Information transmission method and communication device

Abstract

The embodiment of the invention provides an information transmission method and a communication device, and the method comprises the steps that a first network element sends a first request to a second network element, and receives a first response from the second network element; wherein the first request comprises first information, and the first information is used for indicating at least one first sample alignment algorithm supported by the first network element. The first response is used for indicating a second sample alignment algorithm determined by the second network element from at least one first sample alignment algorithm, and the second sample alignment algorithm is used for determining a common sample between a data set, used for longitudinal federated learning model training, of the first network element and a data set, used for longitudinal federated learning model training, of the second network element; or the first response is used for indicating that the execution of sample alignment is not supported. Thus, the first network element and the second network element can negotiate the sample alignment algorithm used in the sample alignment process through the first request and the first response, thereby improving the efficiency and success rate of executing sample alignment between the first network element and the second network element.

Description

Technical Field

[0001] This application relates to the field of communications, and in particular, to an information transmission method and a communication device. Background Art

[0002] Functional network elements in a communication network can utilize artificial intelligence (AI) technology and big data in the communication network, and through model training and inference, output analysis data to assist in formulating communication network strategies and adjusting network resources. For a functional network element to obtain a dataset, due to data privacy protection between different domains of the communication network, the functional network element may be unable to obtain the required data, which may further lead to infeasible model training or low accuracy.

[0003] Therefore, a distributed joint modeling method can be used for model training to break data islands. For example, multiple functional network elements (or called participants) can jointly model through vertical federated learning (VFL) without sharing raw data to achieve AI collaboration.

[0004] Currently, VFL requires that the datasets among all participants have the same sample space (or called a set of sample identifiers). Therefore, before the model training process starts, sample alignment is performed among all participants to determine common sample identifiers. However, how to perform sample alignment specifically to improve the efficiency and success rate of sample alignment remains to be studied. Summary of the Invention

[0005] The information transmission method and communication device provided in the embodiments of this application can improve the efficiency and success rate of sample alignment.

[0006] To achieve the above object, the embodiments of this application adopt the following technical solutions:

[0007] In a first aspect, an information transmission method is provided. This method can be executed by components of a first network element, such as a processor, a chip, or a chip system of the first network element, and can also be implemented by a logic module or software that can implement all or part of the first network element. The method includes: the first network element sends a first request to the second network element and receives a first response from the second network element. The first request includes first information, and the first information is used to indicate at least one first sample alignment algorithm supported by the first network element. The first response is used to indicate a second sample alignment algorithm among at least one first sample alignment algorithm, and the second sample alignment algorithm is used to determine common samples between the dataset of the first network element for vertical federated learning model training and the dataset of the second network element for vertical federated learning model training; or the first response is used to indicate that sample alignment is not supported for execution.

[0008] In the embodiments of the present application, by sending a first request from a first network element to a second network element, the second network element can determine whether the same sample alignment algorithm is supported between the second network element and the first network element according to at least one first sample alignment algorithm to perform sample alignment. Furthermore, when it is determined that the same sample alignment algorithm is supported between the second network element and the first network element, the second network element feeds back a second sample alignment algorithm to the first network element through a first response. Thus, the sample alignment algorithm used in the sample alignment process can be negotiated between the first network element and the second network element, thereby improving the efficiency and success rate of sample alignment between the first network element and the second network element. Further, when the second network element does not support at least one first sample alignment algorithm, the second network element can feed back through the first response that it does not support performing sample alignment, so that the first network element can promptly perform sample alignment with other participants in the vertical federated learning model training, thereby improving the efficiency of the first network element in performing sample alignment. Therefore, based on the information transmission method provided in the embodiments of the present application, the efficiency and success rate of the first network element in performing sample alignment can be improved.

[0009] In a possible implementation manner, the first network element may be a participant in vertical federated learning. For example, the first network element is the main participant in the vertical federated learning model training, and the second network element is the candidate slave participant in the vertical federated learning model training. Another example is that the first network element is a slave participant in the vertical federated learning model training, and the second network element is a candidate slave participant or the main participant in the vertical federated learning model training. That is to say, both the main participant and the slave participant can initiate the sample alignment negotiation process, that is, send the first request. It should be understood that without distinguishing between the main and slave participants, both the first network element and the second network element are participants in the vertical federated learning model training. In addition, since the participants in the vertical federated learning model training can be network function network elements or terminal devices, etc., the first network element can be a network function network element, or a terminal device, or an access network device. For example, the first network element may be a network data analysis function network element, an application function network element, an access and mobility management function network element, a session management function network element, or a policy control function network element, etc. In addition, similar to the first network element, the second network element can be a network function network element, or a terminal device, or an access network device.

[0010] In a possible implementation, the method provided in the first aspect further includes: a first network element receives information of a second network element from a third network element, where the information of the second network element includes an identifier of the second network element and / or address information of the second network element. It can be understood that the third network element may be a trusted third party. For example, the third network element may be a trusted coordinator (such as a VFL server); or, the third network element may be a vertical federated learning support function (VFLSF) network element, and the third network element is used to store the sample alignment capabilities of the participants in the vertical federated learning model training. That is to say, the first network element can obtain the identifier of the second network element and / or the address information of the second network element by receiving the information of the second network element from the third network element, so as to facilitate the first network element to send a first request to the second network element.

[0011] In a possible implementation, the method provided in the first aspect further includes: the first network element sends a second request to the third network element, where the second request includes second information, and the second information is used to indicate a sample alignment requirement. Correspondingly, the information received by the first network element from the second network element includes: the first network element receives a second response from the third network element, and the second response includes information of the second network element. That is to say, the second network element may be a network element whose sample alignment capability meets the sample alignment requirement corresponding to the first network element, thereby reducing the probability that the second network element does not support sample alignment, and thus improving the success rate and efficiency of sample alignment negotiation between the first network element and the second network element, so as to further improve the success rate and efficiency of sample alignment.

[0012] In a possible implementation, the sample alignment requirement includes at least one of the following: a third sample alignment type, a third sample alignment algorithm, or a third time period for sample alignment. That is to say, by one or several of the third sample alignment type, the third sample alignment algorithm, and the third time period for sample alignment included in the sample alignment requirement, a second network element that more meets the sample alignment requirement corresponding to the first network element can be screened out, thereby further reducing the probability that the second network element does not support performing sample alignment, and thus further improving the success rate and efficiency of sample alignment negotiation between the first network element and the second network element.

[0013] In a possible implementation, the information of the second network element further includes at least one of the following: a fourth sample alignment type, a fourth sample alignment algorithm, or a fourth time period supporting sample alignment. That is to say, the first network element can also obtain the sample alignment capability of the second network element, so that the first network element can determine the sample alignment algorithm and / or time period jointly supported by the first network element and the second network element, so as to facilitate the main participant to send a first request to the second network element based on the jointly supported sample alignment algorithm and / or jointly supported time period, and thus can more effectively perform sample alignment negotiation with the second network element.

[0014] In a possible implementation, the method provided by the first aspect further includes: a first network element sending a registration request to a third network element, where the registration request includes third information, and the third information is used to indicate at least one of the following: the sample alignment type supported by the first network element, the supported sample alignment algorithm, or the time period during which sample alignment is supported. That is to say, the first network element can register its sample alignment capabilities with the third network element, so that the third network element can obtain the sample alignment capabilities of the first network element (such as the supported sample alignment type, the supported sample alignment algorithm, or the time period during which sample alignment is supported, etc.), and select a second network element that meets the sample alignment requirements corresponding to the first network element according to the sample alignment capabilities of the first network element.

[0015] In a possible implementation, at least one first sample alignment algorithm is an algorithm jointly supported by the first network element and the second network element. That is to say, the first network element can, through a first request, indicate to the second network element at least one first sample alignment algorithm jointly supported by the first network element and the second network element, thereby reducing the probability that the second network element does not support the execution of sample alignment, and thus improving the success rate and efficiency of sample alignment negotiation.

[0016] In a possible implementation, the first information is further used to indicate at least one of the following: the number of samples supported by the first network element for sample alignment, the first sample alignment type supported by the first network element, or at least one first time period during which the first network element supports sample alignment. That is, by indicating to the second network element through the first information the number of samples supported by the first network element for sample alignment, the first sample alignment type, or at least one first time period, it can be used to assist the second network element in further determining whether it supports performing sample alignment, so as to avoid the failure of sample alignment caused by the local resources not meeting the computing power requirements of the sample alignment algorithm during the actual execution of sample alignment by the second network element, thereby improving the success rate and efficiency of sample alignment. For example, the number of samples supported by the first network element for sample alignment may refer to: the number of all samples in the dataset used by the first network element for vertical federated learning model training, or a subset of the number of all samples in the dataset. Among them, the subset of the number of all samples can be used to indicate that the first network element expects to use the number of samples corresponding to this subset for sample alignment. In addition, the number of samples supported by the first network element for sample alignment can be used by the second network element to determine the computing power requirements of the sample alignment algorithm, and then the second network element can determine whether the local resources can support using the first sample alignment algorithm to perform sample alignment during the execution of sample alignment, thereby improving the success rate and efficiency of sample alignment. Also, for example, the sample alignment type supported by the first network element can be the sample alignment type that the first network element can support, or can also be used to indicate the sample alignment type that the first network element expects to perform sample alignment. Among them, the sample alignment type can support two-party sample alignment and / or multi-party sample alignment. For a sample alignment algorithm that can support both two-party sample alignment and multi-party sample alignment, it is possible that the computing power requirements of this sample alignment algorithm for two-party sample alignment and multi-party sample alignment are different. Therefore, through the sample alignment type supported by the first network element, it can assist the second network element in determining the computing power requirements of the sample alignment algorithm, so as to facilitate the second network element in determining whether it supports sample alignment, thereby improving the success rate and efficiency of sample alignment. Again, for example, at least one time period during which the first network element supports sample alignment can be used by the second network element to determine whether it supports sample alignment. The second network element can determine whether it supports sample alignment according to this at least one time period and the load condition within this at least one time period, thereby improving the success rate and efficiency of sample alignment.

[0017] In a possible implementation, at least one first sample alignment algorithm is associated with at least one first time period. That is, by associating at least one first sample alignment algorithm with at least one first time period, the second network element can determine the first sample alignment algorithms supported by the first network element in different first time periods, and determine whether it supports the computing power requirements of the first sample alignment algorithms corresponding to different first time periods according to the local resources that can be called within different first time periods, so that the second network element can further determine whether it supports performing sample alignment, improving the success rate and efficiency of sample alignment.

[0018] In a possible implementation, the first response is further used to indicate the number of samples supported by the second network element for sample alignment, and / or, the second time period during which the second network element supports sample alignment; or, the first response is further used to indicate the reason why the second network element does not support sample alignment, and / or, the maximum number of samples expected by the second network element for the first network element to use. That is, in the case where the second network element supports performing sample alignment, by further indicating, in the first response, the number of samples supported by the second network element for sample alignment, and / or, the second time period during which the second network element supports sample alignment, it can assist the first network element in further determining whether it can support sample alignment during the process of performing sample alignment, thereby improving the success rate of sample alignment. Additionally, in the case where the second network element does not support performing sample alignment, by further indicating, in the first response, the reason why the second network element does not support sample alignment, and / or, the maximum number of samples expected by the second network element for the first network element to use, it can assist the first network element in adjusting parameters such as the first sample alignment algorithm indicated by the first information, or the number of samples supported by the first network element for sample alignment, or at least one first time period during which the first network element supports sample alignment when the first request is sent again later, improving the success rate of sample alignment negotiation.

[0019] In a possible implementation, the method provided in the first aspect further includes: the first network element sends a first message to the second network element, and the first message is used to indicate a target sample alignment algorithm, and the target sample alignment algorithm is a sample alignment algorithm determined from at least one first sample alignment algorithm according to the first response. That is, the first network element can further determine that sample alignment can be supported between the first network element and the second network element, and the target sample alignment algorithm used for performing sample alignment, according to the sample alignment response message of the second network element, so as to avoid sample alignment failure caused by not meeting the computing power requirements of the sample alignment algorithm during the process of performing sample alignment, and improve the success rate and efficiency of sample alignment.

[0020] In a possible implementation, the method provided by the first aspect further includes: a first network element sending a first request to a fourth network element; the first network element receiving a third response from the fourth network element, the third response being used to indicate a fifth sample alignment algorithm among at least one first sample alignment algorithm, the fifth sample alignment algorithm being used to determine common samples between a data set of the first network element for vertical federated learning model training and a data set of the fourth network element for vertical federated learning model training; the first network element sending a second message to the fourth network element, the second message being used to indicate a target sample alignment algorithm; wherein there are multiple fifth sample alignment algorithms and / or second sample alignment algorithms, and the target sample alignment algorithm is a sample alignment algorithm determined from at least one first sample alignment algorithm according to the second sample alignment algorithm, including: the target sample alignment algorithm is a sample alignment algorithm determined from at least one first sample alignment algorithm according to the second sample alignment algorithm and the fifth sample alignment algorithm.

[0021] That is to say, the first network element can send a first request to multiple network elements to obtain the sample alignment algorithms selected by the multiple network elements from at least one sample alignment algorithm, and then determine the target sample alignment algorithm for which sample alignment can be jointly performed between the first network element and the multiple network elements, thereby realizing multi-party sample alignment.

[0022] In a possible implementation, the first message is further used to indicate a target time period for sample alignment. That is to say, the first network element can also indicate the target time period for performing sample alignment through the first message to notify the second network element to perform the sample alignment process within this target time period, so as to prevent the second network element from performing the sample alignment process during a time period when its computing power requirements for the target sample alignment algorithm may not be met, and thus further improve the success rate and efficiency of sample alignment.

[0023] In a possible implementation, the method provided by the first aspect further includes: when the first network element determines not to perform sample alignment, the first network element sends a third message to the second network element, the third message being used to indicate not to perform sample alignment. That is to say, by sending the third message to the second network element to indicate not to perform sample alignment, the first network element can timely notify the second network element that sample alignment will not be performed next.

[0024] In a possible implementation, the third message is further used to indicate the reason for not performing sample alignment. That is to say, by indicating the reason for not performing sample alignment to the second network element, the first network element can assist the second network element in adjusting the second sample alignment algorithm indicated in the first response, the number of samples supporting sample alignment, or the time period supporting sample alignment when the first network element sends the first request again later, thereby improving the success rate and efficiency of sample alignment negotiation.

[0025] Second aspect, there is provided an information transmission method, which can be executed by components of a first network element, such as a processor, a chip, or a chip system of the first network element, and can also be implemented by a logic module or software that can implement all or part of the first network element. The method includes: the first network element sends a first request to the second network element and receives a first response from the second network element. Among them, the first request includes first indication information, and the first indication information is used to indicate providing sample identifiers for vertical federated learning model training. The first response includes a first sample identifier.

[0026] In the embodiments of the present application, since the first network element can make the second network element feedback the first sample identifier to the first network element by sending the first request to the second network element, the first network element can then determine the common sample identifiers among the participating parties in the vertical federated learning model training according to the first sample identifier, thereby realizing sample alignment. For example, the first network element can be a trusted third party, and the second network element can be the main participating party or the secondary participating party in vertical federated learning. Then, the first network element can send the first request to the second network element to obtain the sample identifiers corresponding to the data set for vertical federated learning model training of the second network element, so as to facilitate the first network element to determine the common samples among the participating parties in the same vertical federated learning model training, and then realize sample alignment. Another example is that the second network element can be a trusted third party for registering the sample identifiers corresponding to the data sets of the participating parties, and the first network element can be the main participating party, the secondary participating party, or the coordination party, etc. in vertical federated learning model training. Then, the second network element can respond to the first request in the first network element and feedback the common sample identifiers of at least two network elements participating in the same vertical federated learning model training to the first network element, thereby realizing sample alignment.

[0027] In a possible implementation manner, the first indication information being used to indicate providing sample identifiers for vertical federated learning model training includes: the first indication information is used to indicate providing the sample identifiers corresponding to the data set for vertical federated learning model training, and the first sample identifier is the sample identifier corresponding to the data set for vertical federated learning model training from the second network element.

[0028] That is to say, through the sample provision indication in the sample alignment request message, the first network element can indicate to the second network element to provide the sample identifiers corresponding to the data set for vertical federated learning model training of the second network element, so that the second network element can feedback the sample identifiers corresponding to the data set for vertical federated learning model training of the second network element to the first network element, and then can trigger the first network element to determine the common sample identifiers between the second network element and other candidate participating parties participating in vertical federated learning model training, thereby realizing sample alignment.

[0029] In a possible implementation, the second network element is a candidate network element participating in the training of the vertical federated learning model; the method provided by the second aspect further includes: the first network element receives the second sample identifier corresponding to the data set for training the vertical federated learning model from the third network element, and sends a third message to the second network element according to the first sample identifier and the second sample identifier. The third network element is other network elements except the second network element among the multiple candidate network elements participating in the training of the vertical federated learning model. The third message includes second indication information, and the second indication information is used to indicate that the second network element participates in the training of the vertical federated learning model, and / or, a third sample identifier, where the third sample identifier is a common sample identifier of at least two network elements participating in the training of the vertical federated learning model, and the at least two network elements are determined from the multiple candidate network elements according to the first sample identifier and the second sample identifier, and the at least two network elements include the third network element; or, the second indication information is used to indicate that the second network element does not participate in the training of the vertical federated learning model, and / or, the reason why the second network element does not participate in the training of the vertical federated learning model. That is to say, after the first network element obtains the sample identifiers corresponding to the data sets for training the vertical federated learning model provided by at least some of the multiple candidate network elements, it can further determine at least two network elements that can participate in the training of the vertical federated learning model and the common sample identifier (i.e., the third sample identifier) of the at least two network elements from the multiple candidate network elements according to the first sample identifier of the second network element and the second sample identifier of the third network element, thereby realizing sample alignment and increasing the probability of the second network element participating in the training of the vertical federated learning model. In addition, for the first network element to determine that the second network element does not participate in the training of the vertical federated learning model, the first network element sends a third message to the second network element to indicate that the second network element does not participate in the training of the vertical federated learning model, so that the second network element can timely determine not to participate in the current training of the vertical federated learning model, facilitating the second network element to prepare to execute other related processes of vertical federated learning. Further, by indicating the reason why the second network element does not participate in the training of the vertical federated learning model through the third message, it can facilitate the second network element to adjust the sample identifier corresponding to the data set provided for training the vertical federated learning model when participating in the training of the vertical federated learning model again, so as to increase the probability of the second network element participating in the training of the vertical federated learning model.

[0030] In a possible implementation, the second network element is a network element for registering sample identifiers corresponding to a data set of network elements participating in vertical federated learning model training; the first indication information is used to indicate the provision of sample identifiers for vertical federated learning model training, including: the first indication information is used to indicate the provision of common sample identifiers, and the common sample identifiers are the common sample identifiers between data sets of at least two network elements participating in the same vertical federated learning model training among multiple network elements, and the first sample identifier is used to indicate the common samples between data sets of at least two network elements participating in the first vertical federated learning model training among multiple network elements. That is to say, for a trusted third party where the second network element is a network element for registering sample identifiers corresponding to a data set of network elements participating in vertical federated learning model training, the first network element can send a first request to the second network element, so that the second network element can feedback to the first network element the common sample identifiers between data sets of at least two network elements participating in the same vertical federated learning model training among multiple network elements, thereby achieving sample alignment.

[0031] In a possible implementation, the first network element is a network element participating in vertical federated learning model training; the method provided in the second aspect further includes: the first network element sends a registration request to the second network element, and the registration request includes fourth indication information, and the fourth indication information is used to indicate the sample identifier corresponding to the data set of the first network element for vertical federated learning model training. That is to say, the first network element can send a registration request to the second network element to register the sample identifier corresponding to the data set of the first network element for vertical federated learning model training with the second network element, so that the second network element can determine the common sample identifiers between the first network element and at least one participant participating in the same vertical federated learning model training as the first network element, thereby achieving sample alignment.

[0032] In a possible implementation, the first request further includes fifth indication information and / or sixth indication information, the fifth indication information is used to indicate the quantity range of the sample identifiers included in the common sample identifiers, and the sixth indication information is further used to indicate the second network element to determine at least two network elements. That is to say, by indicating the quantity range of the samples indicated by the common sample identifiers through the fifth indication information, it is possible to avoid too many or too few common samples determined by the second network element, and thus the probability of the participants corresponding to the common sample identifiers joining the vertical federated learning model training can be increased. In addition, when the first request is not used to request to discover at least two network elements participating in the same vertical federated learning model training among multiple network elements, the second network element can be indicated to determine at least two network elements participating in the same vertical federated learning model training among multiple network elements through the sixth indication information

[0033] In a possible implementation, the first response further includes the identifiers and / or address information of at least two network elements, and / or the number of sample identifiers included in the first sample identifier. That is, by feeding back the identifiers and / or address information of at least two network elements to the first network element through the first response, the second network element can enable the first network element to notify the common sample identifier corresponding to the at least two network elements, thereby completing sample alignment. Additionally, for the case where the first network element is a VFL server and the encryption algorithm or decryption algorithm corresponding to the first sample identifier is not configured, by indicating the number of sample identifiers included in the first sample identifier to the first network element through the first response, the second network element can enable the first network element to determine the first sample identifier within the range of the common sample quantity that meets the requirements of vertical federated learning model training, thereby increasing the probability of the participating party joining the vertical federated learning model training.

[0034] In a third aspect, a communication device is provided for implementing the above various methods. The communication device may be the first network element in any of the above aspects or any of its implementations, or a device including the above first network element, or a device included in the above first network element, such as a chip. The communication device includes corresponding modules, units, or means for implementing the above methods, and the modules, units, or means may be implemented by hardware, software, or by hardware executing corresponding software. The hardware or software includes one or more modules or units corresponding to the above functions.

[0035] In some possible designs, the communication device may include a processing module and a transceiver module. The transceiver module, which may also be referred to as a transceiver unit, is used to implement the sending and / or receiving functions in any of the above aspects and any of its possible implementations. The transceiver module may be composed of a transceiver circuit, a transceiver, a transceiver, or a communication interface. The processing module may be used to implement the processing functions in any of the above aspects and any of its possible implementations.

[0036] In some possible designs, the transceiver module includes a sending module and a receiving module, which are respectively used to implement the sending and receiving functions in any of the above aspects and any of its possible implementations.

[0037] In a fourth aspect, a communication device is provided, including: at least one processor; the processor is used to execute a computer program or instruction, so that the communication device executes the method described in any of the above aspects.

[0038] In a possible implementation, the communication device further includes the memory. Optionally, the memory is coupled to the processor, and the memory may be integrated with the processor, or the memory may be independent of the processor. Optionally, the processor is used to execute the computer program or instruction stored in the memory.

[0039] In one possible implementation, the memory is independent of the communication device.

[0040] In one possible implementation, the communication device further includes a communication interface for communicating with modules outside the communication device.

[0041] The communication device may be the first network element in any of the above aspects or any of its implementation manners, or a device including the above first network element, or a device included in the above first network element, such as a chip.

[0042] In a fifth aspect, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program or instruction. When it runs on a communication device, the communication device can execute the method described in any of the above aspects or any of its implementation manners.

[0043] In a sixth aspect, a computer program product including instructions is provided. When it runs on a communication device, the communication device can execute the method described in any of the above aspects or any of its implementation manners.

[0044] In a seventh aspect, a communication device (for example, the communication device may be a chip or a chip system) is provided. The communication device includes a processor for implementing the functions involved in any of the above aspects or any of its implementation manners.

[0045] In some possible designs, the communication device includes a memory for storing necessary program instructions and data.

[0046] In some possible designs, when the device is a chip system, it may be composed of chips or may include chips and other discrete devices.

[0047] It can be understood that when the communication device provided in any of the third to seventh aspects is a chip, the above-mentioned sending action / function can be understood as output, and the above-mentioned receiving action / function can be understood as input.

[0048] Among them, the technical effects brought by any of the design manners in the third to seventh aspects can refer to the technical effects brought by different design manners in the first aspect above, and will not be elaborated here.

