second message is used to indicate the execution subject of the second model. Exemplarily, if AnLF1 has sufficient local resources to execute multiple models, AnLF1 can indicate through the first indication information that the execution subject of the second model is AnLF1, so MTLF needs to feedback the second model to AnLF1 after determining the second model; if AnLF1 does not have sufficient local resources to execute multiple models, AnLF1 can indicate through the first indication information that the execution subject of the second model is the second device below, so MTLF needs to feedback the second model to the second device after determining the second model. Optionally, AnLF1 may not carry the above-mentioned first indication information in the second message. AnLF1 not carrying the first indication information in the second message may be a default that MTLF feeds back the second model to AnLF1 after determining the second model, or a default that MTLF does not feed back the second model to AnLF1 after determining the second model. This application does not limit this.

[0369] Exemplarily, the filtering information of the second model included in the second message is used to indicate the conditions met for training the second model. When the MTLF does not find a second model that meets the requirements, the MTLF can train a second model that meets the requirements based on the filtering information of the second model.

[0370] Step S616: The MTLF determines the inference speed of the second model, the analysis type of the second model, etc. according to the second message.

[0371] Assuming that the current inference speed of the first model included in the second message is 10 / s and the acceleration factor of the accelerated inference is 2.5, it can be determined that the inference speed of the second model is approximately 100 / s.

[0372] The analysis type of the second model should be consistent with the analysis type of the first model included in the second message.

[0373] In step S618, the MTLF searches for a second model that meets the requirements based on the determined inference speed of the second model, the analysis type of the second model, etc.

[0374] After MTLF finds the second model that meets the requirements, it needs to feed back the second model to the execution entity of the second model.

[0375] Steps S620 to S650 may refer to the above-mentioned steps S520 to S550 and will not be described in detail here.

[0376] The technical solution of method 600 describes in detail how to find another model to assist the local model in accelerating the acquisition of analysis results, promoting collaboration between different network functions of the NWDAF, thereby enhancing the service assurance capabilities of the NWDAF in the network. Furthermore, communication method 600 can use the DCCF to collect data from data sources and dispatch it to the corresponding execution entities of multiple models. This can reduce the number of data collection attempts from data sources during the aforementioned multi-model collaborative accelerated inference process, thereby reducing data collection time.

[0377] AnLF1 in the above-mentioned communication method 300, communication method 400, communication method 500, and communication method 600 determines whether the local model can meet the analysis requirements based on the analysis requirements in the above-mentioned first message, thereby triggering multi-model collaborative reasoning. The present application can also provide another communication method 700, which does not require AnLF1 to predict whether the local model can meet the analysis requirements in the first message, and can reduce the processing complexity of AnLF1.

[0378] As shown in FIG7 , which is a schematic flow chart of a communication method 700, the communication method 700 may include the following steps:

[0379] Step S710: The first device sends a seventh message to AnLF1, where the seventh message includes a second analysis request for feeding back the first analysis result. Accordingly, AnLF1 receives the seventh message from the first device.

[0380] Exemplarily, the seventh message may be a subscription message (Nnwdaf_AnalyticsSubscription_Subscribe or Nnwdaf-_AnalyticsInfo_Request).

[0381] The second analysis requirement in the communication method 700 may include a second time requirement (expected waiting time / time when analytics is needed) and / or a second accuracy requirement. The seventh message may also include an analytics ID and the like.

[0382] Step S712: AnLF1 collects input data of the model according to the first message.

[0383] In step S714 , AnLF1 performs inference using the first model.

[0384] When AnLF1 times out the inference using the first model or the inference accuracy of AnLF1 using the first model is insufficient, in step S716, AnLF1 sends an eighth message to the first device, the eighth message including the third analysis requirement. Accordingly, the first device receives the eighth message from AnLF1.

[0385] The third analysis requirement may include a third time requirement (revised waiting time) and / or a third accuracy requirement. The third time requirement or the third accuracy requirement is a time requirement or an accuracy requirement that can be met by AnLF1 using the first model for reasoning.

