Communication method and apparatus

Abstract

The present disclosure provides a communication method and apparatus for improving performance of a federated learning model, the method includes the steps of: acquiring an i-th first information, the i-th first information indicating an i-th local model of each of a plurality of subnodes, where i is a positive integer; determining an i-th federated learning model of a k-th subnode based on the i-th first information and a type of the i-th federated learning model of the k-th subnode, where k is any positive integer from 1 to K, where K is a number of a plurality of subnodes, and when i is greater than 1, the type of the i-th federated learning model of the k-th subnode is determined based on a type or performance of the (i-1)-th federated learning model of the k-th subnode; and transmitting an i-th second information, the i-th second information indicating an i-th federated learning model of each of a plurality of subnodes.

Description

[Technical field]

[0001] The present disclosure relates to the field of communication technologies, and more particularly to communication methods and devices. [Background technology]

[0002] In wireless communication networks, e.g., mobile communication networks, the services supported by the network are becoming more and more diverse, and therefore the requirements that need to be met are also becoming more and more diverse. For example, the network needs to support ultra-high speed, ultra-low latency, and / or ultra-large connections. These characteristics make network planning, network configuration, and / or resource scheduling more complex. In addition, network capabilities are becoming more powerful, e.g., supporting increasingly higher spectrum, supporting new technologies such as high-order multiple input multiple output (MIMO) technology, beamforming, and / or beam management, so that network energy saving has become a popular research topic. These new requirements, scenarios, and characteristics bring unprecedented challenges to network planning, operation and maintenance, and efficient operation. To address this challenge, artificial intelligence technology can be introduced into wireless communication networks to implement network intelligence. Based on this, how to effectively implement artificial intelligence in networks is a problem worth studying. Summary of the Invention [Problem to be solved by the invention]

[0003] The present disclosure provides a communication method and apparatus for improving the performance of a federated learning model.

[0004] According to a first aspect, the present disclosure provides a communication method including: acquiring an i-th first information, where the i-th first information indicates an i-th local model of each of a plurality of subnodes, where i is a positive integer; determining an i-th federated learning model of the k-th subnode based on the i-th first information and a type of the i-th federated learning model of the k-th subnode, where k is any positive integer from 1 to K, where K is a number of a plurality of subnodes, and when i is greater than 1, the type of the i-th federated learning model of the k-th subnode is determined based on a type or performance of the (i-1)-th federated learning model of the k-th subnode; and transmitting an i-th second information, where the i-th second information indicates the i-th federated learning model of each of the plurality of subnodes.

[0005] In this design, the performance or type of the historical federated learning model is considered, and the federated learning model for each subnode participating in the federated learning model is determined separately to implement personalized federated learning, which helps to improve the performance of the federated learning model.

[0006] In a possible design, the i-th federated learning model of the k-th subnode is the same as or different from the i-th federated learning model of at least one subnode other than the k-th subnode in the plurality of subnodes. By using such a design, data distributions of the subnodes can be matched, and different models are provided for subnodes whose data distributions differ significantly from the data distributions of other subnodes to implement personalized federated learning.

[0007] In a possible design, the type of the ith federated learning model of the kth subnode may be determined in the manner of: obtaining a first test result from the kth subnode, where the first test result indicates the performance of the (i-1)th federated learning model of the kth subnode, where i is greater than 1; and determining the type of the ith federated learning model of the kth subnode based on the first test result. In this design, to help train a federated learning model that matches the data distribution of the subnode, the type of the current federated learning model is determined for the subnode with reference to the performance of the subnode's historical federated learning models.

[0008] In another possible design, the type of the ith federated learning model of the kth subnode may be determined in the following manner: sending information indicating a first model to a subnode having a kth order, where the first model is determined based on the (i-1)th first information and a weight corresponding to the (i-1)th local model of each of the plurality of subnodes, where i is greater than 1; obtaining a second test result from the kth subnode, where the second test result indicates a performance of the first model, or the second test result indicates a performance of the first model and a performance of the (i-1)th federated learning model of the kth subnode; and determining the type of the ith federated learning model of the kth subnode based on the second test result. In this design, the type of the current federated learning model is determined with reference to the first model, i.e., a historical federated learning model of a global type, to help train a federated learning model that matches the data distribution of the subnode.

[0009] In a possible design, determining the i-th federated learning model of the k-th subnode based on the i-th first information and a type of the i-th federated learning model of the k-th subnode includes determining the i-th federated learning model of the k-th subnode based on the i-th first information and a weight corresponding to the i-th local model of each of the multiple subnodes when the type of the i-th federated learning model of the k-th subnode corresponds to a first value, or determining the i-th federated learning model of the k-th subnode based on the i-th first information and a similarity between the i-th local model of the k-th subnode and the i-th local model of each of the multiple subnodes when the type of the i-th federated learning model of the k-th subnode corresponds to a second value. In this design, different model calculation schemes are used for different types of federated learning models to determine the federated learning models of the associated subnodes, so that performance of the federated learning models can be improved.

[0010] In a possible design, the plurality of subnodes include a first group of subnodes and a second group of subnodes, wherein the i-th federated learning models of all subnodes in the first group of subnodes are the same and the i-th federated learning models of all subnodes in the second group of subnodes are different and differ from the i-th federated learning models of all subnodes in the first group of subnodes, and the step of sending the i-th second information includes the steps of sending first instruction information to the first group of subnodes in a broadcast or multicast manner, wherein the first instruction information indicates the i-th federated learning model of each subnode in the first group of subnodes, and sending second instruction information to a p-th subnode in the second group of subnodes in a unicast manner, wherein the second instruction information indicates the i-th federated learning model of the p-th subnode, where p is any positive integer from 1 to P, and P is the number of subnodes included in the second group of subnodes.

[0011] In another possible design, the i-th federated learning model is represented by a first portion of parameter information and a second portion of parameter information, the first portions of the parameter information of all the i-th federated learning models of the plurality of subnodes are the same, the plurality of subnodes include a first group of subnodes and a second group of subnodes, the second portions of the parameter information of all the subnodes in the first group of subnodes are the same and the second portions of the parameter information of all the subnodes in the second group of subnodes are different and the second portions of the parameter information of the i-th federated learning models of all the subnodes in the first group of subnodes are different, and the step of transmitting the i-th second information includes: The method includes the steps of: transmitting a first portion of parameter information of the i-th federated learning model to a plurality of subnodes in a broadcast or multicast manner; transmitting a second portion of the parameter information of the i-th federated learning model of each subnode in the first group of subnodes to the first group of subnodes in a broadcast or multicast manner; and transmitting the second portion of the parameter information of the p-th subnode to the p-th subnode in the second group of subnodes in a unicast manner, where p is any positive integer from 1 to P, and P is the number of subnodes included in the second group of subnodes.

[0012] In the above design of transmitting the i-th second information, the same information corresponding to two or more subnodes is transmitted in a broadcast or multicast manner, so that resources for transmitting the i-th second information can be saved.

[0013] In a possible design, the method further includes a step of transmitting information indicative of the second model to a plurality of subnodes, the information indicative of the second model being used by each of the plurality of subnodes to determine the first local model.

[0014] According to a second aspect, the present disclosure provides a communication method including the steps of: transmitting information indicating an i-th local model of a k-th subnode, where k is any positive integer from 1 to K, K is a number of multiple subnodes participating in federated learning, and i is a positive integer; and acquiring information indicating an i-th federated learning model of the k-th subnode, where the i-th federated learning model of the k-th subnode is determined based on the i-th first information and a type of the i-th federated learning model of the k-th subnode. The i-th first information includes information indicating an i-th local model of each of multiple subnodes, and when i is greater than 1, the type of the i-th federated learning model of the k-th subnode is determined based on a type or performance of the (i-1)-th federated learning model of the k-th subnode.

[0015] In a possible design, the i-th federated learning model of the k-th subnode is the same as or different from the i-th federated learning model of at least one subnode other than the k-th subnode in the plurality of subnodes.

[0016] In a possible design, the method further includes a step of transmitting a first test result, the first test result indicating a performance of the (i-1)th federated learning model of the kth subnode, where i is greater than 1, and a type of the ith federated learning model of the kth subnode is determined based on the first test result.

[0017] In a possible design, the method further includes a step of acquiring information indicative of a first model, where the first model is determined based on the (i-1)th first information and a weight corresponding to the (i-1)th local model of each of the multiple subnodes, where i is greater than 1; and a step of sending second test results, where the second test results indicate performance of the first model, or the second test results indicate performance of the first model and performance of the (i-1)th federated learning model of the kth subnode, and a type of the i-th federated learning model of the kth subnode is determined based on the second test results.

[0018] In a possible design, when the type of the ith federated learning model of the kth subnode corresponds to a first value, the ith federated learning model of the kth subnode is determined based on the ith first information and a weight corresponding to the ith local model of each of the multiple subnodes, or when the type of the ith federated learning model of the kth subnode corresponds to a second value, the ith federated learning model of the kth subnode is determined based on the ith first information and a similarity between the ith local model of the kth subnode and the ith local model of each of the multiple subnodes.

[0019] In a possible design, the method further includes obtaining information indicative of the second model, and determining a first local model of the kth subnode based on the information indicative of the second model.

[0020] According to a third aspect, the present disclosure provides a communication apparatus. The communication apparatus may be a central node, an apparatus in the central node, or an apparatus that can be used with a central node. In one design, the communication apparatus may include modules that correspond one-to-one to methods / operations / steps / actions according to the first aspect. The modules may be implemented by hardware circuits, software, or a combination of hardware circuits and software. In a design, the communication apparatus may include a processing module and a communication module.

[0021] In one example, the communication module is configured to obtain an i-th first information, where the i-th first information indicates an i-th local model of each of the multiple subnodes, where i is a positive integer.

[0022] The processing module is configured to determine an i-th federated learning model of the k-th subnode based on the i-th first information and a type of the i-th federated learning model of the k-th subnode, where k is any positive integer from 1 to K, K is the number of multiple subnodes, and when i is greater than 1, the type of the i-th federated learning model of the k-th subnode is determined based on the type or performance of the (i-1)-th federated learning model of the k-th subnode.

[0023] The communication module is further configured to transmit the i-th second information, where the i-th second information indicates the i-th federated learning model of each of the plurality of subnodes.

[0024] In a possible design, the i-th federated learning model of the k-th subnode is the same as or different from the i-th federated learning model of at least one subnode other than the k-th subnode in the plurality of subnodes.

[0025] In a possible design, the communication module is further configured to obtain a first test result from the kth subnode, the first test result indicating a performance of the (i-1)th federated learning model of the kth subnode, where i is greater than 1, and the processing module is further configured to determine a type of the ith federated learning model of the kth subnode based on the first test result.

[0026] In another possible design, the communication module is further configured to send information indicative of the first model to the kth subnode, where the first model is determined based on the (i-1)th first information and a weight corresponding to the (i-1)th local model of each of the multiple subnodes, where i is greater than 1. The communication module is further configured to obtain a second test result from the kth subnode, where the second test result indicates a performance of the first model, or the second test result indicates a performance of the first model and a performance of the (i-1)th federated learning model of the kth subnode. The processing module is further configured to determine a type of the i-th federated learning model of the kth subnode based on the second test result.

[0027] In a possible design, the processing module is further configured to determine the i-th federated learning model of the k-th subnode based on the i-th first information and a weight corresponding to the i-th local model of each of the multiple subnodes when the type of the i-th federated learning model of the k-th subnode corresponds to a first value, or to determine the i-th federated learning model of the k-th subnode based on the i-th first information and a similarity between the i-th local model of the k-th subnode and the i-th local model of each of the multiple subnodes when the type of the i-th federated learning model of the k-th subnode corresponds to a second value.

[0028] In a possible design, the plurality of subnodes include a first group of subnodes and a second group of subnodes, where the i-th federated learning model of all subnodes in the first group of subnodes is the same and the i-th federated learning model of all subnodes in the second group of subnodes is different and different from the i-th federated learning model of all subnodes in the first group of subnodes. The communication module is further configured to send first instruction information to the first group of subnodes in a broadcast or multicast manner, where the first instruction information indicates the i-th federated learning model of each subnode in the first group of subnodes, and send second instruction information to the p-th subnode in the second group of subnodes in a unicast manner, where the second instruction information indicates the i-th federated learning model of the p-th subnode, where p is any positive integer from 1 to P, and P is the number of subnodes included in the second group of subnodes.

[0029] In another possible design, the i-th federated learning model is represented by a first portion of parameter information and a second portion of parameter information, the first portion of the parameter information of all the i-th federated learning models of the plurality of subnodes are the same, the plurality of subnodes include a first group of subnodes and a second group of subnodes, the second portions of the parameter information of the i-th federated learning models of all subnodes in the first group of subnodes are the same and the second portions of the parameter information of the i-th federated learning models of all subnodes in the second group of subnodes are different and the second portions of the parameter information of the i-th federated learning models of all subnodes in the first group of subnodes are different, and the communication module The method is further configured to: transmit a first portion of the parameter information of the i-th federated learning model to the multiple subnodes in a broadcast or multicast manner; transmit a second portion of the parameter information of the i-th federated learning model of each subnode in the first group of subnodes to the first group of subnodes in a broadcast or multicast manner; and transmit a second portion of the parameter information of the p-th subnode to the p-th subnode in the second group of subnodes in a unicast manner, where p is any positive integer from 1 to P, and P is the number of subnodes included in the second group of subnodes.

[0030] In a possible design, the communication module is further configured to transmit information indicative of the second model to the multiple subnodes, where the information indicative of the second model is used by each of the multiple subnodes to determine the first local model.

[0031] According to a fourth aspect, the present disclosure provides a communication apparatus. The communication apparatus may be a kth subnode, or may be an apparatus in the kth subnode, or may be an apparatus that can be used with the kth subnode. In one design, the communication apparatus may include modules that correspond one-to-one to methods / operations / steps / actions according to the first aspect. The modules may be implemented by hardware circuits, software, or a combination of hardware circuits and software. In a design, the communication apparatus may include a processing module and a communication module.

[0032] In one example, the processing module is configured to transmit information indicating the i-th local model of the k-th subnode by using the communication module, where k is any positive integer from 1 to K, K is the number of multiple subnodes participating in the federated learning, and i is a positive integer, and the communication module is further configured to obtain information indicating the i-th federated learning model of the k-th subnode, where the i-th federated learning model of the k-th subnode is determined based on the i-th first information and the type of the i-th federated learning model of the k-th subnode. The i-th first information includes information indicating the i-th local model of each of the multiple subnodes, and when i is greater than 1, the type of the i-th federated learning model of the k-th subnode is determined based on the type or performance of the (i-1)-th federated learning model of the k-th subnode.

[0033] In a possible design, the i-th federated learning model of the k-th subnode is the same as or different from the i-th federated learning model of at least one subnode other than the k-th subnode in the plurality of subnodes.

[0034] In a possible design, the processing module is further configured to send first test results by using the communication module, the first test results indicating a performance of the (i-1)th federated learning model of the kth subnode, where i is greater than 1, and a type of the ith federated learning model of the kth subnode is determined based on the first test results.

[0035] In a possible design, the communication module is further configured to obtain information indicative of the first model, the first model being determined based on the (i-1)th first information and a weight corresponding to the (i-1)th local model of each of the multiple subnodes, where i is greater than 1, and the processing module is further configured to transmit second test results by using the communication module, the second test results indicating performance of the first model, or the second test results indicating performance of the first model and performance of the (i-1)th federated learning model of the kth subnode, and the type of the i-th federated learning model of the kth subnode is determined based on the second test results.

[0036] For the content description of the type of the i-th federated learning model of the k-th subnode, please refer to the second aspect, and the details will not be described again in this specification.

[0037] In a possible design, the communication module is further configured to obtain information indicative of the second model, and the processing module is further configured to determine a first local model of the kth subnode based on the information indicative of the second model.

