# US 12,184,507 B2 · Data processing method and apparatus > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,184,507 B2 - Document ID: US12184507B2 - Status: Granted - Kind: B2 - Application number: 18/145,129 - Filed: 2022-12-22 - Published: 2023-04-27 - Granted: 2024-12-31 - Earliest claimed: 2020-06-23 - Assignees: Huawei Technologies Co., Ltd. - Applicants: Huawei Technologies Co., Ltd. - Inventors: Weiwei Chong; Yang Xin; Xiaobo Wu - Web page: https://patentgrep.com/patent/US12184507B2 ## CPC - H04L 41/14: Network analysis or design - H04L 41/145: {involving simulating, designing, planning or modelling of a network} - H04L 41/16: using machine learning or artificial intelligence - H04L 41/34: Signalling channels for network management communication - H04L 43/022: by sampling - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) ## Abstract A data processing method and an apparatus relate to the communication field. The data processing method includes: A first network data analytics network element determines identification information of one or more submodels related to a target model. The first network data analytics network element sends the identification information of the submodel to a network element corresponding to the identification information of the submodel. The first network data analytics network element receives, from the network element, a value that reflects impact of the submodel on the target model, and performs training, to obtain the target model. The data processing method and apparatus provide functions such as accurate data collection, data analysis, and data training. ## Claims ### Claim 1 A data processing method comprising: determining, by a first network data analytics network element, identification information of one or more submodels related to a target model; sending, by the first network data analytics network element, a first request message to a network element corresponding to the identification information of the one or more submodels, wherein the first request message comprises the identification information of the one or more submodels; receiving, by the first network data analytics network element, a first response message from the network element, wherein the first response message comprises a first output result obtained through calculation by the network element by using the one or more submodels, and the first output result reflects a value; and obtaining, by the first network data analytics network element, the target model through training that is based on the first output result of the network element corresponding to the identification information of the one or more submodels; wherein the determining, by the first network data analytics network element, of the identification information of one or more submodels related to the target model comprises: determining, by the first network data analytics network element based on a local configuration, the identification information of the one or more submodels related to the target model; or determining, by the first network data analytics network element, the network element that is related to the target model and that has a confidentiality requirement; and obtaining, by the first network data analytics network element, the identification information of the one or more submodels from the network element having the confidentiality requirement. ### Claim 2 (depends on 1) The method according to claim 1, wherein the first request message further comprises one or more of the following: first slice identification information, first service identification information, first area information, first time information, first terminal identification information, first network device identification information, first model training algorithm information, first model feedback validity time information, or first model convergence condition information, wherein the first slice identification information indicates to provide an output result corresponding to the first slice identification information; the first service identification information indicates to provide an output result corresponding to the first service identification information; the first area information indicates to provide an output result corresponding to the first area information; the first time information indicates to provide an output result corresponding to the first time information; the first terminal identification information indicates to provide an output result corresponding to the first terminal identification information; the first network device identification information indicates to provide an output result corresponding to the first network device identification information; the first model training algorithm information is for recommending a model training algorithm to be used for the one or more submodels; the first model feedback validity time information indicates a validity time period for feeding back an output result; and the first model convergence condition information indicates a model training convergence condition. ### Claim 3 (depends on 1) The method according to claim 1, wherein the first response message further comprises one or more of the following: the identification information of the one or more submodels, association information, a quantity N of iterations, model training algorithm information, slice identification information, service identification information, or area information, wherein the association information is for associating output results of a same sample object; the quantity N of iterations indicates that an output result of the network element corresponding to the identification information of the one or more submodels is a result generated in an Nth iteration process; the model training algorithm information indicates an algorithm used by the network element corresponding to the identification information of the one or more submodels; the slice identification information indicates that the output result of the network element corresponding to the identification information of the one or more submodels comprises an output result for the slice identification information; the service identification information indicates that the output result of the network element corresponding to the identification information of the one or more submodels comprises an output result for the service identification information; and the area information indicates that the output result of the network element corresponding to the identification information of the one or more submodels comprises an output result for the area information. ### Claim 4 (depends on 3) The method according to claim 3, wherein the association information comprises one or more of the following: an interface identifier, an N2 tunnel identifier, an N3 tunnel identifier, association time information, a quality of service flow identifier (QFI), an internet protocol IP 5-tuple, an international mobile subscriber identity (IMSI), a generic public subscription identifier (GPSI), or a subscriber permanent identifier (SUPI); or the association information comprises an association indication and one or more of the following information: the interface identifier, the N2 tunnel identifier, the N3 tunnel identifier, the association time information, the quality of service flow identifier (QFI), the internet protocol IP 5-tuple, the international mobile subscriber identity (IMSI), the generic public subscription identifier (GPSI), or the subscriber permanent identifier (SUPI). ### Claim 5 (depends on 1) The method according to claim 1, wherein the obtaining, by the first network data analytics network element, of the target model through training that is based on the first output result of the network element corresponding to the identification information of the one or more submodels comprises: inputting, by the first network data analytics network element, the first output result of the network element to a preset model; and adjusting, by the first network data analytics network element, the preset model based on a loss function difference between sample label data and an output result of the preset model until the loss function difference between the sample label data and the output result of the preset model meets a condition, to obtain the target model. ### Claim 6 (depends on 1) The method according to claim 1, wherein the method further comprises: sending, by the first network data analytics network element, a loss function difference to the network element corresponding to the identification information of the one or more submodels, wherein the network element corresponding to the identification information of the one or more submodels updates the one or more submodels based on the loss function difference. ### Claim 7 (depends on 1) The method according to claim 