Communication method, storage medium, chip system, and communication system

Abstract

Embodiments of the present application disclose a communication method, a storage medium, a chip system, and a communication system. The method comprises: receiving first information, wherein the first information can explicitly or implicitly indicate a usage scope of training data, and the usage scope of the training data can specify one or more of a data validity period of the training data, an available model, an available function, and an available device; and, on the basis of the usage scope of the training data and the training data, performing model training on an artificial intelligence (AI) model to finally obtain a trained AI model. The present application facilitates improvement of the security and reliability of data interaction between different devices during AI model deployment.

Description

Communication method, storage medium, chip system and communication system

[0001] The present application claims priority to the Chinese patent application No. 202411128609.7, filed on August 15, 2024, and entitled "Communication method, storage medium, chip system and communication system", the content of which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[0002] The present application relates to the field of communication technology, and in particular to a communication method, a storage medium, a chip system and a communication system. BACKGROUND

[0003] The progress of artificial intelligence (AI) has driven the development of various industries and laid the foundation for the future development of communication technology. Under the background of rapid development of AI, the 3rd generation partnership project (3GPP) has launched a standardized research on AI communication. One of the important research directions is to apply AI models to air interface technology. AI models need to be trained with a large amount of data before deployment, so that the AI model can converge to appropriate performance. Since the AI model training process often involves data interaction between different devices, how to improve the data security in the data interaction process is a technical problem worthy of study. SUMMARY

[0004] The embodiments of the present application provide a communication method, a storage medium, a chip system and a communication system, which are beneficial to improve the security and reliability of data interaction between different devices in the AI model deployment process.

[0005] In a first aspect, the embodiments of the present application provide a communication method, which can be applied to a second device, or an apparatus (for example, a chip, or a chip system, or a circuit) in the second device, or an apparatus that can be matched with the second device, the second device can be a device responsible for model training, and the method can include: receiving first information; the first information indicates a use range of training data; the use range of the training data includes one or more of a data valid time, a usable model, a usable function and a usable device of the training data; performing model training on an AI model based on the use range of the training data and the training data to obtain a trained AI model.

[0006] Based on the communication method, the device responsible for model training performs model training on the AI model by using the training data according to the use range of the training data, which is beneficial to protect the privacy of the training data and improve the security and reliability of data interaction between different devices.

[0007] With reference to the first aspect, in a possible implementation manner, the method further includes: receiving or sending the training data. In the embodiments of the present application, the training data and the use range of the training data can be carried in different signaling, thereby facilitating the flexibility of data transmission. In addition, the device that sends the training data and the device that sends the first information can also be different devices, which facilitates the flexibility of model training.

[0008] With reference to the first aspect, in a possible implementation manner, the first information can explicitly indicate the use range of the training data; the first information includes one or more of a data validity time of the training data, an available model, an available function, and an available device. In the embodiments of the present application, when the first information explicitly indicates the use range of the training data, the content of the first information can include one or more of the data validity time of the training data, the available model, the available function, and the available device. This facilitates the efficiency of indicating the use range of the training data.

[0009] With reference to the first aspect, in a possible implementation manner, the first information can implicitly indicate the use range of the training data; the first information includes data type of the training data and / or model information of an AI model; at least one of the data type of the training data and the model information of the AI model has a corresponding relationship with the use range of the training data; the model information of the AI model includes one or more of a model identifier, a model function, a data volume, and input and output of the AI model. The corresponding relationship between the data type of the training data and / or the model information of the AI model and the use range of the training data can be preconfigured or predefined.

[0010] In the embodiments of the present application, when the first information implicitly indicates the use range of the training data, the content of the first information can not directly include the use range of the training data, but include the data type of the training data and / or the model information of the AI model. In this way, the second device can obtain the use range of the training data according to the data type of the training data and / or the model information of the AI model. This facilitates the flexibility of indicating the use range of the training data.

[0011] With reference to the first aspect, in a possible implementation manner, the first information further includes a data privacy level of the training data. In the embodiments of the present application, by distinguishing the data privacy levels of different training data, the data security can be finely improved.

[0012] With reference to the first aspect, in a possible implementation manner, the data privacy level is one of a first privacy level, a second privacy level, or a third privacy level; the first privacy level is higher than the second privacy level, and the second privacy level is higher than the third privacy level. In the embodiment of the present application, the data privacy level of the training data can be divided into three levels, the first privacy level can also be described as a high privacy level, the second privacy level can also be described as a medium privacy level, and the third privacy level can also be described as a low privacy level.

[0013] With reference to the first aspect, in a possible implementation manner, the higher the data privacy level, the smaller the indicated use range of the training data. In the embodiment of the present application, the higher the data privacy level of the training data, the shorter the data validity time, the smaller the number of available models, the smaller the number of available functions, and / or the smaller the number of available devices. In this way, the privacy protection of the training data is enhanced.

[0014] With reference to the first aspect, in a possible implementation manner, the receiving or transmitting the training data comprises receiving or transmitting the training data based on a transmission resource; the higher the data privacy level, the higher the transmission efficiency of the transmission resource required. In the embodiment of the present application, the training data with a higher data privacy level can be transmitted in a shorter time, which is beneficial to improve the security of data transmission.

[0015] With reference to the first aspect, in a possible implementation manner, the method further comprises deleting the trained AI model and / or the training data based on a reference time; the reference time is a data validity time of the training data, a cell residence time of the terminal device, or a completion time of the training task. In the embodiment of the present application, the second device can actively delete the trained AI model and / or the training data, thereby improving the data security.

[0016] With reference to the first aspect, in a possible implementation manner, the method further comprises deleting the trained AI model and / or the training data in response to the number of uses of the training data reaching a threshold. In the embodiment of the present application, the second device can delete the trained AI model and / or the training data when the number of uses of the training data reaches the maximum number of uses, thereby improving the data security.

[0017] With reference to the first aspect, in a possible implementation manner, the method further comprises receiving a deletion request message; deleting the trained AI model and / or the training data in response to the deletion request message; and sending a deletion response message. In the embodiment of the present application, the second device can delete the trained AI model and / or the training data when the deletion request message is received, thereby improving the data security.

[0018] In a second aspect, an embodiment of the present application provides another communication method, which can be applied to a first device, a device (for example, a chip, or a chip system, or a circuit) in the first device, or a device that can be used in conjunction with the first device. The first device can collect (or described as collect, measure, etc.) training data of an AI model, or in other words, the first device can be a device responsible for collecting training data. The method can include: sending first information, the first information indicating a usage range of the training data; the usage range of the training data including one or more of a data validity time of the training data, a usable model, a usable function, and a usable device; and the usage range of the training data being used to obtain a trained AI model.

[0019] Based on the communication method, by limiting the usage range of the training data, the privacy of the training data is protected, and the security and reliability of data interaction between different devices are improved.

[0020] In combination with the second aspect, in a possible implementation, the method further includes: sending or receiving the training data. In the embodiment of the present application, the training data and the usage range of the training data can be carried in different signaling, thereby facilitating the flexibility of data transmission.

[0021] In combination with the second aspect, in a possible implementation, the first information can explicitly indicate the usage range of the training data; and the first information includes one or more of the data validity time of the training data, the usable model, the usable function, and the usable device. In the embodiment of the present application, in the case that the first information explicitly indicates the usage range of the training data, the content of the first information can include one or more of the data validity time of the training data, the usable model, the usable function, and the usable device. In this way, the efficiency of indicating the usage range of the training data is improved.

[0022] In combination with the second aspect, in a possible implementation, the first information can implicitly indicate the usage range of the training data; the first information includes a data type of the training data and / or model information of an AI model; at least one of the data type of the training data and the model information of the AI model has a corresponding relationship with the usage range of the training data; and the model information of the AI model includes one or more of a model identifier of the AI model, a model function, a data volume, and an input / output. The corresponding relationship can be pre-configured or pre-defined.