[0038] In a possible design of any one of the first to fourth aspects, 1≦i≦I, I is a positive integer, and the I-th federated learning model of the k-th subnode satisfies a model convergence condition, and when i is greater than 1, the i-th local model of the k-th subnode is determined based on the (i-1)-th federated learning model of the k-th subnode.

[0039] According to a fifth aspect, the present disclosure provides a communication device. The communication device includes a processor configured to perform the method according to the first aspect. The communication device may further include a memory configured to store instructions and data. The memory is coupled to the processor. When executing the instructions stored in the memory, the processor can perform the method according to the first aspect. The communication device may further include a communication interface. The communication interface is used by the device to communicate with another device. For example, the communication interface may be a transceiver, a circuit, a bus, a module, a pin, or another type of communication interface. In a possible device, the communication device may include: a memory configured to store program instructions; a processor configured to receive an i-th first information via a communication interface, the i-th first information indicating an i-th local model of each of a plurality of subnodes, where i is a positive integer; Including, The processor is further configured to determine an i-th federated learning model of the k-th subnode based on the i-th first information and a type of the i-th federated learning model of the k-th subnode, where k is any positive integer from 1 to K, and K is a number of the plurality of subnodes, and when i is greater than 1, the type of the i-th federated learning model of the k-th subnode is determined based on a type or performance of the (i-1)-th federated learning model of the k-th subnode; The processor is further configured to transmit the i-th second information via the communication interface, the i-th second information indicating the i-th federated learning model for each of the plurality of subnodes.

[0040] According to a sixth aspect, the present disclosure provides a communication device. The communication device includes a processor configured to perform the method according to the second aspect. The communication device may further include a memory configured to store instructions and data. The memory is coupled to the processor. When executing the instructions stored in the memory, the processor can perform the method according to the second aspect. The communication device may further include a communication interface. The communication interface is used by the device to communicate with another device. For example, the communication interface may be a transceiver, a circuit, a bus, a module, a pin, or another type of communication interface. In a possible device, the communication device may include: a memory configured to store program instructions; A processor configured to transmit information indicating an i-th local model of a k-th subnode via a communication interface, where k is any positive integer from 1 to K, K is a number of subnodes participating in the federated learning, and i is a positive integer; Including, The processor is further configured to obtain, via the communication interface, information indicative of an ith federated learning model of the kth subnode, where the ith federated learning model of the kth subnode is determined based on the ith first information and a type of the ith federated learning model of the kth subnode, where the ith first information includes information indicative of an ith local model of each of the plurality of subnodes, where when i is greater than 1, the type of the ith federated learning model of the kth subnode is determined based on a type or performance of the (i-1)th federated learning model of the kth subnode.

[0041] According to a seventh aspect, the present disclosure provides a communication system including a communication device according to the third or fifth aspect and a communication device according to the fourth or sixth aspect.

[0042] According to an eighth aspect, the present disclosure further provides a computer program, which when executed on a computer enables the computer to carry out the method according to the first or second aspect.

[0043] According to a ninth aspect, the present disclosure further provides a computer program product comprising instructions which, when executed in a computer, enable the computer to perform a method according to the first or second aspect.

[0044] According to a tenth aspect, the present disclosure further provides a computer readable storage medium, the computer readable storage medium storing a computer program or instructions, which when executed in a computer, enables the computer to perform a method according to the first or second aspect.

[0045] According to an eleventh aspect, the present disclosure further provides a chip, the chip being configured to read a computer program stored in a memory to perform a method according to the first or second aspect.

[0046] According to a twelfth aspect, the present disclosure further provides a chip system. The chip system includes a processor configured to support a computer device in performing a method according to the first or second aspect. In a possible design, the chip system further includes a memory, the memory configured to store programs and data required for the computer device. The chip system may include a chip, or may include a chip and another discrete device. [Brief description of the drawings]

[0047] [Figure 1A] FIG. 1 is a diagram of the structure of a communication system. [Figure 1B] FIG. 1 is a diagram of the structure of a communication system. [Figure 2A]FIG. 1 is a diagram of a neuron structure. [Figure 2B] FIG. 1 is a diagram of the hierarchical relationship of a neural network. [Figure 3A] FIG. 2 is a diagram of a network structure corresponding to the communication method according to the present disclosure. [Figure 3B] FIG. 2 is a diagram of a network structure corresponding to the communication method according to the present disclosure. [Figure 4A] FIG. 1 is a diagram of node interactions in a related federated learning technique. [Figure 4B] FIG. 1 is a diagram of node interactions for training a federated learning model according to the present disclosure. [Diagram 5] 1 is a schematic flow chart of a communication method according to the present disclosure. [Figure 6] FIG. 1 is a diagram of a model structure division. [Figure 7] 1 is a schematic flow chart of a communication method according to the present disclosure. [Figure 8A] 1 is a schematic flow chart of a method for determining a model type according to the present disclosure. [Figure 8B] 1 is a schematic flow chart of a method for determining a model type according to the present disclosure. [Figure 9] 1 is a schematic flow chart of a communication method according to the present disclosure. [Figure 10A] 1 is a schematic flow chart of a method for determining a model type according to the present disclosure. [Figure 10B] 1 is a schematic flow chart of a method for determining a model type according to the present disclosure. [Figure 11] 1 is a schematic flow chart of a communication method according to the present disclosure. [Figure 12] 1 is a diagram of the structure of a communication device according to the invention; [Figure 13] FIG. 2 is a diagram of the structure of a communication device according to the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0048] In order to make the objectives, technical solutions and advantages of the present disclosure clearer, the present disclosure will be further described in detail hereinafter with reference to the accompanying drawings.

[0049] At least one of the following in this disclosure indicates one or more. "Multiple" means two or more. The term "and / or" describes an association relationship to describe related objects and represents that three relationships may exist. For example, A and / or B can represent the following three cases: when only A exists, when both A and B exist, and when only B exists. The character " / " usually indicates an "or" relationship between related objects. In addition, terms such as "first" and "second" may be used in this disclosure to describe objects, but it should be understood that these objects should not be limited by these terms. These terms are used only to distinguish objects from each other.

[0050] The terms "including", "having" and any other variations thereof mentioned in the following description of the present disclosure are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the recited steps or units, but optionally further includes other unrecited steps or units, or optionally further includes other inherent steps or units of the process, method, product, or device. It should be noted that in the present disclosure, terms such as "example" or "for example" are used to represent giving an example, illustration, or explanation. Any method or design manner described in the present disclosure as "example" or "for example" should not be described as being preferred or having more advantages over another method or design manner. Rather, the use of terms such as "example", "for example", and the like is intended to present the relevant concept in a particular way.

[0051] The techniques provided in this disclosure may be applied to various communication systems. For example, the communication system may be a third generation (3G)rd 3G (third generation) communication systems (e.g., universal mobile telecommunications system (UMTS)), fourth generation (4 th 4G (4th generation) communication systems (e.g., long term evolution (LTE) systems), 5th generation (5G) th The 5G communication system may be a 5G (first generation) communication system, a worldwide interoperability for microwave access (WiMAX) or a wireless local area network (WLAN) system, or a system that integrates multiple systems, or a future communication system, for example, a 6G communication system. A 5G communication system may also be referred to as a new radio (NR) system.

[0052] A network element in a communication system can transmit a signal to another network element or receive a signal from another network element. The signal may include information, configuration information, data, etc. A network element may also be referred to as an entity, a network entity, a device, a communication device, a communication module, a node, a communication node, etc. In this disclosure, a network element is used as an example for explanation. For example, a communication system may include at least one terminal device and at least one access network device. The transmitting network element of the configuration information may be an access network device, and the receiving network element of the configuration information may be a terminal device. In addition, it may be understood that when a communication system includes multiple terminal devices, the multiple terminal devices may also transmit signals to each other, i.e., both the transmitting network element of the configuration information and the receiving network element of the configuration information may be terminal devices.

[0053] 1A is a communication system. For example, the communication system includes an access network device 110 and two terminal devices, namely, a terminal device 120 and a terminal device 130. At least one of the terminal device 120 and the terminal device 130 may transmit uplink data to the access network device 110, and the access network device 110 may receive the uplink data. The access network device may transmit downlink data to at least one of the terminal device 120 and the terminal device 130.

[0054] The following describes in detail the terminal device and the access network device in FIG. 1A.

[0055] (1) Access Network Device The access network device may be a base station (BS), and may also be referred to as a network device, an access node (AN), or a radio access node (RAN). The access network device may be connected to a core network (e.g., an LTE core network or a 5G core network), and may provide wireless access services to terminal devices. For example, the access network device includes, but is not limited to, at least one of a base station, a next generation node B (gNB) in 5G, an access network device in an open radio access network (O-RAN), an evolved node B (eNB), a radio network controller (RNC), a node B (NB), a base station controller (BSC), a base transceiver station (BTS), a home base station (e.g., home evolved node B, or home node B, HNB), a base band unit (BBU), a transmitting and receiving point (TRP), a transmitting point (TP), and / or a mobile switching center. Alternatively, the access network device may be a centralized unit (CU), a distributed unit (DU), a central unit control plane (CU-CP) node, or a central unit user plane (CU-UP) node.Alternatively, the access network device may be a relay station, an access point, an in-vehicle device, a wearable device, an access network device in a future evolved public land mobile network (PLMN), etc.

[0056] In the present disclosure, a communication apparatus configured to implement the functions of an access network device may be an access network device, or may be a network device having some functions of an access network device, or may be an apparatus that can support an access network device in implementing functions, such as a chip system, a hardware circuit, a software module, or a hardware circuit and a software module. The apparatus may be installed in an access network device. In the present disclosure, an example in which a communication apparatus configured to implement the functions of an access network device is an access network device is used for explanation.

[0057] (2) Terminal Device A terminal device, also referred to as a terminal, user equipment (UE), mobile station (MS), mobile terminal (MT), etc., is a device that provides voice and / or data connectivity to a user. A terminal device can communicate with one or more core networks via an access network device. A terminal device includes a handheld device with wireless connectivity or other processing device connected to a wireless modem, an in-vehicle device, etc. Alternatively, a terminal device may be a portable, pocket-sized, handheld, computer-embedded, or in-vehicle mobile device.Some examples of terminal devices are personal communication service (PCS) telephones, cordless telephones, session initiation protocol (SIP) telephones, wireless local loop (WLL) stations, personal digital assistants (PDAs), wireless network cameras, mobile phones, tablet computers, notebook computers, palmtop computers, mobile internet devices (MIDs), wearable devices such as smart watches, virtual reality (VR) devices, augmented reality (AR) devices, and wireless terminals in industrial control, terminals in vehicular internet systems, wireless terminals in self driving, wireless terminals in smart grids, wireless terminals in transportation security, wireless terminals in smart cities, such as smart fuel suppliers, terminal devices on high-speed trains, or wireless terminals in smart homes, such as smart stereos, smart coffee machines, or smart printers.

[0058] In the present disclosure, the communication device configured to implement the functions of the terminal device may be a terminal device, or may be a terminal device having some functions of the terminal, or may be a device, such as a chip system, that can support the terminal device in implementing the functions. The device may be installed in the terminal device. In the present disclosure, the chip system may include a chip, or may include a chip and another discrete device. In the technical solution provided in the present disclosure, an example in which the communication device configured to implement the functions of the terminal device is a terminal device or a UE is used for explanation.

[0059] (3) Protocol hierarchy between access network devices and terminal devices The communication between the access network device and the terminal device follows a specified protocol layer structure. The protocol layer structure may include a control plane protocol layer structure and a user plane protocol layer structure. For example, the control plane protocol layer structure may include the functions of protocol layers such as a radio resource control (RRC) layer, a packet data convergence protocol (PDCP) layer, a radio link control (RLC) layer, a media access control (MAC) layer, and a physical layer. For example, the user plane protocol layer structure may include the functions of protocol layers such as a PDCP layer, an RLC layer, a MAC layer, and a physical layer. In a possible implementation, a service data adaptation protocol (SDAP) layer may be further included above the PDCP layer.

[0060] Optionally, the protocol layer structure between the access network device and the terminal may further include an artificial intelligence (AI) layer used to transmit data related to AI functions.

[0061] Data transmission between an access network device and a terminal device is used as an example. Data needs to be transmitted through user plane protocol layers, such as SDAP layer, PDCP layer, RLC layer, MAC layer, and physical layer. The SDAP layer, PDCP layer, RLC layer, MAC layer, and physical layer may be collectively referred to as the access layer. Since the data transmission direction includes a transmission or reception direction, each layer is further divided into a transmission part and a reception part. Downlink data transmission is used as an example. After obtaining data from the upper layer, the PDCP layer transmits data to the RLC layer and the MAC layer, and the MAC layer generates a transport block, and then wireless transmission is performed through the physical layer. Data is encapsulated at each layer. For example, data received by a layer from the upper layer of the layer is regarded as a service data unit (SDU) of the layer, encapsulated by the layer into a protocol data unit (PDU), and then forwarded to the next layer.

[0062] For example, the terminal device may further have an application layer and a non-access layer. The application layer may be used to provide services to applications installed on the terminal device. For example, downlink data received by the terminal device may be sequentially transmitted from the physical layer to the application layer, and then provided to the application by the application layer. In another example, the application layer may obtain data generated by the application, sequentially transmit the data to the physical layer, and transmit the data to another communication device. The non-access layer may be configured to forward user data. For example, the non-access layer forwards uplink data received from the application layer to the SDAP layer, or forwards downlink data received from the SDAP layer to the application layer.

[0063] (4) Access network device structure An access network device may include a central unit (CU) and a distributed unit (DU). Multiple DUs may be centrally controlled by one CU. For example, an interface between a CU and a DU may be referred to as an F1 interface. A control plane (CP) interface may be F1-C, and a user plane (UP) interface may be F1-U. The CU and the DU may be divided based on the protocol layer of a wireless network. For example, the functions of the PDCP layer and the protocol layers above the PDCP layer are configured on the CU, and the functions of the protocol layers below the PDCP layer (e.g., the RLC layer and the MAC layer) are configured on the DU. In another example, the functions of the protocol layers above the PDCP layer are configured on the CU, and the functions of the PDCP layer and the protocol layers below the PDCP layer are configured on the DU.

[0064] It may be understood that the processing function division of the CU and DU based on the protocol layer is only an example, and there may be other divisions. For example, the CU or DU may be divided into functions with more protocol layers. In another example, the CU or DU may be further divided into several processing functions with protocol layers. In a design, some functions of the RLC layer and functions of the protocol layers above the RLC layer are configured on the CU, and the remaining functions of the RLC layer and functions of the protocol layers below the RLC layer are configured on the DU. In another design, the division of the functions of the CU or DU may be alternatively performed based on the service type or other system requirements. For example, the division may be performed based on latency. Functions whose processing time needs to meet the latency requirement are configured on the DU, and functions whose processing time does not need to meet the latency requirement are configured on the CU. In another design, the CU may alternatively have one or more functions of the core network. For example, the CU may be located on the network side to facilitate centralized management. In another design, the radio unit (RU) of the DU is remotely located. The RU has a radio frequency function.

[0065] Optionally, the DU and the RU may be distinguished in the physical layer (PHY). For example, the DU may implement the higher layer functions of the PHY layer, and the RU may implement the lower layer functions of the PHY layer. When used for transmission, the functions of the PHY layer may include cyclic redundancy check (CRC) code addition, channel coding, rate matching, scrambling, modulation, layer mapping, precoding, resource mapping, physical antenna mapping, and / or radio frequency transmission functions. When used for reception, the functions of the PHY layer may include CRC, channel decoding, rate dematching, descrambling, demodulation, layer demapping, channel detection, resource demapping, physical antenna demapping, and / or radio frequency reception functions. The functions of the higher layer of the PHY layer may include some functions of the PHY layer. For example, some functions are closer to the MAC layer. The functions of the lower layer of the PHY layer may include some other functions of the PHY layer. For example, some functions are closer to the radio frequency functions. For example, the upper layer functions of the PHY layer may include CRC code addition, channel coding, rate matching, scrambling, modulation, and layer mapping, and the lower layer functions of the PHY layer may include precoding, resource mapping, physical antenna mapping, and radio frequency transmission functions. Alternatively, the upper layer functions of the PHY layer may include CRC code addition, channel coding, rate matching, scrambling, modulation, layer mapping, and precoding. The lower layer functions of the PHY layer may include resource mapping, physical antenna mapping, and radio frequency transmission functions.

