Communication methods, communication devices and systems

By determining the selection and transfer of network elements for model training based on energy information in network devices, and continuing training using the information of already trained models, the problem of high carbon emissions and energy consumption in network device model training is solved, achieving energy conservation, emission reduction and efficiency improvement.

CN123123931APending Publication Date: 2026-11-17HUAWEI TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202510631894.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-05-15
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

The training process for network equipment models generates significant carbon emissions, leading to increased operating costs and low energy efficiency.

Method used

The first grid element decides whether to stop or transfer model training based on its own and other grid elements' energy information, continues training using the information of the already trained model, and selects grid elements with strong energy-saving capabilities for model training to reduce carbon emissions caused by changes in energy information.

Benefits of technology

It reduces energy consumption and carbon emissions during model training, and improves the efficiency and resource utilization of model training.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123123931A_ABST
    Figure CN123123931A_ABST
Patent Text Reader

Abstract

This application discloses a communication method, communication device, and system. The method includes: a first network element stopping training a first model based on its own energy information; and sending first information to a second network element requesting the second network element to train the first model, the second network element being determined based on its own energy information. This method considers its own energy information and the energy information of other network elements during training, which can reduce carbon emissions generated during model training.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to communication methods, communication devices, and systems. Background Technology

[0002] Carbon emissions from network equipment primarily stem from its power consumption, which directly increases electricity costs. Electricity costs are a major component of operators' operating expenses (OPEX), therefore, reducing carbon emissions can lower these operating costs. Furthermore, with increasing global focus on carbon emissions, data centers and communication networks need to control their emissions and improve energy efficiency. All of this implies a requirement for networks to reduce their carbon emissions.

[0003] Currently, training network devices for models generates significant carbon emissions. Reducing these emissions is a practical problem that needs to be addressed. Summary of the Invention

[0004] This application discloses a communication method, communication device, and system that can reduce carbon emissions generated during model training.

[0005] The present application is described below from different aspects. It should be understood that the different implementation methods and beneficial effects described below can be referenced from each other.

[0006] Firstly, this application provides a communication method that can be executed by a first network element. The first network element is a network device or a chip, chip system, module, or control unit of a network device, such as a server on the network side or a component (e.g., circuit, chip, or chip system) within the server. This application does not specifically limit the definition of the first network element. It should be noted that, in this application, the term "network device" can refer to either the network device itself or the chip, functional module, or integrated circuit that performs the method provided in this application. This application does not specifically limit the definition of the first network element.

[0007] Taking the execution of this method by a first network element as an example, the method may include: the first network element stopping training the first model based on its own energy information; and sending first information to a second network element, the first information being used to request the second network element to train the first model, the second network element being determined based on its own energy information.

[0008] In the above scheme, the first network element considers its own energy information (i.e., the energy information of the first network element mentioned above) during training, and can stop training the first model based on its own energy situation; based on the energy information of other network elements (i.e., the energy information of the second network element mentioned above), it determines the next network element to be trained (i.e., the second network element). This method, by considering its own energy information and the energy information of other network elements during training, can reduce the carbon emissions generated during model training.

[0009] In conjunction with the first aspect, in one possible implementation, the first information includes model information of a first model, which is used for model training of the first model; wherein, before training of the first model is stopped, the training of the first model is not completed, and the model information of the first model is used to indicate the first model that has not completed training.

[0010] Alternatively, the first network element trains the initial model; when training of the first model stops as described above, the first network element has already trained the initial model into a second model, which is an incompletely trained model; the first network element sends the second model (i.e., the model information of the first model) to the second network element so that the second network element can continue training the second model without having to train from the initial model. Here, the initial model and the second model are the models corresponding to the first model at different training stages.

[0011] In this method, the first network element can send the trained model information (i.e., the model information of the first model) to the second network element so that the second network element can continue to train the model based on the model information. In this method, the second network element does not need to start training from the initial model, which can improve the efficiency of model training, save training resources, and reduce training energy consumption.

[0012] Optionally, the model information of the first model may refer to the model trained by the first network element when (or before) training of the first model is stopped. For example, the model information of the first model may be model parameters, model gradients, etc.

[0013] In conjunction with the first aspect, in one possible implementation, the energy information of the first grid element includes at least one of the following: the energy efficiency of the first grid element, the normalized energy consumption of the first grid element, the energy status of the first grid element, the energy consumption of the first grid element within a preset time period, the renewable energy ratio of the first grid element, the carbon emission factor (CEF) of the first grid element, the energy type of the first grid element, the energy efficiency level of the first grid element, or the energy consumption of the first grid element.

[0014] In this method, the energy information of the first network element, such as energy efficiency and energy consumption information (e.g., normalized energy consumption), can be considered to determine whether to stop training, which can reduce the increase in carbon emissions caused by changes in the energy information of the first network element.

[0015] In conjunction with the first aspect, in one possible implementation, based on the energy information of the first network element, stopping the training of the first model includes: stopping the training of the first model if the energy information of the first network element meets a first condition, wherein the first condition includes at least one of the following: the energy efficiency of the first network element is lower than a first energy efficiency threshold; the normalized energy consumption of the first network element exceeds a first normalized energy consumption threshold; the first network element enters an energy-saving state; the energy consumption of the first network element exceeds a first energy consumption threshold within a preset time; the renewable energy ratio of the first network element is lower than a first renewable energy ratio threshold; the carbon emission factor of the first network element is higher than a first carbon emission factor threshold; the energy type of the first network element changes; the energy efficiency level of the first network element is greater than a first energy efficiency level threshold; or, the energy consumption of the first network element exceeds a first network element energy consumption limit level.

[0016] In this method, if the energy information during the training process of the first network element does not meet the first condition, the first network element stops training the first model, which can reduce the situation where changes in the energy information of the first network element lead to an increase in carbon emissions.

[0017] Optionally, the change in the energy type of the first grid element can refer to changing the energy type of the first grid element from a first type to a second type. For example, the energy consumption (e.g., normalized energy consumption) caused by using the second type of energy is greater than the energy consumption caused by using the first type of energy, or the energy efficiency of the second type of energy is less than the energy efficiency of the first type of energy. For instance, the first type of energy is green energy or new energy, while the second type of energy is traditional energy, such as coal energy.

[0018] In conjunction with the first aspect, in one possible implementation, the method further includes: sending a first message to a third network element, the first message being used to request the identification of a second network element; and receiving second information sent by the third network element, the second information being used to instruct the second network element.

[0019] Optionally, the first message used to request the determination of the second network element can mean that the first message is used to request the determination (or discovery) of the network element to perform training. For ease of description, this application refers to the network element to perform training determined based on the first message as the second network element.

[0020] In conjunction with the first aspect, in one possible implementation, the first message includes third information, and the second network element is determined based on the third information. The third information includes at least one of the following: new energy ratio information (such as a second new energy ratio threshold), carbon emission factor information (such as a second carbon emission factor threshold), energy type information (such as green energy), normalized energy consumption information (such as a second normalized energy consumption threshold), energy status information (such as energy-saving status), energy status timetable information, low energy efficiency indication, energy-saving status migration execution indication, energy consumption information (such as a second normalized energy consumption threshold), energy efficiency level performance index information, energy efficiency level information (such as a second energy efficiency level threshold), or network element energy consumption limit level information (such as a second network element energy consumption limit level).

[0021] In conjunction with the first aspect, in one possible implementation, the first message includes third information, which indicates a second condition. The energy information of the second network element meets the second condition, which includes at least one of the following: the energy efficiency of the second network element is lower than a second energy efficiency threshold; the normalized energy consumption of the second network element exceeds a second normalized energy consumption threshold; the second network element enters an energy-saving state; the energy consumption of the second network element exceeds a second energy consumption threshold within a preset time; the renewable energy ratio of the second network element is lower than a second renewable energy ratio threshold; the carbon emission factor of the second network element is higher than a second carbon emission factor threshold; the energy type of the second network element is a first energy type; the energy efficiency level of the second network element is greater than a second energy efficiency level threshold; or, the energy consumption of the second network element exceeds a second network element energy consumption limit level.

[0022] In the above method, selecting network elements whose energy information meets the second condition (i.e., the second network element) can reduce the increase in carbon emissions during training caused by changes in the energy state of network elements.

[0023] Optionally, the second energy efficiency threshold may be the same as or different from the first energy efficiency threshold; the second normalized energy consumption threshold may be the same as or different from the first energy efficiency threshold; the second energy consumption threshold may be the same as or different from the first energy consumption threshold; the second new energy ratio threshold may be the same as or different from the first new energy ratio threshold; the second carbon emission factor threshold may be the same as or different from the first carbon emission factor; the second energy efficiency level threshold may be the same as or different from the first energy efficiency level threshold; and the second grid element energy consumption restriction level may be the same as or different from the first grid element energy consumption restriction level.

[0024] In conjunction with the first aspect, in one possible implementation, sending the first information to the second network element includes: sending a second message to the second network element, the second message being used to request model training, the second message including fourth information, the fourth information being used to instruct the second network element to notify the training result of the model training.

[0025] In conjunction with the first aspect, in one possible implementation, the first condition is either pre-configured or obtained from other network elements.

[0026] Secondly, this application provides a communication method that can be executed by a third network element. The third network element is a network device or a chip, chip system, module, or control unit of the network device, such as a server on the network side or components within the server (e.g., circuits, chips, or chip systems). This application does not specifically limit the definition of the third network element. It should be noted that, in this application, the term "network device" can refer to either the network device itself or the chip, functional module, or integrated circuit that performs the method provided in this application. This application does not specifically limit the definition of the third network element.

[0027] Taking the method executed by a third network element as an example, the method may include: receiving a first message sent by a first network element, the first message being used to request the determination of a second network element; sending third information to the first network element, the third information being used to instruct the second network element, the second network element being determined based on the energy information of the second network element.

[0028] In the above scheme, the network element to be trained (i.e. the second network element) is determined based on the energy information of the network element. The energy information of the network element can represent the energy status of the network element. This method can reduce the large carbon emissions generated by model training due to the energy status of the network element.

[0029] In conjunction with the second aspect, in one possible implementation, the first message includes third information, which indicates a second condition. The energy information of the second network element meets the second condition, which includes at least one of the following: the energy efficiency of the second network element is lower than a second energy efficiency threshold; the normalized energy consumption of the second network element exceeds a second normalized energy consumption threshold; the second network element enters an energy-saving state; the energy consumption of the second network element exceeds a second energy consumption threshold within a preset time; the renewable energy ratio of the second network element is lower than a second renewable energy ratio threshold; the carbon emission factor of the second network element is higher than a second carbon emission factor threshold; the energy type of the second network element is a first energy type; the energy efficiency level of the second network element is greater than a second energy efficiency level threshold; or, the energy consumption of the second network element exceeds a second network element energy consumption limit level.

[0030] Thirdly, this application provides a communication method that can be executed by a fourth network element. The fourth network element is a network device or a chip, chip system, module, or control unit of the network device, such as a server on the network side or components within the server (e.g., circuits, chips, or chip systems). This application does not specifically limit the definition of the fourth network element. It should be noted that, in this application, the term "network device" can refer to either the network device itself or the chip, functional module, or integrated circuit that performs the method provided in this application. This application does not specifically limit the definition of the fourth network element.

[0031] Alternatively, the fourth network element may be a terminal or a terminal's chip, chip system, module, or control unit; this application does not limit the specific details. It should be noted that in this application, the term "terminal" can refer to the terminal itself, or to the chip, functional module, or integrated circuit that performs the method provided in this application; this application does not limit the specific details. For example, the chips in this application include, but are not limited to, modem chips, also known as baseband chips, or system-on-chip (SoC) chips or system-in-package (SIP) chips containing a modem core; these will not be elaborated further below.

[0032] Taking the execution of this method by the fourth network element as an example, the method may include: determining the first network element, which is determined based on the first energy information; sending a first message to the first network element, the first message being used to request training of the first model, the first message including first information, which is related to the energy consumption requirements corresponding to the first model.

[0033] In the above scheme, the network element to be trained (i.e., the second network element) is determined based on the energy information of the network element. The energy information of the network element can represent the energy status of the network element. This method can reduce the large carbon emissions generated by model training due to the energy status of the network element. By indicating the energy consumption requirements of the first model training through the first information, the energy consumption of model training can be reduced, thereby reducing the carbon emissions generated by training.

[0034] In conjunction with the third aspect, in one possible implementation, the method further includes: receiving the training results of a first model sent by a first network element, the training results being obtained based on energy consumption requirements.

[0035] In other words, the training process that yielded the above training results meets the aforementioned energy consumption requirements.

[0036] In the above method, the first information indicates the energy consumption requirements for training the first model, so that the first unit of capital can train the first model based on the energy consumption requirements. This method can reduce the energy consumption of model training, thereby reducing the carbon emissions generated during training.

[0037] In conjunction with the third aspect, in one possible implementation, the first energy information is a first energy-saving capability. Determining a first network element includes: sending second information to a third network element, the second information being used to indicate the first energy-saving capability; receiving third information sent by the third network element, the third information being used to indicate the first network element, the first network element supporting the first energy-saving capability.

[0038] In the above scheme, the network element that supports the first energy-saving capability (i.e., the first network element) is determined as the network element that performs training, which can reduce the large carbon emissions generated by model training due to the limited energy-saving capability of the network element.

[0039] In conjunction with the third aspect, in one possible implementation, the determination of the first network element includes: acquiring the energy-saving capabilities of multiple network elements; determining the first network element based on the energy-saving capabilities of the multiple network elements, wherein the first network element supports the first energy-saving capability, and the first network element belongs to multiple network elements.

[0040] In conjunction with the third aspect, in one possible implementation, the first energy-saving capability includes at least one of the following: training the first model using a first energy type; or reducing energy consumption while maintaining model accuracy at a threshold not lower than a model accuracy threshold; or ensuring that the carbon emission factor during the training process of the first model is not higher than a carbon emission factor threshold; or ensuring that the normalized energy consumption during the training process of the first model is not higher than a normalized energy consumption threshold; or ensuring that the energy state during the training process of the first model is a first energy state; or training the first model based on an energy state schedule, such as training the first model at the time corresponding to a specified energy state; or ensuring that the device training the first model does not have an energy inefficiency indication; or ensuring that the energy consumption during the training process of the first model does not exceed an energy consumption threshold; or ensuring that the device training the first model meets energy efficiency level performance indicators; or ensuring that the energy efficiency level of the energy used during the training process of the first model meets the first energy efficiency level; or ensuring that the device training the first model meets the network element energy consumption limit level.

[0041] In conjunction with the third aspect, in one possible implementation, the first energy information is the network element information of each of at least one network element. The determination of the first network element includes: acquiring the network element information of each of at least one network element, the network element information including the energy consumption and / or hardware capabilities required for a single iteration; and determining the first network element based on the network element information of each of at least one network element, wherein the first network element belongs to at least one network element.

