Method for training machine learning model, terminal device, and network device

Abstract

The present application provides a method for training a machine learning model, a terminal device, and a network device, facilitating improving the training efficiency of a machine learning model. The method comprises: a first terminal device receives a first global model sent by a network device; and the first terminal device divides a local data sample into a first data sample and a second data sample, the first data sample being used by the first terminal device to perform, during a first training period, first training on the first global model, and the second data sample being used by the network device to perform, during the first training period, second training on the first global model.

Description

Machine learning model training methods, terminal devices, and network equipment Technical Field

[0001] The present application relates to the field of machine learning technology, and more specifically, to a training method, terminal device, and network device for a machine learning model. Background Art

[0002] With the development of communication technology, the development of intelligent services requires the support of high-performance machine learning models. Machine learning models can be trained in a distributed manner through a federated learning system, achieving high-performance machine learning models while protecting user privacy.

[0003] However, federated learning systems fail to fully utilize the computing power of network devices such as base stations to further improve the performance of machine learning models, and model uploading and aggregation incur significant latency overhead. Therefore, how to efficiently train machine learning models is an urgent problem that needs to be solved.

[0004] Summary of the Invention

[0005] This application provides a training method for a machine learning model, a terminal device, and a network device. The following describes various aspects of the embodiments of this application.

[0006] In a first aspect, a training method for a machine learning model is provided, including: a first terminal device receives a first global model sent by a network device; the first terminal device divides a local data sample into a first data sample and a second data sample; wherein the first data sample is used by the first terminal device to perform a first training on the first global model within a first training cycle, and the second data sample is used by the network device to perform a second training on the first global model within the first training cycle.

[0007] According to a second aspect, a method for training a machine learning model is provided, comprising: a network device sending a first global model to multiple terminal devices including a first terminal device; the network device receiving multiple second data samples sent by the multiple terminal devices, the multiple second data samples being determined based on local data samples of the multiple terminal devices, and the local data samples being divided into first data samples and second data samples; wherein the multiple first data samples of the multiple terminal devices are respectively used for the multiple terminal devices to perform first training on the first global model within a first training cycle, and the multiple second data samples are used for the network device to perform second training on the first global model within the first training cycle.

[0008] According to a third aspect, a terminal device is provided, which is a first terminal device for training a machine learning model, and the terminal device includes: a receiving unit, which can be used to receive a first global model sent by a network device; and a processing unit, which can be used to divide local data samples into a first data sample and a second data sample; wherein, the first data sample is used for the first terminal device to perform a first training on the first global model within a first training cycle, and the second data sample is used for the network device to perform a second training on the first global model within the first training cycle.

[0009] In a fourth aspect, a network device is provided, which is used to train a machine learning model, and the network device includes: a sending unit, which is used to send a first global model to multiple terminal devices including a first terminal device; a receiving unit, which is used for the network device to receive multiple second data samples sent by the multiple terminal devices, and the multiple second data samples are determined based on local data samples of the multiple terminal devices, and the local data samples are divided into first data samples and second data samples; wherein the multiple first data samples of the multiple terminal devices are respectively used for the multiple terminal devices to perform first training on the first global model within a first training cycle, and the multiple second data samples are used for the network device to perform second training on the first global model within the first training cycle.

[0010] In a fifth aspect, a communication device is provided, comprising a memory and a processor, wherein the memory is used to store a program, and the processor is used to call the program in the memory to execute the method described in the first aspect or the second aspect.

[0011] In a sixth aspect, a device is provided, comprising a processor for calling a program from a memory to execute the method as described in the first aspect or the second aspect.

[0012] In a seventh aspect, a chip is provided, comprising a processor for calling a program from a memory so that a device equipped with the chip executes the method described in the first aspect or the second aspect.

[0013] In an eighth aspect, a computer-readable storage medium is provided, on which a program is stored, wherein the program enables a computer to execute the method as described in the first aspect or the second aspect.

[0014] In a ninth aspect, a computer program product is provided, comprising a program, wherein the program enables a computer to execute the method as described in the first aspect or the second aspect.

[0015] In a tenth aspect, a computer program is provided, which enables a computer to execute the method as described in the first aspect or the second aspect.

[0016] After receiving the first global model, the terminal device of the embodiment of the present application divides the local data sample into a first data sample and a second data sample. The first data sample is used by the terminal device to perform a first training of the first global model, and the second data sample is used by the network device to perform a second training of the first global model. It can be seen that the training method of the embodiment of the present application combines the training performed separately by the terminal device and the network device, making full use of the powerful computing power of the network device and the effect of the terminal device in protecting data privacy during local training, thereby improving training efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] FIG1 is a wireless communication system used in an embodiment of the present application.

[0018] Figure 2 is a flow chart of a training method for a machine learning model provided in an embodiment of the present application.

[0019] FIG3 is a schematic diagram of a possible implementation of the execution sequence of the method shown in FIG2 .

[0020] FIG4 is a schematic diagram of another possible implementation of the execution sequence of the method shown in FIG2 .

[0021] FIG5 is a flow chart of the present federated learning method based on retransmitted over-the-air computation.

[0022] FIG6 is a flow chart of a retransmission over-the-air calculation mechanism.

[0023] FIG7 is a schematic diagram of a federated learning system based on retransmission over-the-air computing provided in an embodiment of the present application.

[0024] FIG8 is a schematic structural diagram of a terminal device provided in an embodiment of the present application.

[0025] FIG9 is a schematic structural diagram of a control device of the terminal device shown in FIG8 .

[0026] FIG10 is a schematic structural diagram of a network device provided in an embodiment of the present application.

[0027] FIG11 is a schematic structural diagram of a control device of the network device shown in FIG10 .

[0028] FIG12 is a schematic structural diagram of an electronic device provided in an embodiment of the present application.

[0029] FIG13 is a schematic block diagram of a communication device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0030] The following will describe the technical solutions in the embodiments of this application in conjunction with the drawings in the embodiments of this application. Obviously, the embodiments described are part of the embodiments of this application, not all of the embodiments. With respect to the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0031] The embodiments of the present application can be applied to various communication systems. For example, the embodiments of the present application can be applied to global system of mobile communication (GSM) system, code division multiple access (CDMA) system, wideband code division multiple access (WCDMA) system, general packet radio service (GPRS), long term evolution (LTE) system, advanced long term evolution (LTE-A) system, new radio (NR) system, NR system evolution system, LTE-based access to unlicensed spectrum (LTE-U) system on unlicensed spectrum, NR-based access to unlicensed spectrum (NR-U) system on unlicensed spectrum, non-terrestrial network (NTN) system, universal mobile telecommunication system (UMTS), wireless local area network (WLAN), wireless fidelity (WiFi), and fifth generation communication (5G) system. The embodiments of the present application may also be applied to other communication systems, such as future communication systems, such as sixth-generation (6G) mobile communication systems or satellite communication systems.

[0032] Traditional communication systems support a limited number of connections and are easy to implement. However, with the development of communication technology, communication systems can not only support traditional cellular communications, but also support one or more other types of communications. For example, a communication system can support one or more of the following communications: device to device (D2D) communication, machine to machine (M2M) communication, machine type communication (MTC), enhanced machine type communication (eMTC), vehicle to vehicle (V2V) communication, and vehicle to everything (V2X) communication, etc. The embodiments of the present application can also be applied to communication systems that support the above-mentioned communication methods.

[0033] The communication system in the embodiment of the present application can be applied to a carrier aggregation (CA) scenario, a dual connectivity (DC) scenario, and a standalone (SA) networking scenario.

[0034] The communication system in the embodiments of the present application can be applied to unlicensed spectrum. The unlicensed spectrum can also be considered a shared spectrum. Alternatively, the communication system in the embodiments of the present application can also be applied to licensed spectrum. The licensed spectrum can also be considered a dedicated spectrum.

[0035] The embodiments of the present application can be applied to NTN systems. As examples, the NTN systems may include 4G-based NTN systems, NR-based NTN systems, Internet of Things (IoT)-based NTN systems, and narrowband Internet of Things (NB-IoT)-based NTN systems.

[0036] A communication system may include one or more terminal devices. The terminal devices mentioned in the embodiments of the present application may also be referred to as user equipment (UE), access terminal, subscriber unit, subscriber station, mobile station, mobile station (MS), mobile terminal (MT), remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent or user device, etc.

[0037] In some embodiments, the terminal device may be a station (ST) in a WLAN. In some embodiments, the terminal device may be a cellular phone, a cordless phone, a session initiation protocol (SIP) phone, a wireless local loop (WLL) station, a personal digital assistant (PDA) device, a handheld device with wireless communication capabilities, a computing device or other processing device connected to a wireless modem, an in-vehicle device, a wearable device, a terminal device in a next-generation communication system (e.g., a NR system), or a terminal device in a future-evolved public land mobile network (PLMN) network.

[0038] In some embodiments, a terminal device may be a device that provides voice and / or data connectivity to a user. For example, the terminal device may be a handheld device, an in-vehicle device, etc. with wireless connection capabilities. As some specific examples, the terminal device may be a mobile phone, a tablet computer, a laptop computer, a PDA, a mobile internet device (MID), a wearable device, a virtual reality (VR) device, an augmented reality (AR) device, a wireless terminal in industrial control, a wireless terminal in self-driving, a wireless terminal in remote medical surgery, a wireless terminal in a smart grid, a wireless terminal in transportation safety, a wireless terminal in a smart city, a wireless terminal in a smart home, etc.

[0039] In some embodiments, the terminal device can be deployed on land. For example, the terminal device can be deployed indoors or outdoors. In some embodiments, the terminal device can be deployed on the water, such as on a ship. In some embodiments, the terminal device can be deployed in the air, such as on an airplane, a balloon, or a satellite.

[0040] In addition to the terminal device, the communication system may also include one or more network devices. The network device in the embodiment of the present application may be a device for communicating with the terminal device, and the network device may also be referred to as an access network device or a radio access network device. The network device may be, for example, a base station. The network device in the embodiment of the present application may refer to a radio access network (RAN) node (or device) that connects the terminal device to a wireless network. The base station can broadly cover various names as follows, or be replaced with the following names, such as: NodeB, evolved NodeB (eNB), next generation NodeB (gNB), relay station, access point, transmission point (transmitting and receiving point, TRP], transmitting point (transmitting point, TP]), master station MeNB, secondary station SeNB, multi-standard radio (MSR) node, home base station, network controller, access node, wireless node, access point (AP), transmission node, transceiver node, baseband unit (BBU), remote radio unit (RRU), active antenna unit (AAU), remote radio head (RRH), central unit (CU), distributed unit (distributed unit, DU), positioning node, etc. The base station can be a macro base station, a micro base station, a relay node, a donor node or the like, or a combination thereof. The base station can also refer to a communication module, a modem or a chip used to be set in the aforementioned equipment or device. The base station can also be a mobile switching center and a device that performs the base station function in D2D, V2X, and M2M communications, a network side device in a 6G network, a device that performs the base station function in a future communication system, etc. The base station can support networks with the same or different access technologies. The embodiments of the present application do not limit the specific technology and specific device form adopted by the network equipment.

[0041] Base stations can be fixed or mobile. For example, a helicopter or drone can be configured to act as a mobile base station, and one or more cells can move based on the location of the mobile base station. In other examples, a helicopter or drone can be configured to act as a device that communicates with another base station.

[0042] In some deployments, the network device in the embodiments of the present application may refer to a CU or a DU, or the network device may include a CU and a DU. The gNB may also include an AAU.

[0043] By way of example and not limitation, in embodiments of the present application, a network device may be mobile, for example, a mobile device. In some embodiments of the present application, the network device may be a satellite or balloon station. In some embodiments of the present application, the network device may also be a base station located on land, water, or the like.