[0023] In the embodiments of the present application, in the case that the first information implicitly indicates the use range of the training data, the content of the first information can not directly include the use range of the training data, but includes the data type of the training data and / or the model information of the AI model. In this way, the second device can learn the use range of the training data according to the data type of the training data and / or the model information of the AI model. In this way, the flexibility of indicating the use range of the training data is improved.

[0024] In combination with the second aspect, in a possible implementation, the first information further includes a data privacy level of the training data. In the embodiments of the present application, by distinguishing the data privacy levels of different training data, the data security is improved in a fine-grained manner.

[0025] In combination with the second aspect, in a possible implementation, the data privacy level is one of a first privacy level, a second privacy level or a third privacy level; the first privacy level is higher than the second privacy level, and the second privacy level is higher than the third privacy level. In the embodiments of the present application, the data privacy level of the training data can be divided into three levels, the first privacy level can also be described as a high privacy level, the second privacy level can also be described as a medium privacy level, and the third privacy level can also be described as a low privacy level.

[0026] In combination with the second aspect, in a possible implementation, the higher the data privacy level, the smaller the indicated use range of the training data. In the embodiments of the present application, the higher the data privacy level of the training data, the shorter the data validity time, the smaller the number of available models, the smaller the number of available functions, and the smaller the number of available devices. In this way, the privacy protection of the training data is enhanced.

[0027] In combination with the second aspect, in a possible implementation, the above sending or receiving the training data includes: sending or receiving the training data based on a transmission resource; the higher the data privacy level, the higher the transmission efficiency of the transmission resource required. In the embodiments of the present application, the training data with a higher data privacy level can be transmitted in a shorter time, and in this way, the security of data transmission is improved.

[0028] In combination with the second aspect, in a possible implementation, the above method further includes: sending a deletion request message; the deletion request is used to request to delete the trained AI model and / or the training data; and receiving a deletion response message. In the embodiments of the present application, the first device sends the deletion request message, so that the second device deletes the trained AI model and / or the training data in response to the deletion request message, thereby improving the data security.

[0029] In a third aspect, an embodiment of the present application provides a communication apparatus, which can comprise a module / unit for performing any of the methods in the first aspect or any possible implementation of the first aspect, the second aspect or any possible implementation of the second aspect.

[0030] In a fourth aspect, an embodiment of the present application provides a communication apparatus, which can comprise a processor coupled with a memory, the memory being configured to store a program or instructions, which when executed by the processor, causes the communication apparatus to perform any of the methods in the first aspect or any possible implementation of the first aspect, the second aspect or any possible implementation of the second aspect.

[0031] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores computer programs or computer instructions, which when executed on a computer, causes the computer to perform any of the methods in the first aspect or any possible implementation of the first aspect, the second aspect or any possible implementation of the second aspect.

[0032] In a sixth aspect, an embodiment of the present application provides a computer program product comprising program instructions, which when executed on a computer, causes the computer to perform any of the methods in the first aspect or any possible implementation of the first aspect, the second aspect or any possible implementation of the second aspect.

[0033] In a seventh aspect, an embodiment of the present application provides a chip system, which can comprise at least one processor and an interface circuit, the interface circuit and the at least one processor being interconnected by a line, the at least one processor being configured to execute computer programs or instructions, so that any of the methods in the first aspect or any possible implementation of the first aspect, the second aspect or any possible implementation of the second aspect is executed. In a possible implementation, the chip system can further comprise at least one memory, the interface circuit, the at least one memory and the at least one processor being interconnected by a line, the at least one memory storing instructions, which when executed by the processor, any of the methods in the first aspect or any possible implementation of the first aspect, the second aspect or any possible implementation of the second aspect is executed. The chip system can be constituted by a chip, or can comprise a chip and other discrete devices.

[0034] In an eighth aspect, an embodiment of the present application provides a communication system, which comprises a first device and a second device, when the first device and the second device operate in the communication system, are configured to execute any of the methods in the first aspect or any possible implementation of the first aspect, the second aspect or any possible implementation of the second aspect. BRIEF DESCRIPTION OF DRAWINGS

[0035] FIG. 1A is a schematic diagram of an architecture of a communication system suitable for embodiments of the application;

[0036] FIG. 1B is a schematic diagram of an architecture of another communication system suitable for embodiments of the application;

[0037] FIG. 2 is a schematic diagram of an application scenario provided by embodiments of the application;

[0038] FIG. 3 is a schematic diagram of another application scenario provided by embodiments of the application;

[0039] FIG. 4 is a schematic diagram of a communication method provided by embodiments of the application;

[0040] FIG. 5 is a schematic diagram of another communication method provided by embodiments of the application;

[0041] FIG. 6 is a schematic diagram of yet another communication method provided by embodiments of the application;

[0042] FIG. 7 is a schematic diagram of a structure of a communication apparatus provided by embodiments of the application;

[0043] FIG. 8 is a schematic diagram of a structure of another communication apparatus provided by embodiments of the application. DETAILED DESCRIPTION

[0044] The technical solutions in embodiments of the application will be apparently and completely described below with reference to the drawings in embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by a person of ordinary skill in the art without any creative effort fall within the protection scope of the application.

[0045] The terms "first", "second", "third", etc. in embodiments of the application are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a series of steps or units are included, or optionally, other steps or units not listed are also included, or optionally, other steps or units inherent to the process, method, product or equipment are also included. The terms "one embodiment" or "some embodiments" mean that the specific features, structures or characteristics described in connection with the embodiment are included in one or more embodiments of the application. Therefore, the statements "in one embodiment", "in some embodiments", "in other some embodiments", "in other embodiments" and the like appearing in different parts of the embodiments of the application do not necessarily all refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized.

[0046] In the description of the present application, "indication" can include direct indication and indirect indication, and can also include explicit indication and implicit indication. If the information (the first information mentioned below) indicated by certain information is referred to as to-be-indicated information, then in the specific implementation process, there are many ways to indicate the to-be-indicated information. For example, the to-be-indicated information can be directly indicated, such as indicating the to-be-indicated information itself or an index of the to-be-indicated information, and the like. For another example, the to-be-indicated information can also be indirectly indicated by indicating other information, and there is an association relationship between the indicated other information and the to-be-indicated information. For another example, only a part of the to-be-indicated information can be indicated, and the other part of the to-be-indicated information is known or agreed in advance. In addition, the indication of a specific information can also be achieved by means of the arrangement order of each information agreed in advance (such as a protocol), thereby reducing the indication overhead to a certain extent.

[0047] In order to facilitate the understanding of the present application, before introducing the embodiments of the present application, the related names or terms involved in the present application are briefly introduced.

[0048] 1. Terminal device

[0049] A terminal device is a device with wireless transceiving function, which can be referred to as terminal, user equipment (UE), mobile station (MS), mobile terminal (MT), access terminal device, Internet of Things terminal device, vehicle-mounted terminal device, industrial control terminal device, UE unit, UE station, mobile station, remote station, remote terminal device, mobile device, UE terminal device, wireless communication device, UE agent or UE apparatus, etc. The terminal device can be fixed or mobile. It should be noted that the terminal device can support at least one wireless communication technology, such as long term evolution (LTE), new radio (NR), wideband code division multiple access (WCDMA), etc. For example, the terminal device can be a mobile phone, a pad, a desktop computer, a notebook computer, an all-in-one machine, a vehicle-mounted terminal, a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless terminal in industrial control, a wireless terminal in self-driving, a wireless terminal in remote medical surgery, a wireless terminal in smart grid, a wireless terminal in transportation safety, a wireless terminal in smart city, a wireless terminal in smart home, a cellular phone, a cordless phone, a session initiation protocol (SIP) phone, a wireless local loop (WLL) station, a personal digital assistant (PDA), a handheld device with wireless communication function, a computing device or other processing device connected to a wireless modem, a wearable device, a terminal device in a future mobile communication network or a terminal device in a future evolved public land mobile network (PLMN), etc. In some embodiments, the terminal device can also be a device with transceiving function, such as a chip module. The chip module can include a chip and can also include other discrete devices. The embodiments of the present application do not limit the specific technology and specific device form of the terminal device.