[0049] In an eighth aspect, a communication system is provided. The communication system includes: the first network element and the second network element in any of the above aspects or any of its implementation manners. The second network element is used to receive a first request and send a first response. Description of the Drawings

[0050] Figure 1 It is a schematic diagram of the feature distribution of a dataset for vertical federated learning provided by an embodiment of the present application;

[0051] Figure 2 It is a schematic diagram of a participant in vertical federated learning provided by an embodiment of the present application;

[0052] Figure 3 It is a schematic diagram of the training process of a vertical federated learning model provided by an embodiment of the present application;

[0053] Figure 4 It is a schematic diagram of sample alignment between two parties provided by an embodiment of the present application;

[0054] Figure 5 It is a schematic diagram of the architecture of a communication system provided by an embodiment of the present application;

[0055] Figure 6 It is a schematic diagram of the process of an information transmission method provided by an embodiment of the present application Figure 1 ;

[0056] Figure 7 It is a schematic diagram of the process of an information transmission method provided by an embodiment of the present application Figure 2 ;

[0057] Figure 8 It is a schematic diagram of the process of an information transmission method provided by an embodiment of the present application Figure 3 ;

[0058] Figure 9 It is a schematic diagram of the process of an information transmission method provided by an embodiment of the present application Figure 4 ;

[0059] Figure 10 It is a schematic diagram of the process of an information transmission method provided by an embodiment of the present application Figure 5 ;

[0060] Figure 11 It is a schematic diagram of the process of an information transmission method provided by an embodiment of the present application Figure 6 ;

[0061] Figure 12 It is a schematic diagram of the process of an information transmission method provided by an embodiment of the present application Figure 7 ;

[0062] Figure 13 It is a schematic diagram of the structure of a communication device provided by an embodiment of the present application Figure 1 ;

[0063] Figure 14 It is a schematic diagram of the structure of a communication device provided by an embodiment of the present application Figure 2 。 Detailed implementation manners

[0064] To facilitate the understanding of the technical solutions provided by the embodiments of this application, a brief introduction to the relevant technical terms of this application is first given. The brief introduction is as follows:

[0065] First, an intelligent network architecture based on the network data analytics function (NWDAF):

[0066] Currently, the 3rd generation partnership project (3GPP) has defined an intelligent network architecture based on NWDAF, aiming to collect a large amount of information in the network and utilize these data using existing big data and AI technologies to output some valuable information to assist operators in formulating strategies and adjusting network resources, thereby improving the user experience, reducing network load, etc.

[0067] For example, Table 1 lists some analysis results that the NWDAF network element can provide, as well as the data that the NWDAF network element needs to collect to provide the corresponding analysis results. For example, the NWDAF network element can provide service experience analysis results. To provide this analysis result, the NWDAF network element can collect service-related information such as service identity (ID) or service experience from the application function (AF) network element or the terminal device (or called user equipment (UE)), and collect information such as signal reception power and signal reception quality from operations, administration and management (OAM). NWDAF can train an AI model based on the collected data, and then obtain the inferred analysis result based on the AI model, such as obtaining the predicted service experience in a certain future time period.

[0068] Table 1

[0069]

[0070]

[0071] Terminal devices, core networks (CNs), or applications (APPs) within the network may not report terminal device or application-related data to NWDAF due to data privacy protection issues, resulting in NWDAF being unable to obtain the required data, making model training infeasible or the accuracy dropping significantly.

[0072] To solve the above problems, federated learning (FL) can be used for model training.

[0073] Second, Federated Learning:

[0074] A machine learning (ML) framework that can effectively help multiple institutions use data and perform machine learning modeling while meeting the requirements of user privacy protection, data security, and government regulations. As a distributed machine learning framework, federated learning can effectively solve the data silo problem, enabling participating parties to jointly model without sharing raw data, technically breaking down data silos and achieving AI collaboration.

[0075] Among them, according to the feature distributions between the data sets used by each participating party for federated learning model training, federated learning can be divided into three categories: horizontal federated learning, vertical federated learning, and federated transfer learning. For the feature distributions between the data sets of each participating party, it mainly refers to the distributions of the sample space and the feature space between the data sets of each participating party.

[0076] It should be understood that a data set can correspond to two dimensions. One dimension is the sample space, which can be used to represent (or characterize): the set of objects reflected or described by the data set; the other dimension is the feature space, which can be used to represent (or characterize): the set of features corresponding to the data reflecting or describing the object in the data set. For example, the object can refer to a user, an account, a terminal device, or an application, etc. The feature can refer to an attribute of the object, such as location, signal reception quality, or speed, etc.

[0077] Exemplarily, the object reflected or described by the data set is a group of UEs, that is, the sample space is this group of UEs. Each UE in this group of UEs can be called a sample, and different UEs in this group of UEs can be distinguished by sample identifiers. The features corresponding to the data reflecting or describing the object in the data set can be the location, signal reception power, or speed, etc. of the UEs in Table 1. For example, referring to Table 2, the data set of the participating party can be composed of the data of 3 samples, which are UE1, UE2, and UE3 respectively. The data of these 3 samples can reflect or describe the location feature and the signal reception power feature of the UEs. Among them, the sample identifier of sample UE1 is U1, and the data of sample UE1 includes location data #1 and signal reception power data #1. The sample identifier of sample UE2 is U2, and the data of sample UE2 includes location data #2 and signal reception power data #2. The sample identifier of sample UE3 is U3, and the data of sample UE3 includes location data #3 and signal reception power data #3.

[0078] In addition, the sample identifier may be an identifier of the UE, such as the UE's subscription concealed identifier (SUCI), subscription permanent identifier (SUPI), or international mobile subscriber identity (IMSI), etc., or it may be other identifiers that can be used to represent the identity of the UE. The embodiments of the present application do not make specific limitations on this.

[0079] It can be understood that since the sample identifier is the UE's SUCI, or SUPI, etc., the sample identifier involves the user's data privacy.

[0080] Table 2

[0081]

[0082] It should be understood that in the embodiments of the present application, the sample can also be understood as the row dimension in Table 2 above. In addition, Table 2 is only an example, and the sample can also be an account, an application, a perception target (such as a vehicle or traffic infrastructure), or an image, etc. The embodiments of the present application do not make specific limitations on this.

[0083] Based on the above descriptions of the data set, the sample space, and the feature space, the feature distribution between the data sets of each party in vertical federated learning is illustrated by examples below.

[0084] Third, vertical federated learning:

[0085] For vertical federated learning, its main feature distribution is that the sample space overlap degree between the data sets of each party is relatively high, and the feature space overlap degree is relatively low.

[0086] Figure 1 It is a schematic diagram of the feature distribution of the data set of a vertical federated learning provided by the embodiments of the present application. As Figure 1 shown, the data set of Party A contains the data of [User 1, User 2, User 3, User 4], and the data of each user contains [Feature 1, Feature 2, Feature 3, Feature 4, Feature 5]. The data set of Party B contains the data of [User 1, User 2, User 3, User 4, User 5], and the data of each user contains [Feature 5, Feature 6, Feature 7, Feature 8, Feature 9]. It can be seen that the intersection of the data features between the data sets of Party A and Party B is very small, but most of the samples are the same. Party A and Party B can select the same samples (i.e., the common samples (Users 1-4)) from their respective data sets for model training of vertical federated learning. In this way, the accuracy of the model can be improved by expanding the number of features.

[0087] It can be understood that the above Figure 1 is only an example. The samples between the data set of Party A and the data set of Party B can be exactly the same (for example, both include Users 1-4). There may be no intersection between the data features of the data set of Party A and the data set of Party B. For example, each sample in the data set of Party A contains features [Feature 1-5], and each sample in the data set of Party B contains data features [Feature 6-10].

[0088] The following introduces the parties involved in vertical federated learning, model training, and sample alignment.

[0089] 3.1. For the parties involved in vertical federated learning:

[0090] Figure 2 is a schematic diagram of the parties involved in vertical federated learning provided by an embodiment of the present application. As Figure 2 shown, the parties involved in vertical federated learning can be divided into three types, namely: the primary party, the secondary party, and the trusted collaborator. Among them, the primary party can refer to: for a specific federated learning task, the party that has the data labels corresponding to this task, and the data labels corresponding to this task are the results to be predicted or output by this task. For example, the federated learning task can be business experience prediction, and the data label can be business experience, that is, the data output by the federated learning task is the data of business experience.

[0091] The secondary party can refer to: for a specific federated learning task, the party that has some of the data features required for this task, that is, the input data required for the model corresponding to this task. For example, taking the data label of UE's business experience as an example, some of the data features required for this task may include location, average throughput rate, or average packet delay, etc.

[0092] The trusted collaborator (or called the trusted coordinator) can refer to: being responsible for maintaining functions such as the vertical federated learning process (VFL process), authorization admission and removal of federated learning members, etc., and also undertaking tasks such as encryption key distribution and decryption of intermediate information. The coordinator is usually served by a third party independent of the federated learning task or an institution with credibility. For example, the trusted coordinator can be the vertical federated learning server (VFL server).

[0093] It should be understood that the primary party or the secondary party participating in the vertical federated learning task can also be called the vertical federated learning client (VFL client).

[0094] In a possible implementation, the trusted coordinator can distribute the vertical federated learning model to multiple parties participating in the vertical federated learning task. These multiple parties can perform model training based on the local vertical federated learning model and upload the intermediate parameters of the model training to the trusted coordinator. The trusted coordinator determines whether to terminate the model training process and distributes the decrypted intermediate parameters, etc., to implement the training of the vertical federated learning model.

[0095] It can be understood that multiple encryption methods can be used in vertical federated learning to ensure the data privacy and security of the parties. Taking homomorphic encryption as an example, the process of training the vertical federated learning model will be described below by way of example.

[0096] 3.2. Model Training of Vertical Federated Learning:

[0097] Exemplarily, taking the participating party as VFL client A, the main participating party as VFL client B, and the trusted coordinator as VFL server, the process of training the vertical federated learning model can be seen in Figure 3 the process shown.

[0098] Figure 3 is a schematic diagram of the process of training a vertical federated learning model provided by an embodiment of the present application. As Figure 3 shown, the process of training the vertical federated learning model can include the following steps:

[0099] S301. Initialization. Among them, the initialization includes that VFL client A initializes the local parameter Θ A (or the model parameter of VFL client A), VFL client B initializes the local parameter Θ B (or the model parameter of VFL client B), and VFL server creates a homomorphic encryption key pair and sends the public key to VFL client A and VFL client B.

[0100] It can be understood that VFL server can also use other encryption methods besides homomorphic encryption to ensure the data privacy and security of the parties, which is not limited herein.

[0101] S302. VFL client A calculates the local parameter and Among them, the symbol represents homomorphic encryption. It should be understood that the result after adding (multiplying) homomorphic encryption ciphertexts is equal to the ciphertext of the result of adding (multiplying) plaintexts. For example

[0102] It can be determined according to the local parameter Θ A and the data of the common samples. Among them, the common samples are the common samples between the dataset for vertical federated learning model training of VFL client A and the dataset for vertical federated learning model training of VFL client B, that is, the same sample space between the two datasets. It can be understood that, as Figure 1 shown, in vertical federated learning, the common samples are first selected from the datasets of multiple parties, that is, Figure 1 the samples in: Users 1-4. In this way, Party A can use the sample data corresponding to Users 1-4 in the dataset of Party A for vertical federated learning model training, and Party B can use the sample data corresponding to Users 1-4 in the dataset of Party B for vertical federated learning model training.

[0103] Suppose the dataset D A provided by VFL client A for vertical federated learning model training B and the dataset D A provided by VFL client B for vertical federated learning model training. The common sample identifiers between them can include: n sample identifiers. In this way, the samples of VFL client A for vertical federated learning model training with VFL client B include the n samples corresponding to the n sample identifiers determined from D . VFL client A determines

[0104] For example, and the relationship between Θ A can be determined according to formula (1).

[0105]

[0106] Among them, represents the data of the i-th sample in.

[0107] It can be understood that formula (1) is only an example, and other methods can also be used to determine A and For example equals the transpose of Θ A or the conjugate transpose multiplied by the product between and . The embodiments of the present application do not make specific limitations on this.

[0108] L AIt can represent the loss function on the VFL client A side. λ represents the regularization parameter.

[0109] S303. The VFL client A sends and Correspondingly, the VFL client B receives from the VFL client A and

[0110] S304: The VFL client B calculates the local parameter intermediate parameter and the loss function

[0111] Among them, The calculation of is similar to and can be determined according to the local parameter Θ B and the data of the common samples for the vertical federated learning model training of the VFL client B (that is, the data of n samples determined from D B in For example,

[0112] The parameter d i can be determined according to the parameter and the data label y provided by the VFL client B i For example

[0113] The loss function The L in A L B and L AB can be determined. Among them,

[0114] S305. The VFL client B sends to the VFL server Correspondingly, the VFL server receives from the VFL client B

[0115] S306. The VFL server uses the private key to decrypt to obtain the decrypted loss function and determine whether to terminate the model training iteration process according to the loss function.

[0116] S307. The VFL client B sends to the VFL client A Accordingly, VFL client A receives from VFL client B

[0117] S308. VFL client A calculates the local gradient and adds the random noise of VFL client A to obtain wherein can be determined according to d i , Θ A , and the regularization parameter λ, for example

[0118] It should be understood that the random noise of VFL client A can also be referred to as the random mask of VFL client A

[0119] S309. VFL client B calculates the local gradient and adds the random noise of VFL client B to obtain wherein is calculated in a similar manner to and can be determined according to d , Θ i , and the regularization parameter λ, for example B

[0120] It should be understood that the random noise of VFL client B can also be referred to as the random mask of VFL client B

[0121] S310. VFL client A sends to the VFL server Accordingly, the VFL server receives from VFL client A

[0122] S311. VFL client B sends to the VFL server Accordingly, the VFL server receives from VFL client B

[0123] S312. The VFL server decrypts using the private key and to obtain the decrypted results and ​

[0124] S313. The VFL server sends to the VFL client A Correspondingly, the VFL client A receives from the VFL server

[0125] S314. The VFL server sends to the VFL client B Correspondingly, the VFL client B receives from the VFL server

[0126] S315. The VFL client A updates the local parameters based on the decrypted gradient where η represents the learning rate.

[0127] S316. The VFL client B updates the local parameters based on the decrypted gradient where η represents the learning rate.

[0128] It can be understood that, as Figure 3 shown, what is exchanged during the model training process are all intermediate calculation results. Therefore, during the entire model training process, the participating parties (such as the VFL client A and the VFL client B) do not know the data and features of the other party, and after the model training is completed, the participating parties only obtain the model parameters on their own side, protecting the local data privacy of the participating parties.

[0129] In addition, Figure 3 the vertical federated learning model training shown is for two-party vertical federated learning model training, that is, the participating parties in vertical federated learning include a main participating party and a subordinate participating party. It can be understood that vertical federated learning can also be extended to scenarios with more participating parties (or called multi-party vertical federated learning). For example, the participating parties in vertical federated learning can include a main participating party and at least two subordinate participating parties.

[0130] It should be understood that the multi-party vertical federated learning model training is similar to Figure 3 the two-party vertical federated learning model training shown, and will not be elaborated here.

[0131] It should also be understood that Figure 3 the two-party vertical federated learning model training algorithm shown is only an example. The model training algorithms for two-party or multi-party vertical federated learning can also adopt other algorithms, such as multi-party multi-classification vertical federated learning based on privacy-preserving label sharing, or multi-party vertical federated learning algorithms based on secret sharing, etc. The embodiments of the present application do not make specific limitations on this.

[0132] ​​It can be understood that vertical federated learning inference is completed through the collaboration between the primary participant and the secondary participants after the model training of vertical federated learning is completed (i.e., the participants have local parameters). For example Figure 3 in the primary participant VFLclient B in Figure 3 receives a business prediction request. VFL client B can send the sample identifier to be predicted (used to indicate the sample to be predicted) and the model type identifier to the secondary participant VFL client A, which can enable the secondary participant VFL client A to be based on the sample identifier to be predicted, the model type identifier, the local parameter Θ4, and the local data x for inference A , to obtain the local inference result u of VFLclient A A , and then VFL client A can send the local inference result u to VFL client B A , so that VFL client B can be based on u A and the local inference result u of VFL client B B , to determine the inference result u A +u B , u A +u B is the output inference result of vertical federated learning inference.

[0133] It should be understood that as Figure 1 shown, vertical federated learning requires the participants to have the same sample space, that is, common samples. Among them, before the start of vertical federated learning model training, it is necessary to first select multiple participants in vertical federated learning (including the primary participant and the secondary participants), and the common sample identifiers between the data sets of the multiple participants for vertical federated learning model training. Then in Figure 3 the step S302 shown, VFL client A can determine the sample data of n common samples to determine the parameter and in Figure 3 the step S304 shown, VFL client B can determine the sample data of n common samples to determine the parameter

[0134] However, the data sets of each participant are different, that is, the participants do not naturally have the same sample space. Therefore, before the start of vertical federated learning model training, an important step is sample alignment, which is specifically introduced below.

[0135] 3.3. Sample Alignment:

[0136] The purpose of sample alignment is to determine the common samples (or common sample identifiers) between the datasets used by each participating party (such as the primary participating party and the secondary participating party) for vertical federated learning model training. Considering data privacy protection, sample alignment also requires meeting the following conditions:

[0137] Condition 1: Each participating party cannot disclose other sample identifiers except the common sample identifiers to each other.

[0138] Condition 2: Each participating party cannot disclose any sample information to an untrusted third party to avoid being attacked by malicious third parties.

[0139] In other words, each participating party cannot perform sample alignment by directly interacting with the sample identifiers of their respective datasets used for vertical federated learning model training. Instead, they encrypt the sample identifiers of the datasets through a sample alignment algorithm and then perform sample alignment.

[0140] In addition, sample alignment can be divided into sample alignment without a third-party participant and sample alignment with a third-party participant. Among them, sample alignment without a third-party participant can refer to the situation where each participating party performs sample alignment based on a sample alignment algorithm and there is no third party independent of each participating party involved in the process of sample alignment. Sample alignment with a third-party participant can refer to the situation where sample alignment can be achieved by introducing a trusted third party independent of the participating parties.

[0141] For sample alignment without a third-party participant, a commonly used technique is private set intersection (PSI). It can be understood that based on the number of participating parties in vertical federated learning, PSI can also be divided into a two-party sample alignment algorithm and a multi-party sample alignment algorithm. The two-party sample alignment algorithm can include the following algorithms: PSI based on Diffie-Hellman, PSI based on homomorphic encryption, PSI based on Rivest-Shamir-Adleman (RSA) and hash algorithm, PSI based on RSA blind signature, or PSI based on oblivious transfer pseudo-random function (OT-PRF).

[0142] The multi-party sample alignment algorithm can include: Freedman multi-party secure intersection protocol, or oblivious programmable pseudo-random function (OPPRF).

[0143] It can be understood that the above method for aligning without a third-party sample is only an exemplary introduction, and other methods or technologies can also be used to achieve alignment without a third-party sample. The embodiments of the present application do not make specific limitations in this regard.

[0144] Figure 4 This is a schematic diagram of two-party sample alignment provided by an embodiment of the present application. As Figure 4 shown, the data set provided by the primary participant for longitudinal federated learning model training is the Figure 4 data set 1 in, and the data set provided by the secondary participant for longitudinal federated learning model training is the Figure 4 data set 2 in. Among them, the data set 1 includes 3 samples, and the sample identifiers of the 3 samples are U1, U2, and U3 respectively. The feature space of the 3 samples includes feature X3, and the label is Y. The data set 2 also includes 3 samples, and the sample identifiers of the 3 samples are U1, U2, and U4 respectively. The feature space of the 3 samples includes feature X4 and feature X5.

[0145] The primary participant and the secondary participant can perform the sample alignment process using a two-party sample alignment algorithm (such as PSI based on Diffie-Hellman, PSI based on homomorphic encryption, or PSI based on RSA blind signature, etc.). By interacting encrypted information between the primary participant and the secondary participant, it can be determined that the common sample identifiers between the data set 1 and the data set 2 are: sample identifier U1 and sample identifier U2.

[0146] In addition, as Figure 4 shown, after the secondary participant determines the common sample identifiers (i.e., U1 and U2), it can determine the data set 3 for longitudinal federated learning model training with the primary participant. The data set 3 includes the data of the sample corresponding to the sample identifier U1 selected from the data set 2, and the data of the sample corresponding to the sample identifier U2. Then, the secondary participant can execute Figure 3 step S302 in. Similarly, the primary participant can determine the data set 4 for longitudinal federated learning model training with the secondary participant according to the common sample identifiers. The data set 4 includes the data of the sample corresponding to the sample identifier U1 selected from the data set 1, and the data of the sample corresponding to the sample identifier U2. Then, the primary participant can execute Figure 3 step S304 in.

[0147] It should be understood that the above Figure 4 is only an exemplary description of the sample alignment process without a third-party participation. In the actual sample alignment process, for different sample alignment algorithms, the computing power (or computational amount) requirements for each participant involved in the sample alignment may be the same or different. The following lists several examples to illustrate the computing power requirements between each participant in the sample alignment process without a third-party participation.

[0148] Example 1:

[0149] For the sample alignment algorithm of Diffie - Hellman - based PSI, a large prime number p is agreed upon between Party A and Party B, and the sample alignment algorithm has the same computing power requirements for Party A and Party B.

[0150] For example, the sample alignment process of Diffie - Hellman - based PSI includes the following steps:

[0151] Step ①: Party A performs hash encryption on the sample identifier x1 to obtain H(x1), and it is required that H(x1) is a primitive root of p.

[0152] Step ②: Party B performs hash encryption on the sample y1 to obtain H(y1), and it is required that H(y1) is a primitive root of p.

[0153] Step ③: Party A randomly generates an integer a (kept secret externally), calculates Ka = (H(x1)) a mod p, and sends Ka to Party B, where mod represents modulo division or taking the remainder.

[0154] Step ④: Party B randomly generates an integer b (kept secret externally), calculates Kb = (H(y1)) b mod p, and sends Kb to Party A.

[0155] Step ⑤: After receiving Kb, Party A calculates Kba = (Kb) a mod p, and sends Kba to Party B.

[0156] Step ⑥: After receiving Ka, Party B calculates Kab = (Ka) b mod p, and sends Kab to Party A.

[0157] It can be understood that if x1 = y1, then according to the principle of the Diffie - Hellman algorithm, Kba = Kab. Therefore, Party A or Party B can determine whether x1 and y1 are the same by comparing whether Kba and Kab are the same.

[0158] Furthermore, when Party A has sample identifiers: x1, x2, …, x n , and Party B has sample identifiers: y1, y2, …, y m, Party A and Party B exchange encrypted values based on the above sample alignment algorithm. Subsequently, Party A calculates [Kba1, Kba2, …, Kbam] and receives [Kab1, Kab2, …, Kabn] from Party B. Thus, Party A takes the intersection of these two sets of data to obtain the common sample identifiers between Party A and Party B. Similarly, Party B can also obtain this common sample identifier, that is, Party B calculates [Kab1, Kab2, …, Kabn] and receives [Kbal, Kba2, …, Kbam] from Party A, and then takes the intersection of these two sets of data.

[0159] It can be understood that according to the relevant description of the sample alignment process of the Diffie-Hellman-based PSI above, the computational complexity between Party A and Party B is the same, that is, the sample alignment algorithm of the Diffie-Hellman-based PSI has the same computing power requirements for both parties.

[0160] Example 2:

[0161] For the RSA and hash algorithm-based PSI, one of Party A and Party B can generate the public key (n, e) and the private key (n, d), and send the public key to the other party. This method has different computing power requirements for Party A and Party B.

[0162] For example, the sample alignment process of the RSA and hash algorithm-based PSI includes the following steps:

[0163] Step 1: Party A uses the RSA algorithm to generate the public key (n, e) and the private key (n, d), and sends the public key (n, e) to Party B.

[0164] Step 2: Party B determines the sample identifier set y B = [y1, y2, …, y i ,.., y k corresponding parameters Y, and sends Y to Party A.

[0165] Among them, y i ∈y B . y B is the sample identifier set of Party B for sample alignment. r i is the random number generated by Party A corresponding to y i , 1 < r i < n, and r i is relatively prime to n. is the result of Party B encrypting r i using the public key (n, e). H(y i) is y i It is obtained by substituting into the hash function H(·).

[0166] Step 3: Party A performs a preliminary calculation on Y according to the private key (n, d) to obtain Y'. Among them,

[0167] Step 4: Party A determines the sample identification set x A = [x1, x2,..., x i ,..., x l corresponding parameter X.

[0168] Among them,

[0169] It can be understood that x A is the sample identification set of Party A for sample alignment, and X is the encrypted information after RSA encryption and hash processing of x A .

[0170] Step 5: Party A sends Y' and X to Party B.

[0171] It can be understood that Step 4 can also be executed before Step 3. For example, Step 4 can be executed simultaneously with Step 1, and Party A can send X and the public key (n, e) to Party B simultaneously. The embodiments of the present application do not make specific limitations on this.

[0172] Step 6: Party B determines the parameter D i according to Y' and r B , and takes the intersection of D B and X to obtain the encrypted common sample identification set I.

[0173] Among them,

[0174] It can be understood that D B is the encrypted information after Y' is de - blinded (i.e., removing r i ), and thus both D B and X are encrypted information after RSA encryption and hash processing. Therefore, the intersection can be taken between D B and X to obtain the encrypted common sample identification set I. Party B can determine the plaintext common sample identification set I' according to the encrypted common sample identification set I.

[0175] Step 7: Party B sends the encrypted common sample identification set I to Party A.

[0176] It can be understood that after Party A receives the encrypted set I of common sample identifiers, it can determine the plaintext set I' of common sample identifiers.

