Information interaction method and apparatus, proxy node, federated learning client and network element

Abstract

The present application belongs to the technical field of communications. Disclosed are an information interaction method and apparatus, a proxy node, a federated learning client and a network element. The information interaction method of the embodiments of the present application comprises: a proxy node receives a first request message from a federated learning server, the proxy node being a device supporting proxy federated learning, and the first request message being used for requesting executing a federated learning-related operation; and the proxy node executes a first operation, the first operation comprising at least one of the following: sending to at least one first federated learning client a second request message which is used for requesting executing the federated learning-related operation, and executing the federated learning-related operation on the basis of the first request message.

Description

Information exchange methods, devices, agent nodes, federated learning clients, and network elements

[0001] Cross-reference to related applications

[0002] This application claims priority to Chinese Patent Application No. 202411917469.1, filed in China on December 24, 2024, the entire contents of which are incorporated herein by reference. Technical Field

[0003] This application belongs to the field of communication technology, specifically relating to an information interaction method, device, proxy node, federated learning client, and network element. Background Technology

[0004] Federated learning can be divided into vertical federated learning (VFL) and horizontal federated learning (HFL). Vertical federated learning essentially involves the jointing of features and is suitable for scenarios with significant user overlap but minimal feature overlap. For example, within a communication network, the core network (CN) and radio access network (RAN) domains serve different services for the same user. By jointly combining the different data features of common samples from participating parties, vertical federation increases the feature dimensions of the training samples, thereby facilitating the training of a better-performing model.

[0005] Currently, in the context of federated learning, there are some inconveniences in the interaction between federated learning devices. For example, external devices participating in federated learning (e.g., application functions (AF)) have difficulty interacting with user equipment (UE) or RAN devices participating in federated learning to exchange information related to the federated learning process, which in turn affects the execution efficiency of the federated learning process. Summary of the Invention

[0006] This application provides an information interaction method, apparatus, proxy node, federated learning client, and network element, which can perform federated learning on behalf of the federated learning device through the proxy node during the federated learning process, thereby improving the execution efficiency of the federated learning process.

[0007] Firstly, an information exchange method is provided, the method comprising:

[0008] The proxy node receives a first request message from the federated learning server, wherein the proxy node is a device that supports proxy federated learning, and the first request message is used to request the execution of federated learning-related operations;

[0009] The proxy node performs a first operation, which includes at least one of the following:

[0010] Send a second request message to at least one first federated learning client, the second request message being used to request the execution of the federated learning-related operations;

[0011] Perform the federated learning-related operations according to the first request message.

[0012] Secondly, an information interaction device is provided for use in a proxy node, the device comprising:

[0013] The receiving module is used to receive a first request message from the federated learning server, wherein the proxy node is a device that supports proxy federated learning, and the first request message is used to request the execution of federated learning-related operations.

[0014] A processing module is configured to perform a first operation, the first operation including at least one of the following:

[0015] Send a second request message to at least one first federated learning client, the second request message being used to request the execution of the federated learning-related operations;

[0016] Perform the federated learning-related operations according to the first request message.

[0017] Thirdly, an information exchange method is provided, which includes:

[0018] The federated learning client performs the second operation;

[0019] The second operation includes at least one of the following:

[0020] Receive a second request message from the proxy node, the second request message being used to request the execution of federated learning related operations;

[0021] Receive a data request message sent by the proxy node, the data request message being used to request data related to the federated learning client;

[0022] Send the data from the federated learning client to the proxy node;

[0023] A first registration request message is sent to the first network element. The first registration request message includes the capability information of the federated learning client, and the capability information of the federated learning client includes at least one of the following:

[0024] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0025] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0026] Information about the agent nodes supported by the federated learning client;

[0027] The identifier of the federated learning task supported by the federated learning client;

[0028] The agent node is a device that supports agent federated learning.

[0029] Fourthly, an information interaction device is provided, the device comprising:

[0030] The processing module is used to perform the second operation;

[0031] The second operation includes at least one of the following:

[0032] Receive a second request message from the proxy node, the second request message being used to request the execution of federated learning related operations;

[0033] Receive a data request message sent by the proxy node, the data request message being used to request data related to the federated learning client;

[0034] Send the data from the federated learning client to the proxy node;

[0035] A first registration request message is sent to the first network element. The first registration request message includes the capability information of the federated learning client, and the capability information of the federated learning client includes at least one of the following:

[0036] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0037] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0038] Information about the agent nodes supported by the federated learning client;

[0039] The identifier of the federated learning task supported by the federated learning client;

[0040] The agent node is a device that supports agent federated learning.

[0041] Fifthly, an information exchange method is provided, which includes:

[0042] The first network element receives the first registration request message from the federated learning client;

[0043] The first registration request message includes the capability information of the federated learning client, which includes at least one of the following:

[0044] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0045] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0046] The information of the proxy nodes supported by the federated learning client, wherein the proxy nodes are devices that support proxy federated learning;

[0047] The identifier of the federated learning tasks supported by the federated learning client.

[0048] Sixthly, an information interaction device is provided, the device comprising:

[0049] The receiving module is used to receive the first registration request message from the federated learning client;

[0050] The first registration request message includes the capability information of the federated learning client, which includes at least one of the following:

[0051] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0052] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0053] The information of the proxy nodes supported by the federated learning client, wherein the proxy nodes are devices that support proxy federated learning;

[0054] The identifier of the federated learning tasks supported by the federated learning client.

[0055] Seventhly, an information exchange method is provided, the method comprising:

[0056] The second network element receives the second registration request message from the first network element;

[0057] The second registration request message includes the capability information of the federated learning client, which includes at least one of the following:

[0058] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0059] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0060] Information about the agent nodes supported by the federated learning client;

[0061] The identifier of the federated learning tasks supported by the federated learning client.

[0062] Eighthly, an information interaction device is provided, the device comprising:

[0063] The receiving module is used to receive the second registration request message from the first network element;

[0064] The second registration request message includes the capability information of the federated learning client, which includes at least one of the following:

[0065] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0066] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0067] Information about the agent nodes supported by the federated learning client;

[0068] The identifier of the federated learning tasks supported by the federated learning client.

[0069] Ninth aspect, an information interaction device is provided, the device being configured to perform the steps of the method described in the first aspect, or implement the steps of the method described in the third aspect, or implement the steps of the method described in the fifth aspect, or implement the steps of the method described in the seventh aspect.

[0070] In a tenth aspect, a proxy node is provided, the proxy node including a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the method as described in the first aspect.

[0071] Eleventhly, a proxy node is provided, including a processor and a communication interface, wherein the communication interface is used to receive a first request message from a federated learning server, wherein the proxy node is a device that supports proxy federated learning, and the first request message is used to request the execution of federated learning-related operations;

[0072] The processor is configured to perform a first operation, the first operation including at least one of the following:

[0073] Send a second request message to at least one first federated learning client, the second request message being used to request the execution of the federated learning-related operations;

[0074] Perform the federated learning-related operations according to the first request message.

[0075] In a twelfth aspect, a federated learning client is provided, the federated learning client including a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the method as described in the third aspect.

[0076] In a thirteenth aspect, a federated learning client is provided, including a processor and a communication interface, wherein the processor is configured to perform a second operation; wherein the second operation includes at least one of the following:

[0077] Receive a second request message from the proxy node, the second request message being used to request the execution of federated learning related operations;

[0078] Receive a data request message sent by the proxy node, the data request message being used to request data related to the federated learning client;

[0079] Send the data from the federated learning client to the proxy node;

[0080] A first registration request message is sent to the first network element. The first registration request message includes the capability information of the federated learning client, and the capability information of the federated learning client includes at least one of the following:

[0081] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0082] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0083] Information about the agent nodes supported by the federated learning client;

[0084] The identifier of the federated learning task supported by the federated learning client;

[0085] The agent node is a device that supports agent federated learning.

[0086] In a fourteenth aspect, a first network element is provided, the first network element including a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the method as described in the fifth aspect.

[0087] In a fifteenth aspect, a first network element is provided, including a processor and a communication interface, wherein the communication interface is used to receive a first registration request message from a federated learning client; wherein the first registration request message includes capability information of the federated learning client, and the capability information of the federated learning client includes at least one of the following:

[0088] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0089] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0090] The information of the proxy nodes supported by the federated learning client, wherein the proxy nodes are devices that support proxy federated learning;

[0091] The identifier of the federated learning tasks supported by the federated learning client.

[0092] In a sixteenth aspect, a second network element is provided, the second network element including a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the method described in the seventh aspect.

[0093] In a seventeenth aspect, a second network element is provided, including a processor and a communication interface, wherein the communication interface is used to receive a second registration request message from a first network element; wherein the second registration request message includes capability information of a federated learning client, and the capability information of the federated learning client includes at least one of the following:

[0094] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0095] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0096] Information about the agent nodes supported by the federated learning client;

[0097] The identifier of the federated learning tasks supported by the federated learning client.

[0098] In an eighteenth aspect, a readable storage medium is provided, on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect, or the steps of the method described in the third aspect, or the steps of the method described in the fifth aspect, or the steps of the method described in the seventh aspect.

[0099] In a nineteenth aspect, a wireless communication system is provided, comprising: a proxy node and a federated learning client, wherein the proxy node can be used to perform the steps of the information interaction method as described in the first aspect, and the federated learning client can be used to perform the steps of the information interaction method as described in the third aspect.

[0100] In a twentieth aspect, a chip is provided, the chip including a processor and a communication interface coupled to the processor, the processor being configured to run a program or instructions to implement the steps of the method described in the first aspect, or the steps of the method described in the third aspect, or the steps of the method described in the fifth aspect, or the steps of the method described in the seventh aspect.

[0101] In a twenty-first aspect, a computer program / program product is provided, the computer program / program product being stored in a storage medium, the computer program / program product being executed by at least one processor to implement the steps of the method as described in the first aspect, or the steps of the method as described in the third aspect, or the steps of the method as described in the fifth aspect, or the steps of the method as described in the seventh aspect.

[0102] In this embodiment, a proxy node receives a first request message from a federated learning server. The proxy node is a device that supports proxy federated learning. The first request message requests the execution of federated learning-related operations. The proxy node executes a first operation, which includes at least one of the following: sending a second request message to at least one first federated learning client, the second request message requesting the execution of the federated learning-related operation; and executing the federated learning-related operation according to the first request message. Since this embodiment can use a proxy node to proxy federated learning clients for federated learning, the execution efficiency of the federated learning process can be improved. Attached Figure Description

[0103] Figure 1 is a block diagram of a wireless communication system applicable to an embodiment of this application;

[0104] Figure 2a is a schematic diagram of vertical federated learning provided in an embodiment of this application;

[0105] Figure 2b is a schematic diagram of horizontal federated learning provided in an embodiment of this application;

[0106] Figure 3a is a schematic diagram of the neural network provided in an embodiment of this application;

[0107] Figure 3b is a schematic diagram of a neuron provided in an embodiment of this application;

[0108] Figure 4 is a flowchart of an information interaction method provided in an embodiment of this application;

[0109] Figure 5 is a flowchart of another information interaction method provided in an embodiment of this application;

[0110] Figure 6 is a flowchart of another information interaction method provided in an embodiment of this application;

[0111] Figure 7 is a flowchart of another information interaction method provided in an embodiment of this application;

[0112] Figure 8 is a flowchart of another information interaction method provided in an embodiment of this application;

[0113] Figure 9 is a flowchart of another information interaction method provided in an embodiment of this application;

[0114] Figure 10 is a flowchart of another information interaction method provided in an embodiment of this application;

[0115] Figure 11 is a structural diagram of an information interaction device provided in an embodiment of this application;

[0116] Figure 12 is a structural diagram of another information interaction device provided in an embodiment of this application;

[0117] Figure 13 is a structural diagram of another information interaction device provided in an embodiment of this application;

[0118] Figure 14 is a structural diagram of another information interaction device provided in an embodiment of this application;

[0119] Figure 15 is a structural diagram of the communication device provided in an embodiment of this application;

[0120] Figure 16 is a structural diagram of a network-side device provided in an embodiment of this application;

[0121] Figure 17 is a structural diagram of another network-side device provided in an embodiment of this application;

[0122] Figure 18 is a structural diagram of the terminal provided in an embodiment of this application. Detailed Implementation

[0123] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0124] The terms "first," "second," etc., used in this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same class, not limited in number; for example, the first object can be one or more. Furthermore, "or" in this application indicates at least one of the connected objects. For example, the scope of protection for "A or B" covers at least three scenarios: Scenario 1: including A but not B; Scenario 2: including B but not A; Scenario 3: including both A and B. In addition, the terms "A and / or B," "at least one of A and B," and "at least one of A or B" also cover at least the above three scenarios. The character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0125] The term "instruction" in this application can be either a direct instruction (or explicit instruction) or an indirect instruction (or implicit instruction). A direct instruction can be understood as the sender explicitly informing the receiver of specific information, the required operation, or the requested result in the instruction sent. An indirect instruction can be understood as the receiver determining the corresponding information based on the instruction sent by the sender, or making a judgment and determining the required operation or requested result based on the judgment result.