[0066] For example, the functions of the CU may be implemented by one entity or by different entities. For example, the functions of the CU may be further divided, that is, the control plane and the user plane are separated and implemented by different entities, which are a control plane CU entity (i.e., a CU-CP entity) and a user plane CU entity (i.e., a CU-UP entity). The CU-CP entity and the CU-UP entity may be combined with the DU to jointly complete the functions of the access network device.

[0067] In the above network architecture, the signaling generated by the CU may be transmitted to the terminal device via the DU, or the signaling generated by the terminal device may be transmitted to the CU via the DU. For example, the signaling at the RRC layer or the PDCP layer is finally processed as the signaling at the physical layer and transmitted to the terminal device, or is converted from the signaling received from the physical layer. Based on such an architecture, the signaling at the RRC layer or the PDCP layer may be considered to be transmitted through the DU, or transmitted through the DU and the RU.

[0068] Optionally, any one of DU, CU, CU-CP, CU-UP, and RU may be a software module, a hardware structure, or a combination of a software module and a hardware structure, but is not limited thereto. Different entities may exist in different forms, but is not limited thereto. For example, DU, CU, CU-CP, and CU-UP are software modules, and RU is a hardware structure. These modules and the methods performed by these modules also fall within the protection scope of the present disclosure.

[0069] It should be understood that the number and types of devices in the communication system shown in FIG. 1A are merely examples, and the present disclosure is not limited thereto. In practical application, the communication system may further include more terminal devices and more access network devices, and may further include another network element, for example, network management, core network devices, and / or network elements configured to implement artificial intelligence functions. Network management can be classified into three types of network management tasks, namely, operation, administration, and maintenance, based on the operator's actual requirements for network operation. Network management can also be referred to as operation administration and maintenance (OAM) network elements, and is simply referred to as OAM. Operation mainly completes the daily tasks performed on networks and services, such as analysis, prediction, planning, and configuration. Maintenance is mainly the daily operation activities, such as testing and fault management, performed on networks and services. Network management can detect network running status, optimize network connection and performance, improve network stability, and reduce network maintenance costs.

[0070] The methods provided in the present disclosure may be applied to communication between an access network device and a terminal device, or may be applied to communication between other communication devices, such as communication between a macro base station and a micro base station in a wireless backhaul link, such as communication between two terminal devices in a sidelink (SL), but are not limited thereto.

[0071] The method provided in this disclosure relates to artificial intelligence (AI). AI may be implemented by using various possible techniques, for example, by using machine learning (ML) techniques. In this disclosure, the communication system may also include a network element that implements an artificial intelligence function. For example, an AI function (e.g., an AI module or an AI entity) may be configured in an existing network element in the communication system to implement AI-related operations. For example, in a 5G new radio (NR) system, the existing network element may be an access network device (e.g., a gNB), a terminal device, a core network device, a network management, etc. Alternatively, an independent network element may be introduced into the communication system to perform AI-related operations. The independent network element may be referred to as an AI network element, an AI node, etc. This name is not limited in this disclosure. In this case, the network element that performs the AI-related operation is a network element (e.g., an AI module or an AI entity) that has a built-in AI function. The AI-related operation may also be referred to as an AI function. For more information on the AI ​​function, please refer to the following description. The AI ​​network element may be directly connected to an access network device in a communication system, or may be indirectly connected to the access network device by using a third party network element. The third party network element may be a core network element, such as an authentication management function (AMF) network element or a user plane function (UPF) network element. For example, as shown in FIG. 1B, an AI network element 140 is introduced into the communication system shown in FIG. 1A.

[0072] For ease of understanding, some terms of AI in the present disclosure will be described below with reference to A1 to A4. It can be understood that this description is not intended to limit the present disclosure.

[0073] A1: AI model An AI model is a specific implementation of an AI function. An AI model may be a function model that represents a mapping relationship between the input and output of the model and maps an input in a dimension to an output in a dimension. A machine learning model may be a neural network or another AI model such as a decision tree or a support vector machine. In this disclosure, an AI model may be simply referred to as a model. In this disclosure, an AI function may include at least one of data collection (collecting training data and / or inference data), data preprocessing, model training (also referred to as model learning), model information release (model information configuration), model inference, or inference result release. Inference may also be referred to as prediction. In this disclosure, an AI model may be simply referred to as a model.

[0074] A2: Machine learning Machine learning is an important technological method for implementing artificial intelligence. For example, machine learning can learn models or rules from raw data. Machine learning is classified into supervised learning, unsupervised learning, and reinforcement learning.

[0075] For supervised learning, based on collected samples (or referred to as sample values) and sample labels, a mapping relationship between samples and sample labels is learned by using a machine learning algorithm, and the learned mapping relationship is represented by using a machine learning model. The process of training a machine learning model is the process of learning a mapping relationship. For example, in signal detection, a sample is a received signal containing noise, and a sample label is an actual constellation point corresponding to the received signal. Machine learning expects to learn a mapping relationship between samples and sample labels through training. During training, model parameters are optimized by calculating the error between the model output (i.e., predicted value) and the sample label. Once the mapping relationship is learned, the sample label of each new sample can be predicted by using the learned mapping relationship. The mapping relationship learned through supervised learning may include linear mapping and nonlinear mapping. The task of classification may be classified into machine learning task and regression task based on the type of sample label.

[0076] For unsupervised learning, the internal patterns of the samples are autonomously explored by using an algorithm based only on the collected samples. In unsupervised learning, certain types of algorithms (such as autoencoders and generative adversarial networks) can use the samples as supervised signals. The model learns the mapping relationship from sample to sample. During training, the model parameters are optimized by calculating the error between the model's predicted value and the sample, thereby implementing self-supervised learning. Self-supervised learning can be used in the application scenarios of signal compression and decompression restoration.

[0077] Reinforcement learning is a type of algorithm that learns a strategy to solve a problem by interacting with the environment. Unlike supervised learning and unsupervised learning, reinforcement learning does not have a specific sample label. The algorithm needs to interact with the environment to obtain a reward signal fed back by the environment and adjust the decision action to obtain a larger reward signal value. For example, in downlink power control, the reinforcement learning model adjusts the downlink transmission power of each terminal based on the total system throughput fed back by the wireless network in the hope of obtaining a higher system throughput. The goal of reinforcement learning is also to learn the mapping relationship between the environment status and the optimal decision action. Reinforcement learning training is implemented through repeated interactions with the environment.

[0078] A3: Neural Networks A neural network is a specific implementation of AI or machine learning technology. According to the general approximation theorem, a neural network can theoretically approximate any continuous function, and as a result, a neural network has the ability to learn any mapping. Traditional communication systems require extensive expertise to design communication modules. However, a neural network-based deep learning communication system can automatically discover implicit pattern structures from large data sets, establish mapping relationships between data, and obtain better performance than traditional modeling methods.

[0079] The concept of neural networks comes from the neuron structure of the brain. For example, each neuron performs a weighted sum operation on its inputs and outputs the result via an activation function. Figure 2A shows a diagram of a neuron structure. If the input of a neuron is x=[x 0 ,x 1 ,…,x n ] and the weights corresponding to the inputs are w=[w,w 1 ,…,w n ], and w i x iis the weight of x i For example, the offset is b to perform a weighted sum on the input values ​​based on the weights. Activation functions can take several forms. The activation function for a neuron is y=f(z)=max(0,z), and the output of the neuron is

number

number

[0080] A neural network generally includes multiple layers, and each layer may include one or more neurons. By increasing the depth and / or width of a neural network, the representation or function fitting ability of the neural network can be improved, providing a more powerful information extraction and abstract modeling ability for complex systems. The depth of a neural network may be the number of layers included in the neural network, and the number of neurons included in each layer may be referred to as the layer width. In one implementation, the neural network includes an input layer and an output layer. The input layer of the neural network performs neuronal processing on the received input information and transfers the processing results to the output layer, which obtains the output layer results of the neural network. In another implementation, the neural network includes an input layer, a hidden layer, and an output layer. The input layer of the neural network performs neuronal processing on the received input information and transfers the processing results to an intermediate hidden layer. The hidden layer performs calculations on the received processing results and obtains the calculation results. The hidden layer transfers the calculation results to the output layer or adjacent hidden layers, and finally, the output layer obtains the output results of the neural network. A neural network may include one hidden layer, or may include multiple hidden layers connected in series, but is not limited thereto.

[0081] As described above, each neuron performs a weighted sum operation on the input values ​​of the neuron, and the weighted sum result is generated according to a nonlinear function and output. The weights of the weighted sum operation between the neuron and the nonlinear function of the neural network are called the parameters of the neural network. For example, the nonlinear function of a neuron is max{0,x}, and the operation performed by the neuron is

number

[0082] The neural network in this disclosure is, for example, a deep neural network (DNN). A DNN usually has multiple hidden layers, and the weights corresponding to each neuron in the DNN are the model parameters of the DNN. The DNN can use a supervised learning strategy or an unsupervised learning strategy to optimize the model parameters. Based on the network construction method, the DNN may include a feedforward neural network (FNN), a convolutional neural networks (CNN), and a recurrent neural network (RNN). The FNN is used as an example. FIG. 2B shows the neural network structure. The characteristic of the FNN is that two neurons in adjacent layers are fully connected.

[0083] CNN may be applied to process data having a grid-like structure. The grid-like structure data may include time series data (time-axis discrete sampling), image data (two-dimensional discrete sampling), etc. The convolutional layer of CNN does not perform a convolution operation using all input information at once. Instead, one or more windows of fixed size are set, and a part of the input information is captured by using each window to perform a convolution operation. Such a design can greatly reduce the amount of calculation of model parameters. Specifically, performing a convolution operation on any one of one or more windows of fixed size may be understood as performing a multiplication operation on the coefficients (e.g., weighting coefficients) of the window and some input information captured by the window, and then performing an addition operation. After the convolution operation, output information corresponding to the window may be obtained. The coefficients of different windows may be independently configured. For example, different coefficients may be configured for different windows so that the CNN can better extract features of the input data. The coefficients of the windows may include convolution kernels. Optionally, the types of some input information captured in different windows may be different. For example, a person and an object in the same secondary image can be understood as different types of information. In two windows with fixed size, one window can capture a person in the image, and the other window can capture an object in the image. RNN is a DNN network that uses feedback time series information. The input of RNN includes a part of the new input value at the current moment and the output value of RNN at the previous moment, and the output value at the previous moment may be determined by the activation function and the input at the previous moment. RNN is suitable for acquiring sequence features with time correlation, and is suitable for application scenarios such as speech recognition and channel encoding and channel decoding.

[0084] Also, a loss function may be defined in the neural network training process. The loss function describes the difference or difference value between the output value of the neural network and an ideal target value, and the present disclosure does not limit the specific form of the loss function. The neural network training process is a process in which the parameters of the neural network are adjusted so that the value of the loss function is less than a threshold value or the value of the loss function meets the target requirement. The parameters of the neural network are adjusted. For example, at least one of the following parameters is adjusted: the number of layers of the neural network, the width of the neural network, the weights of the neurons, or the parameters in the activation function of the neurons.

[0085] A4: Federated learning (FL) Federated learning is a machine learning framework. In this framework, nodes do not need to exchange data, but intermediate results obtained during training are transferred, for example, model parameters or gradients can represent information about the model. That is, federated learning can be used to model machine learning, i.e., train AI models, to meet user privacy protection and data security requirements. As a distributed machine learning paradigm, federated learning can effectively solve the problem of data silos. Nodes participating in federated learning can perform joint modeling without sharing data. This breaks data silos in technology and implements AI collaboration.

[0086] Generally, according to the different distribution of data sources of all participating parties, federated learning can be classified into three types, namely horizontal federated learning, vertical federated learning, and federated transfer learning. Horizontal federated learning means that when the user features of multiple datasets (alternatively, may be understood as sample sets) overlap a lot, but the users overlap little, the dataset is divided horizontally (i.e., in the user dimension), and a model is trained based on some data with users that have the same user features but are incompletely the same. Vertical federated learning means that when the users of multiple datasets overlap a lot, but the user features overlap little, the dataset is divided vertically (i.e., in the feature dimension), and a model is trained based on some data with the same user but are incompletely the same user features. Federated transfer learning means that when the users and user features of multiple datasets overlap little, the data is not divided, but transfer learning is used to overcome the case of lack of data or sample labels.

[0087] The present disclosure specifically relates to the application of federated learning in wireless networks, where data required for training a model is distributed across multiple nodes, and the data may be related to the privacy information of the nodes, making direct data transmission unsuitable. In this case, federated learning techniques may be used to enable multiple nodes to participate in federated learning.

[0088] In a possible design, multiple nodes train local models and send parameter information of the local models trained by the multiple nodes to a central node. The central node can calculate (or train) a federated learning model based on the local model parameters of the multiple nodes and send the federated learning model to the multiple nodes. The multiple nodes keep updating their local models and sending parameter information of the local models to the central node. In this manner, iterations are performed until convergence. FIG. 3A, for example, shows a communication architecture including a central node and three subnodes participating in federated learning, namely subnode 1, subnode 2, and subnode 3. Subnode 1, subnode 2, and subnode 3 can separately collect training data, train local models of subnode 1, subnode 2, and subnode 3, and send parameter information of the local models to the central node. The central node can train the federated learning model.

[0089] In another possible design, multiple nodes train local models and send parameter information of the local models trained by the multiple nodes to a central node. The central node may forward the local model parameters of the multiple nodes to an AI model management node. The AI ​​model management node calculates (or trains) a federated learning model based on the local model parameters of the multiple nodes and sends the federated learning model to the multiple nodes. The multiple nodes keep updating their local models, and the parameter information of the local models is sent to the AI ​​model management node via the central node. In this way, iterations are performed until convergence. In this case, the central node acts as a relay between the AI ​​model management node and the nodes participating in the federated learning. FIG. 3B shows, for example, a communication architecture including an AI model management node, a central node, and three subnodes participating in the federated learning, namely, subnode 1, subnode 2, and subnode 3. Subnode 1, subnode 2, and subnode 3 can separately collect training data, train local models of subnode 1, subnode 2, and subnode 3, and send parameter information of the local models to the central node. The central node can transfer parameter information of the local models from the nodes participating in the federated learning to the AI ​​model management node, and the AI ​​model management node performs training / update operations on the federated learning models and sends the parameter information of the federated learning models to the central node. Then, the central node transfers the parameter information of the federated learning models to the corresponding subnodes 1, 2, and 3.

[0090] It can be understood that the central node is a node configured to train a federated learning model, and the central node may be replaced by another name. Unlike the central node, the aforementioned multiple nodes that can participate in federated learning and train local models may also be called multiple subnodes, multiple distributed nodes, or other names. The present disclosure does not limit the names. The present disclosure is described below by using an example in which the central node trains a federated learning model and the subnode trains a local model.

[0091] Specifically, multiple nodes participating in federated learning can access the central node and communicate with the central node. For example, the central node in FIG. 3A / FIG. 3B may be an access network device, and the nodes participating in federated learning, i.e., subnode 1, subnode 2, and subnode 3, may be termination devices, or the central node may be a core network device, and subnode 1, subnode 2, and subnode 3 may be access network devices. The central node, subnode 1, subnode 2, and subnode 3 may all be AI network elements / modules with AI functions. The AI ​​model management node in FIG. 3B may be an independent AI network element or module with AI functions.