[0042] In the above method, determining the network elements to be trained based on network element information (such as the energy consumption and / or hardware capabilities required for a single iteration) can avoid the situation where the training iteration rounds or computational load are large due to the network element information, thereby causing a large amount of carbon emissions during the training process.

[0043] In conjunction with the third aspect, in one possible implementation, the method further includes: receiving fifth information sent by a first network element, the fifth information being used to indicate model training interruption and / or the reason for model training interruption; and sending a second message to a second network element, the second message being used to request training of the first model, the second message including the first information.

[0044] In this method, the fourth network element obtains the notification of the first network element's model training interruption and / or the reason for the model training interruption, which helps the fourth network element determine the next network element to train the first model and can reduce the number of training interruptions caused by the first network element.

[0045] In conjunction with the third aspect, in one possible implementation, the first information is used to indicate at least one of the following: energy type, model accuracy threshold, carbon emission factor, normalized energy consumption, energy status, energy status timetable, low energy efficiency indication, energy-saving status migration execution indication, energy consumption threshold, energy efficiency level performance index, energy efficiency level, or grid element energy consumption limit level.

[0046] Fourthly, this application provides a communication method that can be executed by a first network element. The first network element is a network device or a chip, chip system, module, or control unit of a network device, such as a server on the network side or a component (e.g., circuit, chip, or chip system) within the server. This application does not specifically limit the type of network element. It should be noted that, in this application, the term "network device" can refer to either the network device itself or the chip, functional module, or integrated circuit that performs the method provided in this application. This application does not specifically limit the type of network device.

[0047] Taking the execution of this method by the first network element as an example, the method may include: receiving a first message sent by the fourth network element, the first message being used to request training of the first model, the first message including first information, the first information being related to the energy consumption requirements corresponding to the first model; and sending the training results of the first model to the fourth network element, the training results being obtained based on the energy consumption requirements.

[0048] In the above scheme, by indicating the energy consumption requirements for training the first model through the first information (i.e., the energy consumption requirements corresponding to the first model), the energy consumption of model training can be reduced, thereby reducing the carbon emissions generated during training.

[0049] In conjunction with the fourth aspect, in one possible implementation, the method further includes: sending fifth information to the fourth network element, the fifth information being used to indicate model training interruption and / or the reason for model training interruption.

[0050] In conjunction with the fourth aspect, in one possible implementation, the first information is used to indicate at least one of the following: energy type, model accuracy threshold, carbon emission factor, normalized energy consumption, energy status, energy status timetable, low energy efficiency indication, energy-saving status migration execution indication, energy consumption threshold, energy efficiency level performance index, energy efficiency level, or network element energy consumption limit level.

[0051] Fifthly, this application provides a communication device, which may be a first network element or a chip / circuit of the first network element. The communication device is used to perform the methods in any possible implementation of the first aspect, the fourth aspect, or any of them. The communication device includes units, modules, or means having functions or units that perform the methods in any possible implementation of the first aspect, the fourth aspect, or any of them. These functions or units or means may be implemented by software, or by hardware, or by hardware executing corresponding software implementations.

[0052] Sixthly, this application provides a communication device, which may be a third network element or a chip / circuit within a third network element. The communication device is used to execute the method in the second aspect or any possible implementation thereof. The communication device includes units, modules, or means having the ability to execute the method in the second aspect or any possible implementation thereof. These functions, units, or means may be implemented in software, hardware, or hardware executing corresponding software.

[0053] In a seventh aspect, this application provides a communication device, which may be a fourth network element or a chip / circuit within a fourth network element. The communication device is used to execute the method in the third aspect or any possible implementation thereof. The communication device includes units, modules, or means having the ability to execute the method in the third aspect or any possible implementation thereof. These functions, units, or means may be implemented in software, hardware, or hardware executing corresponding software implementations.

[0054] In aspects five through seven, the aforementioned communication apparatus may include a transceiver unit and a processing unit. Further details regarding the transceiver unit and processing unit can be found in the apparatus embodiments shown below. The beneficial effects of aspects five through seven can be referenced in the relevant descriptions of aspects one through four, and will not be repeated here.

[0055] Eighthly, this application provides a communication device including a processor for executing the methods described in any possible implementation of the first aspect, the second aspect, or any of the aspects above.

[0056] Ninthly, this application provides a communication device including a processor coupled to a memory storing instructions that, when executed by the processor, cause the communication device to perform the method described in the second aspect or any possible implementation thereof.

[0057] In a tenth aspect, this application provides a communication device including a processor coupled to a memory storing instructions which, when executed by the processor, cause the communication device to perform the method described in the third aspect or any possible implementation thereof.

[0058] In one possible implementation, the communication device further includes a memory. Optionally, the processor and memory are integrated (i.e., the memory is built-in memory). Optionally, the memory and processor are independently configured (i.e., the memory is external memory).

[0059] Eleventhly, this application provides a communication device, which may include a processor and an interface circuit connected together. The interface circuit is used for exchanging (or sending / receiving or inputting / outputting) information or data, and the processor is used to execute program instructions, causing the communication device to perform the methods described in any possible implementation of the first, second, third, or fourth aspect above. The interface circuit may be a communication interface or a transceiver. The transceiver may be a radio frequency module in the communication device, or a combination of a radio frequency module and an antenna, or an input / output interface of a chip or circuit.

[0060] In a twelfth aspect, this application provides a computer-readable storage medium storing program instructions that, when executed on a computer, cause the computer to perform the method described in any possible implementation of the first aspect, or the second aspect, or the third aspect, or the fourth aspect, or any of the aspects described above.

[0061] In a thirteenth aspect, this application provides a computer program product containing program instructions that, when executed, causes the method described in any possible implementation of the first aspect, or the second aspect, or the third aspect, or the fourth aspect, or any one of them, to be performed.

[0062] In a fourteenth aspect, this application provides an apparatus, which can be implemented as a chip or as a device, including a processor. The processor is used to read and execute a program stored in a memory to perform one or more of the first, second, third, or fourth aspects described above, or one or more of any possible implementations of any aspect thereof, providing an information interaction method. Optionally, the apparatus further includes a memory connected to the processor via a circuit. Further optionally, the apparatus includes a communication interface to which the processor is connected. The communication interface is used to receive information to be processed, the processor obtains the information from the communication interface, processes the information, and outputs the processing result through the communication interface. The communication interface can be an input / output interface.

[0063] It is understood that when the communication device provided by any of the third to seventh aspects is a chip, the aforementioned sending action / function can be understood as an output, and the aforementioned receiving action / function can be understood as an input.

[0064] In one possible implementation, the processor and memory described above can be physically independent units, or the memory can be integrated with the processor.

[0065] In a fifteenth aspect, this application provides a communication system comprising a first network element and a third network element; wherein the first network element is used to perform the method described in the first aspect or any possible implementation thereof, and the third network element is used to perform the method described in the second aspect or any possible implementation thereof.

[0066] In a sixteenth aspect, this application provides a communication system comprising a first network element and a fourth network element; wherein the fourth network element is used to perform the method described in the third aspect or any possible implementation of the third aspect, and the first network element is used to perform the method described in the fourth aspect or any possible implementation of the fourth aspect.

[0067] The technical effects achieved in the above aspects can be referred to each other or to the beneficial effects in the method embodiments shown below, which will not be repeated here. Attached Figure Description

[0068] Figure 1 This is a schematic diagram of the architecture of the communication system used in the embodiments of this application;

[0069] Figure 2 This is a schematic diagram of another communication system architecture provided in an embodiment of this application;

[0070] Figure 3This is a schematic diagram of yet another communication system provided by example in the embodiments of this application;

[0071] Figure 4 This is a schematic flowchart of a model training method provided in an embodiment of this application;

[0072] Figure 5 This is a flowchart illustrating a communication method provided in an embodiment of this application;

[0073] Figure 6 This is a schematic diagram of another communication method provided in an embodiment of this application;

[0074] Figure 7 This is a flowchart illustrating another communication method provided in an embodiment of this application;

[0075] Figure 8A and Figure 8B This is a schematic diagram of another communication method provided in an embodiment of this application;

[0076] Figure 9 This is a schematic diagram of the structure of the communication device provided in an embodiment of this application;

[0077] Figure 10 This is another structural schematic diagram of the communication device provided in the embodiments of this application;

[0078] Figure 11 This is another structural schematic diagram of the communication device provided in the embodiments of this application. Detailed Implementation

[0079] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0080] In the description of this application, the terms "first," "second," etc., are used only to distinguish different objects and do not limit the quantity or order of execution, nor do they imply that they are necessarily different. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0081] In the description of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. "And / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Furthermore, "at least one item", "one or more of the following", or similar expressions refer to any combination of these items, including any combination of single or multiple items. For example, at least one of a, b, or c can represent: a, b, c; a and b; a and c; b and c; or a and b and c. Here, a, b, and c can be single or multiple.

[0082] In the description of this application, the words "exemplary" or "for example" are used to indicate that something is an example, illustration, or illustration. Any embodiment or design that is described as "exemplary," "for example," or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Rather, the use of the words "exemplary," "for example," or "for example" is intended to present the relevant concepts in a specific manner.

[0083] It is understood that in the description of this application, "when," "if," and "if" all refer to the device performing a corresponding action under certain objective circumstances, and are not time-limited, nor do they require the device to perform a judgment action when it is implemented, nor do they imply any other limitations. The device performing a corresponding action under certain objective circumstances includes: satisfying the objective circumstances, i.e., being able to perform the corresponding action; or satisfying both the objective circumstances and other circumstances, in order to perform the corresponding action.

[0084] In this application, the use of singular designations for elements is intended to represent "one or more" rather than "one and only one," unless otherwise specified.

[0085] In addition, the terms “system” and “network” are often used interchangeably in this article.

[0086] It is understood that in the various embodiments of this application, expressions such as "A corresponds to B", "A and B correspond", "A corresponds to B" or similar expressions indicate that B is associated with A, and B can be determined based on A. Determining B based on A does not mean that B is determined solely based on A; B can also be determined based on A and / or other information.

[0087] The embodiments of this application can be applied to various communication systems, such as: Long Term Evolution (LTE) systems, LTE Frequency Division Duplex (FDD) systems, LTE Time Division Duplex (TDD) systems, 5G systems, or New Radio (NR), or to future communication systems or other similar communication systems, or Ultra Wide Band (UWB) systems, or Wireless Fidelity (WiFi) systems. The technical solutions provided in this application can also be applied to Internet of Things (IoT) systems, Narrow Band Internet of Things (NB-IoT) systems, etc.

[0088] Please see Figure 1 , Figure 1 This is a schematic diagram of the architecture of the communication system used in the embodiments of this application. It should be noted that... Figure 1 This is a schematic diagram of one possible, non-limiting system. For example... Figure 1 As shown, the communication system 10 includes a radio access network (RAN) 100 and a core network (CN) 200. Optionally, the communication system 10 may also include the Internet 300.

[0089] RAN 100 includes at least one RAN node (such as...) Figure 1 110a and 110b, collectively referred to as 110) and at least one terminal (such as Figure 1 RAN 100, denoted as RAN 120a-120j, is collectively referred to as RAN 120. RAN 100 may also include other RAN nodes, such as wireless relay equipment and / or wireless backhaul equipment. Figure 1 (Not shown in the image). Terminal 120 is connected to RAN node 110 wirelessly. RAN node 110 is connected to core network 200 wirelessly or via wired connection. The core network elements in core network 200 and RAN node 110 in RAN 100 can be different physical devices, or they can be the same physical device integrating core network logical functions and radio access network logical functions, or they can be a single physical device integrating some core network element functions and some RAN node 110 functions. Terminals can be interconnected with each other, and RAN nodes 110 can be interconnected with each other via wired or wireless connection. Figure 1This is just a schematic diagram. The communication system may also include other network devices, such as wireless repeaters and wireless backhaul devices. Each device may also contain different functional units. Figure 1 It is not shown in the middle.

[0090] RAN 100 can be a 3GPP-related cellular system, such as a 4G, 5G mobile communication system, or a future-oriented evolution system. RAN 100 can also be an open access network (O-RAN or ORAN), a cloud radio access network (CRAN), or a wireless fidelity (WiFi) system. RAN 100 can also be a communication system that integrates two or more of the above systems.

[0091] RAN node 110 can be a device or module that is connected to the aforementioned communication system and has corresponding communication and / or sensing functions. Alternatively, RAN node 110 can be a device with communication and / or sensing function modules. RAN node 110 typically contains communication modules, circuits, or chips that perform corresponding communication and / or sensing functions, and it is also configured with program instructions for performing these functions.

[0092] RAN node 110, sometimes also referred to as network equipment, radio access network equipment, access network device, RAN entity, or access node, constitutes part of the communication system and is used to help terminals achieve wireless access. Multiple RAN nodes 110 in the communication system 10 can be of the same type or different types. In some scenarios, the roles of RAN node 110 and terminal 120 are relative, for example... Figure 1 Network element 120i can be a helicopter or a drone, and it can be configured as a mobile base station. For terminals 120j that access RAN 100 through network element 120i, network element 120i is a base station; however, for base station 110a, network element 120i is a terminal. RAN node 110 and terminal 120 are sometimes referred to as communication devices, for example... Figure 1 Network elements 110a and 110b can be understood as communication devices with base station functions, while network elements 120a-120j can be understood as communication devices with terminal functions.

[0093] In one possible scenario, RAN node 110 can be a base station, an evolved NodeB (eNodeB), an access point (AP), a transmission reception point (TRP), a next-generation NodeB (gNB), a base station in a future mobile communication system, or an access node in a WiFi system, etc. Figure 1 110a), micro base stations or indoor stations (such as Figure 1 The RAN node 110 can be a relay node or donor node, or a wireless controller in a CRAN scenario. Optionally, the RAN node 110 can also be a server, wearable device, vehicle, or in-vehicle equipment. For example, the access network device in vehicle-to-everything (V2X) technology can be a roadside unit (RSU). All or part of the functions of the RAN node 110 in this application can also be implemented through software functions running on hardware, or through virtualization functions instantiated on a platform (e.g., a cloud platform). The RAN node 110 in this application can also be a logical node, logical module, or software capable of implementing all or part of the functions of the RAN node 110.

[0094] In another possible scenario, multiple RAN nodes 110 collaborate to assist the terminal in achieving wireless access, with each RAN node 110 implementing a portion of the base station's functions. For example, a RAN node 110 can be a centralized unit (CU), a distributed unit (DU), a CU-control plane (CP), a CU-user plane (UP), or a radio unit (RU), etc. CUs and DUs can be configured separately or included in the same network element, such as a baseband unit (BBU). RUs can be included in radio equipment or radio units, such as remote radio units (RRUs), active antenna units (AAUs), or remote radio heads (RRHs).