[0126] It is worth noting that the technologies described in this application are not limited to Long Term Evolution (LTE) / LTE-Advanced (LTE-A) systems, but can also be used in other wireless communication systems, such as Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Single-carrier Frequency-Division Multiple Access (SC-FDMA), or other systems. The terms "system" and "network" in this application are often used interchangeably, and the described technologies can be used with the systems and radio technologies mentioned above, as well as with other systems and radio technologies. The following description describes New Radio (NR) systems for illustrative purposes, and the term NR is used in most of the following description; however, these technologies can also be applied to systems other than NR systems, such as 6th generation (6G) radio systems. th Generation 6G communication system.

[0127] Figure 1 shows a block diagram of a wireless communication system applicable to an embodiment of this application. The wireless communication system includes a terminal 11 and a network-side device 12. The terminal 11 can also be referred to as User Equipment (UE), and can be a mobile phone, tablet computer, laptop computer, notebook computer, personal digital assistant (PDA), handheld computer, netbook, ultra-mobile personal computer (UMPC), mobile internet device (MID), augmented reality (AR), virtual reality (VR) device, robot, wearable device, flight vehicle, vehicle user equipment (VUE), shipboard equipment, pedestrian user equipment (PUE), smart home (home devices with wireless communication capabilities, such as refrigerators, televisions, washing machines, or furniture), game console, personal computer (PC), ATM, or self-service machine, etc. Wearable devices include: smartwatches, smart bracelets, smart headphones, smart glasses, smart jewelry (smart bracelets, smart chains, smart rings, smart necklaces, smart anklets, smart anklets, etc.), smart wristbands, smart clothing, etc. Among these, in-vehicle devices can also be referred to as in-vehicle terminals, in-vehicle controllers, in-vehicle modules, in-vehicle components, in-vehicle chips, or in-vehicle units, etc. It should be noted that the specific type of terminal 11 is not limited in this application embodiment. Network-side equipment 12 may include access network equipment or core network equipment, wherein access network equipment may also be referred to as Radio Access Network (RAN) equipment, radio access network function, or radio access network unit. Access network equipment may include base stations, Wireless Local Area Network (WLAN) access points (APs), or Wireless Fidelity (WiFi) nodes, etc.Among them, base stations can be referred to as Node B (NB), Evolved Node B (eNB), Next Generation Node B (gNB), New Radio Node B (NR Node B), Access Point, Relay Base Station (RBS), Serving Base Station (SBS), Base Transceiver Station (BTS), Radio Base Station, Radio Transceiver, Basic Service Set (BSS), Extended Service Set (ESS), Home Node B (HNB), Home Evolved Node B, Transmit / Receive Point (TRP), Non-Terrestrial Network (NTN) equipment (such as satellite or high altitude platform stations). The term "base station" can be any suitable term in the field, such as "station" or any other appropriate term in the relevant field, as long as the same technical effect is achieved. The term "base station" is not limited to any specific technical term. It should be noted that the embodiments of this application only use the base station in the NR system as an example for introduction, and do not limit the specific type of base station.

[0128] Core network equipment, also known as core network nodes, core network functions, or core network elements, includes, but is not limited to, at least one of the following: Mobility Management Entity (MME), Access and Mobility Management Function (AMF), Session Management Function (SMF), User Plane Function (UPF), Policy Control Function (PCF), Policy and Charging Rules Function (PCRF), Edge Application Server Discovery Function (EASDF), Unified Data Management (UDM), Unified Data Repository (UDR), Home Subscriber Server (HSS), Centralized network configuration (CNC), Network Repository Function (NRF), Network Exposure Function (NEF), Local NEF (L-NEF), and Binding Support. Functions include BSF, Application Function (AF), Location Management Function (LMF), Gateway Mobile Location Centre (GMLC), Network Data Analytics Function (NWDAF), and Non-Terrestrial Network (NTN) equipment (such as satellite or high altitude platform station).It should be noted that the embodiments of this application only use the core network equipment in the NR system as an example for introduction, and do not limit the specific type of core network equipment. If the name of the core network equipment mentioned in the embodiments of this application changes in subsequent protocol versions (e.g., 6G), it is also within the scope of protection of this application.

[0129] Optionally, the core network equipment can be implemented by one or more functional modules in a single device, or by multiple devices working together; this application does not specifically limit this. It is understood that the aforementioned functional modules can be network elements in hardware devices, software functional modules running on dedicated hardware, or virtualized functional modules instantiated on a platform (e.g., a cloud platform).

[0130] For ease of understanding, the following describes some aspects of the embodiments of this application:

[0131] I. VFL

[0132] Vertical federated learning is essentially the jointing of features, suitable for scenarios with high user overlap and low feature overlap, such as different services (e.g., Mobility Management (MM) and Session Management (SM)) for the same user (e.g., UE) in the CN and RAN domains of a communication network. Here, the same user reflects identical samples, and the different services reflect different features. By jointly combining the different data features of common samples from participating parties, vertical federation increases the feature dimensions of the training samples, resulting in a better model. Figure 2a illustrates vertical federated learning, and Figure 2b illustrates horizontal federated learning.

[0133] One of the characteristics of VFL is that it involves scenarios where users overlap but features differ. Therefore, the two or more parties involved in VFL training / inference need to have overlapping users, i.e., the same sample data.

[0134] II. Artificial Intelligence (AI) and AI Models

[0135] Artificial intelligence has been widely applied in various fields. AI models can be implemented using various algorithms, such as neural networks, decision trees, support vector machines, and Bayesian classifiers. This application uses neural networks as an example for illustration, but it does not limit the specific type of AI module.

[0136] For example, a neural network can be shown in Figure 3a, where X1, X2…Xn are input values, Y is the output result, and a small circle represents a neuron, which is also where the operation is performed. The result is then passed to the next layer. These numerous neurons forming an input layer, hidden layer, and output layer constitute a neural network. The number of hidden layers and the number of neurons in each layer constitute the "network structure" of the neural network.

[0137] The neural network consists of neurons, each of which can be represented as shown in Figure 3b. Here, a1, a2, ..., aK (i.e., X1...Xn from above) are the inputs, w is the weight (multiplicative coefficient), b is the bias (additive coefficient), σ(.) is the activation function, and z is the output value. Common activation functions include Sigmoid, tanh, ReLU (Rectified Linear Unit), etc. The parameters of each neuron, combined with the algorithm used, constitute the "parameter information" of the entire network, which is a crucial part of the AI ​​model file.

[0138] In practical use, an AI model refers to a file containing elements such as network structure and parameter information. The trained AI model can be directly reused by its framework platform without repeated construction or learning, and can directly perform intelligent functions such as judgment and recognition.

[0139] The information interaction method provided in this application will be described in detail below with reference to the accompanying drawings and through some embodiments and application scenarios.

[0140] Please refer to Figure 4, which is a flowchart of an information interaction method provided in an embodiment of this application. This method can be executed by a proxy node, and as shown in Figure 4, it includes the following steps:

[0141] Step 401: The proxy node receives a first request message from the federated learning server, wherein the proxy node is a device that supports proxy federated learning, and the first request message is used to request the execution of federated learning-related operations.

[0142] In this embodiment, the agent node can be any device that supports agent-based federated learning. For example, the agent node can relay messages from the federated learning node or perform federated learning on behalf of the federated learning node. Exemplarily, the agent node can be a newly added network element or an enhanced version of an existing network element, such as an enhanced NEF or NWDAF. Exemplarily, the agent node can directly interact with the federated learning client.

[0143] In some optional embodiments, the agent node may also possess federated learning-related capabilities, such as at least one of the following: federated learning training capabilities, federated learning participant capabilities, federated learning client capabilities, vertical federated learning client capabilities, and horizontal federated learning client capabilities. The aforementioned agent node may also be referred to as an Agent Federated Learning Client, for example, an Agent VFL Client; that is, the aforementioned agent node can be considered a special type of federated learning client. The aforementioned federated learning node may include at least one of a federated learning client and a federated learning server.

[0144] The aforementioned Federated Learning Client refers to the device participating in the Federated Learning process, such as UE, RAN, or core network element.

[0145] The aforementioned Federated Learning Server is a device with Federated Learning Server capabilities. For example, the aforementioned Federated Learning Server can be used for at least one of the following: finding and discovering Federated Learning Clients, triggering Federated Learning training, triggering Federated Learning inference, generating intermediate results for updating the model, and determining the final inference result, etc. For example, the aforementioned Federated Learning Server can be an AF (Automatic Feedback Provider). Exemplarily, the AF can represent a communication device outside the core network, such as a third-party server.

[0146] The aforementioned federated learning-related operations may include at least one of federated learning preparation or preparation for federated learning, federated learning training, and federated learning inference. For example, the federated learning server may decide to perform federated learning-related processes. For instance, the VFL server, based on internal logic, pre-configuration, or requests from other devices, decides to perform VFL-related processes and can then send a first request message to the federated learning client through a proxy node. Federated learning preparation may further include at least one of sample alignment, feature alignment, and confirmation of participation.

[0147] It should be noted that the first request message can be sent directly to the proxy node by the federated learning server, or it can be forwarded to the proxy node through a device such as NEF. For example, if the federated learning server is a trusted device, the first request message can be sent directly to the proxy node by the federated learning server; if the federated learning server is not a trusted device, the first request message can be forwarded to the proxy node by the federated learning server through a device such as NEF.

[0148] For example, when the first request message is a Prepare Request message for requesting accurate federated learning, the first request message may include at least one of the following:

[0149] Federated learning type identifier, used to indicate the type of federated learning, such as vertical federated learning or horizontal federated learning;

[0150] The task type in federated learning can be, for example, a VFL task type, or an analytics ID or federated learning task ID. It's important to note that a task can also refer to an event, analytics, data analysis, or a data analysis task, all conveying the same meaning. For instance, a VFL task could be a VFL-related event or a VFL-related data analysis.

[0151] The first sample identification information is used to indicate the sample identification information that the Federated Learning Server has, can obtain, or can be used in the Federated Learning related process; it can be in the form of a list, that is, there can be multiple sample identification information.

[0152] The first feature information is used to indicate at least one of the following: the feature information that the federated learning server has or can obtain, and the feature information that the federated learning server expects.

[0153] When the aforementioned first request message is a training request message for requesting federated learning training, the aforementioned first request message may include at least one of the following:

[0154] The federated learning correlation ID can be used to indicate the tasks related to federated learning. Subsequent federated learning processes (such as federated learning inference) can use this federated learning correlation ID to determine the model related to the training of this federated learning session.

[0155] The second sample identification information is used to indicate the final sample identification information, that is, the sample identification information used for federated learning training. It should be noted that the final sample identification information can be sent in multiple times, so that each round of training is completed based on different samples.

[0156] The second feature information is used to indicate the feature information used for federated learning training;

[0157] Federated learning task constraint information describes the relevant constraints or conditions for the execution of the federated learning task, such as at least one of the following: execution time, execution area, termination conditions, etc.

[0158] Instructions for requesting participation in federated learning are used to instruct clients requesting participation in federated learning, etc.

[0159] In some optional embodiments, the federated learning involved in the embodiments of this application can be vertical federated learning or horizontal federated learning.

[0160] Step 402: The proxy node performs a first operation, which includes at least one of the following:

[0161] Send a second request message to at least one first federated learning client, the second request message being used to request the execution of the federated learning-related operations;

[0162] Perform the federated learning-related operations according to the first request message.

[0163] In this embodiment, the first federated learning client can be any federated learning client, that is, a device with federated learning client capabilities, or a device that can participate in federated learning (e.g., training and / or inference).

[0164] In some implementations, upon receiving a first request message, the proxy node can send a second request message to at least one first federated learning client. For example, the proxy node can send second request messages to M federated learning clients, where M is a positive integer. The second request message can be the first request message, meaning the proxy node can forward the received first request message to at least one federated learning client. Alternatively, the second request message can be a newly generated request message based on the message content of the first request message. The message content of the second request messages sent to different federated learning clients can be the same or different. In this implementation, the proxy node can relay messages between the federated learning client and the federated learning server, thus enabling convenient interaction between them.

[0165] In some optional embodiments, the proxy node can determine the at least one first federated learning client based on the first request message. For example, the first request message may include identification information of the at least one first federated learning client. In this case, the proxy node can determine the at least one first federated learning client based on the identification information of the at least one first federated learning client. Alternatively, the proxy node can determine the at least one first federated learning client based on at least one of the federated learning type identifier, sample identifier information, and feature information carried in the first request message. For example, a federated learning client that can participate in the federated learning type indicated by the federated learning type identifier can be determined as the at least one first federated learning client. Or, a federated learning client with the sample identifier information can be determined as the at least one first federated learning client.