[0092] For example, in positioning, a terminal device may collect data including a geographical location. However, the data collected by a single terminal device is simple and cannot traverse all possible distributions of real data. In addition, a single terminal device receives a small amount of data. When it is difficult to train a model that meets performance requirements based only on the data collected by a single terminal device. However, when multiple terminal devices transmit data to an access network device or a core network device, such as a location management function (LMF) network element, the access network device or the LMF network element trains a model. Sending data to an access network device or an LMF network element causes privacy leakage of the terminal device because the geographical location of the terminal device is related to privacy and the overhead of sending data by the terminal device to the access network device or the LMF network element is high. In this case, a federated learning technique may be used. For example, multiple terminal devices participate in federated learning. Each terminal device may train a local model based on its respective data, and then send parameter information of the local model to the access network device or the LMF network element. The access network device or the LMF network element trains a federated learning model and sends the federated learning model to multiple terminal devices. The multiple terminal devices continue to update the local models and send parameter information of the local models to the access network device or the LMF network element, and thus iterations are performed until convergence.

[0093] In the related technology, the central node trains only one federated learning model each time, that is, the federated learning model sent to multiple subnodes each time is the same. In the following, the same federated learning model is recorded as a global model, and the related technology is described in detail. In wireless communication scenarios, different terminal devices need to perform distributed training on the same type of data, and horizontal federated learning is suitable. The main steps of horizontal federated learning are as follows:

[0094] S1: The central node distributes the global model to multiple subnodes. Each of the multiple subnodes can train a local model based on local data (i.e., samples acquired by the node) and the global model. In the above positioning, the central node can be a base station or an LMF, and the subnodes in the multiple subnodes can be terminal devices.

[0095] S2: Each of the multiple subnodes trains a local model by using local data and uploads parameters of the local model to the central node, and the central node aggregates the parameters of the local models of each subnode to update global model parameters, i.e., performs update training on the global model in S1.

[0096] Repeat the previous two steps until the model converges.

[0097] As shown in Fig. 4A, an example is used in which multiple subnodes include subnode 1, subnode 2, and subnode 3. First, the central node sends an initial model to each of subnode 1, subnode 2, and subnode 3. Subnode 1 performs update training on the initial model based on the local data of subnode 1 (i.e., data 1 in Fig. 4A) to generate a local model M 1 Subnode 2 performs update training on the initial model based on the local data of subnode 2 (i.e., data 2 in FIG. 4A) to obtain a local model M 2Subnode 3 performs update training on the initial model based on the local data of subnode 3 (i.e., data 3 in FIG. 4A) to obtain a local model M 3 Subnodes 1, 2, and 3 send the parameters of their local models to the central node, which performs weighting on the local models of the subnodes to obtain the global model M 0 = α 1 M 1 +α 2 M 2 +α 3 M 3 where α i is the ratio of the local data amount of the ith subnode of the multiple subnodes to the total local data amount of the multiple subnodes, and M i indicates the local model of the i-th subnode of the multiple subnodes, or can be understood as the parameter of the local model of the i-th subnode. The value of i is 1, 2 or 3. Then, the central node M 0 to each subnode, which continues to perform local training based on the received global model, and then uploads the parameters of the trained local model to the central node. In this way, iterations are performed until the model converges.

[0098] In practical applications, there are distribution differences between the local data of the aforementioned subnodes. For example, in indoor positioning, different terminal devices are located in different areas for a long time or frequently. In this case, by using the aforementioned related technology, multiple subnodes use the same global model to perform update training on local models. For subnodes whose data distribution is significantly different from that of other subnodes, the performance of the local model is poor, and the performance of the federated learning model is degraded.

[0099] Based on this, the present disclosure provides a communication method for improving the performance of a federated learning model. In the present disclosure, when a central node trains a federated learning model based on parameter information of local models of multiple subnodes participating in federated learning, the central node individualizes the federated learning models corresponding to different subnodes, so that multiple subnodes are prevented from using the same global model to perform update training on local models, and the performance of the federated learning model can be improved.

[0100] As shown in Fig. 4B, an example is used in which multiple subnodes include subnode 1, subnode 2, and subnode 3. First, the central node sends an initial model to each of subnode 1, subnode 2, and subnode 3. Subnode 1 performs update training on the initial model based on the local data of subnode 1 (i.e., data 1 in Fig. 4B) to generate a local model M 1 Subnode 2 performs update training on the initial model based on the local data of subnode 2 (i.e., data 2 in FIG. 4B) to obtain a local model M 2 Subnode 3 performs update training on the initial model based on the local data of subnode 3 (i.e., data 3 in FIG. 4B) to obtain a local model M 3Subnodes 1, 2, and 3 send parameters of the local models of subnodes 1, 2, and 3 to the central node. In the present disclosure, the central node may train the same global model for at least two of the multiple subnodes, and train a federated learning model corresponding to the remaining subnode, for example, referred to as an individualized model. Alternatively, the central node may train an individualized model for each of the multiple subnodes. As an example, FIG. 4B shows that the central node trains individualized models for subnodes 1, 2, and 3 separately. Subnodes 1, 2, and 3 send parameters of the local models of subnodes 1, 2, and 3 to the central node. The central node uses different weighting for different subnodes and performs weighting on the local models of the subnodes to obtain the individualized model of the subnode, for example, the individualized model P of the subnode with the kth order. k =β 1k M 1 +β 2k M 2 +β 3k M 3 where the value of k is 1, 2, or 3, and the different P k is different β 1k , β 2k , and β 3k The central node then presents the corresponding individualized model to each subnode, e.g., P 1 Send to subnode 1, and P 2 Send to subnode 2, and P 3 to subnode 3. Each subnode continues to perform local training based on the received federated learning model, and then uploads the parameters of the trained local model to the central node. In this way, iterations are performed until the model converges.

[0101] The communication method provided in the present disclosure will be further described in detail below with reference to the accompanying drawings.

[0102] Figure 5 shows a communication method, which mainly includes the following steps:

[0103] S501: The central node transmits information indicating a second model to a plurality of subnodes participating in the federated learning.

[0104] For example, Figure 5 shows three subnodes participating in federated learning: subnode 1, subnode 2, and subnode 3. For examples of central nodes and subnodes, see the above description in Figure 3A. The details will not be described again herein in this disclosure.

[0105] The information indicating the second model is used by each of the multiple subnodes to determine the first local model of each subnode. The second model may be understood as an initial model. For example, the information indicating the second model may be the second model, or may be parameter information of the second model. The parameter information of the second model may represent the second model. The parameter information of the second model may indicate one or more of the following parameters: a structure type of the second model, a structure parameter of the second model, and a model parameter of the second model. In the following, an example in which the central node transmits the parameter information of the second model is used for explanation.

[0106] Some examples of parameter information of the second model are as follows: The structural type of the second model, e.g., a CNN model, an RNN model, a Transformer model, a random forest, or a support vector machine; The structural parameters of the second model, specifically, models of different structural types may correspond to different structural parameters. For example, when the second model is a CNN model, the structural parameters of the second model may include the number of convolutional layers in the CNN, the number of channels and the size of the convolution kernel in each convolutional layer, the number of fully connected layers, and / or the number of neurons. When the second model is an RNN model, the structural parameters of the second model may include the number of neural network layers of the RNN and the calculation scheme for each state in each layer of the structure, and the state may be a forget gate, an input gate, an output gate, an update gate, an implicit state, etc.; When the model is a Transformer model, the structural parameters of the second model may include the number of encoders and decoders in the Transformer model and the computation schemes for the internal elements in each encoder and decoder, where the internal elements may be a query vector, a key vector, a value vector, etc.; when the second model is a random forest, the structural parameters of the second model may include the number and depth of trees in the random forest; or when the second model is a support vector machine, the structural parameters of the second model may include the type of kernel function in the support vector machine, etc.; Model parameters of the second model, which can be understood as parameters related to constructing the second model, such as weights and offsets corresponding to each neuron in a neural network, decision criteria and thresholds corresponding to each node in a random forest, and corresponding weights and offsets in a support vector machine.

[0107] Optionally, some optional model structures may be specified in the protocol. The model structure includes a structure type and a structure parameter. The parameter information of the second model includes information indicating the structure type and the structure parameter belonging to the model structure specified in the aforementioned protocol of the second model, and further includes information indicating the model parameter. Alternatively, the model structure may be specified in the protocol. In this case, the parameter information of the second model indicates only the model parameter of the second model, and the nodes participating in the federated learning can determine the structure type and the structure parameter of the second model according to the protocol. Alternatively, the combination of the model structure and the model parameter may be specified in the protocol, and the central node may include, in the parameter information of the second model, the indication information of the combination of the model structure and the model parameter specified in the protocol. In addition, it may be understood that the structure type and the structure parameter of the second model may be directly specified in the protocol. In this case, the central node may not need to transmit the parameter information of the second model. That is, in this case, S501 does not need to be executed. S501 is an optional step.

[0108] Specifically, the central node may transmit parameter information of the second model to the multiple subnodes based on the preset resource. For example, the central node or the aforementioned multiple subnodes may configure the parameter information of the model (e.g., the second model) to be transmitted by the central node, or another federated learning management node (e.g., a third-party network element or an independent AI network element) configures resources for the central node and the multiple subnodes to transmit the parameter information of the model by the central node. For example, the central node is an access network device, and the multiple subnodes are terminal devices. The resource used by the central node to transmit the second model may be a downlink resource. The downlink resource may specifically be a control channel resource, e.g., a physical downlink control channel (PDCCH) resource, or the downlink resource may be a data channel resource, e.g., a physical downlink share channel (PDSCH) resource. Specifically, resources may be understood as time domain resources, e.g., slots or symbols, and frequency domain resources, e.g., time-frequency resources including frequency domain resource block numbers, subband numbers, frequency hopping parameters, modulation and coding schemes, and other parameters.

[0109] Specifically, the central node may transmit the parameter information of the second model in a broadcast or multicast manner. For example, the central node may transmit the parameter information of the second model to the cell in which the aforementioned subnodes are located in a broadcast manner. It can be understood that the subnodes may be located in the same cell, or the subnodes may be located in different cells. Another node that does not participate in the federated learning in the cell in which the subnodes are located can also receive the information broadcast by the central node. In another example, the central node can regard the subnodes participating in the federated learning as a group, and the subnodes in the same group correspond to the same group number. The central node transmits the parameter information of the second model to the subnodes in the group based on the same resource in a multicast manner. In another example, the central node can group the subnodes, for example, based on the computing capabilities of all the subnodes, and the computing capabilities of the subnodes in the same group are similar. Different groups are correspondingly configured with different group numbers. The central node may configure different resources for different groups to transmit, in a multicast manner, parameter information of the second model and the group number of the group to subnodes in the group based on the resources corresponding to the group.

[0110] Further, optionally, the central node may further indicate to the multiple subnodes training parameters for training the local model by each subnode. The training parameters may include one or more of the number of iterations required by the subnode to train the local model, a learning rate, a loss function, and a batch size (Batch_Size). The subnodes can train the local model based on the training parameters.

[0111] S502: Repeat the following steps S21 to S24 I times, where I indicates the number of iterations for training the associative learning model, and I is a positive integer.

[0112] S21: Each of a number of subnodes trains a local model.

[0113] Specifically, each subnode can train a local model based on the parameter information of the model sent by the central node. The training data used by the subnode to train the local model includes the local data collected by the subnode. Optionally, the structure type and structure parameters of the local models trained by all the subnodes are the same. For example, if each subnode most recently receives the information indicating the second model described in S501, each subnode may perform a first-time (or referred to as first-time) local model training based on the collected local data and the information indicating the second model. In addition, optionally, the training data used by the subnode to train the local model may further include label data.

[0114] Each subnode may determine training data for a local model for a different application scenario.

[0115] For example, in a positioning scenario, the local data collected by the subnode may be a channel response for training a local model that uses the channel response as an input and the location coordinate as an output. The subnode may further obtain label data, i.e., location coordinates, corresponding to the channel response. Optionally, the subnode may obtain the location coordinates by using a positioning method based on arrival time difference. For example, the subnode is a terminal device, and the central node is an LMF. The terminal device receives Positioning Reference Signals (PRS) from multiple access network devices. The terminal device measures the arrival time difference of the PRS transmitted by different access network devices, and feeds back the aforementioned arrival time difference to the LMF. The LMF can infer the difference in distance from different access network devices to the terminal device based on the arrival time difference, and further determine the location of the terminal device. Then, the LMF feeds back the location of the terminal device to the terminal device by using signaling, and the terminal device can use the location as label data corresponding to the local data. Optionally, the access network device may repeatedly transmit the PRS, and the LMF repeatedly calculates the location of the terminal device in a short time. If the multiple positions of the terminal device obtained through short-time calculations are stable (i.e., the difference between the multiple positions is small, e.g., less than a specified threshold), it may indicate high test positioning accuracy. LMF feeds back the stable position of the terminal device to the UE by using signaling, and the UE uses the position as a label for local data. This can improve the performance of the local model.

[0116] For example, in a beam management scenario, the subnode may be an access network device, and the central node may be a device that may be accessed by the access network device, for example, an independent AI element, or a convergence access network device that allows another access network device to access. The local data that may be collected by the subnode may be channel measurements (e.g., energy or quality of the received signal), and a machine learning model is trained using a small amount of channel measurements as input and an index of the optimal beam as output, where the index of the optimal beam is label data corresponding to the local data. Optionally, the subnode (i.e., the access network device) may transmit a signal to a terminal device that accesses the subnode by using all possible beams (or referred to as beam directions) one by one, and determine the index of the optimal beam based on the feedback of the terminal device. For example, the terminal device may determine the beam that corresponds to the best received signal quality, and the beam is indicated as the (performance) optimal beam. The index of the optimal beam is fed back to the subnode. In addition, the channel measurements collected by the subnode may also be obtained from the terminal device. For example, when feeding back the index of the optimal beam, the terminal device may further feed back a small amount of channel measurements.

[0117] Each subnode can determine the loss function used to train the local model based on the collected local data, the application scenario, or the structure type of the local model. For example, the cross-entropy function can be used as the loss function in classification problems, such as line of sight (LOS) / non-line of sight (NLOS) recognition, and the mean squared error (MSE) or mean absolute error (MAE) can be used as the loss function in regression problems, such as channel estimation.

[0118] S22: Each of a plurality of subnodes transmits information indicative of the local model to the central node.

[0119] The information indicating the local model may be the local model or the parameter information of the local model. In the following, for the purpose of explanation, an example in which the subnode transmits the parameter information of the local model is used. For example, FIG. 5 shows S22. The subnode 1, the subnode 2, and the subnode 3 transmit the parameter information of the local model to the central node separately.

[0120] Specifically, the parameter information of the local model can be understood with reference to the following.

[0121] In an optional implementation, after the local model has been trained by the subnode for any one of I times, the parameter information of the local model sent by the subnode may include model parameters of the local model.

[0122] For example, each time after training a local model, the subnode may send a combination of all model parameters of the local model obtained through the current training to the central node as a vector. The local model is assumed to have a total of G model parameters. Each time after training a local model, the subnode may send a combination of G model parameters of the local model obtained through the current training to the central node as a vector. For example, the parameter information of the local model reported by the k-th subnode of the multiple subnodes is M k =[m k,1 ,m k,2 ,...,m k,G ], where m k,jindicates the j-th model parameter of the local model of the k-th subnode, j is a positive integer, and 1≦j≦G. k is a positive integer from 1 to K, and K is the number of multiple subnodes. The aforementioned k-th subnode may also be described as a subnode having a k-th order, and indicates any one of the multiple subnodes. For the definitions of j, G, k, and K in the following of this disclosure, please refer to this example for understanding. Details will not be described in the following content.

[0123] For example, after each time training a local model, the subnode divides all the model parameters of the local model obtained through the current training into multiple combinations, uses each combination as a vector, and sends the vectors corresponding to the multiple combinations to the central node. Optionally, the multiple combinations obtained through division may be determined based on the structure of the local model. For example, as shown in FIG. 6, the local model may be divided into a global layer and an individualized layer. The multiple combinations obtained by division include the model parameters of the global layer and the model parameters of the individualized layer. This is not limited in the present disclosure.

