WO2024197846A1 - 一种通信的方法、装置、设备、芯片和存储介质 - Google Patents

一种通信的方法、装置、设备、芯片和存储介质 Download PDF

Info

Publication number
WO2024197846A1
WO2024197846A1 PCT/CN2023/085576 CN2023085576W WO2024197846A1 WO 2024197846 A1 WO2024197846 A1 WO 2024197846A1 CN 2023085576 W CN2023085576 W CN 2023085576W WO 2024197846 A1 WO2024197846 A1 WO 2024197846A1
Authority
WO
WIPO (PCT)
Prior art keywords
node
vfl
information
task
network element
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2023/085576
Other languages
English (en)
French (fr)
Inventor
陈景然
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong Oppo Mobile Telecommunications Corp Ltd
Original Assignee
Guangdong Oppo Mobile Telecommunications Corp Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guangdong Oppo Mobile Telecommunications Corp Ltd filed Critical Guangdong Oppo Mobile Telecommunications Corp Ltd
Priority to CN202511499057.5A priority Critical patent/CN121262104A/zh
Priority to EP23929431.7A priority patent/EP4648376A4/en
Priority to JP2025555936A priority patent/JP2026513214A/ja
Priority to KR1020257033412A priority patent/KR20250164227A/ko
Priority to PCT/CN2023/085576 priority patent/WO2024197846A1/zh
Priority to CN202380093174.4A priority patent/CN120642308A/zh
Publication of WO2024197846A1 publication Critical patent/WO2024197846A1/zh
Priority to US19/291,423 priority patent/US20250358200A1/en
Priority to MX2025011026A priority patent/MX2025011026A/es
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/098—Distributed learning, e.g. federated learning
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0803—Configuration setting
    • H04L41/0806—Configuration setting for initial configuration or provisioning, e.g. plug-and-play
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/085—Retrieval of network configuration; Tracking network configuration history
    • H04L41/0853—Retrieval of network configuration; Tracking network configuration history by actively collecting configuration information or by backing up configuration information
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W60/00—Affiliation to network, e.g. registration; Terminating affiliation with the network, e.g. de-registration
    • H04W60/04—Affiliation to network, e.g. registration; Terminating affiliation with the network, e.g. de-registration using triggered events
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W8/00—Network data management
    • H04W8/22—Processing or transfer of terminal data, e.g. status or physical capabilities
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W8/00—Network data management
    • H04W8/22—Processing or transfer of terminal data, e.g. status or physical capabilities
    • H04W8/24—Transfer of terminal data
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/085—Retrieval of network configuration; Tracking network configuration history
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network