[0095] In different systems, CU (or CU-CP and CU-UP), DU, or RU may have different names, but those skilled in the art will understand their meaning. For example, in an ORAN system, CU can also be called O-CU (open CU), DU can also be called O-DU, CU-CP can also be called O-CU-CP, CU-UP can also be called O-CU-UP, and RU can also be called O-RU. For ease of description, this application uses CU, CU-CP, CU-UP, DU, and RU as examples. Any of the units among CU (or CU-CP, CU-UP), DU, and RU in this application can be implemented through software modules, hardware modules, or a combination of software and hardware modules.

[0096] A terminal can be a device or module that accesses the aforementioned communication system and possesses corresponding communication and / or sensing functions. Alternatively, a terminal can be a device with communication and / or sensing function modules. A terminal typically contains communication modules, circuits, or chips that perform the corresponding communication and / or sensing functions, and it also contains program instructions for performing these functions. A terminal can also be referred to as a terminal device, user equipment (UE), user device, access terminal, user unit, user station, mobile station, mobile station (MS), remote station, remote terminal, mobile terminal, mobile device, user terminal, terminal unit, terminal station, terminal device, wireless communication equipment, user agent, or user device, etc. A terminal typically contains communication modules, circuits, or chips that perform the corresponding communication functions. It also contains program instructions for performing these functions. Terminals can be widely used in various scenarios, such as device-to-device (D2D), vehicle-to-everything (V2X) communication, machine-type communication (MTC), Internet of Things (IoT), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grids, smart furniture, smart offices, smart wearables, smart transportation, and smart cities. Terminals can be mobile phones, tablets, computers with wireless transceiver capabilities, wearable devices, vehicles, drones, helicopters, airplanes, ships, robots, robotic arms, smart home devices, transportation vehicles with wireless communication capabilities, communication modules, and roadside units (RSUs) with terminal functions. The embodiments of this application do not limit the device form of the terminal.

[0097] See Figure 2 , Figure 2This is a schematic diagram of another communication system architecture provided in an embodiment of this application. For example... Figure 2 As shown, the communication system architecture 20 includes, but is not limited to: UE, RAN, core network, data network (DN), and application function (AF). The access network can be used to implement functions related to radio access.

[0098] The core network may include various network functions (NFs) or network elements, such as all or some of the following logical functions: user plane function (UPF), network exposure function (NEF), network repository function (NRF), policy control function (PCF), unified data management (UDM), unified data repository (UDR), network data analytics function (NWDAF), authentication server function (AUSF), access and mobility management function (AMF), or session management function (SMF), etc.

[0099] In one possible implementation, the UE can access the data network by establishing a session between the UE and the RAN, then to the UPF, and finally to the DN, i.e., a protocol data unit (PDU) session. Optionally, the service server or information source in this application can be a relevant device in the DN.

[0100] The AMF (Advanced Mobility Management) is primarily responsible for mobility management in mobile networks, such as user location updates, network registration, and handover. Additionally, the AMF is responsible for transmitting user policies between terminal devices and the PCF (Programming Provider Function). The SMF (Supply Provider Function) is primarily responsible for session management in mobile networks, such as session establishment, modification, and release. Specific functions include assigning Internet Protocol (IP) addresses to users and selecting the UPF (User Packet Forwarder) to provide packet forwarding. The PCF provides policies to the AMF and SMF, such as Quality of Service (QoS) policies and slice selection policies. The UPF (User Packet Processor) is primarily responsible for processing user packets, such as forwarding and accounting.

[0101] It should be noted that the name AMF is merely an example. The name of AMF may change as the communication system evolves. Any functional network element with a name similar to AMF can be understood as the AMF of this application. For example, AMF can also be called a mobility management network element or mobility management function, etc., and this application does not limit its specific application. The name UPF may change as the communication system evolves. Any functional network element with a name similar to UPF can be understood as the UPF of this application. For example, UPF can also be called a user plane network element or user plane management network element, etc., and this application does not limit its specific application.

[0102] NWDAF has functions such as data collection, model training, data analysis, and model inference. NWDAF supports collecting data from other network functions and AFs, collecting data from operation administration and maintenance (OAM) network elements, and providing analytical information to other network functions and AFs.

[0103] NWDAF can be used to collect relevant data from network functions, third-party service servers, terminal devices, or network management systems. Based on this data, it performs data analysis to obtain results, which are then provided to these network functions, third-party service servers, terminal devices, or network management systems. These results can assist the network in selecting service quality parameters, performing traffic routing, or choosing background data transmission strategies. Furthermore, NWDAF can also collect relevant data from network functions, third-party service servers, terminal devices, or network management systems, and use this data to train models to obtain artificial intelligence (AI) or machine learning (ML) models. These AI / ML models can then be provided to other NWDAFs. These AI / ML models can assist NWDAFs in generating data analysis results based on relevant data.

[0104] For example, an NWDAF may only support model training, only support data inference, or support both model training and data inference. An NWDAF that supports model training can also be called a training NWDAF, or an NWDAF containing model training logical function (MTLF) (abbreviated as MTLF), or simply an NWDAF.

[0105] In this application, the NWDAF can be trained using acquired data to obtain a trained model. An NWDAF supporting data inference can also be called an inference NWDAF, or an NWDAF containing analytics logical function (AnLF) (abbreviated as AnLF). An inference NWDAF can input input data into the trained model to obtain analysis results or inference data. In this embodiment, a training NWDAF refers to an NWDAF that at least supports model training. As a possible implementation, a training NWDAF can also support data inference. An inference NWDAF refers to an NWDAF that at least supports data inference. As a possible implementation, an inference NWDAF can also support model training. If an NWDAF supports both model training and data inference, it can be called a training NWDAF, an inference NWDAF, a training-inference NWDAF, or simply an NWDAF. In various embodiments of this application, an NWDAF can be a separate network element or co-located with other network elements; this application is not limited to these limitations.

[0106] The NWDAF in this application can simultaneously support model training and / or model (data) inference functions.

[0107] UDM network elements can be responsible for the storage, management, and access control of user data, such as user identification, access authorization, registration, mobility management, subscription management, and SMS management. UDR network elements support the storage and retrieval of multiple network functions (NFs).

[0108] NRF provides registration and discovery functions, enabling network functions (NFs) to discover each other and communicate through an application programming interface (API). NRF can also provide other network element management services, such as network element updates, deregistration, and network element status subscription and push.

[0109] NEF can be used to support the opening of capabilities and events. For example, it can open the capabilities of various network functions (NFs) and transform internal and external information. NEF can provide security guarantees to ensure the security of external applications to the 3GPP network, and provide functions such as opening up the ability to customize the quality of service (QoS) of external applications, subscription to mobility state events, and distribution of AF requests.

[0110] Application Providers (AFs) can be used to convey application-side requests to the network side, such as Quality of Service (QoS) requirements or user state event subscriptions. An AF can be a third-party functional entity or an application server deployed by an operator; this application's embodiments are not limited to these.

[0111] It is understood that the above network elements are examples of one implementation method. This application does not exclude the possibility that in future wireless communication systems, there may be network elements or devices with the above network element functions that have other names or other forms.

[0112] Understandable. Figure 2 The various network functions shown can be referenced in relevant protocols or standards, and will not be elaborated upon in this application. It is understood that... Figure 2 In this context, “Nnef”, “Nnrf”, “Npcf”, “Nudm”, “Nudr”, “Nnwdaf”, “Naf”, “Nausf”, “Namf”, and “Nsmf” represent the names of service interfaces, which can be used to call the corresponding service operations. For details, please refer to the relevant descriptions in the 3GPP standard protocol, which will not be elaborated here.

[0113] It should also be understood that Figure 2N1, N2, N3, N4, and N6, as shown, are all interface sequence numbers. The meanings of these interface sequence numbers are as follows: N1 can represent the interface between the AMF and the terminal equipment, used to transmit non-access stratum (NAS) signaling (such as QoS rules from the AMF) to the terminal equipment. N2 can represent the interface between the AMF and the RAN, used to transmit radio bearer control information from the core network side to the RAN. N3 can represent the interface between the RAN and the UPF, mainly used to transmit uplink and downlink user plane data between the RAN and the UPF. N4 can represent the interface between the SMF and the UPF, used to transmit information between the control plane and the user plane, including the distribution of forwarding rules, QoS rules, traffic statistics rules, etc., from the control plane to the user plane, as well as the reporting of user plane information. N6 can represent the interface between the UPF and the DN, used to transmit uplink and downlink user data streams between the UPF and the DN. For example, the meanings of the above interface sequence numbers can also be found in the definitions in the 3GPP standard protocols; this application does not limit the meanings of the above interface sequence numbers.

[0114] In one possible implementation, the network element or function in this application embodiment can be a network component in a hardware device, a software function running on dedicated hardware, or a virtualization function instantiated on a platform (e.g., a cloud platform). As one possible implementation, the network element or function in this application embodiment can be implemented by one device, multiple devices working together, or a functional module within a single device; this application embodiment does not specifically limit this.

[0115] In this application, the network element can also be replaced by NF, node, device, or equipment, etc., and this application does not limit it.

[0116] Figure 3 This is a schematic diagram of another communication system provided by example in an embodiment of this application. The communication system 30 includes a first network element, a second network element, a third network element, and a fourth network element.

[0117] For example, the first network element, the second network element, and the third network element can be network devices, such as core network devices; the fourth network element can be a network device or a terminal.

[0118] For example, the first and second network elements can be the MTLF described above, and the third network element can be... Figure 2 In the NRF, the fourth network element can be Figure 2The NF in the middle can be PCF, AF, AMF, SMF, or UPF, or it can be RAN or OAM, or it can be UE. Optionally, the fourth network element can also be called NWDAF service consumer (NWDAF consumer or consumer for short).

[0119] For example, the first network element, second network element, third network element, and fourth network element in this application can be referred to the above. Figure 1 Regarding the core network 200 or Figure 2 The description of the relevant network elements of the core network 200 in the above document is provided. For the terminal in this application, please refer to the above description. Figure 1 Regarding terminal 120 or Figure 2 The relevant descriptions of UE in the text will not be repeated hereafter.

[0120] Optionally, Figure 3 The communication system shown may also include more or fewer network elements. For example, the communication system may include only the first network element and the second network element, or the communication system may include only the first network element and the fourth network element, or the communication system may also include a fifth network element, which is used to participate in the training of the first model.

[0121] In this application, "sending information" can be understood as one device sending information to another device, or it can also be understood as one logical module within a device sending information to another logical module. For example, "base station sending information" can be understood as the base station sending information to another device (such as a terminal), or it can be understood as logical module 1 in the base station sending information to logical module 2 in the base station.

[0122] In this application, "receiving information" can be understood as one device receiving information from another device, or it can also be understood as a logical module within a device receiving information from another logical module. For example, "base station receiving information" can be understood as the base station receiving information from another device (such as a terminal), or it can be understood as logical module 1 in the base station receiving information from logical module 2 in the base station.

[0123] The communication between different devices involved in this application can refer to direct communication between different devices (i.e., without the need for relaying or forwarding by other devices), or communication between different devices through other devices (i.e., requiring relaying or forwarding by other devices), or communication between a functional unit within a device and other devices through another functional unit. In other words, "sending information to… (e.g., a terminal)" or the relevant illustrations in the accompanying drawings can be understood as the destination of the information being the terminal. This can include sending information directly or indirectly to the terminal. "Receiving information from… (e.g., a terminal)" or "receiving information from… (e.g., a terminal)" or "receiving information sent (e.g., by a terminal)" or the relevant illustrations in the accompanying drawings can be understood as the source of the information being the terminal. This can include receiving information directly or indirectly from the terminal. Information may undergo necessary processing between the source and destination, such as format changes, analog-to-digital conversion, amplification, filtering, etc., but the destination can understand the valid information from the source. Similar expressions in this application can be understood in a similar way, and will not be elaborated further here.

[0124] Currently, MTLF's model training based on NWDAF consumer requests often results in significant energy consumption and carbon emissions. For example, Figure 4 This is a flowchart illustrating a model training method provided in an embodiment of this application. The method includes the following steps:

[0125] Step S401: The NWDAF consumer sends a request message to the MTLF.

[0126] Correspondingly, MTLF receives request messages sent by the NWDAF consumer.

[0127] This request message is used to subscribe to or unsubscribe from model training.

[0128] For example, the request message can be an Nnwdaf_MLModelTraining_Subscribe / Nnwdaf_MLModelTraining_Unsubscribe message. This message can include one or more of the following:

[0129] Analysis ID; ML interoperability indication information, which is per-operator information, such as model file format, model runtime environment, etc. Interoperability refers to interoperability between different operators. Because interoperability requires information encoding, and the encoding format and other information are independent between different operators, operators negotiate for sharing purposes; Notification address, used to indicate who should receive the model training results, only included in the subscription request; Model information, such as model file address or analysis data repository function (ADRF) ID, which is per-ML model ID; Model file; Model ID; ML model accuracy check flag, which the MTLF receiving this flag uses to test the trained model on the local test set to calculate model accuracy; ML model correlation flag. ID); Training filter information, used to indicate the range of the training set, such as information about a slice; Training report target, used to request training data from the ML model; Use case context; Training report information, such as maximum response time; Iteration round information, used to indicate the current training iteration round of the ML model; Expiration time, used to indicate the time when the subscription expires.

[0130] Step S402: MTLF performs model training.

[0131] Specifically, after receiving the above request message, MTLF determines whether the model training requires data collection or whether the corresponding data is stored locally. After preparing the data, it starts model training.

[0132] Step S403: After MTLF completes model training, it sends a notification message to the NWDAF consumer. This notification message is used to notify the model of information.

[0133] Specifically, after MTLF completes model training, it sends a notification message to the NWDAF consumer. This notification message is used to inform the model information. The notification message can be Nnwdaf_MLModelTraining_Notify, and it includes: ML model information; ML model identifier; global model accuracy information, used to indicate the model accuracy obtained by the federated learning client (FL client) using the local training set; and delay event notification, including delay event indication, reason code, expected time to complete training, etc.

[0134] In the above model training method, MTLF starts model training directly based on NWDAFconsumer request. This training process often results in a large amount of carbon emissions due to MTLF's energy situation or the model's energy consumption.

[0135] In view of this, this application proposes a communication method and apparatus. In this method, a first network element, based on its own energy information, stops training a first model; and sends first information to a second network element requesting the second network element to train the first model. The second network element is determined based on its own energy information. This method considers its own energy information and the energy information of other network elements during training, thus avoiding issues such as… Figure 4 The model training method shown is based solely on model training requests, allowing all models to be trained without considering the increased operating costs and carbon emissions caused by energy information, thus reducing the carbon emissions generated by model training.

[0136] Based on the above system architecture, a communication method provided by an embodiment of this application will be described below.

[0137] Figure 5 This is a flowchart illustrating a communication method provided in an embodiment of this application.

[0138] In this embodiment, the steps performed by the first network element can also be performed by a module of the first network element (e.g., a chip, chip system, integrated circuit, or control unit); in this embodiment, the steps performed by the second communication device can also be performed by a module of the second communication device (e.g., a chip, chip system, integrated circuit, or control unit).

