Model transmission method, device, and storage medium

Abstract

Embodiments of the present application relate to the technical field of communications, and provide a model transmission method, a device, and a storage medium. A network device broadcasts first information, the first information being used for indicating model data of an AI algorithm model, and a terminal obtains the model data from the broadcast first information. Thus, by employing a point-to-multipoint broadcast method, repeated transmission of the AI algorithm model is reduced, thereby reducing waste of air interface resources.

Description

Model transmission method, device and storage medium

[0001] This application claims priority to the Chinese patent application filed with the State Intellectual Property Office of China on February 8, 2024, with application number 202410178096.4 and application name “Model Transmission Method, Device and Storage Medium”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the field of communication technology, and in particular to a model transmission method, device, and storage medium. Background Art

[0003] In communication systems, artificial intelligence (AI) algorithms are used to replace traditional algorithms for local air interface transmission to improve algorithm performance in complex scenarios.

[0004] In a communication system, terminal devices can serve as execution nodes for AI algorithm models, using them to perform operations. Network devices or servers can serve as training nodes for AI algorithm models, providing a large amount of computing power for training them. Because the use and training of AI algorithm models are not performed on the same node, the AI ​​algorithm model needs to be transmitted. The air interface transmission of AI algorithm models is still under discussion, and the solutions being discussed include network devices sending AI models to multiple terminal devices via radio resource control (RRC) messages or user plane (UP) data.

[0005] However, when different terminal devices use the same AI model, the network device sends the AI ​​model to the terminal device through RRC messages or UP data, which causes the same data to be transmitted repeatedly, resulting in a waste of air interface resources. Summary of the Invention

[0006] The embodiments of the present application provide a model transmission method, device, and storage medium, which are applied to the field of communication technology. By broadcasting a model identifier corresponding to an AI algorithm model and model data of the AI ​​algorithm model, repeated data transmission when a network device transmits an AI algorithm model to a terminal device is reduced, thereby saving air interface resources.

[0007] In a first aspect, embodiments of the present application provide a model transmission method. The method includes broadcasting first information, where the first information is used to indicate model data of an AI algorithm model. In the embodiments, a network device broadcasts the model data of the AI ​​algorithm model, thereby reducing repeated transmission of the same AI algorithm model and saving air interface resources.

[0008] In an optional embodiment of the first aspect, the first information is further used to indicate a model identifier corresponding to the AI ​​algorithm model, where the model identifier corresponding to the AI ​​algorithm model is used to uniquely identify the AI ​​algorithm model. In this embodiment, the network device broadcasts the model identifier corresponding to the AI ​​algorithm model, thereby facilitating the terminal device to know the AI ​​algorithm model to which the broadcasted model data belongs. This improves the transmission accuracy of the AI ​​algorithm model, particularly when the network device broadcasts model data corresponding to multiple AI algorithm models.

[0009] In an optional embodiment of the first aspect, a model identifier corresponding to the AI ​​algorithm model has been allocated to a terminal device that supports the algorithm capabilities related to the AI ​​algorithm model. In this embodiment, the model identifier corresponding to the AI ​​algorithm model is allocated in advance to the terminal device that supports the algorithm capabilities related to the AI ​​algorithm model. On the one hand, this improves the rationality of the model identifier allocation. On the other hand, it enables the terminal device to obtain model data of the corresponding AI algorithm model from the broadcasted first information based on the allocated model identifier, thereby improving the data transmission accuracy of the AI ​​algorithm model.

[0010] In an optional embodiment of the first aspect, before broadcasting the first information, the method further includes: obtaining second information, the second information being used to indicate the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device; and indicating third information based on the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device, the third information being used to indicate the allocation of a model identifier corresponding to the AI ​​algorithm model. In this embodiment, the network device allocates a corresponding model identifier to the terminal device based on the algorithm capabilities related to the AI ​​algorithm model reported by the terminal device, thereby improving the accuracy of model identifier allocation and thereby improving the data transmission accuracy of the AI ​​algorithm model.

[0011] In an optional embodiment of the first aspect, the first information is SI. This embodiment provides a specific implementation of the first information, using SI that can be broadcast periodically as the first information.

[0012] In an optional embodiment of the first aspect, the first information includes multiple SIBs, wherein a first SIB among the multiple SIBs is used to indicate that a second SIB includes model data of the AI ​​algorithm model, the first SIB is one of the multiple SIBs, and the second SIB is a different SIB from the first SIB. In this embodiment, when the first information is an SI, the first SIB among the multiple SIBs can be used to indicate that the second SIB includes model data of the AI ​​algorithm model, thereby facilitating the terminal device to accurately obtain the model data of the AI ​​algorithm model from the second SIB.

[0013] In an optional embodiment of the first aspect, before broadcasting the first information, the further step includes: broadcasting fourth information, where the fourth information is used to indicate that the AI ​​algorithm model is ready for distribution and / or that the usage status of the AI ​​algorithm model has changed. In this embodiment, before broadcasting the first information, by broadcasting the fourth information indicating that the AI ​​algorithm model is ready for distribution and / or that the usage status of the AI ​​algorithm model has changed, the terminal device can promptly obtain the first information and / or promptly update the usage status of the AI ​​algorithm model.

[0014] In an optional embodiment of the first aspect, the fourth information is further used to indicate that the SI carrying the AI ​​algorithm model has changed. In this embodiment, by indicating that the SI carrying the AI ​​algorithm model has changed, the terminal device receives the next SI in a timely manner, which is applicable to the case where the first information is SI.

[0015] In an optional embodiment of the first aspect, the fourth information is: a short message of DCI scheduled using a model identifier and / or a paging identifier corresponding to the AI ​​algorithm model, and / or a MAC CE scheduled using a model identifier corresponding to the AI ​​algorithm model. This embodiment provides two implementation methods of the fourth information, one is a short message of DCI, and the other is a MAC CE. The network device transmits the fourth information to the terminal device by using a short message scheduled using a model identifier and / or a paging identifier corresponding to the AI ​​algorithm model, and / or a MAC CE scheduled using a model identifier corresponding to the AI ​​algorithm model.

[0016] In an optional embodiment of the first aspect, before broadcasting the first information, the method further includes: obtaining fifth information, where the fifth information is used to request model data of the AI ​​algorithm model. In this embodiment, the first information is broadcasted after the terminal device requests the model data, thereby saving air interface resources.

[0017] In an optional embodiment of the first aspect, the fifth information is an SI request. This embodiment provides two implementations of the fifth information, and the terminal device can request the model data from the network device through the SI request.

[0018] In an optional embodiment of the first aspect, the terminal device includes at least one of the following: a terminal device in a connected state, a terminal device in an inactive state, and a terminal device in an idle state. In this embodiment, the broadcasted first information can be obtained by the terminal device in the connected state, the terminal device in the inactive state, and the terminal device in the idle state, thereby increasing the scope of applicable terminal devices for AI algorithm model transmission.

[0019] In a second aspect, an embodiment of the present application proposes a model transmission method. The method includes: obtaining a first broadcast message, the first message being used to indicate model data of an AI algorithm model; and obtaining the model data from the first message. In this embodiment, a terminal device can obtain the model data of the AI ​​algorithm model from the first message broadcast by a network device, and the network device does not need to repeatedly broadcast the same AI algorithm model to different terminal devices, thereby saving air interface resources.

[0020] In an optional embodiment of the second aspect, the first information is further used to indicate a model identifier corresponding to the AI ​​algorithm model, where the model identifier corresponding to the AI ​​algorithm model is used to uniquely identify the AI ​​algorithm model. In this embodiment, the network device broadcasts the model identifier corresponding to the AI ​​algorithm model, thereby facilitating the terminal device to know the AI ​​algorithm model to which the broadcasted model data belongs. This improves the transmission accuracy of the AI ​​algorithm model, particularly when the network device broadcasts model data corresponding to multiple AI algorithm models.