[0044] In an embodiment of the present application, the network device can provide services for a cell, and the terminal device communicates with the network device through the transmission resources used by the cell (for example, frequency domain resources, or spectrum resources). The cell can be a cell corresponding to the network device (for example, a base station). The cell can belong to a macro base station or a base station corresponding to a small cell. The small cells here may include: metro cells, micro cells, pico cells, femto cells, etc. These small cells have the characteristics of small coverage and low transmission power, and are suitable for providing high-speed data transmission services.

[0045] For example, Figure 1 is a schematic diagram of the architecture of a communication system provided in an embodiment of the present application. As shown in Figure 1, the communication system 100 may include a network device 110, which may be a device that communicates with a terminal device 120 (or also referred to as a communication terminal or terminal). The network device 110 may provide communication coverage for a specific geographic area and may communicate with terminal devices located within the coverage area.

[0046] Figure 1 exemplarily shows a network device and two terminal devices. In some embodiments of the present application, the communication system 100 may include multiple network devices and each network device may include other numbers of terminal devices within its coverage area, which is not limited in the embodiments of the present application.

[0047] In an embodiment of the present application, the communication system shown in Figure 1 also includes other network entities such as a mobility management entity (MME) and an access and mobility management function (AMF), but this embodiment of the present application does not limit this.

[0048] It should be understood that in the embodiments of the present application, a device having a communication function in a network / system may be referred to as a communication device. Taking the communication system 100 shown in FIG1 as an example, the communication device may include a network device 110 and a terminal device 120 having a communication function. The network device 110 and the terminal device 120 may be the specific devices described above and will not be described in detail here. The communication device may also include other devices in the communication system 100, such as a network controller, a mobility management entity, and other network entities, which are not limited in the embodiments of the present application.

[0049] To facilitate the detailed explanation of the innovative features of the technical solution, we first introduce some relevant technical knowledge involved in the embodiments of this application. The following related technologies can be combined with the technical solutions of the embodiments of this application as optional solutions, and they all fall within the scope of protection of the embodiments of this application. The embodiments of this application include at least part of the following contents.

[0050] With the continuous advancement of communication technology, intelligent services are placing increasingly higher demands on the performance of machine learning models. For example, in future 6G edge networks, the development of edge intelligent services will require the support of high-performance machine learning models. Examples of edge intelligent services include autonomous driving, intelligent traffic management, and security monitoring.

[0051] However, data samples are stored on different terminal devices in edge networks. How to efficiently use distributed data samples to train machine learning models is a challenge that needs to be addressed. Currently, the main methods for training machine learning models in edge networks include federated learning and centralized learning.

[0052] Federated Learning System

[0053] Federated learning systems enable distributed training of machine learning models for edge intelligence services. Typically, a federated learning system in an edge network consists of a base station and multiple terminal devices, and the federated learning process is divided into several rounds. In a round, the federated learning process includes the following steps:

[0054] S1: The base station broadcasts the global model to all terminal devices. The global model is the machine learning model determined in the previous round.

[0055] S2: Multiple terminal devices use local data samples to train the global model to obtain multiple local models, and upload the local models to the base station through the wireless channel.

[0056] S3: The base station aggregates the local models uploaded by all terminal devices to obtain a new global model.

[0057] In the federated learning system, the base station and multiple terminal devices continuously repeat the above process from S1 to S3 until the global model meets the preset convergence conditions, at which point the federated learning is completed.

[0058] However, in federated learning systems, the base station's powerful computing power is not fully utilized for machine learning model training. As mentioned above, the base station in the federated learning system is only responsible for aggregating local models uploaded by terminal devices, and does not undertake the task of training machine learning models, which wastes the base station's powerful computing power.

[0059] Furthermore, in federated learning systems, the uploading and aggregation processes of local models are separated, resulting in low aggregation efficiency. In these federated learning systems, terminal devices typically upload local models using digital communication. For example, the terminal device first encodes the local model into a bitstream and then uploads it to the base station via a wireless channel. Therefore, the base station must decode the local models of all users and then aggregate them to form a global model. This transmission method separates the uploading and aggregation processes of local models, resulting in significant latency overhead and reduced aggregation efficiency.

[0060] Centralized learning system

[0061] The centralized learning system enables centralized training of machine learning models for edge intelligent services. The centralized learning system in the edge network consists of a base station and multiple terminal devices. The centralized learning process is divided into several rounds. In a round, the centralized learning process includes the following steps:

[0062] S1: All terminal devices upload local data samples to the base station.

[0063] S2: The base station uses all received data samples to train a global model, and then obtains a new global model. The global model received for training is the machine learning model determined in the previous round, and the new global model can be used in the next round.

[0064] In a centralized learning system, the base station and multiple terminal devices continuously repeat the above S1 and S2 processes until the global model meets the preset convergence conditions, at which point the centralized learning is completed.

[0065] However, in a centralized learning system, directly uploading locally stored data samples from a terminal device may expose data privacy. In this centralized learning system, local data samples contain private information related to the terminal device. Directly uploading local data samples to the base station will expose the terminal device's private information to the base station, bringing the risk of privacy leakage.

[0066] Air Computing

[0067] Over-the-air computing is a new type of non-orthogonal access method. Traditional orthogonal and non-orthogonal access methods focus solely on how to transmit information from the transmitter to the receiver. Over-the-air computing, however, leverages the superposition characteristics of wireless channels to enable multiple transmitters to transmit information on the same time-frequency resources. This allows the receiver to receive information from each transmitter that has been superimposed or otherwise processed. For example, multiple terminal devices can each transmit multiple transport blocks on the same time-frequency resources. The base station can then receive information from each terminal device, superimposed across multiple transport blocks.

[0068] Furthermore, over-the-air computing requires certain pre-processing and post-processing at both the transmitter and receiver. Through these processes, over-the-air computing can implement various signal computation methods during the communication process. This demonstrates the integration of communication and computation.

[0069] The preceding discussion discusses the challenges of federated learning and centralized learning for training machine learning models. While federated learning effectively protects user device privacy, it fails to fully utilize the powerful computing power of base stations. Furthermore, the uploading and aggregation of local models incurs significant latency, resulting in inefficient training.

[0070] Based on this, the present application embodiment proposes a method for training a machine learning model. Through this method, the powerful computing power of the base station can be fully utilized to undertake the training task of the machine learning model, and the global model performance can also be improved. For ease of understanding, the training method is described in detail below with reference to Figure 2.

[0071] Figure 2 illustrates the interaction between a first terminal device and a network device. The first terminal device is one of the aforementioned terminal devices that possesses certain computing and communication capabilities. In some embodiments, the first terminal device can train a machine learning model based on local data samples. In some embodiments, the first terminal device can send the data samples and the machine learning model to the network device. In some embodiments, the first terminal device can receive the machine learning model broadcasted by the network device.

[0072] In some embodiments, the first terminal device may be any terminal device in an edge network. The first terminal device may store a variety of data samples. The data samples stored by the first terminal device may include local data samples used to train a machine learning model.

[0073] The first terminal device is any terminal device among the multiple terminal devices participating in the machine learning model. The multiple terminal devices can jointly train the machine learning model with the network device. In some embodiments, the multiple terminal devices can provide data samples for the machine learning model.

[0074] The network device is any of the aforementioned communication devices that provide services to multiple terminal devices. In some embodiments, the network device is a communication device with powerful computing capabilities. The network device can be the base station that broadcasts the global model to multiple terminal devices based on federated learning, or the base station that trains the global model based on centralized learning, without limitation.

[0075] The network device can communicate with multiple terminal devices including the first terminal device. In some embodiments, the network device can receive data samples or local models sent by multiple terminal devices. In some embodiments, the network device can send a global model of a certain round to multiple terminal devices.

[0076] 2 , in step S210 , a first terminal device receives a first global model sent by a network device.

[0077] The first global model is a machine learning model that supports multiple intelligent services. In some embodiments, the first global model can be applied to the edge intelligent services described above.

[0078] The first global model can be a variety of machine learning models, which are not limited in the present embodiment. The first global model includes but is not limited to: a convolutional neural network model, a recursive neural network model, a generative adversarial network, etc.

[0079] The first global model can be a machine learning model currently being trained. Machine learning model training methods typically include multiple rounds. A round can refer to a training process or learning process, also known as a learning round or training cycle. For example, in the aforementioned federated learning, one training cycle is used to complete the processes S1 to S3. For another example, in the aforementioned centralized learning, one training cycle is used to complete the processes S1 and S2.

[0080] In some embodiments, during the training of a machine learning model, the first global model may be the global model applied to any training cycle. That is, the first global model may be the model trained within any training cycle. In some embodiments, the first global model may be the model determined in the previous training cycle before any training cycle. That is, the first global model may be the machine learning model determined in any other training cycle except the training cycle in which the last model converged.

[0081] In some embodiments, the first global model may be determined by the network device. For example, the network device may integrate information related to the first global model in the current training cycle to determine the first global model for the next training cycle. For example, the network device may determine the first global model based on the training results of the current training cycle.

[0082] The training of the first global model is performed jointly by the network device and multiple terminal devices including the first terminal device. As mentioned above, the network device can send the first global model to the multiple terminal devices via broadcasting, so that all terminal devices participating in the machine learning model training receive the first global model determined in the previous training cycle.

[0083] In some embodiments, the process of broadcasting the first global model from the network device to multiple terminal devices is included in the current training cycle. Exemplarily, this process can be used to determine the start time of the current training cycle. For example, the network device broadcasting the first global model can indicate the start of the current training cycle. For another example, after the current training cycle begins, the network device broadcasts the first global model, and the first terminal device obtains the first global model by executing step S210.

[0084] In some embodiments, the process of the network device broadcasting the first global model to multiple terminal devices is not part of the process in the current training cycle. For example, the current training cycle begins after the first terminal device obtains the first global model by executing step S210.

[0085] In step S220, the first terminal device divides the local data sample into a first data sample and a second data sample.

[0086] The local data samples may be various data samples used by the first terminal device to train the first global model. For example, the local data samples include but are not limited to pictures, audio, signals, etc., which are not limited in the present embodiment.

[0087] In some embodiments, some of the local data samples involve private information of the first terminal device. In some embodiments, some of the local data samples are information disclosed by the first terminal device. In some embodiments, the local data samples include information that the first terminal device does not wish to be disclosed.

[0088] The first terminal device can obtain local data samples in various ways. In some embodiments, the local data samples can be collected and determined by the first terminal device. In some embodiments, the local data samples can include data samples stored locally by the first terminal device and data samples collected by the first terminal device after receiving the first global model.

[0089] The local data sample is divided into the first data sample and the second data sample, which may mean that the first terminal device divides the local data sample based on a certain division scheme to determine the first data sample and the second data sample.

[0090] The first data sample and the second data sample may be data samples used to train the first global model. In some embodiments, the local data samples may first be filtered for data samples used to train the model before being divided. That is, the local data samples may include not only the first data sample and the second data sample used for training, but also other data samples. In some embodiments, all local data samples are divided to determine the first data sample and the second data sample. That is, the local data sample consists of the first data sample and the second data sample. In some embodiments, when the local data samples are divided, some data samples may be in both the first data sample and the second data sample.

[0091] The data samples in the first data sample and the second data sample may be completely different or partially the same, which is not limited here.

[0092] The purpose of dividing the local data sample into the first data sample and the second data sample is to train the first global model based on different training methods respectively, so as to effectively perform model training. Multiple terminal devices participating in the machine learning model can divide the local data sample into the first data sample and the second data sample.