[0166] In other implementations, upon receiving a first request message, the proxy node can perform the federated learning-related operations based on the first request message. For example, the proxy node itself may act as a federated learning client to perform the federated learning-related operations. In this case, the proxy node performs the federated learning-related operations based on its own data. Alternatively, the proxy node may perform the federated learning-related operations on behalf of the federated learning client. In this case, the proxy node performs the federated learning-related operations based on the data obtained from the replaced federated learning client, and the replaced federated learning device may not participate in the corresponding federated learning process. This reduces the load and computing power consumption of the replaced federated learning device.

[0167] In some other implementations, upon receiving a first request message, the proxy node can send a second request message to at least one first federated learning client and perform the federated learning-related operations according to the first request message. For example, the proxy node can act as a relay node for some federated learning clients (i.e., at least one first federated learning client) to relay messages from the federated learning server to these federated learning clients. The proxy node itself is also a federated learning client and performs the federated learning-related operations. Alternatively, the proxy node can also act as a substitute node for other federated learning clients (e.g., at least one second federated learning client) to perform federated learning on behalf of these federated learning clients. This not only helps to improve the efficiency of the federated learning process but also helps to improve the flexibility of federated learning.

[0168] Optionally, the agent federated learning includes at least one of the following:

[0169] Relay messages from federated learning nodes;

[0170] It replaces federated learning nodes for federated learning.

[0171] In this embodiment, the messages relayed by the federated learning nodes can also be referred to as forwarded, relayed, or bridged messages. For example, a proxy node can send messages from the federated learning server to the federated learning client, and vice versa, enabling convenient interaction between the federated learning server and the federated learning client. In this case, the relayed federated learning node does not need to directly participate in the corresponding federated learning process, thus reducing the risk of privacy data exposure and protecting data privacy.

[0172] It should be noted that when relaying messages from federated learning nodes, proxy nodes can also perform some processing, such as message aggregation, message distribution, or splitting.

[0173] The aforementioned replacement of federated learning nodes for federated learning can be understood as participating in the federated learning process based on the data related to the acquired federated learning nodes. For example, participating in at least one of the federated learning processes such as federated learning preparation, federated learning training, and federated learning inference based on the data related to the federated learning nodes. In this case, the replaced federated learning node may not participate in the corresponding federated learning process, which can reduce the load and computing power consumption of the replaced federated learning node.

[0174] In some alternative embodiments, the replaced federated learning node can be a federated learning client, such as a UE or RAN.

[0175] Optionally, performing the federated learning-related operations based on the first request message includes:

[0176] The proxy node receives data related to at least one second federated learning client.

[0177] The proxy node performs the federated learning-related operations based on the first request message and the data related to the at least one second federated learning client.

[0178] In this embodiment, the second federated learning client can be any federated learning client. The at least one second federated learning client can be different from or the same as the at least one first federated learning client. For example, when the first operation includes sending a second request message to at least one first federated learning client and performing the federated learning-related operation based on the first request message, the at least one second federated learning client can be different from the at least one first federated learning client; when the first operation includes sending a second request message to at least one first federated learning client and performing one of the federated learning-related operations based on the first request message, the at least one second federated learning client can be the same as or different from the at least one first federated learning client.

[0179] The aforementioned proxy node receives data related to at least one second federated learning client and performs federated learning-related operations based on this data. For example, the proxy node receives data related to N federated learning clients and performs the federated learning-related operations based on the N data, where N is a positive integer. The data related to the second federated learning clients can be understood as the federated learning-related data of the respective second federated learning clients. In this embodiment, the proxy node performs federated learning on behalf of at least one second federated learning client, which helps reduce the load and computational overhead of the at least one second federated learning client.

[0180] In some optional embodiments, the proxy node can determine the at least one second federated learning client based on the first request message. For example, the first request message may include identification information of the at least one second federated learning client. In this case, the proxy node can determine the at least one second federated learning client based on the identification information of the at least one second federated learning client. Alternatively, the proxy node can determine the at least one second federated learning client based on at least one of the federated learning type identifier, sample identifier information, and feature information carried in the first request message. For example, a federated learning client that can participate in the federated learning type indicated by the federated learning type identifier can be determined as the at least one second federated learning client, or a federated learning client with the sample identifier information can be determined as the at least one second federated learning client.

[0181] For example, when the first request message is a preparation request message, the proxy node can determine whether to participate in the federated learning process based on the data related to the second federated learning client. For instance, it can compare the data related to the second federated learning client with the first sample identifier information and / or first feature information carried in the first request message to determine whether they are the same or whether there is an intersection. For example, it can determine whether the data related to the second federated learning client has a sample identifier that is the same as the first sample identifier information, or whether there is an intersection between the data related to the second federated learning client and the first sample identifier information, or whether the data related to the second federated learning client has feature information that is the same as the first feature information. If there is an intersection, the proxy node can decide to participate in the federated learning process. When the proxy node determines to participate in the federated learning process, it can send a response message to the federated learning server. The response message can include at least one of the following: third sample identifier information, the intersection of sample identifiers, and third feature information. The third sample identifier information can be used to indicate sample identifier information that the proxy node has, can obtain, or can be used in the VFL-related process. It can be in the form of a list, i.e., multiple sample identifier information. The third feature information can be used to indicate feature information that the proxy node has or can obtain.

[0182] In some alternative embodiments, the proxy node itself may also be a federated learning client. In this case, the proxy node can perform the federated learning-related operations based on data associated with the proxy node and data associated with the at least one second federated learning client.

[0183] Optionally, the method further includes:

[0184] The proxy node sends a data request message to the at least one second federated learning client, the data request message being used to request data related to the second federated learning client.

[0185] In this embodiment, before the proxy node receives data related to at least one second federated learning client, the proxy node can send a data request message to at least one second federated learning client to request the acquisition of data related to the second federated learning client. For example, the proxy node can send a data collection subscribe message to the federated learning client, which may carry a federated learning task ID or a federated learning task type. Exemplarily, the proxy node can decide whether to collect data from federated learning clients such as UEs and / or RANs based on its internal logic, pre-configuration, or requests from other devices.

[0186] Optionally, the data request message includes at least one of the following:

[0187] Federal Learning Task Identifier;

[0188] Types of Federated Learning Tasks;

[0189] Data constraint information is used to indicate the conditions that the required feedback data must meet;

[0190] Feedback condition information indicates the triggering conditions for data feedback.

[0191] For example, the above-mentioned federated learning task ID can be an analytics ID.

[0192] For example, the above-mentioned federated learning task type can be a VFL task type, or an analytics ID, a federated learning task ID, etc.

[0193] The above-mentioned federated learning task types can also be used to indicate federated learning types, such as vertical federated learning or horizontal federated learning.

[0194] The task identifier mentioned above for federated learning can indicate a specific federated learning task or a task type. For example, it could be a VFL task type, an analytics ID, or a federated learning task ID. It should be noted that a task can also refer to an event, analytics, data analysis, or a data analysis task, etc., all conveying the same meaning. For example, a VFL task could be a VFL-related event or a VFL-related data analysis.

[0195] For example, the aforementioned data limitation information may include, but is not limited to, at least one of data type, first time information, and first area information. Specifically, the first time information indicates that the data to be fed back must be data within the time period indicated by the first time information, and the first area information indicates that the data to be fed back must be data within the area indicated by the first area information.

[0196] For example, the feedback condition information described above can be used to indicate at least one of the following: event-based feedback and time-based feedback. Event-based feedback can be understood as the federated learning client providing relevant data in response to a specific event; time-based feedback may include the federated learning client periodically providing relevant data or providing relevant data at a specific time. In some alternative embodiments, the feedback condition information may also be referred to as message notification constraints or data reporting conditions, etc.

[0197] It should be noted that federated learning clients such as UE and RAN can generate and store relevant data before receiving data request messages from agent nodes; they can also generate new relevant data after receiving data request messages. Specifically, federated learning clients such as UE and RAN can generate relevant data based on their own logic or other service requests.

[0198] In this embodiment, the proxy node sends the aforementioned data request message to the federated learning client to request relevant data from the federated learning client. In this way, the federated learning client can provide relevant data based on the aforementioned data request message, which helps to ensure that the data provided by the federated learning client better meets the needs of the proxy node.

[0199] Optionally, the method further includes:

[0200] The proxy node sends a third request message to the first network element. The third request message is used to request information about the federated learning client, and the third request message includes at least one of the following:

[0201] The first instruction information is used to indicate that the obtained federated learning client must support federated learning by proxy nodes;

[0202] The first capability information is used to indicate the federated learning capabilities that the acquired federated learning client needs to support.

[0203] Information about the proxy node.

[0204] In this embodiment, the first network element can be an AMF, NRF, or UDM, etc.

[0205] The obtained federated learning client must support federated learning performed by proxy nodes; in other words, the obtained federated learning client must be willing to have federated learning performed by proxy nodes. This proxy-based federated learning can include the proxy node relaying federated learning-related messages or the proxy node performing the federated learning process on behalf of the client.

[0206] The information of the aforementioned proxy nodes may include the device information of the aforementioned proxy nodes, such as vendor ID or provider ID, so that the first network element can select the corresponding UE and / or RAN when selecting federated learning clients such as UE and / or RAN.

[0207] For example, the proxy node can send a third request message to the first network element through messages such as Namf_EventExposure_Subscribe. This third request message can also carry area information, indication information for requesting access capabilities, such as the Area of ​​Interest (AoI), requesting the first network element to provide feedback on the information of the federated learning client within the AoI.

[0208] In some optional embodiments, the proxy node may send a third request message to the first network element before sending a data request message to at least one second federated learning client, in order to request information on federated learning clients that meet the conditions indicated by the third request message. This can reduce the occurrence of the proxy node requesting federated learning-related data to some unsuitable federated learning clients.

[0209] Optionally, the method further includes: the proxy node receiving a first response message from the first network element, the first response message including information about at least one federated learning client.

[0210] Optionally, the information of the federated learning client includes at least one of the following: the identification information of the federated learning client, and the capability information of the federated learning client;

[0211] The capability information of the federated learning client includes at least one of the following:

[0212] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0213] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0214] Information about the agent nodes supported by the federated learning client;

[0215] The identifier of the federated learning tasks supported by the federated learning client.

[0216] The identification information of the aforementioned federated learning client, such as the ID of the aforementioned federated learning client.

[0217] The proxy nodes supported by the aforementioned federated learning client can be understood as the client supporting federated learning through these proxy nodes, or the proxy nodes acting as proxies for the federated learning client in performing federated learning. The information of the proxy nodes supported by the aforementioned federated learning client may include device information such as vendor ID or provider ID.

[0218] Optionally, the federated learning client supports federated learning capabilities including at least one of the following:

[0219] The federated learning client has the ability to participate in federated learning;

[0220] The federated learning client supports federated learning through proxy nodes.

[0221] The aforementioned federated learning client has the ability to participate in federated learning, which can be understood as the aforementioned federated learning client having the ability to participate in federated learning related processes (e.g., federated learning accuracy, federated learning training, federated learning inference, etc.), or the aforementioned federated learning client having the ability of a federated learning client, the ability of a horizontal federated learning client, or the ability of a vertical federated learning client, etc.

[0222] The aforementioned federated learning client supports federated learning by proxy nodes, which may include the aforementioned federated learning client supporting the relay of messages by proxy nodes or the proxy nodes performing federated learning on behalf of the federated learning client.

[0223] In this embodiment, the agent node sends a third request message to the first network element to request information on a federated learning client that meets the conditions indicated by the third request message, and receives at least one of the identification information and capability information of at least one federated learning client. This facilitates the agent node in selecting a suitable federated learning client for interaction.

[0224] Optionally, the method further includes:

[0225] The proxy node sends a fourth request message to the second network element. The fourth request message includes the identification information of at least one federated learning client and is used to request the capability information of the at least one federated learning client.

[0226] The proxy node receives a second response message from the second network element, the second response message including capability information of the at least one federated learning client.

[0227] In this embodiment, the agent node can obtain the capability information of the federated learning client from the second network element, which makes it easier for the agent node to select a suitable federated learning client for interaction based on the capability information of the federated learning client.

[0228] In some optional embodiments, after receiving a first response message that includes the identification information of at least one federated learning client (i.e., the first response message does not include the capability information of the federated learning client), the proxy node may send a fourth request message to the second network element. This fourth request message includes the identification information of the at least one federated learning client to request the capability information of the at least one federated learning client. For example, after receiving the identification information of the federated learning client from the AMF, the proxy node sends a capability request message to the UDM, which may carry the identification information of the federated learning client. Upon receiving the capability request message from the proxy node, the UDM obtains the capability information of the federated learning client based on its identification information and then returns the capability information.

[0229] It should be noted that the capability information of the aforementioned federated learning client can be found in the relevant descriptions of the foregoing embodiments, and will not be repeated here.

[0230] Optionally, after sending the second request message to at least one first federated learning client, the method further includes:

[0231] The proxy node receives at least one third response message from the at least one first federated learning client;

[0232] The proxy node sends a fourth response message to the federated learning server, the fourth response message being determined based on the at least one third response message.