[0021] In an optional embodiment of the second aspect, a model identifier corresponding to the AI ​​algorithm model has been assigned to a terminal device, and the terminal device supports the algorithm capabilities related to the AI ​​algorithm model. In this embodiment, the model identifier corresponding to the AI ​​algorithm model is assigned in advance to the terminal device that supports the algorithm capabilities related to the AI ​​algorithm model. On the one hand, this improves the rationality of the model identifier assignment. On the other hand, it enables the terminal device to obtain the model data of the corresponding AI algorithm model from the broadcasted first information based on the assigned model identifier, thereby improving the data transmission accuracy of the AI ​​algorithm model.

[0022] In an optional embodiment of the second aspect, before obtaining the first broadcast information, the method further includes: sending second information, the second information being used to indicate the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device; obtaining third information, the third information being used to indicate the model identifier corresponding to the allocation of the AI ​​algorithm model; obtaining the model identifier from the third information, and saving the model identifier. In this embodiment, the network device assigns a corresponding model identifier to the terminal device based on the algorithm capabilities related to the AI ​​algorithm model reported by the terminal device, thereby improving the accuracy of model identifier assignment and thereby improving the data transmission accuracy of the AI ​​algorithm model.

[0023] In an optional embodiment of the second aspect, the first information is SI. This embodiment provides a specific implementation of the first information, using SI that can be broadcast periodically as the first information.

[0024] In an optional embodiment of the second aspect, the first information includes multiple SIBs, and among the multiple SIBs, the first SIB is used to indicate that the second SIB contains model data of the AI ​​algorithm model, the first SIB is one of the multiple SIBs, and the second SIB is a different SIB from the first SIB. Obtaining the model data of the AI ​​algorithm model from the first information includes: obtaining the model data of the AI ​​algorithm model from the second SIB according to the indication of the first SIB. In this embodiment, when the first information is SI, the first SIB among the multiple SIBs can be used to indicate that the second SIB contains the model data of the AI ​​algorithm model, so that the terminal device can obtain the model data of the AI ​​algorithm model from the second SIB.

[0025] In an optional embodiment of the second aspect, before obtaining the first broadcast information, the further step includes: obtaining fourth broadcast information, where the fourth information is used to indicate that the AI ​​algorithm model is ready to be sent and / or that the usage status of the AI ​​algorithm model has changed. In this embodiment, before broadcasting the first information, by broadcasting the fourth information indicating that the AI ​​algorithm model is ready to be sent and / or that the usage status of the AI ​​algorithm model has changed, the terminal device can obtain the first information in a timely manner and / or update the usage status of the AI ​​algorithm model in a timely manner.

[0026] In an optional embodiment of the second aspect, the fourth information is further used to indicate that the SI carrying the AI ​​algorithm model has changed. In this embodiment, by indicating that the SI carrying the AI ​​algorithm model has changed, the terminal device receives the next SI in a timely manner, which is applicable to the case where the first information is SI.

[0027] In an optional embodiment of the second aspect, the fourth information is: a short message of DCI scheduled using a model identifier and / or a paging identifier corresponding to the AI ​​algorithm model, and / or a MAC CE scheduled using a model identifier corresponding to the AI ​​algorithm model. This embodiment provides two implementation methods of the fourth information, one is a short message of DCI, and the other is a MAC CE. The network device transmits the fourth information to the terminal device by using a short message scheduled using a model identifier and / or a paging identifier corresponding to the AI ​​algorithm model, or a MAC CE using a model identifier corresponding to the AI ​​algorithm model.

[0028] In an optional embodiment of the second aspect, before obtaining the first broadcast information, the method further includes: sending fifth information for requesting model data of the AI ​​algorithm model. In this embodiment, the first information is broadcasted after a request for model data is received from a terminal device, thereby conserving air interface resources.

[0029] In an optional embodiment of the second aspect, the fifth information is an SI request. This embodiment provides two implementations of the fifth information, and the terminal device can request the model data from the network device through the SI request.

[0030] In an optional embodiment of the second aspect, the first information is broadcast multiple times. After obtaining the model data from the first information, it also includes: selectively merging the model data obtained from the first information broadcast multiple times to obtain complete model data of the AI ​​algorithm model.

[0031] In an optional embodiment of the second aspect, the terminal device is in a connected state, an inactive state, or an idle state. In this embodiment, the broadcasted first information can be obtained by terminal devices in a connected state, an inactive state, or an idle state, thereby increasing the scope of applicable terminal devices for AI algorithm model transmission.

[0032] In a third aspect, an embodiment of the present application provides a network device, including: a broadcast module, used to broadcast first information, where the first information is used to indicate model data of an AI algorithm model.

[0033] In an optional embodiment of the third aspect, the first information is further used to indicate a model identifier corresponding to the AI ​​algorithm model, and the model identifier corresponding to the AI ​​algorithm model is used to uniquely identify the AI ​​algorithm model.

[0034] In an optional embodiment of the third aspect, the model identifier corresponding to the AI ​​algorithm model has been assigned to a terminal device that supports algorithm capabilities related to the AI ​​algorithm model.

[0035] In an optional embodiment of the third aspect, it further includes: a first acquisition module, used to acquire second information, the second information is used to indicate the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device; an indication module, used to indicate third information based on the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device, the third information is used to indicate the model identifier corresponding to the assigned AI algorithm model.

[0036] In an optional embodiment of the third aspect, the first information is SI.

[0037] In an optional embodiment of the third aspect, the first information includes multiple SIBs, among which the first SIB is used to indicate that the second SIB includes model data of the AI ​​algorithm model, the first SIB is one of the multiple SIBs, and the second SIB is a different SIB from the first SIB.

[0038] In an optional embodiment of the third aspect, the broadcast module can also be used to: broadcast fourth information, where the fourth information is used to indicate that the AI ​​algorithm model is ready to be sent and / or the usage status of the AI ​​algorithm model has changed.

[0039] In an optional embodiment of the third aspect, the fourth information is also used to indicate that the SI carrying the AI ​​algorithm model has changed.

[0040] In an optional embodiment of the third aspect, the fourth information is: a short message of DCI scheduled using a model identifier and / or a paging identifier corresponding to the AI ​​algorithm model, and / or a MAC CE scheduled using a model identifier corresponding to the AI ​​algorithm model.

[0041] In an optional embodiment of the third aspect, it further includes: a second acquisition module, used to obtain fifth information, and the fifth information is used to request model data of the AI ​​algorithm model.

[0042] In an optional embodiment of the third aspect, the fifth information is an SI request.

[0043] In an optional embodiment of the third aspect, the terminal device includes at least one of the following: a terminal device in a connected state, a terminal device in an inactive state, and a terminal device in an idle state.

[0044] It should be understood that the technical solution of the third aspect of the present application corresponds to the technical solution of the first aspect of the present application, and the beneficial effects achieved by each aspect and the corresponding feasible implementation methods are similar and will not be repeated.

[0045] In a fourth aspect, an embodiment of the present application proposes a terminal device, comprising: an acquisition module for acquiring broadcast first information, where the first information is used to indicate model data of an AI algorithm model; and a processing module for acquiring the model data from the first information.

[0046] In an optional embodiment of the fourth aspect, the first information is also used to indicate a model identifier corresponding to the AI ​​algorithm model, and the model identifier corresponding to the AI ​​algorithm model is used to uniquely identify the AI ​​algorithm model.

[0047] In an optional embodiment of the fourth aspect, the model identifier corresponding to the AI ​​algorithm model has been assigned to the terminal device, and the terminal device supports the algorithm capabilities related to the AI ​​algorithm model.

[0048] An optional embodiment of the fourth aspect further includes: a first sending module configured to send second information indicating algorithm capabilities related to the AI ​​algorithm model supported by the terminal device. The acquisition module may also be configured to: obtain third information indicating a model identifier corresponding to the assigned AI algorithm model; obtain the model identifier from the third information, and save the model identifier.

[0049] In an optional embodiment of the fourth aspect, the first information is SI.

[0050] In an optional embodiment of the fourth aspect, the first information includes multiple SIBs, among which the first SIB is used to indicate that the second SIB contains model data of the AI ​​algorithm model, the first SIB is one of the multiple SIBs, and the second SIB is a different SIB from the first SIB. The processing module can also be used to: obtain the model data of the AI ​​algorithm model from the second SIB according to the indication of the first SIB.