[0093] The first data samples are used by the first terminal device to perform first training on the first global model during the first training cycle, and the second data samples are used by the network device to perform second training on the first global model during the first training cycle. In other words, the first training is performed locally on multiple terminal devices, while the second training is performed on the network device. This training method utilizes both multiple terminal devices and the network device for training, thereby leveraging the powerful computing power of the network device while preserving privacy to the greatest extent possible.

[0094] The plurality of terminal devices participating in the machine learning model all perform a first training on the first global model based on the first data sample. The first training can be used to obtain a plurality of local models for the plurality of terminal devices including the first terminal device.

[0095] In some embodiments, the first training is based on federated learning, and the second training is based on centralized learning. That is, the first training is distributed training conducted by multiple terminal devices, including the first terminal device, based on a federated learning system; the second training is centralized training conducted by a network device. In this application, the training method is a semi-federated learning system that combines federated learning and centralized learning. On the one hand, the terminal device uses a portion of local data samples to train a global model to obtain a local model, and then uploads the local model to the network device for aggregation, resulting in the federated learning aggregate model. On the other hand, the terminal uploads a portion of local data samples to the network device, which uses its powerful computing power to train a global model based on this data to obtain a centralized learning model. Finally, the network device combines the federated learning aggregate model and the centralized learning model to obtain a global model. For the sake of simplicity, the following text uses semi-federated learning to describe an embodiment in which the first training is based on federated learning and the second training is based on centralized learning.

[0096] In the above embodiment, the first data sample is used for the first training, that is, the first data sample is used for federated learning. The first data sample can also be referred to as a federated learning data sample. The second data sample is used for the second training, that is, the second data sample is used for centralized learning. The second data sample can also be referred to as a centralized learning data sample. In this scenario, the local data sample is divided into a federated learning data sample and a centralized learning data sample.

[0097] The partitioning scheme for local data samples can be determined in a variety of ways. In some embodiments, the network device can determine a general partitioning strategy, and the first terminal device can determine the final partitioning scheme based on the partitioning strategy and its own capabilities. For example, the local data sample partitioning scheme can be determined by the base station first and then broadcast to all terminal devices. In another example, the local data sample partitioning scheme can be determined independently by each terminal device.

[0098] Exemplarily, the division scheme for local data samples can be determined based on the type of data sample. As an example, local data samples can be divided based on whether the data sample involves privacy. For example, data samples involving the privacy of the first terminal device are all first data samples that are not directly uploaded, thereby preventing data samples sent to the network device from exposing the private information of the first terminal device. In this division scheme, the second data sample is related to the information disclosed by the first terminal device. Therefore, information not disclosed by the first terminal device cannot be attributed to the second data sample.

[0099] For example, the division scheme of the local data samples can be determined based on the number of data samples. As an example, the number of second data samples uploaded to the network device by multiple terminal devices can be equal. As an example, the local data samples can be divided into the same number ratio by multiple terminal devices.

[0100] Exemplarily, the division scheme of local data samples can be determined based on the capabilities of each terminal device. The capabilities of the terminal device may include computing capabilities for training the first global model, and communication capabilities for uploading data samples and local models. As an example, the number of samples of the first data sample and the number of samples of the second data sample can be determined based on the communication capability and / or computing capability of the first terminal device. For example, when the computing capability of the first terminal device is relatively weak, the number of samples of the second data sample can be greater than the number of samples of the first data sample. For another example, the number of samples of the second data sample can be positively correlated with the communication capability of the first terminal device.

[0101] The training methods for the first training and the second training can be gradient descent or other related methods, which are not limited herein. For example, the first terminal device can train the first global model using gradient descent based on the first data sample to obtain a local model. The network device can train the first global model using gradient descent based on multiple second data samples sent by multiple terminal devices to obtain a single model.

[0102] In some embodiments, the first training and the second training are performed in parallel within the first training cycle to improve the efficiency of model training. That is, within the first training cycle, multiple terminal devices and network devices can train the first global model in parallel. Taking federated learning and centralized learning as an example, within the training cycle, after receiving the first global model, multiple terminal devices can perform distributed training of the first global model using federated learning based on the first data sample; multiple terminal devices can also send multiple second data samples to the network device, so that the network device can perform centralized training of the first global model based on the multiple second data samples within the training cycle.

[0103] As can be seen from the foregoing, a training cycle refers to the time it takes to complete one training session when training a machine learning model. Taking the first global model as an example, completing one training session may refer to the process of training the first global model to obtain a second global model after determining the first global model. In the current training cycle, the second global model may be a new machine learning model determined based on the first global model. In the next training cycle, the second global model may be the first global model. By performing repeated training in multiple training cycles until the global model converges, the training (learning) process is terminated. The first training cycle may be any one of the multiple training cycles.

[0104] Because the first and second training are performed in parallel, the duration of the first training cycle needs to comprehensively consider the duration of the two training methods to determine the duration of each learning round of the machine learning model. For example, the period during which multiple terminal devices perform the first training can be called the first sub-period, and the period during which network devices perform the second training can be called the second sub-period. Therefore, the first sub-period can also be called the terminal device period, and the second sub-period can also be called the network device period.

[0105] To complete all training, the first training cycle can be the maximum of the first sub-cycle and the second sub-cycle. Since the first sub-cycle and the second sub-cycle are parallel time cycles, the duration of the first training cycle is the maximum of the first sub-cycle duration and the second sub-cycle duration.

[0106] During the first sub-period, multiple terminal devices perform first training based on the first data sample and transmit the multiple local models obtained from the training to the network device. Therefore, the first sub-period can be determined based on the multiple first durations during which the multiple terminal devices perform the first training and the one or more second durations during which the multiple local models are transmitted to the network device. Exemplarily, when there is one second duration, the duration of the first sub-period is the sum of the maximum value of the multiple first durations and the second duration. Exemplarily, when there are multiple second durations, the multiple first durations and the multiple second durations are added together to obtain multiple sums. The duration of the first sub-period is the maximum value of the multiple sums.

[0107] Exemplarily, the first duration is the duration for the terminal device to perform the first training, that is, the duration for the terminal device to train the first global model. The time for each terminal device to perform the first training is not necessarily the same, so the first training cycle can have multiple first durations of different lengths.

[0108] As an example, when the first terminal device uses the gradient descent method to train the first global model, the first duration can be determined based on the number of times the first terminal device uses the gradient descent method, the number of data samples used in each gradient descent, the number of central processing unit (CPU) cycles required to process a data sample, and the CPU frequency of the first terminal device.

[0109] For example, among K terminal devices (K is an integer greater than or equal to 1), the first duration of the kth terminal device (k is an integer from 1 to K) can be expressed as It can be determined as:

[0110] in, The number of times the gradient descent method is used for the k-th terminal device, is the number of data samples used by the k-th terminal device in each gradient descent, The number of CPU cycles required to process one data sample for the kth terminal device, is the CPU frequency of the kth terminal device.

[0111] Exemplarily, the second duration is the duration for the terminal device to transmit the local model obtained from the first training to the network device. As an example, when each terminal device transmits a local model separately, the second duration may not be the same, and therefore there may be multiple second durations. As an example, when multiple terminal devices transmit multiple local models based on an over-the-air computing mechanism, they are transmitted using the same time-frequency resources, and the second durations corresponding to the multiple terminal devices may be the same.

[0112] As an example, when the parameters of the local model are divided into multiple transmission blocks for uploading, the second duration can be determined based on the total number of parameters contained in the local model, the total number of parameters contained in a model transmission block, the probability of successfully transmitting a model transmission block, and the time length occupied by a transmission block.

[0113] For example, the upload period (second duration) of the local model can be expressed as It can be determined as:

[0114] in, is the total number of parameters contained in the local model, M is the total number of parameters contained in a transmission block, is the probability of successfully transmitting a model transmission block, T s is the length of time a model transmission block takes, is the ceiling function.

[0115] During the second sub-period, the network device must first receive multiple second data samples from multiple terminal devices and then perform the second training based on the multiple second data samples. Therefore, the second sub-period can be determined based on one or more third durations during which multiple terminal devices transmit multiple second data samples to the network device and a fourth duration during which the network device performs the second training. Exemplarily, when there is one third duration, the duration of the second sub-period is the sum of the third duration and the fourth duration. Exemplarily, when there are multiple third durations, the multiple third durations are added to the fourth duration to obtain multiple sums. The duration of the second sub-period is the maximum of the multiple sums.

[0116] Exemplarily, the third duration is the duration for multiple terminal devices to send the second data sample to the network device. As an example, when each terminal device sends the second data sample separately, the third duration may not be the same, and therefore there may be multiple third durations. As an example, when multiple terminal devices send multiple second data samples based on an over-the-air computing mechanism, they are sent using the same time-frequency resources, and the third durations corresponding to the multiple terminal devices may be the same.

[0117] As an example, when the parameters of the second data sample are divided into multiple data transmission blocks for uploading, the third duration can be determined based on the sample number of the second data sample, the total number of parameters contained in one data sample, the total number of parameters contained in one data transmission block, the probability of successfully transmitting one data transmission block, and the time length occupied by one data transmission block.

[0118] For example, the upload period (third duration) of the second data sample can be expressed as It can be determined as:

[0119] in, is the number of samples of the second data sample uploaded by the terminal device, Q D is the total number of parameters contained in a data sample, M is the total number of parameters contained in a transmission block, is the probability of successfully transmitting a data block, T s The length of time a data transmission block takes, is the ceiling function.

[0120] Exemplarily, the fourth duration is the duration for the network device to perform the second training. Since the network device trains the first global model after receiving multiple second data samples, the first training cycle only includes one fourth duration.

[0121] As an example, when the network device uses the gradient descent method to train the first global model, the fourth duration can be determined based on the number of times the network device uses the gradient descent method, the number of mixed data samples used in each gradient descent, the number of CPU cycles required to process a mixed data sample, and the CPU frequency of the network device.

[0122] For example, the fourth duration of the network device can be expressed as It can be determined as:

[0123] in, The number of times the gradient descent method is used for network devices, is the number of mixed data samples used by the network device in each gradient descent, The number of CPU cycles required for a network device to process one sample of mixed data, The CPU frequency of the network device.

[0124] For ease of understanding, this federated learning example is used for illustration. A learning round (training cycle) consists of a federated learning cycle for terminal devices, a local model upload cycle, a centralized learning data sample upload cycle, and a centralized learning cycle for network devices. The length of each cycle depends on the communication and computing capabilities of the terminal and network devices, as well as the data sample partitioning scheme.

[0125] In this example, the terminal device federated learning cycle and the local model upload cycle are in a serial relationship, forming a terminal device cycle; the centralized learning data sample upload cycle and the network device centralized learning cycle are in a serial relationship, forming a network device cycle. Furthermore, the terminal device cycle and the network device cycle are in a parallel relationship.

[0126] For ease of understanding, the following takes the training method based on air computing, hybrid federated learning and centralized learning as an example, and combines Figures 3 and 4 to exemplify the timing relationship in the method shown in Figure 2. The timing of Figures 3 and 4 is used to indicate the timing relationship between the base station and K terminal devices executing the method shown in Figure 2. The communication equipment in Figure 3 includes a base station 310 and a terminal device 301, a terminal device 302, ..., a terminal device 30k, ..., a terminal device 30K. The communication equipment in Figure 4 includes a base station 410 and a terminal device 401, a terminal device 402, ..., a terminal device 40k, ..., a terminal device 40K.