[0233] The aforementioned third response message may be a response message to the aforementioned second request message. For example, if the aforementioned second request message is a preparation request message, the aforementioned third response message may include at least one of the following:

[0234] Instructions regarding participation in the federated learning process;

[0235] Instructions on whether to accept the request for federal learning preparation;

[0236] The sample identification information of the Federated Learning Client is used to indicate the sample identification information that the Federated Learning Client has, can obtain, or can be used in the Federated Learning related process. It can be in the form of a list, that is, there can be multiple sample identification information.

[0237] The intersection of sample identifiers is used to indicate the intersection of sample identifiers possessed by the Federated Learning Client and sample identifiers from the Federated Learning Server.

[0238] The characteristic information of the Federated Learning Client is used to indicate the sample information that the Federated Learning Client has or can obtain.

[0239] In the case that the second request message is a training request message, the third response message may include intermediate results, wherein the intermediate results may be intermediate results generated by the federated learning client for federated learning training.

[0240] For example, the fourth response message may include at least one of the third response messages, or the fourth response message may be a newly generated response message based on the message content of at least one third response message.

[0241] In this embodiment, the proxy node receives at least one third response message from the at least one first federated learning client and sends a fourth response message to the federated learning server. The fourth response message is determined based on the at least one third response message, which ensures the execution of the federated learning process.

[0242] In some alternative embodiments, the proxy node may also participate in federated learning as a federated learning client. In this case, the fourth response message is determined based on at least one third response message and the response message generated by the proxy node.

[0243] Optionally, the method further includes:

[0244] The proxy node aggregates the message content of the at least one third response message to obtain the fourth response message.

[0245] In this embodiment, the proxy node aggregates the message content of the at least one third response message to obtain a fourth response message and then sends it to the federated learning server. This not only helps to save interaction resource overhead, but also makes it easier for the federated learning server to process the message content of at least one federated learning client single feedback.

[0246] In some optional embodiments, the proxy node can also participate in federated learning as a federated learning client. In this case, the proxy node can aggregate the message content of the at least one third response message and the message content of the response message generated by the proxy node to obtain a fourth response message.

[0247] Optionally, the proxy node aggregates the message content of the at least one third response message to obtain the fourth response message, including:

[0248] The proxy node concatenates the message contents of the at least one third response message to obtain the fourth response message;

[0249] or,

[0250] The proxy node obtains the fourth response message based on the message content of the at least one third response message and the preset model.

[0251] In some implementations, the proxy node can concatenate the message content of at least one third response message to obtain a fourth response message. For example, if there are three third response messages, each containing intermediate results, and each intermediate result is a 5*1 matrix, then the three intermediate results can be concatenated into a 15*1 matrix. This implementation achieves message content aggregation through concatenation, which is relatively simple. For another example, when the message content of a third response message includes instructions on whether to participate in federated learning processes or whether to accept a federated learning preparation request, the proxy node can merge the instructions included in at least one third response message into a list, thereby informing the federated learning server of the instructions from each federated learning client at once. Alternatively, when the message content of a third response message includes sample data or feature data, the proxy node can aggregate the sample data or feature data included in at least one third response message and send it to the federated learning server.

[0252] In other implementations, the proxy node obtains a fourth response message based on the message content of the at least one third response message and a preset model. For example, the message content of at least one third response message is input into the preset model, and the output of the preset model can be an aggregation result of the message content of the at least one third response message. The fourth response message may include this aggregation result. The preset model can be a pre-trained model for message content aggregation, and can also be referred to as an intermediate model. This implementation achieves message content aggregation through a preset mode, which can improve the flexibility of message content aggregation.

[0253] In some optional embodiments, the proxy node can also participate in federated learning as a federated learning client. In this case, the proxy node can concatenate the message content of the at least one third response message and the message content of the response message generated by the proxy node to obtain a fourth response message; or, the proxy node can obtain a fourth response message based on the message content of the at least one third response message, the message content of the response message generated by the proxy node, and a preset model.

[0254] Optionally, the third response message includes an intermediate result, which is an intermediate result generated by the at least one first federated learning client performing the federated learning-related operations.

[0255] For example, when the first request message is a training request message, the intermediate result can be an intermediate result generated by performing federated learning training. For instance, upon receiving a training request message, the federated learning client (e.g., RAN and / or UE, etc.) can perform local model training. This can be done using data corresponding to the sample identifiers and / or feature information specified in the training request message, generating corresponding intermediate results. The intermediate result can refer to the output result of the local training; for example, local training can be based on ML model training, and the intermediate result is the model output data / model output result after the data corresponding to the sample identifiers and / or feature information is input as input data to the trained model.

[0256] In some optional embodiments, the third response message may also include at least one of the following: federated learning correlation ID, data identification information corresponding to intermediate results, etc.

[0257] Optionally, the method further includes:

[0258] The proxy node receives intermediate information from the federated learning server, the intermediate information including at least one of the following: information for model updates, information for model training;

[0259] The proxy node determines at least one sub-intermediate information based on the intermediate information, and the at least one sub-intermediate information corresponds to the at least one first federated learning client.

[0260] In this embodiment, the information used for model updating may include, but is not limited to, at least one of loss information and gradient information. For example, upon receiving the fourth response message, the federated learning server can generate loss information and / or gradient information based on the intermediate results carried in the fourth response message, the intermediate results generated by the federated learning server itself, and label data.

[0261] The information used for model training may include at least one of the following: instruction information for model training, sample information, feature information, and federated learning correlation ID. For example, the information used for model training may be information from the next training iteration or the next round of model training. For instance, the intermediate information may carry information generated during the nth training iteration for model updates and information from the (n+1)th training iteration, where n is a positive integer.

[0262] The aforementioned intermediate information can also be referred to as intermediate model training information.

[0263] Specifically, the proxy node can determine at least one sub-intermediate information based on the received intermediate information. This sub-intermediate information corresponds to at least one first federated learning client, and the proxy node can send its corresponding sub-intermediate information to each first federated learning client. Each first federated learning client can then update its model, perform model training, etc., based on the received sub-intermediate information. It should be noted that different sub-intermediate information can be different, or they can be the same.

[0264] In some optional embodiments, the proxy node can also participate in federated learning as a federated learning client. In this case, the proxy node can determine at least two sub-intermediate information based on the received intermediate information. These at least two sub-intermediate information correspond to at least one first federated learning client and the proxy node. For example, if the number of first federated learning clients is M, the proxy node can determine M+1 sub-intermediate information based on the received intermediate information. These M+1 sub-intermediate information correspond to M first federated learning clients and the proxy node, respectively.

[0265] Optionally, the proxy node determines at least one piece of sub-intermediate information based on the intermediate information, including:

[0266] The proxy node splits the intermediate information into at least one sub-intermediate information;

[0267] or,

[0268] The proxy node obtains at least one piece of sub-intermediate information based on the intermediate information and the preset model.

[0269] In some implementations, the proxy node can split the intermediate information into at least one sub-intermediate information. For example, an intermediate information that is a 15*1 matrix can be split into three 5*1 matrices, which is a relatively simple implementation.

[0270] In other implementations, the proxy node can obtain at least one piece of sub-intermediate information based on intermediate information and a preset model. For example, the intermediate information can be input back into the preset model, and at least one piece of sub-intermediate information can be obtained through backpropagation. It is understood that information aggregation based on the preset model and information splitting based on the preset model are two opposite processes. Therefore, at least one piece of sub-intermediate information can be obtained based on the preset model and intermediate information through backpropagation.

[0271] Optionally, the proxy node is a Network Open Function (NEF), and before the proxy node receives the first request message from the federated learning server, the method further includes:

[0272] The NEF receives a lookup request message from the federated learning server, the lookup request message being used to request a lookup of the federated learning client;

[0273] The NEF sends feedback information to the federated learning server, and the feedback information includes at least one of the following:

[0274] Information about at least one federated learning client;

[0275] The second indication information is used to indicate that the NEF is a proxy node;

[0276] The information from NEF.

[0277] The information of at least one federated learning client may include at least one of the following: identification information and capability information of at least one federated learning client. The capability information of the federated learning client can be found in the relevant descriptions of the foregoing embodiments, and will not be repeated here.

[0278] In some optional embodiments, NEF can anonymize the information of at least one federated learning client and then return the anonymized identification information (e.g., a temporary identifier) ​​of the federated learning client to the federated learning server to ensure privacy and security.

[0279] The aforementioned NEF information may include at least one of the following: NEF capability information, supported service information, and service scope information. Optionally, the aforementioned NEF information may include NEF information related to agent-based federated learning, such as NEF-supported capabilities related to agent-based federated learning and NEF-supported services related to agent-based federated learning.

[0280] In this embodiment, when the proxy node is NEF, in addition to reporting the information of the federated learning client it has found, NEF can also report information related to proxy federated learning. This makes it easier for the federated learning server to determine whether to select REF as the proxy node based on the information of NEF related to proxy federated learning.

[0281] Please refer to Figure 5, which is a flowchart of an information interaction method provided in an embodiment of this application. This method can be executed by a federated learning client, and as shown in Figure 5, it includes the following steps:

[0282] Step 501: The Federated Learning Client performs the second operation;

[0283] The second operation includes at least one of the following:

[0284] Receive a second request message from the proxy node, the second request message being used to request the execution of federated learning related operations;

[0285] Receive a data request message sent by the proxy node, the data request message being used to request data related to the federated learning client;

[0286] Send the data from the federated learning client to the proxy node;

[0287] A first registration request message is sent to the first network element. The first registration request message includes the capability information of the federated learning client, and the capability information of the federated learning client includes at least one of the following:

[0288] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0289] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0290] Information about the agent nodes supported by the federated learning client;

[0291] The identifier of the federated learning task supported by the federated learning client;

[0292] The agent node is a device that supports agent federated learning.

[0293] For example, when a federated learning client connects to a network, in addition to registering its network connection and some basic information, it can also register its capability information. After receiving the capability information from the federated learning client, the first network element can send the capability information of the federated learning client to the second network element, requesting the second network element to store it.

[0294] In some optional embodiments, the first network element may send a response message to the federated learning client, which indicates whether the federated learning client's capability information registration was successful or failed.

[0295] Optionally, the federated learning client supports federated learning capabilities including at least one of the following:

[0296] The federated learning client has the ability to participate in federated learning;

[0297] The federated learning client supports federated learning through proxy nodes.

[0298] Optionally, the agent federated learning includes at least one of the following:

[0299] Relay messages from federated learning nodes;

[0300] It replaces federated learning nodes for federated learning.

[0301] It should be noted that the implementation method of this method can be found in the relevant description of the embodiment shown in Figure 4, and will not be repeated here.

[0302] Please refer to Figure 6, which is a flowchart of an information interaction method provided in an embodiment of this application. The method can be executed by a first network element, and as shown in Figure 6, it includes the following steps:

[0303] Step 601: The first network element receives the first registration request message from the federated learning client;

[0304] The first registration request message includes the capability information of the federated learning client, which includes at least one of the following:

[0305] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0306] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0307] The information of the proxy nodes supported by the federated learning client, wherein the proxy nodes are devices that support proxy federated learning;

[0308] The identifier of the federated learning tasks supported by the federated learning client.

[0309] It should be noted that the message content of the second registration request message is the same as the message content of the first registration request message, or the message content of the second registration request message may include a portion of the message content of the first registration request message. It should also be noted that the second registration request message may be used to request that information in the registration message be stored at the second network element.

[0310] Optionally, the agent federated learning includes at least one of the following:

[0311] Relay messages from federated learning nodes;

[0312] It replaces federated learning nodes for federated learning.

[0313] Optionally, the method further includes:

[0314] The first network element receives a third request message from the proxy node. The third request message is used to request information about the federated learning client, and the third request message further includes at least one of the following:

[0315] The first instruction information is used to indicate that the obtained federated learning client must support federated learning by proxy nodes;

[0316] The first capability information is used to indicate the federated learning capabilities that the acquired federated learning client needs to support.

[0317] Information about the proxy node.

[0318] Optionally, the method further includes:

[0319] The first network element sends a second registration request message to the second network element;

[0320] The second registration request message is determined based on the first registration request message, and the second registration request message includes the capability information of the federated learning client.

[0321] Optionally, the method further includes:

[0322] The first network element sends a first response message to the agent node, the first response message including information about at least one federated learning client.

[0323] Optionally, the information of the federated learning client includes at least one of the following: the identifier of the federated learning client, and the capability information of the federated learning client;

[0324] The capability information of the federated learning client includes at least one of the following:

[0325] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0326] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0327] Information about the agent nodes supported by the federated learning client;

[0328] The identifier of the federated learning tasks supported by the federated learning client.

[0329] Optionally, the federated learning client supports federated learning capabilities including at least one of the following:

[0330] The federated learning client has the ability to participate in federated learning;

[0331] The federated learning client supports federated learning through proxy nodes.

[0332] It should be noted that the implementation method of this method can be found in the relevant descriptions of the embodiments shown in Figures 4 and 5, and will not be repeated here.