[0386] In step S718, the first device sends the first message to AnLF1, where the first message includes a first analysis request for feeding back a first analysis result. Accordingly, AnLF1 receives the first message from the first device.

[0387] Illustratively, the first analysis requirement may include a first time requirement (expected waiting time / time when analytics is needed) and / or a first accuracy requirement.

[0388] For example, if AnLF1 cannot meet the second time requirement using the first model reasoning, the time required by the first time requirement is later than the time required by the second time requirement and the time required by the first time requirement is earlier than the time required by the third time requirement.

[0389] For example, if AnLF1 cannot meet the second accuracy requirement using the first model reasoning, the accuracy required by the first accuracy requirement may be equal to or lower than the accuracy required by the second accuracy requirement, and the accuracy required by the first accuracy requirement may be higher than the accuracy required by the third accuracy requirement.

[0390] Step S720: AnLF1 triggers dual-model collaborative reasoning.

[0391] For example, if AnLF1 triggers dual-model collaborative reasoning to meet the above-mentioned first time requirement, then AnLF1 triggers dual-model collaborative accelerated reasoning, corresponding to the above-mentioned first network function requesting the second network function to determine the first and fourth cases of the second model.

[0392] For example, if AnLF1 triggers dual-model collaborative reasoning to meet the above-mentioned first accuracy requirement, then AnLF1 triggers dual-model collaborative reasoning, corresponding to the above-mentioned first network function requesting the second network function to determine the second and third cases of the second model.

[0393] Exemplarily, if AnLF1 triggers dual-model collaborative reasoning to meet the above-mentioned first time requirement and first accuracy requirement, then AnLF1 triggers dual-model collaborative reasoning, corresponding to the fifth situation in which the first network function requests the second network function to determine the second model.

[0394] The steps after AnLF1 triggers dual-model collaborative reasoning can refer to steps S316 to S348 in the communication method 300. Alternatively, the steps after AnLF1 triggers dual-model collaborative reasoning can refer to steps S416 to S448 in the communication method 400.

[0395] Through the above-mentioned communication method 700, another triggering method for AnLF1 to trigger dual-model collaborative reasoning is provided. This method does not require AnLF1 to pre-judge whether the local model can meet the analysis requirements in the first message, which can reduce the processing complexity of AnLF1.

[0396] Finally, the device embodiment of the embodiment of the present application is introduced.

[0397] To implement the various functions of the method provided herein, the first device, the first network function, and the second network function may each include hardware structures and / or software modules, and the aforementioned functions may be implemented in the form of hardware structures, software modules, or a combination of hardware structures and software modules. Whether a particular one of the aforementioned functions is implemented in the form of hardware structures, software modules, or a combination of hardware structures and software modules depends on the specific application and design constraints of the technical solution.

[0398] Figure 8 is a schematic block diagram of a communication device 800 according to an embodiment of the present application. The communication device 800 includes a processor 810 and a communication interface 820. Optionally, the processor 810 and the communication interface 820 may be interconnected via a bus 830. The communication device 800 may be a first network function, a second network function, a first device, or a second device.

[0399] Optionally, the communication device 800 may further include a memory 840. The memory 840 includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or portable read-only memory (CD-ROM), and is used to store relevant instructions and data.

[0400] The processor 810 may be one or more central processing units (CPUs). In the case where the processor 810 is a CPU, the CPU may be a single-core CPU or a multi-core CPU.

[0401] When the communication apparatus 800 is a first network function, illustratively, the communication apparatus 800 is configured to perform the following operations: receiving a first message from a first device, or sending a second message to a second network function, etc.

[0402] When the communication device 800 is a second network function, illustratively, the communication device 800 is configured to perform the following operations: receiving a second message from the first network function, or determining a second model according to the second message, etc.

[0403] When the communication apparatus 800 is a first device, illustratively, the communication apparatus 800 is configured to perform the following operations: sending a first message to a first network function, etc.

[0404] The above contents are merely exemplary descriptions. When the communication apparatus 800 is the first network function / the second network function / the first device, it will be responsible for executing the methods or steps related to the first network function / the second network function / the first device in the aforementioned method embodiments.