[0127] In Figures 3 and 4, T represents the duration of the first training cycle, T 1,k T represents the first duration of the first training of the kth terminal device based on federated learning, T2 represents the second duration of uploading the local model based on over-the-air computing and performing model aggregation, T3 represents the third duration of uploading the second data sample based on over-the-air computing, and T4 represents the fourth duration of the first training of the network device. For the sake of convenience, it is assumed that the first duration of the kth terminal device is the largest among the K terminal devices. Therefore, the duration of the first sub-period is T 1,k The duration of the second sub-period is the sum of T3 and T4 added together with T2.

[0128] 3 , the duration of the first sub-period is greater than that of the second sub-period, and the first training period is the first sub-period, that is, the first training period T=max{T 1,k}+T2.

[0129] 4 , the duration of the first sub-period is shorter than that of the second sub-period, and the first training period is the second sub-period, that is, the first training period T=T3+T4.

[0130] As can be seen from Figures 2 to 4, in an embodiment of the present application, the terminal device divides the local data sample into a first data sample and a second data sample, and uses them for the terminal device and the network device to train the first global model in the current training cycle. This training method can utilize the powerful computing power of the network device. When the second data sample does not involve the privacy information of the terminal device, this training method can also protect the privacy of the terminal device while improving training efficiency.

[0131] As mentioned above, during the first training cycle, the first global model is trained to obtain the second global model. For example, multiple models obtained by training the first global model separately on the terminal device and the network device can be aggregated to the network device, and the network device determines the second global model.

[0132] In some embodiments, during a first training cycle, multiple terminal devices obtain multiple local models through a first training of the first global model, and the network device also obtains a model through a second training of the first global model. The multiple local models and the model obtained through the second training are used by the network device to determine the second global model.

[0133] The second global model may be determined by various information, including one or more of the following: the first global model; the first aggregate model; the second trained model; the first weight corresponding to the first aggregate model; and the second weight corresponding to the second trained model.

[0134] Exemplarily, the first aggregate model can be determined based on multiple local models. For example, some or all of the multiple local models can refer to federated learning to determine the first aggregate model. In another example, the multiple local models can send the federated learning aggregate model to the network device based on an over-the-air computing mechanism.

[0135] Exemplarily, when the second training is based on centralized learning, the model obtained by the second training is a centralized learning model.

[0136] Exemplarily, the first weight corresponding to the first aggregate model is used to determine the share of the first aggregate model in the second global model. The first weight is a non-negative real number less than or equal to 1. When the first training is federated learning-based training, the first aggregate model is a federated learning aggregate model, and the first weight may also be referred to as a hybrid weight of the federated learning aggregate model.

[0137] As an example, the first weight can be determined based on the first data samples obtained by dividing the local data samples of the multiple terminal devices and the mixed data samples used for the second training. For example, the first weight can be determined based on the number of first data samples and the number of mixed data samples obtained by dividing the local data samples of the multiple terminal devices.

[0138] As an example, multiple terminal devices divide local data samples to obtain multiple second data samples, which are used to determine mixed data samples for the second training. When multiple terminal devices upload multiple second data samples to the network device based on the over-the-air calculation mechanism, the network device will receive the mixed multiple data samples. However, the mixed data samples used for the second training of the first global model may only include some of these data samples, thereby improving training efficiency while reducing computational overhead.

[0139] For example, the network device may determine a mixed data sample for the second training from the plurality of second data samples based on a forgetting mechanism. That is, the mixed data sample may include a portion of the plurality of second data samples, the portion of the data sample being determined based on the forgetting mechanism.

[0140] Exemplarily, the second weight corresponding to the model obtained from the second training is used to determine the share of the model in the second global model. The second weight is a non-negative real number less than or equal to 1. When the model obtained from the second training is a centralized learning model, the second weight can also be referred to as a hybrid weight of the centralized learning model.

[0141] As an example, the second weight can be determined based on the first data samples obtained by dividing the local data samples by multiple terminal devices and the mixed data samples used for the second training. For example, the second weight can be determined based on the number of first data samples and the number of mixed data samples obtained by dividing the local data samples by multiple terminal devices.

[0142] Exemplarily, the sum of the first weight and the second weight is 1, and the first weight and the second weight are non-negative real numbers.

[0143] In some embodiments, the second global model can be determined based on the first aggregate model, the first weight, the second trained model, and the second weight. For example, in a semi-federated learning system, the network device can combine the federated learning aggregate model and the centralized learning model according to the first weight and the second weight, respectively, to obtain the second global model.

[0144] In some embodiments, the second global model can be determined based on the first global model, the first aggregate model, the first weight, the second trained model, and the second weight.

[0145] Exemplarily, the network device may add the first global model, the product of the first aggregate model multiplied by the first weight, and the product of the second trained model multiplied by the second weight, thereby determining the second global model.

[0146] For example, the second global model uses w t+1 When expressed, w t+1 It can be determined as:

[0147] Among them, w t is the first global model, is the first aggregation model, is the model obtained by the second training, is the first weight, is the second weight.

[0148] Optionally, in the semi-federated learning method, the first weight and the second weight Can be determined as:

[0149] in, is the number of k-th federated learning data samples, is the number of base station mixed data samples.

[0150] After determining the second global model, the network device can determine whether convergence has been achieved. In other words, the second global model is used by the network device to determine whether the trained machine learning model has reached convergence. In some embodiments, the network device can use specific rules to determine whether the second global model has reached convergence.

[0151] As an embodiment, the second global model is w t+1 The convergence judgment rule of can be determined as: ‖w t+1 -w t ‖≤ε;

[0152] Among them, ‖.‖ represents the calculated vector binorm, and ε represents the preset convergence accuracy.

[0153] As another embodiment, the second global model is represented by w t+1 The convergence judgment rule can also be determined as: |F(w t+1 )-F(w t )|≤ε;

[0154] Among them, F(w) represents the loss function calculated based on a global model w, which can be used to measure the training effect of the global model.

[0155] It should be understood that the above embodiment is merely intended to illustrate how to determine whether the global model has converged. In addition to the convergence judgment rules mentioned above, any rule that can determine whether the second global model has achieved objective convergence can be applied to the embodiments of the present application. This embodiment of the present application does not limit this.

[0156] After the network device determines that the global model meets the convergence conditions, it can broadcast the final global model and the termination training instruction to all terminal devices. All terminal devices can stop collecting data samples and training local models.

[0157] In some embodiments, the network device and all terminal devices may release time-frequency resources used for uploading the local model and the second data sample.

[0158] In some embodiments, the network device may delete received mixed data samples.

[0159] As can be seen from the above, the semi-federated learning system based on retransmission air computing provided by the embodiment of the present application can fully utilize the powerful computing power of the base station to undertake the training task of the machine learning model and improve the performance of the global model. At the same time, by applying the local data sample upload and local model upload based on the retransmission air computing mechanism provided by the embodiment of the present application on orthogonal time-frequency resources, it is possible to combine the upload of the local model with the aggregation process, improve the aggregation efficiency, and protect the privacy of the uploaded data samples.

[0160] In the previous section, we mentioned over-the-air computing in the context of uploading the second data sample and the local model. This mechanism allows for the integration of local model upload and aggregation, improving aggregation efficiency while protecting the privacy of the uploaded data sample. To perform over-the-air computing, multiple devices must use the same resources for upload.

[0161] In some embodiments, the input signals for over-the-air computation are signals sent by various devices that require over-the-air computation. Multiple terminal devices can directly transmit signals based on the over-the-air computation mechanism, which helps improve transmission efficiency. For example, when uploading a local model, the communication device does not need to perform multiple encoding and decoding operations, which can reduce latency overhead.

[0162] Exemplarily, for the uploading process of the second data sample based on the over-the-air calculation mechanism, the terminal device may process the second data sample segment into an over-the-air calculation input signal.

[0163] For example, for a local model upload process based on an over-the-air computing mechanism, the terminal device may process the local model fragments into over-the-air computing input signals.

[0164] In some embodiments, the output signal of the over-the-air calculation is a signal received by the network device and calculated over the air.

[0165] Exemplarily, in semi-federated learning, for the uploading process of the second data sample based on the over-the-air calculation mechanism, the output signal of the over-the-air calculation is a mixed data sample segment.

[0166] For example, in semi-federated learning, for the upload process of a local model based on an over-the-air computing mechanism, the output signal of the over-the-air computing is a federated learning aggregate model fragment.

[0167] In some embodiments, all terminal devices can be scheduled to upload the second data sample using the same time-frequency resources, and over-the-air computing techniques can be used to achieve mixing of local data samples during transmission. The model obtained by the network device directly receiving and using the mixed data samples to train the first global model can be called a centralized learning model. Because the network device only receives the mixed data samples and not the uploaded local original data samples, the privacy of the uploaded data samples can be protected.

[0168] For example, the second data sample can be divided into multiple data transmission blocks based on the upload period (third duration). Each data transmission block carries a mixed data sample segment and occupies a fixed time length (for example, T s ). When a data transmission block is transmitted for the first time, all terminal devices participating in the machine learning model training send the data transmission block to the base station simultaneously on the same time-frequency resources.

[0169] As an embodiment, the second data sample may include one or more sample segments corresponding to one or more data transmission blocks. The one or more sample segments may include a first sample segment. The data transmission block corresponding to the first sample segment is the first data transmission block. During the first training cycle, multiple terminal devices including the first terminal device may send multiple sample segments corresponding to the first data transmission block to the network device on the first resource. In other words, the multiple terminal devices simultaneously send multiple sample segments using the same resource.

[0170] As an example, a sample segment of each terminal device may correspond to a data transmission block. Multiple sample segments of multiple terminal devices may be transmitted via one data transmission block. The first sample segment of the first terminal device is one of the multiple sample segments.

[0171] As an example, the first resource may be a time-frequency resource used to carry a data transmission block, which is not limited here.

[0172] As an example, the first sample segment and other sample segments are used to input a first in-flight calculation. The first in-flight calculation may be used to perform calculations on multiple sample segments on the first resource.

[0173] In some embodiments, all terminal devices can be scheduled to upload local models using the same time-frequency resources. Multiple terminal devices can aggregate local models during transmission using over-the-air computing technology. When the first training is federated learning, network devices can directly receive the aggregated federated learning model, thereby improving aggregation efficiency.

[0174] For example, the local model can be divided into multiple model transmission blocks based on the upload period (second duration). Each model transmission block can carry a federated learning aggregate model fragment and occupy a fixed time length (e.g., T s ). When a model transmission block is transmitted for the first time, all terminal devices participating in the machine learning model training send the model transmission block to the network device simultaneously on the same time-frequency resources.

[0175] As an embodiment, the first local model may include one or more model fragments corresponding to one or more model transmission blocks. These model fragments may also be referred to as model fragments. The one or more model fragments may include a first model fragment. The model transmission block corresponding to the first model fragment is the first model transmission block. During the first training cycle, multiple terminal devices including the first terminal device may send multiple model fragments corresponding to the first model transmission block to the network device via the second resource. In other words, multiple terminal devices simultaneously send multiple model fragments via the same resource.

[0176] As an example, a model segment of each terminal device may correspond to a model transmission block. Multiple model segments of multiple terminal devices may be transmitted via one model transmission block. A first model segment of a first terminal device is a model segment among the multiple model segments.

[0177] As an example, the second resource may be a time-frequency resource used to carry the model transmission block. It should be understood that the second resource is orthogonal to the first resource in terms of time and frequency. That is, the time-frequency resources used by the terminal device during the local model upload period are orthogonal to the time-frequency resources used by the terminal device during the second data sample upload period.

