Communication method, communication node, medium, and program product

Abstract

The invention discloses a communication method, a communication node, a medium and a program product. Comprising the steps of receiving first information and making a decision according to the first information; and sending a federal learning instruction to the at least one second communication node according to the decision result. According to the technical scheme, the decision for adjusting the federated learning state of the at least one second communication node is made based on the received first message, so that the first communication node can dynamically adjust the object participating in federated learning in the federated learning process; and the adaptability of federal learning to the state change is improved.

Description

Communication methods, communication nodes, media and software products Technical Field

[0001] This application relates to the field of communication technology, and in particular to a communication method, communication node, medium, and program product. Background Technology

[0002] In the 5th Generation Mobile Communication (5G) system, the Network Data Analytics Function (NWDAF) is a 5G Core Network Function (5GC NF) located in the control plane, which can perform statistical data and machine learning related tasks in the 5G system.

[0003] Current 5G systems support multiple NWDAFs and / or Application Functions (AFs) as participants for federated learning training and / or inference. However, adjustments to participants can only be made before federated learning training and / or inference begins, and dynamic adjustments cannot be made to federated learning training and / or inference when the state of a participant changes. Summary of the Invention

[0004] This application provides a communication method, communication node, medium, and program product to solve the problem of the inability to dynamically adjust participants during federated learning. It dynamically adjusts the objects participating in federated learning based on the participants' states during federated learning training and / or inference, thereby improving the adaptability of federated learning to state changes.

[0005] To achieve the above objectives, embodiments of this application provide a communication method applied to a first communication node, comprising:

[0006] Receive the first piece of information and make a decision based on it;

[0007] Based on the decision result, a federated learning instruction is sent to at least one second communication node.

[0008] To achieve the above objectives, embodiments of this application provide a communication method applied to a second communication node, comprising:

[0009] Report the first information to the first communication node corresponding to the second communication node;

[0010] Receive federal learning instructions.

[0011] To achieve the above objectives, embodiments of this application provide a communication method applied to a first communication node, comprising:

[0012] Receive the second piece of information and make a decision based on it;

[0013] Based on the decision result, a federated learning instruction is sent to at least one second communication node.

[0014] To achieve the above objectives, embodiments of this application provide a communication node, including: a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for implementing communication between the processor and the memory. When the program is executed by the processor, it implements the steps of the communication method as described in any of the embodiments of this application.

[0015] To achieve the above objectives, embodiments of this application provide a storage medium for computer-readable storage, wherein the storage medium stores one or more programs, which can be executed by one or more processors to implement the steps of the communication method of any embodiment of this application.

[0016] To achieve the above objectives, embodiments of this application provide a computer program product, including a computer program, which, when executed by a processor, implements the steps of any of the communication methods described in the embodiments of this application.

[0017] The communication method, communication node, medium, and program product provided in this application receive first information and make decisions based on the first information; then, based on the decision results, they send federated learning instructions to at least one second communication node. By adopting the above technical solution, decisions are made based on the received first message to adjust the federated learning state of at least one second communication node, enabling the first communication node to dynamically adjust the objects participating in federated learning during the federated learning process, thereby improving the adaptability of federated learning to state changes. Attached Figure Description

[0018] Figure 1 is a schematic diagram of a 5G core network architecture provided in the prior art;

[0019] Figure 2 is a flowchart of a communication method provided in an embodiment of this application;

[0020] Figure 3 is a flowchart of a communication method provided in an embodiment of this application;

[0021] Figure 4 is a flowchart of a communication method provided in an embodiment of this application;

[0022] Figure 5 is a flowchart of a communication method provided in an embodiment of this application;

[0023] Figure 6 is a timing example diagram of dynamic selection of a federated learning client provided in an embodiment of this application;

[0024] Figure 7 is a timing example diagram of dynamic selection of a federated learning client provided in an embodiment of this application;

[0025] Figure 8 is a timing example diagram of a federated learning capability reporting provided in an embodiment of this application;

[0026] Figure 9 is a schematic diagram of the structure of a communication device provided in an embodiment of this application;

[0027] Figure 10 is a schematic diagram of the structure of a communication device provided in an embodiment of this application;

[0028] Figure 11 is a schematic diagram of the structure of a communication device provided in an embodiment of this application;

[0029] Figure 12 is a schematic diagram of the structure of a communication device provided in an embodiment of this application;

[0030] Figure 13 is a schematic diagram of the structure of a communication node provided in an embodiment of this application. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in detail below with reference to the accompanying drawings. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be arbitrarily combined with each other.

[0032] The steps illustrated in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases the steps shown or described may be performed in a different order than that presented here.

[0033] The communication method provided in this application can be applied to 5G systems to perform state adjustments during the federated learning process for multiple NWDAFs and / or AFs that support federated learning training and / or inference. To clearly describe federated learning in 5G systems, a brief introduction is given here of the 5G core network architecture and the NWDAFs that perform federated learning as 5GC NFs based on the 5G core network architecture.

[0034] Figure 1 is a schematic diagram of a 5G core network architecture provided in the prior art, which has the following functions:

[0035] 1) User Equipment (UE).

[0036] 2) Radio Access Network (RAN). The RAN manages radio resources, transmits user data received through the N3 interface to the UE, and transmits user data from the UE through the N3 interface. The RAN maps between Quality of Service (QoS) traffic in Dedicated Radio Bearer (DRB) and Protocol Data Unit (PDU) sessions.

[0037] 3) Access and Mobility Management Function (AMF). This function includes registration management, connection management, reachability management, and mobility management. It also performs access authentication and authorization. AMF is a Network Attached Storage (NAS) security endpoint used to forward SM NAS data between the UE and the Session Management Function (SMF).

[0038] 4) SMF. This function includes the following: session establishment, modification, and release; UE IP address allocation and management (including optional authorization functions); selection and control of User Plane (UP) functions and downlink data notification, etc. The SMF controls the User Plane Function (UPF) through the N4 interface. The SMF provides the UPF with Packet Detection Rules (PDR) to indicate how to detect user data traffic; provides Forwarding and Routing Control Association Rules (FAR); and provides QoS enforcement rules and Usage Reporting Rules (URR) to indicate how the UPF performs user data traffic forwarding, QoS processing, and usage reporting on user data traffic detected using PDR.

[0039] 5) UPF. This function includes the following: serving as an anchor point for intra / inter-radio access type mobility, packet routing and forwarding, traffic usage reporting, QoS processing for the UP, downlink packet buffering, and downlink data notification triggering. The General Packet Radio Service (GPRS) Tunneling Protocol for User Plane (GTP-U) tunnel is used for the N3 interface between the RAN and UPF. The GTP-U tunnel operates on a per-PDU session basis. For downlink traffic, the UPF binds the downlink traffic to the QoS traffic within the PDU session's GTP-U tunnel using the FAR received from the SMF. For uplink traffic, the RAN transmits user plane traffic to the QoS stream identified by the UE.

[0040] 6) Policy Control Function (PCF). The PCF provides QoS policy rules to control plane functions for enforcement. The PCF translates AF requests into PCC rules applicable to PDU sessions.