[0405] The above description is merely exemplary. For details, please refer to the contents of the above method embodiments. The implementation of each operation in FIG8 may also correspond to the corresponding description of the method embodiments shown in FIG2 to FIG7.

[0406] Figure 9 is a schematic block diagram of a communication device 900 according to an embodiment of the present application. Communication device 900 may be a first network function, a second network function, or a first device, or may be a chip or module within the first network function, the second network function, or the first device, configured to implement the methods described in the above embodiments. Communication device 900 includes a transceiver unit 910 and a processing unit 920. The following provides an exemplary description of transceiver unit 910 and processing unit 920.

[0407] The transceiver unit 910 may include a transmitting unit and a receiving unit. The transmitting unit is used to perform the transmitting operation of the communication device 900, and the receiving unit is used to perform the receiving operation of the communication device 900. For ease of description, the embodiment of the present application combines the transmitting unit and the receiving unit into a single transceiver unit. This is described here as a unified description and will not be repeated later.

[0408] When the communication device 900 is a first network function, illustratively, the transceiver unit 910 is used to receive a first message from a first device, and the processing unit 920 is used to determine whether the first time requirement cannot be met using the first model based on the first message, the input data of the model, the time for collecting the input data of the model, and the inference speed of the first model.

[0409] When the communication device 900 is a second network function, illustratively, the transceiver unit 910 is configured to receive a second message from the first network function, and the processing unit 920 is configured to determine a second model according to the second message.

[0410] When the communication apparatus 900 is a first device, illustratively, the transceiver unit 910 is configured to send a first message to a first network function.

[0411] The above contents are merely exemplary descriptions. When the communication apparatus 900 is the first network function / the second network function / the first device, it will be responsible for executing the methods or steps related to the first network function or the second network function or the first device in the aforementioned method embodiments.

[0412] Optionally, the communication device 900 further includes a storage unit 930, which is used to store a program or code for executing the aforementioned method.

[0413] The device embodiments shown in Figures 8 and 9 are used to implement the contents described in Figures 2 to 7. The specific execution steps and methods of the devices shown in Figures 8 and 9 can refer to the contents described in the above method embodiments.

[0414] Figure 10 is a schematic block diagram of a communication device 1000 according to an embodiment of the present application. The communication device 1000 is used to implement the functions of the first network function / the second network function / the first device. The communication device 1000 may be a chip in the first network function / the second network function / the first device.

[0415] Communication device 1000 includes an input / output interface 1020 and a processor 1010. Input / output interface 1020 may be an input / output circuit. Processor 1010 may be a signal processor, a chip, or other integrated circuit capable of implementing the method of the present application. Input / output interface 1020 is used for inputting or outputting signals or data.

[0416] For example, when communication device 1000 is a first network function, input / output interface 1020 is configured to receive a first message from a first device. Processor 1010 is configured to determine, based on the first message, model input data, the time taken to collect the model input data, and the inference speed of the first model, that the first time requirement cannot be met using the first model.

[0417] For example, when the communication device 1000 is the second network function, the input / output interface 1020 is configured to receive a second message from the first network function, and the processor 1010 is configured to determine the second model according to the second message.

[0418] For example, when the communication apparatus 1000 is a first device, the input / output interface 1020 is used to send a first message to a first network function.

[0419] In one possible implementation, the processor 1010 implements the first network function or the second network function or the function implemented by the first device by executing instructions stored in the memory.

[0420] Optionally, the communication device 1000 further includes a memory.

[0421] Optionally, the processor and memory are integrated together.

[0422] Optionally, the memory is outside the communication device 1000 .

[0423] In one possible implementation, the processor 1010 may be a logic circuit, which inputs / outputs messages or signals through the input / output interface 1020. The logic circuit may be a signal processor, a chip, or other integrated circuit that can implement the method of the embodiment of the present application.

[0424] The above description of the communication device 1000 is only an exemplary description. The communication device 1000 can be used to execute the method described in the above embodiment. For specific content, please refer to the description of the above method embodiment, which will not be repeated here.