Definitions

  • the embodiments of the present application relate to the field of communication technology, and specifically to a communication method, apparatus, device, chip and storage medium.
  • VFL Vertical Federated Learning
  • Embodiments of the present application provide a communication method, apparatus, device, chip, and storage medium.
  • an embodiment of the present application provides a communication method, applied to a first node, the method comprising: receiving first information from a first network element, the first information being used to indicate at least one second node that meets the conditions for executing a VFL task; and determining at least one participating node participating in the VFL task from the at least one second node.
  • an embodiment of the present application provides a communication method applied to a second node, the method comprising: sending capability information of the second node to a first network element, the capability information of the second node being used to determine whether the second node meets the conditions for executing a vertical federated learning VFL task.
  • an embodiment of the present application provides a communication method applied to a first network element, the method comprising: obtaining capability information of at least one second node, the capability information of the second node being used to determine whether the second node meets the conditions for executing a vertical federated learning VFL task.
  • an embodiment of the present application provides a communication device, comprising: a first receiving module, configured to receive first information from a first network element, the first information being used to indicate at least one second node that meets the conditions for performing a vertical federated learning VFL task; a determination module, configured to determine at least one participating node participating in the VFL task from at least one second node.
  • an embodiment of the present application provides a communication device, which includes: a first sending module, configured to send capability information of the device to a first network element, and the capability information of the device is used to determine whether the device meets the conditions for performing a vertical federated learning VFL task.
  • an embodiment of the present application provides a communication device, which includes: an acquisition module, configured to obtain capability information of at least one second node, and the capability information of the second node is used to determine whether the second node meets the conditions for performing a vertical federated learning VFL task.
  • an embodiment of the present application provides a communication device, which includes a memory and a processor; wherein the memory is used to store computer-executable instructions; the processor is connected to the memory and is used to implement the method described in any one of the first to third aspects by executing the computer-executable instructions.
  • an embodiment of the present application provides a chip, which includes: a processor, configured to call and run a computer program from a memory, so that a device equipped with the chip executes the method described in any one of the first to third aspects.
  • an embodiment of the present application provides a computer-readable storage medium storing a computer program, which, when executed by at least one processor, implements the method described in any one of the first to third aspects.
  • a first node may receive first information from a first network element, where the first information is used to indicate at least one second node that meets the conditions for executing a VFL task; further, the first node may determine at least one participating node that participates in the VFL task from the at least one second node. In this way, it is ensured that the first node can find a suitable node to perform the VFL task.
  • FIG1 is a schematic diagram of an application scenario of an embodiment of the present application.
  • FIG2 is a schematic diagram of an example of the VFL training process
  • FIG3 is a schematic diagram of an example of the VFL inference process
  • FIG4 is a schematic diagram of an example of a 5G network architecture
  • FIG5 is a flow chart of a communication method provided in an embodiment of the present application.
  • FIG6 is a flow chart of another communication method provided in an embodiment of the present application.
  • FIG. 7 is a schematic diagram of a possible implementation flow of a communication method provided in an embodiment of the present application.
  • FIG8 is a second schematic diagram of a possible implementation flow of the communication method provided in an embodiment of the present application.
  • FIG9 is a third schematic diagram of a possible implementation flow of the communication method provided in an embodiment of the present application.
  • FIG10 is a first structural diagram of a communication device provided in an embodiment of the present application.
  • FIG11 is a second structural diagram of a communication device provided in an embodiment of the present application.
  • FIG12 is a third structural diagram of a communication device provided in an embodiment of the present application.
  • FIG13 is a schematic structural diagram of a communication device provided in an embodiment of the present application.
  • FIG14 is a schematic structural diagram of a chip according to an embodiment of the present application.
  • FIG15 is a schematic block diagram of a communication system provided in an embodiment of the present application.
  • FIG. 1 is a schematic diagram of an application scenario of an embodiment of the present application.
  • the communication system 100 may include a terminal device 110 and a network device 120.
  • the network device 120 may communicate with the terminal device 110 via an air interface.
  • the terminal device 110 and the network device 120 support multi-service transmission.
  • LTE Long Term Evolution
  • TDD LTE Time Division Duplex
  • UMTS Universal Mobile Telecommunication System
  • IoT Internet of Things
  • NB-IoT Narrow Band Internet of Things
  • eMTC enhanced Machine-Type Communications
  • 5G communication system also called New Radio (NR) communication system
  • NR New Radio
  • the network device 120 may be an access network device that communicates with the terminal device 110.
  • the access network device may provide communication coverage for a specific geographical area, and may communicate with the terminal device 110 (eg, UE) located in the coverage area.
  • the network device 120 can be an evolved base station (Evolutional Node B, eNB or eNodeB) in a Long Term Evolution (LTE) system, or a Next Generation Radio Access Network (NG RAN) device, or a base station (gNB) in an NR system, or a wireless controller in a Cloud Radio Access Network (CRAN), or the network device 120 can be a relay station, an access point, an in-vehicle device, a wearable device, a hub, a switch, a bridge, a router, or a network device in a future evolved Public Land Mobile Network (PLMN), etc.
  • Evolutional Node B, eNB or eNodeB in a Long Term Evolution (LTE) system
  • NG RAN Next Generation Radio Access Network
  • gNB base station
  • CRAN Cloud Radio Access Network
  • PLMN Public Land Mobile Network
  • the terminal device 110 may be any terminal device, including but not limited to a terminal device connected to the network device 120 or other terminal devices by wire or wireless connection.
  • the terminal device 110 may refer to an access terminal, a user equipment (UE), a user unit, a user station, a mobile station, a mobile station, a remote station, a remote terminal, a mobile device, a user terminal, a terminal, a wireless communication device, a user agent, or a user device.
  • UE user equipment
  • the access terminal may be a cellular phone, a cordless phone, a Session Initiation Protocol (SIP) phone, an IoT device, a satellite handheld terminal, a Wireless Local Loop (WLL) station, a Personal Digital Assistant (PDA), a handheld device with wireless communication function, a computing device or other processing device connected to a wireless modem, a vehicle-mounted device, a wearable device, a terminal device in a 5G network, or a terminal device in a future evolution network, etc.
  • SIP Session Initiation Protocol
  • IoT IoT device
  • satellite handheld terminal a Wireless Local Loop (WLL) station
  • PDA Personal Digital Assistant
  • PDA Personal Digital Assistant
  • the terminal device 110 can be used for device to device (Device to Device, D2D) communication.
  • D2D Device to Device
  • the wireless communication system 100 may also include a core network device 130 that communicates with the network device 120, and the core network device 130 may be a 5G core network (5G Core, 5GC) device.
  • a core network device 130 that communicates with the network device 120
  • the core network device 130 may be a 5G core network (5G Core, 5GC) device.
  • FIG1 exemplarily shows a network device, a core network device and two terminal devices.
  • the wireless communication system 100 may include multiple network devices and each network device may include other number of terminal devices within its coverage area, which is not limited in the embodiments of the present application.
  • FIG. 1 is only an example of the system to which the present application is applicable.
  • the method shown in the embodiment of the present application can also be applied to other systems.
  • system and “network” are often used interchangeably in this article.
  • the term “and/or” in this article is only a description of the association relationship of the associated objects, indicating that there can be three relationships.
  • a and/or B can represent: A exists alone, A and B exist at the same time, and B exists alone.
  • the character "/" in this article generally indicates that the associated objects before and after are in an "or” relationship.
  • the "indication" mentioned in the embodiment of the present application can be a direct indication, an indirect indication, or an indication of an association relationship.
  • a indicates B which can mean that A directly indicates B, for example, B can be obtained through A; it can also mean that A indirectly indicates B, for example, A indicates C, B can be obtained through C; it can also mean that A and B have an association relationship.
  • the "correspondence” mentioned in the embodiment of the present application can mean that there is a direct or indirect correspondence relationship between the two, or it can mean that there is an association relationship between the two, or it can mean that the relationship between indicating and being indicated, configuring and being configured, etc.
  • predefined or “predefined rules” mentioned in the embodiments of the present application can be implemented by pre-saving corresponding codes, tables or other methods that can be used to indicate relevant information in devices (for example, including terminal devices and network devices), and the present application does not limit its specific implementation method.
  • predefined may refer to the definition in the protocol.
  • protocol may refer to a standard protocol in the field of communications, such as LTE protocols, NR protocols, and related protocols used in future communication systems, and the present application does not limit this.
  • model training requires more dimensional data from users.
  • user data may be distributed on various nodes such as terminals, base stations, core networks, and third-party OTT (Over the Top) application servers.
  • the model training can be performed by combining different feature data of the same user from various nodes, the training effect of the model will be greatly improved, which is of great significance to the training of the model.
  • multi-domain data sharing of multiple nodes will bring great challenges to data privacy. For this reason, the VFL method can be used.
  • VFL allows artificial intelligence systems to use local data of multiple nodes efficiently and accurately while meeting data privacy, security, and regulatory requirements, thereby breaking data silos and realizing cross-domain multi-node data sharing while ensuring privacy and security.
  • the present application provides a communication method, in which a first node can receive first information from a first network element, and can obtain at least one second node that meets the conditions for executing the VFL task through the first information, so as to ensure that the first node can find a suitable node to perform the VFL task.
  • VFL Vertical Federated Learning
  • FIG2 is a schematic diagram of an example of the VFL training process. As shown in FIG2, node A and node B are nodes in different domains, and there is no original data exchange between node A and node B.
  • the VFL training process may include the following steps:
  • VFL is suitable for situations where the training sample IDs of participants overlap a lot, but the data features overlap less.
  • the samples of the participants need to be aligned, so as to increase the feature dimension of each sample without increasing the sample ID.
  • a UE in a certain area generates different feature data at different nodes of the communication system, where the ID of the UE is the sample ID.
  • the feature data generated by the UE at different nodes needs to be aligned (associated).
  • step 202 may include:
  • the third-party coordinator C can send a public key to node A and node B to encrypt the data to be transmitted.
  • the encryption method can adopt a homomorphic encryption algorithm, for example.
  • the homomorphic encryption algorithm the sum of two samples m1 and m2 is homomorphically encrypted, which is equal to the sum of the homomorphic encryption of m1 and the homomorphic encryption of m2.
  • the homomorphic encryption of sample m multiplied by a constant is equal to the homomorphic encryption of sample m multiplied by the constant.
  • the party with the sample label can be the active party (or the demand party), such as node B in Figure 2.
  • Node A can be the passive party (or the data provider), and the passive party does not have a sample label.
  • node A and node B can use their own local data to perform calculations and obtain the intermediate results of the model. Among them, node A can encrypt the intermediate results and send them to node B, and then node B can calculate the overall output error of the model based on its own label and the model output results (intermediate results) of node A and node B, and encrypt the output error and send it to node A.
  • node A and node B can respectively calculate their own encrypted gradients based on the output error in S2, and send the calculation results to coordinator C after adding a mask.
  • coordinator C After decrypting the gradients sent by nodes A and B, coordinator C can send the decrypted gradients back to nodes A and B. Thus, after removing the gradient masks, nodes A and B can update their respective models according to the obtained gradients.
  • FIG3 is a schematic diagram of an example of a VFL inference process. As shown in FIG3 , the VFL inference process may include:
  • the Coordinator C may send a model inference request to node A and node B respectively.
  • the model inference request may include the ID of the model that node A and node B need to adopt to indicate the model that node A and node B need to adopt.
  • Node A and Node B calculate the model results and encrypt them.
  • node A and node B can perform calculations based on their own data and the locally stored model respectively to obtain an intermediate result of the model, and encrypt the intermediate result.
  • node A and node B send the encrypted intermediate result to coordinator C.
  • the coordinator C aggregates the encrypted intermediate results from each node and decrypts them.
  • the coordinator C can aggregate the encrypted intermediate results of nodes A and B to obtain an encrypted model inference result. Then, the coordinator C can decrypt the model inference result and send the decrypted inference result to the demand side node B.
  • service-oriented architecture in which the core network elements (service providers) can provide specific services and can be called by other network elements (consumers) through a defined application programming interface (API).
  • API application programming interface
  • core network element in the embodiment of the present application can also be called “core network function (Network Function, NF)”.
  • core network function Network Function, NF
  • FIG 4 is a schematic diagram of an example of a 5G network architecture.
  • the network architecture may include UE, access network (Access Network, AN)/radio access network (Radio Access Network, RAN) and core network elements.
  • the core network elements include: user plane function (UPF), data network (DN), session management function (SMF), access and mobility management function (AMF), network slice selection function (NSSF), authentication server function (AUSF), network exposure function (NEF), network storage function (NRF), policy control function (PCF), unified data management (UDM) and application function (AF).
  • UPF user plane function
  • DN data network
  • SMF session management function
  • AMF access and mobility management function
  • NSSF network slice selection function
  • AUSF authentication server function
  • NEF network exposure function
  • NRF network storage function
  • PCF policy control function
  • UDM unified data management
  • AF application function
  • the UE and the base station are connected at the access layer (Access Stratum, AS) for exchanging access layer messages and wireless data transmission.
  • the UE and the AMF are connected at the non-access layer (Non-Access Stratum, NAS) for exchanging NAS messages.
  • the AMF is responsible for managing the mobility of the UE, and the SMF is responsible for managing the sessions of the UE.
  • the AMF is also responsible for forwarding session management related messages between the UE and the SMF.
  • the PCF is responsible for formulating policies related to the mobility management, session management, and billing of the UE.
  • the UPF is connected to the base station and the external data network for data transmission.
  • the 5G network has added a network data analysis function (NWDAF) in the core network.
  • NWDAAF network data analysis function
  • This function can collect data from various network elements and network management systems in the core network and perform big data statistics, analysis or intelligent data analysis to obtain analysis and/or prediction data on the network side, thereby assisting each network element to more effectively control UE access based on the data analysis results.
  • UE can be referred to as terminal equipment, access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile equipment, user terminal, terminal, wireless communication equipment, user agent or user device, etc.
  • the embodiments of the present application do not limit the specific technology and specific device form adopted by the UE.
  • (Radio) Access Network ((R)AN) equipment can provide the function of accessing the communication network for authorized users in a specific area, and can specifically include wireless network equipment in the 3rd Generation Partnership Project (3GPP) network and can also include access points in non-3GPP (Non-3GPP) networks.
  • 3GPP 3rd Generation Partnership Project
  • Non-3GPP non-3GPP
  • AMF Mainly used for access control, mobility management, attachment and detachment functions.
  • AMF can also serve as the anchor point for N1 signaling (i.e. signaling of the N1 interface, referred to as N1 signaling for simplicity) and N2 signaling (i.e. signaling of the N2 interface, referred to as N2 signaling for simplicity), and provide routing for N1/N2 session management (SM) messages for SMF.
  • N1 signaling i.e. signaling of the N1 interface, referred to as N1 signaling for simplicity
  • N2 signaling i.e. signaling of the N2 interface, referred to as N2 signaling for simplicity
  • SM N1/N2 session management
  • AMF can also maintain and manage UE status information.
  • SMF Mainly used for user-plane network element selection, user-plane network element redirection, Internet protocol (IP) address allocation for terminal devices, as well as session establishment, modification and release and QoS control.
  • IP Internet protocol
  • UPF Mainly used for receiving and forwarding user plane data. For example, UPF can receive user plane data from DN and send the user plane data to the terminal device through the AN device. UPF can also receive user plane data from the terminal device through the AN device and forward it to DN.
  • PCF A unified policy framework mainly used to guide network behavior and provide policy rule information for control plane network elements (such as AMF, SMF, etc.).
  • AF Mainly used to provide services to the 3GPP network, such as interacting with the PCF for policy control.
  • UDM Mainly used for UE subscription data management, including storage and management of UE identification, UE access authorization, etc. UDM can also generate 3GPP authentication credentials for UE. UDM can also register and maintain the network elements currently serving the UE.
  • NEF Mainly used to securely open the services and capabilities provided by 3GPP network functions to the outside world.
  • DN Mainly used for operator networks that provide data services to UEs, such as the Internet, third-party service networks, and IP Multimedia Service (IMS) networks.
  • IMS IP Multimedia Service
  • AUSF Mainly used for user authentication, such as security authentication of UE when it accesses the network.
  • NSSF Mainly used to select a set of slice instances for the UE, determine the AMF set and allowed network slice selection assistance information (NSSAI) (NSSAIs) for the UE.
  • NSSAI network slice selection assistance information
  • the network elements can communicate with each other through the interfaces shown in the figure, and some interfaces can be implemented in the form of service-oriented interfaces.
  • the UE and AMF can communicate through the N1 interface.
  • the RAN and AMF can communicate through the N2 interface.
  • the RAN and UPF can communicate through the N3 interface, and the N3 interface can be used to transmit user plane data, etc.
  • the SMF and UPF can communicate through the N4 interface.
  • the UPF and DN can communicate through the N6 interface.
  • the UPF and UPF can communicate through the N9 interface, and the N9 interface can be used to transmit uplink and downlink user data flows between UPFs, etc.
  • the relationship between other interfaces and each network element is shown in Figure 4, and for the sake of brevity, it is not described in detail here.
  • interface name between the network elements in FIG4 is only an example, and the interface name in the specific implementation may be other names, which are not specifically limited in this application.
  • name of the message (or signaling) transmitted between the above network elements is only an example and does not constitute any limitation on the function of the message itself.
  • the first node in the embodiment of the present application represents the node that initiates the VFL task, or in other words, the first node is the initiator (or active party/demand party) of the VFL task, that is, the VFL task in the embodiment of the present application can be initiated by the first node.
  • the second node represents a node that can participate in the VFL task as a non-initiator, or in other words, the second node can be the passive party (or data provider) of the VFL task.
  • the number of second nodes can be one, or it can be multiple (two or more), and the embodiment of the present application is not limited to this.
  • the first node of the embodiment of the present application may be, for example, a terminal device, an access network element, a core network element (core network NF) or an AF;
  • the second node may, for example, include at least one of the following: a terminal device, an access network element, a core network element and an AF;
  • the first network element may, for example, be an NRF.
  • the AF may be a trusted AF (such as an AF owned by the operator), or it may also be an untrusted third-party AF (such as an AF from another manufacturer). Among them, whether the AF is a trusted AF can be confirmed or identified by the core network.
  • the core network can confirm or identify the AF belonging to the operator A as a trusted AF, and can confirm or identify the AF that does not belong to the operator A as an untrusted third-party AF.
  • FIG5 shows a communication method provided by an embodiment of the present application, which may include:
  • a first node receives first information from a first network element, where the first information is used to indicate at least one second node that meets a condition for executing a VFL task.
  • the VFL task is initiated by the first node, so “meeting the conditions for executing the VFL task” can also be understood as meeting the conditions for executing the VFL task initiated by the first node.
  • the first node may initiate the VFL task by sending second information to the first network element.
  • the second information may be used to indicate at least one of the following:
  • the time period for executing the VFL task or in other words, the time period that the first node expects to execute the VFL task, such as 9:00 to 21:00.
  • the computing power required to execute the VFL task such as the minimum computing power required to execute the VFL task initiated by the first node
  • a service area for executing the VFL task or in other words, a service area that the first node expects to execute the VFL task.
  • the first network element may send the first information to the first node.
  • At least one second node that meets the conditions for performing the VFL task is indicated to the first node, and accordingly, the first node can receive the first information from the first network element.
  • the first node can learn at least one second node that meets the conditions for performing the VFL task through the first information, thereby ensuring that the first node can find a suitable node to perform the VFL task.
  • the first information may carry, for example, identification and/or address information of at least one second node that meets the conditions for executing the VFL task.
  • whether the second node meets the conditions for executing the VFL task is determined by the first network element based on the capability information of the second node, or in other words, the first network element can determine whether the second node meets the conditions for executing the VFL task based on the capability information of the second node.
  • the capability information of the second node may be used to indicate at least one of the following:
  • the capability information of the second node may include an indication of whether the second node supports executing the VFL task, to indicate whether the second node supports executing the VFL task or does not support executing the VFL task.
  • the time period during which the second node supports the execution of the VFL task For example, the capability information of the second node may be used to indicate that the second node can execute the VFL task during the period from 9:00 to 21:00.
  • the capability information of the second node may include an indication of whether the second node can serve as the initiator of the VFL task, to indicate whether the second node can serve as the initiator of the VFL task or cannot serve as the initiator of the VFL task.
  • the computing power of the second node such as the highest computing power or computing power range that the second node can provide.
  • Service area of the second node When the second node is an access network element, the service area of the second node can also be understood as the signal coverage area of the access network element.
  • the first network element can compare the capability information of each second node collected or stored by the first network element with the content of the second information, determine that it supports performing the VFL task during the period from 9:00 to 21:00, and the service area includes at least one second node in the first service area, and carry the information of the at least one second node (such as identification and/or address information) in the first information and send it to the first node.
  • the capability information of the second node may be carried in the parameter information of the second node, and sent by the second node to the first network element.
  • the method may further include: the second node sends the parameter information of the second node to the first network element, and accordingly, the first network element may receive and store the parameter information of the second node, or in other words, the first network element may collect or determine the parameter information of the second node.
  • the parameter information of the second node may include at least one of the following: capability information of the second node; an identifier (ID) of the second node; address information of the second node; and an identifier of an application (Application) corresponding to the second node.
  • each second node can send parameter information of the second node to the first network element, and the parameter information of each second node includes at least one item of the above parameter information.
  • each second node may send parameter information of each second node to the first network element before the first node initiates the VFL task, so that the first network element may collect parameter information of each second node before the first node initiates the VFL task. Further, when the first node initiates the VFL task, the first network element may obtain capability information of each second node from the collected parameter information of each second node. Furthermore, the first network element may determine at least one second node that meets the conditions for executing the VFL task based on the capability information of each second node, and carry the information of the at least one second node (such as identification and/or address information) in the first information and send it to the first node.
  • the at least one second node such as identification and/or address information
  • the first node determines at least one participating node participating in the VFL task from at least one second node.
  • the first node may determine at least one participating node participating in the VFL task from at least one second node that meets the condition for executing the VFL task.
  • the first node determines at least one participating node participating in the VFL task from at least one second node that meets the conditions for executing the VFL task, including: for a second node that agrees to participate in the VFL task initiated by the first node, the second node is determined as a participating node participating in the VFL task. That is, if a second node meets the conditions for executing the VFL task and agrees to participate in the VFL task initiated by the first node, the first node can be determined as a participating node participating in the VFL task.
  • the method may further include: the first node sends data requirement information to at least one second node, wherein the data requirement information is used to determine data for performing the VFL task. Accordingly, each of the at least one second node may receive the data requirement information from the first node.
  • the data requirement information may include: the type of data required to perform the VFL task, and/or the effective time of the required data.
  • the type of data may include the format of the data (such as 8-bit quantization or 24-bit quantization), and may also include the type of data content (such as application layer data, data related to device power, data related to device communication performance, etc.).
  • the second node may decide whether to agree to participate in the VFL task initiated by the first node based on local configuration or relevant policies. For example, if the second node has data that meets the corresponding data demand locally, it may agree to participate in the VFL task; if the second node does not have data that meets the corresponding data demand locally, it may refuse to participate in the VFL task. Then, the second node may reply to the first node whether to agree to participate in the VFL task.
  • the second node may carry the decision of whether to agree to participate in the VFL task in the third information and reply to the first node.
  • the method may also include: the second node sends the third information to the first node, and the third information is used to indicate whether the second node agrees to participate in the VFL task initiated by the first node.
  • the first node may receive the third information from the second node. Based on the third information, the first node may know whether the second node agrees to participate in the VFL task initiated by the first node.
  • FIG6 shows another communication method provided by an embodiment of the present application, which may include:
  • a second node sends capability information of the second node to a first network element.
  • the first network element receives the capability information of the second node.
  • the second node may send the capability information of the second node to the first network element.
  • the capability information of the second node may be used to characterize the capability of the second node to perform the VFL task.
  • the first network element may determine whether the second node meets the condition for performing the VFL task. In other words, whether the second node meets the condition for performing the VFL task is determined by the first network element based on the capability information of the second node, or in other words, the capability information of the second node may be used by the first network element to determine whether the second node meets the condition for performing the VFL task.
  • the capability information of the second node may be used to indicate at least one of the following:
  • the capability information of the second node may include an indication of whether the second node supports executing the VFL task, to indicate whether the second node supports executing the VFL task or does not support executing the VFL task.
  • the time period during which the second node supports the execution of the VFL task For example, the capability information of the second node may be used to indicate that the second node can execute the VFL task during the period from 9:00 to 21:00.
  • the capability information of the second node may include an indication of whether the second node can serve as the initiator of the VFL task, to indicate whether the second node can serve as the initiator of the VFL task or cannot serve as the initiator of the VFL task.
  • the computing power of the second node such as the highest computing power or computing power range that the second node can provide.
  • Service area of the second node When the second node is an access network element, the service area of the second node can also be understood as the signal coverage area of the access network element.
  • the capability information of the second node may be carried in the parameter information of the second node, and sent by the second node to the first network element.
  • the method may further include: the second node sends the parameter information of the second node to the first network element, and accordingly, the first network element may receive and store the parameter information of the second node, or in other words, the first network element may collect or determine the parameter information of the second node.
  • the parameter information of the second node may include at least one of the following: capability information of the second node; an identifier (ID) of the second node; address information of the second node; and an identifier of an application (Application) corresponding to the second node.
  • the parameter information of each second node includes at least one item of the above parameter information.
  • the second node is a terminal
  • the parameter information of the terminal may include at least one of the following: capability information of the terminal, identification of the terminal, address information of the terminal (such as IP address), and type of the terminal (such as mobile phone, vehicle, etc.).
  • the capability information of the terminal may be used to indicate at least one of the following: whether the terminal supports the execution of VFL tasks, the time period during which the terminal supports the execution of VFL tasks, whether the terminal can be used as the initiator of the VFL task, and the computing power of the terminal.
  • the second node is an access network element
  • the parameter information of the access network element may include at least one of the following: capability information of the access network element, an identifier of the access network element (such as gNB ID), and address information of the access network element.
  • the capability information of the access network element may be used to indicate at least one of the following: whether the access network element supports the execution of the VFL task, the time period during which the access network element supports the execution of the VFL task, whether the access network element can be used as the initiator of the VFL task, the computing power of the access network element, and the service area of the access network element.
  • the second node is a core network element
  • the parameter information of the core network element may include at least one of the following: capability information of the core network element, address information of the core network element, and type of the core network element (such as SMF, AMF, etc.).
  • the capability information of the core network element may be used to indicate at least one of the following: whether the core network element supports the execution of the VFL task, the time period in which the core network element supports the execution of the VFL task, whether the core network element can be used as the initiator of the VFL task, the computing power of the core network element, and the service area of the core network element.
  • the second node is an AF
  • the parameter information of the AF may include at least one of the following: capability information of the AF, an identifier of the AF, address information of the AF, and an identifier of an application corresponding to the AF.
  • the capability information of the AF may be used to indicate at least one of the following: whether the AF supports the execution of a VFL task, the time period during which the AF supports the execution of a VFL task, whether the AF can be used as the initiator of the VFL task, and the computing power of the AF.
  • the second node When the second node is an AF, in one possible case, the second node is a trusted AF, then the second node can directly send the parameter information of the second node to the first network element. In another possible case, the second node is an untrusted third-party AF, then the second node can send (forward) the parameter information of the second node to the first network element through the NEF, and accordingly, the first network element can receive the parameter information of the second node through the NEF. In this case, the second node can map the parameter information of the second node to a certain event (for example, recorded as the second event).
  • a certain event for example, recorded as the second event
  • the second node may generate parameter information of the second event based on the parameter information of the second node, and then the second node may send the parameter information of the second event to the NEF, and the NEF may forward the parameter information of the second event to the first network element, thereby sending the parameter information of the AF to the first network element.
  • the second node when the second node is an untrusted third-party AF, the second node sends parameter information of the second node to the first network element, including: the second node sends parameter information of the second event to the first network element through the NEF, wherein the parameter information of the second event is generated based on the parameter information of the second node.
  • the parameter information of the second event can be used to indicate at least one of the following: whether the AF supports the execution of the second event; the time period in which the AF supports the execution of the second event; whether the AF can be the initiator of the second event; the computing power of the AF; and the identifier of the application corresponding to the AF; wherein the second event is associated with the VFL task, in other words, the execution of the second event can also be understood as the execution of the VFL task.
  • the parameter information of the second node can be carried by, for example, a second registration request message, and the second registration request message is a registration request message sent by the second node to the first network element.
  • the second node can register the parameter information of the second node in the first network element by sending the second registration request message to the first network element.
  • each second node may send parameter information of each second node to the first network element before the first node initiates the VFL task, so that the first network element may receive parameter information of each second node before the first node initiates the VFL task.
  • the method may also include: the first network element obtains capability information of at least one second node. For example, the first network element may obtain capability information of each second node from the received parameter information of each second node. Furthermore, for each second node, the first network element may determine whether the second node meets the conditions for executing the VFL task based on the capability information of the second node.
  • the first node may send (register) the parameter information of the first node to the first network element, and the method may further include: the first node sends the parameter information of the first node to the first network element.
  • the parameter information of the first node may include, for example, at least one of the following: capability information of the first node; an identifier of the first node; address information of the first node; and an identifier of an application corresponding to the first node.
  • the capability information of the first node may be used to indicate at least one of the following: whether the first node supports executing the VFL task; the time period in which the first node supports executing the VFL task; whether the first node can serve as the initiator of the VFL task; the computing power of the first node; and the service area of the first node.
  • each parameter information of the first node is the same as the meaning of each parameter information corresponding to the second node, and they can refer to each other, which will not be repeated here.
  • the first node When the first node is an AF, one possible situation is that the first node is a trusted AF, then the first node can directly send the parameter information of the first node to the first network element. Another possible situation is that the first node is an untrusted third-party AF, then the first node can send (forward) the parameter information of the first node to the first network element through the NEF, and accordingly, the first network element can receive the parameter information of the first node through the NEF.
  • the first node can map the parameter information of the first node to the parameter information of a certain event (for example, recorded as the first event), or the first node can generate the parameter information of the first event based on the parameter information of the first node, and then the first node can send the parameter information of the first event to the NEF, and then the NEF forwards the parameter information of the first event to the first network element, thereby sending the parameter information of the AF to the first network element.
  • a certain event for example, recorded as the first event
  • the first node can generate the parameter information of the first event based on the parameter information of the first node, and then the first node can send the parameter information of the first event to the NEF, and then the NEF forwards the parameter information of the first event to the first network element, thereby sending the parameter information of the AF to the first network element.
  • the first node when the first node is an untrusted third-party AF, the first node sends parameter information of the first node to the first network element, including: the first node sends parameter information of the first event to the first network element through the NEF, and the parameter information of the first event is generated based on the parameter information of the first node.
  • the parameter information of the first event is used to indicate at least one of the following: whether the AF supports the execution of the first event; the time period in which the AF supports the execution of the first event; whether the AF can be the initiator of the first event; the computing power of the AF; and the identifier of the application corresponding to the AF.
  • the first event is associated with the VFL task. In other words, executing the first event can also be understood as executing the VFL task.
  • the parameter information of the first node can be carried by, for example, the first registration request information, which is the registration request information sent by the first node to the first network element.
  • the first node can register the parameter information of the first node in the first network element by sending the first registration request information to the first network element.
  • a first network element sends first information to a first node.
  • the first node receives the first information.
  • the first information may be used to indicate at least one second node that meets the condition for executing a VFL task, where the VFL task is initiated by the first node.
  • the first network element may send first information to the first node to indicate at least one second node that meets the conditions for executing the VFL task.
  • the first node may learn at least one second node that meets the conditions for executing the VFL task through the first information, thereby ensuring that the first node can find a suitable node to perform the VFL task.
  • the first information may carry, for example, identification and/or address information of at least one second node that meets the conditions for executing the VFL task.
  • the method may further include: the first node sends second information to the first network element, where the second information is used to request the first information; accordingly, the first network element may receive the second information from the first node,
  • the first node may send the second information to the first network element to request the first information.
  • the first network element may determine at least one second node that meets the conditions for executing the VFL task initiated by the first node based on the acquired capability information of each second node, and feed back the first information to the first node, thereby indicating the at least one second node to the first node.
  • the first node When the first node is an AF, in one possible case, the first node is a trusted AF, then the first node can directly send the second information to the first network element. In another possible case, the first node is an untrusted third-party AF, then the first node can send (forward) the second information to the first network element through the NEF, and accordingly, the first network element can receive the second information from the first node through the NEF.
  • the second information may be used to indicate at least one of the following:
  • the time period for executing the VFL task or in other words, the time period that the first node expects to execute the VFL task, such as 9:00 to 21:00.
  • the computing power required to execute the VFL task such as the minimum computing power required to execute the VFL task initiated by the first node
  • a service area for executing the VFL task or in other words, a service area that the first node expects to execute the VFL task.
  • the first network element can compare the capability information of each second node stored in the first network element with the content of the second information, determine that it supports performing the VFL task during the period from 9:00 to 21:00, and the service area includes at least one second node in the first service area, and carry the information of the at least one second node (such as identification and/or address information) in the first information and send it to the first node.
  • the first node determines at least one participating node that participates in the VFL task.
  • the first node may determine at least one participating node participating in the VFL task from at least one second node that meets the condition for executing the VFL task.
  • condition for executing the VFL task refers to the conditions for executing the VFL task initiated by the first node.
  • the "at least one second node that meets the conditions for executing the VFL task” is referred to as “at least one second node” hereinafter. That is, the “at least one second node” mentioned hereinafter refers to "at least one second node that meets the conditions for executing the VFL task" without causing ambiguity.
  • determining at least one participating node participating in the VFL task from at least one second node includes: for a second node that agrees to participate in the VFL task initiated by the first node, determining the second node as a participating node participating in the VFL task. That is, if a second node meets the conditions for executing the VFL task and agrees to participate in the VFL task initiated by the first node, the first node can be determined as a participating node participating in the VFL task.
  • the method may further include: the second node sends third information to the first node, the third information being used to indicate whether the second node agrees to participate in the VFL task initiated by the first node. Accordingly, the first node may receive the third information from the at least one second node. Based on the third information, the first node may learn whether each second node agrees to participate in the VFL task initiated by the first node.
  • the method may further include: the first node sends data requirement information to at least one second node, wherein the data requirement information is used to determine data for performing the VFL task. Accordingly, each of the at least one second node may receive the data requirement information from the first node.
  • the data requirement information may include: the type of data required to perform the VFL task, and/or the effective time of the required data.
  • the type of data may include the format of the data (such as 8-bit quantization or 24-bit quantization), and may also include the type of data content (such as application layer data, data related to device power, data related to device communication performance, etc.).
  • the second node that receives the data requirement information can, based on the data requirement information, determine the local data of the application layer that is valid in the first time period and is of 8-bit quantized type as the data for executing the VFL task.
  • the second node may decide whether to agree to participate in the VFL task initiated by the first node based on the local configuration or relevant policies. For example, if the second node has data that meets the corresponding data demand locally, it may agree to participate in the VFL task; if the second node does not have data that meets the corresponding data demand locally, it may refuse to participate in the VFL task. Furthermore, the second node may reply to the first node through the third information about whether to agree to participate in the VFL task.
  • the first node When the first node is an AF, in one possible case, the first node is a trusted AF, then the first node can directly send data demand information to at least one second node. In another possible case, the first node is an untrusted third-party AF, then the first node can send (forward) data demand information to at least one second node through the NEF, and accordingly, the second node can receive the data demand information from the first node through the NEF.
  • S604 The first node sends the VFL model to the participating nodes. Correspondingly, the participating nodes receive the VFL model.
  • the participating node may also be understood as a second node determined as a participating node, or may also be understood as a node participating in the first node sending
  • the first node sends the VFL model to the participating nodes, which can also be understood as the server corresponding to the first node (that is, the server corresponding to AF) sending the VFL model to the participating nodes.
  • the first node may determine at least one participating node participating in the VFL task from at least one second node, wherein each participating node may correspond to a VFL model.
  • the first node may send a corresponding VFL model to each participating node, and the VFL model may be used to execute the VFL task.
  • the VFL model may, for example, include information of the corresponding participating node (such as identification and/or address information of the participating node).
  • the VFL model corresponding to a certain participating node can also be understood as the VFL model required by the participating node.
  • the VFL model can be used as the initial model when the participating node performs the VFL task.
  • the VFL model required by the participating node can be generated by the first node.
  • the participating nodes include terminal devices, access network elements and core network elements, then the first node can generate the VFL model required by the terminal device for the terminal device, and can generate the VFL model required by the access network element for the access network element, and can generate the VFL model required by the core network element for the core network element.
  • the first node can send the VFL model required by each participating node to the corresponding participating node.
  • the first node cannot directly send the VFL model to the corresponding participating node.
  • the first node may forward the VFL model corresponding to the participating node to the participating node through one (or several) intermediate devices.
  • the first node when the first node is AF and the participating nodes are terminal devices or access network devices, AF cannot directly send the VFL model to the terminal device or access network device.
  • the first node sends the corresponding VFL model to the participating nodes, which may include: the first node sends the corresponding VFL model to the terminal device or access network network element through the second network element.
  • the second network element can be a network element, or it can include two or more network elements, which is not limited in the embodiment of the present application.
  • the second network element after receiving the VFL model, the second network element can forward the VFL model to the corresponding participating node based on the identifier and/or address information of the participating node carried in the VFL model.
  • VFL model #1 carries the identifier and address information of the terminal device
  • the first node can send VFL model #1 to the terminal device based on the identifier and address information of the terminal device
  • VFL model #2 carries the identifier and address information of the access network network element
  • the first node can send VFL model #2 to the access network network element based on the identifier and address information of the access network network element.
  • the VFL model may not need to carry the information of the participating nodes (such as identification and/or address information).
  • the method may further include: the first node sends the same association identifier to at least one participating node, and the association identifier can be used to identify different execution results of the same VFL task, and can also be used to aggregate different execution results of the same VFL task.
  • the second node can receive the association identifier from the first node.
  • the first node can send the VFL model and the association identifier to the corresponding participating node through the same message.
  • the different execution results of the same VFL task may include, for example, execution results generated when different nodes (such as the first node; or the second node determined as a participating node) execute the same VFL task.
  • the execution result may include a model training result and a model reasoning result.
  • the association identifier may include a model association identifier and a result association identifier, wherein the model association identifier can be used to identify and/or aggregate different model training results of the same VFL task, and the result association identifier can be used to identify and/or aggregate different model reasoning results of the same VFL task.
  • the received VFL model can be used as the initial model for executing the VFL task, and the data for executing the VFL task determined based on the data requirement information in S603 can be used as input data.
  • the first node can generate (acquire) the VFL model corresponding to the first node, and can determine the local data for the first node to execute the VFL task.