[0051] In an optional embodiment of the fourth aspect, the acquisition module can also be used to: obtain the broadcast fourth information, the fourth information is used to indicate that the AI ​​algorithm model is ready to be sent and / or the usage status of the AI ​​algorithm model changes.

[0052] In an optional embodiment of the fourth aspect, the fourth information is also used to indicate that the SI carrying the AI ​​algorithm model has changed.

[0053] In an optional embodiment of the fourth aspect, the fourth information is: a short message of DCI scheduled using the model identifier and / or paging identifier corresponding to the AI ​​algorithm model, and / or a MAC CE scheduled using the model identifier corresponding to the AI ​​algorithm model.

[0054] In an optional embodiment of the fourth aspect, it also includes: a second sending module, used to send fifth information, and the fifth information is used to request model data of the AI ​​algorithm model.

[0055] In an optional embodiment of the fourth aspect, the fifth information is an SI request.

[0056] In an optional embodiment of the fourth aspect, the first information is broadcast multiple times, and the processing module can also be used to selectively merge the model data obtained from the first information broadcast multiple times to obtain complete model data of the AI ​​algorithm model.

[0057] In an optional embodiment of the fourth aspect, the terminal device is in a connected state, an inactive state or an idle state.

[0058] It should be understood that the technical solution of the fourth aspect of the present application corresponds to the technical solution of the second aspect of the present application, and the beneficial effects achieved by each aspect and the corresponding feasible implementation methods are similar and will not be repeated.

[0059] In a fifth aspect, an embodiment of the present application provides an electronic device, comprising a processor and a memory, the memory being used to store computer-executable instructions, the processor executing the computer-executable instructions stored in the memory, so that the electronic device performs the method described in the first aspect or any possible implementation of the first aspect, or the electronic device performs the method described in the second aspect or any possible implementation of the second aspect.

[0060] In a sixth aspect, an embodiment of the present application provides a computer-readable storage medium, in which a computer program or instruction is stored. When the computer program or instruction is executed by a processor, the method described in the first aspect or any possible implementation of the first aspect is implemented; alternatively, when the computer program or instruction is executed by a processor, the method described in the second aspect or any possible implementation of the second aspect is implemented.

[0061] In a seventh aspect, the present application provides a chip or chip system, which includes at least one processor and a communication interface, the communication interface and the at least one processor being interconnected by a line, and the at least one processor being used to run a computer program or instruction to execute the method described in the first aspect or any possible implementation of the first aspect, or to execute the method described in the second aspect or any possible implementation of the second aspect. The communication interface in the chip can be an input / output interface, a pin, or a circuit, etc.

[0062] In one possible implementation, the chip or chip system described above in this application further includes at least one memory, in which instructions are stored. The memory may be a storage unit within the chip, such as a register, a cache, etc., or a storage unit of the chip (e.g., a read-only memory, a random access memory, etc.).

[0063] In an eighth aspect, an embodiment of the present application provides a computer program product, comprising a computer program, which, when run, enables a computer to execute the method described in the first aspect or any possible implementation of the first aspect, or enables a computer to execute the method described in the second aspect or any possible implementation of the second aspect.

[0064] It should be understood that the fifth to eighth aspects of the present application correspond to the technical solutions of the first to second aspects of the present application, and the beneficial effects achieved by each aspect and the corresponding feasible implementation methods are similar and will not be repeated here.

[0065] An embodiment of the present application provides a model transmission method, device, and storage medium. A network device broadcasts first information, which indicates model data of an AI algorithm model. A terminal device can obtain the model data of the AI ​​algorithm model from the first information. Through a point-to-multipoint broadcast method, the same AI algorithm model can be sent to multiple terminal devices, thereby avoiding or reducing repeated transmission of the same AI algorithm model and saving air interface resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 is a logical block diagram of data collection, model training, model monitoring, model reasoning, and model storage for AI algorithm models;

[0067] FIG2 is a diagram of a communication system architecture provided in an embodiment of the present application;

[0068] FIG3 is a schematic diagram of a model transmission method provided in an embodiment of the present application;

[0069] FIG4 is a schematic diagram of a model transmission method provided in yet another embodiment of the present application;

[0070] FIG5 is a schematic diagram of a model transmission method provided in yet another embodiment of the present application;

[0071] FIG6 is a schematic diagram of a network device provided in an embodiment of the present application;

[0072] FIG7 is a schematic diagram of a terminal device provided in an embodiment of the present application;

[0073] FIG8 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0074] To facilitate a clear description of the technical solutions of the embodiments of the present application, some of the terms and technologies involved in the embodiments of the present application are briefly introduced below:

[0075] Artificial intelligence (AI) algorithm model: An algorithm model with a certain learning ability. By training the AI ​​algorithm model, the AI ​​algorithm model can learn to process massive amounts of data and complex tasks, such as the neural network model in the machine learning (ML) model.

[0076] Figure 1 is a logical block diagram of data collection, model training, model monitoring, model reasoning, and model storage of an AI algorithm model. This logical block diagram shows the life cycle of an AI algorithm model.

[0077] As shown in Figure 1, the life cycle of an AI algorithm model may include a data collection stage 101, a model training stage 102, a model monitoring stage 103, a model reasoning stage 104, and a model storage stage 105: In the data collection stage 101, data collection can obtain training data for model training, monitoring data for model monitoring (i.e., model management), and reasoning data for model reasoning (i.e., model reasoning operations); in the model training stage 102, the training data can be used to train the AI ​​algorithm model to obtain a trained or updated AI algorithm model; in the model monitoring stage 103, the effect of the AI ​​algorithm model can be monitored based on the monitoring data and the parameters output by the model reasoning stage 104. The model effect is fed back to the model training stage 102, or a retraining request is sent to the model training stage 102 to retrain the AI ​​algorithm model; in the model inference stage 104, the AI ​​algorithm model can be inferred based on the inference data to obtain corresponding parameters, and the parameters are output to the model monitoring stage 103. The model monitoring stage 103 can determine the AI ​​algorithm model for model inference in the model inference stage 104 from the stored AI algorithm models through model selection, activation, switching or fallback; in the model storage stage 105, the trained or updated AI algorithm model can be stored, and the AI ​​algorithm model with better model effect in the model monitoring stage 103 can also be stored.

[0078] The training of AI algorithm models requires a lot of computing power support, so in communication scenarios, the training of AI algorithm models can be deployed on the network side or on the server; the application scenarios of AI algorithm models may include the channel state information (CSI) prediction scenario on the terminal device side, and the beam prediction scenario in the time domain and frequency domain. These application scenarios require the terminal device to act as an AI algorithm execution node to perform operations, so the trained AI algorithm model can be deployed on the terminal device side. In this way, when the training and use of the AI ​​algorithm model are not on the same node, the AI ​​algorithm model needs to be transmitted. The embodiment of the present application solves how to transmit the AI ​​algorithm model from the network device to the terminal device.

[0079] It should be noted that in the embodiments of the present application, words such as "first" and "second" are used to distinguish between identical or similar items with substantially the same functions and effects. For example, the first information and the second information are merely used to distinguish different information and do not limit their order. Those skilled in the art will understand that words such as "first" and "second" do not limit the quantity or execution order, and words such as "first" and "second" do not necessarily mean that they are different.

[0080] It should be noted that in the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described in this application as "exemplary" or "for example" should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0081] It should be noted that, in the embodiments of the present application, "at least one" refers to one or more, and "more" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: the existence of A alone, the existence of A and B at the same time, and the existence of B alone, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, a--c, bc, or abc, where a, b, c can be single or multiple.

[0082] In order to better understand the model transmission method provided in the embodiment of the present application, the communication system architecture of the embodiment of the present application is described below in conjunction with Figure 1.

[0083] For example, Figure 2 is a diagram of the communication system architecture provided by an embodiment of the present application. As shown in Figure 2, the communication system includes a terminal device 201 and a network device 202, and the terminal device 201 communicates with the network device 202 wirelessly.