[0139] Figure 5 Some steps are optional, for example, Figure 5 It may include only steps S501 and S504, or, Figure 5 It includes only steps S502 and S503.

[0140] like Figure 5 As shown, the communication method may include some or all of the following steps:

[0141] Step S501: The first network element stops training the first model based on the energy information of the first network element.

[0142] In this application, energy information may refer to energy-related information, such as energy status, energy type, or energy information level, or it may refer to energy consumption information, such as normalized energy consumption or energy efficiency.

[0143] Optionally, the first network element can acquire its own energy information (i.e., the energy information of the first network element), or in other words, the first network element can sense changes in its own energy information, and then, based on its own energy information, determine whether to stop training. This application does not limit the method by which the first network element acquires its energy information.

[0144] For example, the energy information of the first network element includes at least one of the following: the energy efficiency of the first network element, the normalized energy consumption of the first network element, the energy status of the first network element, the energy consumption of the first network element within a preset time, the new energy ratio of the first network element, the carbon emission factor of the first network element, the energy type of the first network element, the energy efficiency level of the first network element, or the energy consumption of the first network element, or the NF profile attributes of each energy status.

[0145] In one possible implementation, if the energy information of the first network element meets a first condition, the first network element stops training the first model. The first condition includes at least one of the following: the energy efficiency of the first network element is lower than a first energy efficiency threshold; the normalized energy consumption of the first network element exceeds a first normalized energy consumption threshold; the first network element enters an energy-saving state; the energy consumption of the first network element exceeds a first energy consumption threshold within a preset time period, such as when the energy consumption of the first network element exceeds the energy consumption limit in the first network element's NF profile within that time window; the renewable energy ratio of the first network element is lower than a first renewable energy ratio threshold; the carbon emission factor of the first network element is higher than a first carbon emission factor threshold; the energy type of the first network element changes, such as switching from green energy to traditional energy; the energy efficiency level of the first network element is greater than a first energy efficiency level threshold; or, the energy consumption of the first network element exceeds the energy consumption limit level of the first network element. Optionally, the first condition is pre-configured or obtained from other network elements. For example, a fourth network element (such as an NWDAF consumer) can send the first condition to the first network element (such as MTLF#1).

[0146] For example, the first network element's decision to stop model training based on the first condition can be based on pre-configuration, such as when the operator signs a green energy agreement or the operator wants to control the first network element's carbon emissions and therefore restricts the first network element's energy information (such as carbon emissions, energy types, etc.), such as when the operator configures the first network element to stop model training when the first condition is met; or, the fourth network element (such as an NWDAF consumer) can send the first condition to the first network element (such as MTLF#1), and the fourth network element instructs the first network element to stop model training when the first condition is met.

[0147] For example, the fourth network element can send a third message to the first network element, the third message including the first condition. Optionally, the third message is used to request model training, and the third message can also be called a model training request; or, the third message is used to request that model training be stopped when the first condition is met.

[0148] In this application, the new energy ratio indicates the proportion of renewable energy consumption in the total energy consumption of the network element, or it can be the type of energy consumed by the network element (e.g., renewable or non-renewable); the carbon emission factor (CEF) indicates the carbon emission relative to resource consumption, which is a relative value or a level (low, medium, high). The CEF can also indicate the relative carbon efficiency level of the network element, such as low, medium, and high indicators, or level values ​​of 1, 2, 3, 4, and 5; normalized energy consumption can refer to energy consumption relative to resource consumption; energy status refers to the energy status of the network element, such as energy-saving status, non-energy-saving status, etc.; the energy status timetable can be a timetable of network element status, such as the time information for entering a certain status; the low energy efficiency indication indicates whether the network element is facing excessive energy consumption, where excessive energy consumption means that the energy consumption exceeds a reasonable range, such as a preset range; the energy-saving status transition execution indication indicates whether the network element can perform different actions based on different energy information, such as stopping the network in the energy-saving status. The network element's services are designed to save energy; energy consumption indicates the energy consumption of a network element within a specified time window; energy efficiency (or energy efficiency) level performance indicators indicate how the network element's energy efficiency level is calculated, such as UPF (total forwarding bit rate), SMF (number of PDU sessions served by the SMF), and AMF (number of UEs served); the energy efficiency level is the energy efficiency level derived by comparing the energy efficiency level performance indicators with the energy consumption of the network element in the execution environment; the network element energy consumption limit level restricts the allowed energy consumption level of the network element within a time window, which can be a specific energy consumption amount or a range of energy consumption; the NF profile attributes for each energy state refer to the configuration file attributes that the NF pre-registers in the NF profile, enabling these configuration files to apply different values ​​according to the energy state, or determining whether to apply them. For example, for MTLF, the NF profile attributes for each energy state indicate that the NF does not perform model training functions in the energy-saving state, or only supports training a limited number of models simultaneously in the energy-saving state.

[0149] Optionally, before step S501, the method may further include: the fourth network element requesting the first network element to perform model training. For example, the fourth network element sends a model training request to the first network element, and correspondingly, the first network element receives the model training request sent by the fourth network element, which is used to request training a first model; the first network element trains the first model, and the first network element sends a response message corresponding to the model training request to the fourth network element. For example, this step can be seen in the description of S601 below.

[0150] Step S502: The first network element sends a first message to the third network element. The first message is used to request the determination of the second network element.

[0151] Correspondingly, the third network element receives the first message sent by the first network element.

[0152] For example, the first message includes third information, which is used to indicate a second condition. The energy information of the second network element meets the second condition, which includes at least one of the following: the energy efficiency of the second network element is lower than a second energy efficiency threshold; the normalized energy consumption of the second network element exceeds a second normalized energy consumption threshold; the second network element enters an energy-saving state; the energy consumption of the second network element exceeds a second energy consumption threshold within a preset time; the new energy ratio of the second network element is lower than a second new energy ratio threshold; the carbon emission factor of the second network element is higher than a second carbon emission factor threshold; the energy type of the second network element is a first energy type; the energy efficiency level of the second network element is greater than a second energy efficiency level threshold; or, the energy consumption of the second network element exceeds a second network element energy consumption limit level.

[0153] Step S503: The third network element sends third information to the first network element. The third information is used to instruct the second network element, which is determined based on the energy information of the second network element.

[0154] Correspondingly, the first network element receives the third information sent by the third network element.

[0155] For example, the third network element can determine the aforementioned second network element based on the energy information of one or more network elements (such as MTLFs) and the first message. For instance, the third network element can obtain the NF profiles registered by one or more MTLFs (such as MTLF#1 and MTLF#2), wherein the NF profiles registered by the MTLFs include the energy information of the MTLFs. For example, MTLF#1 and MTLF#2 have pre-registered NF profiles with energy information to the third network element; based on the NF profiles registered by the MTLFs, the third network element determines the MTLF instance information (i.e., the instance information of the second network element) of the network element that meets the request of the first message, thus determining the second network element.

[0156] For example, the first message is used to request network elements that meet the second condition. Suppose the second condition is that the energy efficiency of the second network element is lower than the second energy efficiency threshold. MTLF#1 and MTLF#2 pre-register NF profiles with energy information to the third network element. The energy information in the NF profile includes information indicating the energy efficiency of the network element. Then, the third network element can determine whether the energy efficiency of MTLF#1 and MTLF#2 is lower than the second energy threshold based on the NF profile. If the energy efficiency of MTLF#1 is lower than the second energy threshold and the energy efficiency of MTLF#2 is not lower than the second energy threshold, then the third network element can determine MTLF#1 as the second network element.

[0157] For example, the first message is used to request network elements that meet the second condition. Suppose the second condition is that the energy type of the second network element is the first energy type (such as green energy). MTLF#1 and MTLF#2 pre-register NF profiles with energy information to the third network element. The energy information in the NF profile includes information indicating the energy type supported by the network element. Then, the third network element can determine whether MTLF#1 and MTLF#2 support the first energy type based on the NF profile. If MTLF#1 supports the first energy type and MTLF#2 does not support the first energy type, then the third network element can determine that MTLF#1 is the second network element.

[0158] Step S504: The first network element sends first information to the second network element. The first information is used to request the second network element to train the first model. The second network element is determined based on the energy information of the second network element.

[0159] For example, the first information includes model information of the first model, which is used for model training of the first model, wherein the training of the first model is not completed before training of the first model is stopped, and the model information of the first model is used to indicate the first model that has not completed training.

[0160] For example, the model information of the first model is used to continue training the first model, including at least one of the following: the analysis identifier corresponding to the first model; the identifier of the first model; the model address of the first model; model parameters, such as the model weights and / or biases; loss function value; optimizer state, such as at least one of learning rate, momentum or cumulative gradient; training progress information, such as training epoch or iteration, random seed, training log, and at least one of intermediate outputs of the model.

[0161] Optionally, the first information used to request the second network element to train the first model can mean that the first information is carried in a second message, and the second message is used to request model training. That is, the first network element can send a second message to the second network element, and the second message is used to request model training. The second message includes the first information.

[0162] For example, the second message includes at least one of the following: analysis identifier, first information (such as model file), model identifier, or fourth information (such as model report address).

[0163] The analysis identifier is used to indicate the analysis corresponding to the model. For example, when the fourth network element (such as the NWDAF consumer) is AnLF, AnLF requests MTLF to train the model required for a certain analysis, and can send the analysis identifier to indicate the model. The model identifier indicates the model that the first network element has not finished training. The model file is used to indicate the relevant information of the model that the first network element has not finished training, including at least one of the following: the weights indicating the current learning state of the model, the optimizer state (such as learning rate, momentum, gradient, etc.) indicating the current training stage of the optimizer, or training progress information (training rounds, current iteration steps, etc.). The model report address is used to indicate the object of the model report obtained from training. For example, the address can be the address of the first network element or the address of the model consumer, i.e., the fourth network element.

[0164] Optionally, the second message may further include fourth information, which is used to instruct the second network element to notify the recipient of the training results of the model training. Then, after the second network element trains the first model based on the first message and obtains the training results, it can send the training results to the notification recipient. For example, the notification recipient can be either the first network element or the fourth network element; that is, the aforementioned fourth information is used to instruct either the first network element or the fourth network element. For instance, if the fourth information can be the address of either the first or the fourth network element, then after the second network element trains the first model and obtains the training results, it can send the training results to either the first or the fourth network element. Exemplarily, the fourth information can be included in the second message, or it can be included in a new message after the second network element responds to the first network element, indicating that model training can continue.

[0165] For example, the second message may be a subscription message or other message, and this application does not limit it.

[0166] For example, the fourth information can be the instance identifier or instance address of the notification object (such as the first network element or the fourth network element), such as an IP address or a fully qualified domain name (FQDN).

[0167] Optionally, the first and fourth messages may be sent via different messages, and this application does not limit this.

[0168] above Figure 5 The method embodiments shown include many possible implementation schemes, which will be discussed below. Figure 6 Some of the implementation schemes will be illustrated with examples. It should be noted that... Figure 6 For any unexplained related concepts, operations, or logical relationships, please refer to [link / reference]. Figure 5 The corresponding description in the illustrated embodiment.

[0169] In this application, Figure 6 The illustrated embodiment can be considered as a standalone embodiment. Figure 6 The embodiments shown can all be implemented without relying on Figure 5 The technical solution; Figure 6 Some of the steps in the illustrated embodiments can also be used as separate embodiments.

[0170] For example, the fourth network element can be an NWDAF consumer, the third network element can be an NRF, and the first and second network elements can be two MTLFs. For ease of description, the first network element is referred to as MTLF#1 and the second network element is referred to as MTLF#2.

[0171] Figure 6 This is a schematic diagram of another communication method provided in an embodiment of this application. The method includes some or all of the following steps:

[0172] Step S601: The fourth network element requests the first network element to perform model training.

[0173] For example, this step may include: the fourth network element sending a model training request to the first network element; correspondingly, the first network element receiving the model training request sent by the fourth network element, the model training request being used to request training a first model; the first network element training the first model; and the first network element sending a response message corresponding to the model training request to the fourth network element. For a detailed example, see [link to example]. Figure 4 The relevant description in the document.

[0174] Optionally, the fourth network element can also send a first instruction message to the first network element, which is used to instruct the reduction of carbon emissions during model training. For example, the model training request sent by the fourth network element to the first network element carries this first instruction message.

[0175] Optionally, the NWDAF profile (or NF profile in this application) of each NWDAF (including the first network element) may include energy-related information (such as energy-related information of the first network element), which may be pre-configured. For example, the operator may configure energy-related information into the NWDAF profile of each NWDAF (including the first network element). The energy-related information includes one or more of the following: renewable energy ratio; carbon emission factor; energy type; normalized energy consumption; energy state schedule; low energy efficiency indication; energy saving state transition execution indication; energy consumption; energy efficiency (energy efficiency) level performance indicators; energy efficiency level; and NF energy consumption limit level.

[0176] For example, energy-related information in the NWDAF profile of the first network element is used to indicate the energy information supported by the first network element, such as the NWDAF of the first network element. The profile includes the following parameters: New Energy Ratio (indicating the ratio of new energy sources supported by the first grid element); Carbon Emission Factor (indicating the carbon emission factor supported by the first grid element); Normalized Energy Consumption (indicating the normalized energy consumption supported by the first grid element); Energy Status (indicating the energy status supported by the first grid element); Energy Status Timetable (which can be a timetable of the first grid element's energy status, such as the time information when the first grid element enters a certain energy status); Low Energy Efficiency Indicator (indicating whether the first grid element faces excessive energy consumption); Energy Saving State Transition Execution Indicator (indicating whether the first grid element can perform different actions based on different energy information, such as stopping the first grid element's service in an energy saving state to save energy); Energy Consumption (indicating the energy consumption of the first grid element within a specified time window); Energy Efficiency Level Performance Index (indicating how to calculate the energy efficiency level of the first grid element); Energy Efficiency Level (indicating the energy efficiency level supported by the first grid element); and NF Energy Consumption Limit Level (limiting the allowed energy consumption level of the first grid element within a time window; the limit level can be a specific energy consumption amount or a range of energy consumption amounts). It should be noted that the above "support" can also be understood as current support. For example, the new energy ratio supported by the first grid element can be understood as the new energy ratio currently supported by the first grid element.

[0177] Step S602: The first network element stops model training based on the energy information of the first network element.

[0178] For example, if the energy information of the first network element meets the first condition during model training, the model training is stopped.

[0179] For example, the first condition may refer to a change in energy state or an increase in carbon emissions. This change in energy state or increase in carbon emissions includes at least one of the following: the first grid element needs to be shut down due to energy consumption constraints, for example, energy consumption constraints include low energy efficiency (e.g., energy efficiency below the energy efficiency forecast), or the first grid element needs to be shut down when it enters an energy-saving state, or the energy consumption of the first grid element exceeds the energy consumption constraint in the first grid element's NF profile within the time window; or, the first grid element switches its energy type, such as switching from green energy to traditional energy, or the carbon emission factor exceeds a first carbon emission factor threshold, or the new energy ratio exceeds a first new energy ratio threshold, etc.