1, wherein the method further comprises: receiving, by the first network data analytics network element, a second request message from a second network data analytics network element, wherein the second request message is for requesting to train the target model, and the second request message comprises an identifier of the target model. ### Claim 8 (depends on 7) The method according to claim 7, wherein the second request message comprises one or more of the following: slice identification information, service identification information, area information, time information, terminal identification information, network device identification information, model training algorithm information, model feedback validity time information, or model convergence condition information. ### Claim 9 (depends on 7) The method according to claim 7, wherein the method further comprises: sending, by the first network data analytics network element, a second response message to the second network data analytics network element, wherein the second response message comprises the target model or address information corresponding to the target model. ### Claim 10 (depends on 9) The method according to claim 9, wherein the second response message further comprises one or more of the following: the slice identification information, the service identification information, the area information, the time information, the terminal identification information, the network device identification information, or model training complete indication information. ### Claim 11 (depends on 1) The method according to claim 1, further comprising: receiving, by the first network data analytics network element, an inference result from a first network element, wherein the inference result is obtained by the first network element based on the one or more submodels in the first network element; outputting, by the first network data analytics network element, a target inference result based on the target model and the inference result of the first network element; and sending, by the first network data analytics network element, the target inference result to a second network element, wherein the second network element is the same as or different from the first network element. ### Claim 12 An apparatus comprising: a memory configured to store program instructions; and a processor configured to execute the program instructions to cause the apparatus to: determine identification information of one or more submodels related to a target model; send a first request message to a network element corresponding to the identification information of the one or more submodels, wherein the first request message comprises the identification information of the one of more submodels; receive a first response message from the network element, wherein the first response message comprises a first output result obtained through calculation by the network element by using the one or more submodels, and the first output result reflects a value; and obtain the target model through training that is based on the first output result of the network element corresponding to the identification information of the one or more submodels; wherein to determine the identification information of the one or more submodels related to the target model, the apparatus further caused to: determine, based on a local configuration, the identification information of the one or more submodels related to the target model; or determine the network element that is related to the target model and that has a confidentiality requirement; and obtain the identification information of the one or more submodels from the network element having the confidentiality requirement. ### Claim 13 (depends on 12) The apparatus according to claim 12, wherein the processor is configured to execute the program instructions in the memory to cause the apparatus to input the first output result of the network element to a preset model; and adjust the preset model based on a loss function difference between sample label data and an output result of the preset model until the loss function difference between the sample label data and the output result of the preset model meets a condition, to obtain the target model. ### Claim 14 (depends on 12) The apparatus according to claim 12, wherein the processor is configured to execute the program instructions in the memory to cause the apparatus to send a loss function difference to the network element corresponding to the identification information of the one or more submodels, wherein the network element corresponding to the identification information of the one or more submodels updates the one or more submodels based on the loss function difference. ### Claim 15 (depends on 12) The apparatus according to claim 12, wherein the processor is configured to execute the program instructions in the memory to cause the apparatus to receive a second request message from a network data analytics network element, wherein the second request message is for requesting to train the target model, and the second request message comprises an identifier of the target model. ### Claim 16 (depends on 15) The apparatus according to claim 15, wherein the processor is configured to execute the program instructions in the memory to cause the apparatus to send a second response message to the network data analytics network element, wherein the second response message comprises the target model or address information corresponding to the target model. ### Claim 17 (depends on 12) The apparatus according to claim 12, wherein the processor is configured to execute the program instructions in the memory to cause the apparatus to receive an inference result from a first network element, wherein the inference result is obtained by the first network element based on the one or more submodels in the first network element; output a target inference result based on the target model and the inference result of the first network element; and send the target inference result to a second network element, wherein the second network element is the same as or different from the first network element. ### Claim 18 A communication system, comprising a first network data analytics network element and a second network data analytics network element, wherein the first network data analytics network element includes a processor and a non-transitory computer-readable storage medium storing instructions executable by the processor such that the first network data analytics network element is at least configured to: determine identification information of one or more submodels related to a target model; send a first request message to a network element corresponding to the identification information of the one or more submodels, wherein the first request message comprises the identification information of the one or more submodels; receive a first response message from the network element, wherein the first response message comprises a first output result obtained through calculation by the network element by using the one or more submodels, and the first output result reflects a value; and perform training that is based on the first output result of the network element corresponding to the identification information of the one or more submodels to obtain the target model; wherein to determine the identification information of the one or more submodels related to the target model, the first network data analytics network element is further configured to: determine, based on a local configuration, the identification information of the one or more submodels related to the target model; or determine the network element that is related to the target model and that has a confidentiality requirement; and obtain the identification information of the one or more submodels from the network element having the confidentiality requirement. ### Claim 19 (depends on 18) The communication system according to claim 18, wherein the first network data analytics network element is further configured to is configured to input the first output result of the network element to a preset model; and adjust the preset model based on a loss function difference between sample label data and an output result of the preset model until the loss function difference between the sample label data and the output result of the preset model meets a condition, to obtain the target model. ### Claim 20 (depends on 18) The communication system according to claim 18, wherein the first network data analytics network element is further configured to receive a second request message from the second network data analytics network element, wherein the second request message is for requesting to train the target model, and the second request message comprises an identifier of the target model.