[0178] The above describes a method for multiple terminal devices to upload local models and second data samples based on an over-the-air computing mechanism. However, the wireless channel changes rapidly, and it is difficult for the communication device to adjust the transceiver configuration scheme in real time, resulting in transmission errors. For example, during the upload of the local model or the second data sample, the transmitter configuration of the terminal device (including transmit power, transmit beamforming, etc.) and the receiver configuration of the network device (including receive beamforming, etc.) need to remain unchanged throughout the upload time. However, the actual wireless channel state changes rapidly and multiple times during the upload time, and it is difficult for the communication device to adjust the transceiver configuration in real time according to the wireless channel state. In the case of a mismatch between the wireless channel state and the transceiver configuration, transmission errors will occur, affecting the quality of the global model.

[0179] In order to solve this problem, an embodiment of the present application proposes a retransmission air calculation mechanism, which eliminates the need for real-time configuration of the transceiver during the upload time by retransmitting air calculation results with large errors, thereby coping with rapidly changing wireless channels. For example, in a semi-federated learning method, multiple terminal devices require federated learning aggregate model fragments and mixed data sample fragments. Through this retransmission air calculation mechanism, for federated learning aggregate model fragments with excessive errors, the network device can initiate a retransmission command to all terminal devices, and the retransmission continues until the error of the fragment is within a tolerable range. Correspondingly, for mixed data sample fragments with excessive errors, the network device can initiate a retransmission command to all terminal devices, and the retransmission continues until the error of the fragment is within a tolerable range.

[0180] Exemplarily, the above-mentioned mechanism based on over-the-air calculation supports the first terminal device to retransmit the sample segments and / or model segments that have failed to be transmitted.

[0181] Exemplarily, based on the retransmission air calculation mechanism, the network device may receive a data transmission block or a model transmission block with a smaller error, and initiate retransmission of the data transmission block or the model transmission block with a larger error.

[0182] As an example, the third air calculation can be any of the multiple air calculations in uploading the data sample and the local model. For example, the third air calculation can be the first air calculation, the second air calculation, or any other air calculation.

[0183] For example, upon receiving data transmission blocks from multiple terminal devices, a network device may evaluate the aggregate quality of the data transmission blocks. For data transmission blocks that pass the aggregate quality assessment, the network device receives them and extracts the mixed data sample fragments carried by the transmission blocks. For data transmission blocks that fail the aggregate quality assessment, the network device discards the data transmission blocks and broadcasts a retransmission instruction to all terminal devices. After receiving the retransmission instruction, all terminal devices simultaneously retransmit the data transmission blocks to the network device using the same time-frequency resources until the network device passes the aggregate quality assessment for the blocks.

[0184] Exemplarily, upon receiving a model transmission block sent by multiple terminal devices, the network device may evaluate the aggregation quality of the model transmission block. For model transmission blocks that pass the aggregation quality assessment, the network device receives them and extracts the aggregated model fragments carried by the transmission block. For model transmission blocks that fail the aggregation quality assessment, the network device discards the model transmission block and broadcasts a retransmission instruction to all terminal devices. After receiving the retransmission instruction, all terminal devices retransmit the model transmission block to the network device simultaneously on the same time-frequency resources until the network device passes the aggregation quality assessment for the block.

[0185] For example, the retransmission over-the-air calculation mechanism may instruct multiple terminal devices to retransmit. During the first training cycle, if the output signal of the third over-the-air calculation does not meet the first condition, the first terminal device may receive a retransmission instruction sent by the network device. The retransmission instruction may instruct the multiple terminal devices to retransmit sample segments or model segments involved in the third over-the-air calculation.

[0186] The first condition may be a preset standard for the quality of the over-the-air calculation output signal. That is, after receiving the over-the-air calculation output signal, the network device evaluates whether the quality of the output signal meets the preset standard.

[0187] In some embodiments, the quality of the output signal calculated over the air can be represented by the mean-square error (MSE) of the signal. The preset evaluation standard can be an evaluation standard based on a threshold.

[0188] For example, the first condition can be expressed as:

[0189] MSE≤ò

[0190] Where, ò is the preset threshold, and MSE can be further determined as:

[0191] in, is the normalized receive beamforming vector of the network device, satisfying ‖b t ‖=1,ζ tis the normalization factor of the base station, satisfying ζ t ≥0,h tk is the channel coefficient vector from the kth terminal device to the network device, is the receiver noise intensity of the network device.

[0192] It should be understood that the above embodiment only provides an evaluation standard for an over-the-air computing output signal. Other evaluation standards that can objectively evaluate the quality of an over-the-air computing output signal can also be applied to the embodiment of the present application. The embodiment of the present application does not limit this.

[0193] The network device can send retransmission instructions to all terminal devices by broadcasting, or send retransmission instructions to multiple terminal devices participating in machine learning model training by signaling.

[0194] The retransmission instruction sent by the network device to all terminal devices may be indicated using one or more bits. In some embodiments, the retransmission instruction sent by the network device to all terminal devices via broadcast may be implemented using only a single "1" bit signal. When a terminal device receives a "1" bit, it retransmits the current over-the-air learning input signal; when it receives a "0" bit, it continues to transmit the next over-the-air learning input signal.

[0195] The above describes the method for transmitting models and data samples based on over-the-air computation and retransmission over-the-air computation in the embodiments of the present application. In the semi-federated learning method, the training method provided by the embodiments of the present application is a semi-federated learning system based on retransmission over-the-air computation.

[0196] For ease of understanding, the following uses a base station as an example network device, and illustrates semi-federated learning based on retransmission over-the-air computation with reference to Figures 5 and 6. Figure 5 is a schematic diagram of a semi-federated learning process based on retransmission over-the-air computation provided in an embodiment of the present application. Figure 6 is a schematic diagram of a retransmission over-the-air computation process provided in an embodiment of the present application.

[0197] Referring to FIG. 5 , in step S510 , when entering a preset learning round, the base station broadcasts the global model of the previous round, and the terminal device divides the collected data samples into federated learning data samples and centralized learning data samples.

[0198] In step S510, the previous round is the previous training cycle, and the global model of the previous round is the first global model. The data samples collected by the terminal device are local data samples. The federated learning data samples are the first data samples, and the centralized learning data samples are the second data samples.

[0199] Steps S522 and S532 are the first training and local model upload based on federated learning, while steps S524 and S534 are the second data sample upload and second training based on centralized learning. As shown in Figure 5, steps S522 and S532 are serial, while steps S524 and S534 are serial; steps S522 and S524 are parallel, while steps S532 and S534 are parallel.

[0200] In step S522, the terminal device uses the federated learning data sample to train the global model of the previous round to obtain a local model. This training process occurs within the terminal device federated learning cycle, that is, within the first duration mentioned above.

[0201] In step S524, the terminal device uploads the centralized learning data samples to the base station using the same time-frequency resources based on the retransmission over-the-air calculation mechanism. The base station receives and accumulates the mixed centralized learning data samples. The centralized learning data sample upload process occurs within the centralized learning data sample upload period, that is, within the third duration described above.

[0202] In step S532, the terminal device uploads the local model to the base station using the same time-frequency resources based on the retransmission air calculation mechanism, and the base station receives the federated learning aggregate model. The local model upload process occurs within the local model upload period, that is, within the second duration mentioned above.

[0203] In step S534, the base station uses the mixed centralized learning data samples to train the global model of the previous round to obtain a centralized learning model. This training process occurs within the base station centralized learning cycle, that is, within the fourth time period mentioned above.

[0204] In step S540, the base station obtains a global model by weightedly mixing the federated learning aggregation model and the centralized learning model. The global model obtained by the base station is the second global model. After obtaining the federated learning aggregation model and the centralized learning model, the base station may multiply the federated learning aggregation model by a specific mixing weight (the first weight) and the centralized learning model by another specific mixing weight (the second weight), and then add the two results together.

[0205] In step S550, the base station determines whether convergence is achieved. If convergence is achieved, step S560 is executed; if not, step S510 is executed.

[0206] At step S560, semi-federated learning based on retransmitted over-the-air computation ends. The base station and terminal device can release the time-frequency resources for uploading local models and centralized data samples based on retransmitted over-the-air computation. The base station can then delete the received hybrid data samples.

[0207] In the semi-federated learning system based on retransmission over-the-air computing shown in Figure 5, the centralized learning data sample upload process based on the retransmission over-the-air computing mechanism and the local model upload process based on the retransmission over-the-air computing mechanism can use two mutually orthogonal time-frequency resources.

[0208] Referring to Figure 6, in step S610, all terminal devices simultaneously transmit over-the-air computation input signals on the same time-frequency resources. For centralized learning data sample uploads using a retransmitted over-the-air computation mechanism, the over-the-air computation input signals are the centralized learning data sample segments of the terminal devices; for local model uploads using a retransmitted over-the-air computation mechanism, the over-the-air computation input signals are the local model segments of the terminal devices.

[0209] In step S620, the base station receives the over-the-air computation output signal and evaluates its quality. The over-the-air computation output signal is the signal received by the base station after undergoing over-the-air computation. For centralized learning data sample uploads using a retransmitted over-the-air computation mechanism, the over-the-air computation output signal is a mixed data sample segment. For local model uploads using a retransmitted over-the-air computation mechanism, the over-the-air computation output signal is a federated learning aggregate model segment.

[0210] In step S630, it is determined whether the quality of the output signal meets the requirements. If so, step S650 is executed; if not, step S640 is executed.

[0211] In step S640, the base station sends a retransmission instruction to all terminal devices. After completing the quality evaluation of the output signal calculated over the air, the base station discards the output signal that fails the evaluation and sends a retransmission instruction to all terminal devices, and then continues to execute step S610.

[0212] In step S650, it is determined whether the over-the-air computing task is complete. The base station counts the number of over-the-air computing output signals that pass the evaluation. If the number is less than the preset total number of over-the-air computing task output signals, the task is not completed. If the task is not completed, the process continues with step S610; if the task is completed, the process proceeds to step S660.

[0213] In step S660, the retransmission of the over-the-air computation process ends. If the number of over-the-air computation output signals that pass the evaluation by the base station equals the preset total number of output signals for the over-the-air computation task, the base station sends a retransmission of the over-the-air computation completion instruction to all terminal devices. All terminal devices stop sending over-the-air computation input signals, and the base station and terminal devices release the time-frequency resources used for retransmission of the over-the-air computation.

[0214] As shown in Figure 6, when applying the retransmission over-the-air calculation mechanism on a rapidly changing wireless channel, the base station does not need to optimize or modify the transceiver configurations of the terminal device and base station in real time to adapt to the rapidly changing wireless channel conditions. Instead, the terminal device and base station maintain a fixed transceiver configuration, and the base station only needs to determine whether the quality of the over-the-air calculation output signal meets the preset evaluation criteria. The base station receives any over-the-air calculation output signal that passes the evaluation. For any signal that fails, the base station discards it and initiates a retransmission, continuing until the signal passes the preset evaluation criteria.

[0215] The following describes the embodiment of the present application in more detail with reference to the specific example FIG7. It should be noted that the examples of FIG2 to FIG6 are merely intended to help those skilled in the art understand the embodiment of the present application, and are not intended to limit the embodiment of the present application to the specific numerical values ​​or specific scenarios illustrated. It is obvious that those skilled in the art can make various equivalent modifications or changes based on the examples of FIG2 to FIG6, and such modifications or changes also fall within the scope of the embodiment of the present application.

[0216] Figure 7 is a schematic diagram of a semi-federated learning system based on a retransmission over-the-air computation mechanism. As shown in Figure 7 , the semi-federated learning system based on a retransmission over-the-air computation mechanism consists of a base station (base station 710) and K terminal devices. The K terminal devices are terminal device 701, terminal device 702, ..., terminal device 70k, ..., terminal device 70K.

[0217] The learning process shown in FIG7 is also divided into several training cycles. When entering the current t-th learning round (the first training cycle), K terminal devices receive the global model w broadcast by the base station 710. t (First global model).