[0425] The present application also provides a chip, including a processor, for calling and executing instructions stored in a memory from the memory, so that a communication device equipped with the chip executes the methods in the above examples.

[0426] The present application also provides a chip, comprising: an input interface, an output interface, and a processor, wherein the input interface, the output interface, and the processor are connected via an internal connection path, and the processor is configured to execute code in a memory. When the code is executed, the processor is configured to execute the methods in the above examples. Optionally, the chip also includes a memory, which is configured to store computer programs or code.

[0427] The present application also provides a processor, which is coupled to a memory and is used to execute the method and function involving the first network function or the second network function or the first device in any of the above embodiments.

[0428] The present application provides a computer program product comprising instructions. When the computer program product is run on a computer, the method of the aforementioned embodiment is implemented.

[0429] The present application also provides a computer program. When the computer program is executed in a computer, the method of the aforementioned embodiment is implemented.

[0430] The present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a computer, the method described in the above embodiment is implemented.

[0431] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0432] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0433] In the several embodiments provided in this application, the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0434] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the technical solutions of the embodiments of the present application.

[0435] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0436] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of each method embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a ROM, a RAM, a magnetic disk, or an optical disk.

[0437] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A communication method, characterized in that, The method includes: A first network function receives a first message from a first device, the first message including a first analysis requirement for feeding back a first analysis result, and the first network function is deployed with a first model; The first network function sends a second message to a second network function based on the first message, the second message being used to determine a second model, and the second model is used to assist the first model in obtaining the first analysis result to meet the first analysis requirement.

2. The method according to claim 1, wherein The second message includes first indication information, and the first indication information is used to indicate the execution entity of the second model.

3. The method according to claim 1 or 2, characterized in that, The second message includes at least one of the following: The inference speed of the second model, the analysis type of the second model.

4. The method according to claim 1 or 2, characterized in that, The second message includes at least one of the following: The inference speed of the first model, the analysis type of the first model, the acceleration multiple for accelerating inference.

5. The method according to any one of claims 1 to 4, characterized in that The method further includes: The first network function receives a third message from the second network function, the third message being used to indicate the second model; The first network function performs inference using the first model and the second model to obtain the first analysis result.

6. The method according to any one of claims 1 to 4, characterized in that The method further includes: The first network function obtains the requirements for the input data of the first model and the requirements for the output data of the second model.

7. The method according to claim 6, wherein The requirements for the input data of the first model include the input data of the first model, and the requirements for the output data of the second model include the number N of output values for each inference during the inference process of the second model, where N is a positive integer greater than or equal to 1.

8. The method according to claim 7, wherein The method further includes: The first network function receives N first output values from a second device, and the second device is the execution entity of the second model; The first network function verifies the N first output values using the first model and determines to reject M of the N first output values; The first network function sends a fourth message to the second device, and the fourth message is used to indicate the M first output values, where M is a positive integer less than or equal to N.

9. The method according to any one of claims 1 to 4, characterized in that, The method further includes: The first network function obtains the requirements for the input data of the first model and the requirements for the output data of the first model.

10. The method according to claim 9, characterized in that, The requirements for the input data of the first model include the input data of the first model, and the requirements for the output data of the first model include the number N of output values for each inference during the inference process of the first model, where N is a positive integer greater than or equal to 1.

11. The method according to claim 10, wherein The method further includes: The first network function sends N first output values to a second device, and the second device is the execution entity of the second model; The first network function receives a fifth message from the second device, the fifth message being used to indicate M first output values, and the M first output values are the output values rejected by the second device among the N first output values, where M is a positive integer less than or equal to N.

12. The method according to any one of claims 1 to 11, characterized in that, Before the first network function receives the first message from the first device, the method further includes: The first network function receives a seventh message from the first device, and the seventh message includes a second analysis requirement for feeding back the first analysis result; When the first network function fails to meet the second analysis requirement, the first network function sends an eighth message to the first device, and the eighth message includes a third analysis requirement for feeding back the first analysis result, the third analysis requirement is lower than the second analysis requirement, and the third analysis requirement is lower than the first analysis requirement.