  • each participating node and the first node can jointly execute the VFL task initiated by the first node based on their respective VFL models and their respective local data.
  • the method may further include: each participating node (or the second node that has participated in the VFL task) sends the execution result of the participating node executing the VFL task to the first node, wherein the execution result may carry an association identifier.
  • each participating node or the second node that has participated in the VFL task
  • sends the execution result of the participating node executing the VFL task to the first node wherein the execution result may carry an association identifier.
  • sending the execution result of the participating node executing the VFL task to the first node can also be understood as sending the execution result of the participating node executing the VFL task to the server corresponding to the first node (that is, the server corresponding to AF).
  • the participating node can directly send the execution result of the VFL task executed by the participating node to the first node.
  • a participating node (hereinafter referred to as participating node #1) cannot directly send the execution result to the first node.
  • the participating node #1 can first send the execution result of the participating node #1 to an intermediate node (such as participating node #2), and then the intermediate node can forward the execution result of the participating node #1 to the first node.
  • implementation #1 After receiving the execution result of participating node #1, participating node #2 can send the execution result of participating node #1 to the first node, and at the same time, participating node #2 can also send its own execution result to the first node.
  • the execution results of participating node #1 and participating node #2 both carry the same association identifier.
  • the participating node #2 may also aggregate the execution result of the participating node #2 and the received execution result of the participating node #1, and then send the obtained aggregated result to the first node.
  • the execution results of the participating node #1 and the participating node #2 may carry the same association identifier, so that when the participating node #2 aggregates the execution results, the execution results with the same association identifier (the execution results of the participating node #1 and the participating node #2) may be aggregated, thereby ensuring that the aggregated execution results are the execution results of the same VFL task, thereby ensuring the correct aggregation of the execution results.
  • the obtained aggregated result may also carry the association identifier.
  • participating node #2 in addition to receiving the execution result of participating node #1, participating node #2 can also receive the execution results of other participating nodes (hereinafter referred to as participating node #3), and can aggregate the execution results of participating node #1, participating node #2 and participating node #3, and then send the obtained aggregated results to the first node.
  • participating node #3 in addition to receiving the execution result of participating node #1, participating node #2 and participating node #3, and then send the obtained aggregated results to the first node.
  • the execution results of participating node #1, participating node #2 and participating node #3 can carry the same association identifier, so that when participating node #2 aggregates the execution results, it can aggregate the execution results with the same association identifier (the execution results of participating node #1, participating node #2 and participating node #3), thereby ensuring that the aggregated execution results are the execution results of the same VFL task, thereby ensuring the correct aggregation of the execution results.
  • the obtained aggregated result can also carry the association identifier.
  • the method may further include: receiving an execution result (such as a second execution result) of at least one participating node (such as participating node #1 and/or participating node #3) executing the VFL task; sending an aggregated result to the first node, the aggregated result being obtained by aggregating the execution result of at least one participating node executing the VFL task (i.e., at least one second execution result) and the execution result of the second node executing the VFL task (i.e., a third execution result), wherein the aggregated result may carry an association identifier.
  • an execution result such as a second execution result
  • the aggregated result may carry an association identifier.
  • each execution result i.e., each second execution result
  • the execution result of the second node executing the VFL task i.e., the third execution result
  • the first node may receive the execution result corresponding to the VFL task initiated by the first node (for example, recorded as the first execution result), wherein the first execution result includes: the execution result of at least one participating node executing the VFL task (for example, the execution result obtained through the above-mentioned implementation method #1, for example, recorded as the second execution result); and/or, at least one aggregation result, each aggregation result in the at least one aggregation result is obtained by aggregation based on the execution results of at least two participating nodes executing the VFL task (that is, at least two second execution results) (for example, obtained through the above-mentioned implementation method #2 or implementation method #3).
  • the first execution result includes: the execution result of at least one participating node executing the VFL task (for example, the execution result obtained through the above-mentioned implementation method #1, for example, recorded as the second execution result); and/or, at least one aggregation result, each aggregation result in the at least one aggregat
  • each execution result in the execution result of at least one participating node executing the VFL task (that is, each second execution result), and/or each aggregation result in the at least one aggregation result, all carry the same association identifier, and each aggregation result in the at least one aggregation result can be obtained by aggregation based on the same association identifier.
  • the method may further include: the first node aggregating execution results corresponding to VFL tasks having the same association identifier and execution results of the VFL tasks executed by the first node.
  • the execution result corresponding to the VFL task and the execution result of the first node itself can be aggregated to obtain the final aggregated result.
  • the execution result corresponding to the VFL task received by the first node and the execution result of the first node itself can both carry the same association identifier.
  • the first node aggregates the execution results it can aggregate the execution results with the same association identifier (the execution result corresponding to the VFL task received by the first node and the execution result of the first node itself). In this way, it can be ensured that the execution results of the same VFL task are aggregated, thereby ensuring the correct aggregation of the execution results.
  • the first network element is NRF
  • the first node is AF
  • the second node includes UE, RAN and core network NF (or referred to as core network network element).
  • UE, RAN or core network NF can be one or more (two or more), which is not limited to this in the embodiments of the present application.
  • the node as the initiator In order to execute VFL, the node as the initiator first needs to discover nodes that have the ability to execute VFL, or nodes that meet the conditions for executing VFL tasks. To this end, before starting to execute VFL tasks, each node needs to register its own information with NRF. Taking AF as the initiator as an example, UE, RAN and core network elements need to register their ability to support VFL with NRF, so that AF can discover suitable nodes through NRF to execute VFL tasks.
  • FIG. 7 is a schematic diagram of a possible implementation flow of a communication method provided in an embodiment of the present application.
  • the implementation flow may include:
  • S701 UE sends a registration request to NRF.
  • the UE can register the UE information in the NRF by sending a registration request to the NRF.
  • the registration request carries the UE's profile (corresponding to the parameter information of the terminal in the aforementioned embodiment), and the UE's profile includes but is not limited to at least one of the following: UE ID, the UE's ability to support VFL, the time period during which the UE can support VFL, the UE's type, the UE's address information (such as IP address), the UE's indication of whether it can support becoming a VFL initiator, and the UE's computing power information.
  • the RAN can register RAN information in NRF by sending a registration request to NRF.
  • the registration request carries RAN files (corresponding to the parameter information of the access network element in the aforementioned embodiment), and the RAN files include but are not limited to at least one of the following: RAN ID (such as gNB ID), RAN's ability to support VFL, RAN's time period for supporting VFL, RAN's address information, indication of whether it can support becoming a VFL initiator, RAN's computing power information, and RAN's service area (coverage area) information.
  • RAN ID such as gNB ID
  • RAN's ability to support VFL such as gNB ID
  • RAN's time period for supporting VFL RAN's address information
  • indication of whether it can support becoming a VFL initiator RAN's computing power information
  • RAN's service area (coverage area) information RAN's service area (coverage area) information.
  • the core network NF sends a registration request to the NRF.
  • the core network NF can register the information of the core network NF with the NRF by sending a registration request to the NRF.
  • the registration request carries the file of the core network NF (corresponding to the parameter information of the core network element in the aforementioned embodiment), and the file of the core network NF includes but is not limited to at least one of the following: the type of NF, whether the NF can support the VFL capability, the time period during which the NF can support the VFL, the address information of the NF, whether the NF can support the indication of becoming the VFL initiator, the computing power information of the NF, and the service area information of the NF.
  • AF can register AF information with NRF by sending a registration request to NRF.
  • the registration request carries AF files (corresponding to the parameter information of AF in the aforementioned embodiment), and AF files include but are not limited to at least one of the following: AF ID, application ID corresponding to AF, whether AF can support VFL capability, time period during which AF can support VFL, address information of AF, whether AF can support becoming the initiator of VFL, and computing power information of AF.
  • AF can register information to NRF through NEF.
  • AF can first send AF's file to NEF, NEF generates a NEF file, and NEF registers the NEF file to NRF.
  • NEF can map AF's information to the information of an event (Event), or to an event ID, so that NEF can register the information of the event (corresponding to the parameter information of the first event in the aforementioned embodiment) as a NEF file to NRF.
  • the information of the event includes but is not limited to at least one of the following: NEF ID, whether AF can support VFL capabilities, the application ID corresponding to AF, the time period supported by the event, whether AF can support an indication of being the initiator of the event, and AF's computing power information.
  • NRF saves the files of each node.
  • NRF After receiving the registration request from each node (UE, RAN, core network NF and AF), NRF can save the file of each node and mark each node as an available valid node. After that, NRF can send a registration response to each node to reply that the registration of each node is accepted.
  • the step of NRF sending a registration response to each node may include S706 to S709:
  • NRF sends a registration response to the core network NF.
  • the NRF sends a registration response to the RAN.
  • the NRF sends a registration response to the UE.
  • the NRF sends a registration response to the AF.
  • each node can register its own information into NRF, so that when executing a VFL task, the node as the initiator can find a suitable node through NRF to perform the VFL task.
  • FIG8 is a second schematic diagram of a possible implementation process of the communication method provided in an embodiment of the present application.
  • the implementation process shown in FIG8 can be executed after the implementation process shown in FIG6, that is, in the implementation process shown in FIG8, it can be considered that the information of each node has been registered in the NRF.
  • the implementation process may include:
  • the AF sends a node discovery request to the NRF (corresponding to the second information in the aforementioned embodiment).
  • the node discovery request may include but is not limited to at least one of the following information: an indication of executing VFL, a time period of the VFL to be executed, an indication of wanting to be the initiator of VFL, and a service area where VFL is to be executed.
  • the AF when the AF is a third-party AF, the AF may send a node discovery request to the NRF through the NEF, and the parameters included in the node discovery request are the same.
  • NRF authorizes AF's node discovery request.
  • the NFR may authorize the node discovery request of the AF.
  • the NRF may first determine whether the AF can support VFL and whether it can support becoming a VFL initiator based on the file information of the AF. Further, if the AF can support VFL and can support becoming a VFL initiator, the node discovery request of the AF is authorized.
  • the NRF sends a node discovery request response to the AF (corresponding to the first information in the aforementioned embodiment).
  • the NRF can determine the nodes that meet the conditions for the AF (such as UE, RAN and core network NF that meet the conditions), and reply the information of the nodes that meet the conditions (such as the ID and/or address information of the node) to the AF through the node discovery request response.
  • the conditions for the AF such as UE, RAN and core network NF that meet the conditions
  • reply the information of the nodes that meet the conditions such as the ID and/or address information of the node
  • the "qualified node” can be understood as a node that meets the conditions included in the node discovery request, or can also be understood as a node that meets the conditions for executing the VFL task initiated by the AF. For ease of explanation, it is assumed below that the UR, RAN and core network NF are all qualified nodes.
  • the NRF may send a node discovery request response to the AF through the NEF.
  • the AF can send a VFL preparation request to the NEF, which then forwards the VFL preparation request to each node through the NEF.
  • the VFL preparation request may include the type of data required to execute the VFL task and the valid time of the data (corresponding to the data requirement information in the aforementioned embodiment).
  • NEF sends a VFL preparation request to each node.
  • NEF After receiving the VFL preparation request from AF, NEF can send the VFL preparation request to each node that meets the conditions (as shown in Figure 8, NEF can send the VFL preparation request to UE, RAN and core network NF respectively), so that each node can prepare the required data for executing the VFL task in advance based on the type of data required for executing the VFL task carried in the VFL preparation request and the validity period of the data.
  • each node sends a VFL preparation request reply message to the NEF (corresponding to the third information in the above embodiment).
  • each node After receiving the VFL preparation request, each node can determine whether to join the VFL task based on the local configuration, and reply to the NEF with the decision of whether to agree to join.
  • the decision can be carried in the VFL preparation request reply message, or in other words, the VFL preparation request message can be used to indicate whether each node agrees to join the VFL task initiated by the AF.
  • NEF can select VFL nodes based on the VFL preparation request reply message fed back by each node and based on the request of AF. For example, NEF can select nodes that meet the conditions for executing the VFL task initiated by AF and agree to join the VFL task.
  • NEF sends the node selection result to AF.
  • NEF may return the node selection result in S807 to AF, wherein the node selection result may include information of the selected node, such as the ID and/or address information of the node, so that AF can determine the selected node as a participating node participating in the VFL task.
  • S804 to S808 are exemplary descriptions based on the AF being a third-party AF.
  • the AF can directly send a VFL preparation request to each node that meets the conditions (such as UE, RAN, and core network NF), and then each node that meets the conditions can directly send a VFL preparation request reply message to the AF.
  • the VFL preparation request message can be used to indicate whether each node agrees to join the VFL task initiated by the AF, and then the AF can determine the nodes that meet the conditions and agree to join the VFL task as participating nodes participating in the VFL task.
  • the AF can discover and determine a suitable node to perform the VFL task by sending a node discovery request to the NRF.
  • FIG9 is a possible implementation flow diagram of the communication method provided in the embodiment of the present application.
  • the implementation flow shown in FIG9 can be executed after the implementation flow shown in FIG7. That is, in the implementation flow shown in FIG9, it can be considered that the AF has determined the nodes participating in the VFL task.
  • the participating nodes determined by the AF are the UE, the RAN, and the core network NF.
  • the implementation flow may include:
  • the server corresponding to the AF sends the VFL model and association identifier corresponding to each node to the core network NF.
  • the VFL model corresponding to each node can be used as the initial model for each node to perform the VFL task.
  • the server corresponding to the AF may first send the (corresponding) VFL model required by each node (UE, RAN, core network NF) participating in the VFL task to the core network NF.
  • Each model may carry the information of the corresponding node.
  • the model required by the UE is model #1
  • the model required by the RAN is model #2.
  • model #1 may carry the information of the UE (such as the ID and/or address information of the UE) to indicate that model #1 needs to be sent to the UE for use
  • model #2 may carry the information of the RAN (such as the ID and/or address information of the RAN) to indicate that model #2 needs to be sent to the RAN for use.
  • AF can also send the same association identifier (association ID) to each node. Since VFL needs to use local models and data on different nodes for model training and result inference, and aggregate them at the initiator, in order to facilitate the association of the VFL execution results (including model training results and/or inference results) generated by each node, AF can send the same association identifier to each node. In this way, after each node calculates the VFL execution result, it can add the association identifier to the VFL execution result of the node, thereby ensuring that the results generated by the same VFL task on different nodes have the same association identifier, thereby ensuring that the aggregation node correctly aggregates the VFL execution results obtained by each node.
  • "aggregation node” can also be understood as a node that aggregates the VFL execution results of at least two nodes.
  • association identifier may be carried in a VFL model corresponding to each node, for example.
  • the core network NF sends the VFL model and association identifier corresponding to the RAN to the RAN.
  • the core network NF After the core network NF obtains the (corresponding) VFL model and associated identifier required by the RAN, it can send the VFL model and associated identifier to the RAN based on the RAN information carried by the VFL model (such as the ID and/or address information of the RAN).
  • the core network NF sends the VFL model and association identifier corresponding to the UE to the UE.
  • the core network NF After the core network NF obtains the (corresponding) VFL model and associated identifier required by the UE, it can send the VFL model and associated identifier to the UE based on the UE information (such as the UE ID and/or address information) carried by the VFL model. As an implementation method, the VFL can send the VFL model and associated identifier corresponding to the UE to the UE through the RAN.
  • the server corresponding to the AF can also directly send the VFL model and association identifier corresponding to the UE through the application layer.
  • the implementation process also includes S904:
  • the server corresponding to the AF sends the VFL model and the association identifier corresponding to the UE to the UE through the application layer.
  • the server corresponding to the AF in S901 may not need to send the VFL model corresponding to the UE to the core network NF, and in this case, there is no need to execute S903.
  • each node executes the VFL task.
  • each node can perform federated learning of the same task based on the association identifier. For example, in the process of executing a VFL task, each node may need to exchange intermediate results (such as gradients, losses, etc.). Since the results generated by each node have the same association identifier, by processing the results with the same association identifier, it can be ensured that each node is performing federated learning of the same task.
  • intermediate results such as gradients, losses, etc.
  • the UE sends the execution result of the UE executing the VFL task to the core network NF.
  • the UE After executing the VFL task, the UE can send the execution result to the core network NF, and the execution result can carry the association identifier.
  • RAN sends the execution result of the VFL task executed by RAN to the core network NF.
  • the RAN may send the obtained execution result to the core network NF, and the execution result may carry the association identifier.
  • the implementation process further includes S908:
  • the core network NF After the core network NF receives the execution results of the VFL task from the UN and RAN, it can aggregate the execution results with the same association identifier (the execution results of the UN, RAN and the core network NF itself) based on the association identifier, and then send the aggregated results to the server corresponding to the AF in S909.
  • the same association identifier the execution results of the UN, RAN and the core network NF itself
  • the core network NF may not aggregate the execution results, but directly send the received execution results (the execution results of the UE and the RAN) and the execution results generated by the core network NF to the server corresponding to the AF.
  • the core network NF sends the execution result corresponding to the VFL task to the server corresponding to the AF.
  • the core network NF receives the execution results of the VFL task from the UN and RAN, it aggregates the execution results of each node.
  • the execution result corresponding to the VFL task may be an aggregated result obtained by aggregating the execution results of each node, and the aggregated result carries an association identifier.
  • the core network NF receives the execution results of the VFL task from the UN and RAN, it does not aggregate the execution results of each node. Then the execution results corresponding to the VFL task may include the execution results of each node, such as the execution results of the UN, RAN and the core network NF, and the execution results of each node carry the same association identifier.
  • the UE may also directly send the execution result of the UE to the server corresponding to the AF through the application layer, and carry the association identifier in the execution result.
  • the implementation process may also include S710:
  • the UE sends the execution result of the VFL task executed by the UE to the server corresponding to the AF through the application layer.
  • the server corresponding to AF aggregates the execution results.
  • the server corresponding to the AF can aggregate the results generated by all nodes to obtain the final result of the VFL.
  • the server corresponding to the AF can aggregate the aggregated result obtained by the core network NF and the execution result generated by the server corresponding to the AF itself based on the association identifier to obtain the final execution result of the VFL task.
  • the aggregated result obtained by the core network NF and the execution result generated by the server corresponding to the AF itself have the same association identifier.
  • AF can send the initial VFL model to each node participating in the VFL task, and can associate the execution results of the VFL task executed by each node through the association identifier, and then aggregate the execution results of the associated nodes to obtain the final execution result of the VFL task.
  • the AF as the initiator of the VFL task in the above process is only exemplary, and in the technical solution of the embodiment of the present application, any node can be the initiator of the VFL task.
  • the UE or AF can usually be the initiator of the VFL task.
  • the embodiment of the present application introduces the process of each node registering, the initiator discovering and determining the participating nodes, the initial model sending and the execution result aggregation, thereby realizing the network architecture's support for cross-domain vertical federated learning.
  • each node can register its own information related to VFL to NRF to ensure that the initiator can find the appropriate node.
  • the initiator can find each node that can participate in VFL through NRF.
  • the initiator can send an initial model and an association identifier to each node to ensure that each node executes the same VFL task and the results are correctly aggregated.
  • This scheme solves the following technical problems that have not yet been solved: 1) How does the node as the initiator (such as UE or AF) select the appropriate node? 1) How the initiator sends the initial model to each node for model training and/or result inference; 2) How the initiator associates the training/inference results generated by each node.
  • the initiator can unite nodes in multiple domains to perform vertical federated learning without leaving the domain, thereby enabling comprehensive analysis and prediction of the research object (such as UE experience) by uniting more dimensional data, so as to optimize the research content in each domain (such as optimizing the UE experience).
  • the research object such as UE experience
  • the embodiments of the present application provide corresponding communication devices.
  • FIG. 10 is a schematic diagram of a structure of a communication device provided in an embodiment of the present application, which is applied to a first node.
  • the communication device 1000 includes:
  • the first receiving module 1001 is configured to receive first information from a first network element, where the first information is used to indicate at least one second node that meets the conditions for performing the VFL task; the determining module 1002 is configured to determine at least one participating node participating in the VFL task from the at least one second node.
  • whether the second node meets the conditions for executing the VFL task is determined by the first network element based on the capability information of the second node, and the capability information of the second node is used to indicate at least one of the following: whether the second node supports executing the VFL task; the time period in which the second node supports executing the VFL task; whether the second node can serve as the initiator of the VFL task; the computing power of the second node; and the service area of the second node.
  • parameter information of at least one second node is collected or determined by a first network element, wherein, for each second node, the parameter information of the second node includes at least one of the following: capability information of the second node; an identifier of the second node; address information of the second node; and an identifier of an application corresponding to the second node.
  • parameter information of the second node is carried by second registration request information, and the second registration request information is registration request information sent by the second node to the first network element.
  • the device 1000 also includes: a first sending module, configured to send parameter information of the device 1000 to the first network element before receiving the first information from the first network element, the parameter information of the device 1000 including at least one of the following: capability information of the device 1000; an identifier of the device 1000; address information of the device 1000; and an identifier of an application corresponding to the device 1000; wherein the capability information of the device 1000 is used to indicate at least one of the following: whether the device 1000 supports executing VFL tasks; the time period in which the device 1000 supports executing VFL tasks; whether the device 1000 can be used as the initiator of the VFL task; the computing power of the device 1000; and the service area of the device 1000.
  • a first sending module configured to send parameter information of the device 1000 to the first network element before receiving the first information from the first network element, the parameter information of the device 1000 including at least one of the following: capability information of the device 1000; an identifier of the device 1000; address information of the device 1000; and an identifier of an application corresponding to the device 1000
  • device 1000 is an application function AF
  • the first sending module is specifically configured to: send parameter information of a first event to a first network element through a network function open NEF, and the parameter information of the first event is generated based on the parameter information of device 1000; the parameter information of the first event is used to indicate at least one of the following: whether AF supports the execution of the first event; the time period in which AF supports the execution of the first event; whether AF can be the initiator of the first event; the computing power of AF; and the identifier of the application corresponding to AF; wherein the first event is associated with a VFL task.
  • parameter information of the device 1000 is carried by first registration request information, and the first registration request information is registration request information sent by the device 1000 to the first network element.
  • the apparatus 1000 further includes: a second sending module configured to send second information to the first network element before receiving the first information from the first network element, wherein the second information is used to request the first information.
  • a second sending module configured to send second information to the first network element before receiving the first information from the first network element, wherein the second information is used to request the first information.
  • the apparatus 1000 is an AF
  • the second sending module is specifically configured to: send the second information to the first network element through the NEF.
  • the second information is used to indicate at least one of the following: executing the VFL task; the time period for executing the VFL task; being the initiator of the VFL task; the computing power requirement for executing the VFL task; and the service area for executing the VFL task.
  • the determination module 1002 is specifically configured to: for a second node that agrees to participate in the VFL task, determine the second node as a participating node that participates in the VFL task.
  • the device 1000 also includes: a third sending module, configured to send data requirement information to at least one second node, the data requirement information is used to determine the data for executing the VFL task, and the data requirement information includes: the type of data required to execute the VFL task, and/or the validity time of the required data.
  • a third sending module configured to send data requirement information to at least one second node, the data requirement information is used to determine the data for executing the VFL task, and the data requirement information includes: the type of data required to execute the VFL task, and/or the validity time of the required data.
  • the apparatus 1000 further includes: a second receiving module configured to receive third information from at least one second node, where the third information is used to indicate whether the second node agrees to participate in the VFL task.
  • the apparatus 1000 is an AF
  • the third sending module is specifically configured to: send data demand information to at least one second node through the NEF.
  • the apparatus 1000 further includes: a fourth sending module configured to send a corresponding VFL model to at least one participating node, the VFL model being used to execute the VFL task; wherein the VFL model includes the identification and/or address information of the corresponding participating node.
  • the apparatus 1000 is an AF
  • the participating node is a terminal device or an access network element
  • the fourth sending module is specifically configured to: send the corresponding VFL model to the terminal device or the access network element through the second network element.
  • the apparatus 1000 further includes: a fifth sending module configured to send an association identifier to at least one participating node, where the association identifier is used to aggregate different execution results of the same VFL task.
  • the device 1000 also includes: a third receiving module, configured to receive a first execution result corresponding to the VFL task, the first execution result including: a second execution result of at least one participating node executing the VFL task; and/or, at least one aggregation result, each aggregation result of the at least one aggregation result being aggregated based on at least two second execution results.
  • a third receiving module configured to receive a first execution result corresponding to the VFL task, the first execution result including: a second execution result of at least one participating node executing the VFL task; and/or, at least one aggregation result, each aggregation result of the at least one aggregation result being aggregated based on at least two second execution results.
  • each second execution result, and/or each aggregated result in at least one aggregated result carries the same association identifier.
  • each aggregated result in the at least one aggregated result is obtained by aggregation based on the same association identifier.
  • the apparatus 1000 further includes: an aggregation module configured to aggregate execution results corresponding to VFL tasks having the same association identifier and execution results of the VFL tasks executed by the apparatus 1000 .
  • the apparatus 1000 is a terminal device, an access network element, a core network element or an AF.
  • the second node includes at least one of the following: a terminal device, an access network element, a core network element, and an AF.
  • the first network element is a network storage function NRF.
  • FIG. 11 is a second schematic diagram of the structure of a communication device provided in an embodiment of the present application, which is applied to a second node.
  • the communication device 1100 includes:
  • the first sending module 1101 is configured to send capability information of the device 1100 to the first network element, where the capability information of the device 1100 is used to determine whether the device 1100 meets the conditions for executing the vertical federated learning VFL task.
  • the capability information of device 1100 is used to indicate at least one of the following: whether device 1100 supports executing VFL tasks; the time period in which device 1100 supports executing VFL tasks; whether device 1100 can act as the initiator of VFL tasks; the computing power of device 1100; and the service area information of device 1100.
  • device 1100 also includes: a second sending module, configured to send parameter information of device 1100 to the first network element, the parameter information of device 1100 including at least one of the following: capability information of device 1100; an identifier of device 1100; address information of device 1100; and an identifier of an application corresponding to device 1100.
  • a second sending module configured to send parameter information of device 1100 to the first network element, the parameter information of device 1100 including at least one of the following: capability information of device 1100; an identifier of device 1100; address information of device 1100; and an identifier of an application corresponding to device 1100.
  • device 1100 is an application function AF
  • the second sending module is specifically configured to: send parameter information of the second event to the first network element through the network function opening NEF, and the parameter information of the second event is generated based on the parameter information of device 1100; the parameter information of the second event is used to indicate at least one of the following: whether AF supports the execution of the second event; the time period in which AF supports the execution of the second event; whether AF can be the initiator of the second event; the computing power of AF; and the identifier of the application corresponding to AF; wherein the second event is associated with the VFL task.
  • the parameter information of the device 1100 is carried by the second registration request information, and the second registration request information is the registration request information sent by the device 1100 to the first network element.
  • the device 1100 also includes: a first receiving module, configured to receive data requirement information from the first node, the data requirement information is used to determine the data for executing the VFL task, the data requirement information includes: the type of data required to execute the VFL task, and/or the validity time of the required data.
  • a first receiving module configured to receive data requirement information from the first node, the data requirement information is used to determine the data for executing the VFL task, the data requirement information includes: the type of data required to execute the VFL task, and/or the validity time of the required data.
  • the apparatus 1100 further includes: a third sending module configured to send third information to the first node, where the third information is used to indicate whether the apparatus 1100 agrees to participate in the VFL task.
  • the first node is an AF
  • the first receiving module is specifically configured to: receive data demand information from the first node through the NEF.
  • the device 1100 further includes: a second receiving module configured to receive a VFL model from the first node, the VFL model being used to perform the VFL task; wherein the VFL model corresponds to the device 1100 , and the VFL model includes an identification and/or address information of the device 1100 .
  • the apparatus 1100 further includes: a third receiving module configured to receive an association identifier from the first node, where the association identifier is used to aggregate different execution results of the same VFL task.
  • the apparatus 1100 further includes: a third sending module configured to send an execution result of the apparatus 1100 executing the VFL task to the first node, where the execution result carries an association identifier.
  • the device 1100 also includes: a fourth receiving module, configured to receive a second execution result of at least one participating node executing the VFL task; a fourth sending module, configured to send an aggregation result to the first node, the aggregation result being obtained by aggregating at least one second execution result and a third execution result of the device 1100 executing the VFL task, wherein the aggregation result carries an association identifier.
  • each second execution result and the third execution result carry the same association identifier, and the aggregated result is obtained by aggregation based on the same association identifier.
  • the first node is a terminal device, an access network element, a core network element or an AF.
  • the apparatus 1100 is a terminal device, an access network element, a core network element or an AF.
  • the first network element is a network storage function NRF.
  • FIG. 12 is a schematic diagram of the third structure of a communication device provided in an embodiment of the present application, which is applied to a first network element.
  • the communication device 1200 includes:
  • the acquisition module 1201 is configured to acquire capability information of at least one second node, where the capability information of the second node is used to determine whether the second node meets the conditions for executing the vertical federated learning VFL task.
  • the capability information of the second node is used to indicate at least one of the following: whether the second node supports executing the VFL task; the time period in which the second node supports executing the VFL task; whether the second node can serve as the initiator of the VFL task; the computing power of the second node; and the service area information of the second node.
  • the device 1200 also includes: a first receiving module, configured to receive parameter information of at least one second node before obtaining the capability information of at least one second node, wherein, for each second node, the parameter information of the second node includes at least one of the following: capability information of the second node; an identifier of the second node; address information of the second node; and an identifier of an application corresponding to the second node.
  • a first receiving module configured to receive parameter information of at least one second node before obtaining the capability information of at least one second node, wherein, for each second node, the parameter information of the second node includes at least one of the following: capability information of the second node; an identifier of the second node; address information of the second node; and an identifier of an application corresponding to the second node.
  • parameter information of the second node is carried by second registration request information, and the second registration request information is registration request information sent by the second node to the device 1200 .
  • the device 1200 also includes: a second receiving module, configured to receive parameter information of the first node, the parameter information of the first node including at least one of the following: capability information of the first node; an identifier of the first node; address information of the first node; and an identifier of an application corresponding to the first node; wherein the capability information of the first node is used to indicate at least one of the following: whether the first node supports executing VFL tasks; the time period in which the first node supports executing VFL tasks; whether the first node can serve as the initiator of the VFL task; the computing power of the first node; and the service area of the first node.
  • a second receiving module configured to receive parameter information of the first node, the parameter information of the first node including at least one of the following: capability information of the first node; an identifier of the first node; address information of the first node; and an identifier of an application corresponding to the first node; wherein the capability information of the first node is
  • the parameter information of the first node is carried by first registration request information, and the first registration request information is registration request information sent by the first node to the device 1200.
  • the apparatus 1200 further includes: a sending module configured to send first information to the first node, where the first information is used to indicate at least one second node that meets the conditions for executing the VFL task.
  • the apparatus 1200 further includes: a third receiving module configured to receive second information from the first node before sending the first information to the first node, where the second information is used to request the first information.
  • the first node is a terminal device, an access network element, a core network element, or an application function AF.
  • the second node includes at least one of the following: a terminal device, an access network element, a core network element, and an AF.
  • the device 1200 is a network storage function NRF.
  • FIG13 is a schematic structural diagram of a communication device 1300 provided in an embodiment of the present application.
  • the communication device may be a terminal device, a network device, or a core network device.
  • the communication device 1300 shown in FIG13 includes a processor 1310, which may call and run a computer program from a memory to implement the method in the embodiment of the present application.
  • the communication device 1300 may further include a memory 1320.
  • the processor 1310 may call and run a computer program from the memory 1320 to implement the method in the embodiment of the present application.
  • the memory 1320 may be a separate device independent of the processor 1310 , or may be integrated into the processor 1310 .
  • the communication device 1300 may further include a transceiver 1330 , and the processor 1310 may control the transceiver 1330 to communicate with other devices, specifically, may send information or data to other devices, or receive information or data sent by other devices.
  • the transceiver 1330 may include a transmitter and a receiver.
  • the transceiver 1330 may further include an antenna, and the number of antennas may be one or more.
  • the communication device 1300 may specifically be the first node of the embodiment of the present application, and the communication device 1300 may implement the corresponding processes implemented by the first node in each method of the embodiment of the present application, which will not be described again for the sake of brevity.
  • the communication device 1300 may specifically be the second node of the embodiment of the present application, and the communication device 1300 may implement the corresponding processes implemented by the second node in each method of the embodiment of the present application, which will not be described in detail here for the sake of brevity.
  • the communication device 1300 may specifically be the first network element of the embodiment of the present application, and the communication device 1300 may implement the corresponding processes implemented by the first network element in each method of the embodiment of the present application, which will not be described again for the sake of brevity.
  • Fig. 14 is a schematic structural diagram of a chip according to an embodiment of the present application.
  • the chip 1400 shown in Fig. 14 includes a processor 1410, which can call and run a computer program from a memory to implement the method according to the embodiment of the present application.
  • the chip 1400 may further include a memory 1420.
  • the processor 1410 may call and run a computer program from the memory 1420 to implement the method in the embodiment of the present application.
  • the memory 1420 may be a separate device independent of the processor 1410 , or may be integrated into the processor 1410 .
  • the chip 1400 may further include an input interface 1430.
  • the processor 1410 may control the input interface 1430 to communicate with other devices or chips, and specifically, may obtain information or data sent by other devices or chips.
  • the chip 1400 may further include an output interface 1440.
  • the processor 1410 may control the output interface 1440 to communicate with other devices or chips, and specifically, may output information or data to other devices or chips.
  • the chip can be applied to the first node in the embodiments of the present application, and the chip can implement the corresponding processes implemented by the first node in each method of the embodiments of the present application. For the sake of brevity, they will not be repeated here.
  • the chip can be applied to the second node in the embodiments of the present application, and the chip can implement the corresponding processes implemented by the second node in each method of the embodiments of the present application. For the sake of brevity, they will not be repeated here.
  • the chip can be applied to the first network element in the embodiment of the present application, and the chip can implement the corresponding processes implemented by the first network element in each method of the embodiment of the present application. For the sake of brevity, they will not be repeated here.
  • the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.
  • the embodiment of the present application further provides a computer storage medium, which stores one or more programs.
  • the one or more programs can be executed by one or more processors to implement the method in the embodiment of the present application.
  • FIG15 is a schematic block diagram of a communication system 1500 provided in an embodiment of the present application. As shown in FIG15 , the communication system 1500 includes a first node 1510 , a second node 1520 , and a first network element 1530 .
  • the first node 1510 can be used to implement the corresponding functions implemented by the first node in the above method
  • the second node 1520 can be used to implement the corresponding functions implemented by the second node in the above method
  • the first network element 1530 can be used to implement the corresponding functions implemented by the first network element in the above method.
  • the first network element 1530 can be used to implement the corresponding functions implemented by the first network element in the above method.
  • the processor of the embodiment of the present application may be an integrated circuit chip with signal processing capabilities.
  • each step of the above method embodiment can be completed by the hardware integrated logic circuit in the processor or the instruction in the form of software.
  • the above processor can be a general processor, a digital signal processor (Digital Signal Processor, DSP), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.
  • DSP Digital Signal Processor
  • ASIC Application Specific Integrated Circuit
  • FPGA Field Programmable Gate Array
  • the methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed.
  • the general processor can be a microprocessor or the processor can also be any conventional processor, etc.
  • the steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to execute, or the hardware and software modules in the decoding processor can be executed.
  • the software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc.
  • the storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.
  • the memory in the embodiment of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories.
  • the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory.
  • the volatile memory can be a random access memory (RAM), which is used as an external cache.
  • RAM Direct Rambus RAM
  • SRAM Static RAM
  • DRAM Dynamic RAM
  • SDRAM Synchronous DRAM
  • DDR SDRAM Double Data Rate SDRAM
  • ESDRAM Enhanced SDRAM
  • SLDRAM Synchlink DRAM
  • DR RAM Direct Rambus RAM
  • the memory in the embodiment of the present application may also be static random access memory (static RAM, SRAM), dynamic random access memory (dynamic RAM, DRAM), synchronous dynamic random access memory (synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (double data rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (enhanced SDRAM, ESDRAM), synchronous link dynamic random access memory (synch link DRAM, SLDRAM) and direct memory bus random access memory (Direct Rambus RAM, DR RAM), etc. That is to say, the memory in the embodiment of the present application is intended to include but not limited to these and any other suitable types of memory.
  • An embodiment of the present application also provides a computer-readable storage medium for storing a computer program.
  • the computer-readable storage medium can be applied to the first node in the embodiments of the present application, and the computer program enables the computer to execute the corresponding processes implemented by the first node in the various methods of the embodiments of the present application. For the sake of brevity, they are not repeated here.
  • the computer-readable storage medium can be applied to the second node in the embodiments of the present application, and the computer program enables the computer to execute the corresponding processes implemented by the second node in the various methods of the embodiments of the present application. For the sake of brevity, they are not repeated here.
  • the computer-readable storage medium can be applied to the first network element in the embodiment of the present application, and the computer program enables the computer to execute the corresponding processes implemented by the first network element in the various methods of the embodiment of the present application. For the sake of brevity, they are not repeated here.
  • An embodiment of the present application also provides a computer program product, including computer program instructions.
  • the computer program product can be applied to the first node in the embodiments of the present application, and the computer program instructions enable the computer to execute the corresponding processes implemented by the first node in the various methods of the embodiments of the present application. For the sake of brevity, they are not repeated here.
  • the computer program product can be applied to the second node in the embodiments of the present application, and the computer program instructions enable the computer to execute the corresponding processes implemented by the second node in the various methods of the embodiments of the present application.
  • the computer program instructions enable the computer to execute the corresponding processes implemented by the second node in the various methods of the embodiments of the present application. For the sake of brevity, they are not repeated here.
  • the computer program product can be applied to the first network element in the embodiment of the present application, and the computer program instructions enable the computer to execute the corresponding processes implemented by the first network element in the various methods of the embodiment of the present application. For the sake of brevity, they are not repeated here.
  • the embodiment of the present application also provides a computer program.
  • the computer program can be applied to the first node in the embodiments of the present application.
  • the computer program runs on a computer, the computer executes the corresponding processes implemented by the first node in the various methods of the embodiments of the present application. For the sake of brevity, they are not repeated here.
  • the computer program can be applied to the second node in the embodiments of the present application.
  • the computer program runs on a computer
  • the computer executes the corresponding processes implemented by the second node in the various methods of the embodiments of the present application. For the sake of brevity, they are not repeated here.
  • the computer program can be applied to the first network element in the embodiment of the present application.
  • the computer program runs on a computer, the computer executes the corresponding processes implemented by the first network element in the various methods of the embodiment of the present application. For the sake of brevity, they are not repeated here.
  • the disclosed systems, devices and methods can be implemented in other ways.
  • the device embodiments described above are only schematic.
  • the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.
  • Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
  • each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.
  • the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art.
  • the computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application.
  • the aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Software Systems (AREA)
  • Databases & Information Systems (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本申请实施例提供一种通信的方法，该方法包括：接收来自第一网元的第一信息，第一信息用于指示符合执行VFL任务的条件的至少一个第二节点；从至少一个第二节点中，确定参与VFL任务的至少一个参与节点。在该方法中，发起VFL任务的第一节点可接收来自第一网元的第一信息，并可通过第一信息获知符合执行该VFL任务的条件的至少一个第二节点，以确保第一节点能够发现合适的节点来进行VFL任务。