[0050] 2. Network device

[0051] The network device is a device that provides a terminal device with a wireless communication function. The network device can be an access network (AN) device or a satellite. The AN device can be a radio access network (RAN) device. The AN device can support at least one wireless communication technology, such as LTE, NR, WCDMA, and the like. Examples of the AN device include, but are not limited to, a generation nodeB (gNB) in the 5th generation mobile communication (5G), an evolved node B (eNB), a radio network controller (RNC), a node B (NB), a base station controller (BSC), a base transceiver station (BTS), a home base station (for example, a home evolved node B or a home node B, HNB), a baseband unit (BBU), a remote radio unit (RRU), an active antenna unit (AAU), a remote radio head (RRH), a transmission and reception point (TRP), a transmitting point (TP), a mobile switching center, and the like. The network device can also be a radio controller, a centralized unit (CU), a distributed unit (DU), and / or a radio unit (RU) in a cloud radio access network (CRAN) scenario. The AN device can be a macro base station, a micro base station, a relay station, an access point, a vehicle-mounted device, a wearable device, and an access network device in future mobile communications or an access network device in a future evolved PLMN, and the like. In some embodiments, the network device can also be a device that provides a terminal device with a wireless communication function, such as a chip module. The chip module can include a chip and other discrete devices. Embodiments of the present application do not limit the specific technology and specific device form of the network device.

[0052] In the embodiments of the present application, an AI module or an AI entity can be configured in a terminal device and / or a network device to implement AI-related operations. Alternatively, a separate device can be introduced to perform AI-related operations. The separate device can be referred to as an AI device or an AI network element or an AI node, etc. Optionally, the separate AI device can be directly connected with the network device, or indirectly connected with the network device through a third-party device. For ease of understanding, the embodiments of the present application take the AI module built-in in the terminal device and / or the network device as an example for description.

[0053] 3. AI model

[0054] An AI model is a specific method to implement an AI function. The AI model represents the mapping relationship between the input and the output of the model. The AI model can be a neural network, a linear regression model, a decision tree model, a clustering SVD model, or other machine learning models. Among them, the AI model can be referred to as an intelligent model, a model, or other names, which are not limited by the present application. AI-related operations can include at least one of the following: data collection, model training, model information publishing, model testing (or model verification), model inference (or model reasoning, reasoning, or prediction, etc.), or inference result publishing, etc.

[0055] The model difference includes at least one of the following: the structural parameters of the model (such as at least one of the number of layers of the model, the width of the model, the connection relationship between layers, the weight of neurons, the activation function of neurons, or the bias in the activation function), the input parameters of the model (such as the type of input parameters and / or the dimension of input parameters), or the output parameters of the model (such as the type of output parameters and / or the dimension of output parameters).

[0056] The AI model design mainly includes a data collection link (e.g., collecting training data and / or inference data), a model training link, and a model inference link. Further, an inference result application link can also be included. In the foregoing data collection link, a data source is configured to provide a training data set and inference data. In the model training link, an AI model is obtained by analyzing or training the training data provided by the data source. The AI model is learned by the model training node, which is equivalent to learning the mapping relationship between the input and output of the AI model by using the training data. In the model inference link, the AI model trained in the model training link is used to perform inference based on the inference data provided by the data source, and an inference result is obtained. This link can also be understood as follows: the inference data is input into the AI model, and the output of the AI model is obtained, which is the inference result. The inference result can indicate a configuration parameter used (executed) by an execution object and / or an operation executed by the execution object. In the inference result application link, the inference result is published, for example, the inference result can be uniformly planned by an execution entity, for example, the execution entity can send the inference result to one or more execution objects (e.g., core network equipment, access network equipment, or terminal equipment, etc.) to execute. For another example, the execution entity can also feed back the performance of the AI model to the data source, so as to facilitate subsequent implementation of the update training of the AI model.

[0057] It can be understood that the implementation of the AI model can be a hardware circuit, a software, or a combination of software and hardware, which is not limited. Non-limiting examples of the software include program code, programs, subprograms, instructions, instruction sets, codes, code segments, software modules, application programs, or software applications, etc.

[0058] The communication system applied in the embodiments of the present application is introduced as follows.

[0059] Embodiments of the present application can be applied to various communication systems. For example, the communication system can be a long term evolution (LTE) system, a 5G system, a 6th generation mobile communication (6G) system, or a communication system evolved after 5G, a satellite communication system, and a short-range wireless communication system. The wireless communication system mentioned in the embodiments of the present application includes, but is not limited to, three major application scenarios of 5G / 6G mobile communication systems, a long range (LoRa) system, or a vehicle-to-everything (V2X) system. The three major application scenarios of the 5G / 6G mobile communication system are: enhanced mobile broadband (eMBB), ultra-reliable low-latency communication (URLLC), and massive machine type of communication (mMTC).

[0060] For example, as shown in FIG. 1A, the communication system includes a radio access network 100 and a core network 200. The radio access network 100 includes at least one network device 110 (110a, 110b, collectively referred to as 110) and at least one terminal device 120 (120a-120g, collectively referred to as 120). The at least one terminal device can be connected to each other or to at least one network device in the radio access network 100.

[0061] In some embodiments, the terminal device 120 in the communication system can have a perception function. The terminal device 120 can determine information of a perception target by transmitting a signal and receiving a signal reflected by the perception target. For example, the terminal device 120 can determine the speed, distance, shape, size, and the like of the perception target. Optionally, the perception target can be a fixed object, such as a mountain, a forest, a building, or the like, or a moving object, such as a vehicle, a drone, a pedestrian, a terminal device, or the like.

[0062] In some embodiments, the devices (such as the network device 110 and the terminal device 120) in the communication system can integrate a module with one or more of the following functions: AI model training function, AI model inference function, and AI model storage function, depending on the implementation of different embodiments.

[0063] In some embodiments, a special AI network element (or AI entity) can be introduced in the communication system to perform AI model training, AI model inference, AI model storage, etc. The AI network element can be a core network element or an operation, administration and maintenance (OAM) element, for example.

[0064] For example, as shown in FIG. 1B, the AI network element is located outside the network device and the terminal device, and can communicate with the network device. Optionally, the network device can forward the data related to the AI model reported by the terminal device to the AI network element, and the AI network element can perform model training, etc., and then forward the trained AI model to the terminal device through the network device.

[0065] In the embodiments of the present application, one device in the communication system can send a signal to another device or receive a signal from another device. The signal can include information, configuration information, data, etc. The device can also be referred to as a communication device, a communication module, a node, a communication node, an entity, a network entity, a network element, etc. The embodiments of the present application take the communication between a first device and a second device as an example for illustration.

[0066] The first device can collect (or described as collect, measure, etc.) the training data of the AI model. Alternatively, the first device can be a device responsible for collecting the training data. The second device can use the training data collected by the first device for analysis or training to obtain the AI model. Alternatively, the second device can be a device responsible for model training. Optionally, the second device can provide the trained AI model (trained AI model) to the first device, so that the first device uses the trained AI model of the second device for model inference to obtain the inference result.

[0067] In one implementation, the first device can be a network device, the second device can be a terminal device, and the AI model can be deployed on the terminal device side, and the terminal device can perform model training according to the training data provided by the network device.

[0068] In another implementation, the first device can be a terminal device, the second device can be a network device, and the AI model can be deployed on the network device side, and the network device can perform model training according to the training data provided by the terminal device.

[0069] Optionally, the terminal device and the network device can also be deployed with AI models on both sides, such as the terminal device deploying an AI model a and the network device deploying an AI model b. The terminal device can send the result output by the AI model a to the network device as training data for the network device to analyze or train the AI model b. Alternatively, the network device can send the result output by the AI model b to the terminal device as training data for the terminal device to analyze or train the AI model a.