[0218] In step S71, the terminal device collects local data samples 71. The local data samples 71 collected by the K terminal devices are D t,1 、D t,2 ,…,D t,k ,…,D t,K .

[0219] In step S72, the K terminal devices divide the local data samples 71 into federated learning data samples 73 (first data samples) and centralized learning data samples 72 (second data samples). t,k Divided into federated learning data samples and centralized learning data samples

[0220] In step S73, during the federated learning cycle of the terminal device, the K terminal devices use the federated learning data sample 73 to train the global model w broadcast by the base station.t , and obtain the local model 74. For example, the k-th terminal device uses the federated learning data sample The global model w broadcast by training base station 710 t Get the local model The local models obtained by K terminal devices are

[0221] In step S74, during the local model upload period, the K terminal devices upload their local models 74 to the base station 710 on the same time-frequency resource (time-frequency resource 2) based on the retransmission air calculation mechanism. Then, the base station 710 receives the federated learning aggregate model fragments with smaller errors and initiates retransmission of the federated learning aggregate model fragments with larger errors. Finally, the base station 710 obtains the federated learning aggregate model 75. The federated learning aggregate model 75 can be expressed as As shown in Figure 7, time-frequency resource 2 can be used to send local model aggregation based on retransmission over-the-air calculations. Time-frequency resource 2 can carry the model fragments of the initial transmission and / or the retransmitted model fragments.

[0222] In step S75, during the centralized learning data sample uploading period, K terminal devices upload their centralized learning data samples 72 to the base station 710 in the same time-frequency resources based on the retransmission air calculation mechanism. The centralized learning data samples uploaded by the K terminal devices are Base station 710 then receives the mixed data sample segments with smaller errors and initiates retransmission of the mixed data sample segments with larger errors. As shown in Figure 7, time-frequency resource 1 can be used to transmit the mixed data samples based on the retransmission over-the-air calculation. Time-frequency resource 1 can carry the initially transmitted sample segments and / or the retransmitted sample segments.

[0223] In step S76, during the centralized learning cycle of the base station 710, the base station 710 trains the global model w using the mixed data samples 76. t Get a centralized learning model Among them, the mixed data sample 76 can be expressed as

[0224] In step S77, at the end of the current t-th learning round, the base station 710 uses the federated learning aggregation model and centralized learning models By weight and weights Mix and get the global model w for the next learning round t+1 (Second global model). As mentioned above, and To satisfy Two non-negative real numbers.

[0225] In step S78, the base station 710 broadcasts the global model w to all terminal devices. t+1 .

[0226] The present application also provides a learning system for a machine learning model. The learning system includes a network device and multiple terminal devices, wherein any of the multiple terminal devices executes the method described above for the terminal device to execute, and the network device executes the method described above for the network device to execute.

[0227] The method embodiment of the present application is described in detail above in conjunction with Figures 1 to 7. The device embodiment of the present application is described in detail below in conjunction with Figures 8 to 13. It should be understood that the description of the device embodiment corresponds to the description of the method embodiment. Therefore, for portions not described in detail, reference can be made to the above method embodiment.

[0228] FIG8 is a schematic block diagram of a terminal device according to an embodiment of the present application. The apparatus 800 may be a first terminal device for training a machine learning model. The first terminal device may be any of the terminal devices described above. The terminal device 800 shown in FIG8 includes a receiving unit 810 and a processing unit 820.

[0229] The receiving unit 810 may be configured to receive a first global model sent by a network device.

[0230] The processing unit 820 can be used to divide the local data sample into a first data sample and a second data sample; wherein the first data sample is used for the first terminal device to perform the first training of the first global model within the first training cycle, and the second data sample is used for the network device to perform the second training of the first global model within the first training cycle.

[0231] Optionally, the sample quantity of the first data sample and the sample quantity of the second data sample are determined according to the communication capability and / or computing capability of the first terminal device.

[0232] Optionally, the second data sample is related to information disclosed by the first terminal device.

[0233] Optionally, the second data sample includes a first sample fragment, and the terminal device 800 also includes a first sending unit, which can be used to send the first sample fragment to the network device on the first resource based on the air calculation mechanism during the first training cycle; wherein the first resource is also used for other terminal devices other than the first terminal device to send other sample fragments to the network device, and the other sample fragments and the first sample fragment are used to input the first air calculation.

[0234] Optionally, the first training is used by the first terminal device to determine a first local model, and the first local model includes a first model fragment. The terminal device 800 also includes a second sending unit, which can be used to send the first model fragment to the network device on the second resource based on the air calculation mechanism during the first training cycle; wherein the second resource is also used by other terminal devices other than the first terminal device to send other model fragments to the network device, and the other model fragments and the first model fragment are used to input the second air calculation.

[0235] Optionally, the over-the-air calculation mechanism supports the first terminal device to retransmit the sample segments and / or model segments that have failed to be transmitted.

[0236] Optionally, the receiving unit 810 is also used to receive a retransmission indication sent by the network device if the output signal of the third air calculation does not meet the first condition during the first training cycle, and the retransmission indication is used to instruct multiple terminal devices to retransmit sample fragments or model fragments participating in the third air calculation.

[0237] Optionally, the first training is used for multiple terminal devices including the first terminal device to determine multiple local models, the multiple local models and the model obtained by the second training are used by the network device to determine the second global model, and the second global model is used by the network device to determine whether the trained machine learning model has reached convergence.

[0238] Optionally, during the first training cycle, multiple local models are used to determine the first aggregate model, and the second global model is determined based on one or more of the following information: the first global model; the first aggregate model; the model obtained by the second training; the first weight corresponding to the first aggregate model; and the second weight corresponding to the model obtained by the second training.

[0239] Optionally, the sum of the first weight and the second weight is 1, and the first weight and the second weight are non-negative real numbers.

[0240] Optionally, the first weight and the second weight are determined according to the number of first data samples obtained by dividing local data samples by multiple terminal devices and the number of mixed data samples used for the second training.

[0241] Optionally, the mixed data samples are partial data samples among a plurality of second data samples obtained by dividing a local data sample by the terminal device, and the partial data samples are determined according to a forgetting mechanism.

[0242] Optionally, the first training and the second training are performed in parallel within the first training cycle.

[0243] Optionally, the first training cycle is a maximum value between the first sub-cycle and the second sub-cycle, the first sub-cycle is related to the first training, and the second sub-cycle is related to the second training.

[0244] Optionally, the first training is used to determine multiple local models for multiple terminal devices including the first terminal device, and the first sub-period is determined based on multiple first durations of the first training performed by the multiple terminal devices and one or more second durations of sending the multiple local models to the network device.

[0245] Optionally, the first terminal device is any terminal device among multiple terminal devices that receive the first global model, and the second sub-period is determined based on one or more third time durations for the multiple terminal devices to send multiple second data samples to the network device and a fourth time duration for the network device to perform second training.

[0246] Optionally, the first training is based on federated learning, and the second training is based on centralized learning.

[0247] Figure 9 is a schematic diagram of the structure of a control device for the terminal device shown in Figure 8. This control device 900 can be used to implement semi-federated learning based on retransmission over-the-air computation. As shown in Figure 9, in a semi-federated learning system based on retransmission over-the-air computation, the control device 900 of the terminal device may include a data acquisition and partitioning module 910, a federated learning module 920, a retransmission instruction receiving module 930, and an over-the-air computation input signal sending module 940.

[0248] The data collection and division module 910 can be used to control the terminal device to collect local data samples, and control the terminal device to divide the collected data samples into federated learning data samples and centralized learning data samples.

[0249] The federated learning module 920 can be used to control the terminal device to use the federated learning data samples to train the global model of the previous round to obtain a local model.

[0250] The retransmission instruction receiving module 930 can be used to control the terminal device to receive the retransmission instruction sent by the base station, and control the terminal device to retransmit the current over-the-air calculation input signal according to the content of the retransmission instruction.

[0251] The over-the-air computing input signal sending module 940 may be used to control the terminal device to process the over-the-air computing task into an over-the-air computing input signal, and control the terminal device to send the over-the-air computing input signal on the same time-frequency resources.

[0252] FIG10 is a schematic block diagram of a network device according to an embodiment of the present application. The network device 1000 may be any of the network devices described above for machine learning model training. The network device 1000 shown in FIG10 includes a sending unit 1010 and a receiving unit 1020.

[0253] The sending unit 1010 may be configured to send the first global model to multiple terminal devices including the first terminal device.

[0254] The receiving unit 1020 can be used for the network device to receive multiple second data samples sent by multiple terminal devices, and the multiple second data samples are determined based on the local data samples of the multiple terminal devices, and the local data samples are divided into first data samples and second data samples; wherein the multiple first data samples of the multiple terminal devices are respectively used for the multiple terminal devices to perform first training on the first global model within the first training cycle, and the multiple second data samples are used for the network device to perform second training on the first global model within the first training cycle.

[0255] Optionally, the second data sample of the first terminal device includes a first sample segment, and the receiving unit 1020 is further used to receive the output signal of the first air calculation based on the air calculation mechanism within the first training cycle, and the input signal of the first air calculation corresponds to multiple sample segments sent by multiple terminal devices through the first resource, and the multiple sample segments include the first sample segment.

[0256] Optionally, the first training is used by the first terminal device to determine a first local model, and the first local model includes a first model fragment. The receiving unit 1020 is also used to receive the output signal of the second air calculation based on the air calculation mechanism within the first training cycle. The input signal of the second air calculation corresponds to multiple model fragments sent by multiple terminal devices through the second resource, and the multiple model fragments include the first model fragment.

[0257] Optionally, the over-the-air calculation mechanism supports multiple terminal devices to retransmit sample segments and / or model segments that have failed to be transmitted.

[0258] Optionally, the sending unit 1010 is also used to send a retransmission indication to multiple terminal devices during the first training cycle if the output signal of the third air calculation does not meet the first condition, and the retransmission indication is used to instruct the multiple terminal devices to retransmit the sample fragments or model fragments participating in the third air calculation.

[0259] Optionally, the first training is used for multiple terminal devices to determine multiple local models, the multiple local models and the model obtained by the second training are used by the network device to determine the second global model, and the second global model is used by the network device to determine whether the trained machine learning model has reached convergence.

[0260] Optionally, during the first training cycle, multiple local models are used to determine the first aggregate model, and the second global model is determined based on one or more of the following information: the first global model; the first aggregate model; the model obtained by the second training; the first weight corresponding to the first aggregate model; and the second weight corresponding to the model obtained by the second training.

[0261] Optionally, the sum of the first weight and the second weight is 1, and the first weight and the second weight are non-negative real numbers.

[0262] Optionally, the first weight and the second weight are determined according to the number of first data samples obtained by dividing local data samples by multiple terminal devices and the number of mixed data samples used for the second training.

[0263] Optionally, the mixed data samples are partial data samples among a plurality of second data samples obtained by dividing a local data sample by the terminal device, and the partial data samples are determined according to a forgetting mechanism.

[0264] Optionally, the first training and the second training are performed in parallel within the first training cycle.

[0265] Optionally, the first training cycle is a maximum value between the first sub-cycle and the second sub-cycle, the first sub-cycle is related to the first training, and the second sub-cycle is related to the second training.

[0266] Optionally, the first training is used for multiple terminal devices to determine multiple local models, and the first sub-period is determined based on multiple first durations of the first training performed by the multiple terminal devices and one or more second durations of sending the multiple local models to the network device.

[0267] Optionally, the second sub-period is determined according to one or more third durations during which the multiple terminal devices send the multiple second data samples to the network device and a fourth duration during which the network device performs the second training.

[0268] Optionally, the first training is based on federated learning, and the second training is based on centralized learning.