[0041] 7) Unified Data Management (UDM). The UDM performs 3GPP AKA authentication credential generation, access authorization based on subscription data, UE service NF registration management (e.g., storing AMF for UE storage services, SMF for UE PDU session storage services), and subscription management. The UDM accesses the UDR to retrieve UE subscription data and stores the UE context in the UDR. The UDM and UDR can be deployed together.

[0042] Based on the 5G core network architecture, the NWDAF is a 5GC NF located in the control plane, performing data statistics and machine learning-related tasks in the 5G system. The NWDAF can interact with different entities for various purposes:

[0043] Data is collected based on event subscriptions provided by AMF, SMF, UPF, PCF, UDM, NSCAF, AF (directly or through NEF) and Operations, Administration and Maintenance (OAM);

[0044] [Optional] Use the Data Collection Coordination Function (DCCF) for analysis and data collection;

[0045] Retrieve information from data repositories (e.g., retrieve UDRs related to users via UDM or retrieve PFD information via NEF (PFDF));

[0046] Collect location information data from the LCS system;

[0047] [Optional] Store and retrieve information from the Analytics Data Storage Function (ADRF);

[0048] [Optional] Analyze and collect data from the Messaging Framework Adaptor Function (MFAF);

[0049] Retrieve information about NF (e.g., retrieve NF-related information from NRF);

[0050] Provide analytics to consumers on demand.

[0051] Provides batch data related to the analysis ID.

[0052] Provides information on the accuracy of the analysis ID.

[0053] Provides information on the accuracy of machine learning (ML) models or indications of ML model accuracy degradation.

[0054] In some examples, a single instance or multiple instances of NWDAF can be deployed in a public terrestrial mobile network. NWDAF may contain the following logical functions:

[0055] Analysis Logic Function (AnLF): A logic function in NWDAF used to perform inference, derive analytical information (i.e., derive statistical data and / or predictions based on analytical consumer requests) and expose analytical services (i.e., Nnwdaf_AnalyticsSubscription or Nnwdaf_AnalyticsInfo).

[0056] Model Training Logic Function (MTLF): A logic function in NWDAF used to train ML models and expose new training services (such as providing pre-trained ML models).

[0057] An NWDAF can contain a Model Training Logical Function (MTLF) or an Analysis Logic Function (AnLF), or both.

[0058] The Data Collection Coordination and Function (DCCF) is also an NF on the 5G core network control plane. The DCCF is responsible for coordinating the collection and distribution of data requested by NF consumers. It prevents data sources from processing multiple subscriptions to the same data and prevents multiple notifications containing the same information from being sent due to incoordination of data consumer requests.

[0059] DCCF is applicable to:

[0060] NWDAF requests data from a data source (such as for computational analysis).

[0061] NF consumers are analyzed from the NWDAF data source.

[0062] An NF consumer that requests data from an ADRF data source.

[0063] ADRF that receives data from NF data sources.

[0064] To clearly describe the application of federated learning in 5G systems, the following is a brief introduction to horizontal federated learning and vertical federated learning:

[0065] Vertical federated learning is suitable for situations where there is little overlap in the data features of participants, but a lot of overlap in sample IDs. In vertical federated learning, the training data for different client models have the same sample space but different feature spaces.

[0066] In the already defined core network vertical federated learning, NWDAF / AF can serve as a vertical federated learning server and / or a vertical federated learning client.

[0067] Horizontal federated learning is suitable for situations where participants' data features overlap significantly, while sample IDs overlap less. In horizontal federated learning, different client models train on data with the same feature space but different sample spaces.

[0068] In the already defined core network horizontal federated learning, only NWDAF can serve as a horizontal federated learning server and / or a horizontal federated learning client.

[0069] However, although current 5G systems support multiple NWDAFs and / or AFs as participants in federated learning training and / or inference, adjustments to participants can only be made before the start of federated learning training and / or inference. Dynamic adjustments to the training and / or inference based on changes in the participants' states during the federated learning process are not possible. To address this issue, this application provides a communication method that can be implemented using a first communication node and / or a second communication node. The first and second communication nodes are respectively nodes acting as federated learning servers and clients among the participants in the 5G system performing federated learning. The first and / or second communication nodes are generally electronic devices with certain computing capabilities. In this embodiment, the first and / or second communication nodes can be NWDAFs and / or AFs. In some possible implementations, the communication method can be implemented by a processor calling computer-executable instructions stored in memory.

[0070] In one exemplary embodiment, Figure 2 is a flowchart of a communication method provided in an embodiment of this application. This method can be applied to the situation where the federated learning state of communication nodes participating in federated learning is adjusted during the federated learning training process of a 5G system. This method can be executed by a communication device, which can be implemented by software and / or hardware and integrated on the communication node. This method can be applied to a first communication node.

[0071] As shown in Figure 2, the communication method provided in this embodiment of the application specifically includes the following steps:

[0072] S101. Receive the first information and make a decision based on the first information.

[0073] In this embodiment, the first information can be specifically understood as information used to characterize the state of federated learning, or information on changes in the state of communication nodes participating in federated learning.

[0074] In a specific example, the first communication node receives first information to characterize the state of the federated learning or the state changes of the communication nodes participating in the federated learning, and determines whether each communication node participating in the federated learning needs to be adjusted based on the first information, thus obtaining a decision result on whether each communication node participating in the federated learning needs to be adjusted.

[0075] S102. Send federated learning instructions to at least one second communication node based on the decision results.

[0076] In this embodiment, the decision result can be specifically understood as including information on the federated learning state that different communication nodes participating in federated learning need to adjust to.

[0077] In this embodiment, the second communication node can be specifically understood as a communication node that participates in the same federated learning as the first communication node and whose federated learning state can be adjusted by the first communication node.

[0078] In this embodiment, the federated learning instruction can be specifically understood as an instruction message used to instruct the second communication node to adjust its own federated learning state.

[0079] In a specific example, the first communication node sends a federated learning instruction to the second communication node, which is participating in the same federated learning process and needs to adjust its federated learning state, based on the decision result.

[0080] The communication method provided in this application receives first information and makes a decision based on the first information; then, based on the decision result, it sends a federated learning instruction to at least one second communication node. By adopting the above technical solution, and making a decision based on the received first information to adjust the federated learning state of at least one second communication node, the first communication node can dynamically adjust the objects participating in the federated learning process, thereby improving the adaptability of the federated learning to state changes.

[0081] In one embodiment, the first information is provided by a Network Repository Function (NRF) and / or at least one second communication node.

[0082] In a specific example, the first communication node may obtain the first information from the NRF, or from multiple second communication nodes participating in the same federated learning, or simultaneously from the NRF and multiple second communication nodes participating in the same federated learning.

[0083] In one embodiment, before receiving the first information, at least one of the following is also included:

[0084] Subscribe to first information from at least one second communication node;

[0085] Subscribe to first information from network storage;

[0086] At least one second communication node periodically sends the first information.