[0333] Please refer to Figure 7, which is a flowchart of an information interaction method provided in an embodiment of this application. The method can be executed by a first network element, and as shown in Figure 7, it includes the following steps:

[0334] Step 701: The second network element receives the second registration request message from the first network element;

[0335] The second registration request message includes the capability information of the federated learning client, which includes at least one of the following:

[0336] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0337] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0338] Information about the agent nodes supported by the federated learning client;

[0339] The identifier of the federated learning tasks supported by the federated learning client.

[0340] Optionally, the federated learning client supports federated learning capabilities including at least one of the following:

[0341] The federated learning client has the ability to participate in federated learning;

[0342] The federated learning client supports federated learning through proxy nodes.

[0343] Optionally, the agent federated learning includes at least one of the following:

[0344] Relay messages from federated learning nodes;

[0345] It replaces federated learning nodes for federated learning.

[0346] Optionally, the method further includes:

[0347] The second network element receives a fourth request message from the agent node. The fourth request message includes identification information of at least one federated learning client and is used to request the capability information of the at least one federated learning client.

[0348] The second network element sends a second response message to the agent node, containing the capability information of the at least one federated learning client.

[0349] It should be noted that the implementation method of this embodiment can be referred to in the relevant descriptions of the embodiments shown in Figures 4, 5 and 6, which will not be repeated here.

[0350] The following uses federated learning as an example (VFL) to illustrate the implementation of this application:

[0351] Example 1: The proxy node is used to relay messages from the federated learning node.

[0352] For example, referring to Figure 8, the information interaction method provided in this application embodiment includes the following steps:

[0353] Step 800: The VFL Server decides to proceed with the VFL-related process.

[0354] For example, the VFL Server determines the relevant VFL processes based on internal logic, pre-configuration, or requests from other devices.

[0355] Step 801: The VFL Server sends a VFL Client lookup request message to devices such as the NEF. This VFL Client lookup request message can include selection criteria, requesting the search / selection of VFL Clients that meet the criteria. Specifically, the VFL can send the NEF with its VFL type identifier (e.g., analytics ID, VFL task ID, etc.), the required device type (e.g., UE, RAN, NWDAF, etc.), VFL capability information, and required feature information, etc., as selection criteria or conditions. For example, the aforementioned VFL Client lookup request message can be an Nnef_NFDiscovery_Request message.

[0356] It should be noted that when the VFL Server is a trusted AF, the AF can send a lookup request message to the NRF, which will then search for / select a VFL Client that meets the criteria. In this case, interaction can be performed through messages such as Nnrf_NFDiscovery_Request.

[0357] Step 802: NEF discovers VFL Clients that meet the requirements. For example, NEF can determine the information of VFL Clients that meet the conditions (i.e., candidate VFL Clients) by interacting with NRF, or it can determine the VFL Clients that meet the conditions based on internal data, internal configuration, etc.

[0358] It should be noted that the proxy node can also be a NEF, that is, by enhancing the NEF, the NEF can directly interact with the RAN and UE to complete the VFL-related processes. In this case, step 802 does not need to perform operations such as VFL Client lookup.

[0359] It should also be noted that when the VFL Server is a trusted AF and can interact directly with the NRF, the NRF can determine the information of the VFL Client that meets the conditions.

[0360] Step 803: NEF sends information about the VFL Clients that meet the criteria to the VFL Server, i.e., the VFLClient response. It should be noted that NEF can anonymize the information of the VFL Clients that meet the criteria and send back the anonymized identification information of the VFL Clients, such as a temporary ID, thereby protecting device information and topology information within the network. Specifically, NEF can send back VFL Client information, such as the VFL Client's identification information and capability information.

[0361] It should be noted that when NEF acts as a proxy node, the information returned by NEF may also include at least one of the following:

[0362] A message indicating that the search is complete;

[0363] NEF supports indication information as a proxy node, which is used to indicate that AF can interact with the NEF to complete VFL-related processes;

[0364] The information about NEF may include at least one of the following: NEF capability information, NEF supported service information, and service scope information.

[0365] It should be noted that when the VFL Server interacts directly with the NRF, the NRF provides feedback to the VFL Client.

[0366] Steps 804-805: The VFL Server sends a VFL preparation request message to the candidate VFL Client to perform sample alignment, feature alignment, etc.

[0367] Specifically, the VFL Server can send a preparation request message via messages such as Nnef_VFLTrainingRequest_Request or Nnwdaf_VFLTrainingRequest_Request. The details of this preparation request message can be found in the descriptions of the foregoing embodiments, and will not be repeated here.

[0368] It should be noted that the above candidate VFL Clients may include proxy nodes, that is, proxy nodes can be regarded as proxy VFL Clients.

[0369] It should also be noted that the VFL preparation request message can be sent directly from the VFL Server to the agent node, or it can be forwarded to the agent node through devices such as NEF.

[0370] Step 806: The agent node sends a VFL Preparation request to VFL Clients such as the RAN and UE. Specifically, after receiving the VFL Preparation request from the VFL Server, the agent node can send a preparation request to one or more VFL Clients, which may carry information similar to that in step 805.

[0371] For example, step 806 above may include steps 806a and 806b, wherein steps 806a and 806b may respectively represent sending accurate request messages to the UE and the RAN.

[0372] Step 807: The VFL Client confirms its eligibility to participate in the VFL-related process. This involves verifying whether it meets the VFL's requirements. For example, the VFL Client can compare its own sample identifier information and / or feature information with the sample identifier information and / or feature information included in the VFL's request to determine if they are identical or if there is any overlap. Specifically, it can determine if the VFL Client's own sample identifier information matches the sample identifier information included in the VFL's request, or if there is any overlap between the VFL Client's own sample identifier information and the sample identifier information included in the VFL's request. Similarly, it can determine if the VFL Client's own feature information matches the feature information included in the VFL's request. If there is overlap, the VFL Client can decide to participate in the federated learning process. It should be noted that the absence of overlap does not necessarily mean that the VFL Client cannot participate in the VFL-related process.

[0373] It should also be noted that when determining whether to participate in VFL-related processes, the VFL Client can also consider its own load, workload, etc.

[0374] It should also be noted that proxy nodes can also participate in VFL-related processes as VFL Clients, and therefore can also decide whether to participate in VFL-related processes. Therefore, step 807 above includes steps 807a and 807b.

[0375] Step 808: The VFL Client (e.g., RAN, UE, etc.) sends back a response message for the VFL preparation request, i.e., a preparation response message.

[0376] Specifically, the VFL Client can provide feedback indicating whether it participates in VFL-related processes, and whether it accepts the VFL preparation request. In addition, the VFL Client can also provide feedback on at least one of the following:

[0377] The sample identification information of the VFL Client is used to indicate the sample identification information that the VFL Client has / can obtain / can be used in VFL-related processes. It can be in the form of a list, that is, there can be multiple sample identification information.

[0378] The intersection of sample identifiers, used to indicate the intersection of sample identifiers possessed by the VFL Client and sample identifiers from the VFL Server;

[0379] VFL Client characteristic information, used to indicate the characteristics that the VFL Client has / can obtain.

[0380] For example, step 808 above may include steps 808a and 808b, wherein steps 808a and 808b may respectively represent the UE and RAN sending a ready response message to the proxy node.

[0381] Step 809: The proxy node aggregates messages from at least one VFL Client. After receiving one or more messages from one or more VFL Clients, the proxy node generates a response message based on the message and / or its own decision. This response message is used to feed back the VFL preparation request message from step 805.

[0382] Specifically, after receiving a message from at least one VFL Client, the proxy node can determine the content of the response message based on the message sent by at least one VFL Client, for example:

[0383] Instructions on whether to participate in VFL-related processes. If the VFL Client reports that it will not participate in VFL-related processes, this instruction may also be set to indicate that it will not participate in VFL-related processes.

[0384] This indicates whether to accept the VFL preparation request. If the VFL Clients respond that they do not accept the VFL preparation request, this instruction may also be set to indicate that they do not accept the VFL preparation request.

[0385] The sample identifier information of a VFL Client can be the intersection of sample identifiers possessed by multiple VFL Clients;

[0386] The intersection of sample identifiers can be the intersection of sample identifiers returned by multiple VFL Clients, that is, the intersection is calculated again based on the intersection of sample identifiers returned by multiple VFL Clients;

[0387] The characteristic information of a VFL Client can be the intersection of the characteristic information of multiple VFL Clients.

[0388] Steps 810-811: The proxy node sends a response message to the VFL Server regarding the VFL preparation request. The proxy node can send the response message directly to the VFL Server or forward it through NEF. For example, when sending the response message to the VFL Server, the proxy node can use the Nnef_VFLTrainingRequest_Request Response message or other service messages. The information in this response message is similar to the information aggregated in step 809.

[0389] Steps 812-813: The VFL Server can send a VFL training request message to the VFL Client through the proxy node to begin VFL training. The VFL Server can determine the final sample identifier or target sample identifier (i.e., the sample identifier information used for VFL training), feature information (feature information used for VFL training), etc., based on the response message from the VFL Client, and send them to the VFL Client. It should be noted that the VFL Server can send the VFL training request message directly to the VFL Client, or it can send the VFL training request message to the VFL Client through NEF. For details on the VFL training request message, please refer to the aforementioned explanation of training request messages; it will not be repeated here.

[0390] Step 814: The proxy node sends a VFL training request message to VFL Clients such as the RAN and UE. After receiving the VFL training request message from the VFL Server, the proxy node sends a VFL training request message to VFL Clients such as the RAN and UE. This message may carry information similar to that in Step 813. It should be noted that the proxy node can send VFL training request messages to one or more VFL Clients such as the RAN and UE.

[0391] Step 815: The VFL Client (e.g., RAN, UE, etc.) performs local model training.

[0392] Specifically, the VFL Client can determine to perform local training based on the VFL training request message from the proxy node. This local training can utilize data corresponding to the sample identifiers and / or feature information specified in the VFL training request message, and generate corresponding intermediate results. These intermediate results can refer to the output of the local training; for example, if the local training is based on ML model training, the intermediate results would be the model output data / model output results after the data corresponding to the sample identifiers and / or feature information are input into the model.

[0393] It should be noted that proxy node devices can also participate in VFL-related processes as VFL Clients, meaning that proxy nodes can also perform local model training, etc.

[0394] Step 816: The VFL Client sends intermediate results to the proxy node. After completing local training, the VFL Client sends intermediate results to the proxy node. In addition, it can also send the VFL correlation ID, data identifier information corresponding to the intermediate results, and other information.

[0395] Step 817: Proxy nodes aggregate intermediate results. After receiving one or more messages from one or more VFL Clients, the proxy node generates an intermediate result based on the messages from the one or more VFL Clients and / or its own local training results, and then feeds back the intermediate results of VFL training to the VFL Server. Specifically, the method by which the proxy node aggregates an intermediate result can be found in the relevant descriptions of the foregoing embodiments, and will not be repeated here.

[0396] Steps 818-819: The proxy node sends the intermediate result to the VFL Server. The proxy node can send the intermediate result directly to the VFL Server, or send it to the VFL Server via NEF forwarding. This intermediate result is the result aggregated by the proxy node in step 817.

[0397] Steps 820-821: The VFL Server sends intermediate information (including information for model updates) to the VFL Client through the proxy node. The VFL Server generates intermediate information based on intermediate results from the proxy node, intermediate results generated by the VFL Server itself, and / or label data, such as difference information and / or gradient information. The aforementioned label data can also be referred to as ground truth.

[0398] Specifically, the VFL Server generates the final result (VFL inference result, VFL output result, etc.) based on the intermediate results from the proxy node and / or its own intermediate results, compares it with the label data, and generates a loss value / gradient value, which can be used to update the model.

[0399] It should be noted that the VFL Server can send this intermediate information directly to the VFL Client, or it can send this intermediate information to the VFL Client through NEF.

[0400] Step 822: The proxy node splits the intermediate information into multiple sub-intermediate information. After receiving the intermediate information from the VFL Server, the proxy node splits / divides the intermediate information into multiple sub-intermediate information (i.e., the split intermediate information). The split intermediate information corresponds to each VFL Client, so that each VFL Client updates its model based on the split intermediate information.

[0401] Specifically, proxy nodes can decide how to split intermediate information based on how they aggregate intermediate results.

[0402] Step 823: The agent node sends sub-intermediate information to the VFL Client (e.g., RAN, UE, etc.). After determining the intermediate information corresponding to each VFL Client, the agent node sends the corresponding sub-intermediate information to each VFL Client.

[0403] Step 824: VFL Clients update their local models. After receiving the sub-intermediate information from the proxy node, the VFL Client updates its local model based on that sub-intermediate information.

[0404] It should be noted that steps 820-823 can carry intermediate information from the nth training iteration and VFL training information from the (n+1)th training iteration, as can be seen in steps 812-814. That is, steps 820-823 can be transmitted using similar service messages as steps 812-814.

[0405] It should be noted that after updating the local model, the VFL Client can perform a new model training based on the updated model (generating intermediate results, etc.). Steps 812-824 can then be repeated to complete VFL training after multiple rounds.