[0124] In another optional implementation, after the local model is trained by the subnode for the first of I times, the parameter information of the local model transmitted by the subnode may include the model parameters of the local model. After the local model is trained by the subnode for any one of the second through I times, the parameter information of the local model transmitted by the subnode may include a difference between the parameter information of the currently trained local model and the parameter information of the previously trained local model. The difference may be described as a gradient of the model parameters of the currently trained local model, or may simply be referred to as the gradient of the currently trained local model.

[0125] For example, it is assumed that the local model has G parameters. After the subnode trains the local model for the first time, the subnode may send a combination of the G parameters of the local model obtained through the current training to the central node as a vector. For details, please refer to the optional implementation mentioned above. The details will not be described again herein in this disclosure. After the subnode trains the local model for any one of the second to I times, the subnode may send a set of differences between the G parameters of the local model obtained through the current training and the G parameters of the local model obtained through the previous training, i.e., the gradient of each parameter, to the central node as a vector. For example, in each of the second to I times, the parameter information of the local model reported by the k-th subnode in the multiple subnodes is ΔM k = [Δm k,1 ,Δm k,2 ,...,Δm k,G ], where Δm k,j denotes the difference between the jth parameter of the local model trained by the kth subnode this time and the jth parameter of the local model trained last time, or Δm k,j It may be understood that G denotes the gradient of the j-th parameter of the local model trained this time by the k-th subnode, where j is a positive integer and 1≦j≦G.

[0126] For example, each time after training a local model, the subnode divides the gradient of the local model obtained through the current training into multiple combinations, uses each combination as a vector, and sends the vector corresponding to the multiple combinations to the central node. Optionally, the multiple combinations obtained through division may be determined based on the structure of the local model. For example, as shown in FIG. 6, the local model is divided into a global layer and an individualized layer. The multiple combinations obtained by division include the gradient of the global layer and the gradient of the individualized layer. This is not limited in the present disclosure.

[0127] Optionally, the sub-node may determine, based on the instruction of the central node, that the parameter information of the local model includes the aforementioned model parameters, gradients of model parameters, other information, etc. Optionally, the central node may further indicate or pre-agree that the parameter types in different combinations may be different when the sub-node divides the parameter information of the local model into multiple combinations and reports the multiple combinations to the central node. The parameter type in this specification may be a model parameter or a gradient of a model parameter. Alternatively, the parameter types of some combinations of the multiple combinations may be set to model parameters, and the parameter types of other combinations may be set to gradients. Specifically, the configuration may be performed with reference to an actual application, which is not limited in the present disclosure.

[0128] In addition, the local model parameter information transmitted by the subnode may further include auxiliary information, where the auxiliary information indicates the number of samples of local data collected by the subnode and / or the parameter type of the local model parameter information.

[0129] Specifically, the subnode may transmit parameter information of the local model to the central node based on the preconfigured resource. For example, the subnode may transmit the parameter information of the local model by using the preconfigured resource. The central node may configure resources for multiple subnodes to transmit the parameter information of the local model by the subnode, or another federated learning management node (e.g., a third-party network element or an independent AI network element) may configure resources for the central node and multiple subnodes to transmit the parameter information of the local model by the subnode. For example, the central node is an access network device, and the multiple subnodes are terminal devices. The resource used by the subnode to transmit the parameter information of the local model may be an uplink resource. The uplink resource may specifically be a control channel resource, for example, a physical uplink control channel (PUCCH) resource, or the uplink resource may be a data channel resource, for example, a physical uplink share channel (PUSCH) resource. Specifically, resources may be understood as time domain resources, e.g., slots or symbols, and frequency domain resources, e.g., time-frequency resources including frequency domain resource block numbers, subband numbers, frequency hopping parameters, modulation and coding schemes, and other parameters.

[0130] Optionally, when the subnode divides the parameter information of the local model into multiple combinations and reports the multiple combinations to the central node, the resources occupied for reporting the parameter information of the local model in different combinations may be different, or the resources occupied for reporting the parameter information of some combinations and the parameter information of other combinations in the multiple combinations are different. In particular, the configuration may be performed with reference to an actual application, which is not limited in the present disclosure.

[0131] S23: The central node trains a federated learning model based on the information representing the local model of each subnode.

[0132] Specifically, the i-th time of the above I times is used as an example. The central node can obtain the i-th first information, and the i-th first information indicates the i-th local model of each of the multiple subnodes. Specifically, the i-th first information may be sent by each of the multiple subnodes and include information indicating the i-th local model.

[0133] Specifically, the kth subnode indicates any one of the multiple subnodes, and the central node can train the i-th federated learning model of the kth subnode according to the i-th first information and the type of the federated learning model of the kth subnode. 1≦i≦I, where i is a positive integer. In FIG. 5, the kth subnode may be subnode 1, subnode 2, or subnode 3.

[0134] In the following, the type of the i-th associative learning model of the k-th subnode is first described in detail.

[0135] In the present disclosure, the type of the i-th federated learning model of the k-th subnode corresponds to a first value or a second value. The model type corresponding to the first value may indicate that at least two of the multiple subnodes correspond to the same federated learning model. The same federated learning model may also be referred to as a global model or another model name. This is not limited in the present disclosure. When the type of the i-th federated learning model of the k-th subnode corresponds to the first value, it may indicate that the i-th federated learning model of the k-th subnode is the same as the i-th federated learning model of at least one subnode other than the k-th subnode in the multiple subnodes. The type of the i-th federated learning model of the k-th subnode corresponds to the first value, and it can also be described that the type of the i-th federated learning model of the k-th subnode is a global model type.

[0136] The model type corresponding to the second value is for a single subnode and indicates a federated learning model that is different from the federated learning model of another subnode in the plurality of subnodes. The federated learning model that is different from the federated learning model of another subnode may be referred to as an individualized model, or another model name that can be distinguished from the above-mentioned global model. This is not limited in the present disclosure. When the type of the i-th federated learning model of the k-th subnode corresponds to the second value, it may indicate that the i-th federated learning model of the k-th subnode is different from the i-th federated learning model of a subnode other than the k-th subnode in the plurality of subnodes. The type of the i-th federated learning model of the k-th subnode corresponds to the second value, and the type of the i-th federated learning model of the k-th subnode may be described as an individualized model type.

[0137] Optionally, when i is a positive integer ranging from 1 to I, the central node determines that the type of the i-th federated learning model of the k-th subnode corresponds to the first value or the second value. Alternatively, the central node can determine whether the type of the first federated learning model corresponding to the k-th subnode corresponds to the first value or the second value. When the value of i is a positive integer ranging from 2 to I, the i-th federated learning model of the k-th subnode matches the type of the (i-1)-th federated learning model of the k-th subnode by default. Alternatively, for some of the I rounds, the central node can determine that in each of the several rounds, the federated learning model corresponding to the k-th subnode corresponds to the first value or the second value, and for other rounds, by default, the type of the current federated learning model of the k-th subnode matches the type of the previous federated learning model of the k-th subnode, or the federated learning model of the k-th subnode corresponds to the first value or the second value.

[0138] The aforementioned several rounds may be periodic, or may be continuous, or may be dynamically triggered by the federated learning management node or subnode. For example, for some rounds in the I times, the central node determines the type of the federated learning model of the subnode, and the type of the federated learning model of another round is the same as the type of the previous federated learning model by default, or when the remaining rounds include the first round, the type of the federated learning model of the first round corresponds to the first value by default. The possible values ​​of i correspond to one or more rounds in the aforementioned part of the rounds, and the value of i is the number of the corresponding rounds in the I times. For example, when I is 10, some rounds are periodic, and the number of some rounds includes 1, 3, 5, 7, and 9, and the value of i is one or more of 1, 3, 5, 7, and 9, or some rounds are continuous, and the number of some rounds is 1, 2, 3, 4, 5, 6, 7, and 9.

number

[0139] For the above-mentioned manner in which the central node determines the type of the federated learning model of the k-th subnode, please refer to the following manner for understanding. The central node can determine a first value or a second value corresponding to the federated learning model of the k-th subnode by the central node, or the central node can determine the type of the federated learning model in the current training by referring to the performance of the federated learning model obtained through past training. For example, when i is greater than 1, the central node can determine the type of the i-th federated learning model of the k-th subnode based on the performance of the (i-1)-th federated learning model of the k-th subnode. Specifically, in the solutions corresponding to Figures 7 and 9, it is described in detail how the central node determines the type of the current federated learning model by referring to the performance of the historical federated learning model.

[0140] For example, the federated learning model is trained in the i-th round. When the k-th subnode is any one of the multiple subnodes, the type of the i-th federated learning model of the k-th subnode is the same as or different from the type of the i-th federated learning model of at least one subnode other than the k-th subnode among the multiple subnodes. Alternatively, it can be understood that the i-th federated learning model of the k-th subnode is the same as or different from the i-th federated learning model of at least one subnode other than the k-th subnode in the multiple subnodes.

[0141] Furthermore, the following describes in detail a solution in which the central node determines the i-th federated learning model of the k-th subnode.

[0142] (1) When the type of the i-th federated learning model of the k-th subnode corresponds to a first value, the central node can determine the i-th federated learning model of the k-th subnode based on the i-th first information and a weight corresponding to the i-th local model of each of the multiple subnodes.

[0143] In an optional embodiment, referring to S22, the multiple sub-nodes report model parameters of the local models obtained through training by the multiple sub-nodes to the central node.

[0144] for example,

number

number

number

number

[0145]

number

number

number

number

[0146] For example, the corresponding subnode divides the model parameters of the local models of the subnode into multiple combinations and reports the combinations to the central node. It is assumed that the subnode reports the model parameters of the global layer and the model parameters of the individualization layer of the local models to the central node. When the kth subnode is any one of the multiple subnodes and the type of the i-th federated learning model of the kth subnode corresponds to a first value, the central node can determine the i-th federated learning model of the kth subnode in the following manner: The central node performs weighting on the model parameters of the global layer of the i-th local models of the multiple subnodes to obtain parameter information of the global layer of the i-th federated learning model of the kth subnode, and the central node performs weighting on the model parameters of the individualization layer of the i-th local models of the multiple subnodes to obtain parameter information of the individualization layer of the i-th federated learning model of the kth subnode.

number

number

number

[0147]

number

number

number

[0148] For example, when the k-th subnode is any one of the multiple subnodes, when the type of the i-th federated learning model of the k-th subnode corresponds to a first value, the central node is represented by the formula

number

number

number

number

[0149] For example, the corresponding subnode reports the model parameters of the global layer and the gradient of the individualization layer of the local model to the central node. When the kth subnode is any one of the multiple subnodes, and the type of the i-th federated learning model of the kth subnode corresponds to a first value, the central node can determine the i-th federated learning model of the kth subnode in the following manner: The central node performs weighting on the model parameters of the global layer of the i-th local model of the multiple subnodes to obtain the model parameters of the global layer of the i-th federated learning model of the kth subnode, and the central node performs weighting on the model parameters of the global layer of the i-th local model of the kth subnode to obtain the gradient of the individualization layer of the i-th local model of the kth subnode.

number

number

number

number

number

number

number

[0150] In addition, optionally, in the i-th round, a type of the i-th federated learning model of a first group of subnodes in the plurality of subnodes is assumed to correspond to a first value, and the central node corresponds to other subnodes in the plurality of subnodes other than the first group of subnodes.

number

number

number

[0151] (2) When the type of the i-th federated learning model of the k-th subnode corresponds to the second value, the central node may determine the i-th federated learning model of the k-th subnode based on the i-th first information and the similarity between the i-th local model of the k-th subnode and the i-th local model of each of the multiple subnodes.

[0152] In an optional embodiment, referring to S22, the multiple sub-nodes report model parameters of the local models obtained through training by the multiple sub-nodes to the central node.

[0153] For example, if the type of the ith federated learning model of the pth subnode among the multiple subnodes is assumed to correspond to a second value, the central node may be expressed by the following formula:

number

number

[0154]

number

number

number

number

number

[0155] For example, the corresponding subnode divides the model parameters of the local model of the subnode into multiple combinations and reports the combinations to the central node. It is assumed that the subnode reports the model parameters of the global layer and the model parameters of the individualization layer of the local model to the central node. It is assumed that the type of the i-th federated learning model of the p-th subnode in the multiple subnodes corresponds to the second value. The central node can determine the i-th federated learning model of the p-th subnode based on the parameter information of the i-th local model of the multiple subnodes in the following manner: The central node performs weighting on the model parameters of the global layer of the i-th local model of the multiple subnodes to obtain the parameter information of the global layer of the i-th federated learning model of the p-th subnode. The central node performs weighting based on the similarity between the model parameters of the individualization layer of the i-th local model of any subnode k among the K subnodes and the model parameters of the individualization layer of the i-th federated learning model of the p-th subnode to obtain the parameter information of the individualization layer of the i-th federated learning model of the p-th subnode.

number

number

number

[0156] In this way, the central node can determine the first model by referring to the model parameters of the global layer and the model parameters of the individualized layer of the first model.

[0157] In another optional implementation, referring to S22, multiple subnodes report gradients of local models obtained through training to a central node.

[0158] For example, if the type of the ith federated learning model of the pth subnode among the multiple subnodes is assumed to correspond to a second value, the central node may be expressed by the following formula:

number

number

number

number

[0159] For example, the corresponding subnode reports the model parameters of the global layer and the gradient of the individualization layer of the local model to the central node. It is assumed that the type of the i-th federated learning model of the p-th subnode in the multiple subnodes corresponds to the second value, and the central node can determine the i-th federated learning model of the p-th subnode in the following manner: The central node performs weighting on the model parameters of the global layer of the i-th local model of the multiple subnodes to obtain the model parameters of the global layer of the i-th federated learning model of the p-th subnode, and the central node performs weighting on the gradient of the individualization layer of the i-th local model of the p-th subnode to obtain the model parameters of the individualization layer of the i-th federated learning model of the p-th subnode.

number

number

number

[0160] In addition, it may be further understood that when the type of the ith associative learning model of the second group of subnodes in the plurality of subnodes corresponds to the second value, the pth subnode may be any one of the second group of subnodes. Alternatively, the number of subnodes included in the second group of subnodes is denoted as P, where p is any positive integer from 1 to P in the second group of subnodes. 1≦P≦K.

[0161] In conclusion, for each of I times of training the federated learning model, the federated learning model of multiple subnodes obtained through training can satisfy one of the following three results. For example, the following three results are explained by using the i-th time of I times as an example.

[0162] Result 1: The types of the i-th federated learning models of the multiple subnodes all correspond to a first value. As a result, the number U of subnodes included in the first group of subnodes in the above example is equal to K, and the i-th federated learning models of the multiple subnodes are the same.

[0163] Result 2: The types of the i-th federated learning models of the multiple subnodes all correspond to a second value. As a result, the number P of subnodes included in the second group of subnodes in the above example is equal to K, and the i-th federated learning models of the multiple subnodes are different.

[0164] Result 3: The type of the i-th federated learning model of the first group of subnodes in the plurality of subnodes corresponds to a first value, and the type of the i-th federated learning model of the second group of subnodes corresponds to a second value. As a result, the number U of subnodes included in the first group of subnodes in the above example is less than K, and the number P of subnodes included in the second group of subnodes is less than K, and K=U+P. The i-th federated learning models of all subnodes in the second group of subnodes are different, and the i-th federated learning models of all subnodes in the first group of subnodes are different. When U is greater than 1, the i-th federated learning models of all subnodes in the first group of subnodes are the same. In addition, corresponding to the solution in the above example in which the federated learning model is divided into a global layer and an individualization layer, the i-th federated learning model may be represented by a first part of parameter information (e.g., parameter information of a global layer) and a second part of parameter information (e.g., individualization layer parameters). The first part of parameter information of all i-th federated learning models of the plurality of subnodes is the same. The second portions of the parameter information of the i-th federated learning models of all subnodes in the second group of subnodes are different and different from the second portions of the parameter information of the i-th federated learning models of all subnodes in the first group of subnodes. When U is greater than 1, the i-th federated learning models of all subnodes in the first group of subnodes are the same.

[0165] S24: The central node sends information indicating the corresponding federated learning model to each subnode.