[0180] Optionally, the decision of the first network element to stop model training may be based on pre-configuration, such as when the operator signs a green energy agreement or when the operator wants to control the carbon emissions of the first network element and therefore restricts the energy information of the first network element (such as carbon emissions, energy types, etc.), such as when the operator configures the first network element to stop model training when the first condition is met; or, based on the first indication information in step S601, such as the first indication information used to indicate the first condition.

[0181] Step S603: The first network element sends a first message to the third network element, which is used to request MTLF discovery.

[0182] For example, the first message can also be called an MTLF discovery request message or an NWDAF discovery request (NWDAFdiscovery), which is used to perform MTLF discovery. For instance, if the first network element determines that it cannot continue model training, it can initiate MTLF discovery, that is, send this message to the third network element.

[0183] For example, the message may include model-related information, such as, but not limited to: analysis identifiers or model data; energy-related information indicators, which are used to indicate target energy information, that is, the MTLF that meets the target energy information belongs to the network element requested by the first message (for convenience, referred to as the target network element). The target energy information may include at least one of the following: energy type indicator, MTLF energy state indicator, MTLF carbon emission level, or MTLF energy efficiency level, etc., wherein the energy type indicator is used to indicate the target energy type, and the target network element supports the target energy type or supports model training using the energy corresponding to the target energy type; the MTLF energy state indicator (NF state indicator) is used to indicate the target energy state, and the target network element meets the target energy state or supports model training under the target energy state; the target network element meets the MTLF carbon emission level or supports model training under the MTLF carbon emission level; or, the target network element meets the MTLF energy efficiency level or supports model training under the MTLF energy efficiency level.

[0184] Among them, energy-related information is an example of the aforementioned third type of information.

[0185] Step S604: The third network element sends a response message to the first network element. The response message includes an instance identifier of at least one MTLF and / or an instance address of at least one MTLF, where at least one MTLF includes the second network element.

[0186] Accordingly, the first network element receives the response message sent by the third network element. For example, this response message can be called an NWDAF discovery response.

[0187] The response message includes an instance identifier and / or an instance address of at least one MTLF, and the at least one MTLF includes a second network element, indicating that the response message includes an instance identifier and / or an instance address of the second network element. The instance identifier and / or instance address of the second network element is an example of the aforementioned second message.

[0188] Among them, at least one MTLF mentioned above can also be referred to as the network element requesting the first message (for ease of description, it is referred to as the target network element).

[0189] For example, a third network element can determine the target network element based on the energy information of one or more MTLFs and the first message. For instance, the third network element can obtain the NF profiles registered by one or more MTLFs, which include the energy information of the MTLF. Based on the analysis identifier and energy-related information indication in the first message, the third network element can obtain the MTLF instance information (i.e., the instance information of the target network element) of the network element that meets the request of the first message.

[0190] Optionally, if the target network element is a single network element, then the second network element is that network element; if the target network element is multiple network elements, then the second network element can be some or all of the target network elements, such as the first network element being able to determine the second network element from the target network elements. Then, the first network element proceeds to step S605.

[0191] Step S605: The first network element sends a second message to the second network element. The second message is used to request model training.

[0192] In this application, the second message can also be referred to as a model training request.

[0193] For example, the content of the second message can be found in the relevant description of step S504 above, and will not be repeated here.

[0194] Step S606: The first network element sends a first notification message to the fourth network element. This message is used to notify the fourth network element of the model training network element change. The message includes the instance identifier or instance address of the second network element.

[0195] This step is optional.

[0196] For example, if the model report address in the message of step S605 is the instance address of the second network element, the first network element can send a message to the fourth network element. This message is used to notify the fourth network element of the model training network element change. The message includes the instance identifier or instance address (such as IP address or FQDN) of the second network element.

[0197] Step S607: The second network element sends a second notification message to the first or fourth network element, which includes the trained model information.

[0198] For example, after the training of the second network element is completed, a second notification message can be sent to the network element (first network element or fourth network element) corresponding to the model report address in step S606. The message includes the trained model information.

[0199] If the model report address in step S606 is the first network element, the first network element will forward the trained model information to the fourth network element after receiving the message.

[0200] It should be noted that during the training process, the second network element may also experience a model training interruption due to changes in energy information, requiring a switch to another MTLF for model training. In this case, the second network element can execute steps S602-S607. This application does not limit the number of model training switches.

[0201] Figure 7 This is a flowchart illustrating another communication method provided in an embodiment of this application.

[0202] In this embodiment, the steps performed by the first network element can also be performed by a module of the first network element (e.g., a chip, chip system, integrated circuit, or control unit); in this embodiment, the steps performed by the second communication device can also be performed by a module of the second communication device (e.g., a chip, chip system, integrated circuit, or control unit).

[0203] Figure 7 Some steps are optional, for example, Figure 7 It may include only steps S701 and S702, or... Figure 7 It includes only steps S702 and S703.

[0204] like Figure 7 As shown, the communication method may include some or all of the following steps:

[0205] Step S701: The fourth grid element determines the first grid element, which is determined based on the first energy information.

[0206] In one possible implementation, the first energy information is the first energy-saving capability; that is, the first network element is determined based on the first energy-saving capability. For example, the first network element can be a network element that supports the first energy-saving capability, i.e., the first network element supports the first energy-saving capability.

[0207] For example, the first energy-saving capability includes at least one of the following (or, in other words, the first network element supports at least one of the following): training the first model using a first energy type; or, reducing energy consumption while maintaining model accuracy at a threshold; or, the carbon emission factor during the training process of the first model is not higher than the carbon emission factor threshold; or, the normalized energy consumption during the training process of the first model is not higher than the normalized energy consumption threshold; or, the energy state during the training process of the first model belongs to the first energy state; or, training the first model based on an energy state timetable, such as training the first model at the time corresponding to a specified energy state; or, the device training the first model does not have an energy inefficiency indication; or, the energy consumption during the training process of the first model does not exceed the energy consumption threshold; or, the device training the first model meets the energy efficiency level performance index; or, the energy efficiency level of the energy used during the training process of the first model meets the first energy efficiency level; or, the device training the first model meets the network element energy consumption limit level. Two methods for determining the first network element are described below as examples.

[0208] For example, the fourth network element sends second information to the third network element, the second information indicating the first energy-saving capability; the fourth network element receives third information sent by the third network element, the third information indicating the first network element that it supports the first energy-saving capability. This process can also be illustrated by the description of method 1 below.

[0209] Optionally, the second information used to indicate the first energy-saving capability may refer to: the second information used to request network elements to support the first energy-saving capability. For example, the fourth network element and the third network element determine, based on a protocol or prior negotiation, that if the fourth network element sends energy-saving capability information (such as the second information) to the third network element, then the third network element, based on the energy-saving capability information, determines the network element that supports the energy-saving capability (such as the first energy-saving capability) indicated by the energy-saving capability information, and sends the network element information (such as the third information) of that network element (such as the first network element mentioned above) to the fourth network element.

[0210] For example, the second information includes a first energy type, used to indicate network elements requesting support for training the first model using the first energy type; or, the second information includes a model accuracy threshold, used to indicate network elements requesting support for reducing energy consumption while maintaining model accuracy at or above the model accuracy threshold; or, the second information includes a carbon emission factor threshold, used to indicate network elements requesting support for reducing energy consumption during the training process of the first model when the carbon emission factor is not higher than the carbon emission factor threshold; or, the second information includes a normalized energy consumption threshold, used to indicate network elements requesting support for reducing energy consumption during the training process of the first model when the normalized energy consumption is not higher than the normalized energy consumption threshold; or, the second information includes a first energy state and a second energy consumption threshold. The second information includes an energy state timetable, which indicates that the energy state during the training process of the first model is in a first energy state; or, the second information includes an energy state timetable, which indicates that the device requesting support is to train the first model based on the energy state timetable; or, the second information includes an energy inefficiency indication, which indicates that the device requesting to train the first model does not have an energy inefficiency indication; or, the energy consumption during the training process of the first model does not exceed an energy consumption threshold; or, the device training the first model meets the energy efficiency level performance index; or, the energy efficiency level of the energy used during the training process of the first model meets the first energy efficiency level; or, the device training the first model meets the network element energy consumption limit level.

[0211] Optionally, the first energy-saving capability can also be called the energy-saving capability requirement, the second information is used to indicate the energy-saving capability requirement, and the first network element is the network element that meets the energy-saving capability requirement.

[0212] For example, the third information can be the instance identifier or instance address of the first network element, such as an IP address or FQDN.

[0213] In another example, the fourth network element acquires the energy-saving capabilities of multiple network elements; based on the energy-saving capabilities of the multiple network elements, a first network element is determined, the first network element supports the first energy-saving capability, and the first network element belongs to multiple network elements. This process can also be illustrated by the description of method 2 below.

[0214] For example, the fourth network element can send a request message to the third network element, the request message including the network element type; the third network element can indicate the energy-saving capabilities of network elements belonging to that network element type (i.e., the aforementioned multiple network elements) to the fourth network element, such as the third network element sending the energy-saving capabilities of the aforementioned multiple network elements to the fourth network element, and correspondingly, the fourth network element receives the energy-saving capabilities of the aforementioned multiple network elements; furthermore, the fourth network element can determine the network element supporting the first energy-saving capability among the multiple network elements as the first network element based on the first energy-saving capability. For example, the network element type can be MTLF or NWDAF, that is, the aforementioned multiple network elements belong to MTLF or NWDAF.

[0215] In another possible implementation, the first energy information is the network element information of each of at least one network element; that is, the first network element is determined based on the network element information of each of at least one network element.

[0216] For example, the fourth network element can obtain network element information of each of the at least one network element, including the energy consumption and / or hardware capabilities required for a single iteration; based on the network element information of each of the at least one network element, a first network element is determined, and the first network element belongs to the at least one network element.

[0217] Optionally, determining the first network element based on the network element information of each of the at least one network element may include: a fourth network element determining the energy consumption information of each network element, the energy consumption information including the number of iterations or the computational cost required for training; and determining the first network element from the at least one network element based on the energy consumption information of each network element. For example, the energy consumption information of the first network element meets carbon emission requirements or energy consumption requirements.

[0218] For example, the fourth network element can request the instance address of the MTLF from the third network element. Assuming that the third network element sends the instance addresses of N MTLFs to the fourth network element, where N is a positive integer, the fourth network element can request network element information from each of the N MTLFs to obtain the network element information of the N MTLFs (i.e., at least one of the above network elements). The network element information includes the energy consumption and / or hardware capabilities required for a single iteration. Based on the network element information of the N MTLFs, the first network element can estimate the number of iterations and / or computational load required for each of the N MTLFs to complete the training of the first model, estimate the carbon emission results or energy consumption information brought about by training the model, and obtain the estimation result. The MTLF whose energy consumption information meets the requirements (such as carbon emission requirements or energy consumption requirements) is determined as the first network element.

[0219] For example, the network element information of MTLF can be the number of iterations and / or computational cost required to complete the training of the first model, that is, the number of iterations and / or computational cost required for MTLF to complete the training of the first model can be estimated by MTLF itself. This process can also be illustrated by the description of method 3 below.

[0220] Step S702: The fourth network element sends a first message to the first network element. The first message is used to request training of the first model. The first message includes first information, which is related to the energy consumption requirements of the first model.

[0221] Correspondingly, the first network element receives the first message sent by the fourth network element.

[0222] For example, the first information is used to indicate at least one of the following: energy type, model accuracy threshold, carbon emission factor, normalized energy consumption, energy status, energy status timetable, low energy efficiency indication, energy saving status migration execution indication, energy consumption threshold, energy efficiency level performance index, energy efficiency level, or network element energy consumption limit level.

[0223] In one possible implementation, the first network element trains the first model based on the first information to obtain the training result of the first model; then, the following step S703 is executed.

[0224] The training result is obtained based on the energy consumption requirement, or it can mean that the training process meets the energy consumption requirement.

[0225] For example, if the first information is used to indicate the energy type, the energy consumption requirement for the first model is: training the first model using the first energy type; if the first information is used to indicate the model accuracy threshold, the energy consumption requirement for the first model is: reducing energy consumption while ensuring the model accuracy is not lower than the model accuracy threshold; if the first information is used to indicate the carbon emission factor threshold, the energy consumption requirement for the first model is: the carbon emission factor during the training process of the first model is not higher than the carbon emission factor indicated by the first information; if the first information is used to indicate normalized energy consumption, the energy consumption requirement for the first model is: the normalized energy consumption during the training process of the first model is not higher than the normalized energy consumption indicated by the first information; if the first information is used to indicate the energy state, the energy consumption requirement for the first model is: the energy state during the training process of the first model belongs to the energy state indicated by the first information; if the first information is used to indicate the energy state timetable, the energy consumption requirement for the first model is: based on the energy state... The timetable is used to train the first model, such as training the first model at a time corresponding to a specified energy state; if the first information is used to indicate a low energy efficiency indication, then the energy consumption requirement for the first model is: the device training the first model does not have a low energy efficiency indication; if the first information is used to indicate an energy consumption threshold, then the energy consumption requirement for the first model is: the energy consumption during the training process of the first model does not exceed the energy consumption threshold; if the first information is used to indicate an energy efficiency level performance index, then the energy consumption requirement for the first model is: the device training the first model meets the energy efficiency level performance index; if the first information is used to indicate an energy efficiency level, then the energy consumption requirement for the first model is: the energy efficiency level of the energy used during the training process of the first model meets the energy efficiency level indicated by the first information; if the first information is used to indicate a network element energy consumption limit level, then the energy consumption requirement for the first model is: the device training the first model meets the network element energy consumption limit level.

[0226] For example, the first information is used to indicate the first energy type, the energy consumption requirement corresponding to the first model is to use the first energy type, and the first message is used to request to train the first model using the first energy type. Then, the fourth network element can use the first energy type to train the first model and obtain the training result of the first model. For a more detailed explanation, please refer to the relevant content of method a below.

[0227] In another example, the first information is used to indicate the model accuracy threshold. The energy consumption requirement for the first model is to use a training method that reduces energy consumption while ensuring that the energy consumption is not lower than the model accuracy threshold. Then, the fourth network element can use a training method that reduces energy consumption (such as model compression) to train the first model while ensuring that the energy consumption is not lower than the model accuracy threshold, and obtain the training result of the first model. For a more detailed explanation, please refer to the relevant content of method b below.

[0228] Step S703: The first network element sends the training results of the first model to the fourth network element. The training results are obtained based on the energy consumption requirements.

[0229] Correspondingly, the fourth network element receives the training results of the first model sent by the first network element.

[0230] In some embodiments of this application, the first network element may also send fifth information to the fourth network element, and correspondingly, the fourth network element receives the fifth information sent by the first network element, wherein the fifth information is used to indicate model training interruption and / or the reason for model training interruption; the fourth network element may send a second message to the second network element, the second message being used to request training of the first model, and the second message including the aforementioned first information. Optionally, the second network element may continue training of the first model based on the first information.