[0406] It should also be noted that the VFL inference process is similar to steps 812-819, except that steps 812-814 send a VFL inference request carrying the target identification information for the inference; step 815 performs local inference based on the data corresponding to the target identification information; step 816 feeds back the local inference result as an intermediate result; step 817 aggregates the intermediate results by the proxy node; and step 818 feeds back the aggregated intermediate result as a response message to the VFL Server.

[0407] Example 2: Proxy nodes are used to perform federated learning in place of federated learning clients.

[0408] For example, referring to Figure 9, the information interaction method provided in this application embodiment includes the following steps:

[0409] Step 900: The VFL Server decides to proceed with the VFL-related process.

[0410] Step 901: The proxy node sends a data collection subscribe message (i.e., the aforementioned data request message) to VFL Clients such as the UE and RAN, requesting the acquisition of relevant data from these VFL Clients. For example, the proxy node decides to collect data from devices such as the UE and RAN based on internal logic, pre-configuration, or requests from other devices. The details of the aforementioned data request message can be found in the descriptions of the foregoing embodiments and will not be repeated here.

[0411] For example, step 901 above may include steps 901a and 901b, wherein steps 901a and 901b may respectively represent sending data collection subscription messages to the UE and the RAN.

[0412] Step 902: VFL Clients such as UE and RAN generate data. Before receiving a data collection request message from the agent node, devices such as UE and RAN can generate and store relevant data; after receiving the data collection request message, they can also generate new relevant data. Specifically, VFL Clients such as UE and RAN can generate relevant data based on their own logic or other service requests.

[0413] Step 903: The UE, RAN, and other VFL Clients send a data collection response message to the agent node. This data collection response message includes the data to be acquired. After receiving the data collection request message from the agent node, the UE, RAN, and other VFL Clients can determine the data to be returned based on the feedback condition information.

[0414] For example, step 903 above may include steps 903a and 903b, wherein steps 903a and 903b may respectively represent receiving data collection response messages from the UE and RAN.

[0415] Step 904: The VFL Server sends a VFL Client lookup request message to devices such as NEF. This step is the same as step 801 above and will not be repeated here.

[0416] Step 905: NEF discovers a VFL Client that meets the requirements. This step is the same as step 802 above and will not be repeated here.

[0417] Step 906: NEF sends information about VFL Clients that meet the criteria to the VFL Server. This step is the same as step 803 above and will not be repeated here.

[0418] Steps 907-908: The VFL Server sends a VFL Ready Request message to the agent node via NEF.

[0419] Step 909: The proxy node confirms whether it can participate in the VFL-related process. The specific implementation method is the same as the implementation method of confirming whether the VFL Client can participate in the VFL-related process mentioned above, and will not be repeated here.

[0420] Steps 910-911: The proxy node sends a response message to the VFL preparation request via NEF, i.e., a preparation response message.

[0421] Steps 912-913: The VFL Server sends a VFL training request message to the agent node via NEF to start VFL training.

[0422] Step 914: The proxy node performs local model training.

[0423] In this step, the proxy node can perform local model training based on at least one VFL Client-related data to obtain intermediate results.

[0424] Steps 915-916: The proxy node sends the intermediate results to the VFL Server via NEF.

[0425] Optionally, the VFL Server can generate intermediate information, such as difference information and / or gradient information, based on intermediate results from the proxy node, intermediate results generated by the VFL Server itself, and / or label data, and send the intermediate information to the proxy node through NEF. The proxy node can then update its local model based on the received intermediate information.

[0426] Example 3: UE, RAN, etc. register their capability information related to federated learning.

[0427] For example, referring to Figure 10, the information interaction method provided in this application embodiment includes the following steps:

[0428] Step 1001: UE and RAN register.

[0429] When connecting to the network, the UE and RAN register their network connection and some basic information. In addition, the UE and RAN may send at least one of the following during registration:

[0430] Supported VFL capability information, used to indicate the VFL capabilities it supports;

[0431] VFL Client capability indicates that the device supports participating in VFL-related processes as a VFL Client.

[0432] A VFL with agent involvement is used to indicate that the device supports VFL-related processes with agent involvement.

[0433] Supports sending instructions for VFL-related data, or instructions for consenting to the collection of VFL-related data, used to indicate that the device consents to the collection of its related data for use in VFL-related processes;

[0434] Supported proxy information, which indicates the device information of the proxy nodes that can proxy it, such as vendor ID, provider ID, etc.;

[0435] Supported VFL task IDs.

[0436] Step 1002: After receiving device information and capability information from UE and RAN, AMF can send UE and RAN capability information and intention information to UDM and request UDM to store them.

[0437] Step 1003: The proxy node sends a request message to the AMF, namely the third request message mentioned above, to request equipment information of UEs, RANs, and other devices within the target area. For example, the proxy node can send this request message to the AMF through messages such as Namf_EventExposure_Subscribe. The details of this request message can be found in the relevant content of the third request message, and will not be elaborated upon here.

[0438] Step 1004: The AMF sends device information of UEs, RANs, etc., that meet the conditions to the proxy node. Specifically, based on the request message from the proxy node in step 1003 and the registration messages of UEs and RANs received in step 1002, the AMF determines the UEs, RANs, etc., that meet the proxy node's request and feeds back the information of the UEs, RANs, etc., that meet the conditions. The information of the UEs, RANs, etc., may include at least one of the following: identification information of the UEs, RANs, etc., and capability information of the UEs, RANs, etc.

[0439] Step 1005: If the message fed back by AMF only contains device identification information or does not contain device capability information, steps 1005-1006 can be executed; otherwise, steps 1005-1006 can be skipped.

[0440] The proxy node sends a capability request message to the UDM, requesting to obtain the capability information of the UE and RAN. Specifically, after receiving the identification information of the UE, RAN, and other devices from the AMF, the proxy node sends a capability request message to the UDM. This capability request message may carry the identification information of the UE, RAN, and other devices.

[0441] Step 1006: UDM feeds back capability information of UE, RAN, and other devices. After receiving the capability request message from the proxy node, UDM determines the capability information corresponding to the UE, RAN, and other devices based on the identification information (i.e., device identification information) of the UE, RAN, and other devices, and feeds back the capability information (i.e., device capability information) of the UE, RAN, and other devices. This capability information is similar to the capability information in the registration request message in step 1002.

[0442] Step 1007: The proxy node sends a data collection request message (i.e., the aforementioned data request message) to devices such as the UE and RAN. After receiving device identification information and / or device capability information from the AMF or UDM, the proxy node determines candidate devices (e.g., candidate UE, RAN, etc.). Based on its own logic (e.g., collecting data before the VFL-related process in Example 2) and requests from other devices (e.g., the VFL preparation request message in Example 1), the proxy node determines the target device to interact with. The proxy node sends a data request message to the target device. The details of this data request message can be found in the descriptions of the foregoing embodiments and will not be repeated here.

[0443] It should be noted that this data request message can be sent through the control plane, for example, by sending a data acquisition request to the target device via the AMF; it can also be sent through the user plane, for example, by sending a data acquisition request to the target device via the UPF; or it can be sent through the data plane (also known as the unified data collection platform / network element, etc.), for example, by sending a data acquisition request to the target device via a data plane function (DPF). It should also be noted that this step can be used not only to send data acquisition requests but also to send VFL-related process messages (e.g., steps 806, 814, and 823 in Example 1).

[0444] Step 1008: The UE, RAN, and other devices send a data collection response message to the proxy node. This data collection response message may include data that meets the conditions. After receiving the data request message from the proxy node, the UE, RAN, and other devices send back data that meets the conditions.

[0445] It should be noted that, similar to step 1007, this message can also be transmitted through the user plane, control plane, or data plane. Similar to step 1007, this message can also be used to send VFL-related process messages (e.g., steps 808, 816, etc. in Example 1).

[0446] It should be noted that the information interaction method provided in this application embodiment can be executed by an information interaction device. This application embodiment uses an information interaction device executing the information interaction method as an example to illustrate the information interaction device provided in this application embodiment.

[0447] This application provides an information interaction device. As an example, the information interaction device may be a communication device or a component within a communication device, such as a chip. The communication device may be a terminal, a network-side device, or a server, etc. Exemplarily, the terminal may include, but is not limited to, the type of terminal 11 listed above, and the network-side device may include, but is not limited to, the type of network-side device 12 listed above. This application does not impose specific limitations.

[0448] The information interaction device includes a receiving module, a transmitting module, and a processing module. These modules can be implemented in software or hardware. When implemented in hardware, the processing module can be implemented by a processor. For example, the processor can include general-purpose processors, special-purpose processors, etc., such as central processing units (CPUs), microprocessors, digital signal processors (DSPs), artificial intelligence (AI) processors, graphics processing units (GPUs), application-specific integrated circuits (ASICs), network processors (NPs), field-programmable gate arrays (FPGAs), or other programmable logic devices, gate circuits, transistors, discrete hardware components, etc. The receiving and transmitting modules can be implemented by a communication interface, which can include one or more of the following: transceivers, pins, circuits, buses, radio frequency units, etc.

[0449] Referring to Figure 11, when the information interaction device is a network-side device or a component of a network-side device, the information interaction device 1100 includes a receiving module 1101, which is used to receive a first request message from the federated learning server, wherein the proxy node is a device that supports proxy federated learning, and the first request message is used to request the execution of federated learning-related operations.

[0450] Processing module 1102 is configured to perform a first operation, the first operation including at least one of the following:

[0451] Send a second request message to at least one first federated learning client, the second request message being used to request the execution of the federated learning-related operations;

[0452] Perform the federated learning-related operations according to the first request message.

[0453] Optionally, the agent federated learning includes at least one of the following:

[0454] Relay messages from federated learning nodes;

[0455] It replaces federated learning nodes for federated learning.

[0456] Optionally, the processing module 1102 is specifically used for:

[0457] Receive data related to at least one second federated learning client;

[0458] The federated learning-related operations are performed based on the first request message and the data associated with the at least one second federated learning client.

[0459] Optionally, the device further includes:

[0460] The sending module is used to send a data request message to the at least one second federated learning client, the data request message being used to request data related to the second federated learning client.

[0461] Optionally, the data request message includes at least one of the following:

[0462] Federal Learning Task Identifier;

[0463] Types of Federated Learning Tasks;

[0464] Data constraint information is used to indicate the conditions that the required feedback data must meet;

[0465] Feedback condition information indicates the triggering conditions for data feedback.

[0466] Optionally, the device further includes:

[0467] The sending module is configured to send a third request message to the first network element, the third request message being used to request information about the federated learning client, and the third request message including at least one of the following:

[0468] The first instruction information is used to indicate that the obtained federated learning client must support federated learning by proxy nodes;

[0469] The first capability information is used to indicate the federated learning capabilities that the acquired federated learning client needs to support.

[0470] Information about the proxy node.

[0471] Optionally, the receiving module is further configured to:

[0472] Receive a first response message from the first network element, the first response message including information about at least one federated learning client.

[0473] Optionally, the information of the federated learning client includes at least one of the following: the identification information of the federated learning client, and the capability information of the federated learning client;

[0474] The capability information of the federated learning client includes at least one of the following:

[0475] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0476] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0477] Information about the agent nodes supported by the federated learning client;

[0478] The identifier of the federated learning tasks supported by the federated learning client.

[0479] Optionally, the federated learning client supports federated learning capabilities including at least one of the following:

[0480] The federated learning client has the ability to participate in federated learning;

[0481] The federated learning client supports federated learning through proxy nodes.

[0482] Optionally, the device further includes:

[0483] The sending module is used to send a fourth request message to the second network element. The fourth request message includes identification information of at least one federated learning client and is used to request the acquisition of capability information of the at least one federated learning client.

[0484] The receiving module is further configured to receive a second response message from the second network element, the second response message including capability information of the at least one federated learning client.

[0485] Optionally, the receiving module is further configured to receive at least one third response message from the at least one first federated learning client;

[0486] The device further includes a sending module for sending a fourth response message to the federated learning server, the fourth response message being determined based on the at least one third response message.

[0487] Optionally, the processing module is further configured to:

[0488] The message content of the at least one third response message is aggregated to obtain the fourth response message.

[0489] Optionally, the processing module is specifically used for:

[0490] The message contents of the at least one third response message are concatenated to obtain the fourth response message;

[0491] or,

[0492] The fourth response message is obtained based on the message content of the at least one third response message and the preset model.

[0493] Optionally, the third response message includes an intermediate result, which is an intermediate result generated by the at least one first federated learning client performing the federated learning-related operations.

[0494] Optionally, the receiving module is further configured to receive intermediate information from the federated learning server, the intermediate information including at least one of the following: information for model updates, and information for model training;

[0495] The processing module is further configured to determine at least one sub-intermediate information based on the intermediate information, wherein the at least one sub-intermediate information corresponds to the at least one first federated learning client.

[0496] Optionally, the processing module is specifically used for:

[0497] The intermediate information is split into at least one sub-intermediate information;

[0498] or,

[0499] Based on the intermediate information and the preset model, at least one piece of sub-intermediate information is obtained.