Description

一种通信的方法、装置、设备、芯片和存储介质 技术领域
本申请实施例涉及通信技术领域，具体涉及一种通信的方法、装置、设备、芯片和存储介质。
背景技术
纵向联邦学习(Vertical Federated Learning，VFL)可在满足数据隐私、安全和监管需求的前提下，让人工智能系统高效、准确地使用多节点的本地数据，从而在保证隐私安全的前提下，打破数据孤岛，实现跨域的多节点数据共享。
然而，目前尚未有方案揭示现有的网络架构(如5G网络架构)如何支持跨域的纵向联邦学习。
发明内容
本申请实施例提供一种通信的方法、装置、设备、芯片和存储介质。
第一方面，本申请实施例提供了一种通信的方法，应用于第一节点，该方法包括：接收来自第一网元的第一信息，第一信息用于指示符合执行VFL任务的条件的至少一个第二节点；从至少一个第二节点中，确定参与VFL任务的至少一个参与节点。
第二方面，本申请实施例提供了一种通信的方法，应用于第二节点，该方法包括：向第一网元发送第二节点的能力信息，第二节点的能力信息用于确定第二节点是否符合执行纵向联邦学习VFL任务的条件。
第三方面，本申请实施例提供了一种通信的方法，应用于第一网元，该方法包括：获取至少一个第二节点的能力信息，第二节点的能力信息用于确定第二节点是否符合执行纵向联邦学习VFL任务的条件。
第四方面，本申请实施例提供了一种通信的装置，该装置包括：第一接收模块，被配置为接收来自第一网元的第一信息，第一信息用于指示符合执行纵向联邦学习VFL任务的条件的至少一个第二节点；确定模块，被配置为从至少一个第二节点中，确定参与VFL任务的至少一个参与节点。
第五方面，本申请实施例提供了一种通信的装置，该装置包括：第一发送模块，被配置为向第一网元发送该装置的能力信息，该装置的能力信息用于确定该装置是否符合执行纵向联邦学习VFL任务的条件。
第六方面，本申请实施例提供了一种通信的装置，该装置包括：获取模块，被配置为获取至少一个第二节点的能力信息，第二节点的能力信息用于确定第二节点是否符合执行纵向联邦学习VFL任务的条件。
第七方面，本申请实施例提供了一种通信设备，该通信设备包括存储器和处理器；其中，存储器，用于存储计算机可执行指令；处理器，与该存储器连接，用于通过执行该计算机可执行指令，实现如第一方面至第三方面中任一方面所述的方法。
第八方面，本申请实施例提供了一种芯片。该芯片包括：处理器，用于从存储器中调用并运行计算机程序，使得安装有该芯片的设备执行如第一方面至第三方面中任一方面所述的方法。
第九方面，本申请实施例提供了一种计算机可读存储介质，该计算机可读存储介质存储有计算机程序，计算机程序被至少一个处理器执行时实现如第一方面至第三方面中任一方面所述的方法。
本申请实施例中，第一节点可接收来自第一网元的第一信息，第一信息用于指示符合执行VFL任务的条件的至少一个第二节点；进而，第一节点可从至少一个第二节点中，确定参与VFL任务的至少一个参与节点。如此，可确保第一节点能够发现合适的节点来进行VFL任务。
附图说明
此处所说明的附图用来提供对本申请的进一步理解，构成本申请的一部分，本申请的示意性实施例及其说明用于解释本申请，并不构成对本申请的不当限定。在附图中：
图1是本申请实施例的一个应用场景的示意图；
图2是VFL训练过程的一例示意图；
图3是VFL推断过程的一例示意图；
图4是5G网络架构的一例示意图；
图5是本申请实施例提供的一种通信的方法的流程示意图；
图6是本申请实施例提供的另一种通信的方法的流程示意图；
图7是本申请实施例提供的通信方法的一种可能的实现流程示意图一；
图8是本申请实施例提供的通信方法的一种可能的实现流程示意图二；
图9是本申请实施例提供的通信方法的一种可能的实现流程示意图三；
图10是本申请实施例提供的一种通信的装置的结构示意图一；
图11是本申请实施例提供的一种通信的装置的结构示意图二；
图12是本申请实施例提供的一种通信的装置的结构示意图三；
图13是本申请实施例提供的一种通信设备示意性结构图；
图14是本申请实施例的芯片的示意性结构图；
图15是本申请实施例提供的一种通信系统的示意性框图。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
图1是本申请实施例的一个应用场景的示意图。
如图1所示，通信系统100可以包括终端设备110和网络设备120。网络设备120可以通过空口与终端设备110通信。终端设备110和网络设备120之间支持多业务传输。
应理解，本申请实施例仅以通信系统100进行示例性说明，但本申请实施例不限定于此。也就是说，本申请实施例的技术方案可以应用于各种通信系统，例如：长期演进(Long Term Evolution，LTE)系统、LTE时分双工(Time Division Duplex，TDD)、通用移动通信系统(Universal Mobile Telecommunication System，UMTS)、物联网(Internet of Things，IoT)系统、窄带物联网(Narrow Band Internet of Things，NB-IoT)系统、增强的机器类型通信(enhanced Machine-Type Communications，eMTC)系统、5G通信系统(也称为新无线(New Radio，NR)通信系统)，或未来的通信系统(如6G通信系统)等。
在图1所示的通信系统100中，网络设备120可以是与终端设备110通信的接入网设备。接入网设备可以为特定的地理区域提供通信覆盖，并且可以与位于该覆盖区域内的终端设备110(例如UE)进行通信。
网络设备120可以是长期演进(Long Term Evolution，LTE)系统中的演进型基站(Evolutional Node B，eNB或eNodeB)，或者是下一代无线接入网(Next Generation Radio Access Network，NG RAN)设备，或者是NR系统中的基站(gNB)，或者是云无线接入网络(Cloud Radio Access Network，CRAN)中的无线控制器，或者该网络设备120可以为中继站、接入点、车载设备、可穿戴设备、集线器、交换机、网桥、路由器，或者未来演进的公共陆地移动网络(Public Land Mobile Network，PLMN)中的网络设备等。
终端设备110可以是任意终端设备，其包括但不限于与网络设备120或其它终端设备采用有线或者无线连接的终端设备。
例如，所述终端设备110可以指接入终端、用户设备(User Equipment，UE)、用户单元、用户站、移动站、移动台、远方站、远程终端、移动设备、用户终端、终端、无线通信设备、用户代理或用户装置。接入终端可以是蜂窝电话、无绳电话、会话启动协议(Session Initiation Protocol，SIP)电话、IoT设备、卫星手持终端、无线本地环路(Wireless Local Loop，WLL)站、个人数字处理(Personal Digital Assistant，PDA)、具有无线通信功能的手持设备、计算设备或连接到无线调制解调器的其它处理设备、车载设备、可穿戴设备、5G网络中的终端设备或者未来演进网络中的终端设备等。
终端设备110可以用于设备到设备(Device to Device，D2D)的通信。
无线通信系统100还可以包括与网络设备120进行通信的核心网设备130，该核心网设备130可以是5G核心网(5G Core，5GC)设备。
图1示例性地示出了一个网络设备、一个核心网设备和两个终端设备，可选地，该无线通信系统100可以包括多个网络设备并且每个网络设备的覆盖范围内可以包括其它数量的终端设备，本申请实施例对此不做限定。
需要说明的是，图1只是以示例的形式示意本申请所适用的系统，当然，本申请实施例所示的方法还可以适用于其它系统。此外，本文中术语“系统”和“网络”在本文中常被可互换使用。本文中术语“和/或”，仅仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。另外，本文中字符“/”，一般表示前后关联对象是一种“或”的关系。还应理解，在本申请的实施例中提到的“指示”可以是直接指示，也可以是间接指示，还可以是表示具有关联关系。举例说明，A指示B，可以表示A直接指示B，例如B可以通过A获取；也可以表示A间接指示B，例如A指示C，B可以通过C获取；还可以表示A和B之间具有关联关系。还应理解，在本申请的实施例中提到的“对应”可表示两者之间具有直接对应或间接对应的关系，也可以表示两者之间具有关联关系，也可以是指示与被指示、配置与被配置等关系。还应理解，在本申请的实施例中提到的“预定义”或“预定义规则”可以通过在设备(例如，包括终端设备和网络设备)中预先保存相应的代码、表格或其他可用于指示相关信息的方式来实现，本申请对于其具体的实现方式不做限定。比如预定义可以是指协议中定义的。还应理解，本申请实施例中，所述“协议”可以指通信领域的标准协议，例如可以包括LTE协议、NR协议以及应用于未来的通信系统中的相关协议，本申请对此不做限定。
为了生成更符合用户需求的人工智能(Artifact Intelligence，AI)模型，模型的训练需要用户更多维度的数据。在实际场景中，用户的数据可能分布在终端、基站、核心网和第三方OTT(Over the Top)应用服务器等各个节点上，如果可以通过联合各节点的同一用户的不同特征数据进行模型训练，将会极大地提升模型的训练效果，对模型的训练具有重要的意义。但是，多节点的多域数据共享，会对数据隐私带来极大的挑战。为此，可采用VFL方法。VFL可在满足数据隐私、安全和监管需求的前提下，让人工智能系统可以高效、准确地使用多节点的本地数据，从而在保证隐私安全的前提下，打破数据孤岛，实现跨域的多节点数据共享。然而，目前尚未有方案揭示现有的网络架构如何支持跨域的VFL。
鉴于此，本申请提供一种通信的方法，在该方法中，第一节点可接收来自第一网元的第一信息，并可通过第一信息获知符合执行该VFL任务的条件的至少一个第二节点，以确保第一节点能够发现合适的节点来进行VFL任务。
为便于理解本申请实施例的技术方案，以下对本申请实施例的相关技术进行说明，以下相关技术作为可选方案与本申请实施例的技术方案可以进行任意结合，其均属于本申请实施例的保护范围。
1、纵向联邦学习(VFL)
图2是VFL训练过程的一例示意图。如图2所示，节点A和节点B为不同域中的节点，且节点A与节点B之间无原始数据交换。为了能够采用多节点的数据进行模型训练，VFL的训练过程可包括以下步骤：
S201，加密样本对齐。
VFL适用于参与者的训练样本ID重叠较多，而数据特征重叠较少的情况。在VFL中，需对参与者的样本进行对齐，从而在不增加样本ID的情况下，增加每个样本的特征维度。例如，某个区域的UE在通信系统的不同节点产生了不同的特征数据，其中，该UE的ID即为样本ID，在该场景中，需将该UE在不同节点产生的特征数据进行对齐(关联)。
S202，加密模型训练。
在样本对齐后，可对对齐的样本进行模型加密训练，进而基于节点A得到的模型A和节点B得到的模型B，训练得到性能更优的联邦模型。作为示例，步骤202可包括：
S1，分发公钥。
如图2所示，可由第三方协调者C向节点A和节点B发送公钥，用于加密需要传输的数据。加密方式例如可采用同态加密算法。在同态加密算法中，对两个样本m1和m2的和进行同态加密，即等于对m1的同态加密和对m2的同态加密之和，对样本m和一个常数相乘的同态加密，即等于对该样本m的同态加密再乘以该常数。
S2，交换中间结果。
在VFL中，拥有样本标签的一方可作为主动方(或称为需求方)，如图2中的节点B。节点A可作为被动方(或称为数据提供方)，被动方不具有样本标签。在该步骤中，节点A和节点B可分别利用自身的本地数据进行计算，得到模型的中间结果。其中，节点A可将得到的中间结果加密后发送给节点B，进而节点B可根据自身标签以及节点A和节点B的模型输出结果(中间结果)，计算得到模型整体的输出误差，并将该输出误差加密后发送给节点A。
S3，计算梯度。
在该步骤中，节点A和节点B可分别根据S2中的输出误差，计算得到各自加密后的梯度，并将计算结果添加掩码后发送给协调者C。
S4，更新模型。
协调者C解密节点A和节点B发送的梯度后，可将解密后的梯度分别回传给节点A和节点B。从而，节点A和节点B去除梯度的掩码后，可根据得到的梯度更新各自的模型。
图3是VFL推断过程的一例示意图。如图3所示，VFL的推断过程可包括：
S301，发送模型推断请求。
协调者C可分别向节点A和节点B发送模型推断请求，该模型推断请求可包括节点A和节点B需采用的模型的ID，以指示节点A和节点B需要采用的模型。
S302，节点A和节点B计算模型结果，并加密。
在该步骤中，节点A和节点B可分别根据自身数据和本地存储的模型进行计算，得到模型的中间结果，并对该中间结果进行加密。
S303，节点A和节点B向协调者C发送加密后的中间结果。
S304，协调者C聚合来自各节点的加密后的中间结果，并解密。
在该步骤中，协调者C可对节点A和节点B的加密后的中间结果进行聚合，得到加密的模型推断结果。进而，协调者C可对该模型推断结果进行解密，并可将解密后的推断结果发送给需求方节点B。
2、第五代(5th Generation，5G)网络架构
在5G网络架构中，一个重要的特征为“服务化架构”。其中，核心网网元(服务提供者)可提供特定的服务，并可通过定义好的应用程序接口(Application Programming Interface，API)供其他网元(消费者)调用。
应理解，本申请实施例中的“核心网网元”还可以称为“核心网网络功能(Network Function，NF)”。
图4是5G网络架构的一例示意图。如图4所示，该网络架构如可以包括UE、接入网(Access Network，AN)/无线接入网(Radio Access Network，RAN)以及核心网网元。其中，核心网网元包括：用户面功能(User Plane Function，UPF)、数据网络(Date Network，DN)、会话管理功能(Session Management Function，SMF)、接入及移动性管理功能(Access and Mobility Management Function，AMF)、网络切片选择功能(Network Slice Selection Function，NSSF)、鉴权服务器功能(Authentication Server Function，AUSF)、网络开放功能(Network Exposure Function，NEF)、网络存储功能(Network Function Repository Function，NRF)、策略控制功能(Policy Control Function，PCF)、统一数据管理(Unified Data Management，UDM)和应用功能(Application Function，AF)。
其中，UE与基站进行接入层(Access Stratum，AS)连接，用于交互接入层消息及无线数据传输。UE与AMF进行非接入层(Non-Access Stratum，NAS)连接，用于交互NAS消息。AMF负责对UE移动性的管理，SMF负责对UE的会话进行管。AMF除了负责对移动终端进行移动性管理之外，还负责将会话管理相关消息在UE和SMF之间进行转发。PCF负责制定对UE的移动性管理、会话管理、计费等相关的策略。UPF与基站及外部数据网络相连并进行数据传输。
此外，5G网络还在核心网中增加了网络数据分析功能(Network Data Analytics Function，NWDAF)，该功能可从核心网各个网元、网管系统等处收集数据并进行大数据统计、分析或智能化的数据分析，从而得出网络侧的分析和/或预测数据，进而辅助各个网元根据数据分析结果对UE接入进行更有效的控制。
下面对图4中示出的各网元做简单介绍。
1)UE：可以称终端设备、接入终端、用户单元、用户站、移动站、移动台、远方站、远程终端、移动设备、用户终端、终端、无线通信设备、用户代理或用户装置等。本申请实施例对UE所采用的具体技术和具体设备形态不做限定。
2)(无线)接入网((Radio)Access Network，(R)AN)设备：可以为特定区域的授权用户提供接入通信网络的功能，具体可以包括第三代合作伙伴计划(3rd Generation Partnership Project，3GPP)网络中无线网络设备也可以包括非3GPP(Non-3GPP)网络中的接入点。本申请实施例对(R)AN设备所采用的具体技术和具体设备形态不做限定。
3)AMF：主要用于接入控制、移动性管理、附着与去附着等功能。AMF还可作为N1信令(即N1接口的信令，为简洁，简称为N1信令)和N2信令(即N2接口的信令，为简洁，简称为N2信令)连接的锚点，为SMF提供N1/N2会话管理(session management，SM)消息的路由。AMF还可维护和管理UE的状态信息。
4)SMF：主要用于用户面网元选择，用户面网元重定向，终端设备的因特网协议(internet protocol，IP)地址分配，以及会话的建立、修改和释放及QoS控制。
5)UPF：主要用于用户面数据的接收和转发。例如，UPF可以从DN接收用户面数据，并通过AN设备将用户面数据发送给终端设备。UPF还可以通过AN设备从终端设备接收用户面数据，并转发到 DN。
6)PCF：主要用于指导网络行为的统一策略框架，为控制面网元(例如AMF，SMF等)提供策略规则信息等。
7)AF：主要用于向3GPP网络提供业务，如与PCF之间交互以进行策略控制等。
8)UDM：主要用于UE的签约数据管理，包括UE标识的存储和管理，UE的接入授权等。UDM还可为UE生成3GPP的认证凭证。UDM还可登记维护当前为UE服务的网元。
9)NEF：主要用于安全地向外部开放由3GPP网络功能提供的业务和能力等。
10)DN：主要用于为UE提供数据服务的运营商网络。例如，因特网(Internet)、第三方的业务网络、IP多媒体服务业务(IP Multi-Media Service，IMS)网络等。
11)AUSF：主要用于用户鉴权等，如UE接入网络时对UE进行安全认证。
12)NSSF：主要用于为UE选择切片实例集合，为UE确定AMF集合、允许的网络切片选择辅助信息(Network Slice Selection Assistance Information，NSSAI)(NSSAIs)。
在图4所示的网络架构中，各网元之间可以通过图中所示的接口通信，部分接口可以采用服务化接口的方式实现。如图4所示，UE和AMF之间可以通过N1接口进行通信。RAN和AMF之间可以通过N2接口进行通信。RAN和UPF之间可以通过N3接口进行通信，N3接口可以用于传输用户面的数据等。SMF和UPF之间可以通过N4接口进行通信。UPF和DN之间可以通过N6接口进行通信。UPF和UPF之间可以通过N9接口进行通信，N9接口可以用于传输UPF之间的上下行用户数据流等。其他接口与各网元之间的关系如图4中所示，为了简洁，这里不一详述。
应理解，上述命名仅为便于区分不同的功能而定义，不应对本申请构成任何限定。本申请并不排除在6G网络以及未来其它的网络中采用其他命名的可能。例如，在6G网络中，上述各个网元中的部分或全部可以沿用5G中的术语，也可能采用其他名称等。
还应理解，图4中的各个网元之间的接口名称只是一个示例，具体实现中接口的名称可能为其他的名称，本申请对此不作具体限定。此外，上述各个网元之间的所传输的消息(或信令)的名称也仅仅是一个示例，对消息本身的功能不构成任何限定。
上面对本申请中涉及到的术语做了简单说明，下文实施例中不再赘述。
为便于理解本申请实施例的技术方案，以下通过具体实施例详述本申请的技术方案。以上相关技术作为可选方案与本申请实施例的技术方案可以进行任意结合，其均属于本申请实施例的保护范围。本申请实施例包括以下内容中的至少部分内容。
应理解，本申请实施例中的第一节点表示发起VFL任务的节点，或者说，第一节点为VFL任务的发起方(或称为主动方/需求方)，也即，本申请实施例中的VFL任务可以是第一节点发起的。相应地，第二节点表示可参与VFL任务的作为非发起方的节点，或者说，第二节点可作为VFL任务的被动方(或称为数据提供方)。其中，第二节点的数量可以为一个，或者也可以为多个(两个或两个以上)，本申请实施例对此不予限定。
还应理解，本申请实施例的第一节点例如可以为终端设备、接入网网元、核心网网元(核心网NF)或AF；第二节点例如可包括以下至少之一：终端设备、接入网网元、核心网网元和AF；第一网元例如可以为NRF。其中，AF可以是可信任的AF(如运营商自有的AF)，或者还可以是不可信的第三方AF(如其他厂商的AF)。其中，AF是否为可信任的AF，可由核心网进行确认或识别。例如，核心网属于运营商A部署的核心网，则该核心网可将属于该运营商A的AF确认或识别为可信任的AF，并可将不属于该运营商A的AF确认或识别为不可信的第三方AF。
图5示出了本申请实施例提供的一种通信的方法，该方法可以包括：
S501，第一节点接收来自第一网元的第一信息，第一信息用于指示符合执行VFL任务的条件的至少一个第二节点。
在一种可能的方式中，该VFL任务是由第一节点发起的，故“符合执行VFL任务的条件”，还可以理解为，符合执行第一节点发起的VFL任务的条件。
作为一种实现方式，第一节点可通过向第一网元发送第二信息的方式，发起VFL任务，示例性地，第二信息可用于指示以下至少一项：
1)执行VFL任务，或者说，第一节点发起的任务类型为VFL任务。
2)执行VFL任务的时间段，或者说，第一节点所期望的执行VFL任务的时间段，如9:00至21:00。
3)作为VFL任务的发起方，或者说，第一节点期望作为VFL任务的发起方；
4)执行VFL任务的算力需求，如执行第一节点发起的VFL任务的最低算力需求；
5)执行VFL任务的服务区域，或者说，第一节点所期望的执行VFL任务的服务区域。
在一种可能的方式中，第一网元接收来自第一节点的第二信息后，可向第一节点发送第一信息，从 而向第一节点指示符合执行VFL任务的条件的至少一个第二节点，相应地，第一节点可接收来自第一网元的第一信息。如此，第一节点可通过第一信息获知符合执行VFL任务的条件的至少一个第二节点，从而可确保第一节点能够发现合适的节点来进行VFL任务。
作为示例，第一信息中例如可携带符合执行VFL任务的条件的至少一个第二节点的标识和/或地址信息。
在一些实施例中，第二节点是否符合执行VFL任务的条件是第一网元基于第二节点的能力信息确定的，或者说，第一网元可基于第二节点的能力信息确定第二节点是否符合执行VFL任务的条件。
示例性地，第二节点的能力信息可用于指示以下至少一项：
1)第二节点是否支持执行VFL任务。例如，第二节点的能力信息可包括第二节点是否支持执行VFL任务的指示，以指示该第二节点支持执行VFL任务或不支持执行VFL任务。
2)第二节点支持执行VFL任务的时间段。例如，第二节点的能力信息可用于指示该第二节点可在9:00至21:00期间执行VFL任务。
3)第二节点是否可作为VFL任务的发起方。例如，第二节点的能力信息可包括第二节点是否可作为VFL任务的发起方的指示，以指示该第二节点可作为VFL任务的发起方或不可作为VFL任务的发起方。
4)第二节点的算力。如第二节点所能提供的最高算力或算力范围。
5)第二节点的服务区域。其中，当第二节点为接入网网元时，第二节点的服务区域还可以理解为该接入网网元的信号覆盖区域。
举例来说，假设第二信息指示了第一节点所期望的执行VFL任务的时间段为9:00至21:00，所期望的执行VFL任务的服务区域为第一服务区域，那么，第一网元在接收到来自第一节点的第二信息之后，可将该第一网元收集或存储的各个第二节点的能力信息与第二信息的内容进行比对，确定出支持在9:00至21:00期间执行VFL任务，且服务区域包括第一服务区域的至少一个第二节点，并将该至少一个第二节点的信息(如标识和/或地址信息)携带于第一信息中发送至第一节点。
在一些实施例中，第二节点的能力信息例如可携带于第二节点的参数信息中，并由第二节点发送至第一网元，此时，该方法还可以包括：第二节点向第一网元发送第二节点的参数信息，相应地，第一网元可接收并存储该第二节点的参数信息，或者说，第一网元可收集或确定该第二节点的参数信息。其中，第二节点的参数信息例如可包括以下至少一项：第二节点的能力信息；第二节点的标识(ID)；第二节点的地址信息；以及，第二节点对应的应用(Application)的标识。
可以理解的是，当第二节点的数量为多个时，各个第二节点均可向第一网元发送该第二节点的参数信息，且每一第二节点的参数信息均包括上述参数信息中的至少一项。
在一些实施例中，各个第二节点可在第一节点发起VFL任务之前，将各个第二节点的参数信息发送至第一网元，从而，第一网元可在第一节点发起VFL任务之前收集各个第二节点的参数信息。进一步地，当第一节点发起VFL任务时，第一网元可从收集的各个第二节点的参数信息中获取各个第二节点的能力信息。进而，第一网元可基于各个第二节点的能力信息，确定符合执行VFL任务的条件的至少一个第二节点，并将该至少一个第二节点的信息(如标识和/或地址信息)携带于第一信息中，发送至第一节点。
S502，第一节点从至少一个第二节点中，确定参与VFL任务的至少一个参与节点。
示例性地，第一节点可从符合执行VFL任务的条件的至少一个第二节点中，确定参与VFL任务的至少一个参与节点。
作为一种实现方式，第一节点从符合执行VFL任务的条件的至少一个第二节点中，确定参与VFL任务的至少一个参与节点，包括：对于同意参与第一节点发起的VFL任务的第二节点，将该第二节点确定为参与该VFL任务的参与节点。也就是说，若某个第二节点符合执行VFL任务的条件，且同意参与第一节点发起的VFL任务，则可将该第一节点确定为参与该VFL任务的参与节点。
在一些实施例中，该方法还可以包括：第一节点向至少一个第二节点发送数据需求信息，其中，数据需求信息用于确定执行VFL任务的数据。相应地，该至少一个第二节点中的每一第二节点可接收来自第一节点的数据需求信息。
作为示例，数据需求信息例如可包括：执行VFL任务所需的数据的类型，和/或，所需的数据的有效时间。其中，数据的类型可以包括数据的格式(如8比特量化或24比特量化)，还可以包括数据内容的种类(如应用层数据、与设备电量相关的数据，与设备通信性能相关的数据等)。
在一种可能的方式中，第二节点在接收到来自第一节点的数据需求信息后，可基于本地配置或相关策略决定是否同意参与第一节点发起的VFL任务。例如，若第二节点本地存在满足相应数据需求的数据，则可同意参与该VFL任务；若第二节点本地不存在满足相应数据需求的数据，则可拒绝参与该VFL 任务。进而，该第二节点可将是否同意参与VFL任务的决定回复给第一节点。
示例性地，该第二节点可将是否同意参与VFL任务的决定携带于第三信息中回复给第一节点。此时，该方法还可以包括：第二节点向第一节点发送第三信息，第三信息用于指示该第二节点是否同意参与第一节点发起的VFL任务。相应地，第一节点可接收来自该第二节点的第三信息。基于第三信息，第一节点可获知该第二节点是否同意参与第一节点发起的VFL任务。
图6示出了本申请实施例提供的另一种通信的方法，该方法可以包括：
S601，第二节点向第一网元发送第二节点的能力信息。相应地，第一网元接收该第二节点的能力信息。
在该步骤中，第二节点可向第一网元发送第二节点的能力信息。其中，第二节点的能力信息可用于表征第二节点执行VFL任务的能力。基于第二节点的能力信息，第一网元可确定该第二节点是否符合执行VFL任务的条件。换句话说，第二节点是否符合执行VFL任务的条件是第一网元基于第二节点的能力信息确定的，或者说，第二节点的能力信息可用于第一网元确定第二节点是否符合执行VFL任务的条件。
示例性地，第二节点的能力信息可用于指示以下至少一项：
1)第二节点是否支持执行VFL任务。例如，第二节点的能力信息可包括第二节点是否支持执行VFL任务的指示，以指示该第二节点支持执行VFL任务或不支持执行VFL任务。