[0269] Figure 11 is a schematic diagram of the structure of a control device for the network device shown in Figure 10. This control device 1100 can be used to implement semi-federated learning based on retransmission over-the-air computation. As shown in Figure 11, in a semi-federated learning system based on retransmission over-the-air computation, the control device 1100 of the network device may include an over-the-air computation output signal receiving module 1110, an over-the-air computation output signal quality assessment module 1120, a retransmission instruction sending module 1130, a centralized learning module 1140, and a global model generation module 1150.

[0270] The over-the-air computation output signal receiving module 1110 can be used to control the base station to receive over-the-air computation output signals. For centralized learning data sample uploads based on a retransmitted over-the-air computation mechanism, the over-the-air computation output signals are mixed data sample segments; for local model uploads based on a retransmitted over-the-air computation mechanism, the over-the-air computation output signals are federated learning aggregate model segments.

[0271] The over-the-air calculation output signal quality evaluation module 1120 may be used to control the base station to evaluate whether the quality of the over-the-air calculation output signal meets a preset standard.

[0272] The retransmission instruction sending module 1130 may be configured to control the base station to send a retransmission instruction to all terminal devices based on the quality evaluation result of the output signal calculated over the air.

[0273] The centralized learning module 1140 may be used to control the base station to train the global model of the previous round using mixed centralized learning data samples to obtain a centralized learning model.

[0274] The global model generation module 1150 can be used to control the base station weighted hybrid federated learning aggregation model and the centralized learning model to obtain a global model.

[0275] Figure 12 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device is used to implement any step in the semi-federated learning process. As shown in Figure 12, the electronic device structure includes a processor 1210, a memory 1220, a communication interface 1230, and a communication bus 1240.

[0276] The processor 1210 may be configured to execute a program stored in the memory 1220 to implement any step in the semi-federated learning process based on retransmission over-the-air computing provided in the above-mentioned embodiment of the present application.

[0277] The memory 1220 may be used to store programs related to semi-federated learning based on retransmission over-the-air computing.

[0278] The communication interface 1230 can be used for external entities to modify the programs stored in the memory 1220. The external entities include, but are not limited to, maintenance personnel of the semi-federated learning system based on retransmission over-the-air computing, system management equipment, etc., which are not specifically limited in this embodiment of the application.

[0279] The communication bus 1240 may be used to facilitate communication among the processor 1210 , the memory 1220 , and the communication interface 1230 .

[0280] Figure 13 shows a schematic block diagram of a communication device according to an embodiment of the present application. The dashed lines in Figure 13 indicate that the unit or module is optional. Device 1300 can be used to implement the method described in the above method embodiment. Device 1300 can be a chip, a terminal device, or a network device.

[0281] The device 1300 may include one or more processors 1310. The processor 1310 may support the device 1300 to implement the method described in the above method embodiment. The processor 1310 may be a general-purpose processor or a special-purpose processor. For example, the processor may be a central processing unit (CPU). Alternatively, the processor may be another general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, etc. The general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0282] The apparatus 1300 may further include one or more memories 1320. The memories 1320 store programs that can be executed by the processor 1310, causing the processor 1310 to perform the methods described in the above method embodiments. The memories 1320 may be independent of the processor 1310 or integrated into the processor 1310.

[0283] The apparatus 1300 may further include a transceiver 1330. The processor 1310 may communicate with other devices or chips via the transceiver 1330. For example, the processor 1310 may transmit and receive data with other devices or chips via the transceiver 1330.

[0284] The present application also provides a computer-readable storage medium for storing a program. The computer-readable storage medium can be applied to a terminal device or network device provided in the present application, and the program enables a computer to execute the method performed by the terminal device or network device in each embodiment of the present application.

[0285] The computer-readable storage medium can be any available medium that can be read by a computer or a data storage device such as a server or data center that includes one or more available media. The available medium can be a magnetic medium, an optical medium, or a semiconductor medium. Examples of computer storage media include, but are not limited to: phase-change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc-read only memory (CD-ROM), solid state disk (SSD), digital versatile disc (DVD) or other optical storage, magnetic cassette, tape / disk storage or other magnetic storage device or any other non-transmission medium. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0286] Computer-readable media includes both permanent and non-permanent, removable and non-removable media, and can be implemented using any method or technology for information storage. Computer-readable media can be used to store information that can be accessed by a computing device. This information can be computer-readable instructions, data structures, program modules, or other data.

[0287] An embodiment of the present application also provides a readable storage medium on which a program or instruction is stored. When the program or instruction is executed by a processor, the various processes of the method embodiments shown in Figures 1 to 7 above can be implemented and the same technical effect can be achieved. To avoid repetition, they will not be repeated here.

[0288] The present application also provides a computer program product. The computer program product includes a program. The computer program product can be applied to a terminal device or network device provided in the present application, and the program causes a computer to execute the method performed by the terminal or network device in each embodiment of the present application.

[0289] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part 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, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means.

[0290] The present application also provides a computer program that can be applied to a terminal device or network device provided in the present application, and enables a computer to execute the method performed by the terminal or network device in each embodiment of the present application.

[0291] The terms "system" and "network" are used interchangeably in this application. Furthermore, the terms used in this application are intended only to explain the specific embodiments of this application and are not intended to limit this application. The terms "first," "second," "third," and "fourth," etc. in the specification and claims of this application and the accompanying drawings are used to distinguish between different objects, rather than to describe a specific order.

[0292] It should be noted that the terms "comprises," "includes," and "has," and any variations thereof, are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, article, or apparatus. In the absence of further qualification, an element defined by the phrase "comprising a..." does not preclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0293] In the embodiments of this application, the term "indication" may refer to a direct indication, an indirect indication, or an indication of an association. For example, "A indicates B" may refer to a direct indication of B, e.g., B can obtain information through A; it may refer to an indirect indication of B, e.g., A indicates C, e.g., B can obtain information through C; or it may refer to an association between A and B.

[0294] In the embodiments of the present application, the term "corresponding" may indicate a direct or indirect correspondence between the two, or an association relationship between the two, or a relationship between indication and indication, configuration and configuration, etc.

[0295] In the embodiments of the present application, the "protocol" may refer to a standard protocol in the communication field, for example, it may include an LTE protocol, a NR protocol, and related protocols used in future communication systems, and this application does not limit this.

[0296] In the embodiments of the present application, determining B based on A does not mean determining B only based on A. B can also be determined based on A and / or other information.

[0297] In the embodiments of this application, the term "and / or" is simply a description of the association relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this document generally indicates that the related objects are in an "or" relationship.

[0298] In the embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0299] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0300] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0301] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0302] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above method embodiments can be implemented by means of software plus the necessary general hardware platform, and of course, by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a service classification device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0303] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A training method for a machine learning model, characterized in that: include: The first terminal device receives the first global model sent by the network device; The first terminal device divides the local data sample into a first data sample and a second data sample; The first data sample is used by the first terminal device to perform first training on the first global model within a first training cycle, and the second data sample is used by the network device to perform second training on the first global model within the first training cycle.

2. The training method according to claim 1, characterized in that: The sample quantity of the first data samples and the sample quantity of the second data samples are determined according to the communication capability and / or computing capability of the first terminal device.

3. The training method according to claim 1, characterized in that: The second data sample is related to information disclosed by the first terminal device.

4. The training method according to any one of claims 1 to 3, characterized in that: The second data sample includes a first sample segment, and the training method further includes: In the first training cycle, the first terminal device sends the first sample segment to the network device on a first resource based on an air calculation mechanism; The first resource is also used by other terminal devices except the first terminal device to send other sample fragments to the network device, and the other sample fragments and the first sample fragments are used to input into a first air calculation.

5. The training method according to any one of claims 1 to 4, characterized in that: The first training is used by the first terminal device to determine a first local model, the first local model includes a first model segment, and the training method further includes: In the first training cycle, the first terminal device sends the first model fragment to the network device on a second resource based on an over-the-air calculation mechanism; The second resource is also used by other terminal devices except the first terminal device to send other model fragments to the network device, and the other model fragments and the first model fragment are used to input into a second air calculation.

6. The training method according to claim 4 or 5, characterized in that: The over-the-air computing mechanism supports the first terminal device to retransmit sample segments and / or model segments that have failed to transmit.

7. The training method according to claim 6, characterized in that: The training method further comprises: During the first training cycle, if the output signal of the third air calculation does not meet the first condition, the first terminal device receives a retransmission indication sent by the network device, and the retransmission indication is used to instruct multiple terminal devices to retransmit sample fragments or model fragments participating in the third air calculation.

8. The training method according to any one of claims 1 to 7, characterized in that: The first training is used for multiple terminal devices including the first terminal device to determine multiple local models, the multiple local models and the model obtained by the second training are used by the network device to determine a second global model, and the second global model is used by the network device to determine whether the trained machine learning model has reached convergence.

9. The training method according to claim 8, characterized in that: In the first training cycle, the multiple local models are used to determine a first aggregate model, and the second global model is determined according to one or more of the following information: the first global model; the first aggregation model; The model obtained by the second training; a first weight corresponding to the first aggregation model; The second weight corresponding to the model obtained by the second training.

10. The training method according to claim 9, characterized in that: A sum of the first weight and the second weight is 1, and the first weight and the second weight are non-negative real numbers.

11. The training method according to claim 9 or 10, characterized in that: The first weight and the second weight are determined according to the number of first data samples obtained by dividing local data samples by the multiple terminal devices and the number of mixed data samples used for the second training.

12. The training method according to claim 11, characterized in that: The mixed data samples are partial data samples among a plurality of second data samples obtained by dividing the local data samples by the terminal device, and the partial data samples are determined according to a forgetting mechanism.

13. The training method according to any one of claims 1 to 12, characterized in that: The first training and the second training are performed in parallel within the first training cycle.

14. The training method according to any one of claims 1 to 13, characterized in that: The first training cycle is a maximum value of a first sub-cycle and a second sub-cycle, the first sub-cycle is related to the first training, and the second sub-cycle is related to the second training.

15. The training method according to claim 14, characterized in that: The first training is used to determine multiple local models for multiple terminal devices including the first terminal device, and the first sub-period is determined based on multiple first durations for the multiple terminal devices to perform the first training and one or more second durations for sending the multiple local models to the network device.

16. The training method according to claim 14 or 15, characterized in that: The first terminal device is any terminal device among multiple terminal devices that receive the first global model, and the second sub-period is determined based on one or more third time durations for the multiple terminal devices to send multiple second data samples to the network device and a fourth time duration for the network device to perform the second training.

17. The training method according to any one of claims 1 to 16, characterized in that: The first training is training based on federated learning, and the second training is training based on centralized learning.

18. A method for training a machine learning model, characterized in that: include: The network device sends the first global model to a plurality of terminal devices including the first terminal device; The network device receives a plurality of second data samples sent by the plurality of terminal devices, the plurality of second data samples are determined according to local data samples of the plurality of terminal devices, the local data samples are divided into first data samples and second data samples; Among them, the multiple first data samples of the multiple terminal devices are respectively used for the multiple terminal devices to perform first training on the first global model within the first training cycle, and the multiple second data samples are used for the network device to perform second training on the first global model within the first training cycle.

19. The training method according to claim 18, characterized in that: The second data sample of the first terminal device includes a first sample segment, and the network device receives a plurality of second data samples sent by the plurality of terminal devices, including: During the first training cycle, the network device receives an output signal of a first air-computation based on an air-computation mechanism, wherein the input signal of the first air-computation corresponds to a plurality of sample segments sent by the plurality of terminal devices through a first resource, and the plurality of sample segments include the first sample segment.