[0087] In a specific example, the first communication node can obtain the first information from the NRF and / or at least one second communication node in different ways. For instance, during the federated learning process, the first communication node can subscribe to federated learning-related information from at least one second communication node participating in the same federated learning process, so that when the federated learning status of at least one second communication node changes, the second communication node can send the first information to the first communication node. Alternatively, during the federated learning process, the first communication node can subscribe to the NRF regarding the capabilities of at least one second communication node participating in the same federated learning process, so that the NRF can send the first information to the first communication node when the capabilities of the second communication node change. Finally, because the first communication node participates in the same federated learning process, at least one second communication node can periodically and proactively send the first information to the first communication node.

[0088] In one embodiment, the first information includes at least one of the following:

[0089] Changes in data samples or feature information of the second communication node;

[0090] Changes in local privacy protection requirements for the second communication node;

[0091] Changes in data quality information of the second communication node;

[0092] The number of outliers in the second communication node or the number of outliers increases;

[0093] Changes in the computing resource information of the second communication node;

[0094] The second communication node is unable to return intermediate training results within the maximum response time specified by the first communication node;

[0095] The capabilities of the second communication node have changed;

[0096] The time period / duration during which federated learning can be conducted can vary.

[0097] In some examples, changes in data samples or feature information may include sample increases, sample decreases, feature increases, and feature decreases.

[0098] In some examples, changes in data quality information may include changes in missing data information (such as changes in the proportion of missing data values) and changes in outlier information (such as changes in the proportion of outliers).

[0099] In some examples, changes in computing resource information may include changes in computing power information (such as changes in CPU / GPU specifications, changes in the amount of computing power available for federated learning) and changes in storage capacity available for federated learning.

[0100] In one embodiment, the decision result includes at least one of the following:

[0101] Discover a new second communication node using network storage functionality;

[0102] A portion of the second communication nodes were identified as exiting federated learning.

[0103] Federal learning terminated;

[0104] The termination of federalized learning, and the time for resuming federalized learning.

[0105] In one embodiment, sending a federated learning instruction to at least one second communication node includes at least one of the following:

[0106] Send training termination instructions to some of the second communication nodes;

[0107] Send training termination instructions to all second communication nodes;

[0108] Send training termination instructions and the time to restart federated learning to some of the second communication nodes;

[0109] Send training termination instructions and the time to perform federated learning again to all second communication nodes;

[0110] Send training continuation instructions to some of the second communication nodes.

[0111] In a specific example, if the decision result is that some second communication nodes need to exit federated learning, the first communication node may send a training termination instruction to the identified second communication nodes that need to exit federated learning, or send a time to resume federated learning at the same time as sending the training termination instruction to the identified second communication nodes that need to exit federated learning, or send a training continuation instruction to the identified second communication nodes that do not need to exit federated learning.

[0112] In a specific example, if the decision result is to terminate the federated learning, the first communication node can send a training termination instruction to all the second communication nodes that participated in the same federated learning.

[0113] In a specific example, if the decision result is to terminate federated learning and specify the time for resuming federated learning, the first communication node can send a training termination instruction and the time for resuming federated learning to all second communication nodes participating in the same federated learning.

[0114] In one embodiment, the training continues instruction includes at least one of the following:

[0115] Updated sample identifiers used for training;

[0116] Updated feature labels used for training;

[0117] Instructions on whether to reuse the initial model for training.

[0118] In one embodiment, if the training termination instruction is not sent to all second communication nodes, the method further includes:

[0119] Receive confirmation response messages returned by the second communication node;

[0120] Among them, some of the second communication nodes were those that did not receive the training termination instruction.

[0121] In a specific example, if the training termination instruction is not sent to all second communication nodes, the second communication nodes that need to continue training and are located in the same federated learning as the first communication node need to provide feedback to the first communication node to indicate that they are still training normally. At this time, the first communication node can receive the confirmation response message returned by the second communication nodes that did not receive the training termination instruction.

[0122] In one embodiment, if the decision result is that a new second communication node is discovered through the network storage function, the method further includes:

[0123] Discover a new second communication node using network storage functionality;

[0124] Send a federated learning join request message to the new second communication node;

[0125] Receive a response message for the request to join a new second communication node.

[0126] In this embodiment, the federated learning join request message can be specifically understood as a message used to request a new second communication node to join the same federated learning as the first communication node.

[0127] In a specific example, if the decision is to discover a new second communication node via NRF, the first communication node needs to use NRF to discover a second communication node that is different from the one in the same federated learning environment as the first communication node, and then designate the discovered second communication node as the new second communication node. The first communication node sends a federated learning join request message to the new second communication node and receives a join request response message from the new second communication node, which includes its decision on whether to join the same federated learning environment as the first communication node.

[0128] In one embodiment, the federated learning join request message includes at least one of the following:

[0129] Sample identifiers used for training;

[0130] Feature identifiers used for training.

[0131] In some examples, a new second communication node may determine whether its own usable feature identifiers and / or sample identifiers for training are the same as or related to those of the first communication node based on the sample identifiers and / or feature identifiers used for training in the federated learning join request message, and then make a decision and feed back a join request response message.

[0132] In one embodiment, the communication method is applied during the federated learning training process.

[0133] In one embodiment, the first communication node is a horizontal federated learning server or a vertical federated learning server in federated learning, and the second communication node is a horizontal federated learning client or a vertical federated learning client in federated learning.

[0134] In one exemplary embodiment, Figure 3 is a flowchart of a communication method provided in an embodiment of this application. This method can be applied to the situation where the federated learning state of communication nodes participating in federated learning is adjusted during the federated learning training process of a 5G system. This method can be executed by a communication device, which can be implemented by software and / or hardware and integrated on the communication node. This method can be applied to a second communication node.

[0135] As shown in Figure 3, the communication method provided in this embodiment of the application specifically includes the following steps:

[0136] S201, Report the first information to the first communication node.

[0137] In a specific example, the second communication node, upon receiving a subscription trigger or according to its own configuration, periodically reports to the first communication node the first information used to characterize the changes in the federated learning state of the second communication node.

[0138] S202, Receive Federal Learning Instructions.

[0139] In a specific example, after reporting the first information, the second communication node will wait for the first communication node to make a decision based on the first information reported by each second communication node. If the first communication node determines that the federated learning state of the second communication node needs to be adjusted, it will receive a federated learning instruction sent by the first communication node.

[0140] In one embodiment, the first information includes at least one of the following:

[0141] Changes in data samples or feature information of the second communication node;

[0142] Changes in local privacy protection requirements for the second communication node;

[0143] Changes in data quality information of the second communication node;

[0144] Changes in the computing resource information of the second communication node;

[0145] The second communication node is unable to return intermediate training results within the longest response time of the first communication node;

[0146] The number of outliers in the second communication node or the number of outliers increases;

[0147] The capabilities of the second communication node have changed;

[0148] The time period / duration during which federated learning can be conducted can vary.

[0149] In one embodiment, receiving a federated learning instruction includes at least one of the following:

[0150] Receive training termination instruction;

[0151] The time to receive training termination instructions and to resume federal learning.

[0152] In some examples, the federated learning instruction received by the second communication node can be an instruction to the second communication node to exit federated learning, that is, a training termination instruction. Simultaneously, if the first communication node instructs the second communication node to terminate training while simultaneously providing a timeframe for the second communication node to resume federated learning, the second communication node will receive both the training termination instruction and the timeframe for resuming federated learning.