2)第二节点支持执行VFL任务的时间段。例如，第二节点的能力信息可用于指示该第二节点可在9:00至21:00期间执行VFL任务。
3)第二节点是否可作为VFL任务的发起方。例如，第二节点的能力信息可包括第二节点是否可作为VFL任务的发起方的指示，以指示该第二节点可作为VFL任务的发起方或不可作为VFL任务的发起方。
4)第二节点的算力。如第二节点所能提供的最高算力或算力范围。
5)第二节点的服务区域。其中，当第二节点为接入网网元时，第二节点的服务区域还可以理解为该接入网网元的信号覆盖区域。
在一些实施例中，第二节点的能力信息例如可携带于第二节点的参数信息中，并由第二节点发送至第一网元，此时，该方法还可以包括：第二节点向第一网元发送第二节点的参数信息，相应地，第一网元可接收并存储该第二节点的参数信息，或者说，第一网元可收集或确定该第二节点的参数信息。其中，第二节点的参数信息例如可包括以下至少一项：第二节点的能力信息；第二节点的标识(ID)；第二节点的地址信息；以及，第二节点对应的应用(Application)的标识。
可以理解的是，当第二节点的数量为多个时，每一第二节点的参数信息均包括上述参数信息中的至少一项。
一示例，第二节点为终端，终端的参数信息例如可包括以下至少一项：终端的能力信息、终端的标识、终端的地址信息(如IP地址)和终端的类型(如手机、车辆等)。其中，终端的能力信息可用于指示以下至少一项：终端是否支持执行VFL任务、终端支持执行VFL任务的时间段、终端是否可作为VFL任务的发起方，以及终端的算力。
另一示例，第二节点为接入网网元，接入网网元的参数信息例如可包括以下至少一项：接入网网元的能力信息、接入网网元的标识(如gNB ID)、接入网网元的地址信息。其中，接入网网元的能力信息可用于指示以下至少一项：接入网网元是否支持执行VFL任务、接入网网元支持执行VFL任务的时间段、接入网网元是否可作为VFL任务的发起方、接入网网元的算力，以及接入网网元的服务区域。
又一示例，第二节点为核心网网元，核心网网元的参数信息例如可包括以下至少一项：核心网网元的能力信息、核心网网元的地址信息和核心网网元的类型(如SMF、AMF等)。其中，核心网网元的能力信息可用于指示以下至少一项：核心网网元是否支持执行VFL任务、核心网网元支持执行VFL任务的时间段、核心网网元是否可作为VFL任务的发起方，核心网网元的算力，以及，核心网网元的服务区域。
再一示例，第二节点为AF，AF的参数信息例如可包括以下至少一项：AF的能力信息、AF的标识、AF的地址信息和AF对应的应用的标识。其中，AF的能力信息可用于指示以下至少一项：AF是否支持执行VFL任务、AF支持执行VFL任务的时间段、AF是否可作为VFL任务的发起方，以及AF的算力。
当第二节点为AF时，一种可能的情况，第二节点为可信任的AF，那么，第二节点可直接向第一网元发送第二节点的参数信息。另一种可能的情况，第二节点为不可信的第三方AF，那么，第二节点可通过NEF向第一网元发送(转发)第二节点的参数信息，相应地，第一网元可通过NEF接收该第二节点的参数信息。在该情况下，第二节点可将该第二节点的参数信息映射为某个事件(例如记为第二事 件)的参数信息，或者说，第二节点可基于第二节点的参数信息生成第二事件的参数信息，进而，第二节点可将第二事件的参数信息发送至NEF，再由NEF将该第二事件的参数信息转发至第一网元，从而实现将该AF的参数信息发送至第一网元。
也就是说，当第二节点为不可信的第三方AF时，第二节点向第一网元发送第二节点的参数信息，包括：第二节点通过NEF向第一网元发送第二事件的参数信息，其中，第二事件的参数信息是基于第二节点的参数信息生成的。示例性地，第二事件的参数信息可用于指示以下至少一项：AF是否支持执行第二事件；AF支持执行第二事件的时间段；AF是否可作为第二事件的发起方；AF的算力；以及，AF对应的应用的标识；其中，第二事件与VFL任务关联，换句话说，执行第二事件还可以理解为执行VFL任务。
作为一种实现方式，对于每一第二节点，第二节点的参数信息例如可通过第二注册请求信息携带，第二注册请求信息是第二节点向第一网元发送的注册请求信息。也就是说，第二节点可通过向第一网元发送第二注册请求信息的方式，将该第二节点的参数信息注册到第一网元中。
在一些实施例中，各个第二节点可在第一节点发起VFL任务之前，将各个第二节点的参数信息发送至第一网元，从而，第一网元可在第一节点发起VFL任务之前接收各个第二节点的参数信息。进一步地，当第一节点发起VFL任务时，该方法还可以包括：第一网元获取至少一个第二节点的能力信息。例如，第一网元可从接收的各个第二节点的参数信息中获取各个第二节点的能力信息。进而，对于每一第二节点，第一网元可基于该第二节点的能力信息，确定该第二节点是否符合执行该VFL任务的条件。
在一些实施例中，第一节点可将该第一节点的参数信息发送(注册)到第一网元中，此时该方法还可包括：第一节点向第一网元发送第一节点的参数信息。其中，第一节点的参数信息例如可包括以下至少一项：第一节点的能力信息；第一节点的标识；第一节点的地址信息；以及，第一节点对应的应用的标识。
示例性地，第一节点的能力信息可用于指示以下至少一项：第一节点是否支持执行VFL任务；第一节点支持执行VFL任务的时间段；第一节点是否可作为VFL任务的发起方；第一节点的算力；以及，第一节点的服务区域。
应理解，第一节点的各个参数信息的具体含义与第二节点对应的各个参数信息的含义相同，可相互参考，这里不再赘述。
当第一节点为AF时，一种可能的情况，第一节点为可信任的AF，那么，第一节点可直接向第一网元发送第一节点的参数信息。另一种可能的情况，第一节点为不可信的第三方AF，那么，第一节点可通过NEF向第一网元发送(转发)第一节点的参数信息，相应地，第一网元可通过NEF接收该第一节点的参数信息。在该情况下，第一节点可将该第一节点的参数信息映射为某个事件(例如记为第一事件)的参数信息，或者说，第一节点可基于第一节点的参数信息生成第一事件的参数信息，进而，第一节点可将第一事件的参数信息发送至NEF，再由NEF将该第一事件的参数信息转发至第一网元，从而实现将该AF的参数信息发送至第一网元。
也就是说，当第一节点为不可信的第三方AF时，第一节点向第一网元发送第一节点的参数信息，包括：第一节点通过NEF向第一网元发送第一事件的参数信息，第一事件的参数信息是基于第一节点的参数信息生成的。示例性地，第一事件的参数信息用于指示以下至少一项：AF是否支持执行第一事件；AF支持执行第一事件的时间段；AF是否可作为第一事件的发起方；AF的算力；以及，AF对应的应用的标识。其中，第一事件与VFL任务关联，换句话说，执行第一事件还可以理解为执行VFL任务。
作为一种实现方式，第一节点的参数信息例如可通过第一注册请求信息携带，第一注册请求信息是第一节点向第一网元发送的注册请求信息。也就是说，第一节点可通过向第一网元发送第一注册请求信息的方式，将该第一节点的参数信息注册到第一网元中。
S602，第一网元向第一节点发送第一信息。相应地，第一节点接收该第一信息。
其中，第一信息可用于指示符合执行VFL任务的条件的至少一个第二节点，该VFL任务是第一节点发起的。
在S602中，第一网元可向第一节点发送第一信息，以指示第一节点符合执行VFL任务的条件的至少一个第二节点。如此，第一节点可通过第一信息获知符合执行该VFL任务的条件的至少一个第二节点，从而可确保第一节点能够发现合适的节点来进行VFL任务。
作为示例，第一信息中例如可携带符合执行VFL任务的条件的至少一个第二节点的标识和/或地址信息。
在一些实施例中，在第一网元向第一节点发送第一信息之前，该方法还可以包括：第一节点向第一网元发送第二信息，第二信息用于请求第一信息；相应地，第一网元可接收来自第一节点的第二信息，
也就是说，当第一节点需要发现符合执行VFL任务的条件的至少一个第二节点时，第一节点可向第一网元发送第二信息，以请求第一信息。第一网元接收到来自第一节点的第二信息后，可基于获取的各个第二节点的能力信息，确定符合执行第一节点发起的VFL任务的条件的至少一个第二节点，并向第一节点反馈第一信息，从而向第一节点指示该至少一个第二节点。
当第一节点为AF时，一种可能的情况，第一节点为可信任的AF，那么，第一节点可直接向第一网元发送第二信息。另一种可能的情况，第一节点为不可信的第三方AF，那么，第一节点可通过NEF向第一网元发送(转发)第二信息，相应地，第一网元可通过NEF接收来自第一节点的第二信息。
示例性地，第二信息可用于指示以下至少一项：
1)执行VFL任务，或者说，第一节点发起的任务类型为VFL任务。
2)执行VFL任务的时间段，或者说，第一节点所期望的执行VFL任务的时间段，如9:00至21:00。
3)作为VFL任务的发起方，或者说，第一节点期望作为VFL任务的发起方；
4)执行VFL任务的算力需求，如执行第一节点发起的VFL任务的最低算力需求；
5)执行VFL任务的服务区域，或者说，第一节点所期望的执行VFL任务的服务区域。
举例来说，假设第二信息指示了第一节点所期望的执行VFL任务的时间段为9:00至21:00，所期望的执行VFL任务的服务区域为第一服务区域，那么，第一网元在接收到来自第一节点的第二信息之后，可将该第一网元中存储的各个第二节点的能力信息与第二信息的内容进行比对，确定出支持在9:00至21:00期间执行VFL任务，且服务区域包括第一服务区域的至少一个第二节点，并将该至少一个第二节点的信息(如标识和/或地址信息)携带于第一信息中发送至第一节点。
S603，第一节点确定参与VFL任务的至少一个参与节点。
示例性地，第一节点可从符合执行VFL任务的条件的至少一个第二节点中，确定参与VFL任务的至少一个参与节点。
应理解，本申请实施例中提及的“符合执行VFL任务的条件”，指的是符合执行第一节点发起的VFL任务的条件。为简洁，下文中将“符合执行VFL任务的条件的至少一个第二节点”简称为“至少一个第二节点”，也就是说，下文中所述的“至少一个第二节点”，在不引起歧义的情况下，指的是“符合执行VFL任务的条件的至少一个第二节点”。
作为一种实现方式，从至少一个第二节点中，确定参与VFL任务的至少一个参与节点，包括：对于同意参与第一节点发起的VFL任务的第二节点，将该第二节点确定为参与该VFL任务的参与节点。也就是说，若某个第二节点符合执行VFL任务的条件，且同意参与第一节点发起的VFL任务，则可将该第一节点确定为参与该VFL任务的参与节点。
在一些实施例中，对于至少一个第二节点中的每个第二节点，该方法还可以包括：第二节点向第一节点发送第三信息，第三信息用于指示该第二节点是否同意参与第一节点发起的VFL任务。相应地，第一节点可接收来自该至少一个第二节点的第三信息。基于第三信息，第一节点可获知各个第二节点是否同意参与第一节点发起的VFL任务。
在一些实施例中，该方法还可以包括：第一节点向至少一个第二节点发送数据需求信息，其中，数据需求信息用于确定执行VFL任务的数据。相应地，该至少一个第二节点中的每一第二节点可接收来自第一节点的数据需求信息。
作为示例，数据需求信息例如可包括：执行VFL任务所需的数据的类型，和/或，所需的数据的有效时间。其中，数据的类型可以包括数据的格式(如8比特量化或24比特量化)，还可以包括数据内容的种类(如应用层数据、与设备电量相关的数据，与设备通信性能相关的数据等)。
举例来说，假设数据需求信息中，执行VFL任务所需的数据的类型为8比特量化的应用层数据，所需的数据的有效时间为第一时间段，那么，接收到该数据需求信息的第二节点可基于该数据需求信息，将在第一时间段内有效、类型为8比特量化的应用层的本地数据确定为执行VFL任务的数据。
在一种可能的方式中，第二节点在接收到来自第一节点的数据需求信息后，可基于本地配置或相关策略决定是否同意参与第一节点发起的VFL任务。例如，若第二节点本地存在满足相应数据需求的数据，则可同意参与该VFL任务；若第二节点本地不存在满足相应数据需求的数据，则可拒绝参与该VFL任务。进而，该第二节点可将是否同意参与VFL任务的决定通过第三信息回复给第一节点。
当第一节点为AF时，一种可能的情况，第一节点为可信任的AF，那么，第一节点可直接向至少一个第二节点发送数据需求信息。另一种可能的情况，第一节点为不可信的第三方AF，那么，第一节点可通过NEF向至少一个第二节点发送(转发)数据需求信息，相应地，第二节点可通过NEF接收来自第一节点的数据需求信息。
S604，第一节点向参与节点发送VFL模型。相应地，参与节点接收该VFL模型。
其中，参与节点还可以理解为被确定为参与节点的第二节点，或者还可以理解为，参与第一节点发 起的VFL任务的第二节点。当第一节点为AF时，第一节点向参与节点发送VFL模型，还可以理解为，第一节点对应的服务器(也即AF对应的服务器)向参与节点发送VFL模型。
在S603中，第一节点可从至少一个第二节点中，确定参与VFL任务的至少一个参与节点，其中，每一个参与节点可对应一个VFL模型。在S604中，第一节点可向各个参与节点发送对应的VFL模型，该VFL模型可用于执行VFL任务。在一种可能的方式中，该VFL模型例如可包括对应的参与节点的信息(如参与节点的标识和/或地址信息)。
其中，某个参与节点对应的VFL模型，还可以理解为，该参与节点所需的VFL模型。该VFL模型可作为该参与节点执行VFL任务时的初始模型。作为一种实现方式，该参与节点所需的VFL模型可由第一节点生成。例如，参与节点包括终端设备、接入网网元和核心网网元，那么第一节点可为终端设备生成该终端设备所需的VFL模型，并可为接入网网元生成该接入网网元所需的VFL模型，以及为核心网网元生成该核心网网元所需的VFL模型。进而，第一节点可将各个参与节点所需的VFL模型发送至相应的参与节点。
在一些场景中，第一节点不能直接向相应的参与节点发送VFL模型，对于该情况，第一节点可通过某个(或某几个)中间设备将该参与节点对应的VFL模型转发至该参与节点。
举例来说，当第一节点为AF、参与节点为终端设备或接入网设备时，AF无法直接向该终端设备或接入网设备发送VFL模型，对于该情况，第一节点向参与节点发送对应的VFL模型，可包括：第一节点通过第二网元向终端设备或接入网网元发送对应的VFL模型。其中，第二网元可以是一个网元，也可以包括两个或两个以上网元，本申请实施例对此不予限定。作为一种实现方式，第二网元接收到VFL模型后，可基于该VFL模型中携带的参与节点的标识和/或地址信息，将该VFL模型转发给相应的参与节点。例如，VFL模型#1中携带终端设备的标识和地址信息，那么，第一节点可基于该终端设备的标识和地址信息，将VFL模型#1发送至该终端设备；又例如，VFL模型#2中携带接入网网元的标识和地址信息，那么，第一节点可基于该接入网网元的标识和地址信息，将VFL模型#2发送至该接入网网元。
需要说明的是，在一些场景中，若第一节点可直接向相应的参与节点发送VFL模型，而无需通过中间设备进行转发，那么VFL模型中也可以不需要携带该参与节点的信息(如标识和/或地址信息)。
在一些实施例中，该方法还可以包括：第一节点向至少一个参与节点发送相同的关联标识，该关联标识可用于标识同一VFL任务的不同执行结果，还可以用于聚合同一VFL任务的不同执行结果。相应地，第二节点可接收来自第一节点的关联标识。作为一种实现方式，第一节点可将VFL模型和该关联标识通过同一条信息发送至对应的参与节点。
其中，同一VFL任务的不同执行结果，例如可包括不同节点(如第一节点；又如被确定为参与节点的第二节点)在执行该同一VFL任务时，所产生的执行结果。
在一种可能的方式中，执行结果可包括模型训练结果和模型推理结果，此时，关联标识可包括模型关联标识和结果关联标识，其中，模型关联标识可用于标识和/或聚合同一VFL任务的不同模型训练结果，结果关联标识可用于标识和/或聚合同一VFL任务的不同模型推理结果。
在本申请实施例中，各个参与节点接收到对应的VFL模型时，可将接收的VFL模型作为执行VFL任务的初始模型，将S603中基于数据需求信息确定的执行VFL任务的数据作为输入数据。同时，第一节点可生成(获取)该第一节点对应的VFL模型，并可确定该第一节点执行VFL任务的本地数据。进一步地，各个参与节点和第一节点可基于各自的VFL模型和各自的本地数据，共同执行第一节点发起的VFL任务。
在一些实施例中，该VFL任务执行完毕后，该方法还可以包括：各个参与节点(或者说参与了VFL任务的第二节点)向第一节点发送该参与节点执行VFL任务的执行结果，其中，执行结果中可携带关联标识。需要说明的是，当第一节点为AF时，向第一节点发送参与节点执行VFL任务的执行结果，还可以理解为，向第一节点对应的服务器(也即AF对应的服务器)发送参与节点执行VFL任务的执行结果。
一种可能的情况，参与节点可直接将该参与节点执行VFL任务的执行结果发送至第一节点。另一种可能的情况，某个参与节点(以下记为参与节点#1)无法直接将执行结果发送至第一节点，对于该情况，该参与节点#1可先将该参与节点#1的执行结果发送至某个中间节点(如参与节点#2)，进而可由该中间节点将参与节点#1的执行结果转发至第一节点。
以中间节点为参与节点#2为例，在第一种实现方式中(以下记为实现方式#1)，参与节点#2在接收到参与节点#1的执行结果后，可将参与节点#1的执行结果发送至第一节点，同时，参与节点#2还可以将自身的执行结果发送至第一节点。其中，参与节点#1和参与节点#2的执行结果中，均携带相同的关联标识。
在第二种实现方式中(以下记为实现方式#2)，参与节点#2还可以对该参与节点#2的执行结果和接收的参与节点#1的执行结果进行聚合，进而将得到的聚合结果发送至第一节点。作为示例，参与节点#1和参与节点#2的执行结果中可携带相同的关联标识，这样，参与节点#2在进行执行结果的聚合时，可将具有相同关联标识的执行结果(参与节点#1和参与节点#2的执行结果)进行聚合，从而确保聚合的是同一VFL任务的执行结果，进而确保执行结果的正确聚合。其中，得到的聚合结果中也可携带该关联标识。
在第三种实现方式中(以下记为实现方式#3)，参与节点#2除了可接收参与节点#1的执行结果之外，还可以接收其他参与节点(以下记为参与节点#3)的执行结果，并可对参与节点#1、参与节点#2和参与节点#3的执行结果进行聚合，进而将得到的聚合结果发送至第一节点。作为示例，参与节点#1、参与节点#2和参与节点#3的执行结果中可携带相同的关联标识，这样，参与节点#2在进行执行结果的聚合时，可将具有相同关联标识的执行结果(参与节点#1、参与节点#2和参与节点#3的执行结果)进行聚合，从而确保聚合的是同一VFL任务的执行结果，进而确保执行结果的正确聚合。其中，得到的聚合结果中也可携带该关联标识。
也就是说，对于某个参与VFL任务的第二节点(如参与节点#2)，在上述实现方式#2和实现方式#3中，该方法还可包括：接收至少一个参与节点(如参与节点#1和/或参与节点#3)执行VFL任务的执行结果(例如记为第二执行结果)；向所述第一节点发送聚合结果，聚合结果基于至少一个参与节点执行VFL任务的执行结果(也即至少一个第二执行结果)和该第二节点执行VFL任务的执行结果(例如记为第三执行结果)进行聚合得到，其中，该聚合结果中可携带关联标识。在一种可能的方式中，至少一个参与节点执行VFL任务的执行结果中的每个执行结果(也即每个第二执行结果)，以及该第二节点执行VFL任务的执行结果(也即第三执行结果)，均携带相同的关联标识，从而，该聚合结果可基于该相同的关联标识聚合得到的。
相应地，第一节点可接收该第一节点发起的VFL任务对应的执行结果(例如记为第一执行结果)，其中，第一执行结果包括：至少一个参与节点执行VFL任务的执行结果(例如通过上述实现方式#1得到的执行结果，例如记为第二执行结果)；和/或，至少一个聚合结果，至少一个聚合结果中的每个聚合结果基于至少两个参与节点执行VFL任务的执行结果(也即至少两个第二执行结果)进行聚合得到(例如通过上述实现方式#2或实现方式#3得到)。其中，至少一个参与节点执行VFL任务的执行结果中每个执行结果(也即每个第二执行结果)，和/或，至少一个聚合结果中每个聚合结果，均携带相同的关联标识，并且，该至少一个聚合结果中的每个聚合结果，可以是基于相同的关联标识聚合得到的。
在一些实施例中，该方法还可以包括：第一节点聚合具有相同关联标识的VFL任务对应的执行结果和第一节点执行VFL任务的执行结果。
作为示例，第一节点接收到该第一节点发起的VFL任务对应的执行结果后，可对该VFL任务对应的执行结果和该第一节点自身的执行结果进行聚合，得到最终的聚合结果。其中，第一节点接收的VFL任务对应的执行结果，以及该第一节点自身的执行结果，均可携带相同的关联标识，这样，第一节点在进行执行结果的聚合时，可将具有相同关联标识的执行结果(第一节点接收的VFL任务对应的执行结果和该第一节点自身的执行结果)进行聚合，如此，可确保聚合的是同一VFL任务的执行结果，进而确保执行结果的正确聚合。
上文结合图5和图6介绍了本申请实施例提供的一种通信的方法。为便于理解本申请的实施例，下面结合图7至图9介绍本申请实施例提供的通信方法的可能的实现流程。为便于说明，以下示例中假设，第一网元为NRF，第一节点为AF，第二节点包括UE、RAN和核心网NF(或者称为核心网网元)。其中，UE、RAN或核心网NF可以为一个，也可以为多个(两个或两个以上)，本申请实施例对此不予限定。
为了执行VFL，作为发起方的节点首先需要发现具备可执行VFL能力的节点，或者说符合执行VFL任务的条件的节点。为此，在开始执行VFL任务之前，各节点需要把各自的信息注册到NRF中。以AF作为发起方为例，UE、RAN和核心网网元，需将自己是否能够支持VFL的能力注册到NRF中，从而AF可通过NRF发现合适的节点以执行VFL任务。
图7是本申请实施例提供的通信方法的一种可能的实现流程示意图一，该实现流程可包括：
S701，UE向NRF发送注册请求。
UE可通过向NRF发送注册请求，将UE的信息注册到NRF中。作为示例，该注册请求中携带UE的文件(Profile)(对应前述实施例中终端的参数信息)，UE的文件包括但不限于以下至少一项：UE ID、UE能否支持VFL的能力、UE可以支持VFL的时间段、UE的类型、UE的地址信息(例如IP地址)、UE能否支持成为VFL发起方的指示、UE的算力信息。
S702，RAN向NRF发送注册请求。
RAN可通过向NRF发送注册请求，将RAN的信息注册到NRF中。作为示例，该注册请求中携带RAN的文件(对应前述实施例中接入网网元的参数信息)，RAN的文件包括但不限于以下至少一项：RAN的ID(例如gNB ID)、RAN能否支持VFL的能力、RAN可以支持VFL的时间段、RAN的地址信息、能否支持成为VFL发起方的指示、RAN的算力信息、RAN的服务区域(覆盖区域)信息。
S703，核心网NF向NRF发送注册请求。
核心网NF可通过向NRF发送注册请求，将核心网NF的信息注册到NRF中。作为示例，该注册请求中携带核心网NF的文件(对应前述实施例中核心网网元的参数信息)，核心网NF的文件包括但不限于以下至少一项：NF的类型、NF能否支持VFL的能力、NF可以支持VFL的时间段、NF的地址信息，NF能否支持成为VFL发起方的指示、NF的算力信息、NF的服务区域信息。
S704，AF向NRF发送注册请求。
AF可通过向NRF发送注册请求，将AF的信息注册到NRF中。作为示例，该注册请求中携带AF的文件(对应前述实施例中AF的参数信息)，AF的文件包括但不限于以下至少一项：AF ID、AF对应的应用(Application)ID、AF能否支持VFL的能力、AF可以支持VFL的时间段、AF的地址信息、AF能否支持成为VFL发起方的指、AF的算力信息。
需要说明的是，当AF为不可信的第三方AF时，AF可通过NEF将信息注册到NRF上。举例来说，AF可先将AF的文件发送至NEF，由NEF生成NEF文件，并由NEF将NEF文件注册到NRF中。NEF可将AF的信息映射为一个事件(Event)的信息，或者说映射为一个事件ID，从而，NEF可将该事件的信息(对应前述实施例中第一事件的参数信息)作为NEF的文件，注册到NRF中。作为示例，该事件的信息中包括但不限于以下至少一项：NEF ID、AF能否支持VFL的能力、AF对应的应用ID、该事件支持的时间段、AF能否支持成为该事件的发起方的指示、AF的算力信息。
应理解，本实施例对于S701至S704的执行顺序不予限定。
S705，NRF保存各节点的文件。
NRF收到各节点(UE、RAN、核心网NF和AF)的注册请求后，可保存各节点的文件，并将各节点标记为可用的有效节点。之后，NRF可向各节点发送注册响应，用于回复各节点注册被接受。NRF向各节点发送注册响应的步骤可包括S706至S709：
S706，NRF向核心网NF发送注册响应。
S707，NRF向RAN发送注册响应。
S708，NRF向UE发送注册响应。
S709，NRF向AF发送注册响应。
应理解，本实施例对于S706至S709的执行顺序不予限定。
通过图7所示的实现流程，各节点可将各自的信息注册到NRF中，以便于在执行VFL任务时，作为发起方的节点能够通过NRF发现合适的节点来进行VFL任务。
图8是本申请实施例提供的通信方法的一种可能的实现流程示意图二。作为一种实现方式，图8所示的实现流程可在图6所示的实现流程之后执行，也就是说，在图8所示的实现流程中，可认为各节点的信息已经注册到NRF中。如图8所示，该实现流程可包括：
S801，AF向NRF发送节点发现请求(对应前述实施例中的第二信息)。
以AF作为VFL任务的发起方为例，当AF想要发现可执行VFL任务的节点时，AF可向NRF发送节点发现请求。作为示例，节点发现请求中可包括但不限于以下信息中的至少一项：执行VFL的指示、想要执行的VFL的时间段、想要作为VFL发起方指示、想要执行VFL的服务区域。
在一些实施例中，当AF为第三方的AF时，AF可通过NEF向NRF发送节点发现请求，节点发现请求中包括的参数相同。
S802，NRF授权AF的节点发现请求。
在该步骤中，NFR可授权AF的节点发现请求。作为一种实现方式，NRF在授权AF的节点发现请求之前，可先基于该AF的文件信息判断该AF能否支持VFL，以及能否支持成为VFL发起方。进一步地，若该AF能够支持VFL，且能够支持成为VFL发起方，则授权该AF的节点发现请求。
S803，NRF向AF发送节点发现请求响应(对应前述实施例中的第一信息)。
由于NRF已经保存有相关节点的文件信息，故基于S801的节点发现请求中携带的信息，NRF可为AF确定满足条件的节点(如满足条件的UE、RAN和核心网NF)，并将满足条件的节点的信息(如节点的ID和/或地址信息)通过节点发现请求响应回复给AF。
其中，“满足条件的节点”可以理解为，满足节点发现请求中包括的条件的节点，或者还可以理解为，符合执行AF发起的VFL任务的条件的节点。为便于说明，以下假设UR、RAN和核心网NF均为满足条件的节点。
在一些实施例中，当AF为第三方的AF时，NRF可通过NEF向AF发送节点发现请求响应。
S804，AF向NEF发送VFL准备请求。
当AF为第三方AF时，AF可向NEF发送VFL准备请求，进而通过NEF向各节点转发该VFL准备请求，该VFL准备请求中可包括执行VFL任务所需的数据的类型和数据的有效时间等(对应前述实施例中的数据需求信息)。
S805，NEF向各节点发送VFL准备请求。
NEF接收到来自AF的VFL准备请求后，可将该VFL准备请求发送给满足条件的各节点(如图8所示，NEF可将VFL准备请求分别发送给UE、RAN和核心网NF)，以便于各节点基于VFL准备请求中携带的执行VFL任务所需的数据的类型和数据的有效时间，提前为执行VFL任务准备好所需的数据。
S806，各节点向NEF发送VFL准备请求回复消息(对应前述实施例中的第三信息)。
各节点收到VFL准备请求后，可基于本地配置确定是否加入该VFL任务，并将是否同意加入的决定回复给NEF，该决定例如可携带于VFL准备请求回复消息中，或者说，VFL准备请求消息可用于指示各节点是否同意加入AF发起的VFL任务。
S807，NEF进行节点选择。
NEF可基于各节点反馈的VFL准备请求回复消息，并基于AF的请求，进行VFL的节点选择。例如，NEF可选择出符合执行AF发起的VFL任务的条件，且同意加入该VFL任务的节点。