[0084] The terminal device involved in the embodiments of the present application can also be referred to as a terminal, which can be a device with wireless transceiver capabilities, which can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; it can also be deployed on the water (such as a ship, etc.); it can also be deployed in the air (for example, on an airplane, a balloon, and a satellite, etc.). The terminal device can be a user equipment (UE), wherein the UE includes a handheld device, a vehicle-mounted device, a wearable device, or a computing device with wireless communication capabilities. Exemplarily, the UE can be a mobile phone, a tablet computer, or a computer with wireless transceiver capabilities. The terminal device can also be a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a mixed reality (MR) terminal device, a wireless terminal in industrial control, a wireless terminal in unmanned driving, a wireless terminal in telemedicine, a wireless terminal in a smart grid, a wireless terminal in a smart city, a wireless terminal in a smart home, and the like. In an embodiment of the present application, the device for realizing the function of the terminal device may be the terminal device; or it may be a device that can support the terminal device to realize the function, such as a chip system, which may be installed in the terminal device.

[0085] The network device 202 involved in the embodiment of the present application includes an access network device 2021 and a core network device 2022.

[0086] Access network (RAN) equipment is the intermediate device that connects terminal devices to core network equipment via wireless communication. It is primarily responsible for radio resource management, quality of service (QoS) management, data compression and encryption, etc. on the air interface side. Examples include NodeBs, evolved eNodeBs, gNodeBs in 5G mobile communication systems or next-generation radio (NR) communication systems, and base stations in future mobile communication systems.

[0087] Core network (CN) equipment includes user plane function (UPF) network elements, access and mobility management function (AMF) network elements, session management function (SMF) network elements, policy control function (PCF) network elements, etc. Among them, the UPF network element is mainly responsible for the transmission of user data, while the other network elements can be called control plane function network elements, which are mainly responsible for authentication, authorization, registration management, session management, mobility management, and policy control to ensure the reliable and stable transmission of user data.

[0088] In an embodiment of the present application, the device for implementing the function of the network device may be a network device, or a device that can support the network device to implement the function, such as a chip system, which may be installed in the network device.

[0089] The technical solutions provided in the embodiments of the present application can be applied to the long term evolution (LTE) architecture, and can also be applied to the universal mobile telecommunications system (UMTS) terrestrial radio access network (UTRAN) architecture, or the global system for mobile communication (GSM) / enhanced data rate for GSM evolution (EDGE) system radio access network (GSM EDGE radio access network, GERAN) architecture. In addition, the technical solutions provided in the embodiments of the present application can also be applied to any other wireless communication system with similar structure and function, such as a public land mobile network (PLMN) system, a 5G communication system or a communication system after 5G, etc., and the embodiments of the present application do not impose any restrictions on this.

[0090] Wireless communication between communication devices may include: wireless communication between network devices and terminal devices, wireless communication between network devices and network devices, and wireless communication between terminal devices. In the embodiments of the present application, the term "wireless communication" may also be referred to as "communication", and the term "communication" may also be described as "data transmission", "information transmission" or "transmission". Those skilled in the art may apply the technical solutions provided in the embodiments of the present application to wireless communication between network devices and terminal devices, such as wireless communication between access network devices and terminal devices.

[0091] In related technologies, in a mobile communication network, terminal devices with the same characteristics in the same cell or the same network device may use the same AI algorithm model. When the network device transmits the AI ​​algorithm model to the terminal device, the same AI algorithm model may be transmitted repeatedly, resulting in a waste of air interface resources.

[0092] In view of the above problems, the present invention proposes a model transmission method, the main inventive ideas of which are as follows:

[0093] The network device broadcasts the model data of the AI ​​algorithm model, and multiple terminal devices that need to use the AI ​​algorithm model can obtain the model data of the AI ​​algorithm model from the broadcast information. In this way, the network device can reduce or avoid repeatedly transmitting the model data of the AI ​​algorithm model to multiple terminal devices, saving air interface resources.

[0094] The model transmission method provided in the embodiment of the present application is applicable to any AI algorithm task or service including channel state information (CSI) processing, beam processing, and positioning processing on the terminal device side.

[0095] Optionally, CSI processing may include CSI prediction and / or CSI feedback enhancement; beam processing may include beam management and / or beam prediction; and positioning processing may include positioning accuracy enhancement.

[0096] It should be noted that the model transmission method provided in the embodiment of the present application is not limited to the above tasks or services.

[0097] The technical solutions shown in this application are described in detail below through specific embodiments. It should be noted that the following embodiments can exist independently or in combination with each other. For the same or similar content, such as the explanation of terms or nouns, and the explanation of steps, etc., different embodiments can refer to each other and will not be repeated.

[0098] First, the technical solution provided by the embodiment of the present application is described in detail through specific embodiments from the terminal device side.

[0099] FIG3 is a schematic diagram of a model transmission method provided in an embodiment of the present application. As shown in FIG3 , the model transmission method provided in this embodiment includes:

[0100] S301, a network device broadcasts first information, where the first information is used to indicate model data of an AI algorithm model.

[0101] Exemplarily, the model data of the AI ​​algorithm model may include weight parameters of the AI ​​algorithm model, model structure parameters, required input data, data format of input data, data format of output data, inference formulas, etc.

[0102] Among them, in the first information, the model data of the AI ​​algorithm model can be presented as one or more data packets.

[0103] Among them, the training of AI algorithm models can be performed on RAN equipment (such as gNodeB), CN equipment and / or operation administration and maintenance (OAM).

[0104] In this embodiment, the network device may broadcast the first information after completing the training or updating of the AI ​​algorithm model, or when the network device is ready to transmit the AI ​​algorithm model.

[0105] S302: The terminal device obtains the broadcasted first information and obtains model data of the AI ​​algorithm model from the first information.

[0106] In this embodiment, a terminal device within the broadcast range of the network device, for example, a terminal device within the cell corresponding to the network device, can receive the first information broadcast by the network device and obtain the model data of the AI ​​algorithm model from the first information. The terminal device can then deploy and execute the AI ​​algorithm model according to business needs.

[0107] In an embodiment of the present application, for the model data of the same AI algorithm model, the network device can transmit it to multiple terminal devices at one time through a point-to-multipoint broadcast method, thereby avoiding or reducing multiple repeated transmissions of the model data and saving air interface resources.

[0108] Regarding the indication content of the first information, the following optional methods may be provided:

[0109] Optionally, the first information is further used to indicate a model identifier corresponding to the AI ​​algorithm model, and the model identifier corresponding to the AI ​​algorithm model is used to uniquely identify the AI ​​algorithm model. In this way, the terminal device can determine the AI ​​algorithm model to which the model data in the first information belongs based on the model identifier corresponding to the AI ​​algorithm model, thereby improving the accuracy of the transmission and application of the AI ​​algorithm model.

[0110] Optionally, the first information is used to indicate model data corresponding to multiple AI algorithm models and model identifiers corresponding to multiple AI algorithm models, thereby achieving transmission of multiple AI algorithm models by broadcasting the first information, thereby improving the transmission efficiency of the AI ​​algorithm models, and distinguishing model data corresponding to multiple AI algorithm models by using the model identifiers corresponding to the multiple AI algorithm models, thereby improving the transmission accuracy of the AI ​​algorithm models.

[0111] Regarding the model identification corresponding to the AI ​​algorithm model, the following optional methods are available:

[0112] Optionally, the model identifier corresponding to the AI ​​algorithm model may include the model name and / or model number corresponding to the AI ​​algorithm model.

[0113] Optionally, the model identifier corresponding to the AI ​​algorithm model may be the radio network temporary identifier (RNTI) corresponding to the AI ​​algorithm model, for example, it may be abbreviated as PAI-RNTI, so as to distinguish different AI algorithm models by the RNTI corresponding to the AI ​​algorithm model.

[0114] Optionally, the model identifier corresponding to the AI ​​algorithm model has been assigned to a terminal device that supports the algorithm capabilities related to the AI ​​algorithm model. Thus, the terminal device can accurately obtain model data of the AI ​​algorithm model that meets the algorithm capabilities supported by itself from the first information based on the model identifier assigned to itself, thereby improving the transmission accuracy of the AI ​​algorithm model.