[0070] Example one, taking the scenario that the first device is a terminal device, the second device is a network device, and the network device side deploys an AI model to perform model training as an example, please refer to FIG. 2, which is a schematic diagram of an application scenario provided by an embodiment of the present application.

[0071] As shown in FIG. 2, the terminal device can be a terminal device with a perception function. Through the perception beam, the terminal device can detect the perception target in the environment and obtain the perception result. The perception result may, for example, include the speed, distance, shape, size, etc. of the perception target. Then, the terminal device can send the collected perception result to the network device as training data. After obtaining the training data, the network device can train the AI model according to the training data provided by the terminal device, obtain the trained model, and distribute the trained AI model to the terminal device.

[0072] Example two, taking the scenario that the terminal device and the network device are both deployed with AI models as an example, please refer to FIG. 3, which is another schematic diagram of an application scenario provided by an embodiment of the present application.

[0073] As shown in FIG. 3, the terminal device can be deployed with an AI model 1, and the network device can be deployed with an AI model 2. The AI model 1 and the AI model 2 can be matched with each other. It can be understood that the output of the AI model 1 can be used as the input of the AI model 2, or the output of the AI model 2 can be used as the input of the AI model 1. Optionally, the AI model 1 and the AI model 2 can be trained by the same device and then deployed on two devices, or can be distributedly trained by two devices.

[0074] In the scenario that the AI model is applied to air interface technology, the device responsible for AI model training and the device responsible for collecting training data can not be the same device, and the model training process can also involve the case of multiple parties participating together. Therefore, the security and privacy of data are problems worthy of attention in the AI model deployment process. Based on this, the present embodiment provides a communication method, which helps to improve the security of data interaction between different devices in the AI model deployment process.

[0075] The communication method provided by the embodiments of the present application is described in detail below. In the following embodiments, the device (such as the first device or the second device) can be a terminal device or a component (such as a chip or a circuit) of the terminal device, or can also be a network device or a component (such as a chip or a circuit) of the network device. The embodiments of the present application are uniformly described here, and will not be described again later.

[0076] Please refer to FIG. 4, which is a flowchart of a communication method provided by an embodiment of the present application. As shown in FIG. 4, the method can include but is not limited to the following steps:

[0077] S401, the first device sends first information to the second device. Correspondingly, the second device receives the first information from the first device.

[0078] The first information can indicate the usage range of the training data. The usage range of the training data can include one or more of the data validity time of the training data, the available model, the available function, and the available device. The available model can be understood as an AI model that allows the use of training data, the available function can be understood as an AI function that allows the use of training data, and the available device can be understood as a device that allows the use of training data. The first device can be a device that provides training data, and the second device can be a device that performs model training. The first device can limit the usage range of the training data by sending the data usage range corresponding to the training data to the second device, thereby facilitating the improvement of the security of the training data.

[0079] In an implementation manner, the first information can explicitly indicate the usage range of the training data. Optionally, the manner in which the first information explicitly indicates the usage range of the training data can be that the first information includes one or more of the data validity time of the training data, the available model, the available function, and the available device. It can be understood that the content of the first information includes one or more of the data validity time of the training data, the available model, the available function, and the available device. For example, the first information can include the available model of the training data, and specifically can include the model type or the model identifier (ID) of the available model, such as the model type can be the channel state information (CSI) compression feedback type, the intelligent transceiver type, etc.; the model ID can be model 1, model 2, model 3, etc. For another example, the first information can include the available function of the training data, such as the available function can be the CSI related function (including CSI acquisition, CSI compression feedback, etc.), the positioning function, the ranging function, the speed measurement function, the environmental imaging function, etc.

[0080] Optionally, the first information can further include a data privacy level of the training data. By distinguishing the data privacy levels of different training data, it is beneficial to improve data security in detail. For example, the data privacy level of the training data can be divided into three levels: a first privacy level, a second privacy level, and a third privacy level. The first privacy level is higher than the second privacy level, and the second privacy level is higher than the third privacy level. The first privacy level can also be described as a high privacy level, the second privacy level can also be described as a medium privacy level, and the third privacy level can also be described as a low privacy level. It can be understood that the embodiments of the present application take the data privacy level as an example to be divided into three levels for description, but it does not constitute a specific limitation to the present application.

[0081] The training data of the first privacy level can be training data related to user privacy, which can include user information that can be extracted from the data. For example, the position, heart rate, posture, distance, speed, angle, and the like of the user extracted from the perception signal. The training data of the second privacy level can be data that only the receiving end and the sending end can know, and other devices cannot use. For example, for a bilateral AI model, such as CSI compression feedback based on deep learning, an intelligent transceiver, the AI model adopted by the terminal device can contain the privacy data of the terminal device manufacturer, and the AI model adopted by the network device can contain the privacy data of the network device manufacturer; for CSI compression feedback based on deep learning, the terminal device can feed back the compressed CSI output by the AI model A to the network device, and the network device can use the AI model B to recover the compressed CSI; for an intelligent transceiver, the signal sent by the sending end can only be decoded by the corresponding receiving end. The training data of the third privacy level can be training data unrelated to user privacy, which can include data obtained by measurement, such as signal strength, waveform, frequency band, bandwidth, perception parameter, communication parameter, gradient parameter of model interaction in the process of federated learning training, and the like.

[0082] In some embodiments, the higher the data privacy level of the training data, the smaller the scope of use of the training data, for example, one or more of the shorter data validity time, the smaller number of available models, the smaller number of available functions, and the smaller number of available devices. Taking the available devices as an example, the number of available devices of the training data of the first privacy level can be p, the number of available devices of the training data of the second privacy level can be k, and p can be less than k. Taking the data validity time as another example, the data validity time of the training data of the first privacy level can be a, the data validity time of the training data of the second privacy level can be b, and a can be less than b. Optionally, the data validity time of the training data can be indicated in an explicit manner. For example, the data validity time of the training data can be directly indicated as n time slots. Alternatively, the data validity time of the training data can also be indicated in an implicit manner. For example, there can be a correspondence between the data type of the training data and the data validity time of the training data. Assuming that the data validity time corresponding to the positioning data can be c time slots, and the data validity time corresponding to the breathing data can be b time slots, and c can be greater than b. In this way, by indicating the data type of the training data, the data validity time of the training data can be implicitly indicated.

[0083] In the case where the first information includes the data privacy level of the training data, the scope of use of the training data can be indirectly indicated by the data privacy level. The data privacy level can indicate part of the scope of use of the training data, or can indicate the entire scope of use of the training data. Optionally, the first information can include one or more of the data privacy level, the data validity time, the available models, the available functions, and the available devices of the training data.

[0084] In another implementation manner, the first information can implicitly indicate the scope of use of the training data. Optionally, the manner in which the first information implicitly indicates the scope of use of the training data can be that the first information includes the data type of the training data and / or the model information of the AI model. At least one of the data type of the training data and the model information of the AI model can have a correspondence with the scope of use of the training data. Optionally, the correspondence can be pre-configured or pre-defined. The model information of the AI model can include one or more of the model identifier, the model function, the data volume, and the input / output of the AI model. That is, the content of the first information can not include the scope of use of the training data, but include the data type of the training data and / or the model information of the AI model. The second device can obtain the scope of use of the training data according to the data type of the training data and / or the model information of the AI model. Taking the model information of the AI model including the model identifier of the AI model as an example, the data type of the training data, the model identifier of the AI model, and the scope of use of the training data can have a correspondence as shown in Table 1:

[0085] Table 1

[0086] If the data type of the training data is a1 and the model identifier of the AI model is b1, it can be determined that the use range of the training data is c1; if the data type of the training data is a2 and the model identifier of the AI model is b2, it can be determined that the use range of the training data is c2.