[0166] Specifically, the i-th time out of the above I times is used as an example. The central node can transmit the i-th second information, and the i-th second information indicates the i-th federated learning model of each of the multiple subnodes. For example, the i-th second information includes information indicating the i-th federated learning model of each subnode, and the information indicating the i-th federated learning model of one subnode may be parameter information of the i-th federated learning model of the subnode. The parameter information of the federated learning model of the subnode may include model parameters of the federated learning model. Alternatively, the difference between the parameter information of the federated learning model trained this time for the subnode and the parameter information of the federated learning model trained last time may be included. The difference may also be described as the gradient of the model parameters of the federated learning model trained this time, or may be simply referred to as the gradient of the federated learning model trained this time. 1≦i≦I, where i is a positive integer.

[0167] Referring to S23, in any one of I times, if the type of the federated learning model of the subnode corresponds to a first value or a second value, when the type of the federated learning model corresponds to the first value, the parameter information of the federated learning model may be understood by referring to the description of (1) in S23, and when the type of the federated learning model corresponds to the second value, the parameter information of the federated learning model may be understood by referring to the description of (2) in S23. Details are not described again in this specification in the present disclosure.

[0168] When the type of the federated learning model corresponds to a first value, the central node can transmit information used for the federated learning model in a broadcast or multicast manner. When the type of the federated learning model corresponds to a second value, the central node can transmit information used for the federated learning model to the corresponding subnode in a unicast manner. The result 3 corresponding to the federated learning model of the multiple subnodes trained in the i-th time described in S23 is used as an example as follows.

[0169] In an optional manner, the federated learning model is not divided into a global layer and an individualized layer, and the central node may transmit the i-th second information according to the following steps: transmitting first instruction information to a first group of subnodes in a broadcast or multicast manner, where the first instruction information indicates the i-th federated learning model of each subnode in the first group of subnodes; and transmitting second instruction information to a p-th subnode in a second group of subnodes in a unicast manner, where the second instruction information indicates the i-th federated learning model of the p-th subnode.

[0170] In another optional manner, the federated learning model is divided into a global layer and an individualized layer. The i-th federated learning model may be represented by a first part of parameter information (e.g., global layer parameter information) and a second part of parameter information (e.g., individualized layer parameters). The central node may transmit the i-th second information according to the following steps: transmitting the first part of parameter information of the i-th federated learning model to a plurality of subnodes in a broadcast or multicast manner; transmitting the second part of parameter information of the i-th federated learning model of each subnode in the first group of subnodes to the first group of subnodes in a broadcast or multicast manner; and transmitting the second part of parameter information of the p-th subnode to the p-th subnode in the second group of subnodes in a unicast manner, where p is any positive integer from 1 to P, and P is the number of subnodes included in the second group of subnodes.

[0171] The resource used by the central node to transmit information through broadcast or multicast may be the resource used by the central node to transmit the second model in S501, and the resource used by the central node to transmit information through broadcast or multicast may be the resource reconfigured by the central node or the federated learning management node. Similarly, the resource used by the central node to transmit information through unicast may be the resource configured by the central node or the federated learning management node. The broadcast or multicast resource and the unicast resource may be different, partially overlapping, or the same. This is not limited in the present disclosure.

[0172] In addition, the central node may further transmit auxiliary information. The auxiliary information may indicate to each subnode whether the type of the federated learning model transmitted by the central node to each subnode corresponds to a first value or a second value. For example, the first value is 1 and the second value is 0, or the first value is 0 and the second value is 1, or the first value indicates a global type and the second value indicates an individualized type. The auxiliary information may further include relevant parameters used to assist the subnode in training the local model, for example, a loss function available for training the local model.

[0173] Further, referring to S23, it can be understood that the kth subnode indicates any one of a plurality of subnodes. When the value of i is a positive integer ranging from 1 to I-1, the kth subnode can obtain parameter information of the ith federated learning model, and determine the (i+1)th local model of the kth subnode based on the parameter information of the ith federated learning model. Alternatively, it can be understood that when i is greater than 1, the parameter information of the (i-1)th federated learning model of the kth subnode is used to determine the ith local model of the kth subnode.

[0174] Optionally, when i is greater than 1, the corresponding loss function used when the kth subnode trains the ith local model may be related to the local data collected by the kth subnode and determined by the type of the (i-1)th federated learning model of the kth subnode.

[0175] If the type of the (i-1)-th federated learning model of the k-th subnode corresponds to a first value, the loss function is primarily related to the local data set collected by the k-th subnode. For example, an MSE loss function may be applied to the local model training by the k-th subnode,

number

number

[0176] When the type of the (i-1)th federated learning model of the kth subnode corresponds to the second value, a part of the loss function of the local model trained by the kth subnode is related to the local data collected by the kth subnode, and another part is related to the difference between the parameter information of the trained local model and the parameter information of the (i-1)th federated learning model. The difference may be expressed by parameters such as Euclidean distance and squared Euclidean distance.

[0177] For example, if the parameter information of the (i-1)-th federated learning model of the k-th subnode contains all the model parameters of the model, the loss function of the k-th subnode is

number

number

number

number

number

number

[0178] For example, if the parameter information of the (i-1)th federated learning model of the kth subnode includes model parameters of the global layer and model parameters of the individualization layer of the model, the loss function of the kth subnode is

number

number

number

number

number

number

[0179] In conclusion, it can be understood that the value of I in S502 may be pre-configured, or the value of I is determined based on a model convergence condition, and the I-th federated learning model corresponding to the k-th subnode satisfies the model convergence condition. The model convergence condition can be that the performance of the I-th federated learning model of the k-th subnode is equal to or greater than a preset performance threshold, or the difference between the I-th federated learning model of the k-th subnode and the (I-1)-th federated learning model of the k-th subnode is less than a first threshold, or the difference between the I-th federated learning model of the k-th subnode and the (I-1)-th federated learning model of the k-th subnode and the (I-2)-th federated learning model of the k-th subnode is less than a first threshold, that is, the difference between the models tends to be stable.

[0180] In the above solution provided in the present disclosure, the central node can train a federated learning model for each subnode, and provide an individualized model for the subnode whose data distribution is different from that of another subnode, which can help improve the performance of the local models of these subnodes and improve the performance of the federated learning model.

[0181] In the following, the training method of the federated learning model in the present disclosure is further described in detail by using an example in which the central node needs to refer to the performance of the (i-1)th federated learning model of the kth subnode to determine the i-th federated learning model of the kth subnode. The value of i in this specification may be any of the 2nd to Ith times shown in FIG. 5, or any of the several times described in S23.

[0182] Figure 7 shows a communication method. The method mainly includes the following steps:

[0183] S701: The central node obtains the i-th first information.

[0184] In particular, details can refer to the solution in S22 for implementation, and will not be described again herein in this disclosure.

[0185] S702: The central node obtains a first test result from the k-th subnode.

[0186] The kth subnode is any one of the multiple subnodes. The first test result indicates the performance of the (i-1)th associative learning model of the kth subnode, where i is greater than 1.

[0187] Specifically, after the kth subnode can test the performance of the (i-1)th federated learning model of the kth subnode, the kth subnode actively transmits the first test result to the central node. Alternatively, the kth subnode can test the performance of the (i-1)th federated learning model under the instruction of the central node. For example, the central node may send a test signaling to the kth subnode, and the kth subnode may report the first test result to the central node after the kth subnode performs a performance test based on the test signaling. The test signaling in this specification instructs the kth subnode to test the performance of the (i-1)th federated learning model of the kth subnode.

[0188] Specifically, the kth subnode can test the performance of the federated learning model in the following manner: the kth subnode uses the local data as input data, inputs the local data into the federated learning model to be tested, obtains the output of the federated learning model, i.e., the test result, and calculates the loss of the federated learning model by using the test result and the label of the local data. A larger loss indicates a poorer performance of the (i-1)th federated learning model.

[0189] Optionally, the first test result may include at least one of the following parameters: a type of the (i-1)th federated learning model; information indicating whether the loss of the (i-1)th federated learning model is greater than a loss threshold. For example, the information may be a first value or a second value, where the first value indicates that the loss is greater than the loss threshold, indicating a large loss, and the second value indicates that the loss is less than the loss threshold, indicating a small loss, which is the loss of the (i-1)th federated learning model.

[0190] In addition, optionally, the kth subnode may transmit the first test result including the loss of the (i-1)th federated learning model only when the loss of the (i-1)th federated learning model is large. When the loss of the (i-1)th federated learning model is small, the first test result does not include the loss of the (i-1)th federated learning model, but includes information indicating that the performance of the (i-1)th federated learning model meets the requirement, or when the loss of the (i-1)th federated learning model is small, the first test result is not transmitted, in which case, when the central node does not receive the first test result, it may be considered by default that the performance of the (i-1)th federated learning model meets the requirement. When the kth subnode does not receive the first test result, or when the first test result indicates that the loss of the (i-1)th federated learning model is small, it may be determined that the type of the ith federated learning model is consistent with the type of the (i-1)th federated learning model.

[0191] Specifically, the central node may send a test signaling to the kth subnode, and after the kth subnode performs a test based on the test signaling, the kth subnode reports a first test result to the central node. Some optional implementation forms of sending the test signaling by the central node are further described below.

[0192] In an optional implementation, the central node may periodically send a test signaling to the kth subnode, where the test signaling indicates to the kth subnode to test the performance of the federated learning model recently received by the kth subnode. In addition, the test signaling may further include a loss threshold and a loss function type.

[0193] The central node periodically transmits the test signaling may be understood as follows: Some rounds described in S23 are periodic, and the central node transmits the test signaling before training the federated learning model in these rounds. The i-th round may be one of the several rounds. Based on this, the central node obtaining the first test result from the k-th subnode may include the central node transmitting the test signaling to the k-th subnode in the i-th round, and the test signaling may instruct the k-th subnode to test the performance of the (i-1)-th federated learning model of the k-th subnode, and the k-th subnode obtains the performance of the (i-1)-th federated learning model through a test based on the test signaling, and the k-th subnode transmits the first test result to the central node.

[0194] Alternatively, the central node periodically transmitting the test signaling may be understood as follows: the test periodicity is configured when a resource for transmitting information to the k-th subnode is configured by the central node, and the k-th subnode tests the performance of the received federated learning model based on the test periodicity, and the testing includes a test operation on the performance of the (i-1)-th federated learning model. On this basis, the central node obtaining the first test result from the k-th subnode may include that when the k-th subnode transmits parameter information of the i-th local model, the k-th subnode needs to further transmit the first test result.

[0195] In another optional implementation, the central node can send a test signaling to the k-th subnode under the triggering of the k-th subnode, and the test signaling indicates the k-th subnode to test the performance of the federated learning model most recently received by the k-th subnode. For the method of testing the performance of the federated learning model by the k-th subnode, please refer to the above implementation for understanding. The details will not be described again in this disclosure. In addition, the test signaling may further include a loss threshold and a loss function type.

[0196] Specifically, when the kth subnode determines that the performance of the (i-1)th federated learning model is abnormal, the kth subnode can report information indicating that the performance is abnormal to the central node. The kth subnode can input local data into the (i-1)th federated learning model, and when the calculated loss is greater than a specified threshold, the kth subnode can determine that the performance of the (i-1)th federated learning model is abnormal; or when the kth subnode can refer to environmental information or another interference factor to find that the (i-1)th federated learning model is not applicable, the kth subnode can determine that the performance of the (i-1)th federated learning model is abnormal. For example, in positioning, the subnode is a terminal device. When the terminal device moves from one environment to another, the channel information of the two environments is significantly different, and the result of positioning performed by using the same federated learning model is insufficient. In this case, the terminal device can find that a sudden change occurs in the positioning result and determine that the performance of the federated learning model is abnormal.

[0197] The central node receives the above information indicating that the performance is abnormal and sends test signaling to the kth subnode in the ith round, where the test signaling can instruct the kth subnode to test the performance of the (i-1)th federated learning model of the kth subnode, the kth subnode obtains the performance of the (i-1)th federated learning model through testing based on the test signaling, and the kth subnode sends the first test result to the central node.

[0198] S703: The central node determines the type of the i-th federated learning model of the k-th subnode based on the first test result.

[0199] First, the central node can determine, based on the first test result, whether the type of the i-th federated learning model of the k-th subnode corresponds to a first value or a second value.

[0200] Specifically, if the first test result indicates that the performance of the (i-1)th federated learning model of the kth subnode is poor, the central node can determine that the type of the i-th federated learning model of the kth subnode is different from the type of the (i-1)th federated learning model, or if the first test result indicates that the performance of the (i-1)th federated learning model of the kth subnode is good, the central node can determine that the type of the i-th federated learning model of the kth subnode is the same as the type of the (i-1)th federated learning model. In addition, for the case of S702, if the kth subnode does not report the first test result, the central node can consider that the type of the i-th federated learning model of the kth subnode is the same as the type of the (i-1)th federated learning model by default.

[0201] For example, Figure 8A shows a logic for determining the type of the federated learning model. When the type of the (i-1)th federated learning model corresponds to a first value, if the first test result shows a small loss, i.e., the performance of the (i-1)th federated learning model is good, the central node can determine that the type of the i-th federated learning model corresponds to a first value. Alternatively, if the first test result shows a large loss, i.e., the performance of the (i-1)th federated learning model is insufficient, the central node can determine that the type of the i-th federated learning model corresponds to a second value.

[0202] For example, Figure 8B shows the logic for determining the type of the federated learning model. When the type of the (i-1)th federated learning model corresponds to the second value, if the first test result shows a small loss, i.e., the performance of the (i-1)th federated learning model is good, the central node can determine that the type of the i-th federated learning model corresponds to the second value. Alternatively, if the first test result shows a large loss, i.e., the performance of the (i-1)th federated learning model is insufficient, the central node can determine that the type of the i-th federated learning model corresponds to the first value.

[0203] It can be seen that in the above scheme, when the central node determines that the performance of the (i-1)th federated learning model of any subnode is insufficient, the central node can adjust the type of the i-th federated learning model of any subnode. In addition, as another optional implementation form, a threshold of the number of adjusted subnodes can also be specified. If the number of subnodes of the (i-1)th federated learning model with insufficient performance exceeds the threshold, the type of the i-th federated learning model can be preferentially adjusted to be different from the type of the (i-1)th federated learning model for the subnode with large loss within the threshold.

[0204] S704: The central node can determine the i-th federated learning model of the k-th subnode based on the i-th first information and the type of the i-th federated learning model of the k-th subnode.

[0205] Specifically, please refer to the solution described in S23 for the implementation form, and the details will not be described again here in this disclosure.

[0206] S705: The central node sends the i-th second information.

[0207] For understanding of this step, please refer to the solution described in S24, the details will not be described again here in this disclosure.

[0208] 9 shows a communication method. The method can be applied when the type of the (i-1)-th federated learning model of the k-th subnode corresponds to a second value. The method includes the following steps:

[0209] S901: The central node obtains the i-th first information.

[0210] Specifically, please refer to the solution of S22 for implementation, the details of which will not be described again here in this disclosure.

[0211] S902: The central node sends information indicating the first model to the k-th subnode.

[0212] The first model is determined based on the (i-1)th first information and the weight corresponding to the (i-1)th local model of each of the plurality of subnodes, where i is greater than 1. Specifically, when the kth subnode is any one of the plurality of subnodes, if the type of the (i-1)th federated learning model trained by the central node for at least two of the plurality of subnodes in the (i-1)th round corresponds to a first value, in S902, the central node can transmit parameter information of the (i-1)th federated learning model of the at least two subnodes. Alternatively, the central node may determine the parameter information of the first model to be transmitted in S902 based on the acquired (i-1)th first information and the weight corresponding to the (i-1)th local model of each of the plurality of subnodes. The (i-1)th first information includes parameter information of the (i-1)th local model of each of the plurality of subnodes.

[0213] S903: The central node obtains a second test result from the k-th subnode.

[0214] The second test result indicates the performance of the first model, or the second test result indicates the performance of the first model and the performance of the (i-1)th associative learning model of the kth subnode, or the second test result indicates the performance of the (i-1)th associative learning model of the kth subnode.