[0115] For example, the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device can be determined based on information such as the application scenarios related to the AI ​​algorithm model supported by the terminal device and the computing capabilities related to the AI ​​algorithm model.

[0116] Optionally, the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device include scenario identifiers of application scenarios related to the AI ​​algorithm model supported by the terminal device. For example, terminal devices supporting the same application scenario have the same algorithm capabilities related to the AI ​​algorithm model, and the same model identifier can be assigned to terminal devices supporting the same application scenario.

[0117] Optionally, application scenarios related to AI algorithm models may include one or more of the following: CSI processing scenarios, beam processing scenarios, and positioning processing scenarios. CSI processing scenarios may include CSI prediction and / or CSI feedback enhancement; beam processing may include beam management and / or beam prediction; and positioning processing may include positioning accuracy enhancement.

[0118] For example, for multiple terminal devices that support CSI prediction using an AI algorithm model, the model identifier corresponding to the AI ​​algorithm model used for CSI prediction can be allocated; for multiple terminal devices that support beam prediction using an AI algorithm model, the model identifier corresponding to the AI ​​algorithm model used for beam prediction can be allocated; and for multiple terminal devices that support positioning accuracy enhancement using an AI algorithm model, the model identifier corresponding to the AI ​​algorithm model used for positioning accuracy enhancement can be allocated. This improves the accuracy of model identifier allocation.

[0119] Regarding the allocation of model identifiers, the following embodiments may be provided:

[0120] FIG4 is a schematic diagram of a model transmission method provided by another embodiment of the present application. As shown in FIG4 , the model transmission method provided by this embodiment includes:

[0121] S401, the terminal device sends second information, where the second information is used to indicate the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device.

[0122] Among them, the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device can be referred to the above content and will not be repeated here.

[0123] In this embodiment, the terminal device can report the algorithm capabilities related to the AI ​​algorithm model that it supports by sending the second information to the network device.

[0124] Optionally, the terminal device sends the second information when accessing the cell where the network device is located.

[0125] Optionally, the terminal device sends the second information when the supported algorithm capabilities related to the AI ​​algorithm model are updated.

[0126] Optionally, the terminal device periodically sends the second information.

[0127] Optionally, the terminal device sends the second information when obtaining information from the network device indicating that the terminal device is querying the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device.

[0128] Among them, S401 is an optional step. The network device can obtain the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device by obtaining the second information. The network device can also obtain the algorithm capabilities through other means.

[0129] Optionally, in the event of a cell handover of a terminal device, the network device corresponding to the target cell may obtain algorithm capability information related to the AI ​​algorithm model supported by the terminal device from the network device corresponding to the source cell during the cell handover process.

[0130] S402, the network device indicates third information based on the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device, where the third information is used to indicate the model identifier corresponding to the allocated AI algorithm model.

[0131] In this embodiment, the network device determines the model identifier corresponding to the AI ​​algorithm model assigned to the terminal device based on the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device. For example, for a terminal device that supports CSI prediction using the AI ​​algorithm model, the network device determines the model identifier corresponding to the AI ​​algorithm model assigned for CSI prediction. After determining the model identifier assigned to the terminal device, the network device sends third information to the terminal device, which includes the model identifier corresponding to the AI ​​algorithm model, to instruct the terminal device to assign the model identifier corresponding to the AI ​​algorithm model.

[0132] Optionally, the second information and the third information are transmitted via RRC signaling.

[0133] In this optional manner, dedicated RRC signaling may be used to carry the second information and the third information.

[0134] S403: The terminal device obtains the model identifier from the third message and saves the model identifier.

[0135] In this embodiment, the terminal device obtains the model identifier assigned by the network device from the third message and saves the model identifier so that the terminal device can subsequently accurately obtain corresponding model data based on the model identifier.

[0136] S402 and S403 are optional steps. For example, when the network device broadcasts the first information, the terminal device that has just accessed the cell where the network device is located does not obtain the model identifier assigned to it by the network device.

[0137] S404, the network device broadcasts first information, where the first information is used to indicate model data of the AI ​​algorithm model and a model identifier corresponding to the AI ​​algorithm model.

[0138] In this embodiment, when a model identifier has been assigned to the terminal device, the network device may broadcast the model data of the AI ​​algorithm model and the model identifier corresponding to the AI ​​algorithm model after training or updating the AI ​​algorithm model, or when the network device is ready to transmit the AI ​​algorithm model.

[0139] S405: The terminal device obtains the broadcasted first information and obtains model data of the AI ​​algorithm model from the first information.

[0140] In this embodiment, the terminal device receives the broadcasted first information and can obtain the model data of the AI ​​algorithm model corresponding to the model identifier assigned to the terminal device from the first information.

[0141] The implementation principles and technical effects of S404 and S405 can be referred to the aforementioned embodiments and will not be described in detail here.

[0142] Optionally, if the terminal device is not assigned a model identifier, the terminal device may obtain the broadcasted first information, compare the AI ​​algorithm model associated with the first information with the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device, and if the terminal device supports the AI ​​algorithm model associated with the first information, the model data of the AI ​​algorithm model may be saved. The AI ​​algorithm model associated with the first information is the AI ​​algorithm model corresponding to the model data or model identifier indicated in the first information.

[0143] In an embodiment of the present application, a terminal device may report the algorithm capabilities related to the AI ​​algorithm model that it supports. The network device may assign a model identifier corresponding to the corresponding AI algorithm model to the terminal device based on the algorithm capabilities reported by the terminal device. When the network device broadcasts a first message, the terminal device may obtain model data of the AI ​​algorithm model from the broadcasted first message based on the model identifier. Thus, by assigning model identifiers based on the algorithm capabilities of the terminal device, the accuracy of AI algorithm model transmission is improved. By broadcasting the same AI algorithm model to terminal devices with the same algorithm capabilities, repeated transmission of the same AI algorithm model is avoided or reduced, effectively saving air interface resources.

[0144] Based on any of the foregoing embodiments, the following optional methods may be provided regarding the first information:

[0145] Optionally, the first information is system information (SI), so as to transmit the AI ​​algorithm model through SI.

[0146] Among them, the cell corresponding to the network device can transmit some attribute information, configuration information, etc. related to the cell by broadcasting SI. After the terminal device achieves downlink synchronization with the cell, it can obtain the SI broadcast by the network device corresponding to the cell.

[0147] In this optional method, the first information is SI, which means that in addition to some cell-related attribute information and configuration information, the SI may also contain model data of the AI ​​algorithm model. Therefore, by broadcasting SI, the AI ​​algorithm model is transmitted, avoiding or reducing repeated transmission of the same AI algorithm model in the same cell, saving air interface resources.

[0148] Optionally, when the first information is SI, the first information includes multiple system information blocks (SIBs). Among the multiple SIBs, the first SIB is used to indicate that the second SIB contains model data of the AI ​​algorithm model. The first SIB is one of the multiple SIBs, and the second SIB is a different SIB from the first SIB.

[0149] In one possible implementation, the terminal device may obtain the model data of the AI ​​algorithm model from the second SIB according to the instruction of the first SIB. Thus, by carrying the model data of the AI ​​algorithm model in the second SIB, the model data of the AI ​​algorithm model indicated in the first SIB is located in the second SIB. In the scenario where the model data of the AI ​​algorithm model is broadcast through the SI, the terminal can accurately obtain the model data of the AI ​​algorithm model from the SIB.

[0150] Optionally, the second SIB is located after the first SIB in the multiple SIBs. Thus, the terminal device first obtains the first SIB and then obtains the model data of the AI ​​algorithm model from the second SIB obtained later according to the instruction of the first SIB.

[0151] Optionally, multiple SIBs may be represented in sequence as SIB1, SIB2, SIB3, ..., the first SIB may be SIB1, the second SIB may be SIBx, that is, the xth SIB among the multiple SIBs, and SIB1 indicates that SIBx contains model data of the AI ​​algorithm model.

[0152] Where x is greater than 1.