[0177] It can also be understood that according to the above description of the sample alignment process of PSI based on RSA and hash algorithms, the computational complexity of Party B is greater than that of Party A, that is, the sample alignment algorithm based on RSA and hash algorithms has different computing power requirements for the two parties.

[0178] It should be understood that for PSI based on homomorphic encryption, it is similar to PSI based on RSA and hash algorithms, and this method has different computing power requirements between the two parties.

[0179] In addition, for the multi-party sample alignment algorithm, similar to the two-party sample alignment algorithm, the computing power requirements between each participating party may be the same or different, specifically depending on the actual multi-party sample alignment algorithm used.

[0180] However, the following problems exist in sample alignment, namely:

[0181] Problem 1: For sample alignment without a third-party participation, since sample alignment requires hash encryption, homomorphic encryption, or RSA encryption, etc. for all sample identifiers of the dataset used for vertical federated learning, the computing power requirements for the participating parties are relatively high, and there may be differences in the sample space size (i.e., the number of sample identifiers) between the participating parties. In this case, different sample alignment algorithms may have different requirements for the computational amount of each participating party, which may cause one or more of the participating parties to be unable to support the computing power required for sample alignment, and thus unable to achieve sample alignment or unable to effectively perform sample alignment. For example, for a sample alignment algorithm with different computing power requirements for each participating party, if the computational amount of sample alignment is large, the participating party with a heavier load may be unable to support the completion of the sample alignment process.

[0182] Problem 2: For sample alignment with a third-party participation, when a trusted third party independent of the vertical federated learning participants participates in sample alignment, there is currently no implementation solution for how to specifically implement sample alignment.

[0183] In view of the above technical problems, the embodiments of the present application propose the following technical solutions. The technical solutions in the present application will be described below in conjunction with the accompanying drawings.

[0184] The technical solutions of the embodiments of this application can be applied to various communication systems, such as wireless network systems, vehicle-to-everything (V2X) communication systems, device-to-device (D2D) communication systems, vehicle networking communication systems, 4G mobile communication systems, such as long term evolution (LTE) systems, worldwide interoperability for microwave access (WiMAX) communication systems, 5G mobile communication systems, such as new radio (NR) systems, and future communication systems, etc.

[0185] In the embodiments of this application, "indication" can include direct indication and indirect indication, and can also include explicit indication and implicit indication. If the information indicated by a certain piece of information (such as the first indication information, the second indication information, or the third indication information, etc. below) is called the information to be indicated, then in the specific implementation process, there are many ways to indicate the information to be indicated. For example, but not limited to, the information to be indicated can be directly indicated, such as the information to be indicated itself or the index of the information to be indicated, etc. It is also possible to indirectly indicate the information to be indicated by indicating other information, where there is an association relationship between the other information and the information to be indicated. It is also possible to only indicate a part of the information to be indicated, while the other parts of the information to be indicated are known or pre-agreed. For example, it is also possible to use the arrangement order of each piece of information pre-agreed (such as stipulated by a protocol) to achieve the indication of specific information, thereby reducing the indication overhead to a certain extent. At the same time, the common parts of each piece of information can be identified and indicated uniformly to reduce the indication overhead caused by indicating the same information separately.

[0186] In addition, the specific indication method can also be various existing indication methods. For example, but not limited to, the above indication methods and their various combinations, etc. The specific details of various indication methods can refer to the prior art and will not be elaborated herein. As can be seen from the above, for example, when it is necessary to indicate multiple pieces of information of the same type, there may be a situation where the indication methods of different pieces of information are different. In the specific implementation process, the required indication method can be selected according to specific needs. The embodiments of this application do not limit the selected indication method. In this way, the indication methods involved in the embodiments of this application should be understood to cover various methods that can enable the party to be indicated to obtain the information to be indicated.

[0187] "Pre - defined" or "pre - configured" can be achieved by pre - storing corresponding codes, tables or other means that can be used to indicate relevant information in the device. The embodiments of the present application do not limit the specific implementation methods thereof. Among them, "storing" may refer to storing in one or more memories. The one or more memories may be separately provided, or may be integrated in an encoder or a decoder, a processor, or a communication device. The one or more memories may also be partly separately provided and partly integrated in a decoder, a processor, or a communication device. The type of the memory may be any form of storage medium, and the embodiments of the present application do not limit this.

[0188] The "protocol" involved in the embodiments of the present application may refer to a protocol family in the communication field, a standard protocol with a frame structure similar to that of a protocol family, or a relevant protocol applied to a future communication system. The embodiments of the present application do not make specific limitations on this.

[0189] In the embodiments of the present application, descriptions such as "when...", "in the case of...", "if", and "if" all mean that the device will perform corresponding processing under a certain objective situation, which does not limit time, and does not require the device to have a judgment action during implementation, nor does it mean other limitations exist.

[0190] In the description of the embodiments of the present application, unless otherwise specified, " / " indicates that the objects associated before and after are in an "or" relationship. For example, A / B may represent A or B. The "and / or" in the embodiments of the present application is merely a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. Also, in the description of the embodiments of the present application, unless otherwise specified, "a plurality of" means two or more than two. "At least one (item)" or its similar expression refers to any combination of these items, including any combination of a single item or plural items. For example, at least one (item) of a, b, or c may represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, and c can be single or multiple. Additionally, for the convenience of clearly describing the technical solutions of the embodiments of the present application, in the embodiments of the present application, terms such as "first" and "second" are used to distinguish the same items or similar items with basically the same functions and roles. Those skilled in the art can understand that terms such as "first" and "second" do not limit the quantity and execution order, and "first", "second", etc. do not necessarily mean different. At the same time, in the embodiments of the present application, words such as "exemplary" or "for example" are used to represent examples, illustrations, or explanations. Any embodiment or design solution described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or having more advantages than other embodiments or design solutions. Exactly speaking, using words such as "exemplary" or "for example" aims to present relevant concepts in a specific way for easy understanding.

[0191] The network architecture and service scenarios described in the embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those of ordinary skill in the art know that with the evolution of the network architecture and the emergence of new service scenarios, the technical solutions provided by the embodiments of the present application are equally applicable to similar technical problems.

[0192] To facilitate the understanding of the embodiments of the present application, first, Figure 5 Take the communication system shown in Figure 5 as an example to illustrate in detail the communication system applicable to the embodiments of the present application. Exemplarily,

[0193] Figure 5 is a schematic diagram of the architecture of a communication system provided by the embodiments of the present application. As shown in Figure 5As shown, the communication system may include one or more terminal devices, access network devices, and core network devices. Among them, the core network device may deploy network function network elements, such as network data analysis function network elements (or NWDAF network elements), application function network elements (or AF network elements), access and mobility management function (AMF) network elements, session management function (SMF) network elements, policy control function (PCF) network elements, network repository function (NRF) network elements, or network exposure function (NEF) network elements, etc., without limitation.

[0194] Among them, Figure 5 The terminal device may be within the beam / cell coverage range of the access network device, and the access network device may provide communication services for the terminal device.

[0195] Among them, Figure 5 The terminal device in it may be a device with wireless transceiver functions or a chip or chip system that can be set in the device, which can allow users to access the network and is a device for providing voice and / or data connectivity to users. The terminal device may also be referred to as UE, subscriber unit, terminal, mobile station (MS), or mobile terminal (MT), etc.

[0196] Exemplarily, Figure 5The terminal device in it can be a mobile phone, a tablet computer, or a computer with wireless transceiver functions. The terminal device can also be a user station, a mobile station, a remote station, a remote terminal device, a mobile terminal device, a user terminal device, a wireless communication device, a user agent, a user device, a cellular phone, a cordless phone, a session initiation protocol (SIP) phone, a wireless local loop (WLL) station, a personal digital assistant (PDA), a handheld device with wireless communication functions, a computing device, a processing device connected to a wireless modem, a vehicle-mounted device, a wearable device, a terminal device in the Internet of Things, a household appliance, a virtual reality (VR) terminal, an augmented reality (AR) terminal, a wireless terminal in industrial control, a wireless terminal in unmanned driving, a wireless terminal in telemedicine, a wireless terminal in a smart grid, a wireless terminal in a smart city, a wireless terminal in a smart home, a vehicle with vehicle-to-vehicle (V2V) communication capabilities, a connected vehicle, a drone with unmanned aerial vehicle to unmanned aerial vehicle (U2U) communication capabilities, a terminal device in a future network, or a terminal device in a future evolved public land mobile network (PLMN), etc., without limitation.

[0197] Among them, Figure 5 The access network device in it can be any device deployed in the access network that can perform wireless communication with the terminal device, can also be a chip or a chip system that can be set in the above device, can also be a logical node or a logical module or a function implemented in software, and can be used to implement functions such as wireless physical control functions, resource scheduling and wireless resource management, wireless access control, and mobility management. Specifically, the network device can be a device supporting wired access or a device supporting wireless access.

[0198] Exemplarily, the access network device may consist of one or more access network (AN) / radio access network (RAN) nodes. The AN / RAN nodes may be: a gNodeB (gNB), a transmission reception point (TRP), an evolved NodeB (eNB), a radio network controller (RNC), a NodeB (NB), a base station controller (BSC), a base transceiver station (BTS), a home base station (e.g., a home evolved NodeB, or a home NodeB, HNB), a baseband unit (BBU), or a wireless fidelity (Wi-Fi) access point (AP), etc.

[0199] In another example, the access network device may also be a device including a centralized unit (CU) node, or including a distributed unit (DU) node, or including a CU node and a DU node. For example, the access network device can be logically divided into a CU and a DU from the perspective of logical functions. The functions of some protocol layers are centrally controlled by the CU, and the functions of the remaining part or all protocol layers are distributed in the DU, and the DU is centrally controlled by the CU. Further, the centralized unit CU can be further divided into a control plane (CU-CP) and a user plane (CU-UP). In different systems, the CU (including CU-CP or CU-UP), or the DU may also have different names. For example, in an open radio access network (O-RAN) system, the CU may also be called an O-CU (open CU), the DU may also be called an O-DU, the CU-CP may also be called an O-CU-CP, and the CU-UP may also be called an O-CU-UP.

[0200] Among them, Figure 5The network data analysis function network element in it has functions such as data collection, model training, data analysis, and model inference. It can be used to collect relevant data from network function network elements, third-party service servers, terminal devices, or network management systems (such as OAM), perform data analysis or model training based on the relevant data, and provide data analysis results to network function network elements, third-party service servers, terminal devices, or network management systems, or provide the trained model to other network data analysis function network elements. The network data analysis function network element can be divided into an analysis logic function and a model training logic function according to its functions. Among them, the analysis logic function is the logic function in the network data analysis function network element, which is used to execute model inference, derive analysis results (that is, derive statistical or predictive analysis results according to the requests of analysis consumers), and open the analysis results. The model training logic function is the logic function in the network data analysis function network element, which is used to train the model and open the training service (for example, provide the trained model). A network data analysis function network element may only contain the analysis logic function or only contain the model training logic function, or may contain both the analysis logic function and the model training logic function. The embodiments of this application do not make specific limitations on this.

[0201] Among them, Figure 5 The application function network element in it is mainly an intermediate function entity that provides the interaction between the application server in the data network (DN) and the network elements in the core network, and transfers the requirements of the application side for the network side (such as quality of service requirements or user status event subscriptions, etc.). The application server can use it to achieve dynamic control of network service quality and billing, obtain the operation information of a certain network element in the core network, etc. In the embodiments of this application, the application function network element can be a function entity deployed by an operator or a function entity deployed by a service provider. The service provider can be a third-party service provider or a service provider within the operator, without limitation.

[0202] Among them, Figure 5 The access and mobility management function network element in it is mainly responsible for terminal device access authentication, mobility management, signaling interaction between various function network elements, and termination of non-access stratum (NAS) layer signaling security, etc., such as: managing the user's registration status, reachability status, N1 / N2 interface signaling transmission, access authentication and authorization, user connection status, user registration to the network, tracking area update, cell handover user authentication, key security, etc.

[0203] Among them, Figure 5The session management function network element therein mainly provides session management for the sessions of terminal devices (such as session establishment, modification, and release), Internet Protocol (IP) address allocation and management, and selection and control of user plane network elements, etc.

[0204] Among them, Figure 5 The policy control function network element therein is mainly responsible for generating policies such as access policies for terminal devices and service quality flow control policies, and can also provide the generated policies to the access and mobility management function network element, or the session management function network element, etc.

[0205] Among them, Figure 5 The network storage function network element therein can provide the storage ability for subscription data, policy data, and data related to capability opening.

[0206] Among them, Figure 5 The network opening function network element therein is mainly responsible for opening the capabilities and events of the network to external access entities (such as AF network elements), and receiving relevant external information (such as receiving information provided by the AF network element).

[0207] Next, in combination with Figure 5 Describe the participants in vertical federated learning.

[0208] The main participant or slave participant in vertical federated learning can be the above-mentioned network function network elements such as the network data analysis function network element, application function network element, access and mobility management function network element, session management function network element, or policy control function network element, etc. It can be understood that the main participant or slave participant in vertical federated learning can also be a terminal device, or an access network device, etc., and the embodiments of the present application do not make specific limitations thereto.

[0209] Among them, Figure 5 The trusted coordinator among the participants in vertical federated learning shown can be a function network element independently deployed in the core network, or can be deployed in Figure 5 The network function network element therein (such as the network storage function network element, or the network data analysis function network element, etc.). It can be understood that the trusted coordinator can also be referred to as the vertical federated learning server (VFLserver), and the embodiments of the present application do not make specific limitations thereto.

[0210] Optionally, Figure 5 The communication system shown further includes a vertical federated learning support function (VFL support function, VFLSF) network element, which is responsible for the registration and discovery of the participants in vertical federated learning. It should be understood that the VFLSF network element can be independently deployed, or can be deployed on the above-mentioned network data analysis function network element, application function network element, network storage function network element, or other network elements (such as the network opening function network element).

[0211] In addition, the trusted coordinator may be deployed together with the VFLSF network element in the same network element, or may be deployed separately. The embodiments of the present application do not make specific limitations on this.

[0212] It should be noted that the terminal device, access network device, and core network device in the embodiments of the present application may each be one or more chips, or may be a system on chip (SOC), etc. Figure 5 These are only exemplary drawings, and the number of devices or equipment included therein is not limited. Figure 5 The names of the various devices or equipment and the naming of the various links in [the drawings] are not limited. Except Figure 5 for the names shown, the various devices or equipment and the various links may also be named other names, which are not limited.

[0213] In addition, except Figure 5 for the devices or network elements shown, the communication system may further include other devices or network elements, such as Figure 5 the OAM, user plane function (UPF) network element, or application server shown, etc., which are not limited.

[0214] It should be understood that with the evolution of the network, Figure 5 the communication system shown may also support or include other network functions, such as functions related to AI or perception, etc. In addition, Figure 5 the names of the network elements shown may also change in the future evolved network. The embodiments of the present application do not make specific limitations on this.

[0215] Regarding Problem 1, the embodiments of the present application provide an information transmission method. The execution subject of this method may be a first network element. The first network element may be a participant in the vertical federated learning model training (such as a main participant or a secondary participant). The first network element may be Figure 5 a network data analysis function network element, application function network element, access and mobility management function network element, session management function network element, or policy control function network element, etc. in [the network], or may be a terminal device, or an access network device, etc. The embodiments of the present application do not make specific limitations on this. It should be understood that the second network element may also be a participant in the vertical federated learning model training.

[0216] In a possible implementation, a first network element sends a first request to a second network element and receives a first response from the second network element. The first request includes first information, which is used to indicate at least one first sample alignment algorithm supported by the first network element. The first response is used to indicate the second sample alignment algorithm determined by the second network element from the at least one first sample alignment algorithm, or that sample alignment is not supported. The second sample alignment algorithm is used to determine the common samples between the dataset for longitudinal federated learning model training of the first network element and the dataset for longitudinal federated learning model training of the second network element. In this way, by sending the first request to the second network element, the first network element can enable the second network element to determine whether the same sample alignment algorithm is supported between the second network element and the first network element according to the at least one first sample alignment algorithm to perform sample alignment. Furthermore, when it is determined that the same sample alignment algorithm is supported between the second network element and the first network element, the second network element feeds back the second sample alignment algorithm to the first network element through the first response. Then, the sample alignment algorithm used in the sample alignment process can be negotiated between the first network element and the second network element, thereby improving the efficiency and success rate of sample alignment between the first network element and the second network element. Further, when the second network element does not support the at least one first sample alignment algorithm, the second network element can feedback through the first response that sample alignment is not supported, so that the first network element can perform sample alignment with other participants in the longitudinal federated learning model training in a timely manner, thereby improving the efficiency of the first network element in performing sample alignment. Therefore, based on the information transmission method provided in the embodiments of the present application, the efficiency and success rate of the first network element in performing sample alignment can be improved.

[0217] For problem 2, an embodiment of the present application provides an information transmission method. The execution subject of this method can be a first network element, and the first network element can be a trusted third party, such as Figure 5 the trusted coordinator in, or a VFLSF network element, etc.; or the first network element can be a participant in the longitudinal federated learning model training (such as the main participant or the secondary participant).

[0218] In a possible implementation, the first network element sends a first request to the second network element and receives a first response from the second network element. The first request includes first indication information for indicating a sample identifier for vertical federated learning model training, and the first response includes a first sample identifier. In this way, by sending the first request to the second network element, the first network element can cause the second network element to feedback the first sample identifier to the first network element. Furthermore, the first network element can determine the common sample identifiers among the participating parties involved in vertical federated learning model training based on the first sample identifier, thereby achieving sample alignment. For example, the first network element may be a VFL server, and the second network element may be the primary participant or secondary participant in vertical federated learning. Furthermore, the first network element can send the first request to the second network element to obtain the sample identifier corresponding to the dataset for vertical federated learning model training of the second network element, so as to facilitate the first network element to determine the common samples among the participating parties involved in the same vertical federated learning model training, and thus achieve sample alignment. Another example is that the second network element may be a VFLSF network element, and the first network element may be the primary participant, secondary participant, or coordinator in vertical federated learning model training, etc. Furthermore, the second network element can respond to the first request in the first network element and feedback the common sample identifiers of at least two network elements involved in the same vertical federated learning model training to the first network element, thereby achieving sample alignment.

[0219] For ease of understanding, the technical terms involved in the embodiments of the present application are first introduced uniformly below:

[0220] 1) Sample: The sample in the embodiments of the present application refers to the sample included in the dataset for vertical federated learning model training. Among them, the sample may be, for example, a user, an account, an application, a terminal device, a perception target (such as a vehicle or traffic infrastructure), or an image, etc. The sample can be indicated or characterized by a sample identifier. Specifically, reference can be made to Figure 1 Table 1 and Table 2.

[0221] 2) Common sample: A common sample refers to the common sample among the datasets for vertical federated learning model training of each participating party. For example, as Figure 1 shown, the common samples are Users 1 to 4.

[0222] In addition, the common sample can be indicated by a common sample identifier. For example, as Figure 4 shown, the common sample identifiers are Sample Identifiers U1 and U2, and the common samples are the samples indicated by Sample Identifier U1 and the samples indicated by Sample Identifier U2.

[0223] 3) Sample alignment: The purpose of sample alignment is to determine the common samples (or common sample identifiers) among the datasets of each participating party for training the vertical federated learning model. At the same time, each participating party cannot disclose other sample identifiers except the common sample identifiers, and each participating party cannot disclose any sample information to an untrusted third party.

[0224] It can be understood that when each participating party determines the common samples, each participating party can determine the samples used for training the vertical federated learning model, and then can perform the training of the vertical federated learning model.

[0225] It can also be understood that the common sample identifier or the name of the common sample is only an example. The common sample identifier can also be called an intersection sample ID or a public sample identifier, and the common sample can be called an intersection sample or a public sample, etc. The embodiments of the present application do not make specific limitations on this.

[0226] 4) Sample alignment types: Sample alignment types can be divided into two-party sample alignment and multi-party sample alignment. Two-party sample alignment can mean that the participating parties in the vertical federated learning model training include a primary participating party and a secondary participating party. Multi-party sample alignment can mean that the participating parties in the vertical federated learning model training include a primary participating party and at least two secondary participating parties. The participating parties in the vertical federated learning model training can support two-party sample alignment and / or multi-party sample alignment.

[0227] It can be understood that whether a participating party supports two-party sample alignment and / or multi-party sample alignment depends on the sample alignment algorithm supported by the participating party. If participating party A supports programmable oblivious pseudorandom functions and homomorphic encryption-based PSI, then participating party A supports two-party sample alignment and multi-party sample alignment; if participating party A supports two-party sample alignment algorithms such as Diffie-Hellman-based PSI or homomorphic encryption-based PSI, then participating party A only supports two-party sample alignment; if participating party A supports multi-party sample alignment algorithms such as Freedman multi-party secure intersection protocol or programmable oblivious pseudorandom functions, then participating party A only supports multi-party sample alignment.

[0228] 5) Sample alignment algorithm: The sample alignment algorithm is used to determine the common samples among the datasets of at least two participating parties for training the vertical federated learning model. For example, the sample alignment algorithm can encrypt the sample identifiers, and at least two participating parties determine the common samples among the datasets of at least two participating parties by interacting with the encrypted intermediate results, so as to avoid disclosing the sample identifiers and protect the user's private data.

[0229] The following will be combined with Figures 6 to 12, the interaction process between each network element / device in the above communication system is specifically introduced through method embodiments. The information transmission method provided in the embodiments of this application can be applied to the above Figure 5 shown communication system.

[0230] Figure 6 is a flowchart of an information transmission method provided in the embodiments of this application Figure 1 . The information transmission method is a sample alignment method for Problem 1 applicable to the case without a third party's participation, mainly involving the interaction between the primary participating party, at least one candidate secondary participating party, a trusted coordinator (i.e., VFL server), and the vertical federated learning support function (VFLSF) network element. Among them, sample alignment without a third party's participation means that there is no third party's participation in the sample alignment process between the primary participating party and at least one candidate secondary participating party. In other vertical federated learning-related processes (such as vertical federated learning model training and / or inference), a third party can participate or not participate, and the embodiments of this application do not make specific limitations on this.

[0231] In Figure 6 the information transmission method process shown, the primary participating party initiates vertical federated learning, and the VFL server is responsible for the discovery and selection of participating parties (or called vertical federated learning clients (VFL clients)).

[0232] As Figure 6 shown, the process of the information transmission method includes:

[0233] S601. The participating party sends a registration request message to the VFLSF network element. Correspondingly, the VFLSF network element receives the registration request message from the participating party.

[0234] Among them, the registration request message is used to indicate the sample alignment ability of the participating party. That is to say, the participating party sends a registration request message to the VFLSF network element to register the sample alignment ability of the participating party with the VFLSF network element, so that the VFL server or other participating parties initiating vertical federated learning can discover the participating party, improving the efficiency and success rate of discovering the participating party.

[0235] In addition, the VFLSF network element in step S601 can be Figure 5 the VFLSF network element in Figure 5 , or a network storage function network element, or a network open function network element. The participating party in step S601 can be

[0236] In a possible implementation, the sample alignment ability includes at least one of the following: the type of sample alignment supported by the participating party, the supported sample alignment algorithm, or the time period during which sample alignment is supported.

[0237] Among them, the type of sample alignment can include two-party sample alignment and multi-party sample alignment. Furthermore, the type of sample alignment supported by the participating party can refer to: the participating party only supports two-party sample alignment, or the participating party only supports multi-party sample alignment, or the participating party supports both two-party sample alignment and multi-party sample alignment. Additionally, the type of sample supported by the participating party can also be used to indicate that the participating party does not support sample alignment. For example, the protocol can stipulate that type a of the type of sample alignment supported by the participating party indicates no support for sample alignment. In other words, the type of sample alignment supported by the participating party can be used to indicate whether the participating party supports sample alignment, or whether the participating party supports two-party sample alignment, or whether the participating party supports multi-party sample alignment, or whether the participating party supports both two-party sample alignment and multi-party sample alignment simultaneously.

[0238] It can be understood that for a participating party that does not support sample alignment, for example, it may be that the participating party has not pre-configured the sample alignment algorithm.

[0239] For example, the type of sample alignment supported by the participating party can be indicated by parameter a included in the registration request message. The several possibilities of the type of sample alignment supported by the participating party above can be represented by different parameters. For instance, parameter a1 is used to indicate whether the participating party supports sample alignment, parameter a2 is used to indicate whether the participating party supports two-party sample alignment, parameter a3 is used to indicate whether the participating party supports multi-party sample alignment, parameter a4 is used to indicate whether the participating party supports both two-party sample alignment and multi-party sample alignment, etc. Additionally, the several possibilities of the type of sample alignment supported by the participating party above can also be indicated by other parameters or combinations of parameters, and the embodiments of this application do not make specific limitations in this regard.

[0240] Furthermore, the type of sample alignment algorithm supported by the participating party can be determined through the type of sample alignment supported by the participating party. For example, by the participating party supporting two-party sample alignment, it can be determined that the sample alignment algorithm supported by the participating party includes the two-party sample alignment algorithm. Another example is that by the participating party supporting multi-party sample alignment, it can be determined that the sample alignment algorithm supported by the participating party includes the multi-party sample alignment algorithm.