[0153] In one embodiment, receiving a federated learning instruction further includes:

[0154] Receive instructions to continue training.

[0155] In some examples, in addition to being instructed to terminate federated learning, the second communication node can also be instructed to continue federated learning. In this case, the second communication node will be instructed to continue training by the first communication node.

[0156] Understandably, if the second communication node needs to continue training, it may need to adjust the samples, features, or models in the original federated learning process. This adjustment will be determined based on the content contained in the training continuation instruction.

[0157] In one embodiment, the training continues instruction includes at least one of the following:

[0158] Updated sample identifiers used for training;

[0159] Updated feature labels used for training;

[0160] Instructions on whether to reuse the initial model for training.

[0161] In one embodiment, if the second communication node is a new second communication node discovered by the first communication node through the network storage function, the method further includes:

[0162] Receive the federated learning join request message sent by the first communication node;

[0163] Return a join request response message to the first communication node.

[0164] In a specific example, after making a decision based on the received first information, the first communication node may discover a new second communication node to join the federated learning through NRF. If the second communication node is indeed the aforementioned new second communication node, it will receive a federated learning join request message sent by the first communication node. After analyzing the content of the federated learning join request message, the second communication node can decide whether to join the same federated learning as the first communication node and feed back the decision result to the first communication node through a join request response message, ultimately deciding whether to join or refuse to join the federated learning.

[0165] In one embodiment, the federated learning join request message includes at least one of the following:

[0166] Sample identifiers used for training;

[0167] Feature identifiers used for training.

[0168] In one embodiment, the communication method is applied during the federated learning training process.

[0169] In one embodiment, the second communication node is a horizontal federated learning client or a vertical federated learning client in federated learning, and the first communication node is a horizontal federated learning server or a vertical federated learning server in federated learning.

[0170] In one exemplary embodiment, Figure 4 is a flowchart of a communication method provided in an embodiment of this application. This method can be applied to the situation where the federated learning state of communication nodes participating in federated learning is adjusted during the federated learning inference process of a 5G system. This method can be executed by a communication device, which can be implemented by software and / or hardware and integrated on the communication node. This method can be applied to a first communication node.

[0171] As shown in Figure 4, the communication method provided in this embodiment of the application specifically includes the following steps:

[0172] S301. Receive the second information and make a decision based on the second information.

[0173] In this embodiment, the second information can be specifically understood as information used to characterize the federated learning state, or information on changes in the inference state of communication nodes participating in federated learning.

[0174] In a specific example, during the inference process of federated learning, the first communication node receives second information that characterizes changes in the state of federated learning or the inference state of communication nodes participating in federated learning, and determines whether each communication node participating in federated learning needs to adjust its state based on the second information, thus obtaining a decision result for whether each communication node participating in federated learning needs to make adjustments.

[0175] S302. Based on the decision result, send a federated learning instruction to at least one second communication node.

[0176] In this embodiment, the decision result can be specifically understood as including information on the federated learning state that different communication nodes participating in federated learning need to adjust to.

[0177] In this embodiment, the second communication node can be specifically understood as a communication node that participates in the same federated learning as the first communication node and whose federated learning state can be adjusted by the first communication node.

[0178] In this embodiment, the federated learning instruction can be specifically understood as an instruction message used to instruct the second communication node to adjust its own federated learning state.

[0179] In a specific example, the first communication node sends a federated learning instruction to the second communication node, which is participating in the same federated learning process and needs to adjust its federated learning inference state, based on the decision result.

[0180] The communication method provided in this application receives second information and makes a decision based on the second information; then, based on the decision result, it sends a federated learning instruction to at least one second communication node. By adopting the above technical solution, and making a decision based on the received second information to adjust the federated learning state of at least one second communication node, the first communication node can dynamically adjust the objects participating in the federated learning process, thereby improving the adaptability of the federated learning to state changes.

[0181] In one embodiment, the second information is provided by a network storage function and / or at least one second communication node.

[0182] In one embodiment, before receiving the second information, at least one of the following is also included:

[0183] Subscribe to second information from at least one second communication node;

[0184] Subscribe to second information from network storage;

[0185] At least one second communication node periodically sends a second message.

[0186] In one embodiment, the second information includes at least one of the following:

[0187] The second communication node lacks local inference data;

[0188] The number of outliers in the second communication node or the number of outliers increases;

[0189] Changes in sample data or feature data of the second communication node;

[0190] Changes in data quality information of the second communication node;

[0191] Changes in the computing resource information of the second communication node;

[0192] The second communication node is unable to return intermediate inference results within the maximum response time specified by the first communication node;

[0193] The capabilities of the second communication node have changed;

[0194] The time period / duration during which federated learning can be conducted can vary.

[0195] In some examples, missing local inference data may include the loss of local inference data or the inability to collect the corresponding inference data.

[0196] In some examples, changes in data quality information may include changes in missing data information (such as changes in the proportion of missing data values) and changes in outlier information (such as changes in the proportion of outliers).

[0197] In some examples, changes in computing resource information may include changes in computing power information (such as changes in CPU / GPU specifications, changes in computing power available for federated learning) and changes in storage capacity available for federated learning.

[0198] In one embodiment, the decision result includes at least one of the following:

[0199] Federalized learning reasoning terminated;

[0200] It was determined that some of the second communication nodes exited the federated learning inference process, and the inference process continued.

[0201] In one embodiment, if the decision result is that the federated learning inference terminates, a federated learning instruction is sent to at least one second communication node based on the decision result, including:

[0202] Send inference termination instructions and the time to perform federated learning inference again to all second communication nodes.

[0203] In a specific example, if the decision result determined by the first communication node is that the federated learning inference is terminated, it can be assumed that the federated learning inference of the first communication node and each of the second communication nodes will be terminated. At this time, the first communication node will send an inference termination instruction to all the second communication nodes so that each of the second communication nodes stops the federated learning inference. At the same time, it will send a time to each of the second communication nodes to resume the federated learning inference so that each of the second communication nodes can re-enter the federated learning inference.

[0204] In one embodiment, if the decision result determines that some second communication nodes will exit the federated learning inference process and continue the inference process, a federated learning instruction is sent to at least one second communication node according to the decision result, including:

[0205] Send inference termination instructions to some second communication nodes and the time to perform federated learning inference again;

[0206] Send a reasoning continuation instruction to the second communication nodes, excluding some of the second communication nodes.

[0207] In a specific example, if the decision determined by the first communication node is that some second communication nodes will withdraw from federated learning inference, the first communication node needs to notify the selected second communication nodes to stop federated learning inference, while ensuring that the other second communication nodes can continue federated learning inference. In this case, the first communication node will send an inference termination instruction and a timer for restarting federated learning inference to the selected second communication nodes, and will also send an inference continuation instruction to the other second communication nodes that need to continue federated learning inference, ensuring that the nodes that need to continue can proceed with subsequent federated learning inference normally.

[0208] In some examples, after the first communication node sends a federated learning instruction to at least one second communication node to complete the federated learning status adjustment of each participant in this federated learning, the first communication node may also decide to restart the federated learning preparation / training process to start a new federated learning cycle.