[0153] FIG5 is a schematic diagram of a model transmission method provided by another embodiment of the present application. As shown in FIG5 , the model transmission method provided by this embodiment includes:

[0154] S501, the terminal device sends second information, where the second information is used to indicate the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device.

[0155] S502, the network device indicates third information based on the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device, where the third information is used to indicate the model identifier corresponding to the allocated AI algorithm model.

[0156] S503: The terminal device obtains the model identifier from the third message and saves the model identifier.

[0157] Among them, S501 to S503 are optional steps.

[0158] The implementation principles and technical effects of S501 to S503 may refer to the aforementioned embodiments and will not be described in detail here.

[0159] S504, the network device broadcasts fourth information, where the fourth information is used to indicate that the AI ​​algorithm model is ready to be sent and / or to indicate a change in the usage status of the AI ​​algorithm model.

[0160] Among them, S504 is an optional step.

[0161] Optionally, the fourth information is used to indicate that an unissued AI algorithm model is ready to be issued.

[0162] Optionally, the fourth information is used to indicate that the AI ​​algorithm model that has been issued is ready for issuance after retraining and modification.

[0163] Optionally, the fourth information is used to indicate that the usage status of the issued AI algorithm model has changed to allowed or unavailable, including indicating that a previously available AI algorithm model is unavailable or that a previously unavailable AI algorithm model is allowed to be used.

[0164] In this embodiment, the network device notifies the terminal device of the issuance or use of the AI ​​algorithm model by broadcasting the fourth information. After obtaining the fourth information, the terminal device may wait to obtain the first information and notify the terminal device in advance through the fourth information to prevent the terminal device from missing the first information; or, if the fourth information is used to indicate a change in the use status of the AI ​​algorithm model, after obtaining the fourth information, the terminal device may directly change the use status of the AI ​​algorithm model based on the indication of the fourth information.

[0165] Optionally, the fourth information includes a model identifier corresponding to the AI ​​algorithm model, and the fourth information is used to indicate that the AI ​​algorithm model corresponding to the model identifier is ready to be issued and / or to indicate a change in the usage status of the AI ​​algorithm model corresponding to the model identifier.

[0166] Optionally, the fourth information is used to indicate that the SI carrying the AI ​​algorithm model has changed. When the fourth information is also used to indicate that the SI carrying the AI ​​algorithm model has changed, the first information is the SI carrying the changed AI algorithm model. Thus, through the notification information of the SI change, the terminal device is told to prepare to receive the SI carrying the AI ​​algorithm model. The terminal device that supports the algorithm capabilities related to the AI ​​algorithm model can prepare to obtain the next SI broadcast by the network device, obtain the model data of the AI ​​algorithm model from the SI, and realize the broadcast of the AI ​​algorithm model by broadcasting the SI change notification and SI.

[0167] As an example, before the SI is changed, it contains some attribute information and configuration information of the cell. If the network device wants to transmit the AI ​​algorithm model by broadcasting SI, it is necessary to add the model data of the AI ​​algorithm model in the SI, and it may be necessary to add the model identifier corresponding to the AI ​​algorithm model, which will cause the SI to change. Therefore, before broadcasting the changed SI, the notification information of the SI change can be broadcast, that is, the fourth information can be broadcast, so that the terminal device is ready to receive the changed SI.

[0168] Optionally, the fourth information is: a short message of downlink control information (DCI) scheduled using the model identifier and / or paging identifier corresponding to the AI ​​algorithm model, and / or a media access control (MAC) control element (CE) scheduled using the model identifier corresponding to the AI ​​algorithm model.

[0169] The DCI short message scheduled using the model identifier corresponding to the AI ​​algorithm model refers to a DCI short message that carries or includes the model identifier corresponding to the AI ​​algorithm model. The MAC CE scheduled using the model identifier corresponding to the AI ​​algorithm model refers to a MAC CE that carries or includes the model identifier corresponding to the AI ​​algorithm model.

[0170] Exemplarily, the paging identifier may be a radio network temporary identifier (RNTI) used for paging, which may be referred to as P-RNTI for short.

[0171] Among them, the short message can indicate that one or more AI algorithm models are ready to be issued or the usage status has changed, and the correspondence between the short message and the AI ​​algorithm model can be carried in the first information.

[0172] When the fourth information is a short message, the following options are available:

[0173] Option 1: When the fourth information is a DCI short message scheduled using a paging identifier, the fourth information has multiple bits, and one of the multiple bits (such as the first bit) is used to indicate whether the SI has changed; after the terminal device obtains the fourth information, if the bit in the fourth information indicates that the SI has changed, the SI broadcast by the network device is obtained, and based on the first SIB included in the SI, it can be determined whether the second SIB carrying the AI ​​algorithm model has changed. If it is determined that the second SIB has changed, the terminal device can further obtain the second SIB through the scheduling information corresponding to the second SIB indicated by the first SIB (such as the time-frequency resources corresponding to the second SIB) to obtain the model data of the AI ​​algorithm model in the second SIB.

[0174] Option 2: When the fourth information is a DCI short message scheduled using a paging identifier, the fourth information has N bits, and the fourth information can be used to indicate that 1 to N AI algorithm models are ready to be sent and / or indicate a change in the usage status of 1 to N AI algorithm models, for example, the first bit corresponds to AI algorithm model a1, the second bit corresponds to AI algorithm model a2, and so on. Among them, the first information can also be used to indicate the mapping relationship between the N bits in the fourth information and the AI ​​algorithm model, so that the terminal device can combine the fourth information and the mapping relationship. After obtaining the fourth information, the terminal can obtain the SI broadcast by the network device, further obtain the first SIB in the SI, and can obtain model data from the second SIB based on the indication of the first SIB.

[0175] Option 3: When the fourth information is a DCI short message scheduled using the model identifier corresponding to the AI ​​algorithm model, the fourth information can be used to indicate that the AI ​​algorithm model corresponding to the model identifier is ready for delivery or a change in usage status. After obtaining the fourth information, the terminal can obtain the SI broadcast by the network device, further obtain the first SIB in the SI, and obtain model data from the second SIB based on the indication of the first SIB.

[0176] Exemplarily, the model identifier corresponding to the AI ​​algorithm model may be the RNTI that identifies the AI ​​algorithm model, referred to as PAI-RNTI.

[0177] Optionally, after obtaining the fourth information, the terminal obtains the first information at the next paging occasion (PO), thereby accurately receiving the first information, that is, accurately obtaining the model data of the AI ​​algorithm model.

[0178] S505, the network device broadcasts first information, where the first information is used to indicate model data of the AI ​​algorithm model.

[0179] Optionally, the AI ​​algorithm model is the AI ​​algorithm model that the fourth information indicates is ready to be issued.

[0180] Optionally, the AI ​​algorithm model is an AI algorithm model whose usage status changes as indicated by the fourth information.

[0181] Optionally, the first information is also used to indicate the model identifier corresponding to the AI ​​algorithm model.

[0182] S506: The terminal device obtains the broadcasted first information and obtains model data of the AI ​​algorithm model from the first information.

[0183] The implementation principles and technical effects of S505 and S506 can be referred to the aforementioned embodiments and will not be described in detail.

[0184] In an embodiment of the present application, the network device broadcasts the fourth information to inform the terminal device that it is preparing to send the AI ​​algorithm model or that the usage status of the AI ​​algorithm model has changed. After obtaining the fourth information, the terminal device can prepare to receive the model data of the AI ​​algorithm model. The network device then transmits the model data of the AI ​​algorithm model to the terminal device by broadcasting the first information, so that the terminal device receives the AI ​​algorithm model in a timely and accurate manner. This avoids or reduces the repeated transmission of the AI ​​algorithm model, saves air interface resources, and improves the accuracy and success rate of the AI ​​algorithm model transmission.

[0185] Based on any of the above embodiments, the following optional solutions can be provided:

[0186] Optionally, before the network device broadcasts the first information, the terminal device may send a fifth information, which the network device may obtain. The fifth information is used to request model data for the AI ​​algorithm model. Thus, the terminal can actively request model data for the AI ​​algorithm model by sending the fifth information to the network device.