[0241] The sample alignment algorithms supported by the participating parties may include at least one sample alignment algorithm. That is, through the sample alignment algorithms supported by the participating parties, the specific sample alignment algorithms supported by the participating parties can be determined, so as to facilitate the subsequent discovery of the participating parties by the VFL server. In addition, the sample alignment algorithms may be the two-party sample alignment algorithms and / or multi-party sample alignment algorithms in the aforementioned "sample alignment algorithms", such as PSI based on Diffie-Hellman, PSI based on homomorphic encryption, PSI based on RSA and hash algorithms, or Freedman multi-party secure intersection protocol, etc.

[0242] Exemplarily, the sample alignment algorithms supported by the participating parties can be represented by the sample alignment algorithms supported by parameters. The parameter can be a list of values used to represent a set of sample alignment algorithms. For example, the parameter can be used to indicate [PSI based on Diffie-Hellman, PSI based on homomorphic encryption, PSI based on RSA and hash algorithms], and thus through this parameter, it can be determined that the participating parties support using the above three algorithms for sample alignment. Another example is that the protocol can stipulate, or the participating parties and the VFLSF network element can negotiate in advance to use symbols (such as a set of numbers and / or characters) to represent a sample alignment algorithm. For example, use "01" to represent PSI based on Diffie-Hellman, use "02" to represent PSI based on homomorphic encryption, and use "03" to represent PSI based on RSA and hash algorithms. If the participating party only supports PSI based on Diffie-Hellman and PSI based on RSA and hash algorithms, then this parameter can be represented as [01, 03].

[0243] It should be understood that the sample alignment algorithms may also be other sample alignment algorithms other than PSI, and the embodiments of the present application do not make specific limitations in this regard.

[0244] The time period during which the participating party supports sample alignment can be used to represent the time period during which the participating party supports (or expects) to perform sample alignment. This time period can be determined by the participating party according to changes in its own computing power or load, etc. For example, if the participating party determines that its load is low during the time periods of [9:00 - 11:00] and [14:00 - 17:00] and can support the execution of sample alignment, the participating party can determine that the time periods for supporting the execution of sample alignment are [9:00 - 11:00] and [14:00 - 17:00]. That is, through the time when the participating party supports sample alignment, the specific time periods supported by the participating party for sample alignment can be determined, so as to further improve the efficiency of subsequent discovery of the participating party by the VFL server.

[0245] It can be understood that the time period during which the participating party supports sample alignment can also be indicated by a parameter. For example, the time period during which the participating party supports sample alignment can be indicated by the parameter "supported sample alignment time period", which can indicate a start time (e.g., 9:00) and an end time (e.g., 11:00); or, this parameter can indicate a start time (e.g., 9:00 or 14:00) and a duration (e.g., 1 hour, or 2 hours, etc.); or, this parameter can also indicate a time offset and a reference time, and the time offset can be a time shift relative to the reference time, and thus the two time endpoints corresponding to the time period for performing sample alignment can be determined through the time offset and the reference time. For example, one of the two time endpoints can be the reference time, and the other time endpoint can be the difference or sum value between the reference time and the time offset; or, the two time endpoints are respectively the difference between the reference time and the time offset and the sum value between the reference time and the time offset.

[0246] In addition, the time period regarding sample alignment involved in the embodiments of the present application can be understood as the time period during which the participating party can perform sample alignment. For example, the participating party can start performing sample alignment at the start time of the above time period, or can also start performing sample alignment at other moments within the above time period other than the start time. The embodiments of the present application do not make specific limitations on this.

[0247] It should be understood that the time indicated above can be standard time, and the standard time can be coordinated universal time (UTC), global positioning system (GPS) time, long range navigator (LORAN) time, or international atomic time (TAI), etc. The embodiments of the present application do not make specific limitations on this. In addition, the standard time can also be referred to as absolute time, which objectively exists and is independent of any special reference system. Each participating party in vertical federated learning and the VFLSF network element have the same understanding of the same absolute time.

[0248] In addition, the above implementation manners for indicating the time period during which the participating party supports sample alignment are only examples, and other ways can also be used to indicate the time period during which the participating party supports sample alignment. The embodiments of the present application do not make specific limitations on this.

[0249] Exemplarily, the above sample alignment capability can be indicated by a parameter sample alignment capability indication, which can be used to indicate the type of sample alignment supported by the above parties, the supported sample alignment algorithm, or the time period during which sample alignment is supported.

[0250] It can be understood that a participating party can send a registration request message to the VFLSF network element according to the identifier or address of the VFLSF network element. Among them, the identifier or address of the VFLSF network element can be pre-configured by the protocol or received by the participating party from other network elements. For example, the participating party can receive the identifier or address of the VFLSF network element from the network storage function network element. The embodiments of the present application do not make specific limitations on this.

[0251] For example, the address of the VFLSF network element can be an Internet Protocol (IP) address, or a Fully Qualified Domain Name (FQDN), or a Uniform Resource Locator (URL).

[0252] It should be understood that in the embodiments of the present application, the participating parties can be classified into Figure 2 the primary participating party and the secondary participating party as shown, or the participating parties in the embodiments of the present application can also be classified into the primary participating party and the secondary participating party, and this is not limited herein.

[0253] For ease of understanding, the following takes the classification of participating parties into the primary participating party and the secondary participating party as an example for illustration.

[0254] In a possible implementation manner, the participating parties in the embodiments of the present application are multiple participating parties, and the multiple participating parties can be divided into the primary participating party and the secondary participating party. Accordingly, step S601 may include:

[0255] S601a. The primary participating party sends a registration request message to the VFLSF network element. Correspondingly, the VFLSF network element receives the registration request message from the primary participating party. Among them, the registration request message is used to indicate the sample alignment capability of the primary participating party.

[0256] S601b. The secondary participating party sends a registration request message to the VFLSF network element. Correspondingly, the VFLSF network element receives the registration request message from the secondary participating party. Among them, the registration request message is used to indicate the sample alignment capability of the secondary participating party.

[0257] It can be understood that the primary participating party can be the Figure 5 application function network element, the network data analysis function network element, or other network function network elements in Figure 5The access and mobility management function network element, or the session management function network element) in it, and the participating party can also be an application function network element, a network data analysis function network element, or other network function network elements. The embodiments of the present application do not make specific limitations on this.

[0258] S602. The VFLSF network element sends a registration response message to the participating party. Correspondingly, the participating party receives the registration response message from the VFLSF network element.

[0259] Among them, the registration response message is used to indicate whether the registration is successful. For example, the registration response message may include a result indication, and this parameter can be used to indicate whether the registration is successful.

[0260] It can be understood that, as described in step S601, the participating party is multiple participating parties, and these multiple participating parties can be divided into a primary participating party and a secondary participating party. Further, step S602 may include:

[0261] S602a. The VFLSF network element sends a registration response message to the primary participating party. Correspondingly, the primary participating party receives the registration response message from the VFLSF network element.

[0262] S602b. The VFLSF network element sends a registration response message to the secondary participating party. Correspondingly, the secondary participating party receives the registration response message from the VFLSF network element.

[0263] It can be understood that the above steps S601 and S602 are the sample alignment ability registration processes. After the sample alignment ability of the participating party is successfully registered, it can prepare for the subsequent discovery and selection of the participating party in order to initiate vertical federated learning. In addition, when the sample alignment ability of the participating party (such as the sample alignment type supported by the participating party, the supported sample alignment algorithm, or the time period supporting sample alignment) changes, the participating party can update the sample alignment ability of the participating party by executing the above sample alignment ability registration process again.

[0264] It should be understood that the registration request message in the above step S601 can also be used to register other information of the participating party in addition to the sample alignment ability, such as the identifier of the participating party, the address, the type of vertical federated learning ability supported by the participating party, the vertical federated learning group identifier (VFL group ID) supported by the participating party, or the model type identifier (model type ID(s)) supported by the participating party, etc. The embodiments of the present application do not make specific limitations on this.

[0265] It can be understood that the type of vertical federated learning capabilities supported by a participant can be used to indicate whether the participant supports being a participant in vertical federated learning, or to indicate whether the participant supports being the primary participant and / or secondary participant in vertical federated learning. Among them, whether a participant supports being a participant in vertical federated learning can refer to whether the participant supports being a participant in vertical federated learning model training and / or vertical federated learning inference. Additionally, whether a participant supports being the primary participant and / or secondary participant in vertical federated learning can refer to whether the participant supports providing data labels and / or data features for the federated learning task.

[0266] In addition, the VFL server can also send a registration request to the VFLSF network element to request registration of the VFL server's information. The VFL server's information can include: the identifier of the VFL server, the address, the type of vertical federated learning capabilities supported by the VFL server, the vertical federated learning group identifier (VFL group ID) supported by the VFL server, or the model type identifier(s) (model type ID(s)) supported by the VFL server, etc.; alternatively, the VFLSF network element can pre-configure the VFL server's information, and the embodiments of this application do not make specific limitations in this regard.

[0267] It can be understood that the type of vertical federated learning capabilities supported by the VFL server can be used to indicate whether the VFL server supports being the coordinator in vertical federated learning, that is, whether the VFL server supports functions such as being responsible for maintaining the vertical federated learning process, authorizing access to and removing federated learning members.

[0268] In addition, in the embodiments of this application, the vertical federated learning process can include: a sample alignment process, a vertical federated learning model training process, or a vertical federated learning inference process. This is explained uniformly here and will not be elaborated further below.

[0269] It should be understood that in Figure 6In the process shown, the primary participant can initiate vertical federated learning. For example, the primary participant determines to train a model through vertical federated learning (for example, the primary participant discovers that the training data cannot be collected due to privacy protection issues), and then initiates vertical federated learning. In addition, in the vertical federated learning process, in addition to being responsible for decrypting and transmitting intermediate calculation results in the vertical federated learning model training process, the VFL server can also be responsible for discovering (or selecting) secondary participants. The primary participant can discover a VFL server that meets the vertical federated learning requirements initiated by the primary participant through the VFLSF network element, and then discover secondary participants through the VFL server to perform subsequent processes such as sample alignment, vertical federated learning model training, or vertical federated learning inference.

[0270] The following introduces step S603 and step S604, which are used for the primary participant to discover a VFL server that meets the vertical federated learning requirements initiated by the primary participant.

[0271] S603. The primary participant sends a discovery request message to the VFLSF network element. Correspondingly, the VFLSF network element receives the discovery request message from the primary participant. The discovery request message is used to request the discovery of a VFL server that meets the vertical federated learning requirements initiated by the primary participant.

[0272] In a possible implementation manner, the discovery request message includes at least one of the following: a vertical federated learning group identifier, a vertical federated learning model type identifier, or a first vertical federated learning capability type indication.

[0273] Among them, the vertical federated learning group identifier (VFL group ID) may refer to the vertical federated learning group expected by the primary participant among multiple candidate vertical federated learning groups. The vertical federated learning group identifier can be used to determine the vertical federated learning group expected by the primary participant from multiple candidate vertical federated learning groups. Each candidate vertical federated learning group among multiple candidate vertical federated learning groups may include multiple secondary participants and / or at least one VFL server. It can be understood that multiple candidate vertical federated learning groups can be pre-configured, or negotiated in advance among the primary participant, secondary participants, and the VFLSF network element, or indicated by the VFLSF network element. The embodiments of the present application do not make specific limitations in this regard.

[0274] In addition, the discovery request message may include one or more vertical federated learning group identifiers. The discovery request message includes multiple vertical federated learning group identifiers, that is, the primary participant expects to perform vertical federated learning based on multiple vertical federated learning groups.

[0275] The vertical federated learning model type identifier (model type ID(s)) may refer to the type of vertical federated learning model that the primary participant expects to train, such as a business experience model or a personalized recommendation model, etc. It can be understood that the type of vertical federated learning model and the corresponding identifier may be pre-configured by the protocol or provided by the VFL server. The embodiments of this application do not make specific limitations in this regard.

[0276] In addition, there may be multiple vertical federated learning model type identifiers, that is, the primary participant can obtain one or more VFL servers that support multiple vertical federated learning model type identifiers.

[0277] The first vertical federated learning capability type indication is used to indicate the discovery of the VFL server, and this indication information can be represented by the parameter VFL capability type = VFL server.

[0278] Optionally, the discovery request message can also indicate: the time period expected to perform vertical federated learning and the supported training mode. Among them, the time period for performing vertical federated learning can include: the time period for training the vertical federated learning model, the time period for performing vertical federated learning inference, or the time period for performing sample alignment. The supported training mode may refer to the mode or process of training the vertical federated learning model, and this mode or process can be Figure 3 the model training process shown, or it can be other model training processes. The embodiments of this application do not make specific limitations in this regard.

[0279] It can be understood that the expected time period for performing vertical federated learning and the supported training mode can be used by the VFLSF network element to better discover the VFL server that meets the needs of the primary participant.

[0280] In addition, the primary participant can send the actual code corresponding to the training mode; or the protocol can stipulate, or the participant and the VFLSF network element can negotiate in advance to use symbols (such as numbers and / or characters) to represent a training mode. For example, the symbol c1 can represent training mode 1, the symbol c2 can represent training mode 2, the symbol c3 can represent training mode 3, etc., or other methods can also be used to indicate the training mode. The embodiments of this application do not make specific limitations in this regard.

[0281] S604. The VFLSF network element sends a discovery response message to the primary participant. Correspondingly, the primary participant receives the discovery response message from the VFLSF network element.

[0282] Among them, the discovery response message is used to indicate information about at least one VFL server. The at least one VFL server can be at least one VFL server determined by the VFLSF network element from multiple VFL servers according to the discovery request message. The information about the at least one VFL server may include at least one VFL server instance, and each VFL server instance is represented by the identifier and / or address of the VFL server.

[0283] It can be understood that after the primary participant discovers the VFL server through steps S603 and S604, the secondary participant can be discovered through the VFL server, so as to facilitate subsequent sample alignment and vertical federated learning model training.

[0284] The following introduces steps S605 to S609. Steps S605 to S607 are used to discover candidate secondary participants (or potential secondary participants) who can perform sample alignment with the primary participant.

[0285] S605: The primary participant sends a vertical federated learning initiation request message to the VFL server. Correspondingly, the VFL server receives the vertical federated learning initiation request message from the primary participant.

[0286] Among them, the VFL server is the VFL server determined by the primary participant from at least one VFL server. For example, in the case where the discovery response message indicates multiple VFL servers, the primary participant can randomly select one VFL server from the multiple VFL servers. In addition, the primary participant can also select one VFL server from the multiple VFL servers in other ways, and the embodiments of the present application do not make specific limitations in this regard.

[0287] It can be understood that the vertical federated learning initiation request message is used to initiate the vertical federated learning process and may also include the meaning of indicating the selection of secondary participants or requesting the selection of secondary participants.

[0288] In a possible implementation manner, the vertical federated learning initiation request message includes the following parameters: a vertical federated learning group identifier (VFL group ID), and a vertical federated learning model type identifier (mode type ID(s)). It can be understood that the vertical federated learning group identifier and the vertical federated learning model type identifier can be used to select secondary participants. In addition, for the specific meanings of the vertical federated learning group identifier and the vertical federated learning model type identifier, reference can be made to the foregoing step S603, and details will not be elaborated here.

[0289] Optionally, the vertical federated learning initiation request message further includes the following parameters:

[0290] Master VFL client indicator, which is used to indicate that the initiator of the vertical federated learning initiation request message is the master participant;

[0291] VFL client selection flag, which is used to indicate the selection of participants or to indicate the selection of slave participants;

[0292] Alternatively, slave VFL client requirement, which is used to indicate the requirements for slave participants or the conditions for selecting slave participants.

[0293] It can be understood that vertical federated learning can be initiated by the master participant or by the slave participant. Thus, through the master VFL client indicator, it can be determined whether the initiator of the vertical federated learning initiation request message is the master participant or the slave participant, so as to facilitate the VFL server to determine whether it is necessary to select the master participant.

[0294] For example, if the master VFL client indicator is carried in the vertical federated learning initiation request message, the VFL server determines to select slave participants; if the slave VFL client indicator is carried in the vertical federated learning initiation request message to indicate that the initiator of the vertical federated learning initiation request message is the slave participant, the VFL server determines to select the master participant, or the VFL server determines to select the master participant and at least one slave participant.

[0295] For the VFL client selection flag, it is used to indicate the request to select participants (i.e., select VFL clients) or to indicate the selection of slave participants (i.e., select slave VFL clients). In the case where the vertical federated learning initiation request message includes the master VFL client indicator or the slave VFL client indicator, the vertical federated learning initiation request message may not carry the VFL client selection flag.

[0296] For the slave VFL client requirement, it may specifically include: time intervals supporting VFL, and / or, the desired training mode. Among them, the time intervals supporting VFL may include: the time intervals supporting the training of the vertical federated learning model, the time intervals supporting the inference of the vertical federated learning, or the time intervals supporting sample alignment. The desired training mode may refer to a specific training mode or training process expected to be used by the master participant.

[0297] It should be understood that the time periods supporting vertical federated learning can be one or more time periods, and the desired training modes can be one or more training modes. Among them, when the time periods supporting vertical federated learning are multiple time periods, the VFL server can select participants according to one of the time periods. Similarly, the desired training modes can be multiple training modes, and the VFL server can select participants according to one of the training modes.

[0298] It can be understood that after the VFL server receives the vertical federated learning initiation request message from the primary participant, the VFL server can determine the conditions for selecting secondary participants based on the parameters carried in the vertical federated learning initiation request message, that is, trigger the participant discovery process (or referred to as the VFL client discovery process). In addition, the participant discovery process can also be understood as the participant selection process (or referred to as the VFL client selection process).

[0299] The following introduces steps S606 - S607, which are the participant discovery process.

[0300] S606. The VFL server sends a participant discovery request message to the VFLSF network element. Correspondingly, the VFLSF network element receives the participant discovery request message from the VFL server.

[0301] Among them, the participant discovery request message is used to request the discovery of participants. The participant discovery request message may include: vertical federated learning group identifier, vertical federated learning model type identifier, requirements for secondary participants, indication of the second vertical federated learning capability type, and sample alignment requirements.

[0302] It can be understood that for the vertical federated learning group identifier and the vertical federated learning model type identifier, please refer to the relevant description in step S603 for details and will not be elaborated here.

[0303] In addition, when the vertical federated learning initiation request message in step S605 includes requirements for secondary participants, the requirements for secondary participants in step S606 are the same as those in step S605, or the requirements for secondary participants in step S606 are determined from the requirements for secondary participants in step S605. The embodiments of the present application do not make specific limitations on this.

[0304] When the vertical federated learning initiation request message in step S605 does not include requirements for secondary participants, the VFL server can determine the requirements for secondary participants by itself, or the VFL server can negotiate the requirements for secondary participants with the primary participant. The embodiments of the present application do not make specific limitations on this.

[0305] The second vertical federated learning capability type indication is used to indicate the discovery request for participating parties, for example, it can be represented by VFL capability type=slave VFL client.

[0306] The sample alignment requirement (or the required sample alignment capability) can be used to indicate the sample alignment capability that the requested participating party should support. This sample alignment capability can include, for example, the sample alignment type, the sample alignment algorithm, or the time period for sample alignment.

[0307] For example, the sample alignment requirement can indicate that the requested participating party should support the sample alignment type. For instance, it can indicate that the requested participating party should support sample alignment, or indicate that the requested participating party should support two-party sample alignment, or indicate that the requested participating party should support multi-party sample alignment, or indicate that the requested participating party should support both two-party sample alignment and multi-party sample alignment simultaneously.

[0308] It can be understood that the VFLSF network element can select participating parties based on the sample alignment type supported by the participating parties carried in the registration request message sent by the participating parties in step S601, and the sample alignment type that the above-mentioned requested participating party should support.

[0309] Again, for example, the sample alignment requirement can indicate the required sample alignment algorithm that the requested participating party should support. If the sample alignment algorithm that the requested participating party should support only includes one sample alignment algorithm, the discovered participating party supports this sample alignment algorithm. If the sample alignment algorithm that the requested participating party should support includes multiple sample alignment algorithms, the discovered participating party only needs to support any one of the multiple sample alignment algorithms, or it can also mean that the discovered participating party should support all of these multiple sample alignment algorithms simultaneously.

[0310] It can be understood that the VFLSF network element can select participating parties based on the sample alignment algorithm supported by the participating parties carried in the registration request message sent by the participating parties in step S601, and the sample alignment algorithm that step S606 indicates the requested participating party should support.

[0311] For another example, the sample alignment requirement may indicate an expected sample alignment time period. It can be understood that the VFLSF network element can select a participating party based on the supported sample alignment time period carried in the registration request message sent by the participating party in step S601, and the above-mentioned expected sample alignment time period.

[0312] It can be understood that when a participating party discovers that the request message carries a sample alignment requirement, it can enable the VFLSF network element to select a candidate secondary participating party (or potential secondary participating party) that can perform sample alignment with the primary participating party, thereby improving the success rate and efficiency of negotiating the sample alignment method (such as the sample alignment type, algorithm, and the time period for performing sample alignment) between the primary participating party and the candidate secondary participating party.

[0313] It should be understood that for the sample alignment capability, please refer to the relevant description of the sample alignment capability in step S601 for details, and it will not be elaborated here.

[0314] In addition, the VFL server can pre-configure multiple sample alignment types and multiple sample alignment algorithms to facilitate the VFL server to indicate the sample alignment capabilities that the requested participating parties should support. It can be understood that the multiple sample alignment types and multiple sample alignment algorithms can also be negotiated in advance between the VFL server and the VFLSF network element, or indicated by the VFLSF network element. The embodiments of this application do not make specific limitations in this regard.

[0315] It should be understood that the vertical federated learning initiation request message in step S605 may also include a sample alignment requirement, and the sample alignment requirement included in the request message discovered by the participating party in step S606 is one or more of the sample alignment requirements included in the vertical federated learning initiation request message in step S605. If the vertical federated learning initiation request message in step S605 does not include a sample alignment requirement, then the sample alignment requirement included in the request message discovered by the participating party in step S606 is determined by the VFL server, or the VFL server can negotiate the sample alignment requirement with the primary participating party. The embodiments of this application do not make specific limitations in this regard.

[0316] S607. The VFLSF network element sends a participating party discovery response message to the VFL server. Correspondingly, the VFL server receives the participating party discovery response message from the VFLSF network element.

[0317] Among them, the participating party discovers that the response message includes information of at least one candidate secondary participating party. The information of at least one candidate secondary participating party may include: the identifier or address of at least one candidate secondary participating party. The identifier or address of at least one candidate secondary participating party can be used by the VFL server or the primary participating party to communicate with the at least one candidate secondary participating party.

[0318] It should be understood that the at least one candidate secondary participating party may include n candidate secondary participating parties, where n is an integer greater than or equal to 1. For example, n can be 1, 2, 3, or a larger value. The embodiments of the present application do not make specific limitations on this.

[0319] In a possible implementation manner, the information of at least one candidate secondary participating party may further include: the sample alignment ability of each candidate secondary participating party among the at least one candidate secondary participating parties. The sample alignment ability of each candidate secondary participating party may include: the type of sample alignment supported by each candidate secondary participating party, the sample alignment algorithm supported, or the time period during which sample alignment is supported.

[0320] It can be understood that for the sample alignment ability, reference can be made to the relevant description in the foregoing step S601 for details, and no further elaboration will be provided here.

[0321] It should be understood that after the VFL server determines the candidate secondary participating parties that can perform sample alignment with the primary participating party, the VFL server can trigger the primary participating party and the at least one candidate secondary participating party to perform vertical federated learning through steps S608 and S609.

[0322] The following introduces steps S608 and S609.

[0323] S608. The VFL server sends a first vertical federated learning request message to the at least one candidate secondary participating party. Correspondingly, the at least one candidate secondary participating party receives the first vertical federated learning request message from the VFL server.

[0324] Among them, the first vertical federated learning request message is used to initiate a vertical federated learning process, or the first vertical federated learning request message is used to indicate the initiation of a vertical federated learning preparation process. That is to say, by sending the first vertical federated learning request message to the at least one candidate secondary participating party, the VFL server can notify the at least one candidate secondary participating party that a vertical federated learning process will be carried out next, so that the at least one candidate secondary participating party can prepare the relevant configurations for executing the vertical federated learning process in advance.

[0325] It can be understood that the at least one candidate secondary participating party is a candidate secondary participating party (or potential secondary participating party) that can perform sample alignment with the primary participating party.

[0326] In a possible implementation, the first vertical federated learning request message includes information about the primary participant. Among them, the information of the primary participant may include the identifier or address of the primary participant. In other words, the VFL server can notify the information of the primary participant to at least one candidate secondary participant capable of sample alignment with the primary participant through the first vertical federated learning request message. It should be understood that at least one candidate secondary participant can verify whether the next received message about sample alignment is sent by the primary participant based on the identifier or address of the primary participant to improve security.