[0087] For example, the data type of the training data can be a CSI type, the model identifier of the AI model indicates that the AI model is a CSI compression feedback model, and the use range of the training data corresponding to the AI model can be an AI model of a specified function, such as an AI model of a CSI function (which can include a channel estimation AI model, a pilot design AI model, a CSI compression feedback AI model, a beam management AI model, etc.), that is, the AI model of the CSI function can all use the training data of the CSI type. Alternatively, the use range of the training data can also be an AI model of a specified type, such as a CSI compression feedback AI model, a CSI compression AI model, a CSI feedback AI model, and a channel prediction AI model. Alternatively, the use range of the data can also be an AI model of a specified model ID, such as model 1, model 2, and model 3.

[0088] Optionally, the first information can be transmitted through radio resource control (RRC) layer signaling, a medium access control (MAC) layer MAC control element (MAC CE), uplink control information (UCI) of a physical uplink control channel (PUCCH) of a physical layer, an uplink data shared channel (PUSCH), or a random access procedure.

[0089] In some embodiments, the training data and the usage range of the training data can be carried in the same signaling or in different signaling. For the case that the training data and the usage range of the training data are carried in the same signaling, the first information can further include the training data. For the case that the training data and the usage range of the training data are carried in different signaling, the first information can further include a data identifier of the training data, so as to identify which data the usage range of the first information specifically indicates. In this case, the first device can send the training data to the second device before sending the first information, or send the training data to the second device after sending the first information. Correspondingly, the second device can receive the training data from the first device. Alternatively, the training data can also be sent by another device other than the first device, that is, the device providing the training data and the device indicating the usage range of the training data can be different devices. For example, for the case that the terminal device performs model training, the usage range of the training data can be indicated by the network device, and the training data used by the terminal device for performing model training can be provided by other terminal devices in the cell.

[0090] In some embodiments, the first device can send the training data to the second device based on the transmission resource. Correspondingly, the second device can receive the training data from the first device based on the transmission resource. Wherein, the higher the data privacy level of the training data is, the higher the transmission efficiency of the transmission resource required is. That is, the training data with a higher data privacy level needs to be transmitted in a shorter time, which is beneficial to improve the security of data transmission.

[0091] Optionally, in the case that the first device is a terminal device and the second device is a network device, the second device can configure the transmission resource based on the data privacy level of the training data. For example, assuming that the data privacy level of the training data C is higher than that of the training data D, the second device configures the transmission resource for the training data C to indicate that the training data C needs to be transmitted within a first time period, and configures the transmission resource for the training data D to indicate that the training data D needs to be transmitted within a second time period, then the time length of the first time period can be less than the time length of the second time period.

[0092] In some embodiments, after the first device sends the training data to the second device, the first device can delete the training data. That is, the first device can not save the training data, thereby being beneficial to improve the privacy of the training data. Optionally, the first device can delete the training data according to a preconfigured deletion policy, which can include, for example, deleting the training data at the moment when the first device successfully sends the training data, or deleting the training data within a preset time period after the first device successfully sends the training data, or deleting the training data in the case that the first device receives deletion indication information from the second device.

[0093] S402, the second device performs model training on the AI model based on the use range of the training data and the training data, to obtain a trained AI model.

[0094] After the second device obtains the use range of the training data and the training data, the second device can perform model training on the AI model based on the use range of the training data and the training data, to obtain a trained AI model. The trained AI model can be understood as an AI model that meets the model convergence condition.

[0095] Optionally, the AI model on which the second device performs model training can be provided by the first device. The first device can send the AI model to the second device. Correspondingly, the second device can receive the AI model from the first device.

[0096] In some embodiments, the second device needs to obtain the trained AI model within the data valid time of the training data. That is, the time at which the second device obtains the trained AI model cannot exceed the data valid time of the training data. Optionally, the second device can independently perform model training on the AI model to obtain the trained AI model, or can perform model training on the AI model in cooperation with other devices to obtain the trained AI model.

[0097] For the case where the second device independently performs model training, the time from the start time at which the second device obtains the training data to the end time at which the second device obtains the trained AI model does not exceed the data valid time of the training data. Optionally, after the second device obtains the training data, the second device can perform data preprocessing on the training data before starting model training. The data preprocessing can include one or more of the following: splitting, combining, transforming, discretizing, normalizing, outlier processing, redundant value processing, missing value processing, dimension increasing or decreasing, data augmentation, etc. For example, if the time length of the data valid time of the training data is t1, the time length of the data preprocessing performed by the second device on the training data is t2, and the time length of the model training performed using the training data after data preprocessing is t3, then (t2+t3)≤t1.

[0098] For the case that the second device cooperates with other devices (e.g., the third device) to perform model training, the third device should belong to the use range of the training data. Optionally, the second device can send the training data to the third device. Correspondingly, the third device can receive the training data from the second device. The sum of the time that the second device performs model training and the time that the third device performs model training can be less than or equal to the data valid time of the training data. For example, assuming that the use range of the training data indicates that the available devices of the training data include device 1, device 2, and device 3, the second device is device 1, and the third device can be any one of device 2 or device 3. If the time length of the valid time of the training data is t4, the time length that the second device uses the training data to perform model training is t5, and the time length that the third device performs model training is t6, then (t5+t6)≤t4. Alternatively, both the second device and the third device satisfy that the time of performing model training is less than or equal to the valid time of the training data, i.e., max(t5, t6)≤t4.

[0099] In some embodiments, after the second device obtains the trained AI model, the second device can perform model inference based on the trained AI model to obtain a model inference result. Alternatively, the second device can send the trained AI model to the first device. Correspondingly, the first device can receive the trained AI model from the second device. Then, the first device can perform model inference based on the trained AI model to obtain a model inference result.

[0100] Optionally, after the second device obtains the model inference result, or after the second device sends the trained AI model to the first device, the second device can also delete the trained AI model and / or the training data. In this way, it is helpful to protect the privacy of data.

[0101] In an implementation manner, the second device can actively delete the trained AI model and / or the training data based on a reference time. Optionally, the reference time can be the data valid time of the training data, and the second device can delete the trained AI model and / or the training data at the end of the data valid time. Alternatively, the reference time can be the cell residence time of the terminal device. In the case that the first device is a terminal device and the second device is a network device, the second device can delete the trained AI model and / or the training data at the moment when the terminal device leaves the cell coverage. Alternatively, the reference time can be the completion time of the training task. The process that the second device performs model training can be understood as the process of performing the training task. That is, the second device can delete the training data at the end of the model training.

[0102] In another implementation, the second device can delete the trained AI model and / or the training data in response to the number of times of using the training data reaching a threshold. For example, the threshold can be set to 3, and if the second device has used the training data 3 times, the trained AI model and / or the training data can be deleted.

[0103] In yet another implementation, the second device can delete the trained AI model and / or the training data upon receiving a deletion request message from the first device. The first device can send the deletion request message to the second device. Accordingly, the second device can receive the deletion request message from the first device. In response to the deletion request message, the second device can delete the trained AI model and / or the training data. Optionally, after completing the deletion action, the second device can send a deletion response message to the first device to notify the first device that the deletion of the trained AI model and / or the training data has been completed.

[0104] In the embodiments of the present application, the trained AI model is obtained by performing model training using the training data according to the usage range of the training data, which is beneficial to protect the privacy of the training data and improve the security and reliability of data interaction between different devices.

[0105] The above embodiments introduce the overall process of the communication method provided by the embodiments of the present application. The communication method provided by the embodiments of the present application is introduced from the perspective of specific application.

[0106] In some embodiments, the embodiments of the present application can be applied to the scene of model updating. For example, referring to FIG. 5, which is a flowchart of another communication method provided by the embodiments of the present application. As shown in FIG. 5, the method can include but is not limited to the following steps:

[0107] S501, the first device sends a model update request to the second device. Accordingly, the second device receives the model update request from the first device.

[0108] The model update request can be used to request to update the AI model. The first device can send the model update request to the second device in the case that the model accuracy of the current AI model is less than or equal to the accuracy threshold, or in the case that the model accuracy of the current AI model does not meet the inference requirement.