[0500] Optionally, the proxy node is a Network Open Function (NEF).

[0501] The receiving module is further configured to receive a search request message from the federated learning server, the search request message being used to request a search for the federated learning client;

[0502] The device further includes a sending module for sending feedback information to the federated learning server, the feedback information including at least one of the following:

[0503] Information about at least one federated learning client;

[0504] The second indication information is used to indicate that the NEF is a proxy node;

[0505] The information from NEF.

[0506] The information interaction device provided in this application embodiment can implement the various processes implemented in the method embodiment of FIG4 and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0507] Specifically, referring to Figure 12, when the information interaction device is a terminal or a component in a terminal or a network-side device or a component in a network-side device, the information interaction device 1200 includes a processing module 1201 for performing a second operation;

[0508] The second operation includes at least one of the following:

[0509] Receive a second request message from the proxy node, the second request message being used to request the execution of federated learning related operations;

[0510] Receive a data request message sent by the proxy node, the data request message being used to request data related to the federated learning client;

[0511] Send the data from the federated learning client to the proxy node;

[0512] A first registration request message is sent to the first network element. The first registration request message includes the capability information of the federated learning client, and the capability information of the federated learning client includes at least one of the following:

[0513] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0514] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0515] Information about the agent nodes supported by the federated learning client;

[0516] The identifier of the federated learning task supported by the federated learning client;

[0517] The agent node is a device that supports agent federated learning.

[0518] Optionally, the federated learning client supports federated learning capabilities including at least one of the following:

[0519] The federated learning client has the ability to participate in federated learning;

[0520] The federated learning client supports federated learning through proxy nodes.

[0521] Optionally, the agent federated learning includes at least one of the following:

[0522] Relay messages from federated learning nodes;

[0523] It replaces federated learning nodes for federated learning.

[0524] The information interaction device provided in this application embodiment can implement the various processes implemented in the method embodiment of FIG5 and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0525] Referring to Figure 13, when the information interaction device is a network-side device or a component in a network-side device, the information interaction device 1300 includes a receiving module 1301, which is used to receive a first registration request message from the federated learning client.

[0526] The first registration request message includes the capability information of the federated learning client, which includes at least one of the following:

[0527] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0528] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0529] The information of the proxy nodes supported by the federated learning client, wherein the proxy nodes are devices that support proxy federated learning;

[0530] The identifier of the federated learning tasks supported by the federated learning client.

[0531] Optionally, the agent federated learning includes at least one of the following:

[0532] Relay messages from federated learning nodes;

[0533] It replaces federated learning nodes for federated learning.

[0534] Optionally, the device further includes:

[0535] The sending module is used to send a second registration request message to the second network element;

[0536] The second registration request message is determined based on the first registration request message, and the second registration request message includes the capability information of the federated learning client.

[0537] Optionally, the receiving module is further configured to receive a third request message from the proxy node, the third request message being used to request information about the federated learning client, and the third request message further including at least one of the following:

[0538] The first instruction information is used to indicate that the obtained federated learning client must support federated learning by proxy nodes;

[0539] The first capability information is used to indicate the federated learning capabilities that the acquired federated learning client needs to support.

[0540] Information about the proxy node.

[0541] Optionally, the apparatus further includes a sending module for sending a first response message to the proxy node, the first response message including information about at least one federated learning client.

[0542] Optionally, the information of the federated learning client includes at least one of the following: the identifier of the federated learning client, and the capability information of the federated learning client;

[0543] The capability information of the federated learning client includes at least one of the following:

[0544] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0545] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0546] Information about the agent nodes supported by the federated learning client;

[0547] The identifier of the federated learning tasks supported by the federated learning client.

[0548] Optionally, the federated learning client supports federated learning capabilities including at least one of the following:

[0549] The federated learning client has the ability to participate in federated learning;

[0550] The federated learning client supports federated learning through proxy nodes.

[0551] The information interaction device provided in this application embodiment can implement the various processes implemented in the method embodiment of FIG6 and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0552] Referring to Figure 14, when the information interaction device is a network-side device or a component in a network-side device, the information interaction device 1400 includes a receiving module 1401, used to receive a second registration request message from the first network element;

[0553] The second registration request message includes the capability information of the federated learning client, which includes at least one of the following:

[0554] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0555] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0556] Information about the agent nodes supported by the federated learning client;

[0557] The identifier of the federated learning tasks supported by the federated learning client.

[0558] Optionally, the federated learning client supports federated learning capabilities including at least one of the following:

[0559] The federated learning client has the ability to participate in federated learning;

[0560] The federated learning client supports federated learning through proxy nodes.

[0561] Optionally, the agent federated learning includes at least one of the following:

[0562] Relay messages from federated learning nodes;

[0563] It replaces federated learning nodes for federated learning.

[0564] Optionally, the receiving module is further configured to receive a fourth request message from the proxy node, the fourth request message including identification information of at least one federated learning client, for requesting to obtain capability information of the at least one federated learning client;

[0565] The device further includes a sending module for sending a second response message to the proxy node, containing capability information of the at least one federated learning client.

[0566] The information interaction device provided in this application embodiment can implement the various processes implemented in the method embodiment of FIG7 and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0567] This application also provides a network-side device, including a processor and a communication interface. The communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the steps of the method embodiments shown in Figures 4, 5, 6, or 7. This network-side device embodiment corresponds to the above-described network-side device method embodiments. All implementation processes and methods of the above-described method embodiments can be applied to this network-side device embodiment and achieve the same technical effects.

[0568] As shown in Figure 15, this application embodiment also provides a communication device 1500, including a processor 1501 and a memory 1502. The memory 1502 stores programs or instructions that can run on the processor 1501. For example, when the communication device 1500 is a proxy node, when the program or instructions are executed by the processor 1501, they implement the various steps of the above-described proxy node-side information interaction method embodiment and achieve the same technical effect. When the communication device 1500 is a federated learning client, when the program or instructions are executed by the processor 1501, they implement the various steps of the above-described federated learning client-side information interaction method embodiment and achieve the same technical effect. When the communication device 1500 is a first network element, when the program or instructions are executed by the processor 1501, they implement the various steps of the above-described first network element-side information interaction method embodiment and achieve the same technical effect. When the communication device 1500 is a second network element, when the program or instructions are executed by the processor 1501, they implement the various steps of the above-described second network element-side information interaction method embodiment and achieve the same technical effect. To avoid repetition, these will not be described again here.

[0569] Specifically, this application also provides a network-side device. As shown in FIG16, the network-side device 1600 includes a processor 1601, a network interface 1602, and a memory 1603. The network-side device may be the information interaction device shown in FIG11, FIG13, or FIG14. The network interface 1602 is, for example, a Common Public Radio Interface (CPRI).

[0570] The network interface 1602 is used to receive a first request message from the federated learning server, wherein the proxy node is a device that supports proxy federated learning, and the first request message is used to request the execution of federated learning-related operations.

[0571] Processor 1601 is configured to perform a first operation, the first operation including at least one of the following:

[0572] Send a second request message to at least one first federated learning client, the second request message being used to request the execution of the federated learning-related operations;

[0573] Execute the federated learning-related operations according to the first request message;

[0574] or,

[0575] Network interface 1602 is used to receive the first registration request message from the federated learning client;

[0576] The first registration request message includes the capability information of the federated learning client, which includes at least one of the following:

[0577] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0578] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0579] The information of the proxy nodes supported by the federated learning client, wherein the proxy nodes are devices that support proxy federated learning;

[0580] The identifier of the federated learning task supported by the federated learning client;

[0581] or,

[0582] Network interface 1602 is used to receive a second registration request message from the first network element;

[0583] The second registration request message includes the capability information of the federated learning client, which includes at least one of the following:

[0584] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0585] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0586] Information about the agent nodes supported by the federated learning client;

[0587] The identifier of the federated learning tasks supported by the federated learning client.

[0588] In addition, the network-side device 1600 of this application embodiment also includes: a program or instructions stored in the memory 1603 and executable on the processor 1601. The processor 1601 calls the program or instructions in the memory 1603 to execute the methods executed by the modules shown in FIG11, FIG13 or FIG14 and achieve the same technical effect. To avoid repetition, it will not be described in detail here.

[0589] Specifically, this application embodiment also provides a network-side device, which can be the information interaction device shown in FIG12. As shown in FIG17, the network-side device 1700 includes: an antenna 1701, a radio frequency device 1702, a baseband device 1703, a processor 1704, and a memory 1705. The antenna 1701 is connected to the radio frequency device 1702. In the uplink direction, the radio frequency device 1702 receives information through the antenna 1701 and sends the received information to the baseband device 1703 for processing. In the downlink direction, the baseband device 1703 processes the information to be transmitted and sends it to the radio frequency device 1702, which processes the received information and then transmits it through the antenna 1701.

[0590] The method executed by the network-side device in the above embodiments can be implemented in the baseband device 1703, which includes a baseband processor.

[0591] The baseband device 1703 may include at least one baseband board, on which multiple chips are disposed, as shown in FIG17. One of the chips is, for example, a baseband processor, which is connected to the memory 1705 via a bus interface to call the program or instructions in the memory 1705 to execute the network-side device operation shown in the above method embodiment.

[0592] The network-side device may also include a network interface 1706, such as a Common Public Radio Interface (CPRI).

[0593] The processor 1704 is used to perform the second operation;

[0594] The second operation includes at least one of the following:

[0595] Receive a second request message from the proxy node, the second request message being used to request the execution of federated learning related operations;

[0596] Receive a data request message sent by the proxy node, the data request message being used to request data related to the federated learning client;

[0597] Send the data from the federated learning client to the proxy node;

[0598] A first registration request message is sent to the first network element. The first registration request message includes the capability information of the federated learning client, and the capability information of the federated learning client includes at least one of the following:

[0599] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0600] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0601] Information about the agent nodes supported by the federated learning client;

[0602] The identifier of the federated learning task supported by the federated learning client;

[0603] The agent node is a device that supports agent federated learning.

[0604] In addition, the network-side device 1700 of this application embodiment also includes: a program or instructions stored in memory 1705 and executable on processor 1704. The processor 1704 calls the program or instructions in memory 1705 to execute the methods executed by each module shown in FIG12 and achieve the same technical effect. To avoid repetition, it will not be described in detail here.

[0605] This application also provides a terminal, including a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the steps in the method embodiment shown in FIG5. This terminal embodiment corresponds to the above-described terminal-side method embodiment, and all implementation processes and methods of the above-described method embodiments can be applied to this terminal embodiment and can achieve the same technical effect. The terminal can be the information interaction device shown in FIG12. Specifically, FIG18 is a schematic diagram of the hardware structure of a terminal implementing an embodiment of this application.

[0606] The terminal 1800 includes, but is not limited to, at least some of the following components: radio frequency unit 1801, network module 1802, audio output unit 1803, input unit 1804, sensor 1805, display unit 1806, user input unit 1807, interface unit 1808, memory 1809, and processor 1810.

[0607] Those skilled in the art will understand that the terminal 1800 may also include a power supply (such as a battery) for powering various components. The power supply can be logically connected to the processor 1810 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The terminal structure shown in Figure 18 does not constitute a limitation on the terminal. The terminal may include more or fewer components than shown, or combine certain components, or have different component arrangements, which will not be elaborated here.

[0608] It should be understood that, in this embodiment, the input unit 1804 may include a graphics processor 18041 and a microphone 18042. The graphics processor 18041 processes image data of still images or videos obtained by an image capture device (such as a camera) in video capture mode or image capture mode. The display unit 1806 may include a display panel 18061, which may be configured in the form of a liquid crystal display, an organic light-emitting diode, or the like. The user input unit 1807 includes at least one of a touch panel 18071 and other input devices 18072. The touch panel 18071 is also called a touch screen. The touch panel 18071 may include a touch detection device and a touch controller. Other input devices 18072 may include, but are not limited to, physical keyboards, function keys (such as volume control buttons, power buttons, etc.), trackballs, mice, and joysticks, which will not be described in detail here.

[0609] In this embodiment, after receiving downlink data from the network-side device, the radio frequency unit 1801 can transmit it to the processor 1810 for processing; in addition, the radio frequency unit 1801 can send uplink data to the network-side device. Typically, the radio frequency unit 1801 includes, but is not limited to, antennas, amplifiers, transceivers, couplers, low-noise amplifiers, duplexers, etc.

[0610] The memory 1809 can be used to store software programs or instructions, as well as various data. The memory 1809 may primarily include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area may store the operating system, application programs or instructions required for at least one function (such as sound playback, image playback, etc.). Furthermore, the memory 1809 may include volatile memory or non-volatile memory. The non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DRRAM). The memory 1809 in this embodiment includes, but is not limited to, these and any other suitable types of memory.

[0611] Processor 1810 may include one or more processing units; optionally, processor 1810 integrates an application processor and a modem processor, wherein the application processor mainly handles operations involving the operating system, user interface, and applications, and the modem processor mainly handles wireless communication signals, such as a baseband processor. It is understood that the aforementioned modem processor may also not be integrated into processor 1810.