[0215] Specifically, in S902, after the kth subnode tests the performance of the first model of the kth subnode and / or the performance of the (i-1)th federated learning model, the kth subnode actively transmits the second test result to the central node. Alternatively, the kth subnode may perform a performance test under the instruction of the central node. For example, the central node may send a test signaling to the kth subnode, and the kth subnode may report the second test result to the central node after the kth subnode performs a performance test based on the test signaling. The test signaling in this specification instructs the kth subnode to test the performance of the first model in S902 and / or the performance of the (i-1)th federated learning model of the kth subnode. For optional implementations related to transmitting test signaling by the central node, please refer to S702. Details will not be described again in this specification in this disclosure.

[0216] Specifically, please refer to the solution described in S702. It can be understood that the kth subnode tests the performance of the first model or the (i-1)th federated model. Details are not described again in this disclosure. In the following, the information contained in the second test result is described by using an example.

[0217] In the first scheme, the kth subnode tests the performance of the first model and the performance of the (i-1)th associative learning model based on a decision made by the kth subnode or based on a test signaling of the central node.

[0218] For example, the kth subnode can test the performance of the first model and the performance of the (i-1)th federated learning model, i.e., calculate the loss of the first model and the loss of the (i-1)th federated learning model.The kth subnode then includes at least one of the following information in the second test result: the loss of the first model and the loss of the (i-1)th federated learning model, the third indication information indicating that the loss of the first model is greater than the loss of the (i-1)th federated learning model or the third indication information indicating that the loss of the first model is smaller than the loss of the (i-1)th federated learning model, and the fourth indication information indicating the model type having the large loss or small loss and the difference between the loss of the first model and the loss of the (i-1)th federated learning model.

[0219] For example, as shown in FIG. 10A, after the kth subnode calculates the loss of the first model and the loss of the (i-1)th federated learning model, the kth subnode determines whether the loss of the (i-1)th federated learning model is greater than the loss of the first model. If so, the kth subnode can include the fifth instruction information in the second test result. The fifth instruction information can indicate that the performance of the first model is better than the performance of the (i-1)th federated learning model, and can implicitly indicate that the type of the i-th federated learning model corresponds to the first value. If not, the kth subnode continues to determine whether the loss of the first model is greater than the loss threshold. When the loss of the first model is greater than the loss threshold, the kth subnode can include the sixth instruction information in the second test result. The sixth instruction information can indicate that the performance of the (i-1)th federated learning model is better than the performance of the first model, and can implicitly indicate that the type of the i-th federated learning model is the same as the type of the (i-1)th federated learning model. When the loss of the first model is equal to or less than the specified threshold, the kth subnode may include seventh indication information in the second test result. The seventh indication information may indicate that the performance of the (i-1)th federated learning model is better than that of the first model, but the loss of the first model is equal to or less than the loss threshold. The seventh indication information may implicitly indicate that the type of the ith federated learning model may correspond to the first value or the second value.

[0220] For example, as shown in FIG. 10B, the kth subnode may first calculate the loss of the (i-1)th federated learning model, and determine whether the loss of the (i-1)th federated learning model is greater than the loss threshold. If so, the kth subnode may include the eighth instruction information in the second test result, and the eighth instruction information may indicate that the loss of the (i-1)th federated learning model is greater than the loss threshold, and may implicitly indicate that the type of the i-th federated learning model corresponds to the first value. If not, the kth subnode may continue to calculate the loss of the first model, and determine whether the loss of the (i-1)th federated learning model is greater than the loss of the first model. If so, the kth subnode may include the fifth instruction information in the second test result, and the fifth instruction information may indicate that the performance of the first model is better than the performance of the (i-1)th federated learning model, and may implicitly indicate that the type of the i-th federated learning model corresponds to the first value. If not, the kth subnode may include the sixth instruction information in the second test result. The sixth instruction information may indicate that the performance of the (i-1)th associative learning model is better than the performance of the first model, and may implicitly indicate that the type of the i-th associative learning model is the same as the type of the (i-1)th associative learning model.

[0221] In addition, optionally, for the procedure shown in FIG. 10B, the central node may transmit the first model only when the loss of the (i-1)th federated learning model is equal to or less than the loss threshold. For example, after executing S901, the central node may first execute S903 to obtain the second test result. If the second test result obtained by the central node includes the eighth indication information, the central node may not transmit the first model, that is, after executing S901 and S903, the central node directly executes S904 to S906. In another example, after executing S901, the central node may first obtain the third test result. The third test result indicates that the loss of the (i-1)th federated learning model is equal to or less than the loss threshold. In this case, the central node may execute S902 to S906 based on the third test result. Such a design may reduce the signaling overhead for transmitting the first model by the central node.

[0222] In the second scheme, the kth subnode only tests the performance of the first model based on the decision made by the kth subnode or based on the test signaling of the central node. In this way, the second test result can indicate the performance of the first model. For example, the second test result can include information indicating the loss of the first model, or whether the loss of the first model is greater than the loss threshold. The information can be a third value or a fourth value. The third value indicates that the loss is greater than the loss threshold, indicating a large loss, and implicitly indicates that the type of the ith federated learning model is the same as the type of the (i-1)th federated learning model. The fourth value indicates that the loss is equal to or less than the loss threshold, indicating a small loss, and implicitly indicates that the type of the ith federated learning model corresponds to the first value.

[0223] In the third scheme, the kth subnode only tests the performance of the (i-1)th federated learning model based on the decision made by the kth subnode or based on the test signaling of the central node. In this way, the second test result can indicate the performance of the (i-1)th federated learning model. For example, the second test result can include information indicating the loss of the (i-1)th federated learning model, or whether the loss of the (i-1)th federated learning model is greater than a loss threshold, where the information can be a fifth value or a sixth value. The fifth value indicates that the loss is greater than the loss threshold, indicating a large loss, and implicitly indicates that the type of the ith federated learning model corresponds to the first value. The sixth value indicates that the loss is equal to or less than the loss threshold, indicating a small loss, and implicitly indicates that the type of the ith federated learning model is the same as the type of the (i-1)th federated learning model.

[0224] S904: The central node determines the type of the i-th associative learning model of the k-th subnode based on the second test result.

[0225] Specifically, the central node can determine, based on the second test result, whether the type of the i-th federated learning model of the k-th subnode corresponds to the first value or the second value.

[0226] For example, referring to the first method in S903, When the second test result includes the loss of the first model and the loss of the (i-1)th federated learning model, the central node can determine that the type of the model with the small loss is the type of the i-th federated learning model; or when the second test result includes the third indication information, the type of the i-th federated learning model of the k-th subnode corresponds to the second value; or when the second test result includes the fourth indication information and the difference is greater than a specified difference threshold, the type of the model with the small loss is determined to be the type of the i-th federated learning model; or when the difference is equal to or less than the specified difference threshold, the type of the i-th federated learning model is determined to correspond to the first value or the second value; Or, when the second test results include the fifth instruction information, the type of the i-th associative learning model of the k-th subnode corresponds to the first value, or, when the second test results include the sixth instruction information, the type of the i-th associative learning model of the k-th subnode corresponds to the second value, or, when the second test results include the fifth instruction information, it is determined that the type of the i-th associative learning model corresponds to the first value or the second value, or, when the second test results include the sixth instruction information, the type of the i-th associative learning model of the k-th subnode corresponds to the first value.

[0227] For example, referring to the second manner of S903, when the central node determines, based on the second test result, that the loss of the first model is greater than a loss threshold, the central node determines that the type of the i-th federated learning model of the k-th subnode corresponds to a second value, or when the central node determines that the loss of the first model is less than or equal to a loss threshold, the central node determines that the type of the i-th federated learning model of the k-th subnode corresponds to a first value.

[0228] For example, referring to the third scheme of S903, when the central node determines based on the second test result that the loss of the (i-1)th federated learning model is greater than a loss threshold, the central node determines that the type of the i-th federated learning model of the k-th subnode corresponds to a first value; or when the central node determines that the loss of the (i-1)th federated learning model is less than or equal to a loss threshold, the central node determines that the type of the i-th federated learning model of the k-th subnode corresponds to a second value.

[0229] In the above scheme, it can be understood that when the central node determines that the performance of the (i-1)th federated learning model of any subnode is insufficient, the central node can adjust the type of the i-th federated learning model of any subnode. In addition, as another optional implementation form, a threshold of the number of adjusted subnodes can also be specified. If the number of subnodes of the (i-1)th federated learning model with insufficient performance exceeds the threshold, the type of the i-th federated learning model can be preferentially adjusted to be different from the type of the (i-1)th federated learning model for the subnode with large loss within the threshold.

[0230] S905: The central node can determine the i-th federated learning model of the k-th subnode based on the i-th first information and the type of the i-th federated learning model of the k-th subnode.

[0231] Specifically, please refer to the solution described in S23 for the implementation form, and the details will not be described again here in this disclosure.

[0232] S906: The central node sends the i-th second information.

[0233] For understanding of this step, please refer to the solution described in S24, the details will not be described again here in this disclosure.

[0234] In the solution of FIG. 7 or FIG. 9, the type of federated learning model for current training is determined with reference to the performance of the global model and / or the individualized model obtained through past training, and a corresponding model is trained based on the type of federated learning model, so that the performance of the model can be improved.

[0235] In addition, the present disclosure does not limit the number of subnodes participating in federated learning. The number of subnodes may be a fixed number of subnodes or may be dynamically changed. For example, in the I-th training process, a new subnode is allowed to participate in federated learning.

[0236] Figure 11 shows a communication method. The method mainly includes the following steps:

[0237] S1101: A new subnode sends third information to a central node, where the third information is used to request to participate in federated learning.

[0238] S1102: The central node sends resource configuration information to the new subnode.

[0239] The resource configuration information is used to configure resources used by the central node to transmit the third model. The third model may be understood as a federated learning model whose type obtained through the most recent training corresponds to the first value based on the local models of the existing subnodes participating in the federated learning. Optionally, the resource configuration information is further used to configure resources used by the new subnode to transmit parameter information of the local model for the new subnode.

[0240] Alternatively, the federated learning management node can configure resources used by the central node to send the third model to the new subnode and resources used by the new subnode to send parameter information of the local model.

[0241] S1103: The central node sends information indicating the third model to the new subnode.

[0242] Specifically, in S1102, the central node may send the parameter information of the third model to the new subnode in a broadcast or multicast manner based on the resource configuration information. For the parameter information of the third model, please refer to S501 for understanding. The details will not be described again in this disclosure.

[0243] Additionally, the central node may instruct the new subnode to run performance tests against a third model.

[0244] S1104: The new subnode sends a fourth test result to the central node, where the fourth test result indicates the performance of the third model.

[0245] For the new subnode, the performance of the third model may be tested with reference to the solution in S702. Details will not be described again herein in this disclosure. The fourth test result includes the loss of the third model, or information indicating whether the loss of the third model is greater than a loss threshold.

[0246] S1105: The central node sends fourth information to the new subnode based on the fourth test result.

[0247] In one scheme, when the central node determines based on the fourth test result that the loss of the third model is greater than the loss threshold, the central node may determine that the type of federated learning model suitable for the new subnode corresponds to the second value, i.e., the new subnode is suitable for an individualized type of federated learning model. In this case, the central node may include in the fourth information the following information, i.e., information indicating the new subnode using the individualized type of federated learning model, and related auxiliary information for training the local model by the new subnode, such as parameter information of the initial model (i.e., the second model) used to train the local model, the current iteration number for training the local model, the learning rate, the loss function, the batch size, and the parameter type (model parameter or gradient) of the reported local model.

[0248] The current number of iterations for training the local model is related to the number of iterations for training the local model for the remaining subnodes participating in the federated learning before the new subnode receives the fourth information. For example, in each of the above I rounds of training the federated learning model, the number of iterations for training the local model by the corresponding subnode is 20. If the new subnode has performed four federated learning model trainings before receiving the fourth information, in this case, the current number of iterations for training the local model by the new subnode is 80.

[0249] In another manner, when the central node determines based on the fourth test result that the loss of the third model is equal to or less than the loss threshold, the central node may determine that the type of federated learning model suitable for the new subnode corresponds to the first value, that is, the new subnode is suitable for a global type federated learning model. In this case, the central node may include the following information in the fourth information: Information indicating a new subnode that uses an individualized type of federated learning model; and related auxiliary information for training a local model based on the third model by the new subnode, such as the number of iterations, learning rate, loss function, batch size, and parameter type (model parameters or gradients) of the reported local model for training the local model in each round of training the federated learning model.

[0250] In addition, the central node then sends the individualized model to the new subnode in a unicast manner, and the unicast resource can be configured by the central node or the federated learning management node. In the above manner, the related auxiliary information for training the local model by the new subnode may further include the configured unicast resource. Alternatively, the central node sends the global model to the new subnode in a broadcast or multicast manner. The broadcast or multicast resource may be the resource allocated in S1103, or may be reconfigured by the central node or the federated learning management node.

[0251] S1106: The new subnode trains a local model based on the fourth information.

[0252] S1107: The new subnode transmits information indicating the local model to the central node.

[0253] This step may be understood with reference to S21, and details will not be described in this disclosure.

[0254] In this case, the new subnode can be considered to have successfully participated in the federated learning. For the subsequent procedures, please refer to one or more solutions in Figures 5 to 10B. The details will not be described again in this disclosure.

[0255] In the solution of FIG. 11, the newly added subnode is supported to use an individualized type model, so that the problem of the poor performance of the newly added subnode in the federated learning can be avoided when the newly added subnode uses a global type model, which can improve the model performance.

[0256] Based on the same concept, please refer to FIG. 12. The present disclosure provides a communication device 1200. The communication device 1200 includes a processing module 1201 and a communication module 1202. The communication device 1200 may be a k-th subnode, or a communication device used in or with the k-th subnode, and can implement a communication method performed on the k-th subnode side. Alternatively, the communication device 1200 may be a central node, or a communication device used in or with the central node, and can implement a communication method performed on the central node side.

[0257] The communication module may also be referred to as a transceiver module, a transceiver, a transceiver machine, a transceiver device, etc. The processing module may also be referred to as a processor, a processing board, a processing unit, a processing device, etc. Optionally, a device in the communication module and configured to implement a receiving function may be regarded as a receiving unit. It should be understood that the communication module is configured to perform a transmitting operation and a receiving operation at the k-th subnode side or the central node side in the embodiment of the above method, and a device in the communication module and configured to implement a transmitting function is regarded as a transmitting unit. In other words, the communication module includes a receiving unit and a transmitting unit.

[0258] When the communication device 1200 is used in a central node, the receiving unit included in the communication module 1202 of the communication device 1200 is configured to perform a receiving operation on the central node side, for example, to receive information from the k-th subnode. The transmitting unit included in the communication module 1202 is configured to perform a transmitting operation on the central node side, for example, to transmit information to the k-th subnode. When the communication device 1200 is used in a k-th subnode, the receiving unit included in the communication module 1202 of the communication device 1200 is configured to perform a receiving operation on the k-th subnode side, for example, to receive information from the central node, and the transmitting unit included in the communication module 1202 of the communication device 1200 is configured to perform a transmitting operation on the k-th subnode side, for example, to transmit information to the central node. When the communication device 1200 is used in a central node, a receiving unit included in the communication module 1202 of the communication device 1200 is configured to perform a receiving operation on the central node side, e.g., receive information from the central node, and a transmitting unit included in the communication module 1202 of the communication device 1200 is configured to perform a transmitting operation on the central node side, e.g., transmit information to the central node.

[0259] In addition, it should be noted that if the device is implemented by using a chip / chip circuit, the communication module may be an input / output circuit and / or a communication interface, performing input operations (corresponding to the receiving operations described above) and output operations (corresponding to the transmitting operations described above). The processing module is an integrated processor, a microprocessor, or an integrated circuit.