[0109] Optionally, the model update request can include one or more of the following: training data, usage range indication information, AI model training precision, maximum model distribution time. The usage range indication information can explicitly or implicitly indicate the usage range of the training data. Here, the specific implementation of the usage range indication information explicitly or implicitly indicating the usage range of the training data is the same as the implementation process of the first information explicitly or implicitly indicating the usage range of the training data in the embodiment shown in FIG. 4, and specific reference can be made to the related description in the embodiment shown in FIG. 4, which will not be repeated here.

[0110] Optionally, the usage range indication information can also be sent independently of the model update request. In the RRC connected state, the usage range indication information can be carried by the RRC configuration signaling. In the RRC non-connected state, the usage range indication information can be carried by the information 3 (message 3, Msg3) in the random access process.

[0111] Optionally, the device providing the training data and the device sending the model update request can also be different devices. For example, the model update request can be sent by the first device to the second device, and the training data can be sent by the third device to the second device.

[0112] S502, the second device determines whether to perform model update.

[0113] After the second device receives the update request message, it can determine whether the conditions for performing model update are met. Optionally, the second device can determine whether there is an AI model for model update, or it can also determine whether there is sufficient computing power resource for performing model update. For example, for a smart transceiver, both the first device and the second device need to have an AI model. In this case, the second device can determine whether it has a candidate AI model that matches the AI model of the first device. If the second device does not have a candidate AI model that matches the AI model of the first device, it cannot complete the model update.

[0114] S503, the second device sends a model update response to the first device. Correspondingly, the first device receives the model update response from the second device.

[0115] If the second device has the conditions for performing model update, the model update response can be used to notify the confirmation of performing model update. If the second device does not have the conditions for performing model update, the model update response can be used to notify the reason why the model update cannot be performed.

[0116] In some embodiments, for the case that the first device is a terminal device and the second device is a network device, the second device can further configure a transmission resource, which can be used to transmit the data related to updating the AI model. The second device can configure the transmission resource based on a data privacy level of the training data used to train the AI model. The higher the data privacy level of the training data, the higher the transmission efficiency of the transmission resource to be configured.

[0117] Optionally, S504, the first device sends the AI model and / or the training data to the second device. Correspondingly, the second device receives the AI model and / or the training data from the first device.

[0118] In the case that the second device does not have the AI model used to perform the model updating, the second device can send a model request message to the first device. After receiving the model request message, the first device can send the AI model to the second device in response to the model request message.

[0119] Optionally, the first device can provide the training data to the second device. Alternatively, the device sending the training data can also be another device other than the first device. For example, the second device can send a data request message to a third device, and the third device can send the training data to the second device in response to the data request message. For the case that multiple devices are required to provide the training data for the model training, for example, multiple terminal devices in a cell are required to provide the training data for the training of a positioning model, the second device can also send data request messages to multiple devices and receive the training data from the multiple devices.

[0120] S505, the second device performs model training on the AI model based on the usage range of the training data and the training data, to obtain a trained AI model.

[0121] The specific implementation process of step S505 can be referred to the related description of step S402 in the embodiment shown in FIG. 4 described above, which will not be repeated here.

[0122] Optionally, S506, the second device sends the trained AI model to the first device. Correspondingly, the first device receives the trained AI model from the second device.

[0123] Optionally, the second device can send the trained AI model to the first device, so that the first device performs model inference based on the updated trained AI model to obtain a model inference result.

[0124] S507, the second device deletes the trained AI model and / or the training data.

[0125] In an implementation, the second device can configure a timer, and delete the trained AI model and / or the training data at a time when the timer expires. The time length of the timer can be the time length of the reference time. Optionally, the reference time can be the data validity time of the training data, or the cell residence time of the terminal device, or the completion time of the training task.

[0126] In another implementation, the second device can configure a counter, and delete the trained AI model and / or the training data when the counter reaches a maximum count (max count). The counter can be used to record the number of times the training data is used. That is, when the number of times the training data is used reaches a threshold, the second device can delete the trained model and / or the training data.

[0127] Optionally, after the second device completes deleting the trained AI model and / or the training data, the second device can send a deletion response message to the first device to inform the first device that the trained AI model and / or the training data has been deleted.

[0128] In some embodiments, the embodiment shown in FIG. 5 can be applied to model updating between terminal devices, and the first device and the second device can both be terminal devices; or can be applied to model updating between a terminal device and a network device, and the first device can be a terminal device and the second device can be a network device.

[0129] For example, taking the first device as a terminal device and the second device as a network device as an example, the model updating scenarios to which the embodiment shown in FIG. 5 can be applied include but are not limited to the following cases: ① the terminal device provides training data, the network device is responsible for performing model training, and sends the trained AI model to the terminal device, and the terminal device performs model inference; ② the terminal device provides training data, and the network device is responsible for performing model training and performing model inference; ③ the terminal device and the network device can both perform model training, and the terminal device and the network device share training data.

[0130] For another example, taking the first device as a network device and the second device as a terminal device as an example, the model updating scenarios to which the embodiment shown in FIG. 5 can be applied include but are not limited to the following cases: ① the network device provides training data, the terminal device is responsible for performing model training, and sends the trained AI model to the network device, and the network device performs model inference; ② the terminal device and the network device can both perform model training, and the terminal device and the network device share training data.

[0131] In some implementation scenarios, the second device can also be a core network element, or other network element for AI model training. Optionally, the device providing the training data can include the first device and a third device, for example, the first device is a terminal device, and the third device is a network device. For example, refer to FIG. 6, which is a flow diagram of another communication method provided by the embodiments of the present application. As shown in FIG. 6, the method can include but is not limited to the following steps:

[0132] S601, the core network element sends a data request message #1 to the network device. Correspondingly, the network device receives the data request message #1 from the core network element.

[0133] The data request message #1 can be used to request training data. Optionally, the training data can be provided by a terminal device and / or a network device.

[0134] S602, the network device sends a data request message #2 to the terminal device. Correspondingly, the terminal device receives the data request message #2 from the network device.

[0135] If the device providing the training data includes the terminal device, after receiving the data request message #1, the network device can send the data request message #2 to the terminal device.

[0136] S603, the terminal device sends a data response message #1 to the network device. Correspondingly, the network device receives the data response message #1 from the terminal device.

[0137] The data response message #1 can include information indicating that the terminal device agrees to provide the training data.

[0138] S604, the network device sends a data response message #2 to the core network element. Correspondingly, the core network element receives the data response message #2 from the network device.

[0139] The data response message #2 can include information indicating that the terminal device and / or the network device agrees to provide the training data.

[0140] Optionally, the data response message #2 can also include usage range indication information. The usage range indication information can explicitly or implicitly indicate the usage range of the training data. Here, the specific implementation of the usage range indication information explicitly or implicitly indicating the usage range of the training data is the same as the implementation process of the first information explicitly or implicitly indicating the usage range of the training data in the embodiment shown in FIG. 4, and specific reference can be made to the related description in the embodiment shown in FIG. 4, which will not be repeated here.

[0141] Optionally, S605, the terminal device sends the AI model and / or the training data to the network device. Accordingly, the network device receives the AI model and / or the training data from the terminal device.

[0142] Optionally, S606, the network device sends the AI model and / or the training data to the core network element. Accordingly, the core network element receives the AI model and / or the training data from the network device.

[0143] S607, the core network element performs model training on the AI model based on the usage range of the training data and the training data, to obtain a trained AI model.

[0144] The specific implementation process of step S607 can refer to the related description of step S402 in the embodiment shown in FIG. 4 described above, and will not be repeated here.

[0145] Optionally, S608, the core network element sends the trained AI model to the network device. Accordingly, the network device receives the trained AI model from the core network element.

[0146] If the network device and / or the terminal device is responsible for performing model inference, the core network element can send the trained AI model to the network device.

[0147] Optionally, S609, the network device sends the trained AI model to the terminal device. Accordingly, the terminal device receives the trained AI model from the network device.

[0148] If the terminal device is responsible for performing model inference, the network device sends the trained AI model to the terminal device after receiving the trained AI model.