13. The method according to any one of claims 1 to 12, characterized in that, The second analysis requirement includes a time requirement and / or an accuracy requirement.

14. A communication method, characterized in that, The method includes: A second network function receives a second message from the first network function; The second network function determines a second model according to the second message, and the second model is used to assist the first model to obtain a first analysis result to meet a first analysis requirement, and the first model is deployed on the first network function.

15. The method according to claim 14, wherein The second message includes first indication information, and the first indication information is used to indicate the execution entity of the second model.

16. The method according to claim 14 or 15, characterized in that The second message includes at least one of the following: The inference speed of the second model, the analysis type of the second model.

17. The method according to claim 14 or 15, characterized in that, The second message includes at least one of the following: The inference speed of the first model, the analysis type of the first model, the acceleration multiple for accelerating inference.

18. The method according to any one of claims 14 to 17, characterized in that The method further includes: The second network function sends a third message to the first network function, and the third message is used to indicate the second model.

19. The method according to claim 15 or 16, characterized in that, The method further includes: The second network function sends a sixth message to a second device, and the sixth message is used to indicate the first network function and the second model, and the second device is the execution entity of the second model.

20. The method according to any one of claims 14 to 19, characterized in that, The first analysis requirement includes a time requirement and / or an accuracy requirement.

21. A communication method, characterized in that, The method includes: A second device receives a sixth message from the second network function, and the sixth message is used to indicate a first network function and a second model, and the first network function deploys a first model; The second device performs inference using the second model, and the second model is used to assist the first model to obtain a first analysis result to meet a first analysis requirement.

22. The method according to claim 21, wherein The method further includes: The second device obtains the requirements for the input data of the second model and the requirements for the output data of the second model.

23. The method according to claim 22, wherein The requirements for the input data of the second model include the input data of the second model, and the requirements for the output data of the second model include the number N of output values for each inference during the inference process of the second model, where N is a positive integer greater than or equal to 1.

24. The method according to claim 23, wherein The method further includes: The second device sends N first output values to the first network function; The second device receives a fourth message from the first network function, and the fourth message is used to indicate M first output values, and the M first output values are the output values rejected by the first network function among the N first output values, where M is a positive integer less than or equal to N.

25. The method according to claim 21, wherein The method further includes: The second device obtains the requirements for the input data of the second model and the requirements for the output data of the first model.

26. The method according to claim 25, wherein The requirements for the input data of the second model include the input data of the second model, and the requirements for the output data of the first model include the number N of output values for each inference during the inference process of the first model. Wherein, N is a positive integer greater than or equal to 1.

27. The method according to claim 26, wherein The method further includes: The second device receives N first output values from the first network function; The second device uses the second model to verify the N first output values and determines to reject M first output values among the N first output values; The second device sends a fifth message to the first network function, and the fifth message is used to indicate the M first output values. Wherein, M is a positive integer less than or equal to N.

28. The method according to any one of claims 21 to 27, characterized in that, The first analysis requirement includes a time requirement and / or an accuracy requirement.

29. A communication device, characterized in that, Including a processor, the processor is configured to, by executing a computer program or instruction, cause the communication device to execute the method according to any one of claims 1 to 13, or cause the communication device to execute the method according to any one of claims 14 to 20, or cause the communication device to execute the method according to any one of claims 21 to 28.

30. The communication device according to claim 29, characterized in that, The communication device further includes a memory for storing the computer program or instruction.

31. The communication device according to claim 29, characterized in that, The communication device further includes a communication interface for inputting and / or outputting signals.

32. A computer-readable storage medium, characterized in that, A computer program or instruction is stored on the computer-readable storage medium, and when the computer program or the instruction runs on a computer, cause the method according to any one of claims 1 to 13 to be executed, or cause the method according to any one of claims 14 to 20 to be executed, or cause the method according to any one of claims 21 to 28 to be executed.

33. A computer program product, characterized in that, Containing instructions, when the instructions run on a computer, cause the method according to any one of claims 1 to 13 to be executed, or cause the method according to any one of claims 14 to 20 to be executed, or cause the method according to any one of claims 21 to 28 to be executed.

Images