[0260] The division into modules in the present disclosure is an example, and is merely a division into logical functions, and may be other divisions in actual implementation. In addition, the functional modules in the embodiments of the present disclosure may be integrated into one processor, or may exist physically alone, or two or more modules may be integrated into one module. The integrated module may be implemented in the form of hardware, or in the form of a software functional module.

[0261] Based on the same technical concept, the present disclosure further provides a communication device 1300. The communication device 1300 may be a chip or a chip system. Optionally, in the present disclosure, the chip system may include a chip, or may include a chip and another discrete device.

[0262] The communication device 1300 may be configured to implement the functions of any network element in the communication system shown in FIG. 1A, FIG. 3A, and FIG. 3B. The communication device 1300 may include at least one processor 1310. The processor 1310 is coupled to a memory. Optionally, the memory may be located in the device, the memory may be integrated with the processor, or the memory may be located outside the device. For example, the communication device 1300 may further include at least one memory 1320. The memory 1320 stores computer programs, configuration information, computer programs or instructions, and / or data required to implement any one of the aforementioned embodiments. The processor 1310 may execute the computer programs stored in the memory 1320 to complete the method in any one of the aforementioned embodiments.

[0263] The communication device 1300 may further include a communication interface 1330, through which the communication device 1300 may exchange information with other devices. For example, the communication interface 1330 may be a transceiver, a circuit, a bus, a module, a pin, or another type of communication interface. If the communication device 1300 is a chip device or circuit, the communication interface 1330 in the communication device 1300 may be an input / output circuit, which may input information (or receive information) and output information (or transmit information). The processor may be an integrated processor, a microprocessor, an integrated circuit, or a logic circuit. The processor may determine output information based on the input information.

[0264] The coupling in this disclosure may be an indirect coupling or communication connection between devices, units, or modules in an electrical, mechanical, or other form, and is used for information exchange between the devices, units, or modules. The processor 1310 can cooperate with the memory 1320 and the communication interface 1330. The specific connection medium between the processor 1310, the memory 1320, and the communication interface 1330 is not limited in this disclosure.

[0265] Optionally, referring to FIG. 13, the processor 1310, the memory 1320, and the communication interface 1330 are connected to each other via a bus 1340. The bus 1340 can be a peripheral component interconnect (PCI) bus, an extended industry standard architecture (EISA) bus, etc. The bus can be classified into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is shown in FIG. 13, but this does not mean that there is only one bus or only one type of bus.

[0266] In this disclosure, a processor may be a general purpose processor, a digital signal processor, an application specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware assembly, which may implement or execute the methods, steps, and logic block diagrams disclosed in this disclosure. A general purpose processor may be a microprocessor, any conventional processor, etc. The steps of the methods disclosed with reference to this disclosure may be performed directly by a hardware processor, or may be performed by a combination of hardware and software modules in a processor.

[0267] In the present disclosure, the memory may be a non-volatile memory, such as a hard disk drive (HDD) or a solid-state drive (SSD), or a volatile memory, such as a random-access memory (RAM). The memory is any other medium capable of carrying or storing expected program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto. The memory in the present disclosure may alternatively be a circuit or any other device capable of implementing a storage function and configured to store program instructions and / or data.

[0268] In a possible implementation, the communication device 1300 may be used in a central node. Specifically, the communication device 1300 may be a central node or a device that supports a central node and can implement the functionality of the central node in any one of the aforementioned embodiments. The memory 1320 stores computer programs, computer programs or instructions, and / or data required to implement the functionality of the central node in any one of the aforementioned embodiments. The processor 1310 can execute the computer programs stored in the memory 1320 to complete the method performed by the central node in any one of the aforementioned embodiments. When used in a central node, the communication interface in the communication device 1300 may be configured to interact with the kth subnode and transmit information to or receive information from the kth subnode.

[0269] In another possible implementation, the communication device 1300 may be used in a kth subnode. Specifically, the communication device 1300 may be a kth subnode or a device that supports the kth subnode and can implement the functionality of the kth subnode in any one of the aforementioned embodiments. The memory 1320 stores computer programs, computer programs or instructions, and / or data required to implement the functionality of the kth subnode in any one of the aforementioned embodiments. The processor 1310 may execute the computer programs stored in the memory 1320 to complete the method performed by the kth subnode in any one of the aforementioned embodiments. When used in the kth subnode, the communication interface in the communication device 1300 may be configured to interact with a central node and transmit information to or receive information from the central node.

[0270] The communication device 1300 provided in this embodiment may be used in a central node to complete a method performed by a central node, or may be used in a k-th subnode to complete a method performed by a k-th subnode. Therefore, for technical effects that can be achieved by the communication device 1300, please refer to the above-mentioned method embodiments. Details will not be described again in this specification.

[0271] Based on the above-mentioned embodiment, the present disclosure further provides a computer program. When the computer program is executed in a computer, the computer is enabled to execute the communication method provided in the embodiments shown in Figures 5 to 7, Figures 8A and 8B, Figures 9, Figures 10A and 10B, and Figure 11 from the viewpoint of the central node side or the kth subnode side.

[0272] Based on the foregoing embodiment, the present disclosure further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program. When the computer program is executed by a computer, the computer is enabled to execute the communication method provided in the embodiments shown in Figures 5 to 7, 8A and 8B, 9, 10A and 10B, and 11 from the perspective of the central node side or the kth subnode side. The storage medium may be any available medium that can be accessed by a computer. The following are used as examples, but are not limited to them. The computer-readable medium may include RAM, read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), CD-ROM or other optical disk storage device, magnetic storage medium or another magnetic storage device, or any other medium that can be used to carry or store expected program code in the form of instructions or data structures and can be accessed by a computer.

[0273] Based on the above-mentioned embodiment, the present disclosure provides a communication system including a central node and a plurality of subnodes. The central node and the plurality of subnodes can implement the communication methods provided in the embodiments shown in Figures 5 to 7, 8A and 8B, 9, 10A and 10B, and 11.

[0274] Based on the above embodiment, the present disclosure further provides a chip, which is configured to read a computer program stored in a memory, and implement the communication method provided in the embodiments shown in Figures 5 to 7, Figures 8A and 8B, Figures 9, Figures 10A and 10B, and Figure 11 from the perspective of the central node side or the kth subnode side.

[0275] Based on the aforementioned embodiment, the present disclosure provides a chip system. The chip system includes a processor configured to support a computer device in implementing the functions of the kth subnode or central node in the embodiments shown in Figures 5 to 7, 8A and 8B, 9, 10A and 10B, and 11. In a possible design, the chip system further includes a memory, the memory configured to store programs and data required for the computer device. The chip system may include a chip, or may include a chip and another discrete device.

[0276] All or part of the technical solutions provided in the present disclosure may be implemented by using software, hardware, firmware, or any combination thereof. When software is used to implement the embodiments, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the procedures or functions according to the present disclosure are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, a k-th subnode, a central node, or another programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from a computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website, computer, server, or data center to another website, computer, server, or data center in a wired (e.g., coaxial cable, fiber optic, or digital subscriber line (DSL)) or wireless (e.g., infrared, radio, or microwave) manner. A computer-readable storage medium can be any available medium accessible by a computer, or a data storage device that integrates one or more available media, such as a server or a data center. The available media may be magnetic media (e.g., floppy disks, hard disks, or magnetic tapes), optical media (e.g., digital video discs (DVDs)), semiconductor media, etc.

[0277] In this disclosure, where there is no logical contradiction, embodiments may be cross-referenced. For example, methods and / or terms in a method embodiment may be cross-referenced. For example, features and / or terms in an apparatus embodiment may be cross-referenced. For example, features and / or terms between an apparatus embodiment and a method embodiment may be cross-referenced.

[0278] It is obvious that those skilled in the art can make various modifications and variations to the present disclosure without departing from the scope of the present disclosure. The present disclosure is intended to cover these modifications and variations of the present disclosure, provided that they fall within the scope of protection defined by the following claims and their equivalent technologies. [Explanation of symbols]

[0279] 110 Access Network Devices 120 Terminal Devices 130 Terminal Devices 140 AI network elements 1200 Communication Equipment 1201 Processing Module 1202 Communication Module 1300 Communication Equipment 1310 Processor 1320 Memory 1330 Communication Interface 1340 Bus

Claims

1. 1. A communication method comprising: acquiring an i-th first piece of information, the i-th first piece of information indicating an i-th local model of each of a plurality of subnodes, where i is a positive integer; determining an i-th federated learning model of a k-th subnode based on the i-th first information and a type of the i-th federated learning model of the k-th subnode, where k is any positive integer from 1 to K, K is the number of the plurality of subnodes, and when i is greater than 1, the type of the i-th federated learning model of the k-th subnode is determined based on a type or performance of the (i-1)-th federated learning model of the k-th subnode; transmitting an i-th second information, wherein the i-th second information indicates an i-th federated learning model of each of the plurality of subnodes; A communication method comprising:

2. 2. The method of claim 1, wherein the i-th federated learning model of the k-th subnode is the same as or different from the i-th federated learning model of at least one subnode other than the k-th subnode among the plurality of subnodes.

3. obtaining a first test result from the k-th subnode, the first test result indicating a performance of the (i-1)-th associative learning model of the k-th subnode, where i is greater than 1; determining a type of the i-th federated learning model of the k-th subnode based on the first test result; 3. The method of claim 1 or 2, further comprising:

4. sending information indicating a first model to the k-th subnode, the first model being determined based on the (i-1)-th first information and a weight corresponding to the (i-1)-th local model of each of the subnodes, where i is greater than 1; obtaining a second test result from the kth subnode, the second test result being indicative of a performance of the first model, or the second test result being indicative of a performance of the first model and a performance of an (i-1)th associative learning model of the kth subnode; determining a type of the i-th federated learning model of the k-th subnode based on the second test result; 3. The method of claim 1 or 2, further comprising:

5. The step of determining an i-th federated learning model of a k-th subnode based on the i-th first information and a type of the i-th federated learning model of the k-th subnode includes: determining the i-th federated learning model of the k-th subnode based on the i-th first information and a weight corresponding to the i-th local model of each of the plurality of subnodes when the type of the i-th federated learning model of the k-th subnode corresponds to a first value; or determining the i-th federated learning model of the k-th subnode based on the i-th first information and a similarity between the i-th local model of the k-th subnode and the i-th local model of each of the plurality of subnodes when the type of the i-th federated learning model of the k-th subnode corresponds to a second value; 5. The method according to claim 1 , comprising:

6. the plurality of subnodes includes a first group of subnodes and a second group of subnodes, wherein the i-th federated learning models of all subnodes in the first group of subnodes are the same, and the i-th federated learning models of all subnodes in the second group of subnodes are different and distinct from the i-th federated learning models of all the subnodes in the first group of subnodes; The step of transmitting the i-th second information includes: sending first instruction information to the first group of subnodes in a broadcast or multicast manner, the first instruction information indicating the i-th federated learning model of each subnode in the first group of subnodes; sending second instruction information to a p-th subnode in the second group of subnodes in a unicast manner, the second instruction information indicating an i-th federated learning model of the p-th subnode, where p is any positive integer from 1 to P, and P is the number of subnodes included in the second group of subnodes; 6. The method according to any one of claims 1 to 5, comprising:

7. the i-th associative learning model is represented by a first portion of parameter information and a second portion of parameter information, and the first portion of parameter information of all the i-th associative learning models of the plurality of subnodes is the same; The plurality of subnodes include a first group of subnodes and a second group of subnodes, wherein the second portions of the parameter information of the i-th federated learning model of all subnodes in the first group of subnodes are the same, and the second portions of the parameter information of the i-th federated learning model of all subnodes in the second group of subnodes are different and are different from the second portions of the parameter information of the i-th federated learning model of all the subnodes in the first group of subnodes; The step of transmitting the i-th second information includes: transmitting the first portion of parameter information of the i-th federated learning model to the plurality of subnodes in a broadcast or multicast manner; transmitting the second portion of parameter information of the i-th federated learning model of each subnode in the first group of subnodes to the first group of subnodes in a broadcast or multicast manner; transmitting a second portion of parameter information of the ith federated learning model of the pth subnode to the pth subnode in the second group of subnodes in a unicast manner, where p is any positive integer from 1 to P, and P is the number of subnodes included in the second group of subnodes; 5. The method according to claim 1 , comprising:

8. transmitting information indicative of a second model to the plurality of subnodes, the information indicative of the second model being used by each of the plurality of subnodes to determine a first local model; 7. The method of claim 1, further comprising:

9. 8. The method of claim 1, wherein 1≦i≦I, I is a positive integer, the I-th federated learning model of the k-th subnode satisfies a model convergence condition, and when i is greater than 1, the i-th local model of the k-th subnode is determined based on the (i-1)-th federated learning model of the k-th subnode.

10. 1. A communication method comprising: transmitting information indicating the i-th local model of the k-th subnode, where k is any positive integer from 1 to K, K is the number of subnodes participating in the federated learning, and i is a positive integer; Obtaining information indicating an i-th federated learning model of the k-th subnode, wherein the i-th federated learning model of the k-th subnode is determined based on the i-th first information and a type of the i-th federated learning model of the k-th subnode; Including, A communication method, wherein the i-th first information includes information indicating the i-th local model of each of the multiple subnodes, and when i is greater than 1, the type of the i-th federated learning model of the k-th subnode is determined based on the type or performance of the (i-1)-th federated learning model of the k-th subnode.

11. 11. The method of claim 10, wherein the i-th federated learning model of the k-th subnode is the same as or different from the i-th federated learning model of at least one subnode other than the k-th subnode among the plurality of subnodes.

12. sending a first test result, the first test result indicating a performance of the (i-1)-th federated learning model of the k-th subnode, where i is greater than 1, and a type of the i-th federated learning model of the k-th subnode is determined based on the first test result; 12. The method of claim 10 or 11, further comprising:

13. acquiring information indicative of a first model, the first model being determined based on the (i-1)th first information and a weight corresponding to the (i-1)th local model of each of the plurality of subnodes, where i is greater than 1; sending a second test result, wherein the second test result indicates a performance of the first model, or the second test result indicates a performance of the first model and a performance of the (i-1)-th federated learning model of the k-th subnode, and a type of the i-th federated learning model of the k-th subnode is determined based on the second test result; 12. The method of claim 10 or 11, further comprising:

14. When a type of the i-th federated learning model of the k-th subnode corresponds to a first value, the i-th federated learning model of the k-th subnode is determined based on the i-th first information and a weight corresponding to the i-th local model of each of the plurality of subnodes; or When the type of the i-th federated learning model of the k-th subnode corresponds to a second value, the i-th federated learning model of the k-th subnode is determined based on the i-th first information and a similarity between the i-th local model of the k-th subnode and the i-th local model of each of the multiple subnodes.

14. The method according to any one of claims 10 to 13.

15. obtaining information indicative of a second model; determining a first local model of the k-th subnode based on information indicative of the second model; 15. The method of any one of claims 10 to 14, further comprising:

16. 16. The method of claim 10, wherein 1≦i≦I, I is a positive integer, the I-th federated learning model of the k-th subnode satisfies a model convergence condition, and when i is greater than 1, the i-th local model of the k-th subnode is determined based on the (i-1)-th federated learning model of the k-th subnode.

17. A communication device comprising a module or unit configured to carry out the method according to any one of claims 1 to 9 or to carry out the method according to any one of claims 10 to 16.

18. A communication device, comprising:

17. A communications device comprising a processor coupled to a memory, the processor configured to execute program instructions in the memory to perform a method according to any one of claims 1 to 9 or to perform a method according to any one of claims 10 to 16.

19. A computer-readable storage medium storing instructions which, when executed in a computer, enable the computer to perform the method according to any one of claims 1 to 9 or to perform the method according to any one of claims 10 to 16.

20. A computer program product comprising instructions, which when executed in a computer, cause the computer to perform the method according to any one of claims 1 to 9 or to be enabled to perform the method according to any one of claims 10 to 16.

21. A communication system comprising an apparatus configured to carry out the method according to any one of claims 1 to 9 and / or an apparatus configured to carry out the method according to any one of claims 10 to 16.

Images