S808，NEF向AF发送节点选择结果。
在该步骤中，NEF可将S807中的节点选择结果返回给AF，其中，节点选择结果可包括选择的节点的信息，如节点的ID和/或地址信息，从而AF可将选择的节点确定为参与VFL任务的参与节点。
需要说明的是，S804至S808是以AF为第三方AF进行的示例性说明。当AF为可信的AF时，AF可直接将VFL准备请求发送至满足条件的各节点(如UE、RAN和核心网NF)，随后，满足条件的各节点可直接将VFL准备请求回复消息发送给AF，该VFL准备请求消息可用于指示各节点是否同意加入AF发起的VFL任务，进而，AF可将满足条件且同意加入VFL任务的节点确定为参与VFL任务的参与节点。
通过图8所示的实现流程，AF可通过向NRF发送节点发现请求，从而发现并确定合适的节点以进行VFL任务。
图9是本申请实施例提供的通信方法的一种可能的实现流程示意图三。作为一种实现方式，图9所示的实现流程可在图7所示的实现流程之后执行，也就是说，在图9所示的实现流程中，可认为AF已经确定了参与VFL任务的各节点。在以下实现流程中，假设AF确定的参与节点为UE、RAN和核心网NF。如图9所示，该实现流程可包括：
S901，AF对应的服务器向核心网NF发送各节点对应的VFL模型和关联标识。
其中，各节点对应的VFL模型可作为各节点执行VFL任务的初始模型。
在该步骤中，AF对应的服务器可将参与VFL任务的各节点(UE、RAN、核心网NF)所需的(对应的)VFL模型首先发送至核心网NF。其中，每个模型可携带对应的节点的信息。例如，UE所需的模型为模型#1，RAN所需的模型为模型#2，那么，模型#1中可携带UE的信息(如UE的ID和/或地址信息)，用于指示模型#1需发送给UE使用，模型#2中可携带RAN的信息(如RAN的ID和/或地址信息)，用于指示模型#2需发送给RAN使用。
在该步骤中，AF还可以向各节点发送相同的关联标识(关联ID)。由于VFL需在不同的节点上采用本地模型和数据进行模型训练和结果推理，并在发起方进行聚合，因此，为了便于关联各个节点产生的VFL执行结果(包括模型的训练结果和/或推理结果)，AF可向各个节点发送相同的关联标识，这样，各个节点在计算得到VFL执行结果后，可为该节点的VFL执行结果添加该关联标识，从而保证同一个VFL任务在不同节点产生的结果具有相同的关联标识，进而保证汇聚节点对各节点得到的VFL执行结果进行正确的聚合。其中，“汇聚节点”还可以理解为对至少两个节点的VFL执行结果进行聚合的节点。
作为一种实现方式，该关联标识例如可携带于各节点对应的VFL模型中。
S902，核心网NF向RAN发送RAN对应的VFL模型和关联标识。
核心网NF得到RAN所需的(对应的)VFL模型和关联标识后，可基于该VFL模型所携带的RAN的信息(如RAN的ID和/或地址信息)，将该VFL模型和关联标识发送至RAN。
S903，核心网NF向UE发送UE对应的VFL模型和关联标识。
核心网NF得到UE所需的(对应的)VFL模型和关联标识后，可基于该VFL模型所携带的UE的信息(如UE的ID和/或地址信息)，将该VFL模型和关联标识发送至UE。作为一种实现方式，VFL可将UE对应的VFL模型和关联标识通过RAN发送至UE。
在一些实施例中，AF对应的服务器还可以通过应用层直接将UE对应的VFL模型和关联标识发送 给UE。在该情况下，该实现流程法还包括S904：
S904，AF对应的服务器通过应用层向UE发送UE对应的VFL模型和关联标识。
可以理解的是，在执行S904的情况下，在S901中AF对应的服务器可以不需要向核心网NF发送UE对应的VFL模型，且在该情况下可以不需要执行S903。
S905，各节点执行VFL任务。
VFL模型下发完成后，各节点可基于关联标识，进行同一个任务的联邦学习。举例来说，在执行VFL任务的过程中，各节点可能需要交互中间结果(如梯度、损失等)，由于各节点产生的结果具有相同的关联标识，因此通过对具有相同关联标识的结果进行处理，可保证各节点进行的是同一个任务的联邦学习。
S906，UE向核心网NF发送UE执行VFL任务的执行结果。
UE在执行完VFL任务后，可将得到的执行结果发送至核心网NF，该执行结果中可携带关联标识。
S907，RAN向核心网NF发送RAN执行VFL任务的执行结果。
RAN在执行完VFL任务后，可将得到的执行结果发送至核心网NF，该执行结果中可携带关联标识。
在一些实施例中，该实现流程还包括S908：
S908，核心网NF聚合执行结果。
核心网NF接收来自UN和RAN执行VFL任务的执行结果后，可基于关联标识，对具有相同关联标识的执行结果(UN、RAN以及核心网NF自身的执行结果)进行聚合，进而在S909中将聚合得到的聚合结果发送至AF对应的服务器。
应理解，在一些实施例中，核心网NF接收来自UN和RAN的执行结果后，也可以不进行执行结果的聚合，而是直接将接收的执行结果(UE和RAN的执行结果)和该核心网NF产生的执行结果发送给AF对应的服务器。
S909，核心网NF向AF对应的服务器发送VFL任务对应的执行结果。
一种可能的情况，核心网NF接收来自UN和RAN执行VFL任务的执行结果后，对各节点的执行结果进行了聚合，那么在该步骤中，VFL任务对应的执行结果可以是对各节点的执行结果进行聚合得到的聚合结果，且该聚合结果中携带关联标识。
另一种可能的情况，核心网NF接收来自UN和RAN执行VFL任务的执行结果后，未对各节点的执行结果进行聚合，那么VFL任务对应的执行结果中可包括各节点的执行结果，如UN、RAN和核心网NF的执行结果，且各节点的执行结果中携带相同的关联标识。
需要说明的是，在一些实施例中，UE还可以通过应用层直接将UE的执行结果发送给AF对应的服务器，并在执行结果中携带关联标识，此时，该实现流程还可以包括S710：
S910，UE通过应用层向AF对应的服务器发送UE执行VFL任务的执行结果。
可以理解的是，在执行S910的情况下，可以不需要执行S906，且在S908中，核心网NF在对各节点的执行结果进行聚合时，可以不需要对UE的执行结果进行聚合。
S911，AF对应的服务器聚合执行结果。
在该步骤中，AF对应的服务器可聚合所有节点产生的结果，得到VFL最后的结果。例如，在核心网NF对UR、RAN和该核心网NF的执行结果进行了聚合的情况下，AF对应的服务器可基于关联标识，将核心网NF聚合得到的聚合结果和该AF对应的服务器自身产生的执行结果进行聚合，得到最终的VFL任务的执行结果。其中，核心网NF聚合得到的聚合结果和该AF对应的服务器自身产生的执行结果具有相同的关联标识。
通过图9所示的实现流程，AF可向各个参与VFL任务的节点发送初始VFL模型，并可通过关联标识将各节点执行VFL任务的执行结果进行关联，进而聚合相关联的各节点的执行结果，得到最终的VFL任务的执行结果。
应理解，上述流程中AF作为VFL任务的发起方仅是示例性地，在本申请实施例的技术方案中，任何节点都可以作为VFL任务的发起方。作为一种实现方式，在VFL的实际应用场景中，由于UE或AF可能存在标签，因此通常可由UE或AF作为VFL任务的发起方。
本申请实施例介绍了各节点进行注册、发起方发现并确定参与节点、初始模型发送和执行结果汇聚的过程，实现了网络架构对跨域纵向联邦学习的支持。
在本申请实施例的方案中，首先，各节点可将自身与VFL相关的信息注册到NRF上，以保证发起方能够发现合适的节点。从而，发起方可通过NRF，发现可参与VFL的各个节点。进而，发起方可向各节点发送初始模型以及关联标识，以保证各节点执行同一个VFL任务以及结果的正确汇聚(聚合)。该方案解决了以下目前尚未解决的技术问题：1)作为发起方的节点(如UE或AF)如何选择合适的其 他域的节点来进行VFL；2)发起方如何将初始模型发送给各节点进行模型训练和/或结果推理；3)发起方如何关联各节点产生的训练/推理结果。
通过本申请实施例的方案，发起方可以联合多个域的节点，在防止数据出本域的情况下进行纵向联邦学习，从而可通过联合更多维的数据对研究对象(如UE的体验)进行全面的分析和预测，以便于在各个域内对研究内容进行优化(如对UE的体验进行优化)。
基于前述的实施例，本申请实施例提供相应的通信的装置。
图10是本申请实施例提供的通信的装置的结构组成示意图一，应用于第一节点，如图10所示，所述通信的装置1000包括：
第一接收模块1001，被配置为接收来自第一网元的第一信息，所述第一信息用于指示符合执行VFL任务的条件的至少一个第二节点；确定模块1002，被配置为从所述至少一个第二节点中，确定参与所述VFL任务的至少一个参与节点
在一些实施例中，第二节点是否符合执行VFL任务的条件，是第一网元基于第二节点的能力信息确定的，第二节点的能力信息用于指示以下至少一项：第二节点是否支持执行VFL任务；第二节点支持执行VFL任务的时间段；第二节点是否可作为VFL任务的发起方；第二节点的算力；以及，第二节点的服务区域。
在一些实施例中，至少一个第二节点的参数信息由第一网元收集或确定，其中，对于每一第二节点，第二节点的参数信息包括以下至少一项：第二节点的能力信息；第二节点的标识；第二节点的地址信息；以及，第二节点对应的应用的标识。
在一些实施例中，对于每一第二节点，第二节点的参数信息通过第二注册请求信息携带，第二注册请求信息是第二节点向第一网元发送的注册请求信息。
在一些实施例中，装置1000还包括：第一发送模块，被配置为在接收来自第一网元的第一信息之前，向第一网元发送装置1000的参数信息，装置1000的参数信息包括以下至少一项：装置1000的能力信息；装置1000的标识；装置1000的地址信息；以及，装置1000对应的应用的标识；其中，装置1000的能力信息用于指示以下至少一项：装置1000是否支持执行VFL任务；装置1000支持执行VFL任务的时间段；装置1000是否可作为VFL任务的发起方；装置1000的算力；以及，装置1000的服务区域。
在一些实施例中，装置1000为应用功能AF，第一发送模块，具体被配置为：通过网络功能开放NEF向第一网元发送第一事件的参数信息，第一事件的参数信息是基于装置1000的参数信息生成的；第一事件的参数信息用于指示以下至少一项：AF是否支持执行第一事件；AF支持执行第一事件的时间段；AF是否可作为第一事件的发起方；AF的算力；以及，AF对应的应用的标识；其中，第一事件与VFL任务关联。
在一些实施例中，装置1000的参数信息通过第一注册请求信息携带，第一注册请求信息是装置1000向第一网元发送的注册请求信息。
在一些实施例中，装置1000还包括：第二发送模块，被配置为在接收来自第一网元的第一信息之前，向第一网元发送第二信息，第二信息用于请求第一信息。
在一些实施例中，装置1000为AF，第二发送模块，具体被配置为：通过NEF向第一网元发送第二信息。
在一些实施例中，第二信息用于指示以下至少一项：执行VFL任务；执行VFL任务的时间段；作为VFL任务的发起方；执行VFL任务的算力需求；以及，执行VFL任务的服务区域。
在一些实施例中，确定模块1002，具体被配置为：对于同意参与VFL任务的第二节点，将第二节点确定为参与VFL任务的参与节点。
在一些实施例中，装置1000还包括：第三发送模块，被配置为向至少一个第二节点发送数据需求信息，数据需求信息用于确定执行VFL任务的数据，数据需求信息包括：执行VFL任务所需的数据的类型，和/或，所需的数据的有效时间。
在一些实施例中，装置1000还包括：第二接收模块，被配置为接收来自至少一个第二节点的第三信息，第三信息用于指示第二节点是否同意参与VFL任务。
在一些实施例中，装置1000为AF，第三发送模块，具体被配置为：通过NEF向至少一个第二节点发送数据需求信息。
在一些实施例中，装置1000还包括：第四发送模块，被配置为向至少一个参与节点发送对应的VFL模型，VFL模型用于执行VFL任务；其中，VFL模型包括对应的参与节点的标识和/或地址信息。
在一些实施例中，装置1000为AF，参与节点为终端设备或接入网网元，第四发送模块，具体被配置为：通过第二网元向终端设备或接入网网元发送对应的VFL模型。
在一些实施例中，装置1000还包括：第五发送模块，被配置为向至少一个参与节点发送关联标识，关联标识用于聚合同一VFL任务的不同执行结果。
在一些实施例中，装置1000还包括：第三接收模块，被配置为接收VFL任务对应的第一执行结果，第一执行结果包括：至少一个参与节点执行VFL任务的第二执行结果；和/或，至少一个聚合结果，至少一个聚合结果中的每个聚合结果基于至少两个第二执行结果进行聚合得到。
在一些实施例中，每个第二执行结果，和/或，至少一个聚合结果中的每个聚合结果，均携带相同的关联标识。
在一些实施例中，至少一个聚合结果中的每个聚合结果，是基于相同的关联标识聚合得到的。
在一些实施例中，装置1000还包括：聚合模块，被配置为聚合具有相同关联标识的VFL任务对应的执行结果和装置1000执行VFL任务的执行结果。
在一些实施例中，装置1000为终端设备、接入网网元、核心网网元或AF。
在一些实施例中，第二节点包括以下至少之一：终端设备、接入网网元、核心网网元和AF。
在一些实施例中，第一网元为网络存储功能NRF。
图11是本申请实施例提供的通信的装置的结构组成示意图二，应用于第二节点，如图11所示，通信的装置1100包括：
第一发送模块1101，被配置为向第一网元发送装置1100的能力信息，装置1100的能力信息用于确定装置1100是否符合执行纵向联邦学习VFL任务的条件。
在一些实施例中，装置1100的能力信息用于指示以下至少一项：装置1100是否支持执行VFL任务；装置1100支持执行VFL任务的时间段；装置1100是否可作为VFL任务的发起方；装置1100的算力；以及，装置1100的服务区域信息。
在一些实施例中，装置1100还包括：第二发送模块，被配置为向第一网元发送装置1100的参数信息，装置1100的参数信息包括以下至少一项：装置1100的能力信息；装置1100的标识；装置1100的地址信息；以及，装置1100对应的应用的标识。
在一些实施例中，装置1100为应用功能AF，第二发送模块，具体被配置为：通过网络功能开放NEF向第一网元发送第二事件的参数信息，第二事件的参数信息是基于装置1100的参数信息生成的；第二事件的参数信息用于指示以下至少一项：AF是否支持执行第二事件；AF支持执行第二事件的时间段；AF是否可作为第二事件的发起方；AF的算力；以及，AF对应的应用的标识；其中，第二事件与VFL任务关联。
在一些实施例中，装置1100的参数信息通过第二注册请求信息携带，第二注册请求信息是装置1100向第一网元发送的注册请求信息。
在一些实施例中，装置1100还包括：第一接收模块，被配置为接收来自第一节点的数据需求信息，数据需求信息用于确定执行VFL任务的数据，数据需求信息包括：执行VFL任务所需的数据的类型，和/或，所需的数据的有效时间。
在一些实施例中，装置1100还包括：第三发送模块，被配置为向第一节点发送第三信息，第三信息用于指示装置1100是否同意参与VFL任务。
在一些实施例中，第一节点为AF，第一接收模块，具体被配置为：通过NEF接收来自第一节点的数据需求信息。
在一些实施例中，装置1100还包括：第二接收模块，被配置为接收来自第一节点的VFL模型，VFL模型用于执行VFL任务；其中，VFL模型与装置1100对应，VFL模型包括装置1100的标识和/或地址信息。
在一些实施例中，装置1100还包括：第三接收模块，被配置为接收来自第一节点的关联标识，关联标识用于聚合同一VFL任务的不同执行结果。
在一些实施例中，装置1100还包括：第三发送模块，被配置为向第一节点发送装置1100执行VFL任务的执行结果，执行结果携带关联标识。
在一些实施例中，装置1100还包括：第四接收模块，被配置为接收至少一个参与节点执行VFL任务的第二执行结果；第四发送模块，被配置为向第一节点发送聚合结果，聚合结果基于至少一个第二执行结果和装置1100执行VFL任务的第三执行结果进行聚合得到，其中，聚合结果携带关联标识。
在一些实施例中，每个第二执行结果，以及第三执行结果，均携带相同的关联标识，聚合结果是基于相同的关联标识聚合得到的。
在一些实施例中，第一节点为终端设备、接入网网元、核心网网元或AF。
在一些实施例中，装置1100为终端设备、接入网网元、核心网网元或AF。
在一些实施例中，第一网元为网络存储功能NRF。
图12是本申请实施例提供的通信的装置的结构组成示意图三，应用于第一网元，如图12所示，通信的装置1200包括：
获取模块1201，被配置为获取至少一个第二节点的能力信息，第二节点的能力信息用于确定第二节点是否符合执行纵向联邦学习VFL任务的条件。
在一些实施例中，第二节点的能力信息用于指示以下至少一项：第二节点是否支持执行VFL任务；第二节点支持执行VFL任务的时间段；第二节点是否可作为VFL任务的发起方；第二节点的算力；以及，第二节点的服务区域信息。
在一些实施例中，装置1200还包括：第一接收模块，被配置为在获取至少一个第二节点的能力信息之前，接收至少一个第二节点的参数信息，其中，对于每一第二节点，第二节点的参数信息包括以下至少一项：第二节点的能力信息；第二节点的标识；第二节点的地址信息；以及，第二节点对应的应用的标识。
在一些实施例中，对于每一第二节点，第二节点的参数信息通过第二注册请求信息携带，第二注册请求信息是第二节点向装置1200发送的注册请求信息。
在一些实施例中，装置1200还包括：第二接收模块，被配置为接收第一节点的参数信息，第一节点的参数信息包括以下至少一项：第一节点的能力信息；第一节点的标识；第一节点的地址信息；以及，第一节点对应的应用的标识；其中，第一节点的能力信息用于指示以下至少一项：第一节点是否支持执行VFL任务；第一节点支持执行VFL任务的时间段；第一节点是否可作为VFL任务的发起方；第一节点的算力；以及，第一节点的服务区域。
在一些实施例中，第一节点的参数信息通过第一注册请求信息携带，第一注册请求信息是第一节点向装置1200发送的注册请求信息。
在一些实施例中，装置1200还包括：发送模块，被配置为向第一节点发送第一信息，第一信息用于指示符合执行VFL任务的条件的至少一个第二节点。
在一些实施例中，装置1200还包括：第三接收模块，被配置为在向第一节点发送第一信息之前，接收来自第一节点的第二信息，第二信息用于请求第一信息。
在一些实施例中，第一节点为终端设备、接入网网元、核心网网元或应用功能AF。
在一些实施例中，第二节点包括以下至少之一：终端设备、接入网网元、核心网网元和AF。
在一些实施例中，装置1200为网络存储功能NRF。
本领域技术人员应当理解，本申请实施例的上述通信的装置的相关描述可以参照本申请实施例的通信的方法的相关描述进行理解。
图13是本申请实施例提供的一种通信设备1300示意性结构图。该通信设备可以是终端设备，也可以是网络设备，还可以是核心网设备。图13所示的通信设备1300包括处理器1310，处理器1310可以从存储器中调用并运行计算机程序，以实现本申请实施例中的方法。
可选地，如图13所示，通信设备1300还可以包括存储器1320。其中，处理器1310可以从存储器1320中调用并运行计算机程序，以实现本申请实施例中的方法。
其中，存储器1320可以是独立于处理器1310的一个单独的器件，也可以集成在处理器1310中。
可选地，如图13所示，通信设备1300还可以包括收发器1330，处理器1310可以控制该收发器1330与其他设备进行通信，具体地，可以向其他设备发送信息或数据，或接收其他设备发送的信息或数据。
其中，收发器1330可以包括发射机和接收机。收发器1330还可以进一步包括天线，天线的数量可以为一个或多个。
可选地，该通信设备1300具体可为本申请实施例的第一节点，并且该通信设备1300可以实现本申请实施例的各个方法中由第一节点实现的相应流程，为了简洁，在此不再赘述。
可选地，该通信设备1300具体可为本申请实施例的第二节点，并且该通信设备1300可以实现本申请实施例的各个方法中由第二节点实现的相应流程，为了简洁，在此不再赘述。
可选地，该通信设备1300具体可为本申请实施例的第一网元，并且该通信设备1300可以实现本申请实施例的各个方法中由第一网元实现的相应流程，为了简洁，在此不再赘述。
图14是本申请实施例的芯片的示意性结构图。图14所示的芯片1400包括处理器1410，处理器1410可以从存储器中调用并运行计算机程序，以实现本申请实施例中的方法。
可选地，如图14所示，芯片1400还可以包括存储器1420。其中，处理器1410可以从存储器1420中调用并运行计算机程序，以实现本申请实施例中的方法。
其中，存储器1420可以是独立于处理器1410的一个单独的器件，也可以集成在处理器1410中。
可选地，该芯片1400还可以包括输入接口1430。其中，处理器1410可以控制该输入接口1430与其他设备或芯片进行通信，具体地，可以获取其他设备或芯片发送的信息或数据。
可选地，该芯片1400还可以包括输出接口1440。其中，处理器1410可以控制该输出接口1440与其他设备或芯片进行通信，具体地，可以向其他设备或芯片输出信息或数据。
可选地，该芯片可应用于本申请实施例中的第一节点，并且该芯片可以实现本申请实施例的各个方法中由第一节点实现的相应流程，为了简洁，在此不再赘述。
可选地，该芯片可应用于本申请实施例中的第二节点，并且该芯片可以实现本申请实施例的各个方法中由第二节点实现的相应流程，为了简洁，在此不再赘述。
可选地，该芯片可应用于本申请实施例中的第一网元，并且该芯片可以实现本申请实施例的各个方法中由第一网元实现的相应流程，为了简洁，在此不再赘述。
应理解，本申请实施例提到的芯片还可以称为系统级芯片，系统芯片，芯片系统或片上系统芯片等。
本申请实施例还提供了一种计算机存储介质，所述计算机存储介质存储有一个或者多个程序，所述一个或者多个程序可被一个或者多个处理器执行，以实现本申请实施例中的方法。
图15是本申请实施例提供的一种通信系统1500的示意性框图。如图15所示，该通信系统1500包括第一节点1510、第二节点1520和第一网元1530。
其中，该第一节点1510可以用于实现上述方法中由第一节点实现的相应的功能，该第二节点1520可以用于实现上述方法中由第二节点实现的相应的功能，以及该第一网元1530可以用于实现上述方法中由第一网元实现的相应的功能，为了简洁，在此不再赘述。
应理解，本申请实施例的处理器可能是一种集成电路芯片，具有信号的处理能力。在实现过程中，上述方法实施例的各步骤可以通过处理器中的硬件的集成逻辑电路或者软件形式的指令完成。上述的处理器可以是通用处理器、数字信号处理器(Digital Signal Processor，DSP)、专用集成电路(Application Specific Integrated Circuit，ASIC)、现成可编程门阵列(Field Programmable Gate Array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成，或者用译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器，处理器读取存储器中的信息，结合其硬件完成上述方法的步骤。
可以理解，本申请实施例中的存储器可以是易失性存储器或非易失性存储器，或可包括易失性和非易失性存储器两者。其中，非易失性存储器可以是只读存储器(Read-Only Memory，ROM)、可编程只读存储器(Programmable ROM，PROM)、可擦除可编程只读存储器(Erasable PROM，EPROM)、电可擦除可编程只读存储器(Electrically EPROM，EEPROM)或闪存。易失性存储器可以是随机存取存储器(Random Access Memory，RAM)，其用作外部高速缓存。通过示例性但不是限制性说明，许多形式的RAM可用，例如静态随机存取存储器(Static RAM，SRAM)、动态随机存取存储器(Dynamic RAM，DRAM)、同步动态随机存取存储器(Synchronous DRAM，SDRAM)、双倍数据速率同步动态随机存取存储器(Double Data Rate SDRAM，DDR SDRAM)、增强型同步动态随机存取存储器(Enhanced SDRAM，ESDRAM)、同步连接动态随机存取存储器(Synchlink DRAM，SLDRAM)和直接内存总线随机存取存储器(Direct Rambus RAM，DR RAM)。应注意，本文描述的系统和方法的存储器旨在包括但不限于这些和任意其它适合类型的存储器。
应理解，上述存储器为示例性但不是限制性说明，例如，本申请实施例中的存储器还可以是静态随机存取存储器(static RAM，SRAM)、动态随机存取存储器(dynamic RAM，DRAM)、同步动态随机存取存储器(synchronous DRAM，SDRAM)、双倍数据速率同步动态随机存取存储器(double data rate SDRAM，DDR SDRAM)、增强型同步动态随机存取存储器(enhanced SDRAM，ESDRAM)、同步连接动态随机存取存储器(synch link DRAM，SLDRAM)以及直接内存总线随机存取存储器(Direct Rambus RAM，DR RAM)等等。也就是说，本申请实施例中的存储器旨在包括但不限于这些和任意其它适合类型的存储器。
本申请实施例还提供了一种计算机可读存储介质，用于存储计算机程序。
可选的，该计算机可读存储介质可应用于本申请实施例中的第一节点，并且该计算机程序使得计算机执行本申请实施例的各个方法中由第一节点实现的相应流程，为了简洁，在此不再赘述。
可选地，该计算机可读存储介质可应用于本申请实施例中的第二节点，并且该计算机程序使得计算机执行本申请实施例的各个方法中由第二节点实现的相应流程，为了简洁，在此不再赘述。
可选地，该计算机可读存储介质可应用于本申请实施例中的第一网元，并且该计算机程序使得计算机执行本申请实施例的各个方法中由第一网元实现的相应流程，为了简洁，在此不再赘述。
本申请实施例还提供了一种计算机程序产品，包括计算机程序指令。
可选的，该计算机程序产品可应用于本申请实施例中的第一节点，并且该计算机程序指令使得计算机执行本申请实施例的各个方法中由第一节点实现的相应流程，为了简洁，在此不再赘述。
可选地，该计算机程序产品可应用于本申请实施例中的第二节点，并且该计算机程序指令使得计算机执行本申请实施例的各个方法中由第二节点实现的相应流程，为了简洁，在此不再赘述。
可选地，该计算机程序产品可应用于本申请实施例中的第一网元，并且该计算机程序指令使得计算机执行本申请实施例的各个方法中由第一网元实现的相应流程，为了简洁，在此不再赘述。
本申请实施例还提供了一种计算机程序。
可选的，该计算机程序可应用于本申请实施例中的第一节点，当该计算机程序在计算机上运行时，使得计算机执行本申请实施例的各个方法中由第一节点实现的相应流程，为了简洁，在此不再赘述。
可选地，该计算机程序可应用于本申请实施例中的第二节点，当该计算机程序在计算机上运行时，使得计算机执行本申请实施例的各个方法中由第二节点实现的相应流程，为了简洁，在此不再赘述。
可选地，该计算机程序可应用于本申请实施例中的第一网元，当该计算机程序在计算机上运行时，使得计算机执行本申请实施例的各个方法中由第一网元实现的相应流程，为了简洁，在此不再赘述。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统、装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。
所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(Read-Only Memory，)ROM、随机存取存储器(Random Access Memory，RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (57)