[0209] In one embodiment, the communication method is applied to the federated learning inference process.

[0210] In one embodiment, the first communication node is a horizontal federated learning server or a vertical federated learning server in federated learning, and the second communication node is a horizontal federated learning client or a vertical federated learning client in federated learning.

[0211] In one exemplary embodiment, Figure 5 is a flowchart of a communication method provided in an embodiment of this application. This method can be applied to the situation where the federated learning state of communication nodes participating in federated learning is adjusted during the federated learning inference process of a 5G system. This method can be executed by a communication device, which can be implemented by software and / or hardware and integrated on the communication node. This method can be applied to a second communication node.

[0212] As shown in Figure 5, the communication method provided in this embodiment of the application specifically includes the following steps:

[0213] S401, Report the second information to the first communication node.

[0214] S402, Receive Federal Learning Instructions.

[0215] In one embodiment, the second information includes at least one of the following:

[0216] The second communication node lacks local inference data;

[0217] The number of outliers in the second communication node or the number of outliers increases;

[0218] Changes in sample data or feature data of the second communication node;

[0219] Changes in data quality information of the second communication node;

[0220] Changes in the computing resource information of the second communication node;

[0221] The second communication node is unable to return intermediate inference results within the maximum response time specified by the first communication node;

[0222] The capabilities of the second communication node have changed;

[0223] The time period / duration during which federated learning can be conducted can vary.

[0224] In one embodiment, receiving a federated learning instruction includes:

[0225] The time to receive the inference termination instruction and to resume federated learning inference.

[0226] In one embodiment, receiving a federated learning instruction further includes:

[0227] Receive the instruction to continue reasoning.

[0228] In one embodiment, the communication method is applied to the federated learning inference process.

[0229] In one embodiment, the second communication node is a horizontal federated learning client or a vertical federated learning client in federated learning, and the first communication node is a horizontal federated learning server or a vertical federated learning server in federated learning.

[0230] The communication method of this application is illustrated below through some exemplary schemes. In the following schemes, the federated learning server refers to the first communication node, and the federated learning client refers to the second communication node, as an example for illustration.

[0231] Solution 1: A specific example of dynamically selecting federated learning clients participating in the vertical federated learning training process is provided. Figure 6 is a timing example diagram of dynamic selection of federated learning clients provided in an embodiment of this application. As shown in Figure 6, the specific steps may include the following:

[0232] 1. The federated learning server subscribes to federated learning-related information from the federated learning client;

[0233] In some examples, reports from the federated learning client can be triggered periodically by the federated learning client or requested by the federated learning server.

[0234] 2. The federated learning client sends the following information to the federated learning server:

[0235] 1) Instructions and reasons for exiting vertical federated learning training (Insufficient computing power; vertical federated learning training is not supported at this time).

[0236] 2) Changes in data samples or feature information (increase in samples, decrease in samples, increase in features, decrease in features)

[0237] 3) Changes in local privacy protection requirements

[0238] 4) Changes in data quality information (changes in missing data information (e.g., changes in the proportion of missing data values), changes in outlier information (e.g., changes in the proportion of outliers)).

[0239] 5) The amount of outlier data or an increase in the amount of outlier data

[0240] 6) Changes in computing resource information (changes in computing power information (changes in CPU / GPU specifications, changes in computing power available for federated learning), changes in storage capacity available for federated learning)

[0241] 7) Unable to return intermediate training results within the maximum response time specified by the federated learning server.

[0242] 3. The federated learning server subscribes to NRF for capability information related to the federated learning client;

[0243] 4. NRF sends the following information to the federated learning server:

[0244] 1) The capabilities of federated learning clients have changed.

[0245] 2) Changes in the time period / duration during which federated learning can be conducted.

[0246] 5. The federated learning server can make the following decisions:

[0247] 1) Discover new federated learning clients using NRF to perform federated learning.

[0248] 2) The decision-making part of the federated learning client exits federated learning.

[0249] 3) Termination of Federal Learning

[0250] 4) Time between termination of federated learning and resumption of federated learning

[0251] 6. The federated learning server discovers new federated learning clients through NRF;

[0252] 7. The federated learning server sends a federated learning join request to the new federated learning client;

[0253] The federated learning join request may include sample IDs and / or feature IDs used for training.

[0254] 8. The new federated learning client decides whether to join the federated learning system;

[0255] 9. The federated learning server sends a stop federated learning instruction plus the time to restart federated learning to some or all of the old federated learning clients.

[0256] or

[0257] The Federated Learning server sends instructions to some older Federated Learning clients to continue Federated Learning training.

[0258] The instructions for continued training in federated learning may include updating the sample IDs and feature IDs used for training, as well as instructions on whether to reuse the initial model for training.

[0259] Among them, the old federated learning client is the federated learning client that is in the same federated learning process as the federated learning server during the initial training process.

[0260] 10. If the federated learning server does not send a stop federated learning instruction to all old federated learning clients in step 9, the old federated learning clients will return an acknowledgment response message to the federated learning server.

[0261] Solution 2: A specific example of dynamically selecting federated learning clients participating in vertical federated learning inference is provided. Figure 7 is a timing example diagram of dynamic selection of federated learning clients provided in an embodiment of this application. As shown in Figure 7, the specific steps may include the following:

[0262] 1. The federated learning server subscribes to federated learning-related information from the federated learning client;

[0263] 2. The federated learning client sends the following information to the federated learning server:

[0264] 1) Instructions and reasons for exiting vertical federated learning inference (Insufficient computing power; vertical federated learning inference is not supported during this period).

[0265] 2) Local inference data is missing (local inference data is lost, and the corresponding inference data cannot be collected).

[0266] 3) The number of outliers or an increase in the number of outliers

[0267] 4) Changes in sample data or feature data

[0268] 5) Changes in data quality information (changes in missing data information (e.g., changes in the proportion of missing data values), changes in outlier information (e.g., changes in the proportion of outliers)).

[0269] 6) Changes in computing resource information (changes in computing power information (such as changes in CPU / GPU specifications, changes in computing power available for federated learning), changes in storage capacity available for federated learning)

[0270] 7) Unable to return intermediate inference results within the maximum response time specified by the Federated Learning Server.

[0271] 3. The federated learning server subscribes to NRF for capability information related to the federated learning client;

[0272] 4. NRF sends the following information to the federated learning server:

[0273] 1) The capabilities of federated learning clients have changed.

[0274] 2) It allows for changes in the time period / duration of horizontal / vertical federated learning.

[0275] 5. If the federated learning server determines that the accuracy of the federated learning inference results has been affected, it may decide to stop the federated learning inference; or allow some federated learning clients to exit the federated learning inference process and continue the inference process.

[0276] 6. The federated learning server notifies all or some federated learning clients to exit federated learning inference, which may include the following two situations:

[0277] 6a. If federated learning inference continues, the federated learning server sends a federated learning inference stop instruction to some federated learning clients, including the time when inference may be performed again; the federated learning server sends an inference continue instruction to some federated learning clients.

[0278] 6b. If federated learning inference terminates, the federated learning server sends a federated learning inference stop instruction to all federated learning clients, including the time when inference may be performed again.

[0279] 7. The federated learning server may decide to trigger the federated learning preparation / training process again.