[0231] For example, if the training process of the first network element does not meet the energy consumption requirements (e.g., an energy switch occurs or the network enters an energy-saving state), the first network element can return the training results of the first model (e.g., the trained model parameters) and fifth information (e.g., a training interruption indication and a reason value) to the fourth network element. The reason value may include an energy type switch, an increase in energy consumption level, or entering an energy-saving state. The fourth network element can then re-determine the network element to perform the training (this process can be found in step S701). Assuming the re-determined network element is the second network element, the fourth network element can send a second message to the second network element. The second message requests training of the first model. The second message includes the aforementioned first information and the training results of the first model returned by the first network element. Accordingly, after receiving the second message, the second network element can continue training the first model based on the first information and the training results of the first model.

[0232] In this embodiment of the application, the total carbon emissions from the entire model training are controlled by switching between different MTLFs that meet energy requirements (such as the first energy-saving requirement mentioned above) on the model training side, such as switching from the first grid element to the second grid element for training.

[0233] above Figure 7 The method embodiments shown include many possible implementation schemes, which will be discussed below. Figure 8A and Figure 8B Some of the implementation schemes will be illustrated with examples. It should be noted that... Figure 8A and Figure 8B For any unexplained related concepts, operations, or logical relationships, please refer to [link / reference]. Figure 7 The corresponding description in the illustrated embodiment.

[0234] In this application, Figure 8A or Figure 8B The illustrated embodiment can be considered as a standalone embodiment. Figure 8A or Figure 8B The embodiments shown can all be implemented without relying on Figure 7 The technical solution; Figure 8A or Figure 8B Some of the steps in the illustrated embodiments can also be used as separate embodiments.

[0235] For example, the fourth network element can be an NWDAF consumer, the third network element can be an NRF, and the first and second network elements can be two MTLFs. For ease of description, the first network element is referred to as MTLF#1 and the second network element is referred to as MTLF#2.

[0236] In this application, MTLF can also be replaced with MTLF instance, NWDAF, or NWDAF instance, and this application does not limit this.

[0237] Figure 8A This is a schematic diagram of another communication method provided in an embodiment of this application. The method includes some or all of the following steps:

[0238] Step S801: N MTLFs register their energy-saving capabilities with the third network element, where N is a positive integer and the N MTLFs include the first network element.

[0239] For example, the MTLF can send a registration message to a third network element. This message is used to register its energy-saving capability with the third network element. The message includes an NF profile and indication information indicating the MTLF's energy-saving capability. In this application, energy-saving capability can also be referred to as energy capacity.

[0240] For example, information indicating energy-saving capabilities can be stored in the NF profile.

[0241] For example, energy-saving capabilities may include at least one of the following: energy type, such as clean energy, and whether clean energy is supported; whether training methods (or model compression methods) that reduce energy consumption or computational load are supported, such as pruning or distillation.

[0242] Step S802: The fourth network element interacts with the third network element to determine the first network element.

[0243] For example, step S802 may include method 1, method 2, or method 3 described below.

[0244] Method 1 may include the following steps S11 to S12:

[0245] Step S11: The fourth network element sends the second information to the third network element. The second information is used to indicate the first energy-saving capability.

[0246] Correspondingly, the third network element receives the second information sent by the fourth network element.

[0247] For example, a fourth network element can send a network element discovery request to a third network element. This request is used to discover the MTLF (Mean Transmission Default Allocation Requirement). The request may include the network element type (NFtype) and the aforementioned second information, which can also be referred to as an energy-saving capability requirement. The NF type indicates either the NWDAF (Non-Wideband Default Allocation Requirement) or the MTLF. Correspondingly, the third network element receives the network element discovery request sent by the fourth network element.

[0248] For example, the specific content of the second information can be found above. Figure 7 The relevant descriptions (such as those in step S701) will not be repeated here.

[0249] Step S12: The third network element sends third information to the fourth network element. The third information is used to instruct the first network element that the first network element supports the first energy-saving capability.

[0250] Correspondingly, the fourth network element receives the third information sent by the third network element.

[0251] For example, the third information may be the instance identifier and / or address of the first network element.

[0252] Optionally, if the third network element determines that multiple network elements support the first energy-saving capability, the third network element may select one of them as the first network element. Alternatively, the third network element may instruct the fourth network element to select the first network element from the multiple network elements. This application does not limit the method for selecting the first network element from the multiple network elements.

[0253] In this method, the fourth network element searches for an MTLF that meets the energy-saving capability requirements from the third network element to determine the first network element.

[0254] Method 2 includes the following steps S21 to S22:

[0255] Step S21: The fourth network element obtains the energy-saving capabilities of multiple network elements from the third network element.

[0256] For example, a fourth network element can send a network element discovery request to a third network element. This request is used to discover MTLFs (Medium-Terminal Element Levels), and the request may include a network element type (NFtype). The NFtype indicates either NWDAF (Non-Terminal Element Level) or MTLF. Correspondingly, the third network element receives the network element discovery request from the fourth network element. Then, the third network element can send energy-saving capability information of network elements conforming to the network element type (i.e., NWDAF or MTLF) to the fourth network element. This energy-saving capability information indicates the energy-saving information of multiple network elements. Correspondingly, the fourth network element receives the energy-saving capability information of multiple network elements sent by the third network element; that is, the fourth network element obtains the energy-saving capabilities of multiple network elements from the third network element.

[0257] For example, when a third network element sends energy-saving capability information to a fourth network element, it may include: instance identifiers and / or addresses of all MTLFs and energy-saving capability information corresponding to each MTLF, wherein the energy-saving capability information may be energy-related information.

[0258] It should be understood that the difference between the network element discovery request in step S21 and the network element discovery request in step S11 is that the network element discovery request in step S21 does not include the aforementioned second information.

[0259] Step S22: The fourth network element determines the first network element based on the energy-saving capabilities of multiple network elements. The first network element supports the first energy-saving capability and belongs to multiple network elements.

[0260] For example, if a third network element returns instance identifiers and / or addresses of all MTLFs and energy-saving capability information corresponding to each MTLF to a fourth network element, then the fourth network element can select a first network element from the MTLFs returned by the third network element based on energy-saving capability requirements. For instance, the energy-saving capability requirement might be: "This network element supports the first energy-saving capability." An exemplary description of the first energy-saving capability can be provided below. Figure 7 The relevant description in the document.

[0261] Method 3 includes the following steps S31 to S35:

[0262] S31: The fourth network element requests the instance address of the MTLF from the third network element.

[0263] S32: The third network element sends N instance addresses of MTLF to the fourth network element, where N is a positive integer.

[0264] S33: The fourth network element sends a first request message to N MTLFs. This message is used to request network element information, which includes the energy consumption and / or hardware capabilities required for a single iteration.

[0265] For example, network element information can also be called model energy consumption related information, that is, all MTLF request model energy consumption related information contained in the message returned by the fourth network element to the third network element.

[0266] For example, the first request message may include at least one of model-related information, a network element energy type request indication, or a network element energy type request indication. The model-related information indicates the model to be trained, such as a model file or model identifier, and MTLF can estimate the energy consumption required for a single iteration based on the model-related information. The network element energy type request indication instructs MTLF to return its energy type, and the model energy consumption information request indication instructs MTLF to return model energy consumption-related information.

[0267] Optionally, the first request message may not include a model energy consumption information request indication. The first request message may implicitly instruct MTLF to return model energy consumption related information, such as by implicitly instructing MTLF to return model energy consumption related information through the name of the first request message.

[0268] S34: The fourth network element receives network element information from N MTLFs.

[0269] For example, after receiving the first request message, the MTLF returns its energy type and model energy consumption information to the fourth network element, such as the energy consumption required for a single iteration and / or the MTLF hardware capabilities, such as computing power.

[0270] S35: The fourth network element determines the first network element based on the network element information of N MTLFs.

[0271] For example, the first network element can estimate the number of iterations and / or computational cost required for each of the N MTLFs to complete the training of the first model based on the network element information of the N MTLFs, and estimate the carbon emission results or energy consumption information brought about by training the model to obtain the estimation result; the MTLF whose estimation result meets the requirements (such as carbon emission requirements or energy consumption requirements) is determined as the first network element.

[0272] Step S803: The fourth network element sends a first message to the first network element. The first message is used to request training of the first model. The first message includes first information, which is related to the energy consumption requirements of the first model.

[0273] Correspondingly, the first network element receives the first message sent by the fourth network element.

[0274] Step S804: The first network element trains the first model based on the energy consumption requirements corresponding to the first model.

[0275] For example, by Figure 8B This paper introduces two possible implementation methods in the model training process. For ease of description, these two possible implementation methods will be referred to as method a and method b.

[0276] For example, in method a, the first information is used to indicate the first energy type, and the energy consumption requirement corresponding to the first model is to use the first energy type. In method a, the first network element can be the NWDAF that supports the first energy type determined by the fourth network element. Alternatively, in method b, the first information is used to indicate the model accuracy threshold, and the energy consumption requirement corresponding to the first model is to adopt a training method that reduces energy consumption (such as a compressed model) when it is not lower than the model accuracy threshold. In method b, the first network element can be the NWDAF that supports the compressed model determined by the fourth network element.

[0277] For example, methods a and b are described in detail below.

[0278] Method a may include the following steps S1a to S5a:

[0279] S1a: The fourth network element sends a model training request to the first network element. This request requests model training using the first energy type. The request may include model energy-related information (such as the first energy type). Optionally, the request may also include analysis identifiers, model information, or model files.

[0280] For example, the model training request can reuse the model subscription request or be a new message type, such as an energy-saving model training request.

[0281] S2a: The first network element uses the first energy type to train the first model.

[0282] S3a: If the first network element switches its power or enters an energy-saving state, it sends the trained model parameters (which can be included in the model information or sent directly as the model file) to the fourth network element. Optionally, the first network element can also return a training interruption indication and a reason value to the fourth network element, such as a power type switch, an increase in energy consumption level, or entering an energy-saving state.

[0283] S4a: The fourth network element requests other MTLFs (such as the second network element) to continue training the model.

[0284] For example, the fourth network element selects an MTLF from the MTLFs obtained from the third network element, or selects another MTLF based on the energy-related information corresponding to the MTLF obtained from the third network element, to continue model training. Assuming the selected network element is the second network element, the fourth network element can send a portion of the trained model file to the second network element, and the second network element can continue model training based on this portion of the model file.

[0285] S5a: The second network element can send the training results to the fourth network element.

[0286] For example, the training process of the second network element can refer to the training process of the first network element. If there is no interruption, the second network element can send the training result obtained after training to the fourth network element. If the training of the second network element is interrupted, the operation in step S3a can be performed to send the training result obtained when the training was interrupted to the fourth network element.

[0287] In method a, step S804 may include steps S1a to S2a as described above, and step S805 may include step S3a as described above.

[0288] For example, if the requirement for the fourth network element is to reduce model training energy consumption and the requirement for model accuracy is relatively relaxed, then step S804 may include method b, which may include the following steps S1b to S3b:

[0289] S1b: If the first network element supports reducing model energy consumption through model compression, the fourth network element sends a model training request to the first network element. The model training request includes a model accuracy threshold and is used to instruct the first model to be trained.

[0290] For example, the model training request may also include at least one of the following: analysis identifier, model information, or model file.

[0291] S2b: After receiving a model training request, the first network element can reduce the energy consumption of model training as much as possible, such as by compressing the model, while ensuring that the model accuracy exceeds the above-mentioned model accuracy threshold.

[0292] S3b: The first network element sends the training results of the first model to the fourth network element.

[0293] In method b, step S804 may include steps S1b to S2b as described above, and step S805 may include step S3b as described above.

[0294] Step S805: The first network element sends the training results of the first model to the fourth network element.

[0295] Correspondingly, the fourth network element receives the training results of the first model sent by the first network element.

[0296] Optionally, the training result of the first model sent by the first network element to the fourth network element can be the training result obtained after training is completed; or it can be the training result obtained after training is interrupted.

[0297] For example, if the training of the first network element is interrupted in methods 1 to 2 above, the first network element can also send a fifth message to the fourth network element. Accordingly, the fourth network element receives the fifth message sent by the first network element, wherein the fifth message is used to indicate the interruption of model training and / or the reason for the interruption of model training. Furthermore, the fourth network element can re-determine the network element to perform training (this process can be referred to in steps S701 and S802). Assuming that the network element re-determined by the fourth network element is the second network element, the fourth network element can send a second message to the second network element. The second message is used to request training of the first model. The second message includes the first message and the training result of the first model returned by the first network element. Accordingly, after receiving the second message, the second network element can continue the training of the first model based on the first message and the training result of the first model.

[0298] In this embodiment of the application, the fourth network element (such as the consumer) notifies the MTLF (such as the first network element) of its energy requirements for model training (i.e., the energy consumption requirements corresponding to the first model mentioned above) during model training. By controlling the MTLF to meet the energy requirements throughout the entire model training process, the energy consumption or carbon emissions of the entire model training are controlled.

[0299] The foregoing details the method provided in this application. To facilitate the implementation of the above-described solutions in the embodiments of this application, corresponding apparatus or devices are also provided in the embodiments of this application.

[0300] This application divides the first network element, second network element, third network element, and fourth network element into functional modules according to the above method embodiments. For example, each function can be divided into its own functional module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this application is illustrative and only represents one logical functional division; other division methods may be used in actual implementation. The following will combine... Figures 9 to 11 The communication device of the embodiments of this application is described in detail.

[0301] See Figure 9 , Figure 9 This is a schematic diagram of the structure of a communication device provided in an embodiment of this application. Figure 9 As shown, the communication device may include a transceiver unit 901 and a processing unit 902.

[0302] In some embodiments of this application, the communication device may be the first network element shown above, or a chip or circuit disposed in the first network element. That is, the communication device may be used to perform the steps or functions performed by the first network element in the above method embodiments.

[0303] In one design, processing unit 902 is used to: stop training the first model based on the energy information of the first network element; transceiver unit 901 is used to: send first information to a second network element, the first information being used to request the second network element to train the first model, the second network element being determined based on the energy information of the second network element.

[0304] For example, the first information includes model information of the first model, which is used for model training of the first model; wherein, the training of the first model is not completed before training of the first model is stopped, and the model information of the first model is used to indicate the first model that has not completed training.

[0305] For example, the energy information of the first grid element includes at least one of the following: the energy efficiency of the first grid element, the normalized energy consumption of the first grid element, the energy status of the first grid element, the energy consumption of the first grid element within a preset time period, the new energy ratio of the first grid element, the carbon emission factor of the first grid element, the energy type of the first grid element, the energy efficiency level of the first grid element, or the energy consumption of the first grid element.