[0327] Exemplarily, the information of the primary participant may further include the identifier of the vertical federated learning group supported by the primary participant, the identifier of the vertical federated learning model type supported, the supported training mode, or the supported sample alignment ability, etc. The supported sample alignment ability may include at least one of the following: the supported sample type, the supported sample alignment algorithm, or the time period for sample alignment. The embodiments of the present application do not make specific limitations in this regard.

[0328] Optionally, the first vertical federated learning request message may include a vertical federated learning preparation process identifier (VFLpreparation flag), and this vertical federated learning preparation process identifier can be used to indicate the initiation of the vertical federated learning preparation process. It can be understood that if the first vertical federated learning request message contains the meaning of indicating the initiation of the vertical federated learning preparation process, the first vertical federated learning request message may not carry the vertical federated learning preparation process identifier.

[0329] It can also be understood that the VFL server sending the first vertical federated learning request message to at least one candidate secondary participant may mean that the VFL server sends the first vertical federated learning request message to each candidate secondary participant among at least one candidate secondary participant.

[0330] S609. The VFL server sends a second vertical federated learning request message to the primary participant. Correspondingly, the primary participant receives the second vertical federated learning request message from the VFL server.

[0331] Among them, the similarity between the second vertical federated learning request message and the first vertical federated learning request message is that the second vertical federated learning request message can be used to indicate the initiation of the vertical federated learning process, or can also be used to indicate the initiation of the vertical federated learning preparation process. Additionally, in the case where the second vertical federated learning request message does not contain the meaning of indicating the initiation of the vertical federated learning preparation process, the second vertical federated learning request message may include the vertical federated learning preparation process identifier.

[0332] It should be understood that the difference between the second vertical federated learning request message and the first vertical federated learning request message is that the second vertical federated learning request message does not include the information of the primary participant, but includes the information of at least one candidate secondary participant in the above step S607. Among them, the information of at least one candidate secondary participant may include: the identifier and / or address of each candidate secondary participant in at least one candidate secondary participant. In addition, the information of at least one candidate secondary participant may also include the supported sample alignment capabilities.

[0333] The sample alignment capabilities supported by each candidate secondary participant in at least one candidate secondary participant may include at least one of the following: supported sample types, supported sample alignment algorithms, or time periods for supporting sample alignment. Additionally, for the sample alignment capabilities supported by each candidate secondary participant, specific reference may be made to the relevant description in step S601, which will not be elaborated here.

[0334] It can be understood that the identifier or address of each candidate secondary participant can be used for the primary participant to communicate with each candidate secondary participant. The sample alignment capabilities supported by each candidate secondary participant can be used for the primary participant to determine the sample alignment algorithm and / or time period jointly supported between the primary participant and at least one candidate secondary participant, so that the primary participant can more effectively negotiate the sample alignment algorithm and / or time period with at least one candidate secondary participant.

[0335] It should be understood that after the primary participant obtains the relevant information of at least one candidate secondary participant through the second vertical federated learning request message, the primary participant can initiate a sample alignment process. Among them, the sample alignment process may include a sample alignment negotiation process and a sample alignment execution process. The sample alignment negotiation process can be used to determine the target secondary participant for actually performing sample alignment, the actual target sample alignment algorithm used, and the target time period for actually performing sample alignment, etc. The sample alignment execution process may refer to the process of performing sample alignment between the primary participant and the target secondary participant, and the target secondary participant is the candidate secondary participant determined by the primary participant from at least one candidate secondary participant.

[0336] It can be understood that the sample alignment negotiation process is only an exemplary name, and it can also be called a sample alignment assistance process, or other names. The embodiments of the present application do not make specific limitations in this regard.

[0337] The following introduces steps S610 - S612, and steps S610 - S612 are the sample alignment negotiation process.

[0338] S610. The primary participant sends a sample alignment request message to at least one candidate secondary participant. Correspondingly, at least one candidate secondary participant receives the sample alignment request message from the primary participant.

[0339] Among them, the sample alignment request message can be used to request the execution of sample alignment, or the sample alignment request message can be used to negotiate the sample alignment method, and the sample alignment method can include the sample alignment type, the sample alignment algorithm, or the time period for sample alignment, etc.

[0340] In a possible implementation, the sample alignment request message includes first information, and the first information is used to indicate the first common sample alignment algorithm supported between the primary participant and at least one candidate secondary participant. That is to say, the primary participant can determine the first common sample alignment algorithm supported between the primary participant and at least one candidate secondary participant, and send it to at least one candidate secondary participant through the sample alignment request message to indicate that the first common sample alignment algorithm can be used to perform sample alignment, thereby reducing the probability that at least one candidate secondary participant does not support sample alignment, and thus improving the success rate and efficiency of sample alignment negotiation.

[0341] For example, the primary participant can determine the sample alignment algorithms supported by each candidate secondary participant among at least one candidate secondary participant according to the second vertical federated learning request message in step S609, and take the intersection between the sample alignment algorithms supported by the primary participant and the sample alignment algorithms supported by each candidate secondary participant, thereby obtaining an intersection containing the first common sample alignment algorithm.

[0342] In addition, the aforementioned intersection may include one or more first common sample alignment algorithms. Exemplarily, for the sake of convenience of expression, "01" is used to represent PSI based on Diffie-Hellman, "02" is used to represent PSI based on homomorphic encryption, "03" is used to represent PSI based on RSA and hash algorithm, and "04" is used to represent PSI based on OT-PRF. Suppose at least one candidate secondary participant includes candidate secondary participant 1 and candidate secondary participant 2, the sample alignment algorithms supported by candidate secondary participant 1 are [01, 02, 03], the sample alignment algorithms supported by candidate secondary participant 2 are [02, 03, 04], and the sample alignment algorithms supported by the primary participant are [01, 02, 03, 04]. Thus, the intersection among the primary participant, candidate secondary participant 1, and candidate secondary participant 2 is [02, 03], that is, the intersection contains 2 first common sample alignment algorithms.

[0343] It should be understood that if the aforementioned intersection only includes one sample alignment algorithm, the first information can indicate the one sample alignment algorithm; if the aforementioned intersection includes multiple sample alignment algorithms, the first information can indicate the multiple sample alignment algorithms, or indicate at least one of the multiple sample alignment algorithms (i.e., a subset of the intersection). The embodiments of the present application do not make specific limitations on this.

[0344] It can be understood that the primary participant can select at least one sample alignment algorithm with a relatively small computing power requirement for the candidate secondary participant from multiple sample alignment algorithms (for example, the computing power requirement for Participant A in Example 2 is relatively small). In this way, the probability that the candidate secondary participant does not support sample alignment due to a large computing power requirement can be reduced, thereby improving the success rate and efficiency of sample alignment negotiation.

[0345] In addition, the name of the first common sample alignment algorithm is only an example. The first common sample alignment algorithm can also be called available sample alignment algorithms, or other names, which are not specifically limited in the embodiments of this application.

[0346] In a possible implementation, the first information is further used to indicate at least one of the following: the number of samples supported by the primary participant for sample alignment, the type of sample alignment supported by the primary participant, or at least one time period during which the primary participant supports sample alignment.

[0347] Among them, the number of samples supported by the primary participant for sample alignment can be used by the candidate secondary participant to determine the computing power requirement of the sample alignment algorithm. For example, for a sample alignment algorithm with the same computing power requirement for each participant (such as the sample alignment algorithm based on Diffie-Hellman PSI in Example 1), the computing power requirement of the sample alignment algorithm is not only related to the number of samples of the candidate secondary participant itself, but also related to the number of samples of the primary participant. Furthermore, the candidate secondary participant can determine the computing power requirement of the first common sample alignment algorithm based on its own number of samples and the number of samples of the primary participant, so that the candidate secondary participant can determine whether its computing power can support using this first common sample alignment algorithm to perform sample alignment, improving the success rate of sample alignment.

[0348] In addition, the number of samples supported by the primary participant for sample alignment can refer to: the number of all samples in the dataset of the primary participant for vertical federated learning model training, or a subset of the number of all samples in this dataset. It can be understood that a subset of the number of all samples can be used to indicate that the primary participant expects to perform sample alignment using the number of samples corresponding to this subset.

[0349] It can be understood that in the case where the first common sample alignment algorithm is a sample alignment algorithm with the same computing power requirement for each participant, the primary participant indicates the number of samples supported by the primary participant for sample alignment through the first information. Or, in the case where the first common sample alignment algorithm is a sample alignment algorithm with different computing power requirements for each participant, the primary participant indicates the number of samples supported by the primary participant for sample alignment through the first information.

[0350] The sample alignment types supported by the primary participant can be the sample alignment types that the primary participant can support, or can be used to indicate the sample alignment types that the primary participant expects to perform sample alignment. For example, the sample alignment types supported by the primary participant can be supporting two-party sample alignment and / or multi-party sample alignment. It can be understood that for a sample alignment algorithm that can support both two-party sample alignment and multi-party sample alignment, perhaps the computing power requirements of this sample alignment algorithm for two-party sample alignment and multi-party sample alignment are different. Furthermore, through the sample alignment types supported by the primary participant, it can assist candidate secondary participants in determining the computing power requirements of the sample alignment algorithm, so as to facilitate candidate secondary participants in determining whether to support sample alignment.

[0351] At least one time period during which the primary participant supports sample alignment can be used for candidate secondary participants to determine whether to support sample alignment. For example, candidate secondary participants can determine whether to support sample alignment based on this at least one time period and the load conditions of the candidate secondary participants within this at least one time period. For instance, if the load of the candidate secondary participant within this at least one time period is less than threshold 1, the candidate secondary participant can provide sufficient local resources for sample alignment within this at least one time period, that is, the candidate secondary participant determines to support sample alignment within this at least one time period. Another example is that if the load of the candidate secondary participant within this at least one time period is greater than threshold 2, the local resources that the candidate secondary participant can call within this at least one time period do not match the computing power requirements of the sample alignment algorithm, that is, the candidate secondary participant determines that it cannot support sample alignment within this at least one time period.

[0352] It can be understood that the specific values of the above threshold 1 and threshold 2 depend on the actual implementation, and the embodiments of the present application do not make specific limitations on this.

[0353] Exemplarily, local resources can include computing resources, storage resources, and network resources. Among them, computing resources can include the frequency of the processor, storage resources can include the memory size and the storage capacity of the local memory, and network resources can include throughput.

[0354] Optionally, at least one first common sample alignment algorithm can be associated with at least one time period.

[0355] It can be understood that considering that the main participant has different loads in different time periods, and thus the computing power supported by the local resources that can be called in different time periods is also different, that is, the first common sample alignment algorithms supported by the main participant in different time periods may be different. That is to say, since the first common sample alignment algorithms supported by the main participant in different time periods may be different, by associating at least one first common sample alignment algorithm with at least one time period, the candidate secondary participants can determine the first common sample alignment algorithms supported by the main participant for use in different time periods, and determine whether they support the computing power requirements of the first common sample alignment algorithms corresponding to different time periods according to the local resources that can be called in different time periods, so as to further determine whether to support the execution of sample alignment.

[0356] For example, the first information indicates the first common sample alignment algorithm 1 and the first common sample alignment algorithm 2, at least one time period includes time period 1 and time period 2, and the first common sample alignment algorithm 1 is associated with time period 1, which can indicate that the main participant can support using the first common sample alignment algorithm 1 to execute sample alignment in time period 1, and the first common sample alignment algorithm 2 is associated with time period 2, which can indicate that the main participant can support using the first common sample alignment algorithm 2 to execute sample alignment in time period 2.

[0357] In addition, one first common sample alignment algorithm can be associated with at least two time periods. For example, the first common sample alignment algorithm 1 is respectively associated with time period 1 and time period 2, to indicate that the main participant supports using the first common sample alignment algorithm 1 to execute sample alignment in time period 1 and time period 2.

[0358] Alternatively, one time period can be associated with at least two first common sample alignment algorithms. For example, time period 1 is associated with the first common sample alignment algorithm 1 and the first common sample alignment algorithm 2, to indicate that the main participant supports using the first common sample alignment algorithm 1 or the first common sample alignment algorithm 2 to execute sample alignment in time period 1.

[0359] It should be understood that the number of samples supported by the main participant for sample alignment, the sample alignment types supported by the main participant, or at least one time period during which the main participant supports sample alignment can be indicated using other parameters in addition to the first information. The embodiments of the present application do not make specific limitations on this.

[0360] S611. At least one candidate secondary participant sends a sample alignment response message to the main participant. Correspondingly, the main participant receives the sample alignment response messages from at least one candidate secondary participant.

[0361] Among them, the sample alignment response message can be used to indicate support for performing sample alignment and / or a second common sample alignment algorithm, where the second common sample alignment algorithm is a sample alignment algorithm determined from multiple first common sample alignment algorithms; alternatively, the sample alignment response message can be used to indicate non-support for performing sample alignment and / or the reason for non-support for performing sample alignment.

[0362] It can be understood that at least one candidate slave sending a sample alignment response message to the master can include: each candidate slave among at least one candidate slave sending a sample alignment response message to the master. Correspondingly, the master receiving sample alignment response messages from at least one candidate slave can include: the master receiving sample alignment response messages from each candidate slave among at least one candidate slave.

[0363] It should be understood that each candidate slave among at least one candidate slave can determine whether to support performing sample alignment using the first common sample alignment algorithm indicated by the first information according to the sample alignment request message and the local information of each candidate slave. Among them, the local information of the candidate slave can include the usage of local resources or locally available resources that can be called. For specific information about local resources, reference can be made to the description of local resources in step S610, which will not be elaborated here.

[0364] In addition, if a candidate slave supports performing sample alignment using the first common sample alignment algorithm, the second common sample alignment algorithm indicated by this candidate slave can be this first common sample alignment algorithm.

[0365] It can be understood that a candidate slave determines whether to support performing sample alignment according to the sample alignment request message and local information. For the specific principle, reference can be made to the relevant description of the first information in step S610. Two examples are listed below for further illustration.

[0366] Exemplarily, for the first common sample alignment algorithm being the PSI based on RSA and hash algorithm in the aforementioned Example 2, the secondary party can determine the computing power requirement of the first common sample alignment algorithm according to the first common sample alignment algorithm and the number of samples included in the dataset of the secondary party for vertical federated learning model training, and determine whether to support using the first common sample alignment algorithm to perform sample alignment according to whether the available local resources meet the computing power requirement. Additionally, in the case where the first information further indicates at least one time period during which the primary party supports sample alignment, the secondary party can also estimate the corresponding load situation for the at least one time period, determine whether the available local resources corresponding to the at least one time period meet the computing power requirement, and further determine whether to support using the first common sample alignment algorithm to perform sample alignment within the at least one time period. If the secondary party determines to support, then the first common sample alignment algorithm is the second common sample alignment algorithm; if not, then the first common sample alignment algorithm is not the second common sample alignment algorithm.

[0367] Exemplarily, for the first common sample alignment algorithm being the PSI based on Diffie-Hellman in the aforementioned Example 1, the secondary party can determine the computing power requirement of the first common sample alignment algorithm according to the number of samples supported by the primary party for sample alignment indicated by the first information and the number of samples included in the dataset of the secondary party for vertical federated learning model training, and determine whether to support using the first common sample alignment algorithm to perform sample alignment according to whether the available local resources meet the computing power requirement. If the secondary party determines to support using the first common sample alignment algorithm to perform sample alignment, then the first common sample alignment algorithm is the second common sample alignment algorithm; if not, then the first common sample alignment algorithm is not the second common sample alignment algorithm. Additionally, in the case where the first information further indicates at least one time period during which the primary party supports sample alignment, the secondary party can also estimate the corresponding load situation for the at least one time period, determine whether the available local resources corresponding to the at least one time period meet the above computing power requirement, and further determine whether to support using the first common sample alignment algorithm to perform sample alignment within the at least one time period. If the secondary party determines to support, then the first common sample alignment algorithm is the second common sample alignment algorithm; if not, then the first common sample alignment algorithm is not the second common sample alignment algorithm.

[0368] It can be understood that in the case where the first information indicates multiple first common sample alignment algorithms, the secondary party can traverse each first common sample alignment algorithm among the multiple first common sample alignment algorithms according to the above examples of determining whether to use the first common sample alignment algorithm to perform sample alignment, and determine whether to support performing sample alignment.

[0369] In addition, there can be multiple second common sample alignment algorithms. For example, multiple first common sample alignment algorithms are: sample alignment algorithm 1, sample alignment algorithm 2, and sample alignment algorithm 3. The second common sample alignment algorithm can be sample alignment algorithm 2 and sample alignment algorithm 3.

[0370] It should be understood that the above-related implementation of determining whether the participating party supports performing sample alignment is only an example. The participating party can use other methods to make a judgment based on the first information. For the specific implementation, reference can be made to step S610 of this embodiment of the present application, and no specific limitation is made here.

[0371] Next, based on whether the participating party supports performing sample alignment, the sample alignment response message will be introduced in two cases.

[0372] Case 1: The participating party supports performing sample alignment.

[0373] For the case where the participating party supports performing sample alignment, the sample alignment response message can be used to indicate support for performing sample alignment and / or the second common sample alignment algorithm.

[0374] It can be understood that for the first information in step S610 that only indicates one first common sample alignment algorithm, to save network overhead, the sample alignment response message can only indicate support for performing sample alignment. For the first information that indicates multiple first common sample alignment algorithms, if the participating party supports these multiple first sample alignment algorithms, to save network overhead, the sample alignment response message can only indicate support for performing sample alignment; if the participating party supports the second common sample alignment algorithm among the multiple first common sample alignment algorithms, to save network overhead, the sample alignment response message can only indicate the second common sample alignment algorithm to indicate that the participating party supports using the second common sample alignment algorithm among the multiple first common sample alignment algorithms to perform sample alignment. Of course, the sample alignment response message can also indicate support for performing sample alignment and the second common sample alignment algorithm. No specific limitation is made here in the embodiments of the present application.

[0375] Optionally, the sample alignment response message can also be used to indicate: the time period during which the participating party supports sample alignment and / or the number of samples that the participating party supports for sample alignment.

[0376] It can be understood that the time period during which the secondary participant supports sample alignment can be used by the primary participant to determine the time period during which the secondary participant can perform sample alignment, so that the primary participant can determine whether to select the secondary participant to perform sample alignment based on this time period. Among them, the time period during which the secondary participant supports sample alignment can include one or more time periods, and the one or more time periods can be associated with the second common sample alignment algorithm. For example, the time period during which the secondary participant supports sample alignment includes time period 3 and time period 4, and multiple second common sample alignment algorithms include sample alignment algorithm 1 and sample alignment algorithm 2. Sample alignment algorithm 1 can be associated with time period 3 to indicate that the secondary participant supports performing sample alignment using sample alignment algorithm 1 during time period 3. Similarly, sample alignment algorithm 2 can be associated with time period 4 to indicate that the secondary participant supports performing sample alignment using sample alignment algorithm 2 during time period 4. Of course, sample alignment algorithm 1 can be associated with time period 3 and time period 4 respectively to indicate that the secondary participant supports performing sample alignment using sample alignment algorithm 1 during time period 3 and time period 4.

[0377] In addition, when the first information in step S610 also indicates at least one time period during which the primary participant supports sample alignment (the at least one time period can be associated with the first common sample alignment algorithm), the time period during which the secondary participant supports sample alignment can be one or more time periods among the at least one time period, or a combination of subsets of one or more time periods among the at least one time period.

[0378] Exemplarily, assuming that at least one time period during which the primary participant supports sample alignment includes: [9:00 - 12:00], and [14:00 - 18:00], then the time period during which the secondary participant supports sample alignment can be one or more time periods among the at least one time period, for example, it can be [9:00 - 12:00], or [14:00 - 18:00], or [9:00 - 12:00] and [14:00 - 18:00]; or, the time period during which the secondary participant supports sample alignment can be a subset of one or more time periods among the at least one time period, for example, it can be the subset [9:30 - 11:00] of [9:00 - 12:00], or the subset [15:00 - 17:30] of [14:00 - 18:00], or the subset [9:45 - 11:25] of [9:00 - 12:00] and the subset [14:30 - 17:45] of [14:00 - 18:00], etc.

[0379] The number of samples supported by the subordinate party for sample alignment can be used by the primary party to determine the computing power requirements of the second common sample alignment algorithm. It can be understood that for a sample alignment algorithm where the computing power requirements for each participating party are the same in the second common sample alignment algorithm (such as the Diffie-Hellman-based PSI in Example 1), the computing power requirements of the sample alignment algorithm are related not only to the number of samples of the primary party itself but also to the number of samples of the subordinate party. Furthermore, the primary party can determine the computing power requirements of the second common sample alignment algorithm based on the number of its own samples and the number of samples of the subordinate party, so as to determine whether sample alignment can be supported during the actual execution of sample alignment and improve the success rate of sample alignment.

[0380] In addition, the number of samples supported by the subordinate party for sample alignment can refer to: the number of all samples in the dataset used by the subordinate party for vertical federated learning model training, or a subset of the number of all samples in this dataset. It can be understood that a subset of the number of all samples can be used to indicate that the subordinate party expects to use the samples corresponding to this subset for vertical federated learning model training. For example, if the subordinate party determines that its local computing power cannot support sample alignment for all samples using the second common sample alignment algorithm, the subordinate party can use the samples corresponding to a subset of all samples for sample alignment and subsequent vertical federated learning model training.

[0381] Exemplarily, assume that the first information indicates two first common sample alignment algorithms, namely sample alignment algorithm 1 and sample alignment algorithm 2. The first information also indicates that the number of samples supported by the primary party for sample alignment is 1 billion, and the subordinate party determines that the number of samples included in the dataset used for vertical federated learning model training is 100,000. Based on the above conditions, the subordinate party determines that sample alignment algorithm 1 can be supported, but sample alignment algorithm 2 cannot be supported. Thus, the sample alignment response message can include: sample alignment algorithm 1 (i.e., the second common sample alignment algorithm), the number of samples supported by the subordinate party for sample alignment, which is 100,000, and the time period during which the subordinate party supports the execution of sample alignment using sample alignment algorithm 1.

[0382] Case 2: The subordinate party does not support the execution of sample alignment.

[0383] For the case where the subordinate party does not support the execution of sample alignment, the sample alignment response message can be used to indicate the non-support for the execution of sample alignment and / or the reason for the non-support for the execution of sample alignment.

[0384] It can be understood that after the slave party determines that it does not support sample alignment, it can send a response message to the master party to indicate that it does not support sample alignment, so that the master party can be notified that the slave party cannot support sample alignment. Alternatively, after the slave party determines that it does not support sample alignment, the slave party can send the reason for not supporting sample alignment to assist the master party in adjusting the first common sample alignment algorithm indicated by the first information, or the number of samples supported by the master party for sample alignment, or the time period for which the master party supports sample alignment when the master party subsequently initiates a sample alignment request again, so as to improve the success rate of negotiating the relevant configuration of sample alignment (such as sample alignment algorithm, number of samples, or time period for sample alignment, etc.).

[0385] In addition, the reasons for not supporting sample alignment may include one or more reasons. For example, computing power is not supported, or time period is not supported. Among them, computing power not supported may include that the number of samples supported by the main participant for sample alignment is too large, or the number of samples corresponding to the data set of the slave participant for longitudinal federated learning model training is too large. It can be understood that the above reasons for not supporting sample alignment are only examples, and other reasons may also be included, and the embodiments of the present application do not specifically limit this.

[0386] Exemplarily, the sample alignment response message may include a parameter cause value (or failure indication with cause code), where the cause value is used to indicate the reason why sample alignment is not supported. It can be understood that indicating the reason for not performing sample alignment through the cause value can reduce indication overhead and improve reliability.

[0387] For example, the protocol can pre-configure reason values, such as the pre-configured symbol "01" is used to indicate that the reason for not supporting sample alignment is that the computing power is not supported, and "02" is used to indicate that the reason for not supporting sample alignment is that the time period is not supported. For another example, the pre-configured symbol "01" is used to indicate that the reason for not supporting sample alignment is that the number of samples supported by the main participant for sample alignment is too large, and "02" is used to indicate that the reason for not supporting sample alignment is that the time period is not supported.

[0388] It can be understood that the above protocol pre-configured cause values ​​are only examples. The symbols corresponding to the cause values ​​can also be negotiated in advance between the slave participants, the master participants, and the VFLSF network elements. The symbols corresponding to the cause values ​​can also be represented in other ways (such as bits). The embodiments of the present application do not specifically limit this.

[0389] In addition, the sample alignment response message may include two parameters, namely a failure indication and a reason value. The failure indication is used to indicate that the participating party does not support performing sample alignment, and the reason value indicates the reason for not supporting the execution of sample alignment. The embodiments of the present application do not make specific limitations on this.

[0390] Optionally, the sample alignment response message may be used to indicate the maximum number of samples expected to be used by the primary participating party.