[0149] S610, the core network element deletes the trained AI model and / or the training data.

[0150] The specific implementation process of step S610 can refer to the related description of step S507 in the embodiment shown in FIG. 5 described above, and will not be repeated here.

[0151] In the embodiments of the present application, on the one hand, by specifying the usage range of the training data, the device responsible for performing model training performs model training within the usage range of the training data, which is conducive to enhancing the privacy protection of the training data; on the other hand, after obtaining the trained model, the device responsible for performing model training can delete the training data and / or the trained model, which is conducive to further protecting the security of the data. By implementing the embodiments of the present application, it is conducive to improving the security and reliability of data interaction between different devices in the AI model deployment process.

[0152] The foregoing describes the method embodiments provided by the present application. In order to better implement the foregoing scheme of the embodiments of the present application, the embodiments of the present application further provide a corresponding communication device.

[0153] In some embodiments, the communication device comprises a hardware structure and / or a software module for implementing the above functions. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of the examples described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a certain function is implemented in hardware or computer software driven hardware depends on the specific application of the technical solution and the design constraints. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0154] The embodiments of the present application can divide the functional modules of the communication device according to the above method examples. For example, each functional module can be divided according to each function, or two or more functions can be integrated into one processing module. The integrated module can be implemented in the form of hardware or software functional module. It should be noted that the division of modules in the embodiments of the present application is illustrative, and is only a logical functional division. Actual implementation can have another division manner.

[0155] Please refer to FIG. 7, which is a structural schematic diagram of a communication device provided by an embodiment of the present application. The communication device 70 can be the first device or the second device in the above method embodiments, or a device (such as a chip, or a chip system, or a circuit) in the first device or the second device. As shown in FIG. 7, the communication device 70 at least includes a communication unit 701 and a processing unit 702.

[0156] For the case that the communication device is used to implement the functions of the second device in the embodiments of the present application:

[0157] The communication unit 701 is configured to receive first information. The first information indicates a usage range of training data. The usage range of the training data includes one or more of a data valid time, a usable model, a usable function, and a usable device of the training data.

[0158] The processing unit 702 is configured to perform model training on an AI model based on the usage range of the training data and the training data, to obtain a trained AI model.

[0159] In a possible implementation, the communication unit 701 is further configured to receive or send the training data.

[0160] In a possible implementation, the first information explicitly indicates the use range of the training data; the first information includes one or more of a data validity time, an available model, an available function, and an available device of the training data.

[0161] In a possible implementation, the first information implicitly indicates the use range of the training data; the first information includes data type of the training data and / or model information of the AI model; at least one of the data type of the training data and the model information of the AI model has a corresponding relationship with the use range of the training data; the model information of the AI model includes one or more of a model identifier, a model function, a data volume, and an input / output of the AI model.

[0162] In a possible implementation, the first information further includes a data privacy level of the training data.

[0163] In a possible implementation, the data privacy level is one of a first privacy level, a second privacy level, or a third privacy level; the first privacy level is higher than the second privacy level, and the second privacy level is higher than the third privacy level.

[0164] In a possible implementation, the higher the data privacy level, the smaller the indicated use range of the training data.

[0165] In a possible implementation, the communication unit 701 is further configured to receive or send the training data based on a transmission resource; the higher the data privacy level, the higher the transmission efficiency of the transmission resource required.

[0166] In a possible implementation, the processing unit 702 is further configured to delete the trained AI model and / or the training data based on a reference time; the reference time is a data validity time of the training data, a cell residence time of the terminal device, or a completion time of the training task.

[0167] In a possible implementation, the processing unit 702 is further configured to delete the trained AI model and / or the training data in response to the use number of the training data reaching a threshold.

[0168] In a possible implementation, the communication unit 701 is further configured to receive a deletion request message; the processing unit 702 is further configured to delete the trained AI model and / or the training data in response to the deletion request message; and the communication unit 701 is further configured to send a deletion response message.

[0169] For the case where the communication apparatus is used to implement the function of the first device in the embodiments of the present application:

[0170] The communication unit 701 is configured to send first information, the first information indicating a usage range of the training data; the usage range of the training data includes one or more of a data validity time, a usable model, a usable function, and a usable device of the training data; and the usage range of the training data is used to obtain the trained AI model.

[0171] In a possible implementation, the communication unit 701 is further configured to send or receive the training data.

[0172] In a possible implementation, the first information explicitly indicates the usage range of the training data; and the first information includes one or more of the data validity time, the usable model, the usable function, and the usable device of the training data.

[0173] In a possible implementation, the first information implicitly indicates the usage range of the training data; the first information includes data type of the training data and / or model information of an AI model; at least one of the data type of the training data and the model information of the AI model has a corresponding relationship with the usage range of the training data; and the model information of the AI model includes one or more of a model identifier, a model function, a data volume, and an input / output of the AI model.

[0174] In a possible implementation, the first information further includes a data privacy level of the training data. In a possible implementation, the data privacy level is one of a first privacy level, a second privacy level, or a third privacy level; the first privacy level is higher than the second privacy level, and the second privacy level is higher than the third privacy level.

[0175] In a possible implementation, the higher the data privacy level, the smaller the indicated usage range of the training data.

[0176] In a possible implementation, the communication unit 701 is further configured to send or receive the training data based on a transmission resource; and the higher the data privacy level, the higher the transmission efficiency of the transmission resource required.

[0177] In a possible implementation, the communication unit 701 is further configured to send a deletion request message; the deletion request is used to request deletion of the trained AI model and / or the training data; and a deletion response message is received.

[0178] For more detailed descriptions of the communication unit 701 and the processing unit 702, reference can be made to the related descriptions of the first device and the second device in the above method embodiments, which are not repeated here.

[0179] Please refer to FIG. 8, which is a structural schematic diagram of another communication apparatus provided in the embodiments of the present application. As shown in FIG. 8, the communication apparatus 80 can include one or more processors 801, which can also be referred to as processing units, and can implement certain control functions. The processor 801 can be a general processor or a special-purpose processor, etc. For example, it can be a baseband processor or a central processing unit. The baseband processor can be used to process communication protocols and communication data, and the central processing unit can be used to control the communication apparatus (such as a base station, a baseband chip, a terminal, a terminal chip, a DU or a CU, etc.), execute software programs, and process data of the software programs.

[0180] In an alternative design, the processor 801 can also store instructions 803 and / or data, which can be executed by the processor to enable the communication apparatus 80 to perform the methods described in the above method embodiments.

[0181] In another alternative design, the processor 801 can include a transceiver unit for implementing receiving and transmitting functions. For example, the transceiver unit can be a transceiver circuit, or an interface, or an interface circuit, or a communication interface. The transceiver circuit, interface or interface circuit for implementing receiving and transmitting functions can be separate or integrated together. The above transceiver circuit, interface or interface circuit can be used for reading and writing codes / data, or the above transceiver circuit, interface or interface circuit can be used for signal transmission or transfer.

[0182] In yet another possible design, the communication apparatus 80 can include a circuit that can implement the functions of sending or receiving or communicating in the foregoing method embodiments.

[0183] Optionally, the communication apparatus 80 can include one or more memories 802, which can store instructions 804 and / or data that can be executed by the processor to enable the communication apparatus 80 to perform the methods described in the above method embodiments. Optionally, the memory can also store data. Optionally, the processor can also store instructions and / or data. The processor and the memory can be separately arranged or integrated together. For example, the correspondence described in the above method embodiments can be stored in the memory or in the processor.

[0184] Optionally, the communication apparatus 80 can further include a transceiver 805 and / or an antenna 806. The processor 801 can be referred to as a processing unit and can control the communication apparatus 80. The transceiver 805 can be referred to as a transceiving unit, a transceiver, a transceiver circuit, a transceiving apparatus or a transceiving module, etc., and can be used to implement receiving and transmitting functions.

[0185] Optionally, the communication apparatus 80 in the embodiments of the present application can be used to perform the methods described in the above method embodiments.