[0187] Optionally, the fifth information is also used to indicate the model identifier corresponding to the AI ​​algorithm model, and to request model data of the AI ​​algorithm model corresponding to the model identifier, so as to accurately inform the network device of the AI ​​algorithm model that needs to be transmitted.

[0188] Optionally, the fifth information is an SI request. In the case where the fifth information is an SI request, the first information is SI. In this case, the terminal device requests the model data of the AI ​​algorithm model from the network device through the SI request, and the network device can transmit the model data of the AI ​​algorithm model to the terminal device by broadcasting SI carrying the model data of the AI ​​algorithm model.

[0189] Optionally, the terminal device sends the fifth information when a cell handover occurs. When a cell handover occurs, the terminal device may miss the model data of the AI ​​algorithm model broadcast by the new cell, and can actively request to obtain the model data of the AI ​​algorithm model by actively sending the fifth information to the network device.

[0190] Optionally, the terminal device sends the fifth information when the data packet in the model data of the AI ​​algorithm model fails to be received. Thus, the terminal device actively requests to obtain the model data of the AI ​​algorithm model when the model data reception fails.

[0191] Optionally, after obtaining the fourth information, the terminal device sends the fifth information. Thus, when the network device notifies the AI ​​algorithm model of the change, the terminal device actively requests to obtain the model data of the AI ​​algorithm model, effectively improving the transmission efficiency of the AI ​​algorithm model.

[0192] Optionally, if the network device does not receive the fifth information after broadcasting the fourth information, it means that the terminal devices in the cell corresponding to the network device may not need the AI ​​algorithm model changed as indicated by the fourth information, and the network device may not broadcast the first information to save air interface resources. If the network device receives the fifth information after broadcasting the fourth information, it means that the terminal devices in the cell corresponding to the network device need the AI ​​algorithm model changed as indicated by the fourth information, and the network device may broadcast the first information.

[0193] Optionally, the first information is broadcast multiple times, and after the terminal device obtains the model data of the AI ​​algorithm model from the first information, it also includes: the terminal device selectively merges the model data obtained from the first information broadcast multiple times to obtain the complete model data of the AI ​​algorithm model.

[0194] For example, considering that the model data of the AI ​​algorithm model may have a large amount of data, each broadcast of the first information may include a portion of the data packets in the model data of the AI ​​algorithm model. The terminal device can obtain a portion of the data packets in the model data of the AI ​​algorithm model from the first broadcast information, and obtain the complete model data of the AI ​​algorithm model by selectively merging the partial data packets obtained multiple times. Among them, duplicate data packets may appear between the first information broadcasted at different times, and the selective merging includes deduplicating and then merging the duplicate data packets.

[0195] As another example, when the first information is broadcast once, the terminal device may not successfully obtain the first information, or the terminal device may successfully parse and obtain part of the data packet in the model data from the first information after obtaining the first information. The model data of the AI ​​algorithm model can be broadcast multiple times by broadcasting the first information multiple times, thereby improving the success rate of the AI ​​algorithm model transmission.

[0196] Optionally, the terminal device includes at least one of the following: a terminal device in a connected state, a terminal device in an inactive state, and a terminal device in an idle state. Thus, the model transmission method provided in the embodiment of the present application can be applied to terminal devices in different states and can transmit AI algorithm models to terminal devices in different states.

[0197] The model transmission method of the embodiment of the present application has been described above. The device for executing the above method provided by the embodiment of the present application is described below. Those skilled in the art will understand that the method and device can be combined and referenced with each other, and the relevant device provided by the embodiment of the present application can perform the steps in the above-mentioned list sorting method.

[0198] FIG6 is a schematic diagram of a network device according to an embodiment of the present application. As shown in FIG6 , the network device 60 includes a broadcast module 61 .

[0199] Among them, the broadcast module 61 is used to broadcast the first information, and the first information is used to indicate the model data of the AI ​​algorithm model.

[0200] In an optional embodiment, the first information is also used to indicate a model identifier corresponding to the AI ​​algorithm model, and the model identifier corresponding to the AI ​​algorithm model is used to uniquely identify the AI ​​algorithm model.

[0201] In an optional embodiment, the model identifier corresponding to the AI ​​algorithm model has been assigned to a terminal device that supports algorithm capabilities related to the AI ​​algorithm model.

[0202] In an optional embodiment, the network device 60 further includes: a first acquisition module (not shown) for acquiring second information, the second information being used to indicate the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device; an indication module (not shown) for indicating third information based on the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device, the third information being used to indicate a model identifier corresponding to the assigned AI algorithm model.

[0203] In an optional embodiment, the first information is SI.

[0204] In an optional embodiment, the first information includes multiple SIBs, among which the first SIB is used to indicate that the second SIB includes model data of the AI ​​algorithm model, the first SIB is one of the multiple SIBs, and the second SIB is a different SIB from the first SIB.

[0205] In an optional embodiment, the broadcast module 61 can also be used to: broadcast fourth information, where the fourth information is used to indicate that the AI ​​algorithm model is ready to be sent and / or that the usage status of the AI ​​algorithm model has changed.

[0206] In an optional embodiment, the fourth information is also used to indicate that the SI carrying the AI ​​algorithm model has changed.

[0207] In an optional embodiment, the fourth information is: a short message of DCI scheduled using the model identifier and / or paging identifier corresponding to the AI ​​algorithm model, and / or a MAC CE scheduled using the model identifier corresponding to the AI ​​algorithm model.

[0208] In an optional embodiment, the network device further includes: a second acquisition module (not shown), configured to acquire fifth information, wherein the fifth information is used to request model data of the AI ​​algorithm model.

[0209] In an optional embodiment, the fifth information is an SI request.

[0210] In an optional embodiment, the terminal device includes at least one of the following: a terminal device in a connected state, a terminal device in an inactive state, and a terminal device in an idle state.

[0211] The network device provided in this embodiment is used to implement the technical solution of the network device in the aforementioned method embodiment. Its implementation principle and technical effects are similar and will not be described in detail here.

[0212] FIG7 is a schematic diagram of a terminal device provided in an embodiment of the present application. As shown in FIG7 , the terminal device 70 includes an acquisition module 71 and a processing module 72 .

[0213] Among them, the acquisition module 71 is used to obtain the first broadcast information, and the first information is used to indicate the model data of the AI ​​algorithm model; the processing module 72 is used to obtain the model data from the first information.

[0214] In an optional embodiment, the first information is also used to indicate a model identifier corresponding to the AI ​​algorithm model, and the model identifier corresponding to the AI ​​algorithm model is used to uniquely identify the AI ​​algorithm model.

[0215] In an optional embodiment, a model identifier corresponding to the AI ​​algorithm model has been assigned to a terminal device, and the terminal device supports algorithm capabilities related to the AI ​​algorithm model.

[0216] In an optional embodiment, the terminal device 70 further includes: a first sending module (not shown) for sending second information, the second information being used to indicate the algorithm capabilities related to the AI ​​algorithm model supported by the terminal device. The acquisition module 71 can also be used to: obtain third information, the third information being used to indicate the model identifier corresponding to the assigned AI algorithm model; obtain the model identifier from the third information, and save the model identifier.

[0217] In an optional embodiment, the first information is SI.

[0218] In an optional embodiment, the first information includes multiple SIBs, among which the first SIB is used to indicate that the second SIB contains model data of the AI ​​algorithm model, the first SIB is one of the multiple SIBs, and the second SIB is a different SIB from the first SIB. The processing module 72 can also be used to: obtain the model data of the AI ​​algorithm model from the second SIB according to the indication of the first SIB.

[0219] In an optional embodiment, the acquisition module 71 can also be used to: obtain the broadcasted fourth information, where the fourth information is used to indicate that the AI ​​algorithm model is ready to be sent and / or the usage status of the AI ​​algorithm model has changed.

[0220] In an optional embodiment, the fourth information is also used to indicate that the SI carrying the AI ​​algorithm model has changed.