[0306] In one possible implementation, the processing unit 902 is specifically configured to: stop training the first model if the energy information of the first network element meets a first condition, wherein the first condition includes at least one of the following: the energy efficiency of the first network element is lower than a first energy efficiency threshold; the normalized energy consumption of the first network element exceeds a first normalized energy consumption threshold; the first network element enters an energy-saving state; the energy consumption of the first network element exceeds a first energy consumption threshold within a preset time; the new energy ratio of the first network element is lower than a first new energy ratio threshold; the carbon emission factor of the first network element is higher than a first carbon emission factor threshold; the energy type of the first network element changes; the energy efficiency level of the first network element is greater than a first energy efficiency level threshold; or, the energy consumption of the first network element exceeds a first network element energy consumption limit level.

[0307] In one possible implementation, the transceiver unit 901 is further configured to: send a first message to a third network element, the first message being used to request the determination of a second network element; and receive second information sent by the third network element, the second information being used to instruct the second network element.

[0308] For example, the first message includes third information, which is used to indicate a second condition. The energy information of the second network element meets the second condition, which includes at least one of the following: the energy efficiency of the second network element is lower than a second energy efficiency threshold; the normalized energy consumption of the second network element exceeds a second normalized energy consumption threshold; the second network element enters an energy-saving state; the energy consumption of the second network element exceeds a second energy consumption threshold within a preset time; the new energy ratio of the second network element is lower than a second new energy ratio threshold; the carbon emission factor of the second network element is higher than a second carbon emission factor threshold; the energy type of the second network element is a first energy type; the energy efficiency level of the second network element is greater than a second energy efficiency level threshold; or, the energy consumption of the second network element exceeds a second network element energy consumption limit level.

[0309] In one possible implementation, the transceiver unit 901 is specifically used to: send a second message to the second network element, the second message being used to request model training, the second message including fourth information, the fourth information being used to instruct the second network element to notify the training result of the model training object.

[0310] For example, the first condition is either pre-configured or obtained from other network elements.

[0311] Specific details in the embodiments of this application can be exemplarily referred to above. Figures 5 to 6 The descriptions in the method embodiments shown will not be repeated here.

[0312] It is understood that the specific descriptions of the transceiver unit 901 and the processing unit 902 shown in the embodiments of this application are merely examples. For the specific functions or execution steps of the transceiver unit 901 and the processing unit 902, please refer to the above descriptions. Figures 5 to 6 The method embodiments shown are not described in detail here. Furthermore, the technical effects of the embodiments of this application are explained in the foregoing. Figures 5 to 6 The technical effects of the method embodiments shown are not elaborated here for the sake of simplicity.

[0313] Reuse Figure 9 In some embodiments of this application, the communication device may be the third network element shown above, or a chip or circuit disposed in the third network element. That is, the communication device may be used to perform the steps or functions performed by the third network element in the above method embodiments.

[0314] In one design, the transceiver unit 901 is used to: receive a first message sent by a first network element, the first message being used to request the determination of a second network element; and send third information to the first network element, the third information being used to instruct the second network element, the second network element being determined based on the energy information of the second network element.

[0315] For example, the first message includes third information, which is used to indicate a second condition. The energy information of the second network element meets the second condition, which includes at least one of the following: the energy efficiency of the second network element is lower than a second energy efficiency threshold; the normalized energy consumption of the second network element exceeds a second normalized energy consumption threshold; the second network element enters an energy-saving state; the energy consumption of the second network element exceeds a second energy consumption threshold within a preset time; the new energy ratio of the second network element is lower than a second new energy ratio threshold; the carbon emission factor of the second network element is higher than a second carbon emission factor threshold; the energy type of the second network element is a first energy type; the energy efficiency level of the second network element is greater than a second energy efficiency level threshold; or, the energy consumption of the second network element exceeds a second network element energy consumption limit level.

[0316] Specific details in the embodiments of this application can be exemplarily referred to above. Figures 5 to 6 The descriptions in the method embodiments shown will not be repeated here.

[0317] It is understood that the specific descriptions of the transceiver unit 901 and the processing unit 902 shown in the embodiments of this application are merely examples. For the specific functions or execution steps of the transceiver unit 901 and the processing unit 902, please refer to the above descriptions. Figures 5 to 6 The method embodiments shown are not described in detail here. Furthermore, the technical effects of the embodiments of this application are explained in the foregoing. Figures 5 to 6 The technical effects of the method embodiments shown are not elaborated here for the sake of simplicity.

[0318] Reuse Figure 9 In some embodiments of this application, the communication device may be the second network element shown above, or a chip or circuit disposed in the second network element. That is, the communication device may be used to perform the steps or functions performed by the second network element in the above method embodiments.

[0319] In one design, the transceiver unit 901 is used to: receive first information sent by a first network element, the first information being used to request a second network element to train a first model.

[0320] For example, the processing unit 902 is further configured to train the first model to obtain training results; the transceiver unit 901 is configured to send the training results.

[0321] Specific details in the embodiments of this application can be exemplarily referred to above. Figures 5 to 6 The descriptions in the method embodiments shown will not be repeated here.

[0322] It is understood that the specific descriptions of the transceiver unit 901 and the processing unit 902 shown in the embodiments of this application are merely examples. For the specific functions or execution steps of the transceiver unit 901 and the processing unit 902, please refer to the above descriptions. Figures 5 to 6The method embodiments shown are not described in detail here. Furthermore, the technical effects of the embodiments of this application are explained in the foregoing. Figures 5 to 6 The technical effects of the method embodiments shown are not elaborated here for the sake of simplicity.

[0323] Reuse Figure 9 In some embodiments of this application, the communication device may be the fourth network element shown above, or a chip or circuit disposed in the fourth network element. That is, the communication device may be used to perform the steps or functions performed by the fourth network element in the above method embodiments.

[0324] In one design, the processing unit 902 is used to: determine a first network element, the first network element being determined based on first energy information; the transceiver unit 901 is used to: send a first message to the first network element, the first message being used to request training of a first model, the first message including first information, the first information being related to the energy consumption requirements corresponding to the first model.

[0325] In one possible implementation, the transceiver unit 901 is also used to: receive the training results of the first model sent by the first network element, the training results being obtained based on energy consumption requirements.

[0326] For example, the first energy information is the first energy-saving capability. The transceiver unit 901 is specifically used to: send second information to the third network element, the second information being used to indicate the first energy-saving capability; and receive third information sent by the third network element, the third information being used to indicate the first network element that the first network element supports the first energy-saving capability.

[0327] In one possible implementation, the processing unit 902 is specifically used to: acquire the energy-saving capabilities of multiple network elements; and, based on the energy-saving capabilities of the multiple network elements, determine a first network element, wherein the first network element supports the first energy-saving capability and the first network element belongs to multiple network elements.

[0328] For example, the first energy-saving capability includes at least one of the following: training the first model using a first energy type; or reducing energy consumption while maintaining model accuracy at a threshold; or ensuring that the carbon emission factor during the training process of the first model is not higher than the carbon emission factor threshold; or ensuring that the normalized energy consumption during the training process of the first model is not higher than the normalized energy consumption threshold; or ensuring that the energy state during the training process of the first model is a first energy state; or training the first model based on an energy state schedule, such as training the first model at the time corresponding to a specified energy state; or ensuring that the device training the first model does not have an energy inefficiency indication; or ensuring that the energy consumption during the training process of the first model does not exceed the energy consumption threshold; or ensuring that the device training the first model meets the energy efficiency level performance indicators; or ensuring that the energy efficiency level of the energy used during the training process of the first model meets the first energy efficiency level; or ensuring that the device training the first model meets the network element energy consumption limit level.

[0329] For example, the first energy information is the network element information of each of the at least one network element. The processing unit 902 is specifically used to: obtain the network element information of each of the at least one network element, the network element information including the energy consumption and / or hardware capabilities required for a single iteration; and determine the first network element based on the network element information of each of the at least one network element, the first network element belonging to the at least one network element.

[0330] In one possible implementation, the transceiver unit 901 is further configured to: receive fifth information sent by the first network element, the fifth information being used to indicate model training interruption and / or the reason for model training interruption; and send a second message to the second network element, the second message being used to request training of the first model, the second message including the first information.

[0331] For example, the first information is used to indicate at least one of the following: energy type, model accuracy threshold, carbon emission factor, normalized energy consumption, energy status, energy status timetable, low energy efficiency indication, energy saving status migration execution indication, energy consumption threshold, energy efficiency level performance index, energy efficiency level, or network element energy consumption limit level.

[0332] Specific details in the embodiments of this application can be exemplarily referred to above. Figures 7 to 8B The descriptions in the method embodiments shown will not be repeated here.

[0333] It is understood that the specific descriptions of the transceiver unit 901 and the processing unit 902 shown in the embodiments of this application are merely examples. For the specific functions or execution steps of the transceiver unit 901 and the processing unit 902, please refer to the above descriptions. Figures 7 to 8B The method embodiments shown are not described in detail here. Furthermore, the technical effects of the embodiments of this application are explained in the foregoing. Figures 7 to 8BThe technical effects of the method embodiments shown are not elaborated here for the sake of simplicity.

[0334] Reuse Figure 9 In some embodiments of this application, the communication device may be the first network element shown above, or a chip or circuit disposed in the first network element. That is, the communication device may be used to perform the steps or functions performed by the first network element in the above method embodiments.

[0335] In one design, the transceiver unit 901 is used to: receive a first message sent by a fourth network element, the first message being used to request training of a first model, the first message including first information, the first information being related to the energy consumption requirements corresponding to the first model; and send the training results of the first model to the fourth network element, the training results being obtained based on the energy consumption requirements.

[0336] In one possible implementation, the transceiver unit 901 is further configured to: send fifth information to the fourth network element, the fifth information being used to indicate model training interruption and / or the reason for model training interruption.

[0337] For example, the first information is used to indicate at least one of the following: energy type, model accuracy threshold, carbon emission factor, normalized energy consumption, energy status, energy status timetable, low energy efficiency indication, energy saving status migration execution indication, energy consumption threshold, energy efficiency level performance index, energy efficiency level, or network element energy consumption limit level.

[0338] Specific details in the embodiments of this application can be exemplarily referred to above. Figures 7 to 8B The descriptions in the method embodiments shown will not be repeated here.

[0339] It is understood that the specific descriptions of the transceiver unit 901 and the processing unit 902 shown in the embodiments of this application are merely examples. For the specific functions or execution steps of the transceiver unit 901 and the processing unit 902, please refer to the above descriptions. Figures 7 to 8B The method embodiments shown are not described in detail here. Furthermore, the technical effects of the embodiments of this application are explained in the foregoing. Figures 7 to 8B The technical effects of the method embodiments shown are not elaborated here for the sake of simplicity.

[0340] The first network element, second network element, third network element, and fourth network element of this application embodiment have been described above. The following describes the possible product forms of the first network element, second network element, third network element, and fourth network element. It should be understood that any network element possessing the above-described... Figure 9 Any product in any form that incorporates the functions of any one of the first, second, third, and fourth network elements described herein falls within the protection scope of this application. It should also be understood that the following description is merely illustrative and does not limit the product form of the communication device in this application to these specific examples.

[0341] In one possible implementation, Figure 9 In the communication device shown, processing unit 902 can be a processing circuit, and transceiver unit 901 can be a communication circuit. The processing circuit can be one or more processors, or all or part of the control or processing circuitry within one or more processors. When the communication device is any one of a first network element, a second network element, a third network element, and a fourth network element, the communication circuit can be a transceiver circuit, which can be a transceiver. When the communication device is a chip or a chip system, the communication circuit can be an interface circuit. When the communication device is a server, the communication circuit can be an interface circuit or a transceiver circuit. Transceiver unit 901 can also be a sending unit and / or a receiving unit. The sending unit can be a sending circuit, and the receiving unit can be a receiving circuit, integrated into a single device. In this embodiment, the processing circuit and the communication circuit can be coupled, etc., and the connection method between the processing circuit and the communication circuit is not limited in this embodiment. During the execution of the above method, the process of sending information in the above method can be understood as the process of the processing circuit outputting the above information. When outputting the above information, the processing circuit outputs the above information to the communication circuit so that the communication circuit can transmit it. After the aforementioned information is output by the processing circuit, it may require further processing before reaching the communication circuit. Similarly, the process of receiving information in the above method can be understood as the process by which the processing circuit receives the input information. When the processing circuit receives the input information, the communication circuit receives the information and inputs it into the processing circuit. Furthermore, after the communication circuit receives the information, it may require further processing before being input into the processing circuit. In one possible implementation, Figure 9In the communication device shown, the processing unit 902 can be one or more processors, and the transceiver unit 901 can be a transceiver, or the transceiver unit 901 can also be a transmitting unit and / or a receiving unit. The transmitting unit can be a transmitter, and the receiving unit can be a receiver. The transmitting unit and the receiving unit are integrated into one device, such as a transceiver. In the embodiments of this application, the processor and the transceiver can be coupled, etc., and the connection method between the processor and the transceiver is not limited in the embodiments of this application. In the process of executing the above method, the process of sending information in the above method can be understood as the process of the processor outputting the above information. When outputting the above information, the processor outputs the above information to the transceiver so that the transceiver can transmit it. After the above information is output by the processor, it may need to undergo other processing before reaching the transceiver. Similarly, the process of receiving information in the above method can be understood as the process of the processor receiving the input above information. When the processor receives the input information, the transceiver receives the above information and inputs it into the processor. Furthermore, after the transceiver receives the above information, the above information may need to undergo other processing before being input into the processor.

[0342] See Figure 10 , Figure 10 This is another structural schematic diagram of the communication device provided in the embodiments of this application. For example... Figure 10 As shown, the communication device provided in this application embodiment can be used to implement the methods described in the above method embodiments, and the description in the above method embodiments can be referred to. The communication device can be a second communication device, a first network element, or a chip. For example, the communication device includes one or more processors 1001. The communication device may further include a memory 1003. Optionally, the communication device may further include a transceiver 1002. In one implementation, the communication device further includes an input / output device (…). Figure 10 (Not indicated).

[0343] The processor 1001 is mainly used to process communication protocols and communication data, control the entire communication device, execute software programs, and process the data of the software programs. The memory 1003 is mainly used to store software programs and data. The transceiver 1002 may include control circuitry and an antenna. The control circuitry is mainly used for converting baseband signals to radio frequency signals and processing radio frequency signals. The antenna is mainly used for transmitting and receiving radio frequency signals in the form of electromagnetic waves. Input / output devices, such as touchscreens, displays, and keyboards, are mainly used to receive user input data and output data to the user.

[0344] When the communication device is powered on, the processor 1001 can read the software program in the memory 1003, interpret and execute the instructions of the software program, and process the data of the software program. When data needs to be transmitted wirelessly, the processor 1001 performs baseband processing on the data to be transmitted and outputs the baseband signal to the radio frequency (RF) circuit. The RF circuit processes the baseband signal and transmits the RF signal outward in the form of electromagnetic waves through the antenna. When data is sent to the communication device, the RF circuit receives the RF signal through the antenna, converts the RF signal into a baseband signal, and outputs the baseband signal to the processor 1001. The processor 1001 converts the baseband signal into data and processes the data.