[0612] The processor 1810 is used to perform the second operation;

[0613] The second operation includes at least one of the following:

[0614] Receive a second request message from the proxy node, the second request message being used to request the execution of federated learning related operations;

[0615] Receive a data request message sent by the proxy node, the data request message being used to request data related to the federated learning client;

[0616] Send the data from the federated learning client to the proxy node;

[0617] A first registration request message is sent to the first network element. The first registration request message includes the capability information of the federated learning client, and the capability information of the federated learning client includes at least one of the following:

[0618] Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client;

[0619] The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data.

[0620] Information about the agent nodes supported by the federated learning client;

[0621] The identifier of the federated learning task supported by the federated learning client;

[0622] The agent node is a device that supports agent federated learning.

[0623] It is understood that the implementation process of each implementation method mentioned in this embodiment can refer to the relevant description of the federated learning client-side method embodiment and achieve the same or corresponding technical effects. To avoid repetition, it will not be described again here.

[0624] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described information interaction method embodiments and achieve the same technical effects. To avoid repetition, they will not be described again here.

[0625] The processor mentioned above is either the processor in the terminal described in the above embodiments or the processor in the network-side device. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk. In some examples, the readable storage medium may be a non-transient readable storage medium.

[0626] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described information interaction method embodiments and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0627] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0628] This application also provides a computer program / program product, which is stored in a storage medium and executed by at least one processor to implement the various processes of the above-described information interaction method embodiments, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0629] This application also provides a wireless communication system, including: a proxy node and a federated learning client. The proxy node can be used to execute the steps of the information interaction method described above, and the federated learning client can be used to execute the steps of the information interaction method described above.

[0630] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0631] From the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of computer software products plus necessary general-purpose hardware platforms, and of course, they can also be implemented by hardware. The computer software product is stored in a storage medium (such as ROM, RAM, magnetic disk, optical disk, etc.), and the computer software product includes several instructions to cause the terminal or network-side device to execute the methods described in the various embodiments of this application.

[0632] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other implementations under the guidance of this application without departing from the spirit and scope of the claims. All of these implementations are within the protection scope of this application.

Claims

1. An information exchange method, comprising: The proxy node receives a first request message from the federated learning server, wherein the proxy node is a device that supports proxy federated learning, and the first request message is used to request the execution of federated learning-related operations; The proxy node performs a first operation, which includes at least one of the following: Send a second request message to at least one first federated learning client, the second request message being used to request the execution of the federated learning-related operations; Perform the federated learning-related operations according to the first request message.

2. The method according to claim 1, wherein, The agent-fed learning includes at least one of the following: Relay messages from federated learning nodes; It replaces federated learning nodes for federated learning.

3. The method according to claim 1 or 2, wherein, The step of performing the federated learning-related operations according to the first request message includes: The proxy node receives data related to at least one second federated learning client. The proxy node performs the federated learning-related operations based on the first request message and the data related to the at least one second federated learning client.

4. The method according to claim 3, wherein, The method further includes: The proxy node sends a data request message to the at least one second federated learning client, the data request message being used to request data related to the second federated learning client.

5. The method of claim 4, wherein, The data request message includes at least one of the following: Federal Learning Task Identifier; Types of Federated Learning Tasks; Data constraint information is used to indicate the conditions that the required feedback data must meet; Feedback condition information indicates the triggering conditions for data feedback.

6. The method of any one of claims 1 to 5, wherein, The method further includes: The proxy node sends a third request message to the first network element. The third request message is used to request information about the federated learning client, and the third request message includes at least one of the following: The first instruction information is used to indicate that the obtained federated learning client must support federated learning by proxy nodes; The first capability information is used to indicate the federated learning capabilities that the acquired federated learning client needs to support. Information about the proxy node.

7. The method according to claim 6, wherein, The method further includes: The proxy node receives a first response message from the first network element, the first response message including information about at least one federated learning client.

8. The method of claim 6 or 7, wherein, The information of the federated learning client includes at least one of the following: the identification information of the federated learning client, and the capability information of the federated learning client; The capability information of the federated learning client includes at least one of the following: Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client; The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data. Information about the agent nodes supported by the federated learning client; The identifier of the federated learning tasks supported by the federated learning client.

9. The method of claim 8, wherein, The federated learning client supports at least one of the following federated learning capabilities: The federated learning client has the ability to participate in federated learning; The federated learning client supports federated learning through proxy nodes.

10. The method of any one of claims 6-9, wherein, The method further includes: The proxy node sends a fourth request message to the second network element. The fourth request message includes the identification information of at least one federated learning client and is used to request the capability information of the at least one federated learning client. The proxy node receives a second response message from the second network element, the second response message including capability information of the at least one federated learning client.

11. The method of any one of claims 1 to 10, wherein, After sending the second request message to at least one first federated learning client, the method further includes: The proxy node receives at least one third response message from the at least one first federated learning client; The proxy node sends a fourth response message to the federated learning server, the fourth response message being determined based on the at least one third response message.

12. The method of claim 11, wherein, The method further includes: The proxy node aggregates the message content of the at least one third response message to obtain the fourth response message.

13. The method of claim 12, wherein, The proxy node aggregates the message content of the at least one third response message to obtain the fourth response message, including: The proxy node concatenates the message contents of the at least one third response message to obtain the fourth response message; or, The proxy node obtains the fourth response message based on the message content of the at least one third response message and the preset model.

14. The method of any one of claims 11 to 13, wherein, The third response message includes intermediate results, which are intermediate results generated by the at least one first federated learning client performing the federated learning-related operations.

15. The method of any one of claims 11 to 14, wherein, The method further includes: The proxy node receives intermediate information from the federated learning server, the intermediate information including at least one of the following: information for model updates, information for model training; The proxy node determines at least one sub-intermediate information based on the intermediate information, and the at least one sub-intermediate information corresponds to the at least one first federated learning client.

16. The method of claim 15, wherein, The proxy node determines at least one piece of sub-intermediate information based on the intermediate information, including: The proxy node splits the intermediate information into at least one sub-intermediate information; or, The proxy node obtains at least one piece of sub-intermediate information based on the intermediate information and the preset model.

17. The method of any one of claims 1 to 16, wherein, The proxy node is a Network Open Function (NEF) node. Before the proxy node receives the first request message from the federated learning server, the method further includes: The NEF receives a lookup request message from the federated learning server, the lookup request message being used to request a lookup of the federated learning client; The NEF sends feedback information to the federated learning server, and the feedback information includes at least one of the following: Information about at least one federated learning client; The second indication information is used to indicate that the NEF is a proxy node; The information from NEF.

18. An information exchange method, comprising: The federated learning client performs the second operation; The second operation includes at least one of the following: Receive a second request message from the proxy node, the second request message being used to request the execution of federated learning related operations; Receive a data request message sent by the proxy node, the data request message being used to request data related to the federated learning client; Send the data from the federated learning client to the proxy node; A first registration request message is sent to the first network element. The first registration request message includes the capability information of the federated learning client, and the capability information of the federated learning client includes at least one of the following: Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client; The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data. Information about the agent nodes supported by the federated learning client; The identifier of the federated learning task supported by the federated learning client; The agent node is a device that supports agent federated learning.

19. The method of claim 18, wherein, The federated learning client supports at least one of the following federated learning capabilities: The federated learning client has the ability to participate in federated learning; The federated learning client supports federated learning through proxy nodes.

20. The method of claim 18 or 19, wherein, The agent-fed learning includes at least one of the following: Relay messages from federated learning nodes; It replaces federated learning nodes for federated learning.

21. An information exchange method, comprising: The first network element receives the first registration request message from the federated learning client; The first registration request message includes the capability information of the federated learning client, which includes at least one of the following: Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client; The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data. The information of the proxy nodes supported by the federated learning client, wherein the proxy nodes are devices that support proxy federated learning; The identifier of the federated learning tasks supported by the federated learning client.

22. The method of claim 21, wherein, The agent-fed learning includes at least one of the following: Relay messages from federated learning nodes; It replaces federated learning nodes for federated learning.

23. The method of any one of claims 21-22, wherein, The method further includes: The first network element sends a second registration request message to the second network element; The second registration request message is determined based on the first registration request message, and the second registration request message includes the capability information of the federated learning client.

24. The method of any one of claims 21-23, wherein, The method further includes: The first network element receives a third request message from the proxy node. The third request message is used to request information about the federated learning client, and the third request message further includes at least one of the following: The first instruction information is used to indicate that the obtained federated learning client must support federated learning by proxy nodes; The first capability information is used to indicate the federated learning capabilities that the acquired federated learning client needs to support. Information about the proxy node.

25. The method of claim 24, wherein, The method further includes: The first network element sends a first response message to the agent node, the first response message including information about at least one federated learning client.

26. The method according to claim 24 or 25, wherein, The information of the federated learning client includes at least one of the following: the identifier of the federated learning client, and the capability information of the federated learning client; The capability information of the federated learning client includes at least one of the following: Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client; The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data. Information about the agent nodes supported by the federated learning client; The identifier of the federated learning tasks supported by the federated learning client.

27. The method according to any one of claims 21 to 26, wherein, The federated learning client supports at least one of the following federated learning capabilities: The federated learning client has the ability to participate in federated learning; The federated learning client supports federated learning through proxy nodes.

28. An information exchange method, wherein, include: The second network element receives the second registration request message from the first network element; The second registration request message includes the capability information of the federated learning client, which includes at least one of the following: Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client; The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data. Information about the agent nodes supported by the federated learning client; The identifier of the federated learning tasks supported by the federated learning client.

29. The method according to claim 28, wherein, The federated learning client supports at least one of the following federated learning capabilities: The federated learning client has the ability to participate in federated learning; The federated learning client supports federated learning through proxy nodes.

30. The method according to claim 29, wherein, The agent-fed learning includes at least one of the following: Relay messages from federated learning nodes; It replaces federated learning nodes for federated learning.

31. The method according to any one of claims 28 to 30, wherein, The method further includes: The second network element receives a fourth request message from the agent node. The fourth request message includes identification information of at least one federated learning client and is used to request the capability information of the at least one federated learning client. The second network element sends a second response message to the agent node, containing the capability information of the at least one federated learning client.

32. An information interaction device, applied to a proxy node, comprising: The receiving module is used to receive a first request message from the federated learning server, wherein the proxy node is a device that supports proxy federated learning, and the first request message is used to request the execution of federated learning-related operations. A processing module is configured to perform a first operation, the first operation including at least one of the following: Send a second request message to at least one first federated learning client, the second request message being used to request the execution of the federated learning-related operations; Perform the federated learning-related operations according to the first request message.

33. An information interaction device, comprising: The processing module is used to perform the second operation; The second operation includes at least one of the following: Receive a second request message from the proxy node, the second request message being used to request the execution of federated learning related operations; Receive a data request message sent by the proxy node, the data request message being used to request data related to the federated learning client; Send the data from the federated learning client to the proxy node; A first registration request message is sent to the first network element. The first registration request message includes the capability information of the federated learning client, and the capability information of the federated learning client includes at least one of the following: Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client; The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data. Information about the agent nodes supported by the federated learning client; The identifier of the federated learning task supported by the federated learning client; The agent node is a device that supports agent federated learning.

34. An information interaction device, comprising: The receiving module is used to receive the first registration request message from the federated learning client; The first registration request message includes the capability information of the federated learning client, which includes at least one of the following: Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client; The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data. The information of the proxy nodes supported by the federated learning client, wherein the proxy nodes are devices that support proxy federated learning; The identifier of the federated learning tasks supported by the federated learning client.

35. An information interaction device, comprising: The receiving module is used to receive the second registration request message from the first network element; The second registration request message includes the capability information of the federated learning client, which includes at least one of the following: Federated learning capability information, used to indicate the federated learning capabilities supported by the federated learning client; The third instruction information is used to instruct the federated learning client to support the feedback of federated learning-related data or for the federated learning client to agree to the collection of federated learning-related data. Information about the agent nodes supported by the federated learning client; The identifier of the federated learning tasks supported by the federated learning client.

36. A proxy node comprising a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the information interaction method as claimed in any one of claims 1 to 17.

37. A federated learning client, comprising a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the information interaction method as claimed in any one of claims 18 to 20.

38. A network element comprising a processor and a memory, the memory storing a program or instructions executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the information interaction method as described in any one of claims 21 to 27, or implement the steps of the information interaction method as described in any one of claims 28 to 31.

39. A readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the information interaction method as claimed in any one of claims 1 to 17, or the steps of the information interaction method as claimed in any one of claims 18 to 20, or the steps of the information interaction method as claimed in any one of claims 21 to 27, or the steps of the information interaction method as claimed in any one of claims 28 to 31.

40. A computer program product, said computer program product being executed by at least one processor to implement the steps of the information interaction method as claimed in any one of claims 1 to 17, or to implement the steps of the information interaction method as claimed in any one of claims 18 to 20, or to implement the steps of the information interaction method as claimed in any one of claims 21 to 27, or to implement the steps of the information interaction method as claimed in any one of claims 28 to 31.

Images