[0280] Solution 3: A specific example is provided whereby, before federated learning, an NWDAF / AF, acting as a federated learning client and / or server, reports its federated learning capabilities. Figure 8 is a timing example diagram of federated learning capability reporting provided by an embodiment of this application. As shown in Figure 8, the specific steps may include the following:

[0281] 1. NWDAF and / or AF register or update their federated learning capabilities with NRF, as well as the time periods during which they support various types of federated learning.

[0282] The federated learning capabilities include: whether it supports being used as a horizontal federated learning server, whether it supports being used as a vertical federated learning server, whether it supports being used as a horizontal federated learning client, and whether it supports being used as a vertical federated learning client.

[0283] 2. NRF stores the above capabilities.

[0284] 3. NRF returns a registration or update instruction.

[0285] In one exemplary embodiment, FIG9 is a schematic diagram of a communication device provided in an embodiment of this application. The communication device is applied to a first communication node. As shown in FIG9, the device includes:

[0286] The first decision module 510 is configured to receive first information and make a decision based on the first information.

[0287] The first instruction sending module 520 is configured to send federated learning instructions to at least one second communication node based on the decision result.

[0288] In one embodiment, the first information is provided by a network storage function and / or at least one second communication node.

[0289] In one embodiment, before receiving the first information, at least one of the following is also included:

[0290] Subscribe to first information from at least one second communication node;

[0291] Subscribe to first information from network storage;

[0292] At least one second communication node periodically sends the first information.

[0293] In one embodiment, the first information includes at least one of the following:

[0294] Changes in data samples or feature information of the second communication node;

[0295] Changes in local privacy protection requirements for the second communication node;

[0296] Changes in data quality information of the second communication node;

[0297] The number of outliers in the second communication node or the number of outliers increases;

[0298] Changes in the computing resource information of the second communication node;

[0299] The second communication node is unable to return intermediate training results within the maximum response time specified by the first communication node;

[0300] The capabilities of the second communication node have changed;

[0301] The time period / duration during which federated learning can be conducted can vary.

[0302] In one embodiment, the decision result includes at least one of the following:

[0303] Discover a new second communication node using network storage functionality;

[0304] A portion of the second communication nodes were identified as exiting federated learning.

[0305] Federal learning terminated;

[0306] The termination of federalized learning, and the time for resuming federalized learning.

[0307] In one embodiment, sending a federated learning instruction to at least one second communication node includes at least one of the following:

[0308] Send training termination instructions to some of the second communication nodes;

[0309] Send training termination instructions to all second communication nodes;

[0310] Send training termination instructions and the time to restart federated learning to some of the second communication nodes;

[0311] Send training termination instructions and the time to perform federated learning again to all second communication nodes;

[0312] Send training continuation instructions to some of the second communication nodes.

[0313] In one embodiment, the training continues instruction includes at least one of the following:

[0314] Updated sample identifiers used for training;

[0315] Updated feature labels used for training;

[0316] Instructions on whether to reuse the initial model for training.

[0317] In one embodiment, if the training termination instruction is not sent to all second communication nodes, the method further includes:

[0318] Receive confirmation response messages returned by the second communication node;

[0319] Among them, some of the second communication nodes were those that did not receive the training termination instruction.

[0320] In one embodiment, if the decision result is that a new second communication node is discovered through the network storage function, the method further includes:

[0321] Discover a new second communication node using network storage functionality;

[0322] Send a federated learning join request message to the new second communication node;

[0323] Receive a response message for the request to join a new second communication node.

[0324] In one embodiment, the federated learning join request message includes at least one of the following:

[0325] Sample identifiers used for training;

[0326] Feature identifiers used for training.

[0327] In one embodiment, the communication device is used in the federated learning training process.

[0328] In one embodiment, the first communication node is a horizontal federated learning server or a vertical federated learning server in federated learning, and the second communication node is a horizontal federated learning client or a vertical federated learning client in federated learning.

[0329] In one exemplary embodiment, FIG10 is a schematic diagram of a communication device provided in an embodiment of this application. The communication device is applied to a second communication node. As shown in FIG10, the device includes:

[0330] The first information reporting module 610 is configured to report first information to the first communication node;

[0331] The first instruction receiving module 620 is configured to receive federated learning instructions.

[0332] In one embodiment, the first information includes at least one of the following:

[0333] Changes in data samples or feature information of the second communication node;

[0334] Changes in local privacy protection requirements for the second communication node;

[0335] Changes in data quality information of the second communication node;

[0336] Changes in the computing resource information of the second communication node;

[0337] The second communication node is unable to return intermediate training results within the longest response time of the first communication node;

[0338] The number of outliers in the second communication node or the number of outliers increases;

[0339] The capabilities of the second communication node have changed;

[0340] The time period / duration during which federated learning can be conducted can vary.

[0341] In one embodiment, the first indication receiving module 620 is configured to be at least one of the following:

[0342] Receive training termination instruction;

[0343] The time to receive training termination instructions and to resume federal learning.

[0344] In one embodiment, the first instruction receiving module 620 is further configured to receive a training continuation instruction.

[0345] In one embodiment, the training continues instruction includes at least one of the following:

[0346] Updated sample identifiers used for training;

[0347] Updated feature labels used for training;

[0348] Instructions on whether to reuse the initial model for training.

[0349] In one embodiment, if the second communication node is a new second communication node discovered by the first communication node through the network storage function, the method further includes:

[0350] Receive the federated learning join request message sent by the first communication node;

[0351] Return a join request response message to the first communication node.

[0352] In one embodiment, the federated learning join request message includes at least one of the following:

[0353] Sample identifiers used for training;

[0354] Feature identifiers used for training.

[0355] In one embodiment, the communication device is used in the federated learning training process.

[0356] In one embodiment, the second communication node is a horizontal federated learning client or a vertical federated learning client in federated learning, and the first communication node is a horizontal federated learning server or a vertical federated learning server in federated learning.

[0357] In one exemplary embodiment, FIG11 is a schematic diagram of a communication device provided in an embodiment of the present application. The communication device is applied to a first communication node. As shown in FIG11, the device includes:

[0358] The second decision module 710 is configured to receive second information and make a decision based on the first information.

[0359] The second instruction sending module 720 is configured to send federated learning instructions to at least one second communication node based on the decision result.

[0360] In one embodiment, the second information is provided by a network storage function and / or at least one second communication node.

[0361] In one embodiment, before receiving the second information, at least one of the following is also included:

[0362] Subscribe to second information from at least one second communication node;

[0363] Subscribe to second information from network storage;

[0364] At least one second communication node periodically sends a second message.

[0365] In one embodiment, the second information includes at least one of the following:

[0366] The second communication node lacks local inference data;

[0367] The number of outliers in the second communication node or the number of outliers increases;

[0368] Changes in sample data or feature data of the second communication node;

[0369] Changes in data quality information of the second communication node;

[0370] Changes in the computing resource information of the second communication node;

[0371] The second communication node is unable to return intermediate inference results within the maximum response time specified by the first communication node;

[0372] The capabilities of the second communication node have changed;

[0373] The time period / duration during which federated learning can be conducted can vary.