[0186] In one embodiment, the communication apparatus 80 can be a first device or a device (e.g., a chip, or a chip system, or a circuit) in the first device. When computer program instructions stored in the memory 802 are executed, the transceiver 805 is configured to perform operations performed by the communication unit 701 in the above-described embodiments. The transceiver 805 is further configured to send information to other communication apparatuses outside the communication apparatus. The first device or the device in the first device can also be configured to perform various methods performed by the first device in the above-described method embodiments, which will not be repeated.

[0187] In one embodiment, the communication apparatus 80 can be a second device or a device (e.g., a chip, or a chip system, or a circuit) in the second device. When computer program instructions stored in the memory 802 are executed, the transceiver 805 is configured to perform operations performed by the communication unit 701 in the above-described embodiments. The second device or the device in the second device can also be configured to perform various methods performed by the second device in the above-described method embodiments, which will not be repeated.

[0188] The embodiments of the present application further provide a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the method provided by the above-described method embodiments.

[0189] The embodiments of the present application further provide a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the method provided by the above-described method embodiments.

[0190] The embodiments of the present application further provide a computer program product, which, when executed on a computer or a processor, causes the computer or the processor to perform one or more steps in the above-described any one method. The components of the devices involved in the above description can be stored in a computer readable storage medium if they are realized in the form of software function units and sold or used as independent products.

[0191] The embodiments of the present application further provide a chip system, which comprises at least one processor and an interface circuit, the interface circuit and the at least one processor are interconnected through a line, and the at least one processor is configured to execute a computer program or an instruction, so that any one of the steps described in the above method embodiments is executed partially or entirely. In a possible implementation, the chip system can further comprise at least one memory, the interface circuit, the at least one memory and the at least one processor are interconnected through a line, and the at least one memory stores an instruction, which is executed by the processor, so that any one of the steps described in the above method embodiments is executed partially or entirely. The chip system can be composed of a chip, or can comprise a chip and other discrete devices.

[0192] The embodiments of the present application further provide a communication system, which comprises a first device and a second device, and the specific description can refer to the method shown in the above method embodiments.

[0193] It should be appreciated that the memory mentioned in the embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a hard disk drive (HDD), a solid-state drive (SSD), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically EPROM (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM) used as an external cache. By way of example, but not limitation, many forms of RAM can be used, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous dynamic RAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchlink DRAM (SLDRAM), and direct rambus RAM (DR RAM). The memory is any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited to this. The memory in the embodiments of the present application can also be a circuit or any other device capable of realizing a storage function, used for storing program instructions and / or data.

[0194] It should also be understood that the processor mentioned in the embodiments of the present application can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0195] It should be noted that when the processor is a general processor, DSP, ASIC, FPGA or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, the memory (storage module) is integrated in the processor. It should be noted that the memory described herein is intended to include but not limited to these and any other suitable type of memory.

[0196] It should be understood that the size of the sequence number of the above-mentioned processes does not mean the order of execution in various embodiments of the present application. The execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0197] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in conjunction with the embodiments provided herein can be realized in electronic hardware, or a combination of computer software and electronic hardware. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0198] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working process of the above-mentioned system, device and unit can be referred to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0199] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division, and actual implementation can have another division manner. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0200] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0201] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit.

[0202] If the functions are implemented in the form of software function units and sold or used as independent products, the functions can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application or the parts of the technical solutions that essentially contribute to the prior art or the parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes various media that can store program codes, such as a U disk, a mobile hard disk, a ROM, a RAM, a magnetic disk, or an optical disk.

[0203] The steps in the method embodiments of the present application can be adjusted, combined, and reduced in sequence according to actual needs.

[0204] The modules / units in the device embodiments of the present application can be combined, divided, and reduced according to actual needs.

[0205] The above only discloses one preferred embodiment of the present application, which is only a part of the embodiments of the present application, and cannot limit the scope of the rights of the present application.

Claims

1. A communication method characterized by comprising: The method comprises: receiving first information; the first information indicates a use range of training data; the use range of the training data comprises one or more of a data validity time, available models, available functions, and available devices of the training data; based on the use range of the training data and the training data, performing model training on an artificial intelligence (AI) model to obtain a trained AI model.

2. The method of claim 1, wherein, The method further comprises: receiving or sending the training data.

3. The method of claim 1 or 2, wherein, The first information explicitly indicates the use range of the training data; the first information comprises one or more of a data validity time, available models, available functions, and available devices of the training data.

4. The method of claim 1 or 2, wherein, The first information implicitly indicates the use range of the training data; the first information comprises a data type of the training data and / or model information of the AI model; at least one of the data type of the training data and the model information of the AI model corresponds to the use range of the training data; the model information of the AI model comprises one or more of a model identifier, a model function, a data volume, and an input / output of the AI model.

5. The method of claim 3, wherein, The first information further comprises a data privacy level of the training data.

6. The method of claim 5, wherein, The data privacy level is one of a first privacy level, a second privacy level, or a third privacy level; the first privacy level is higher than the second privacy level, and the second privacy level is higher than the third privacy level.

7. The method of claim 5 or 6, wherein, The higher the data privacy level, the smaller the indicated use range of the training data.

8. The method according to any one of claims 5 to 7, wherein, The receiving or sending of the training data comprises: receiving or sending the training data based on transmission resources; the higher the data privacy level, the higher the transmission efficiency of the transmission resources required.

9. The method according to any one of claims 1 to 8, wherein, The method further comprises: deleting the trained AI model and / or the training data based on a reference time; the reference time is a data validity time of the training data, a cell residence time of a terminal device, or a completion time of a training task; or in response to a use number of the training data reaching a threshold, deleting the trained AI model and / or the training data; or receiving a deletion request message; in response to the deletion request message, deleting the trained AI model and / or the training data; sending a deletion response message.

10. A communication method characterized by comprising: The method comprises: sending first information, the first information indicating a use range of training data; the use range of the training data comprises one or more of a data validity time, available models, available functions, and available devices of the training data; the use range of the training data is used to obtain a trained AI model.

11. The method of claim 10, wherein, The method further comprises: sending or receiving the training data.

12. The method of claim 10 or 11, wherein, The first information explicitly indicates the use range of the training data; the first information comprises one or more of a data validity time, available models, available functions, and available devices of the training data.

13. The method of claim 10 or 11, wherein, The first information implicitly indicates a use range of the training data; the first information includes data type of the training data and / or model information of the AI model; at least one of the data type of the training data and the model information of the AI model has a corresponding relationship with the use range of the training data; the model information of the AI model includes one or more of model identification, model function, data volume, and input and output of the AI model.

14. The method of claim 12, wherein, The first information further includes a data privacy level of the training data.

15. The method of claim 14, wherein, The data privacy level is one of a first privacy level, a second privacy level, or a third privacy level; the first privacy level is higher than the second privacy level, and the second privacy level is higher than the third privacy level.

16. The method of claim 14 or 15, wherein, The higher the data privacy level, the smaller the indicated use range of the training data.

17. The method of any one of claims 14-16, wherein, The sending or receiving the training data comprises: Sending or receiving the training data based on a transmission resource; the higher the data privacy level, the higher the transmission efficiency of the transmission resource required.

18. The method of any one of claims 10-17, wherein, The method further comprises: Sending a deletion request message; the deletion request is used to request deletion of the trained AI model and / or the training data; Receiving a deletion response message.

19. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer programs or computer instructions, when the computer programs or computer instructions are executed by the processor, the second device executes the method in any one of claims 1-9, or the first device executes the method in any one of claims 10-18.

20. A chip system, characterized by The at least one processor, the at least one memory, and the interface circuit are interconnected by a line, the at least one memory stores instructions; when the instructions are executed by the processor, the second device executes the method in any one of claims 1-9, or the first device executes the method in any one of claims 10-18.

21. A communication system, characterized by The first device and the second device are included, the second device is used to execute the method in any one of claims 1-9, and the first device is used to execute the method in any one of claims 10-18.

Images