[0345] In another implementation, the radio frequency circuitry and antenna can be set up independently of the processor performing baseband processing. For example, in a distributed scenario, the radio frequency circuitry and antenna can be arranged remotely, independent of the communication device.

[0346] The processor 1001, transceiver 1002, and memory 1003 can be connected via a communication bus.

[0347] For example, when the communication device is used to perform the above... Figure 5 In the illustrated embodiment, when the first network element executes a step, method, or function, the transceiver 1002 can be used to perform... Figure 5 In steps S502 to S504, the processor 1001 can be used to execute step S501.

[0348] For example, when the communication device is used to perform the above... Figure 5 In the illustrated embodiment, when the third network element executes a step, method, or function, the transceiver 1002 can be used to perform... Figure 5 Steps S502 and S503 in the text.

[0349] For example, when the communication device is used to perform the above... Figure 7 In the illustrated embodiment, when the first network element executes a step, method, or function, the transceiver 1002 can be used to perform... Figure 7 Steps S702 to S703 in the process.

[0350] For example, when the communication device is used to perform the above... Figure 7 In the illustrated embodiment, when the fourth network element executes a step, method, or function, the transceiver 1002 can be used to perform... Figure 7 In steps S702 and S703, the processor 1001 can be used to execute step S701.

[0351] In any of the above implementations, the processor 1001 may include a transceiver for implementing receiving and transmitting functions. For example, the transceiver may be a transceiver circuit, an interface, or an interface circuit. The transceiver circuit, interface, or interface circuit for implementing receiving and transmitting functions may be separate or integrated. The aforementioned transceiver circuit, interface, or interface circuit can be used for reading and writing code / data, or it can be used for transmitting or relaying signals.

[0352] In any of the above implementations, the processor 1001 may store instructions, which may be computer programs. These computer programs, running on the processor 1001, cause the communication device to execute the methods described in the above method embodiments. The computer program may be embedded in the processor 1001; in this case, the processor 1001 may be implemented in hardware.

[0353] In one implementation, the communication device may include a circuit that can perform the functions of transmitting, receiving, or communicating in the aforementioned method embodiments. The processor and transceiver described in this application can be implemented on integrated circuits (ICs), analog ICs, radio frequency integrated circuits (RFICs), mixed-signal ICs, application-specific integrated circuits (ASICs), printed circuit boards (PCBs), electronic devices, etc. The processor and transceiver can also be manufactured using various IC process technologies, such as complementary metal-oxide semiconductors (CMOS), n-metal-oxide-semiconductor (NMOS), positive-channel metal-oxide semiconductors (PMOS), bipolar junction transistors (BJTs), bipolar CMOS (BiCMOS), silicon-germanium (SiGe), gallium arsenide (GaAs), etc.

[0354] It is understood that the communication device shown in the embodiments of this application may also have more than Figure 10 This application does not limit the use of other components or other related elements. The methods performed by the processor and transceiver shown above are merely examples; for the specific steps performed by the processor and transceiver, please refer to the description of the method embodiments above.

[0355] In another possible implementation, the communication device provided in this application embodiment may include one or more processors and a memory. The processor executes a program stored in the memory, and when the program is executed, the method embodiment described above is performed. Exemplarily, the processor and memory may also be integrated into a single device, i.e., the processor and memory may be integrated together. For more details regarding the processor and memory, please refer to... Figure 10 The relevant information regarding the processor 1001 and memory 1003.

[0356] See Figure 11 , Figure 11 This is another structural schematic diagram of the communication device provided in the embodiments of this application. For example... Figure 11 As shown, Figure 11 The communication device shown includes logic circuit 1101 and interface 1102. That is, the processing unit can be implemented using logic circuit 1101, and the transceiver unit 901 can be implemented using interface 1102. The logic circuit 1101 can be a chip, processing circuit, integrated circuit, or system-on-chip (SoC) chip, etc., and the interface 1102 can be a communication interface, input / output interface, pins, etc. For example, Figure 11 The above-mentioned communication device is used as an example of a chip, which includes a logic circuit 1101 and an interface 1102.

[0357] In this embodiment, the logic circuit and the interface can also be coupled to each other. The specific connection method between the logic circuit and the interface is not limited in this embodiment.

[0358] For example, when the communication device is used to perform the above... Figure 5 When the first network element executes a step, method, or function in the method embodiment shown, interface 1102 is used to transmit a first message, third information, and first information.

[0359] For example, when the communication device is used to perform the above... Figure 5 When the third network element executes a step, method, or function in the method embodiment shown, interface 1102 is used to transmit the first message and the third information.

[0360] For example, when the communication device is used to perform the above... Figure 5 When the second network element executes a step, method, or function in the method embodiment shown, interface 1102 is used to transmit first information.

[0361] In the embodiments of this application, the descriptions of the first message, the third information, and the first information, etc., can be found above. Figure 5The method embodiments shown are described in detail here, and will not be repeated here. It is understood that for a more detailed explanation of the logic circuit 1101 and interface 1102, please refer to [the relevant documentation / reference needed]. Figure 9 The descriptions of the processing unit and transceiver unit shown will not be repeated here.

[0362] It is understood that the communication device shown in the embodiments of this application can implement the method provided in the embodiments of this application in hardware form or in software form, etc., and the embodiments of this application do not limit it in this way.

[0363] for Figure 11 For specific implementations of the various embodiments shown, please refer to the above embodiments, which will not be described in detail here.

[0364] This application also provides a communication system, which includes a first network element and a third network element that can be used to execute any of the foregoing method embodiments. Figures 5 to 6 The method in ).

[0365] This application also provides a communication system, which includes a first network element and a fourth network element that can be used to execute any of the foregoing method embodiments. Figures 7 to 8B The method in ).

[0366] In addition, this application also provides a computer program for implementing the operations and / or processes performed by a communication device (such as any of the first network element, second network element, third network element, or fourth network element described above) in the method provided in this application.

[0367] This application also provides a computer-readable storage medium storing computer code, which, when executed on a computer, causes the computer to perform the operations and / or processes performed by a communication device (such as any of the first, second, third, or fourth network elements described above) in the method provided in this application.

[0368] This application also provides a computer program product, which includes computer code or a computer program. When the computer code or computer program is run on a computer, the operations and / or processes performed by the communication device (such as any of the first network element, second network element, third network element and / or fourth network element described above) in the method provided in this application are executed.

[0369] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, devices, or units, or it may be an electrical, mechanical, or other form of connection.

[0370] The units described as separate components may or may not be physically separate. 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 can be selected according to actual needs to achieve the technical effects of the solutions provided in the embodiments of this application.

[0371] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0372] If the integrated unit is implemented as 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 this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a computer-readable storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a first network element, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned computer-readable storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0373] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A communication method, characterized in that, Applied to the first network element, the method further includes: Based on the energy information of the first network element, training of the first model is stopped; Send a first message to the second network element, the first message being used to request the second network element to train the first model, the second network element being determined based on the energy information of the second network element.

2. The method according to claim 1, characterized in that, The first information includes model information of the first model, and the model information is used for model training of the first model; Wherein, the training of the first model was not completed before the training of the first model was stopped, and the model information of the first model is used to indicate the first model that has not completed training.

3. The method according to claim 1 or 2, characterized in that, The energy information of the first network element includes at least one of the following: the energy efficiency of the first network element, the normalized energy consumption of the first network element, the energy status of the first network element, the energy consumption of the first network element within a preset time, the new energy ratio of the first network element, the carbon emission factor of the first network element, the energy type of the first network element, the energy efficiency level of the first network element, or the energy consumption of the first network element.

4. The method according to claim 3, characterized in that, The step of stopping training the first model based on the energy information of the first network element includes: If the energy information of the first network element meets the first condition, training of the first model is stopped, wherein the first condition includes at least one of the following: The energy efficiency of the first network element is lower than the first energy efficiency threshold; the normalized energy consumption of the first network element exceeds the first normalized energy consumption threshold; the first network element enters an energy-saving state; the energy consumption of the first network element exceeds the first energy consumption threshold within a preset time; the renewable energy ratio of the first network element is lower than the first renewable energy ratio threshold; the carbon emission factor of the first network element is higher than the first carbon emission factor threshold; the energy type of the first network element changes; the energy efficiency level of the first network element is greater than the first energy efficiency level threshold; or, the energy consumption of the first network element exceeds the first network element energy consumption limit level.

5. The method according to any one of claims 1-4, characterized in that, The method further includes: Send a first message to the third network element, the first message being used to request the identification of the second network element; The system receives second information sent by the third network element, the second information being used to instruct the second network element.

6. The method according to claim 5, characterized in that, The first message includes third information, which indicates a second condition that the energy information of the second network element satisfies the second condition, wherein the second condition includes at least one of the following: The energy efficiency of the second grid element is lower than the second energy efficiency threshold; the normalized energy consumption of the second grid element exceeds the second normalized energy consumption threshold; the second grid element enters an energy-saving state; the energy consumption of the second grid element exceeds the second energy consumption threshold within a preset time; the renewable energy ratio of the second grid element is lower than the second renewable energy ratio threshold; the carbon emission factor of the second grid element is higher than the second carbon emission factor threshold; the energy type of the second grid element is the first energy type; the energy efficiency level of the second grid element is greater than the second energy efficiency level threshold; or, the energy consumption of the second grid element exceeds the energy consumption limit level of the second grid element.

7. The method according to any one of claims 1-6, characterized in that, Sending the first information to the second network element includes: A second message is sent to the second network element. The second message is used to request model training. The second message includes fourth information, which is used to instruct the second network element to notify the recipient of the training results of the model training.

8. The method according to any one of claims 4-7, characterized in that, The first condition is either pre-configured or obtained from other network elements.

9. A communication method, characterized in that, Applied to a third network element, the method further includes: Receive a first message sent by the first network element, the first message being used to request the determination of the second network element; A third message is sent to the first network element, the third message being used to instruct the second network element, the second network element being determined based on the energy information of the second network element.

10. The method according to claim 9, characterized in that, The first message includes third information, which indicates a second condition that the energy information of the second network element satisfies the second condition, wherein the second condition includes at least one of the following: The energy efficiency of the second grid element is lower than the second energy efficiency threshold; the normalized energy consumption of the second grid element exceeds the second normalized energy consumption threshold; the second grid element enters an energy-saving state; the energy consumption of the second grid element exceeds the second energy consumption threshold within a preset time; the renewable energy ratio of the second grid element is lower than the second renewable energy ratio threshold; the carbon emission factor of the second grid element is higher than the second carbon emission factor threshold; the energy type of the second grid element is the first energy type; the energy efficiency level of the second grid element is greater than the second energy efficiency level threshold; or, the energy consumption of the second grid element exceeds the energy consumption limit level of the second grid element.

11. A communication method, characterized in that, Applied to the fourth network element, the method further includes: The first network element is determined based on the first energy information; A first message is sent to the first network element. The first message is used to request training of the first model. The first message includes first information, which is related to the energy consumption requirements of the first model.

12. The method according to claim 11, characterized in that, The method further includes: The system receives the training results of the first model sent by the first network element, wherein the training results are obtained based on the energy consumption requirements.

13. The method according to claim 11 or 12, characterized in that, The first energy information is the first energy-saving capability, and determining the first network element includes: Send a second message to the third network element, the second message being used to indicate the first energy-saving capability; The system receives third information sent by the third network element, the third information being used to instruct the first network element that the first network element supports the first energy-saving capability.

14. The method according to claim 12, characterized in that, Determining the first network element includes: Acquire the energy-saving capabilities of multiple network elements; Based on the energy-saving capabilities of the multiple network elements, the first network element is determined, the first network element supports the first energy-saving capability, and the first network element belongs to the multiple network elements.

15. The method according to claim 13 or 14, characterized in that, The first energy-saving capability includes at least one of the following: The first model is trained using the first energy type; Alternatively, reduce energy consumption while ensuring the model accuracy is not lower than the model accuracy threshold. Alternatively, the carbon emission factor during the training process of the first model is not higher than the carbon emission factor threshold. Alternatively, the normalized energy consumption during the training process of the first model is not higher than the normalized energy consumption threshold. Alternatively, the energy state during the training process of the first model belongs to the first energy state; Alternatively, the first model can be trained based on the energy state timetable; Alternatively, the device used to train the first model does not have an energy inefficiency indication; Alternatively, the energy consumption during the training process of the first model does not exceed the energy consumption threshold. Alternatively, the device used to train the first model meets the energy efficiency rating performance indicators. Alternatively, the energy efficiency level of the energy used in the training process of the first model meets the first energy efficiency level; Alternatively, the device used to train the first model meets the network element energy consumption limit level.

16. The method according to claim 12, characterized in that, The first energy information is the network element information of each of at least one network element, and determining the first network element includes: Obtain network element information for each of the at least one network element, wherein the network element information includes the energy consumption and / or hardware capabilities required for a single iteration; Based on the network element information of each of the at least one network element, the first network element is determined, and the first network element belongs to the at least one network element.

17. The method according to any one of claims 12-16, characterized in that, The method further includes: Receive the fifth information sent by the first network element, the fifth information being used to indicate model training interruption and / or the reason for model training interruption; A second message is sent to the second network element. The second message is used to request training of the first model. The second message includes the first information.

18. The method according to any one of claims 12-17, characterized in that, The first information is used to indicate at least one of the following: energy type, model accuracy threshold, carbon emission factor, normalized energy consumption, energy status, energy status timetable, low energy efficiency indication, energy saving status migration execution indication, energy consumption threshold, energy efficiency level performance index, energy efficiency level, or network element energy consumption limit level.

19. A communication method, characterized in that, Applied to the first network element, the method further includes: Receive a first message sent by the fourth network element. The first message is used to request training of the first model. The first message includes first information, which is related to the energy consumption requirements of the first model. The training results of the first model are sent to the fourth network element, the training results being obtained based on the energy consumption requirements.

20. The method according to claim 19, characterized in that, The method further includes: Send a fifth message to the fourth network element, the fifth message being used to indicate the interruption of model training and / or the reason for the interruption of model training.

21. The method according to claim 19 or 20, characterized in that, The first information is used to indicate at least one of the following: energy type, model accuracy threshold, carbon emission factor, normalized energy consumption, energy status, energy status timetable, low energy efficiency indication, energy saving status migration execution indication, energy consumption threshold, energy efficiency level performance index, energy efficiency level, or network element energy consumption limit level.

22. A communication device, characterized in that, Includes units or modules for performing the method as described in any one of claims 1 to 21.

23. A communication device, characterized in that, The communication device includes at least one processor; the at least one processor is used to perform the method as described in any one of claims 1 to 21.

24. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a program or instructions that, when executed on the device, cause the device to perform the method as described in any one of claims 1 to 21.

25. A computer program product, characterized in that, The computer program product includes a program or instructions that, when executed by a device, cause the device to perform the method as described in any one of claims 1 to 21.