  1. 一种通信的方法，应用于第一节点，所述方法包括：
    接收来自第一网元的第一信息，所述第一信息用于指示符合执行VFL任务的条件的至少一个第二节点；
    从所述至少一个第二节点中，确定参与所述VFL任务的至少一个参与节点。
  2. 根据权利要求1所述的方法，其中，所述第二节点是否符合执行所述VFL任务的条件，是所述第一网元基于所述第二节点的能力信息确定的，所述第二节点的能力信息用于指示以下至少一项：
    所述第二节点是否支持执行VFL任务；
    所述第二节点支持执行VFL任务的时间段；
    所述第二节点是否可作为VFL任务的发起方；
    所述第二节点的算力；以及，
    所述第二节点的服务区域。
  3. 根据权利要求2所述的方法，其中，所述至少一个第二节点的参数信息由所述第一网元收集或确定，其中，对于每一第二节点，所述第二节点的参数信息包括以下至少一项：
    所述第二节点的能力信息；
    所述第二节点的标识；
    所述第二节点的地址信息；以及，
    所述第二节点对应的应用的标识。
  4. 根据权利要求3所述的方法，其中，
    对于每一第二节点，所述第二节点的参数信息通过第二注册请求信息携带，所述第二注册请求信息是所述第二节点向所述第一网元发送的注册请求信息。
  5. 根据权利要求1至4中任一项所述的方法，其中，在所述接收来自第一网元的第一信息之前，所述方法还包括：
    向所述第一网元发送所述第一节点的参数信息，所述第一节点的参数信息包括以下至少一项：
    所述第一节点的能力信息；
    所述第一节点的标识；
    所述第一节点的地址信息；以及，
    所述第一节点对应的应用的标识；
    其中，所述第一节点的能力信息用于指示以下至少一项：
    所述第一节点是否支持执行VFL任务；
    所述第一节点支持执行VFL任务的时间段；
    所述第一节点是否可作为VFL任务的发起方；
    所述第一节点的算力；以及，
    所述第一节点的服务区域。
  6. 根据权利要求5所述的方法，其中，所述第一节点为应用功能AF，所述向所述第一网元发送所述第一节点的参数信息，包括：
    通过网络功能开放NEF向所述第一网元发送第一事件的参数信息，所述第一事件的参数信息是基于所述第一节点的参数信息生成的；
    所述第一事件的参数信息用于指示以下至少一项：
    所述AF是否支持执行所述第一事件；
    所述AF支持执行所述第一事件的时间段；
    所述AF是否可作为所述第一事件的发起方；
    所述AF的算力；以及，
    所述AF对应的应用的标识；其中，所述第一事件与VFL任务关联。
  7. 根据权利要求5或6所述的方法，其中，
    所述第一节点的参数信息通过第一注册请求信息携带，所述第一注册请求信息是所述第一节点向所述第一网元发送的注册请求信息。
  8. 根据权利要求1至7中任一项所述的方法，其中，在所述接收来自第一网元的第一信息之前，所述方法还包括：
    向所述第一网元发送第二信息，所述第二信息用于请求所述第一信息。
  9. 根据权利要求8所述的方法，其中，所述第一节点为AF，所述向所述第一网元发送第二信息，包括：
    通过NEF向所述第一网元发送所述第二信息。
  10. 根据权利要求8或9所述的方法，其中，所述第二信息用于指示以下至少一项：
    执行VFL任务；
    执行VFL任务的时间段；
    作为VFL任务的发起方；
    执行VFL任务的算力需求；以及，
    执行VFL任务的服务区域。
  11. 根据权利要求1至10中任一项所述的方法，其中，所述从所述至少一个第二节点中，确定参与所述VFL任务的至少一个参与节点，包括：
    对于同意参与所述VFL任务的第二节点，将所述第二节点确定为参与所述VFL任务的参与节点。
  12. 根据权利要求1至11中任一项所述的方法，其中，所述方法还包括：
    向所述至少一个第二节点发送数据需求信息，所述数据需求信息用于确定执行所述VFL任务的数据，所述数据需求信息包括：执行所述VFL任务所需的数据的类型，和/或，所需的数据的有效时间。
  13. 根据权利要求1至12中任一项所述的方法，其中，所述方法还包括：
    接收来自所述至少一个第二节点的第三信息，所述第三信息用于指示所述第二节点是否同意参与所述VFL任务。
  14. 根据权利要求12或13所述的方法，其中，所述第一节点为AF，所述向所述至少一个第二节点发送数据需求信息，包括：
    通过NEF向所述至少一个第二节点发送所述数据需求信息。
  15. 根据权利要求1至14中任一项所述的方法，其中，所述方法还包括：
    向所述至少一个参与节点发送对应的VFL模型，所述VFL模型用于执行所述VFL任务；其中，所述VFL模型包括对应的参与节点的标识和/或地址信息。
  16. 根据权利要求15所述的方法，其中，所述第一节点为AF，所述参与节点为终端设备或接入网网元，所述向所述参与节点发送对应的VFL模型，包括：
    通过第二网元向所述终端设备或所述接入网网元发送对应的VFL模型。
  17. 根据权利要求1至16中任一项所述的方法，其中，所述方法还包括：
    向所述至少一个参与节点发送关联标识，所述关联标识用于聚合同一VFL任务的不同执行结果。
  18. 根据权利要求1至17中任一项所述的方法，其中，所述方法还包括：
    接收所述VFL任务对应的第一执行结果，所述第一执行结果包括：至少一个参与节点执行所述VFL任务的第二执行结果；和/或，至少一个聚合结果，所述至少一个聚合结果中的每个聚合结果基于至少两个第二执行结果进行聚合得到。
  19. 根据权利要求18所述的方法，其中，
    每个第二执行结果，和/或，所述至少一个聚合结果中的每个聚合结果，均携带相同的关联标识。
  20. 根据权利要求18或19所述的方法，其中，所述至少一个聚合结果中的每个聚合结果，是基于相同的关联标识聚合得到的。
  21. 根据权利要求18至20中任一项所述的方法，其中，所述方法还包括：
    聚合具有相同关联标识的所述VFL任务对应的执行结果和所述第一节点执行所述VFL任务的执行结果。
  22. 根据权利要求1至5、7至8、10至13、15或17至21中任一项所述的方法，其中，所述第一节点为终端设备、接入网网元、核心网网元或AF。
  23. 根据权利要求1至22中任一项所述的方法，其中，所述第二节点包括以下至少之一：终端设备、接入网网元、核心网网元和AF。
  24. 根据权利要求1至23中任一项所述的方法，其中，所述第一网元为网络存储功能NRF。
  25. 一种通信的方法，应用于第二节点，所述方法包括：
    向第一网元发送所述第二节点的能力信息，所述第二节点的能力信息用于确定所述第二节点是否符合执行纵向联邦学习VFL任务的条件。
  26. 根据权利要求25所述的方法，其中，所述第二节点的能力信息用于指示以下至少一项：
    所述第二节点是否支持执行VFL任务；
    所述第二节点支持执行VFL任务的时间段；
    所述第二节点是否可作为VFL任务的发起方；
    所述第二节点的算力；以及，
    所述第二节点的服务区域信息。
  27. 根据权利要求25或26所述的方法，其中，所述方法还包括：
    向所述第一网元发送所述第二节点的参数信息，所述第二节点的参数信息包括以下至少一项：
    所述第二节点的能力信息；
    所述第二节点的标识；
    所述第二节点的地址信息；以及，
    所述第二节点对应的应用的标识。
  28. 根据权利要求27所述的方法，其中，所述第二节点为应用功能AF，所述向所述第一网元发送所述第二节点的参数信息，包括：
    通过网络功能开放NEF向所述第一网元发送第二事件的参数信息，所述第二事件的参数信息是基于所述第二节点的参数信息生成的；
    所述第二事件的参数信息用于指示以下至少一项：
    所述AF是否支持执行第二事件；
    所述AF支持执行所述第二事件的时间段；
    所述AF是否可作为所述第二事件的发起方；
    所述AF的算力；以及，
    所述AF对应的应用的标识；其中，所述第二事件与VFL任务关联。
  29. 根据权利要求27或28所述的方法，其中，
    所述第二节点的参数信息通过第二注册请求信息携带，所述第二注册请求信息是所述第二节点向所述第一网元发送的注册请求信息。
  30. 根据权利要求25至29中任一项所述的方法，其中，所述方法还包括：
    接收来自所述第一节点的数据需求信息，所述数据需求信息用于确定执行所述VFL任务的数据，所述数据需求信息包括：执行所述VFL任务所需的数据的类型，和/或，所需的数据的有效时间。
  31. 根据权利要求25至30中任一项所述的方法，其中，所述方法还包括：
    向所述第一节点发送第三信息，所述第三信息用于指示所述第二节点是否同意参与所述VFL任务。
  32. 根据权利要求30或31所述的方法，其中，所述第一节点为AF，所述接收来自所述第一节点的数据需求信息，包括：
    通过NEF接收来自所述第一节点的所述数据需求信息。
  33. 根据权利要求25至32中任一项所述的方法，其中，所述方法还包括：
    接收来自所述第一节点的VFL模型，所述VFL模型用于执行所述VFL任务；其中，所述VFL模型与所述第二节点对应，所述VFL模型包括所述第二节点的标识和/或地址信息。
  34. 根据权利要求25至33中任一项所述的方法，其中，所述方法还包括：
    接收来自所述第一节点的关联标识，所述关联标识用于聚合同一VFL任务的不同执行结果。
  35. 根据权利要求34所述的方法，其中，所述方法还包括：
    向所述第一节点发送所述第二节点执行所述VFL任务的执行结果，所述执行结果携带所述关联标识。
  36. 根据权利要求34所述的方法，其中，所述方法还包括：
    接收至少一个参与节点执行所述VFL任务的第二执行结果；
    向所述第一节点发送聚合结果，所述聚合结果基于至少一个第二执行结果和所述第二节点执行所述VFL任务的第三执行结果进行聚合得到，其中，所述聚合结果携带所述关联标识。
  37. 根据权利要求36所述的方法，其中，每个第二执行结果，以及所述第三执行结果，均携带相同的关联标识，所述聚合结果是基于所述相同的关联标识聚合得到的。
  38. 根据权利要求25至31或33至37中任一项所述的方法，其中，所述第一节点为终端设备、接入网网元、核心网网元或AF。
  39. 根据权利要求25至27或29至38中任一项所述的方法，其中，所述第二节点为终端设备、接入网网元、核心网网元或AF。
  40. 根据权利要求25至39中任一项所述的方法，其中，所述第一网元为网络存储功能NRF。
  41. 一种通信的方法，应用于第一网元，所述方法包括：
    获取至少一个第二节点的能力信息，所述第二节点的能力信息用于确定所述第二节点是否符合执行纵向联邦学习VFL任务的条件。
  42. 根据权利要求41所述的方法，其中，所述第二节点的能力信息用于指示以下至少一项：
    所述第二节点是否支持执行VFL任务；
    所述第二节点支持执行VFL任务的时间段；
    所述第二节点是否可作为VFL任务的发起方；
    所述第二节点的算力；以及，
    所述第二节点的服务区域信息。
  43. 根据权利要求41或42所述的方法，其中，在所述获取至少一个第二节点的能力信息之前，所述方法还包括：
    接收所述至少一个第二节点的参数信息，其中，对于每一第二节点，所述第二节点的参数信息包括以下至少一项：
    所述第二节点的能力信息；
    所述第二节点的标识；
    所述第二节点的地址信息；以及，
    所述第二节点对应的应用的标识。
  44. 根据权利要求43所述的方法，其中，
    对于每一第二节点，所述第二节点的参数信息通过第二注册请求信息携带，所述第二注册请求信息是所述第二节点向所述第一网元发送的注册请求信息。
  45. 根据权利要求41至44中任一项所述的方法，其中，所述方法还包括：
    接收所述第一节点的参数信息，所述第一节点的参数信息包括以下至少一项：
    所述第一节点的能力信息；
    所述第一节点的标识；
    所述第一节点的地址信息；以及，
    所述第一节点对应的应用的标识；
    其中，所述第一节点的能力信息用于指示以下至少一项：
    所述第一节点是否支持执行VFL任务；
    所述第一节点支持执行VFL任务的时间段；
    所述第一节点是否可作为VFL任务的发起方；
    所述第一节点的算力；以及，
    所述第一节点的服务区域。
  46. 根据权利要求45所述的方法，其中，
    所述第一节点的参数信息通过第一注册请求信息携带，所述第一注册请求信息是所述第一节点向所述第一网元发送的注册请求信息。
  47. 根据权利要求40至46中任一项所述的方法，其中，所述方法还包括：
    向所述第一节点发送第一信息，所述第一信息用于指示符合执行所述VFL任务的条件的至少一个第二节点。
  48. 根据权利要求47所述的方法，其中，在所述向所述第一节点发送第一信息之前，所述方法还包括：
    接收来自所述第一节点的第二信息，所述第二信息用于请求所述第一信息。
  49. 根据权利要求41至48中任一项所述的方法，其中，所述第一节点为终端设备、接入网网元、核心网网元或应用功能AF。
  50. 根据权利要求41至49中任一项所述的方法，其中，所述第二节点包括以下至少之一：终端设备、接入网网元、核心网网元和AF。
  51. 根据权利要求41至50中任一项所述的方法，其中，所述第一网元为网络存储功能NRF。
  52. 一种通信的装置，所述装置包括：
    第一接收模块，被配置为接收来自第一网元的第一信息，所述第一信息用于指示符合执行VFL任务的条件的至少一个第二节点；
    确定模块，被配置为从所述至少一个第二节点中，确定参与所述VFL任务的至少一个参与节点。
  53. 一种通信的装置，所述装置包括：
    第一发送模块，被配置为向第一网元发送所述装置的能力信息，所述装置的能力信息用于确定所述装置是否符合执行纵向联邦学习VFL任务的条件。
  54. 一种通信的装置，所述装置包括：
    获取模块，被配置为获取至少一个第二节点的能力信息，所述第二节点的能力信息用于确定所述第二节点是否符合执行纵向联邦学习VFL任务的条件。
  55. 一种通信设备，所述通信设备包括：
    存储器，用于存储计算机可执行指令；
    处理器，与所述存储器连接，用于通过执行所述计算机可执行指令，实现权利要求1至24中任一项所述的方法，或者，实现权利要求25至40中任一项所述的方法，或者，实现权利要求41至51中任一项所述的方法。
  56. 一种芯片，所述芯片包括：
    处理器，用于从存储器中调用并运行计算机程序，使得安装有所述芯片的设备执行如权利要求1至24中任一项所述的方法，或者，执行如权利要求25至40中任一项所述的方法，或者，执行如权利要求41至51中任一项所述的方法。
  57. 一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，所述计算机程序被至少一个处理器执行时实现如权利要求1至24中任一项所述的方法，或者，实现如权利要求25至40中任一项所述的方法，或者，实现如权利要求41至51中任一项所述的方法。
PCT/CN2023/085576 2023-03-31 2023-03-31 一种通信的方法、装置、设备、芯片和存储介质 Ceased WO2024197846A1 (zh)