[0391] It can be understood that the maximum number of samples expected to be used by the primary participating party can be used to indicate the maximum number of samples that the participating party can accept for the primary participating party to perform sample alignment according to local resources. Among them, if the primary participating party can use a data set corresponding to a number of samples less than or equal to the maximum number of samples expected to be used by the primary participating party to perform sample alignment, the participating party can support the execution of sample alignment.

[0392] In addition, when the first information indicates multiple first common sample alignment algorithms, the sample alignment response message may also be used to indicate the first common sample alignment algorithm associated with the maximum number of samples expected to be used by the primary participating party. In this way, the primary participating party can determine which sample alignment algorithm the maximum number of samples expected to be used by the primary participating party indicated by the sample alignment response message corresponds to, and further assist the primary participating party in adjusting the number of samples supported by the primary participating party for sample alignment when initiating a sample alignment request again (i.e., step S610) later for the first common sample alignment algorithm.

[0393] Exemplarily, assume that the first information indicates two first common sample alignment algorithms, namely sample alignment algorithm 1 and sample alignment algorithm 2. The first information also indicates that the number of samples supported by the primary participating party for sample alignment is 1 billion, and the number of samples included in the data set determined by the participating party for vertical federated learning model training is 100,000. Based on the above conditions, the participating party determines that due to the excessive number of samples supported by the primary participating party for sample alignment, the computing power is not supported. If the number of samples supported by the primary participating party for sample alignment is 500 million, the participating party can support sample alignment algorithm 1; if the number of samples supported by the primary participating party for sample alignment is 700 million, the participating party can support sample alignment algorithm 2. Furthermore, the sample alignment response message may include: a failure indication, a reason value (such as the excessive number of samples supported by the primary participating party for sample alignment), the first maximum number of samples expected to be used by the primary participating party associated with sample algorithm 1 is 500 million, and the second maximum number of samples expected to be used by the primary participating party associated with sample algorithm 2 is 700 million.

[0394] In addition, the above is only an example. The sample alignment response message may be a combination of one or more of a failure indication, a reason value, and the maximum number of samples expected to be used by the participating party for the primary participating party. The embodiments of the present application do not make specific limitations on this.

[0395] It should be understood that for the case where the participating party does not support sample alignment execution, the sample alignment response message may further include other information such as the time period during which the participating party expects to execute sample alignment. The embodiments of this application do not make specific limitations on this.

[0396] It can be understood that if the sample alignment response information of each candidate participating party among at least one candidate participating party indicates that sample alignment execution is not supported, the sample alignment negotiation fails. The primary participating party can initiate the sample alignment negotiation process again. For example, the primary participating party can adjust one or more of the first common sample alignment algorithm, the number of samples supported by the primary participating party for sample alignment, and at least one time period during which the primary participating party supports sample alignment, and send an adjusted sample alignment request message to improve the success rate of negotiating the relevant configuration of sample alignment. Additionally, if the sample alignment negotiation fails again, the primary participating party can initiate the sample alignment negotiation process again until the sample alignment negotiation is successful, or the number of repeated executions of the sample alignment negotiation process reaches the upper limit value. The upper limit value can be pre-configured by the protocol, or negotiated in advance by at least two of the primary participating party, the participating party, the VFL server, or the VFLSF network element. The embodiments of this application do not make specific limitations on this.

[0397] It can also be understood that if the sample alignment response messages of at least some of the candidate participating parties among at least one candidate participating party indicate support for sample alignment execution, the primary participating party can execute step S612, and step S612 is used to indicate the target sample alignment algorithm to at least one target participating party that meets the conditions for sample alignment execution among the at least some participating parties. In addition, the primary participating party can also indicate to the participating parties that do not meet the conditions for sample alignment execution among the at least some participating parties that sample alignment is not to be executed, so as to inform the participating parties that do not meet the conditions for sample alignment execution that sample alignment will not be executed next.

[0398] Step S612 is introduced below.

[0399] S612. The primary participating party sends a first sample alignment negotiation result message to the target participating party. Correspondingly, the target participating party receives the first sample alignment negotiation result message from the primary participating party. The first sample alignment negotiation result message is used to indicate the target sample alignment algorithm.

[0400] It can be understood that the primary participating party can determine the target secondary participating party and the target sample alignment algorithm according to the sample alignment response messages of at least one candidate secondary participating party. For example, in the case where there is only one secondary participating party that indicates support for performing sample alignment in the sample alignment response messages of at least one candidate secondary participating party, the target secondary participating party is the secondary participating party that supports performing sample alignment, and the target sample alignment algorithm is the second common sample alignment algorithm indicated in the sample alignment response message that indicates support for performing sample alignment. Additionally, in the case where there is only one secondary participating party that indicates support for performing sample alignment in the sample alignment response messages of at least one candidate secondary participating party, and the sample alignment response message of this secondary participating party indicates multiple second common sample alignment algorithms, the primary participating party can determine one second common sample alignment algorithm from the multiple second common sample alignment algorithms as the target sample alignment algorithm. Among them, there can be multiple implementation manners for the primary participating party to determine one second common sample alignment algorithm from the multiple second common sample alignment algorithms. For example, the primary participating party randomly selects one second common sample alignment algorithm from the multiple second common sample alignment algorithms, or the primary participating party selects the second common sample alignment algorithm with the smallest computing power requirement (i.e., computational complexity) from the multiple second common sample alignment algorithms, etc. The embodiments of the present application do not make specific limitations on this.

[0401] Additionally, in the case where there are multiple secondary participating parties that indicate support for performing sample alignment in the sample alignment response messages of at least one candidate secondary participating party, the primary participating party can determine the target secondary participating party and the target sample alignment algorithm from these multiple secondary participating parties according to the sample alignment response messages of these multiple secondary participating parties. For example, the sample alignment response messages of at least one candidate secondary participating party are the sample alignment response messages of at least one candidate secondary participating party in step S611. The primary participating party can determine multiple secondary participating parties that support performing sample alignment according to the sample alignment response messages of at least one candidate secondary participating party, and determine the target secondary participating party and the target sample alignment algorithm according to the intersection of the second common sample alignment algorithms indicated by these multiple participating parties.

[0402] That is to say, the primary participating party can determine the secondary participating parties that support performing sample alignment according to the sample alignment response messages of at least one candidate secondary participating party, and further determine the target sample alignment algorithm used for performing sample alignment and the target secondary participating party that meets the computing power requirement of this target sample alignment algorithm according to the multiple secondary participating parties that support performing sample alignment, avoiding the failure of sample alignment caused by not meeting the computing power requirement of the sample alignment algorithm during the process of performing sample alignment, and improving the success rate and efficiency of sample alignment.

[0403] It should be understood that the primary participating party can adopt multiple ways to determine the target secondary participating party and the target sample alignment algorithm according to the sample alignment response messages of at least one candidate secondary participating party, which will be described separately below.

[0404] Example A: The primary participant aligns the response messages according to the sample alignment of at least one candidate secondary participant, determines a combination of multiple secondary participants that support the same sample alignment algorithm, and selects the combination with the largest number of supported secondary participants as the target sample alignment algorithm and the target secondary participants.

[0405] For example, the at least one candidate secondary participant in step S611 includes: candidate secondary participants 1 to 5. The second common sample alignment algorithms indicated by the sample alignment response message of candidate secondary participant 1 are: sample alignment algorithm 1 and sample alignment algorithm 2. The second common sample alignment algorithms indicated by the sample alignment response message of candidate secondary participant 2 are: sample alignment algorithm 2 and sample alignment algorithm 3. The sample alignment response message of candidate secondary participant 3 indicates that it does not support performing sample alignment. The second common sample alignment algorithms indicated by the sample alignment response message of candidate secondary participant 4 are: sample alignment algorithm 3 and sample alignment algorithm 4. The second common sample alignment algorithm indicated by the sample alignment response message of candidate secondary participant 5 is: sample alignment algorithm 2. Furthermore, among the intersections of the second common sample alignment algorithms of the above different participants, the combinations of multiple participants that support the same sample alignment algorithm include: candidate secondary participant 1, candidate secondary participant 2, and candidate secondary participant 5 support sample alignment algorithm 2, and candidate secondary participant 2 and candidate secondary participant 4 support sample alignment algorithm 3.

[0406] Based on the above two combinations, the primary participant can select the combination with the largest number of supported candidate secondary participants as the target sample alignment algorithm and the target secondary participants. That is, the primary participant can determine that the target secondary participants are respectively: candidate secondary participant 1, candidate secondary participant 2, and candidate secondary participant 5, and the target sample alignment algorithm is sample alignment algorithm 2. Further, the primary participant can send a first sample alignment negotiation result message to candidate secondary participant 1, candidate secondary participant 2, and candidate secondary participant 5 respectively. The first sample alignment negotiation result message is used to indicate that the target sample alignment algorithm for performing sample alignment is sample alignment algorithm 2.

[0407] It can be understood that for vertical federated learning model training, the purpose is to improve the accuracy of model training by expanding the number of features. Furthermore, in the sample alignment process, the primary participant can increase the number of expanded features and thus improve the accuracy of model training by determining more target secondary participants for sample alignment.

[0408] Example B: In the case where the combinations with the largest number of supported candidate secondary participants are at least two different combinations, the primary participant can randomly select one of the combinations to determine the target secondary participants and the target sample alignment algorithm.

[0409] For example, assume that the difference between the sample alignment response messages of at least one candidate secondary participating party between Example B and Example A is as follows: the second common sample alignment algorithms indicated by the sample alignment response messages of candidate secondary participating party 5 in Example B are: sample alignment algorithm 2 and sample alignment algorithm 3. That is, among the intersections of the second common sample alignment algorithms of the above different participating parties, the combinations of multiple participating parties that support the same sample alignment algorithm include: candidate secondary participating party 1, candidate secondary participating party 2, and candidate secondary participating party 5 support sample alignment algorithm 2, and candidate secondary participating party 2, candidate secondary participating party 4, and candidate secondary participating party 5 support sample alignment algorithm 3.

[0410] Since the number of candidate secondary participating parties corresponding to the above two combinations is the same, the primary participating party can randomly select one of the combinations to determine the target secondary participating party and the target sample alignment algorithm. For example, the primary participating party can send a first sample alignment negotiation result message to candidate secondary participating party 1, candidate secondary participating party 2, and candidate secondary participating party 5 respectively, and this first sample alignment negotiation result message is used to indicate that the target sample alignment algorithm is sample alignment algorithm 2; or, the primary participating party can send a first sample alignment negotiation result message to candidate secondary participating party 2, candidate secondary participating party 4, and candidate secondary participating party 5 respectively, and this first sample alignment negotiation result message is used to indicate that the target sample alignment algorithm is sample alignment algorithm 3.

[0411] Example C: The primary participating party can determine the target secondary participating party and the target sample alignment algorithm according to the range of the number of secondary participating parties for performing sample alignment and the sample alignment response messages of at least one candidate secondary participating party.

[0412] It can be understood that the range of the number of secondary participating parties for performing sample alignment can include: the maximum number of secondary participating parties for performing sample alignment, and / or, the minimum number. For example, the maximum number of secondary participating parties for performing sample alignment can be indicated by a first threshold, then the range of the number of secondary participating parties for performing sample alignment is [1, the first threshold]. Another example, the minimum number of secondary participating parties for performing sample alignment can be indicated by a second threshold, then the range of the number of secondary participating parties for performing sample alignment is [the second threshold, +∞). Still another example, the minimum number of secondary participating parties for performing sample alignment is the second threshold, and the maximum number is the first threshold, then the range of the number of secondary participating parties for performing sample alignment is [the second threshold, the first threshold].

[0413] For example, assume that the difference between the sample alignment response messages of at least one candidate secondary participating party between Example C and Example B is as follows: the second common sample alignment algorithm indicated by the sample alignment response message of candidate secondary participating party 1 in Example C is: sample alignment algorithms 1 to 3. Furthermore, among the intersections of the second common sample alignment algorithms of the above different participating parties, multiple combinations that support the same sample alignment algorithm include: candidate secondary participating parties 1, 2, 4, and 5 support sample alignment algorithm 3, and candidate secondary participating parties 1, 2, and 5 support sample alignment algorithms 2 to 3.

[0414] Taking the range of the number of secondary participating parties performing sample alignment with a minimum number of 4 as an example, the primary participating party can select candidate secondary participating parties 1, 2, 4, and 5 to support sample alignment algorithm 3 as the target sample alignment algorithm. That is, the primary participating party can send the first sample alignment negotiation result message to candidate secondary participating parties 1, 2, 4, and 5 respectively, and this first sample alignment negotiation result message indicates that the target sample alignment algorithm is sample alignment algorithm 3.

[0415] Taking the range of the number of secondary participating parties performing sample alignment with a minimum number of 3 as an example, the primary participating party can determine the target secondary participating party and the target sample alignment algorithm according to any one of the above two combinations.

[0416] Taking the range of the number of secondary participating parties performing sample alignment with a maximum number of 3 as an example, the primary participating party can determine the target secondary participating party and the target sample alignment algorithm according to the fact that candidate secondary participating parties 1, 2, and 5 among the above combinations support sample alignment algorithms 2 to 3. For example, the target secondary participating parties are candidate secondary participating parties 1, 2, and 5 respectively, and the sample alignment algorithm can be sample alignment algorithm 2 and / or sample alignment algorithm 3.

[0417] Taking the range of the number of secondary participating parties performing sample alignment with a minimum number of 4 and a maximum number of 6 as an example, the primary participating party can select candidate secondary participating parties 1, 2, 4, and 5 to support sample alignment algorithm 3 as the target sample alignment algorithm.

[0418] It can be understood that the primary participating party can determine the target secondary participating party and the target sample alignment algorithm that meet the sample number requirements for performing sample alignment according to the range of the number of secondary participating parties performing sample alignment, so as to further avoid the problem that the excessive number of target secondary participating parties participating in sample alignment leads to excessive computing power requirements and affects the efficiency and success rate of sample alignment, thereby improving the success rate and efficiency of sample alignment.

[0419] It should be understood that the above Examples A to C are examples based on the premise that the sample alignment response message indicates the second common sample alignment algorithm. When the sample alignment response message of the candidate secondary participant only indicates support for performing sample alignment, the primary participant may default that the candidate secondary participant supports the first common sample alignment algorithm indicated in the sample alignment request message in step S610.

[0420] In addition, on the basis that the sample alignment response message indicates the second common sample alignment algorithm, it also indicates the time period during which the secondary participant supports sample alignment and / or the number of samples supported by the secondary participant for sample alignment. In this case, the primary participant may determine the target secondary participant and the target sample alignment algorithm according to the second common sample alignment algorithm indicated by each candidate secondary participant, the time period during which the secondary participant supports sample alignment, or the number of samples supported by the secondary participant for sample alignment. The following is a separate explanation.

[0421] Example D: The primary participant may determine the target secondary participant and the target sample alignment algorithm according to the second common sample alignment algorithm indicated by the sample alignment response messages of at least one candidate secondary participant and the time period during which the secondary participant supports sample alignment.

[0422] For example, on the basis that the target sample alignment algorithm is determined to be Sample Alignment Algorithm 2 in Example A, assume that the difference between the sample alignment response messages of at least one candidate secondary participant between Example D and Example A is as follows: the sample alignment response message of candidate secondary participant 1 also indicates the time period 1 associated with Sample Alignment Algorithm 2: [14:00 - 17:00], the sample alignment response message of candidate secondary participant 2 also indicates the time period 2 associated with Sample Alignment Algorithm 2: [9:00 - 11:00], and the sample alignment response message of candidate secondary participant 5 also indicates the time period 3 associated with Sample Alignment Algorithm 2: [15:00 - 17:00]. Further, the primary participant further determines that there is no overlap between time period 2 and time period 1 and time period 3 respectively according to whether there is an intersection among the time periods of candidate secondary participants 1, 2, and 5. That is, candidate secondary participant 2 is not the target secondary participant. That is, the primary participant determines that the target secondary participants are candidate secondary participants 1 and 5 respectively, the target sample alignment algorithm is Sample Alignment Algorithm 2, and the target time period for sample alignment is time period 3.

[0423] In addition, assume that the time period is 2: [16:30 - 17:00]. Then, there is an intersection among the time periods during which candidate secondary participants 1, 2, and 5 support sample alignment, but the intersection time period is only [16:30 - 17:00], that is, the intersection time period is too short. Thus, the primary participant may also determine that candidate secondary participant 2 is not suitable as the target secondary participant.

[0424] It should be understood that assuming that the time periods indicated by the sample response messages of each candidate slave participating party are not a subset of at least one time period supported by the master participating party, the master participating party should first determine the intersection between the time periods indicated by each candidate slave participating party and the at least one time period, and determine the target slave participating party and the target time period based on this intersection.

[0425] Example E: The master participating party may determine the target slave participating party and the target sample alignment algorithm according to the second common sample alignment algorithm indicated by the sample alignment response messages of at least one candidate slave participating party and the number of samples supported by the slave participating party for sample alignment.

[0426] For example, assume that the difference between the sample alignment response messages of at least one candidate slave participating party between Example E and Example A is as follows: the sample alignment response message of candidate slave participating party 1 also indicates that the number of samples supported by candidate slave participating party 1 for sample alignment is 100,000, the sample alignment response message of candidate slave participating party 2 also indicates that the number of samples supported by candidate slave participating party 2 for sample alignment is 500,000, and the sample alignment response message of candidate slave participating party 5 also indicates that the number of samples supported by candidate slave participating party 5 for sample alignment is 2 million. If the master participating party determines that the computing power requirement of sample alignment algorithm 2 is not met when the number of samples of candidate slave participating party 5 is 2 million, the master participating party determines that candidate slave participating party 5 is not the target slave participating party, that is, the target slave participating parties are candidate slave participating parties 1 and 2 respectively, and the target sample alignment algorithm is sample alignment algorithm 2.

[0427] Example F: The master participating party may determine the target slave participating party and the target sample alignment algorithm according to the second common sample alignment algorithm indicated by the sample alignment response messages of at least one candidate slave participating party, the time period supported by the slave participating party for sample alignment, and the number of samples supported by the slave participating party for sample alignment.

[0428] For example, on the basis that in Example A, the target sample alignment algorithm is determined to be Sample Alignment Algorithm 2, it is assumed that the difference between Example F and the sample alignment response messages of at least one candidate secondary participating party with respect to Example A is as follows: The sample alignment response message of candidate secondary participating party 1 also indicates the time period 1 associated with Sample Alignment Algorithm 2: [14:00 - 15:00], and the number of samples supported by candidate secondary participating party 1 for sample alignment is 2 million; the sample alignment response message of candidate secondary participating party 2 also indicates the time period 2 associated with Sample Alignment Algorithm 2: [14:00 - 17:00], and the number of samples supported by candidate secondary participating party 2 for sample alignment is 100,000; the sample alignment response message of candidate secondary participating party 5 also indicates the time period 3 associated with Sample Alignment Algorithm 2: [15:00 - 17:00], and the number of samples supported by candidate secondary participating party 5 for sample alignment is 500,000; furthermore, the primary participating party determines that candidate secondary participating parties 1, 2, and 5 jointly support Sample Alignment Algorithm 2. The intersection between time period 1 of candidate secondary participating party 1 and time period 2 of candidate secondary participating party 2 is [14:00 - 15:00], and the intersection between time period 2 of candidate secondary participating party 2 and time period 3 of candidate secondary participating party 5 is [15:00 - 17:00]. And the primary participating party determines that when the number of samples of the secondary participating party is 2 million, the computing power requirement of Sample Alignment Algorithm 2 is not met. Thus, the primary participating party can determine that candidate secondary participating party 1 is not the target secondary participating party, the target secondary participating parties are candidate secondary participating parties 2 and 5 respectively, the target sample alignment algorithm is Sample Alignment Algorithm 2, and the target time period for sample alignment is [15:00 - 17:00].

[0429] It can be understood that on the basis of the second common sample alignment algorithm, the primary participating party also determines the target secondary participating party and the target sample alignment algorithm according to the time period supported by the secondary participating party for sample alignment, and / or, the number of samples supported by the secondary participating party for sample alignment, which can further reduce the probability of sample alignment failure or efficiency reduction caused by not meeting the computing power requirement of the sample alignment algorithm.

[0430] It should be understood that the above Examples A to F are only examples. The primary participating party can also determine the target secondary participating party and the target sample alignment algorithm in other ways, and the embodiments of the present application do not make specific limitations in this regard.

[0431] Optionally, the first sample alignment negotiation result message is also used to indicate the target time period for sample alignment.

[0432] It can be understood that the target time period can be the intersection between the time periods indicated by the sample alignment response messages of each target secondary participating party, such as the target time period in Example D.

[0433] That is to say, the participants can also indicate the target time period for performing sample alignment through the first sample alignment negotiation result message, so that each participant can perform the sample alignment process within the target time period, avoiding each participant from performing the sample alignment process in a time period when other participants may not be able to meet the computing power requirements of the target sample alignment algorithm, thereby further improving the success rate and efficiency of sample alignment.

[0434] Optionally, Figure 6 The method flow shown also includes step S613.

[0435] S613: The master participant sends a second sample alignment negotiation result message to other slave participants except the target slave participant among the multiple slave participants supporting the sample alignment. Correspondingly, other slave participants except the target slave participant among the multiple slave participants supporting the sample alignment receive the second sample alignment negotiation result message from the master participant.

[0436] Among them, the second sample alignment negotiation result message is used to indicate that sample alignment is not performed, and / or the reason for not performing sample alignment. That is to say, the main participant sends a second sample alignment negotiation result message to other slave participants to indicate that sample alignment is not performed, and can promptly notify other slave participants that sample alignment will not be performed next. In addition, by indicating the reason for not performing sample alignment through the second sample alignment negotiation result message, when the main participant subsequently initiates the sample alignment negotiation process again, it can assist in adjusting the second common sample alignment algorithm indicated by the sample alignment response message fed back by the slave participant, the number of samples supported by the slave participant for sample alignment, or the time period supported by the slave participant for sample alignment, etc., so as to improve the success rate and efficiency of the sample alignment negotiation.

[0437] It can be understood that the multiple slave participants supporting the sample alignment can be, for example, at least one candidate slave participant that sends a sample alignment response message indicating support for the execution of the sample alignment algorithm. The other slave participants except the target slave participant among the multiple slave participants can, for example, refer to the slave participants among the multiple participants that are not selected as the target slave participants by the master participant, such as candidate slave participant 4 in Examples A to F, candidate slave participant 2 in Example D, candidate slave participant 5 in Example E, and candidate slave participant 1 in Example F.

[0438] In addition, there can be various reasons for not performing sample alignment. For example, for candidate slave participant 4 in Examples A to F, the reason for not performing sample alignment is that sample alignment algorithms 3 and 4 supported by candidate slave participant 4 are not used. Another example is that for candidate slave participant 2 in Example D, the reason for not performing sample alignment is that there is no intersection between time period 2 supported by candidate slave participant 2 for sample alignment and time periods 1 and 3. Another example is that for candidate slave participant 5 in Example E, the reason for not performing sample alignment is that the master participant determines that the computing power requirement of sample alignment algorithm 2 is not met when the number of samples of candidate slave participant 5 is 2 million. Another example is that for candidate slave participant 1 in Example F, the reason for not performing sample alignment is that there is no intersection between time period 1 supported by candidate slave participant 1 for sample alignment and time period 3, and the master participant determines that the computing power requirement of sample alignment algorithm 2 is not met when the number of samples of candidate slave participant 1 is 2 million.

[0439] It should be understood that the above reasons for not performing sample alignment are only examples, and there can be other reasons for not performing sample alignment in actual implementation, which are not specifically limited in the embodiments of the present application.

[0440] In addition, the second sample alignment negotiation result message can indicate the above reasons for not performing sample alignment through the parameter reason value. For specific details, refer to Case 2 in step S611 above, which will not be elaborated here.

[0441] It can be understood that step S613 is an optional step, that is, the master participant can also not send the second sample alignment negotiation result message to other slave participants among the multiple slave participants that support performing sample alignment except the target slave participant. These other slave participants can start timing when sending the sample alignment response message to the master participant. If the first sample alignment negotiation result message from the master participant is not received after exceeding the time threshold, it can be determined that the master participant instructs not to perform sample alignment.

[0442] In addition, the above time threshold can be pre-agreed by the protocol, or negotiated in advance between the candidate slave participant and the master participant, or indicated by other network elements (such as the VFL server or the master participant, etc.), which are not specifically limited in the embodiments of the present application.

[0443] It should be understood that in the above step S611, when there are multiple slave participating parties indicating support for performing sample alignment in the sample alignment response messages of at least one candidate slave participating party, the master participating party may determine, based on the sample alignment response messages of these multiple slave participating parties, that sample alignment cannot be performed between the master participating party and these multiple slave participating parties, that is, the sample alignment negotiation fails. For example, there is no intersection between the second common sample alignment algorithms indicated in the sample alignment response messages of multiple slave participating parties. Another example is that the number (or quantity) of slave participating parties whose sample alignment response messages indicate an intersection between the second common sample alignment algorithms is less than or equal to the second threshold, that is, the minimum number of slave participating parties for performing sample alignment. Still another example is that there is no intersection between the time periods supported for sample alignment indicated in the sample alignment response messages of multiple slave participating parties, or the number of slave participating parties with an intersection is less than or equal to the second threshold, etc. The embodiments of the present application do not make specific limitations in this regard.