20. The training method according to claim 18 or 19, characterized in that: The first training is used by the first terminal device to determine a first local model, the first local model includes a first model segment, and the training method further includes: During the first training cycle, the network device receives an output signal of a second air-computation based on an air-computation mechanism, and the input signal of the second air-computation corresponds to a plurality of model fragments sent by the plurality of terminal devices through a second resource, and the plurality of model fragments include the first model fragment.

21. The training method according to claim 19 or 20, characterized in that: The over-the-air calculation mechanism supports the multiple terminal devices to retransmit the sample segments and / or model segments that have failed to be transmitted.

22. The training method according to claim 21, characterized in that: The training method further comprises: During the first training cycle, if the output signal of the third air calculation does not meet the first condition, the network device sends a retransmission indication to the multiple terminal devices, and the retransmission indication is used to instruct the multiple terminal devices to retransmit the sample fragments or model fragments participating in the third air calculation.

23. The training method according to any one of claims 18 to 22, characterized in that: The first training is used for the multiple terminal devices to determine multiple local models, the multiple local models and the model obtained by the second training are used for the network device to determine a second global model, and the second global model is used for the network device to determine whether the trained machine learning model has reached convergence.

24. The training method according to claim 23, characterized in that: In the first training cycle, the multiple local models are used to determine a first aggregate model, and the second global model is determined according to one or more of the following information: the first global model; the first aggregation model; The model obtained by the second training; a first weight corresponding to the first aggregation model; The second weight corresponding to the model obtained by the second training.

25. The training method according to claim 24, characterized in that: A sum of the first weight and the second weight is 1, and the first weight and the second weight are non-negative real numbers.

26. The training method according to claim 24 or 25, characterized in that: The first weight and the second weight are determined according to the number of first data samples obtained by dividing local data samples by the multiple terminal devices and the number of mixed data samples used for the second training.

27. The training method according to claim 26, characterized in that: The mixed data samples are partial data samples among a plurality of second data samples obtained by dividing the local data samples by the terminal device, and the partial data samples are determined according to a forgetting mechanism.

28. The training method according to any one of claims 18 to 27, characterized in that: The first training and the second training are performed in parallel within the first training cycle.

29. The training method according to any one of claims 18 to 28, characterized in that: The first training cycle is a maximum value of a first sub-cycle and a second sub-cycle, the first sub-cycle is related to the first training, and the second sub-cycle is related to the second training.

30. The training method according to claim 29, characterized in that: The first training is used for the multiple terminal devices to determine multiple local models, and the first sub-period is determined according to multiple first durations for the multiple terminal devices to perform the first training and one or more second durations for sending the multiple local models to the network device.

31. The training method according to claim 29 or 30, characterized in that: The second sub-period is determined according to one or more third durations during which the multiple terminal devices send multiple second data samples to the network device and a fourth duration during which the network device performs the second training.

32. The training method according to any one of claims 18 to 31, characterized in that: The first training is training based on federated learning, and the second training is training based on centralized learning.

33. A terminal device, characterized in that: The terminal device is a first terminal device for training a machine learning model, and the terminal device includes: A receiving unit, configured to receive a first global model sent by a network device; A processing unit, configured to divide the local data sample into a first data sample and a second data sample; The first data sample is used by the first terminal device to perform first training on the first global model within a first training cycle, and the second data sample is used by the network device to perform second training on the first global model within the first training cycle.

34. The terminal device according to claim 33, characterized in that: The sample quantity of the first data samples and the sample quantity of the second data samples are determined according to the communication capability and / or computing capability of the first terminal device.

35. The terminal device according to claim 33, characterized in that: The second data sample is related to information disclosed by the first terminal device.

36. The terminal device according to any one of claims 33 to 35, characterized in that: The second data sample includes a first sample segment, and the terminal device further includes: A first sending unit, configured to send the first sample segment to the network device on a first resource based on an air calculation mechanism within the first training cycle; The first resource is also used by other terminal devices except the first terminal device to send other sample fragments to the network device, and the other sample fragments and the first sample fragments are used to input into a first air calculation.

37. The terminal device according to any one of claims 33 to 36, characterized in that: The first training is used by the first terminal device to determine a first local model, where the first local model includes a first model segment, and the terminal device further includes: a second sending unit, configured to send the first model fragment to the network device on a second resource based on an over-the-air calculation mechanism during the first training cycle; The second resource is also used by other terminal devices except the first terminal device to send other model fragments to the network device, and the other model fragments and the first model fragment are used to input into a second air calculation.

38. The terminal device according to claim 36 or 37, characterized in that: The over-the-air calculation mechanism supports the first terminal device to retransmit the sample segments and / or model segments that have failed to be transmitted.

39. The terminal device according to claim 38, characterized in that: The receiving unit is also used to receive a retransmission indication sent by the network device during the first training cycle if the output signal of the third air calculation does not meet the first condition, and the retransmission indication is used to instruct multiple terminal devices to retransmit sample fragments or model fragments participating in the third air calculation.

40. The terminal device according to any one of claims 33 to 39, characterized in that: The first training is used for multiple terminal devices including the first terminal device to determine multiple local models, the multiple local models and the model obtained by the second training are used by the network device to determine a second global model, and the second global model is used by the network device to determine whether the trained machine learning model has reached convergence.

41. The terminal device according to claim 40, characterized in that: In the first training cycle, the multiple local models are used to determine a first aggregate model, and the second global model is determined according to one or more of the following information: the first global model; the first aggregation model; The model obtained by the second training; a first weight corresponding to the first aggregation model; The second weight corresponding to the model obtained by the second training.

42. The terminal device according to claim 41, characterized in that: A sum of the first weight and the second weight is 1, and the first weight and the second weight are non-negative real numbers.

43. The terminal device according to claim 41 or 42, characterized in that: The first weight and the second weight are determined according to the number of first data samples obtained by dividing local data samples by the multiple terminal devices and the number of mixed data samples used for the second training.

44. The terminal device according to claim 43, characterized in that: The mixed data samples are partial data samples among a plurality of second data samples obtained by dividing the local data samples by the terminal device, and the partial data samples are determined according to a forgetting mechanism.

45. The terminal device according to any one of claims 33-44, characterized in that: The first training and the second training are performed in parallel within the first training cycle.

46. ​​The terminal device according to any one of claims 33 to 45, characterized in that: The first training cycle is a maximum value of a first sub-cycle and a second sub-cycle, the first sub-cycle is related to the first training, and the second sub-cycle is related to the second training.

47. The terminal device according to claim 46, characterized in that: The first training is used to determine multiple local models for multiple terminal devices including the first terminal device, and the first sub-period is determined based on multiple first durations for the multiple terminal devices to perform the first training and one or more second durations for sending the multiple local models to the network device.

48. The terminal device according to claim 46 or 47, characterized in that: The first terminal device is any terminal device among multiple terminal devices that receive the first global model, and the second sub-period is determined based on one or more third time durations for the multiple terminal devices to send multiple second data samples to the network device and a fourth time duration for the network device to perform the second training.

49. The terminal device according to any one of claims 33 to 48, characterized in that: The first training is training based on federated learning, and the second training is training based on centralized learning.

50. A network device, characterized in that: The network device is used to train a machine learning model, and the network device includes: A sending unit, configured to send the first global model to a plurality of terminal devices including the first terminal device; A receiving unit, configured to receive, by the network device, a plurality of second data samples sent by the plurality of terminal devices, wherein the plurality of second data samples are determined according to local data samples of the plurality of terminal devices, and the local data samples are divided into first data samples and second data samples; Among them, the multiple first data samples of the multiple terminal devices are respectively used for the multiple terminal devices to perform first training on the first global model within the first training cycle, and the multiple second data samples are used for the network device to perform second training on the first global model within the first training cycle.

51. The network device according to claim 50, characterized in that The second data sample of the first terminal device includes a first sample segment, and the receiving unit is also used to receive the output signal of the first air calculation based on the air calculation mechanism during the first training cycle, and the input signal of the first air calculation corresponds to multiple sample segments sent by the multiple terminal devices through the first resource, and the multiple sample segments include the first sample segment.

52. The network device according to claim 50 or 51, characterized in that: The first training is used for the first terminal device to determine a first local model, and the first local model includes a first model fragment. The receiving unit is also used to receive an output signal of a second air-computation based on an air-computation mechanism during the first training cycle, and the input signal of the second air-computation corresponds to multiple model fragments sent by the multiple terminal devices through the second resource, and the multiple model fragments include the first model fragment.

53. The network device according to claim 51 or 52, characterized in that: The over-the-air calculation mechanism supports the multiple terminal devices to retransmit the sample segments and / or model segments that have failed to be transmitted.

54. The network device according to claim 53, characterized in that The sending unit is also used to send a retransmission indication to the multiple terminal devices during the first training cycle if the output signal of the third air calculation does not meet the first condition, and the retransmission indication is used to instruct the multiple terminal devices to retransmit the sample fragments or model fragments participating in the third air calculation.

55. The network device according to any one of claims 50-54, characterized in that: The first training is used for the multiple terminal devices to determine multiple local models, the multiple local models and the model obtained by the second training are used for the network device to determine a second global model, and the second global model is used for the network device to determine whether the trained machine learning model has reached convergence.

56. The network device according to claim 55, characterized in that In the first training cycle, the multiple local models are used to determine a first aggregate model, and the second global model is determined according to one or more of the following information: the first global model; the first aggregation model; The model obtained by the second training; a first weight corresponding to the first aggregation model; The second weight corresponding to the model obtained by the second training.

57. The network device according to claim 56, characterized in that A sum of the first weight and the second weight is 1, and the first weight and the second weight are non-negative real numbers.

58. The network device according to claim 56 or 57, characterized in that: The first weight and the second weight are determined according to the number of first data samples obtained by dividing local data samples by the multiple terminal devices and the number of mixed data samples used for the second training.

59. The network device according to claim 58, characterized in that The mixed data samples are partial data samples among a plurality of second data samples obtained by dividing the local data samples by the terminal device, and the partial data samples are determined according to a forgetting mechanism.

60. The network device according to any one of claims 50-59, characterized in that: The first training and the second training are performed in parallel within the first training cycle.

61. The network device according to any one of claims 50 to 60, characterized in that: The first training cycle is a maximum value of a first sub-cycle and a second sub-cycle, the first sub-cycle is related to the first training, and the second sub-cycle is related to the second training.

62. The network device according to claim 61, characterized in that The first training is used for the multiple terminal devices to determine multiple local models, and the first sub-period is determined according to multiple first durations for the multiple terminal devices to perform the first training and one or more second durations for sending the multiple local models to the network device.

63. The network device according to claim 61 or 62, characterized in that: The second sub-period is determined according to one or more third durations during which the multiple terminal devices send multiple second data samples to the network device and a fourth duration during which the network device performs the second training.

64. The network device according to any one of claims 50 to 63, characterized in that: The first training is training based on federated learning, and the second training is training based on centralized learning.

65. A communication device, characterized in that: It comprises a memory and a processor, wherein the memory is used to store a program, and the processor is used to call the program in the memory to execute the method as claimed in any one of claims 1 to 32.

66. A device, characterized in that The invention comprises a processor, which is used to call a program from a memory to execute the method according to any one of claims 1 to 32.

67. A chip, characterized in that: It comprises a processor, which is used to call a program from a memory, so that a device equipped with the chip executes a method as claimed in any one of claims 1 to 32.

68. A computer-readable storage medium, characterized in that A program is stored thereon, the program causing a computer to execute the method according to any one of claims 1 to 32.

69. A computer program product, characterized in that The method comprises a program which causes a computer to execute the method according to any one of claims 1 to 32.

70. A computer program, characterized in that The computer program causes a computer to execute the method according to any one of claims 1 to 32.

Images