Priority Applications (8)

Application Number Priority Date Filing Date Title
CN202511499057.5A CN121262104A (zh) 2023-03-31 2023-03-31 一种通信的方法、装置、设备、芯片和存储介质
EP23929431.7A EP4648376A4 (en) 2023-03-31 2023-03-31 COMMUNICATION METHODS AND DEVICES, DEVICES, CHIPS AND STORAGE MEDIA
JP2025555936A JP2026513214A (ja) 2023-03-31 2023-03-31 通信方法、装置、機器、チップおよび記憶媒体
KR1020257033412A KR20250164227A (ko) 2023-03-31 2023-03-31 통신 방법, 장치, 기기, 칩 및 저장 매체
PCT/CN2023/085576 WO2024197846A1 (zh) 2023-03-31 2023-03-31 一种通信的方法、装置、设备、芯片和存储介质
CN202380093174.4A CN120642308A (zh) 2023-03-31 2023-03-31 一种通信的方法、装置、设备、芯片和存储介质
US19/291,423 US20250358200A1 (en) 2023-03-31 2025-08-05 Communication methods and apparatuses, device, chip and storage medium
MX2025011026A MX2025011026A (es) 2023-03-31 2025-09-18 Metodos y aparatos de comunicacion, dispositivo, chip y medio de almacenamiento

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2023/085576 WO2024197846A1 (zh) 2023-03-31 2023-03-31 一种通信的方法、装置、设备、芯片和存储介质

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US19/291,423 Continuation US20250358200A1 (en) 2023-03-31 2025-08-05 Communication methods and apparatuses, device, chip and storage medium

Publications (1)

Publication Number Publication Date
WO2024197846A1 true WO2024197846A1 (zh) 2024-10-03

Family

ID=92903143

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/085576 Ceased WO2024197846A1 (zh) 2023-03-31 2023-03-31 一种通信的方法、装置、设备、芯片和存储介质

Country Status (7)

Country Link
US (1) US20250358200A1 (zh)
EP (1) EP4648376A4 (zh)
JP (1) JP2026513214A (zh)
KR (1) KR20250164227A (zh)
CN (2) CN120642308A (zh)
MX (1) MX2025011026A (zh)
WO (1) WO2024197846A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN119363745A (zh) * 2024-10-29 2025-01-24 中国电信股份有限公司技术创新中心 联邦学习的方法、服务器、客户端和系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113573331A (zh) * 2020-04-29 2021-10-29 华为技术有限公司 一种通信方法、装置及系统
CN114666221A (zh) * 2020-12-07 2022-06-24 中兴通讯股份有限公司 网络切片子网运维管理方法、装置、系统、设备及介质
WO2022183496A1 (zh) * 2021-03-05 2022-09-09 华为技术有限公司 模型训练的方法和装置
US20220350898A1 (en) * 2021-04-29 2022-11-03 Jiangsu Superfluidity Information Technology Co., Ltd Model training method, model using method, system, trusted node and device

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11836583B2 (en) * 2019-09-09 2023-12-05 Huawei Cloud Computing Technologies Co., Ltd. Method, apparatus and system for secure vertical federated learning
CN114691167A (zh) * 2020-12-31 2022-07-01 华为技术有限公司 机器学习模型更新的方法和装置
CN116917907A (zh) * 2021-02-15 2023-10-20 瑞典爱立信有限公司 关于分布式分割属性的高能效深度神经网络训练
CN117121516A (zh) * 2021-04-22 2023-11-24 Oppo广东移动通信有限公司 应用于移动通信系统的联邦学习方法、装置、终端及介质

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113573331A (zh) * 2020-04-29 2021-10-29 华为技术有限公司 一种通信方法、装置及系统
CN114666221A (zh) * 2020-12-07 2022-06-24 中兴通讯股份有限公司 网络切片子网运维管理方法、装置、系统、设备及介质
WO2022183496A1 (zh) * 2021-03-05 2022-09-09 华为技术有限公司 模型训练的方法和装置
US20220350898A1 (en) * 2021-04-29 2022-11-03 Jiangsu Superfluidity Information Technology Co., Ltd Model training method, model using method, system, trusted node and device

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4648376A4 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN119363745A (zh) * 2024-10-29 2025-01-24 中国电信股份有限公司技术创新中心 联邦学习的方法、服务器、客户端和系统

Also Published As

Publication number Publication date
EP4648376A1 (en) 2025-11-12
JP2026513214A (ja) 2026-04-23
US20250358200A1 (en) 2025-11-20
CN121262104A (zh) 2026-01-02
CN120642308A (zh) 2025-09-12
EP4648376A4 (en) 2026-02-25
KR20250164227A (ko) 2025-11-24
MX2025011026A (es) 2025-10-01

Similar Documents

Publication Publication Date Title
CN116193441B (zh) 一种通信方法、通信装置和通信系统
US20250274359A1 (en) Communication method and communications device
EP3906647B1 (en) Flexible authorization in 5g service based core network
AU2017423732B2 (en) Network security management method, and apparatus
TWI820696B (zh) 通訊方法、裝置及電腦可讀儲存介質
EP3833150A1 (en) User plane security policy implementation method, apparatus, and system
CN115244892B (zh) 安全认证方法、装置、设备及存储介质
WO2024032226A1 (zh) 通信方法和通信装置
WO2022222745A1 (zh) 一种通信方法及装置
US20240380730A1 (en) Enabling distributed non-access stratum terminations
US20250358200A1 (en) Communication methods and apparatuses, device, chip and storage medium
CN113543112B (zh) 网络漫游认证方法、装置、电子设备及存储介质
CN119997014B (zh) 一种通信方法和通信装置
KR20260059574A (ko) 서비스 api 호출 방법 및 장치
WO2025026205A1 (zh) 服务api调用的方法和装置
JP2026507148A (ja) 通信方法および通信装置
WO2025209408A1 (zh) 通信方法及装置
CN120934778A (zh) 通信方法、系统及装置
CN122002374A (zh) 数据传输方法、装置、介质、电子设备及程序产品
CN120475377A (zh) 一种安全通信方法和通信装置
WO2022237838A1 (zh) 通信方法和通信装置
WO2025146024A9 (zh) 信息传输方法及通信装置
WO2025026183A1 (zh) 通信方法和通信装置
WO2022252658A1 (zh) 一种漫游接入方法及装置
CN121312173A (zh) 提供网络服务的方法和通信装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23929431

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 202380093174.4

Country of ref document: CN

WWE Wipo information: entry into national phase

Ref document number: 2023929431

Country of ref document: EP

ENP Entry into the national phase

Ref document number: 2023929431

Country of ref document: EP

Effective date: 20250805

WWE Wipo information: entry into national phase

Ref document number: 202517079771

Country of ref document: IN

REG Reference to national code

Ref country code: BR

Ref legal event code: B01A

Ref document number: 112025018556

Country of ref document: BR

WWP Wipo information: published in national office

Ref document number: 202517079771

Country of ref document: IN

Ref document number: 202380093174.4

Country of ref document: CN

ENP Entry into the national phase

Ref document number: 2025555936

Country of ref document: JP

Kind code of ref document: A

WWE Wipo information: entry into national phase

Ref document number: 2025555936

Country of ref document: JP

ENP Entry into the national phase

Ref document number: 1020257033412

Country of ref document: KR

Free format text: ST27 STATUS EVENT CODE: A-0-1-A10-A15-NAP-PA0105 (AS PROVIDED BY THE NATIONAL OFFICE)

WWE Wipo information: entry into national phase

Ref document number: KR1020257033412

Country of ref document: KR

NENP Non-entry into the national phase

Ref country code: DE

WWP Wipo information: published in national office

Ref document number: 2023929431

Country of ref document: EP

WWP Wipo information: published in national office

Ref document number: 1020257033412

Country of ref document: KR