[0444] In addition, when the master participating party determines that sample alignment cannot be performed between it and these multiple slave participating parties, the master participating party may send a second sample alignment response message to the multiple slave participating parties that support performing sample alignment, to indicate non - execution of sample alignment and / or the reason for non - execution of sample alignment.

[0445] It can be understood that in the case of sample alignment negotiation failure, the master participating party can initiate the sample alignment negotiation process again. For example, the master participating party can adjust one or more of the first common sample alignment request, the number of samples supported by the master participating party for sample alignment, and at least one time period supported by the master participating party for sample alignment, and send an adjusted sample alignment request message, to increase the success rate of negotiating the relevant configuration for sample alignment.

[0446] S614. Perform a sample alignment process between the master participating party and the target slave participating party.

[0447] It can be understood that based on step S612, the master participating party and the target slave participating party can determine the target sample alignment algorithm for performing sample alignment (for example, it can also include the target time period for sample alignment). The master participating party and the target slave participating party can perform the sample alignment process based on the target sample alignment algorithm to obtain a sample alignment result. Among them, the sample alignment result may include the common samples between the master participating party and multiple target slave participating parties.

[0448] In addition, the sample alignment process may include a two - party sample alignment process and / or a multi - party sample alignment process. The specific implementation of the two - party sample alignment process and / or the multi - party sample alignment process can refer to the relevant descriptions in the preamble part "Sample Alignment" of the specific implementation manners (such as Example 1 and Example 2), and will not be elaborated here.

[0449] It should be understood that the primary participant and the target secondary participant can determine the common samples and / or common sample identifiers between the primary participant and the target secondary participant according to the sample alignment result.

[0450] S615. The participant sends a vertical federated learning response message to the VFL server. Correspondingly, the VFL server receives the vertical federated learning response message from the participant.

[0451] Among them, the vertical federated learning response message is used to indicate whether the participant joins the vertical federated learning. It should be understood that joining the vertical federated learning may refer to joining the vertical federated learning model training process and / or the vertical federated learning inference process. The embodiments of the present application do not make specific limitations on this.

[0452] In addition, joining the vertical federated learning has the same meaning as participating in the vertical federated learning and can be used interchangeably. It is uniformly explained here and will not be repeated hereinafter.

[0453] It can be understood that the participant may include a primary participant and a target secondary participant, and the vertical federated learning response message may include a first vertical federated learning response message and a second vertical federated learning response message. Further, step S615 may include:

[0454] S615a. The target secondary participant sends a first vertical federated learning response message to the VFL server. Correspondingly, the VFL server receives the first vertical federated learning response message from the target secondary participant.

[0455] It can be understood that the first vertical federated learning response message can be understood as the response message of the first vertical federated learning request message in step S608, that is, the first vertical federated learning response message can also be understood as the response message in the vertical federated learning preparation process, and is used to indicate whether the target secondary participant joins the vertical federated learning.

[0456] S615b. The primary participant sends a second vertical federated learning response message to the VFL server. Correspondingly, the VFL server receives the second vertical federated learning response message from the primary participant.

[0457] It can be understood that the second vertical federated learning response message can be understood as the response message of the second vertical federated learning request message in step S609, that is, the second vertical federated learning response message can also be understood as the response message in the vertical federated learning preparation process, and is used to indicate whether the primary participant joins the vertical federated learning.

[0458] It should be understood that the primary participant and the target secondary participant can determine whether to participate in vertical federated learning based on the sample alignment result. For example, if there are only a small number of common samples between the first participant (any one of the primary participant and multiple target secondary participants) and other participants, or the number of common samples is less than or equal to the third threshold, the first participant can determine not to participate in vertical federated learning. The third threshold can be 10%, 20%, 30%, or other values of the samples corresponding to the dataset provided by the first participant for vertical federated learning model training, which specifically depends on the actual implementation, and the embodiments of the present application do not make specific limitations in this regard.

[0459] In addition, the first participant can also determine whether to participate in vertical federated learning based on other information. For example, if the first participant determines that its own load is too high, it can determine not to participate in vertical federated learning. The embodiments of the present application do not make specific limitations in this regard.

[0460] S616. The VFL server determines the target participants who participate in vertical federated learning according to the vertical federated learning response messages of the participants.

[0461] It can be understood that the target participants who participate in vertical federated learning can be, for example, the participants who participate in the training of the vertical federated learning model. Then, after completing the sample alignment process, each target participant can determine the common samples and can perform the vertical federated learning model training process. Among them, the specific process of vertical federated learning model training can be referred to Figure 3 the process shown, which will not be elaborated here.

[0462] Since in the embodiments of the present application Figure 6For the information transmission method process shown, each participating party can register its sample alignment ability with the VFLSF network element. Then, when the main participating party initiates vertical federated learning, the main participating party can discover the VFL server, which is the coordinator of vertical federated learning, through the VFLSF network element, and send a discovery request message to the VFL server. As a result, the VFL server can discover at least one candidate secondary participating party that meets the sample alignment requirements through the VFLSF network element, obtain the sample alignment ability (such as the sample alignment algorithm) of the at least one candidate secondary participating party, and send the sample alignment ability of the at least one candidate secondary participating party to the main participating party. Subsequently, the main participating party can negotiate sample alignment with the at least one candidate secondary participating party based on the first common sample alignment algorithm jointly supported between the main participating party and the at least one candidate secondary participating party, so as to improve the success rate and efficiency of sample alignment negotiation. Further, through sample alignment negotiation, the main participating party can determine the target secondary participating party and the target sample alignment algorithm for which the local computing power can support the execution of sample alignment based on the responses of the at least one candidate secondary participating party, so as to avoid the problem of sample alignment failure caused by the computing power requirement of the sample alignment algorithm not being met during the execution of sample alignment, and improve the success rate and efficiency of sample alignment.

[0463] It can be understood that in the information transmission method process shown above, Figure 6 as the initiator of vertical federated learning, in another implementation, the secondary participating party can also initiate vertical federated learning, and the VFL server is responsible for discovering the main participating party and candidate secondary participating parties. For example, when the secondary participating party discovers that the training data cannot be collected due to privacy protection issues and then initiates vertical federated learning, the secondary participating party can execute step S603 to send a discovery request message to the VFLSF network element and determine the VFL server through step S604. Further, the secondary participating party can execute step S605, that is, the secondary participating party can send a vertical federated learning initiation request message to the VFL server, so that the VFL server and the VFLSF network element execute steps S606 - S607 to determine the main participating party and candidate secondary participating parties. After that, the VFL server sends vertical federated learning request messages to the secondary participating party, the main participating party, and the candidate secondary participating parties (i.e., steps S608 and S609), so as to facilitate the subsequent execution of the sample alignment negotiation process (i.e., steps S610 - S613), the sample alignment execution process (i.e., step S614), and the process of the VFL server determining the target participating parties joining vertical federated learning (steps S615 - S616) between the main participating party and the secondary participating party.

[0464] In addition, in the above example of the vertical federated learning initiated by the participating party, the participating party can initiate a sample alignment negotiation process, that is, the participating party can execute step S610, that is, the participating party can send a sample alignment request message to the primary participating party and the candidate secondary participating parties.

[0465] It should be understood that in the information transmission method flow shown above, the participating party (such as the primary participating party or the secondary participating party) can be a trusted network element within the core network (or a network element located in the same trusted domain as the VFLSF network element). For example, the participating party can be an application function network element within the operator, and the participating party can directly interact with network function network elements such as the VFLSF network element (or the network storage function network element), the VFL server, or the network data analysis function network element. When the participating party is a third-party application function network element (or an untrusted application function (untrusted AF) network element), the participating party and the above network function network elements are not in the same trusted domain, and the network open function network element can provide the participating party with secure access to the network functions of the above network function network elements, that is, the participating party can interact with the above network function network elements through the network open function network element. Figure 6

[0466] Next, taking the primary participating party in Figure 6 as an untrusted application function network element, the candidate secondary participating party in at least one candidate secondary participating party as a network data analysis function network element, the VFL server as the VFL server in Figure 5 , and the VFLSF network element as the VFLSF network element or the network storage function network element in Figure 5 as an example, the information transmission method flow shown in Figure 7 will be further described in combination with Figure 6 .

[0467] Figure 7 is a schematic flow of an information transmission method provided by an embodiment of the present application. Figure 2 . As Figure 7 shown, the information transmission method flow includes: steps S701 to S728, steps S705 to S706 are the same as steps S601 to S602, S713 to S715 are the same as steps S606 to S608, and step S728 is the same as step S616, which will not be elaborated here.

[0468] Next, steps S701 to S704, S707 to S712, and S716 to S727 will be introduced.

[0469] S701. The primary participating party sends a first registration request message to the network open function network element. Correspondingly, the network open function network element receives the first registration request message from the primary participating party.

[0470] ​It can be understood that the specific content of the first registration request message can be referred to the registration request message in step S601 of Figure 6 , which will not be elaborated here.

[0471] It should be understood that the network exposure function network element can support the following functions: capability and event exposure, authentication and authorization, and internal and external information conversion. Among them, capability and event exposure can refer to, for example, a network function network element within the network can securely expose capabilities and events to an untrusted application function network element through the network exposure function network element. Authentication and authorization can refer to, for example, the network exposure function network element can authenticate and authorize the behavior of the untrusted application function network element to achieve the function of the untrusted application function network element securely sending information to the network function network element. Internal and external information conversion can refer to, for example, converting the information exchanged between the untrusted application function network element and the network function network element. Additionally, to improve security, the network exposure function network element can also mask the information of the untrusted application function network element and the sensitive information of the network function network element.

[0472] Exemplarily, the primary participant can send the first registration request information by invoking the first service operation provided by the network exposure function network element, and this first service operation can be a service operation related to vertical federated learning registration. For example, the first service operation can be a vertical federated learning registration request service operation.

[0473] In addition, the primary participant can interact with the network exposure function network element through the service interface Nnef of the network exposure function network element.

[0474] It should be understood that the above example of the primary participant sending the first registration request message by invoking the first service operation provided by the network exposure function network element is just an example. The primary participant can also send the first registration request message by invoking other service functions or service operations provided by the network exposure function network element, or send the first registration request message in other ways. The embodiments of the present application do not make specific limitations on this.

[0475] It can be understood that the network exposure function network element can verify and authorize the first registration request message of the primary participant. Then, when the authorization is passed, the network exposure function network element can provide the information contained in the first registration request message to the VFLSF network element (i.e., step S702).

[0476] In addition, the execution subject of step S701 being the primary participant is just an example. As Figure 6 the relevant description of step S601 in, in the embodiments of the present application for vertical federated learning, the participants can also not distinguish between the primary participant and the secondary participant, that is, the execution subject of step S701 can be an untrusted network element, and the primary participant and the secondary participant can be not distinguished. In addition, as Figure 6Regarding the relevant description of step S605, in the case where the participating parties are distinguished into the primary participating party and the secondary participating party, the initiator of vertical federated learning can also be the secondary participating party, that is, the execution entity of step S701 can also be the secondary participating party. This is uniformly explained here and will not be elaborated further below.

[0477] S702. The network exposure function network element sends a second registration request message to the VFLSF network element. Correspondingly, the VFLSF network element receives the second registration request message from the network exposure function network element.

[0478] It can be understood that the second registration request message may include the information contained in the first registration request message.

[0479] Exemplarily, the network exposure function network element may send the second registration request message by invoking the second service operation provided by the VFLSF network element. Among them, if the VFLSF network element is the Figure 5 VFLSF network element in, the second service operation may be a service operation related to vertical federated learning registration provided by the VFLSF network element (such as Nvflsf Management RegisterRequest). If the VFLSF network element is the Figure 5 network storage function network element in, the second service operation may be a network function management service operation provided by the network storage function network element (such as Nnrf_NFManagement NFRegister Request).

[0480] In addition, the above-mentioned second service operation may be different from the first service operation in step S701, that is, the primary participating party sends the first registration request message to the network exposure function network element, and the network exposure function network element sends the second registration request message to the VFLSF network element, and the service operations used between the two are different.

[0481] It should be understood that the above-mentioned network exposure function network element sending the second registration request message by invoking the second service operation provided by the VFLSF network element is only an example. The network exposure function network element may also send the second registration request message by invoking other service functions or service operations provided by the VFLSF network element, or send the second registration request message in other ways. The embodiments of the present application do not make specific limitations on this.

[0482] S703. The VFLSF network element sends a second registration response message to the network exposure function network element. Correspondingly, the network exposure function network element receives the second registration response message from the VFLSF network element.

[0483] It can be understood that the second registration response message is used to indicate whether the registration is successful.

[0484] In addition, the VFLSF network element can send a second registration response message by invoking a third service operation. For example, for the VFLSF network element in Figure 5 , the third service operation can be a service operation related to vertical federated learning registration provided by the VFLSF network element (such as Nvflsf Management Register Response). Another example, for the VFLSF network element being a network storage function network element, the third service operation can be a network function management service operation provided by the network storage function network element (such as Nnrf_NFManagement NFRegister Response).

[0485] S704. The network openness function network element sends a first registration response message to the primary participant. Correspondingly, the primary participant receives the first registration response message from the network openness function network element.

[0486] It can be understood that the information included in the first registration response message can include the information included in the second registration response message.

[0487] In addition, the network openness function network element can send a first registration response message by invoking a fourth service operation. For example, the fourth service operation can be a vertical federated learning registration response service operation provided by the network openness function network element.

[0488] S707. The primary participant sends a first discovery request message to the network openness function network element. Correspondingly, the network openness function network element receives the first discovery request message from the primary participant.

[0489] It can be understood that the first discovery request message can specifically refer to the discovery request message in step S603 of Figure 6 , which will not be elaborated here.

[0490] In addition, the primary participant can send a first discovery request message by invoking a relevant service operation provided by the network openness function network element. For example, the primary participant can send a first discovery request message by invoking the vertical federated learning participant discovery request (Nnef VFL ParticipantDiscovery Request) service operation provided by the network openness function.

[0491] It can be understood that when the network openness function network element verifies and authorizes the first discovery request message sent by the primary participant, and then when the authorization is passed, the network openness function network element can provide the information included in the first discovery request message to the VFLSF network element (i.e., step S708).

[0492] The network openness function network element sends a second discovery request message to the VFLSF network element. Correspondingly, the VFLSF network element receives the second discovery request message from the network openness function network element.

[0493] It can be understood that the second discovery request message may include the information contained in the first discovery request message.

[0494] Exemplarily, for the VFLSF network element in Figure 5 , the network openness function network element may send a second discovery request message by invoking the vertical federated learning participant discovery request (Nvflsf VFL ParticipantDiscoveryRequest) service operation provided by the VFLSF network element. For the network storage function network element in Figure 5 , the network openness function network element may send a second discovery request message by invoking the network function discovery request (NnrfNFDiscovery Request) service operation provided by the network storage function network element.

[0495] S709. The VFLSF network element sends a second discovery response message to the network openness function network element. Correspondingly, the network openness function network element receives the second discovery response message from the VFLSF network element.

[0496] It can be understood that for the specific content of the second discovery response message, reference can be made to Figure 6 the discovery response message in step S604, which will not be elaborated here.

[0497] S710. The network openness function network element sends a first discovery response message to the primary participant. Correspondingly, the primary participant receives the first discovery response message from the network openness function network element.

[0498] In a possible implementation manner, the first discovery response message includes the information contained in the second discovery response message. In other words, the first discovery response message may include the identifier or address of at least one VFLserver contained in the discovery response message in step S604.

[0499] In another possible implementation, the first discovery response message includes a first identifier or a first address assigned (or configured or provided, etc.) by the network exposure function network element for at least one VFL server indicated by the second discovery response message. The first identifier of each VFL server in the at least one VFL server is different from the identifier of each VFL server, and the first address of each VFL server is different from the address of each VFL server. That is to say, the network exposure function network element can assign a first identifier or a first address to each VFL server in the at least one VFL server indicated by the second discovery response message to avoid exposing the true identifier or the true address of the VFL server, thereby improving security.

[0500] For example, assume that the at least one VFL server indicated by the second discovery response message includes: VFL server1, VFL server2, and VFL server3. The identifier of VFL server1 is identifier A, the identifier of VFL server2 is identifier B, and the identifier of VFL server3 is identifier C. Based on the above conditions, the first identifier assigned by the network exposure function network element to VFL server1 is identifier #1, the first identifier assigned to VFL server2 is identifier #2, and the first identifier assigned to VFL server3 is identifier #3. Furthermore, the first discovery response message includes identifier #1, identifier #2, and identifier #3. Thus, what the primary participant obtains is the first identifier of the VFL server rather than the true identifier of the VFL server, which can avoid exposing the true identifier of the VFL server.

[0501] In addition, the principle of the network exposure function network element assigning the first address to the at least one VFL server is similar to the above example of assigning the first identifier, and will not be elaborated here.

[0502] It can be understood that the network exposure function network element can save or maintain the correspondence between the identifier of the VFL server and the first identifier of the VFL server, and / or the correspondence between the address of the VFL server and the first address of the VFL server. In this way, the network exposure function network element can, based on the above correspondence, determine the true identifier of the VFL server according to the first identifier of the VFL server, and / or determine the true address of the VFL server according to the first address of the VFL server.

[0503] It can be understood that the above description of the first identifier is only an example. The network openness function network element can also allocate the first identifier in other ways, and the embodiments of the present application do not make specific limitations in this regard.

[0504] S711. The primary participant sends a first vertical federated learning initiation request message to the network openness function network element. Correspondingly, the network openness function network element receives the first vertical federated learning initiation request message from the primary participant.

[0505] It can be understood that for the specific content of the first vertical federated learning initiation request message, reference can be made to Figure 6 the vertical federated learning initiation request message in step S605, which will not be elaborated here.

[0506] In addition, as described in step S605, the primary participant can determine a VFL server from at least one VFL server. Accordingly, the first vertical federated learning initiation request message in step S711 may include the identifier and / or address of the VFL server. In this way, the network openness function network element can determine which VFL server among the at least one VFL server the information included in the first vertical federated learning initiation request message is sent to according to the identifier and / or address of the VFL server.

[0507] It can be understood that in the case where the first discovery response message in step S710 includes the first identifier and / or the first address of each VFL server among at least one VFL server, the first vertical federated learning initiation request message includes the first identifier and / or the first address of the VFL server determined by the primary participant from among at least one VFL server.

[0508] It should be understood that the VFL server indicated by the first vertical federated learning initiation request message in step S711 corresponds to the primary participant, that is, the VFL server is the VFL server corresponding to the vertical federated learning process initiated by the primary participant. Since the network openness function network element may provide services for the initiators of other vertical federated learning processes in addition to the primary participant, to distinguish the VFL server corresponding to the primary participant from the VFL servers corresponding to other initiators, the network openness function network element can maintain the VFL servers corresponding to different initiators. That is to say, by maintaining the correspondence between the primary participant and the VFL server selected by the primary participant, the network openness function network element can determine which VFL server among the multiple VFL servers to send the message of the primary participant to according to this correspondence, and which initiator among the multiple initiators to send the message of the VFL server to.

[0509] In a possible implementation, the first vertical federated learning initiation request message further includes a correlation ID that associates the vertical federated learning process between the primary participant and the VFL server. It can be understood that the primary participant may execute multiple different vertical federated learning processes in parallel, and the multiple different vertical federated learning processes can be distinguished by the correlation ID. For example, taking the vertical federated learning process as a sample alignment process as an example, assume that the primary participant and the secondary participant 1 execute the sample alignment process #1, and the primary participant and the secondary participant 2 execute the sample alignment process #2. These two sample alignment processes are different. For example, the common samples determined by these two sample alignment processes are used for training two different vertical federated learning models, or for another example, the sample alignment algorithms used in these two sample alignment processes and / or the data sets used for vertical federated learning model training are different. Additionally, the vertical federated learning process can also be a vertical federated learning model training process or a vertical federated learning inference process. The primary participant can execute multiple different vertical federated learning model training processes or vertical federated learning inference processes in parallel with different secondary participants, etc.

[0510] Furthermore, among the multiple different vertical federated learning processes executed in parallel by the primary participant, they can correspond to the same VFL server. Thus, by adding the correlation ID to the messages exchanged between the primary participant, the VFL server, or the secondary participant through the network function virtualization (NFV) network element, the primary participant, the VFL server, the secondary participant, or the NFV network element can distinguish the multiple different vertical federated learning processes executed in parallel. For example, the NFV network element can determine which vertical federated learning process of the primary participant the first vertical federated learning initiation request message belongs to according to the correlation ID. When the NFV network element sends the information included in the first vertical federated learning initiation request message to the VFL server corresponding to this vertical federated learning process of the primary participant, it can also send the correlation ID. In this way, the VFL server can determine which vertical federated learning process the received message is for according to the correlation ID.

[0511] It can be understood that the above correlation ID is assigned (or generated) by the primary participant. In another implementation, the NFV network element can also assign the correlation ID. For details, refer to step S712 below.

[0512] It should be understood that the above examples of the vertical federated learning process are only a limited number of examples. The vertical federated learning process can also include a vertical federated learning inference process, a sample alignment process, etc. The embodiments of the present application do not make specific limitations in this regard.

[0513] The network openness function network element sends a second vertical federated learning initiation request message to the VFL server. Correspondingly, the VFL server receives the second vertical federated learning initiation request message from the network openness function network element.

[0514] It can be understood that the VFL server may be the VFL server in step S605, and the second vertical federated learning initiation request message includes the information contained in the first vertical federated learning initiation request message.

[0515] In a possible implementation manner, the second vertical federated learning initiation request message may include a second identifier and / or a second address assigned by the network openness function network element to the primary participant. It can be understood that similar to the network openness function network element assigning a first identifier and / or a first address to at least one VFL server in step S710, the network openness function network element may also assign a second identifier and / or a second address to the primary participant to avoid exposing the true identifier and true address of the primary participant.

[0516] In addition, when the network openness function network element sends a message from the primary participant to other network elements in the following steps, the network openness function network element may add the second identifier and / or the second address of the primary participant to the message. This is uniformly described here and will not be repeated below.

[0517] It can be understood that in the case where the first vertical federated learning initiation request message does not include an association identifier (i.e., the case where the primary participant has not allocated an association identifier), the network openness function network element can allocate the association identifier and carry the association identifier in the second vertical federated learning initiation request message. For example, the network openness function network element can determine the vertical federated learning process #1 of the primary participant corresponding to the first vertical federated learning initiation request message according to the information included in the first vertical federated learning initiation request message, and allocate an association identifier for the vertical federated learning process #1. In addition, the network openness function network element can determine the VFL server selected by the primary participant according to the first vertical federated learning initiation request message, and add the association identifier to the second vertical federated learning initiation request message sent to the VFL server. In this way, the VFL server can determine that the second vertical federated learning initiation request message belongs to the vertical federated learning process #1 according to the association identifier. In addition, after the network openness function network element allocates the association identifier, it can send the association identifier to each participant in the vertical federated learning process corresponding to the association identifier (such as the primary participant, the VFL server, or at least one candidate secondary participant, etc.), so that each participant in the vertical federated learning process can determine which vertical federated learning process the sent or received message belongs to according to the association identifier. For example, for the VFL server, the network openness function network element can send the association identifier to the VFL server through the second vertical federated learning initiation request message in step S713. Then, the VFL server can determine that the second vertical federated learning initiation request message is a request message from the vertical federated learning process of the primary participant. In addition, the VFL server executes steps S713 to S714 according to the second vertical federated learning initiation request message (i.e., the VFL server obtains information of at least one candidate secondary participant), and then can trigger the VFL server to send a second vertical federated learning request message to the network openness function network element (i.e., step S716). Thus, when the network openness function network element receives the second vertical federated learning request message from the VFL server and triggers sending a third vertical federated learning request message to the primary participant (i.e., step S717), it can send the association identifier to the primary participant through the third vertical federated learning request message.

[0518] For another example, in the subsequent sample alignment negotiation process, the network openness function network element can send the association identifier to at least one candidate secondary participant through the second sample alignment request message in step S719.

[0519] S716. The VFL server sends a second vertical federated learning request message to the network openness function network element. Correspondingly, the network openness function network element receives the second vertical federated learning request message from the VFL server.

[0520] It can be understood that the specific content of the second vertical federated learning request message can be referred to Figure 6 the second vertical federated learning request message in step S609 of

[0521] In addition, according to the relevant descriptions of the association identifier in steps S711 and S712, the second vertical federated learning request message may further include the association identifier corresponding to the vertical federated learning process among the primary participant, the VFL server, and at least one candidate secondary participant indicated by the second vertical federated learning request message, so that the network openness function network element can determine that the second vertical federated learning request message belongs to the vertical federated learning process corresponding to this association identifier.