[0374] In one embodiment, the decision result includes at least one of the following:

[0375] Federalized learning reasoning terminated;

[0376] It was determined that some of the second communication nodes exited the federated learning inference process, and the inference process continued.

[0377] In one embodiment, if the decision result is that the federated learning inference is terminated, the second instruction sending module 720 is specifically configured to send an inference termination instruction and the time for re-performing federated learning inference to all second communication nodes.

[0378] In one embodiment, if the decision result determines that some second communication nodes are exiting federated learning inference and the inference process continues, the second instruction sending module 720 is specifically configured to send an inference termination instruction and a time for resuming federated learning inference to some second communication nodes; and send an inference continuation instruction to the other second communication nodes.

[0379] In one embodiment, the communication device is used in the federated learning inference process.

[0380] In one embodiment, the first communication node is a horizontal federated learning server or a vertical federated learning server in federated learning, and the second communication node is a horizontal federated learning client or a vertical federated learning client in federated learning.

[0381] In one exemplary embodiment, FIG12 is a schematic diagram of a communication device provided in an embodiment of the present application. The communication device is applied to a second communication node. As shown in FIG12, the device includes:

[0382] The second information reporting module 810 is configured to report second information to the first communication node;

[0383] The second instruction receiving module 820 is configured to receive federated learning instructions.

[0384] In one embodiment, the second information includes at least one of the following:

[0385] The second communication node lacks local inference data;

[0386] The number of outliers in the second communication node or the number of outliers increases;

[0387] Changes in sample data or feature data of the second communication node;

[0388] Changes in data quality information of the second communication node;

[0389] Changes in the computing resource information of the second communication node;

[0390] The second communication node is unable to return intermediate inference results within the maximum response time specified by the first communication node;

[0391] The capabilities of the second communication node have changed;

[0392] The time period / duration during which federated learning can be conducted can vary.

[0393] In one embodiment, the second instruction receiving module 820 is specifically configured to receive an inference termination instruction and the time for resuming federated learning inference.

[0394] In one embodiment, the second indication receiving module 820 is further configured to receive a reasoning continuation indication.

[0395] In one embodiment, the communication method is applied to the federated learning inference process.

[0396] In one embodiment, the second communication node is a horizontal federated learning client or a vertical federated learning client in federated learning, and the first communication node is a horizontal federated learning server or a vertical federated learning server in federated learning.

[0397] This application embodiment also provides a communication node. Figure 13 is a structural schematic diagram of a communication node provided in this application embodiment. As shown in Figure 13, the communication node provided in this application embodiment includes a memory 920, a processor 910, and a computer program stored in the memory and executable on the processor. When the processor 910 executes the program, it implements the above-mentioned communication method.

[0398] The communication node may also include a memory 920; the processor 910 in the communication node may be one or more, with one processor 910 as an example in Figure 13; the memory 920 is used to store one or more programs; the one or more programs are executed by the one or more processors 910, so that the one or more processors 910 implement the communication method as described in the embodiments of this application.

[0399] The communication node also includes: a communication device 930, an input device 940, and an output device 950.

[0400] The processor 910, memory 920, communication device 930, input device 940 and output device 950 in the communication node can be connected by a bus or other means. Figure 13 shows an example of connection by bus.

[0401] Input device 940 can be used to receive input digital or character information, and to generate key signal inputs related to user settings and function control of the communication node. Output device 950 may include display devices such as a display screen.

[0402] The communication device 930 may include a receiver and a transmitter. The communication device 930 is configured to perform information transmission and reception communication under the control of the processor 910.

[0403] The memory 920, as a computer-readable storage medium, can be configured to store software programs, computer-executable programs, and modules, such as program instructions / modules corresponding to the communication method described in the embodiments of this application (e.g., a first decision module 510, a first instruction sending module 520; or a first information reporting module 610, a first instruction receiving module 620; or a second decision module 710, a second instruction sending module 720; or a second information reporting module 810, a second instruction receiving module 820). The memory 920 may include a program storage area and a data storage area, wherein the program storage area may store the operating system, an application program required for at least one function, and the data storage area may store data created according to the use of the communication node, etc. Furthermore, the memory 920 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 920 may further include memory remotely located relative to the processor 910, and these remote memories can be connected to the communication node via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0404] This application also provides a storage medium storing a computer program, which, when executed by a processor, implements any of the communication methods described in this application.

[0405] Optionally, the communication method, applied to a first communication node, includes: receiving first information and making a decision based on the first information; and sending a federated learning instruction to at least one second communication node based on the decision result.

[0406] Optionally, the communication method, applied to the second communication node, includes: reporting first information to the first communication node corresponding to the second communication node; and receiving federated learning instructions.

[0407] Optionally, the communication method, applied to a first communication node, includes: receiving second information and making a decision based on the second information; and sending a federated learning instruction to at least one second communication node based on the decision result.

[0408] Optionally, the communication method, applied to the second communication node, includes: reporting second information to the first communication node; and receiving federated learning instructions.

[0409] The computer storage medium in this application embodiment can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. For example, a computer-readable storage medium can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable CD-ROM, optical storage device, magnetic storage device, or any suitable combination thereof. The computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0410] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit programs for use by or in connection with an instruction execution system, apparatus, or device.

[0411] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, radio frequency (RF), etc., or any suitable combination thereof.

[0412] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0413] Optionally, embodiments of the present invention also provide a computer program product, including a computer program that, when executed by a processor, implements the communication method provided in any embodiment of the present invention.

[0414] The above description is merely an exemplary embodiment of this application and is not intended to limit the scope of protection of this application.

[0415] Those skilled in the art will understand that the term user terminal encompasses any suitable type of wireless user equipment, such as mobile phones, portable data processing devices, portable web browsers, or vehicle-mounted mobile stations.

[0416] Generally, the various embodiments of this application can be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. For example, some aspects can be implemented in hardware, while others can be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device, although this application is not limited thereto.

[0417] Embodiments of this application can be implemented by executing computer program instructions through the data processor of a mobile device, for example, in a processor entity, or through hardware, or through a combination of software and hardware. The computer program instructions can be assembly instructions, Instruction Set Architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages.

[0418] Any block diagram of logical flow in the accompanying drawings of this application may represent program steps, or may represent interconnected logic circuits, modules, and functions, or may represent a combination of program steps and logic circuits, modules, and functions. The computer program may be stored on memory. Memory may be of any type suitable to the local technical environment and may be implemented using any suitable data storage technology, such as, but not limited to, read-only memory (ROM), random access memory (RAM), optical storage devices and systems (Digital Video Disc (DVD) or Compact Disk (CD), etc.). Computer-readable media may include non-transitory storage media. Data processors may be of any type suitable to the local technical environment, such as, but not limited to, general-purpose computers, special-purpose computers, microprocessors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), and processors based on multi-core processor architectures.

[0419] A detailed description of exemplary embodiments of this application has been provided above through exemplary and non-limiting examples. However, various modifications and adjustments to the above embodiments will be apparent to those skilled in the art when considered in conjunction with the accompanying drawings and claims, without departing from the scope of this application. Therefore, the proper scope of this application will be determined by the claims.