[0522] It can be understood that on the basis of the network openness function network element maintaining the corresponding relationship between the primary participant and the VFL server in the foregoing step S711, the network openness function network element can also maintain the corresponding relationship among the primary participant, the VFL server, and at least one candidate secondary participant indicated by the second vertical federated learning request message. Furthermore, when the primary participant (untrusted application function network element) interacts with at least one candidate secondary participant (such as a network data analysis function network element) through the network openness function network element during the vertical federated learning process, the network openness function network element can determine which network data analysis function network element to send the message of the primary participant to and which application function network element to send the message of the network data analysis function network element to according to this corresponding relationship.

[0523] S717. The network openness function network element sends a third vertical federated learning request message to the primary participant. Correspondingly, the primary participant receives the third vertical federated learning request message from the network openness function network element.

[0524] In a possible implementation manner, the third vertical federated learning request message includes the third identifier and / or the third address assigned by the network openness function network element for at least one candidate secondary participant. It can be understood that similar to the network openness function network element assigning the first identifier and / or the first address for at least one VFL server in step S710, the third identifier and / or the third address assigned by the network openness function network element can avoid exposing the real identifiers and real addresses of at least one candidate secondary participant, thereby improving security.

[0525] For example, assume that at least one candidate slave participating party includes Network Data Analysis Function Network Element 1 and Network Data Analysis Function Network Element 2. The second identifier assigned by the Network Open Function Network Element to the master participating party is ID#1, the third identifier assigned to Network Data Analysis Function Network Element 1 is ID#2, and the third identifier assigned to Network Data Analysis Function Network Element 2 is ID#3. When the master participating party sends a message (such as the first sample alignment request message in step S718 below) to Network Data Analysis Function Network Element 1 through the Network Open Function Network Element, the Network Open Function Network Element carries ID#1 when sending the message to Network Data Analysis Function Network Element 1. In this way, Network Data Analysis Function Network Element 1 determines which participating party sent the message based on ID#1. When Network Data Analysis Function Network Element 2 sends a message to the master participating party through the Network Open Function Network Element, the Network Open Function Network Element carries ID#3 when sending the message to the master participating party. In this way, the master participating party determines that the message is sent by Network Data Analysis Function Network Element 2 based on ID#3.

[0526] It can be understood that according to the descriptions of the associated identifiers in the above steps S711, S712, and the above step S716, the third vertical federated learning request message may also include the associated identifier in step S716, so as to further distinguish multiple different vertical federated learning processes executed in parallel by the master participating party. For specific details, please refer to the relevant description in step S711, which will not be elaborated here.

[0527] In addition, in the following steps involving the interaction between the master participating party and at least one candidate slave participating party (such as the steps in the sample alignment negotiation process), the messages exchanged between the master participating party and at least one candidate slave participating party (including the target slave participating party) may carry the associated identifier, the third identifier, and / or the third address, etc. in the above step S716. This is uniformly explained here and will not be elaborated below.

[0528] The following introduces the sample alignment negotiation process, which includes steps S718 to S724.

[0529] S718. The master participating party sends a first sample alignment request message to the Network Open Function Network Element. Correspondingly, the Network Open Function Network Element receives the first sample alignment request message from the master participating party.

[0530] It can be understood that for the specific content of the first sample alignment request message, please refer to Figure 6 the sample alignment request message in step S610, which will not be elaborated here.

[0531] In addition, the first sample alignment request message may also include the associated identifier in step S717, so that the Network Open Function Network Element can determine which vertical federated learning process of the master participating party the first sample alignment request message belongs to.

[0532] In addition, the first sample alignment request message may further include the third identifier and / or the third address of each candidate secondary participant among at least one candidate secondary participant in step S717, so as to facilitate the network exposure function network element to determine which network function network elements (such as network data analysis function network elements) among multiple network function network elements in the network to send the information included in the first sample alignment request message.

[0533] S719. The network exposure function network element sends a second sample alignment request message to at least one candidate secondary participant. Accordingly, at least one candidate secondary participant receives the second sample alignment request message from the network exposure function network element.

[0534] It can be understood that the second sample alignment request message includes the information included in the first sample alignment request message.

[0535] In addition, the second sample alignment request message may include the second identifier and / or the second address of the primary participant, so as to facilitate at least one candidate secondary participant to determine that the second sample alignment request message is requested by the primary participant.

[0536] In addition, considering that at least some of the at least one candidate secondary participant may also participate in multiple different vertical federated learning processes in parallel, furthermore, the second sample alignment request message may further include the association identifier in step S718, so that at least one candidate secondary participant can determine the multiple different vertical federated learning processes in which it participates in parallel through the association identifier.

[0537] S720. At least one candidate secondary participant sends a second sample alignment response message to the network exposure function network element. Accordingly, the network exposure function network element receives the second sample alignment response message from at least one candidate secondary participant.

[0538] It can be understood that for the second sample alignment response message, reference can be specifically made to Figure 6 the sample alignment response message in step S611, which will not be elaborated here.

[0539] In addition, the second sample alignment response message may further include the association identifier in step S719, so as to facilitate the network exposure function network element to distinguish which vertical federated learning process the second sample alignment response message belongs to.

[0540] S721. The network exposure function network element sends a first sample alignment response message to the primary participant. Accordingly, the primary participant receives the first sample alignment response message from the network exposure function network element.

[0541] It can be understood that the first sample alignment response message includes the information included in the second sample alignment response message.

[0542] In addition, the first sample alignment response message may include the third identifier and / or the third address of at least one candidate slave participant, so as to facilitate the master participant to determine which candidate slave participant responds to the first sample alignment response message.

[0543] In addition, the first sample alignment response message may further include the association identifier in step S720, so as to facilitate the master participant to distinguish which vertical federated learning process the first sample alignment response message belongs to.

[0544] S722. The master participant sends a first sample alignment negotiation result message to the network openness function network element. Correspondingly, the network openness function network element receives the first sample alignment negotiation result message from the master participant.

[0545] It can be understood that for the specific content of the first sample alignment negotiation result message, reference can be made to Figure 6 the first sample alignment negotiation result message in step S612, which will not be elaborated here.

[0546] In addition, the first sample alignment negotiation result message may further include the association identifier in step S721, or the third identifier and / or the third address of the target slave participant, etc. The embodiments of the present application do not make specific limitations in this regard.

[0547] S723. The network openness function network element sends a second sample alignment negotiation result message to the target slave participant. Correspondingly, the target slave participant receives the second sample alignment negotiation result message from the network openness function network element.

[0548] It can be understood that the second sample alignment negotiation result message includes the information contained in the first sample alignment negotiation result message.

[0549] In addition, the second sample alignment negotiation result message may further include the second identifier and / or the second address of the master participant, or the association identifier in step S722, etc. The embodiments of the present application do not make specific limitations in this regard.

[0550] Optionally, Figure 7 the information transmission method flow shown may further include steps S724 and S725.

[0551] S724. The master participant sends a third sample alignment negotiation result message to the network openness function network element. Correspondingly, the network openness function network element receives the third sample alignment negotiation result message from the master participant.

[0552] It can be understood that for the specific content of the third sample alignment negotiation result message, reference can be made to Figure 6 the second sample alignment negotiation result message in step S613, which will not be elaborated here.

[0553] In addition, the third sample alignment negotiation result message may further include the third identifier and / or the third address, or the associated identifier, etc. of other slave participating parties among at least one candidate slave participating party except the target slave participating party. The embodiments of the present application do not make specific limitations thereto.

[0554] S725. The network opening function network element sends a fourth sample alignment negotiation result message to other slave participating parties among multiple slave participating parties that support sample alignment except the target slave participating party. Correspondingly, other slave participating parties among multiple slave participating parties that support sample alignment except the target slave participating party receive the fourth sample alignment negotiation result message from the network opening function network element.

[0555] It can be understood that the fourth sample alignment negotiation result message includes the information included in the third sample alignment negotiation result message.

[0556] In addition, the fourth sample alignment negotiation result message may further include the second identifier and / or the second address, or the associated identifier, etc. of the master participating party. The embodiments of the present application do not make specific limitations thereto.

[0557] S726. The sample alignment process is executed among the master participating party, the network opening function network element, and the target slave participating party.

[0558] It can be understood that step S726 is similar to step S614. The difference between the two is that: since the master participating party is an untrusted application function network element and the target slave participating party is a network data analysis function network element, the information interaction in the sample alignment process between the master participating party and the target slave participating party is forwarded and coordinated through the network opening function network element. In other words, the network opening function network element is responsible for maintaining the corresponding relationship between the master participating party and one or more target slave participating parties, or the distinction of the vertical federated learning process, etc. For details, reference can be made to the relevant description of step S716, which will not be elaborated here.

[0559] In addition, the messages exchanged among the master participating party, the network opening function network element, and the target slave participating party may carry the associated identifier, the second identifier, the third identifier, the second address, or the third address, etc. involved in the above steps. The embodiments of the present application do not make specific limitations thereto.

[0560] S727. The participating party sends a vertical federated learning response message to the VFL server. Correspondingly, the VFL server receives the vertical federated learning response message from the participating party.

[0561] It can be understood that step S727 is similar to Figure 6 step S615. In the case where the participating party includes the master participating party and the target slave participating party, step S727 may include:

[0562] S727a. The target slave participant sends a first vertical federated learning response message to the VFL server. Correspondingly, the VFL server receives the first vertical federated learning response message from the target slave participant.

[0563] It can be understood that for the specific implementation of step S727a, reference can be made to step S615a, which will not be elaborated here.

[0564] S727b. The master participant sends a second vertical federated learning response message to the network openness function network element. Correspondingly, the network openness function network element receives the second vertical federated learning response message from the master participant.

[0565] It can be understood that for the second vertical federated learning response message, reference can be made to Figure 6 the second vertical federated learning response message in step S615b, which will not be elaborated here.

[0566] In addition, the second vertical federated learning response message may further include the first identifier and / or the first address of the VFL server, or an associated identifier, etc. The embodiments of the present application do not make specific limitations thereon.

[0567] S727c. The network openness function network element sends a third vertical federated learning response message to the VFL server. Correspondingly, the VFL server receives the third vertical federated learning response message from the network openness function network element.

[0568] It can be understood that the third vertical federated learning response message includes the information contained in the second vertical federated learning response message.

[0569] In addition, the third vertical federated learning response message may include the second identifier and / or the second address of the master participant, or an associated identifier, etc. The embodiments of the present application do not make specific limitations thereon.

[0570] Since in the embodiments of the present application Figure 7For the information transmission method process shown, each participating party can register its sample alignment capability with the VFLSF network element. The primary participating party can discover the VFL server, which is the coordinator of vertical federated learning, through the VFLSF network element, and send a discovery request message through the VFL server to obtain at least one candidate secondary participating party and the sample alignment capabilities (such as sample alignment algorithms) of the at least one candidate secondary participating party. Furthermore, the primary participating party can perform sample alignment negotiation with the at least one candidate secondary participating party based on the first common sample alignment algorithm jointly supported between the primary participating party and the at least one candidate secondary participating party, so as to improve the success rate and efficiency of sample alignment negotiation. And through sample alignment negotiation, the primary participating party can determine, based on the responses of the at least one candidate secondary participating party, the target secondary participating party and the target sample alignment algorithm that the local computing power can support for performing sample alignment, so as to avoid the problem of sample alignment failure caused by not meeting the computing power requirements of the sample alignment algorithm during the sample alignment process, and improve the success rate and efficiency of sample alignment. Further, the above-mentioned primary participating party can be an untrusted network element, and the network open function network element can provide functions such as coordination, identity assignment, and trusted interaction for the interaction between the primary participating party and other network elements within a trusted domain, so as to better realize the interaction between the above-mentioned primary participating party, at least one candidate secondary participating party, VFL server, and VFLSF network element.

[0571] It can be understood that as Figure 5 the description of the VFLSF network element in Figure 7 shows, the VFLSF network element can be a network open function network element, or rather, the network open function network element has the functions of the VFLSF network element. Thus

[0572] In addition, for steps S711 - S712 and subsequent steps S716 - S728, the VFLSF network element replaces the network open function network element. That is, the VFLSF network element maintains the corresponding relationships among the primary participating party, the VFL server, and at least one candidate secondary participating party, the association identifier corresponding to the vertical federated learning process among the primary participating party, the VFL server, and at least one candidate secondary participating party, assigns a second identifier and / or a second address to the primary participating party, and assigns a third identifier and / or a third address to at least one candidate secondary participating party, etc. For specific details, please refer to Figure 7 the relevant descriptions in

[0573] and will not be elaborated here. Figure 6 It should be understood that in the information transmission method process shown above Figure 7 and Figure 8 , the primary participating party can also replace the VFL server to be responsible for coordination, that is, discover candidate secondary participating parties. The following will be described in combination with the information transmission method process shown in

[0574] Figure 8 is a schematic flow of an information transmission method provided by an embodiment of the present application. Figure 3 As shown in Figure 8 , the process of this information transmission method includes: steps S801 - S812. Steps S801 - S804 are the same as steps S601 - S604, steps S807 - S810 are the same as steps S610 - S613, and step S811 is the same as step S614, and will not be elaborated here.

[0575] It should be understood that the above steps S803, S804, and S812 are optional steps, that is, the VFL server may not participate in the sample alignment and the vertical federated learning model training process.

[0576] The following will separately describe steps S805, S806, and S812.

[0577] S805. The primary participating party sends a participating party discovery request message to the VFLSF network element. Correspondingly, the VFLSF network element receives the participating party discovery request message from the primary participating party. Among them, the participating party discovery request message is used to request to discover participating parties. The participating party discovery request message may include: the vertical federated learning group identifier, the vertical federated learning model type identifier, the requirements for secondary participating parties, the second vertical federated learning ability type indication, and the sample alignment requirements. For specific details, please refer to Figure 6 step S606 in

[0578] and will not be elaborated here. Figure 5When it comes to the VFLSF network element or the network storage function network element, the primary participant sends a participant discovery request message to the VFLSF network element, which is similar to the information transmission process shown in Figure 7 The primary participant can send a participant discovery request message to the VFLSF network element through the network exposure function network element. For example, the primary participant sends a first participant discovery request message to the network exposure function network element. The network exposure function network element verifies and authorizes the request of the primary participant. If the authorization is passed, then the network exposure function network element sends a second participant discovery request message to the VFLSF network element, and the second participant discovery request message includes the content contained in the first participant discovery request message.

[0579] In addition, similar to Figure 7 step S707, the primary participant can send the first participant discovery request message by invoking the Nnef VFL ParticipantDiscovery Request service operation of the network exposure function network element, and the network exposure function network element can send the second participant discovery request message by invoking the NnrfNFDiscovery Request service operation of the network storage function network element.

[0580] It can also be understood that when the primary participant is an untrusted application function network element and the VFLSF network element is the network exposure function network element, the primary participant can execute step S805, that is, directly send a participant discovery request message to the VFLSF network element.

[0581] S806: The VFLSF network element sends a participant discovery response message to the primary participant. Correspondingly, the primary participant receives the participant discovery response message from the VFLSF network element. Among them, the participant discovery response message includes information of at least one candidate secondary participant. The information of at least one candidate secondary participant may include: the identifier or address of at least one candidate secondary participant. The identifier or address of at least one candidate secondary participant can be used for the primary participant to communicate with the at least one candidate secondary participant.

[0582] It can be understood that for the participant discovery response message in step S806, specifically, reference can be made to Figure 6 step S607 therein, which will not be elaborated here.

[0583] It should be understood that when the primary participant is an untrusted application function network element and the VFLSF network element is the Figure 5 VFLSF network element or the network storage function network element in Figure 7Similar to the information transmission process shown in , the VFLSF network element can send a participant discovery response message to the primary participant through the network open function network element. For example, the VFLSF network element sends a second participant discovery response message to the network open function network element, and the network open function network element sends a first participant discovery response message to the primary participant, and the first participant discovery response message includes the information contained in the second participant discovery response message.

[0584] Understandably, Figure 7 Similar to the network open network element in, the network open functional network element maintains the corresponding relationship between the main participant and at least one candidate slave participant, and then when the main participant (untrusted application functional network element) interacts with at least one candidate slave participant (such as the network data analysis functional network element) through the network open functional network element in the vertical federated learning process (such as the sample alignment process, the vertical federated learning model training process, or the vertical federated learning reasoning process), the network open functional network element can determine to which network analysis network data analysis functional network element to send the message of the main participant, and to which application functional network element to send the message of the network data analysis functional network element according to the corresponding relationship.

[0585] In addition, the network open function network element can assign an association identifier to the vertical federated learning process between the main participant and at least one candidate slave participant to distinguish multiple vertical federated learning processes executed in parallel by the main participant. For details, please refer to steps S711 and S712, which will not be repeated here.

[0586] In addition, the network open function network element can also assign a second identifier and / or a second address to the main participant, and assign a first identifier and / or a first address to at least one slave participant, thereby avoiding exposure of the true identifier and true address of the main participant, and avoiding exposure of the true identifier and true address of at least one candidate slave participant. For details, please refer to the relevant instructions of step S712 and step S717, which will not be repeated here.

[0587] It should be understood that in the following steps involving the interaction between the master participant and at least one candidate slave participant, the messages interacting between the master participant and at least one candidate slave participant (including t...

Claims

1. An information transmission method, characterized in that, The method includes: A first network element sends a first request to a second network element, the first request including first information for indicating at least one first sample alignment algorithm supported by the first network element; The first network element receives a first response from the second network element, the first response being used to indicate a second sample alignment algorithm among the at least one first sample alignment algorithms, the second sample alignment algorithm being used to determine common samples between a data set of the first network element for vertical federated learning model training and a data set of the second network element for vertical federated learning model training; or the first response is used to indicate that sample alignment is not supported.

2. The method according to claim 1, characterized in that, The method further includes: The first network element receives information of the second network element from a third network element, the information of the second network element including an identifier of the second network element and / or address information of the second network element.

3. The method according to claim 2, wherein The method further includes: The first network element sends a second request to the third network element, the second request including second information for indicating a sample alignment requirement; The first network element receives information of the second network element from the third network element, including: the first network element receives a second response from the third network element, the second response including the information of the second network element.

4. The method according to claim 3, wherein The sample alignment requirement includes at least one of the following: a third sample alignment type, a third sample alignment algorithm, or a third time period for sample alignment.

5. The method according to any one of claims 2-4, characterized in that, The information of the second network element further includes at least one of the following: a fourth sample alignment type, a fourth sample alignment algorithm, or a fourth time period supporting sample alignment.

6. The method according to any one of claims 2-5, characterized in that, The method further includes: The first network element sends a registration request to the third network element, the registration request including third information for indicating at least one of the following: a sample alignment type supported by the first network element, a sample alignment algorithm supported, or a time period supporting sample alignment.

7. The method according to any one of claims 2 - 6, characterized in that The at least one first sample alignment algorithm is an algorithm jointly supported by the first network element and the second network element.

8. The method according to any one of claims 1 to 7, characterized in that, The first information is further used to indicate at least one of the following: the number of samples supported by the first network element for sample alignment, a first sample alignment type supported by the first network element, or at least one first time period during which the first network element supports sample alignment.

9. The method according to claim 8, wherein The at least one first sample alignment algorithm is associated with the at least one first time period.

10. The method according to any one of claims 1-9, characterized in that, The first response is further used to indicate the number of samples supported by the second network element for sample alignment, and / or a second time period during which the second network element supports sample alignment; Or, the first response is further used to indicate the reason why the second network element does not support sample alignment, and / or the maximum number of samples that the second network element expects the first network element to use.

11. The method according to any one of claims 1-10, characterized in that, The method further includes: The first network element sends a first message to the second network element, the first message being used to indicate a target sample alignment algorithm, the target sample alignment algorithm being a sample alignment algorithm determined from the at least one first sample alignment algorithm according to the second sample alignment algorithm.

12. The method according to claim 11, wherein The method further includes: The first network element sends the first request to a fourth network element; The first network element receives a third response from the fourth network element, where the third response is used to indicate a fifth sample alignment algorithm among the at least one first sample alignment algorithm, and the fifth sample alignment algorithm is used to determine common samples between the dataset of the first network element for vertical federated learning model training and the dataset of the fourth network element for vertical federated learning model training; The first network element sends a second message to the fourth network element, where the second message is used to indicate the target sample alignment algorithm; wherein, there are multiple fifth sample alignment algorithms and / or second sample alignment algorithms, and the target sample alignment algorithm is a sample alignment algorithm determined from the at least one first sample alignment algorithm according to the second sample alignment algorithm, including: the target sample alignment algorithm is a sample alignment algorithm determined from the at least one first sample alignment algorithm according to the second sample alignment algorithm and the fifth sample alignment algorithm.

13. The method according to claim 11 or 12, characterized in that, The first message is further used to indicate a target time period for sample alignment.

14. The method according to any one of claims 1 to 13, characterized in that, The method further includes: When the first network element determines not to perform sample alignment, the first network element sends a third message to the second network element, where the third message is used to indicate not performing sample alignment.

15. The method according to claim 14, wherein The third message is further used to indicate the reason for not performing sample alignment.

16. An information transmission method, characterized in that, The method includes: A first network element sends a first request to a second network element, where the first request includes first indication information, and the first indication information is used to indicate providing sample identifiers for vertical federated learning model training; The first network element receives a first response from the second network element, where the first response includes first sample identifiers.

17. The method according to claim 16, wherein The first indication information is used to indicate providing sample identifiers for vertical federated learning model training, including: the first indication information is used to indicate providing sample identifiers corresponding to the dataset for vertical federated learning model training, and the first sample identifiers are sample identifiers corresponding to the dataset of the second network element for vertical federated learning model training.

18. The method according to claim 16 or 17, characterized in that, The second network element is a candidate network element participating in the vertical federated learning model training; the method further includes: The first network element receives second sample identifiers corresponding to the dataset for vertical federated learning model training from a third network element, where the third network element is another network element other than the second network element among the multiple candidate network elements participating in the vertical federated learning model training; The first network element sends a third message to the second network element according to the first sample identifiers and the second sample identifiers, where the third message includes second indication information, and the second indication information is used to indicate that the second network element participates in the vertical federated learning model training, and / or, third sample identifiers, where the third sample identifiers are common sample identifiers of at least two network elements participating in the vertical federated learning model training, the at least two network elements are determined from the multiple candidate network elements according to the first sample identifiers and the second sample identifiers, and the at least two network elements include the third network element; Alternatively, the second indication information is used to indicate that the second network element does not participate in vertical federated learning model training, and / or the reason why the second network element does not participate in vertical federated learning model training.

19. The method according to claim 16, wherein The second network element is a network element corresponding to sample identifiers of a data set of network elements registered to participate in vertical federated learning model training; The first indication information is used to indicate providing sample identifiers for vertical federated learning model training, including: the first indication information is used to indicate providing common sample identifiers, where the common sample identifiers are common sample identifiers between data sets of at least two network elements participating in the same vertical federated learning model training among multiple network elements, and the first sample identifier is used to indicate common samples between data sets of at least two network elements participating in the first vertical federated learning model training among multiple network elements.

20. The method according to claim 19, wherein The first network element is a network element participating in vertical federated learning model training; the method further includes: The first network element sends a registration request to the second network element, and the registration request includes fourth indication information, where the fourth indication information is used to indicate sample identifiers corresponding to a data set of the first network element for vertical federated learning model training.

21. The method according to claim 19 or 20, characterized in that The first request further includes fifth indication information and / or sixth indication information, where the fifth indication information is used to indicate the quantity range of sample identifiers included in the common sample identifiers, and the sixth indication information is further used to indicate that the second network element determines the at least two network elements.

22. The method according to any one of claims 19-21, characterized in that, The first response further includes identifiers and / or address information of the at least two network elements, and / or the number of sample identifiers included in the first sample identifier.

23. A communication device, characterized in that, The communication device includes a module or unit for executing the method according to any one of claims 1-15, or includes a module or unit for executing the method according to any one of claims 16-22.

24. A communication device, characterized in that, The communication device includes a processor, and the processor is used to cause the communication device to execute the method according to any one of claims 1-15 through logic circuits and / or executing instructions, or cause the communication device to execute the method according to any one of claims 16-22.

25. The communication device according to claim 24, wherein The communication device further includes a memory, and the memory is used to store the instructions.

26. The communication device according to claim 24 or 25, characterized in that, The communication device further includes a communication interface, and the communication interface is used to input and / or output signaling and / or data.

27. A computer-readable storage medium, characterized in that The computer-readable storage medium includes instructions, and when the instructions are run by a processor, the method according to any one of claims 1-15 is implemented, or the method according to any one of claims 16-22 is implemented.

28. A computer program product, characterized in that, The computer program product includes instructions, and when the instructions are run on a computer, the computer is caused to execute the method according to any one of claims 1-15, or the computer is caused to execute the method according to any one of claims 16-22.

Images