Claims

1. A communication method, characterized in that, Applied to a first communication node, the method includes: receiving first information and making a decision based on the first information; and sending a federated learning instruction to at least one second communication node based on the decision result.

2. The communication method according to claim 1, characterized in that, The first information is provided by network storage functionality and / or the at least one second communication node.

3. The communication method according to claim 2, characterized in that, Before receiving the first information, the method includes at least one of the following: subscribing to the first information from the at least one second communication node; subscribing to the first information from the network storage function; and the at least one second communication node periodically sending the first information.

4. The communication method according to claim 1, characterized in that, The first information includes at least one of the following: changes in the data samples or feature information of the second communication node; changes in the local privacy protection requirements of the second communication node; changes in the data quality information of the second communication node; an increase in the number of outliers in the second communication node; changes in the computing resource information of the second communication node; the second communication node being unable to return intermediate training results within the longest response time specified by the first communication node; changes in the capabilities of the second communication node; and changes in the time period / duration during which federated learning can be performed.

5. The communication method according to claim 1, characterized in that, The decision result includes at least one of the following: discovering a new second communication node through network storage function; determining that some of the second communication nodes have exited federated learning; Federal learning terminated; The termination of federalized learning, and the time for resuming federalized learning.

6. The communication method according to claim 1, characterized in that, Sending federated learning instructions to at least one second communication node includes at least one of the following: sending training termination instructions to some of the second communication nodes; sending training termination instructions to all the second communication nodes; sending training termination instructions and the time for resuming federated learning to some of the second communication nodes; sending training termination instructions and the time for resuming federated learning to all the second communication nodes; and sending training continue instructions to some of the second communication nodes.

7. The communication method according to claim 6, characterized in that, The training continuation instruction includes at least one of the following: updated sample identifiers for training; updated feature identifiers for training; and an instruction on whether to retrain using the initial model.

8. The communication method according to claim 6, characterized in that, If the training termination instruction is not sent to all the second communication nodes, the method further includes: receiving confirmation response messages returned by some of the second communication nodes; wherein, the second communication nodes are those that did not receive the training termination instruction.

9. The communication method according to claim 5, characterized in that, If the decision result is that a new second communication node is discovered through the network storage function, the method further includes: discovering a new second communication node through the network storage function; sending a federated learning join request message to the new second communication node; and receiving a join request response message from the new second communication node.

10. The communication method according to claim 9, characterized in that, The federated learning join request message includes at least one of the following: sample identifiers for training; feature identifiers for training.

11. The communication method according to any one of claims 1-10, characterized in that, The communication method is applied during the federated learning training process.

12. The communication method according to any one of claims 1-10, characterized in that, The first communication node is a horizontal federated learning server or a vertical federated learning server in federated learning, and the second communication node is a horizontal federated learning client or a vertical federated learning client in federated learning.

13. A communication method, characterized in that, Applied to the second communication node, it includes: reporting first information to the first communication node; and receiving federated learning instructions.

14. The communication method according to claim 13, characterized in that, The first information includes at least one of the following: changes in the data samples or feature information of the second communication node; changes in the local privacy protection requirements of the second communication node; changes in the data quality information of the second communication node; changes in the computing resource information of the second communication node; the second communication node being unable to return intermediate training results within the longest response time of the first communication node; an increase in the number of outliers in the second communication node; changes in the capabilities of the second communication node; and changes in the time period / duration during which federated learning can be performed.

15. The communication method according to claim 13, characterized in that, The receipt of federated learning instructions includes at least one of the following: receiving a training termination instruction; receiving a training termination instruction and the time for resuming federated learning.

16. The communication method according to claim 13, characterized in that, The receiving of federated learning instructions also includes receiving training continuation instructions.

17. The communication method according to claim 16, characterized in that, The training continuation instruction includes at least one of the following: updated sample identifiers for training; updated feature identifiers for training; and an instruction on whether to retrain using the initial model.

18. The communication method according to claim 13, characterized in that, If the second communication node is a new second communication node discovered by the first communication node through the network storage function, the method further includes: receiving a federated learning join request message sent by the first communication node; and returning a join request response message to the first communication node.

19. The communication method according to claim 18, characterized in that, The federated learning join request message includes at least one of the following: sample identifiers for training; feature identifiers for training.

20. The communication method according to any one of claims 13-19, characterized in that, The communication method is applied during the federated learning training process.

21. The communication method according to any one of claims 13-19, characterized in that, The second communication node is a horizontal federated learning client or a vertical federated learning client in federated learning, and the first communication node is a horizontal federated learning server or a vertical federated learning server in federated learning.

22. A communication method, characterized in that, The method is applied to a first communication node and includes: receiving second information and making a decision based on the second information; and sending a federated learning instruction to at least one second communication node based on the decision result.

23. The communication method according to claim 22, characterized in that, The second information is provided by network storage functionality and / or by the at least one second communication node.

24. The communication method according to claim 23, characterized in that, Before receiving the second information, at least one of the following is included: subscribing to the second information from the at least one second communication node; subscribing to the second information from the network storage function; and the at least one second communication node periodically sending the second information.

25. The communication method according to claim 22, characterized in that, The second information includes at least one of the following: missing local inference data of the second communication node; an increase in the number of outliers of the second communication node; changes in the sample data or feature data of the second communication node; changes in the data quality information of the second communication node; changes in the computing resource information of the second communication node; the second communication node is unable to return intermediate inference results within the longest response time specified by the first communication node; changes in the capabilities of the second communication node; and changes in the time period / duration during which federated learning can be performed.

26. The communication method according to claim 22, characterized in that, The decision result includes at least one of the following: the federated learning inference is terminated; some of the second communication nodes are determined to have exited the federated learning inference and the inference process continues.

27. The communication method according to claim 26, characterized in that, If the decision result is that the federated learning inference is terminated, the step of sending a federated learning instruction to at least one second communication node based on the decision result includes: sending an inference termination instruction to all second communication nodes and the time for re-performing federated learning inference.

28. The communication method according to claim 26, characterized in that, If the decision result determines that some of the second communication nodes exit the federated learning inference and continue the inference process, the step of sending a federated learning instruction to at least one second communication node according to the decision result includes: sending an inference termination instruction and a time for re-performing federated learning inference to the portion of the second communication nodes; and sending an inference continuation instruction to the second communication nodes other than the portion of the second communication nodes.

29. The communication method according to any one of claims 22-28, characterized in that, The communication method is applied in the federated learning reasoning process.

30. The communication method according to any one of claims 22-28, characterized in that, The first communication node is a horizontal federated learning server or a vertical federated learning server in federated learning, and the second communication node is a horizontal federated learning client or a vertical federated learning client in federated learning.

31. A communication node, characterized in that, include: The program includes a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for implementing communication between the processor and the memory, wherein the program, when executed by the processor, implements the steps of the communication method as described in any one of claims 1-30.

32. A storage medium for computer-readable storage, characterized in that, The storage medium stores one or more programs, which can be executed by one or more processors to implement the steps of the communication method as described in any one of claims 1-30.

33. A computer program product comprising a computer program that, when executed by a processor, implements the steps of the communication method as described in any one of claims 1-30.

Images