[0221] In an optional embodiment, the fourth information is: a short message of DCI scheduled using the model identifier and / or paging identifier corresponding to the AI ​​algorithm model, and / or a MAC CE scheduled using the model identifier corresponding to the AI ​​algorithm model.

[0222] In an optional embodiment, the terminal device 70 further includes: a second sending module (not shown), configured to send fifth information, wherein the fifth information is used to request model data of the AI ​​algorithm model.

[0223] In an optional embodiment, the fifth information is an SI request.

[0224] In an optional embodiment, the first information is broadcast multiple times, and the processing module 72 can also be used to selectively merge the model data obtained from the first information broadcast multiple times to obtain complete model data of the AI ​​algorithm model.

[0225] In an optional embodiment, the terminal device is in a connected state, an inactive state, or an idle state.

[0226] The terminal device provided in this embodiment is used to implement the technical solution of the terminal device in the aforementioned method embodiment. Its implementation principle and technical effects are similar and will not be repeated here.

[0227] It should be noted that the module names involved in the embodiments of the present application can be defined as other names as long as the functions of each module can be achieved, and there is no specific restriction on the names of the modules.

[0228] FIG8 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. As shown in FIG8 , the electronic device 80 includes: at least one processor 81, a memory 82, a communication interface 83, and a system bus 84. The memory 82 and the communication interface 83 are connected to the processor 81 via the system bus 84 and communicate with each other. The memory 82 is used to store instructions, the communication interface 83 is used to communicate with other devices, and the processor 81 is used to call instructions in the memory to execute the method steps provided in the above-mentioned method embodiment. The specific implementation methods and technical effects are similar and will not be repeated here.

[0229] The system bus 84 mentioned in FIG8 may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The system bus 184 may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, the figure shows only one thick line, but this does not mean that there is only one bus or only one type of bus.

[0230] The communication interface 83 is used to implement communication between the database access device and other devices (such as clients, read-write libraries, and read-only libraries).

[0231] The memory 82 may include a random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0232] The processor 81 can be a general-purpose processor, including a central processing unit, a network processor (NP), etc.; a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0233] The embodiment of the present application also provides a computer-readable storage medium, in which a computer program is stored, and when the computer program is executed by a processor, it is used to implement the method steps in the above method embodiment. The method described in the above embodiment can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. If implemented in software, the function can be stored as one or more instructions or codes on a computer-readable medium or transmitted on a computer-readable medium. Computer-readable media can include computer storage media and communication media, and can also include any medium that can transfer a computer program from one place to another. The storage medium can be any target medium that can be accessed by a computer.

[0234] In one possible implementation, computer-readable media may include RAM, ROM, compact disc read-only memory (CD-ROM) or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium designed to carry or store the desired program code in the form of instructions or data structures and accessible by a computer. Furthermore, any connection is appropriately referred to as a computer-readable medium. For example, if software is transmitted from a website, server or other remote source using a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL) or wireless technologies such as infrared, radio and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL or wireless technologies such as infrared, radio and microwave are included in the definition of medium. Disk and disc as used herein include optical disc, laser disc, optical disc, digital versatile disc (DVD), floppy disk and Blu-ray disc, where disks typically reproduce data magnetically, while optical discs reproduce data optically using lasers. Combinations of the above should also be included within the scope of computer-readable media.

[0235] The present application also provides a computer program product, which includes a computer program. When the computer program is executed, the computer executes the method steps in the above method embodiment.

[0236] An embodiment of the present application also provides a chip system, including at least one processor and a communication interface, the communication interface and the at least one processor are interconnected through lines, and the at least one processor is used to run computer programs or instructions to execute the method steps in the above method embodiment.

[0237] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0238] The present application embodiment is described with reference to the flow chart and / or block diagram of the method, device (system) and computer program product according to the embodiment of the present application.It should be understood that each flow process and / or box in the flow chart and / or block diagram and the combination of the flow process and / or box in the flow chart and / or block diagram can be realized by computer program instructions.These computer program instructions can be provided to the processing unit of general-purpose computer, special-purpose computer, embedded processing machine or other programmable device to produce a machine, so that the instruction executed by the processing unit of computer or other programmable data processing device produces the device for realizing the function specified in one flow chart flow or multiple flows and / or one block or multiple blocks of block diagram.

[0239] The above specific implementation methods further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above are only specific implementation methods of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent replacements, improvements, etc. made on the basis of the technical solutions of the present invention should be included in the scope of protection of the present invention.

Claims

1. A model transmission method, characterized in that: include: Broadcast first information, where the first information is used to indicate model data of an artificial intelligence (AI) algorithm model.

2. The method according to claim 1, characterized in that The first information is also used to indicate a model identifier corresponding to the AI algorithm model, and the model identifier is used to uniquely identify the AI algorithm model.

3. The method according to claim 2, characterized in that The model identifier has been assigned to a terminal device that supports algorithm capabilities related to the AI algorithm model.

4. The method according to claim 3, characterized in that Before broadcasting the first information, the method further includes: Acquire second information, where the second information is used to indicate algorithm capabilities related to the AI algorithm model supported by the terminal device; According to the algorithm capabilities related to the AI algorithm model supported by the terminal device, third information is indicated, and the third information is used to indicate the allocation of the model identifier.

5. The method according to any one of claims 1 to 4, characterized in that The first information is system information SI.

6. The method according to claim 5, characterized in that The first information includes multiple system information blocks SIBs. Among the multiple SIBs, the first SIB is used to indicate that the second SIB includes the model data. The first SIB is one of the multiple SIBs, and the second SIB is different from the first SIB.

7. The method according to any one of claims 1 to 6, characterized in that Before broadcasting the first information, the method further includes: Broadcast fourth information, where the fourth information is used to indicate that the AI algorithm model is ready to be sent and / or that the usage status of the AI algorithm model has changed.

8. The method according to claim 7, characterized in that The fourth information is used to indicate that the SI carrying the AI algorithm model has changed.

9. The method according to claim 7 or 8, characterized in that The fourth information is: a short message of downlink control information DCI scheduled using the model identifier and / or paging identifier corresponding to the AI algorithm model, and / or a media access control MAC control element CE scheduled using the model identifier corresponding to the AI algorithm model.

10. The method according to any one of claims 1 to 8, characterized in that Before broadcasting the first information, the method further includes: Acquire fifth information, where the fifth information is used to request the model data.

11. The method according to claim 10, characterized in that The fifth information is an SI request.

12. The method according to any one of claims 1 to 11, characterized in that The terminal device includes at least one of the following: a terminal device in a connected state, a terminal device in an inactive state, and a terminal device in an idle state.

13. A model transmission method, characterized in that: include: Obtaining first broadcast information, where the first information is used to indicate model data of an AI algorithm model; The model data is obtained from the first information.

14. The method according to claim 13, characterized in that The first information is also used to indicate a model identifier corresponding to the AI algorithm model, and the model identifier is used to uniquely identify the AI algorithm model.

15. The method according to claim 14, characterized in that The model identifier has been assigned to a terminal device, and the terminal device supports algorithm capabilities related to the AI algorithm model.

16. The method according to claim 15, characterized in that Before acquiring the first broadcast information, the method further includes: Sending second information, where the second information is used to indicate algorithm capabilities related to the AI algorithm model supported by the terminal device; Acquire third information, where the third information is used to indicate allocation of the model identifier; The model identifier is obtained from the third information and saved.

17. The method according to any one of claims 13 to 16, characterized in that The first information is SI.

18. The method according to claim 17, characterized in that The first information includes multiple SIBs, where a first SIB among the multiple SIBs is used to indicate that a second SIB includes the model data, the first SIB is one of the multiple SIBs, and the second SIB is a different SIB from the first SIB. Acquiring the model data from the first information includes: According to an instruction of the first SIB, the model data is obtained from the second SIB.

19. The method according to any one of claims 13 to 18, characterized in that Before acquiring the first broadcast information, the method further includes: Obtain the fourth broadcast information, where the fourth information is used to indicate that the AI algorithm model is ready to be sent and / or that the usage status of the AI algorithm model has changed.

20. The method according to claim 19, characterized in that The fourth information is also used to indicate that the SI carrying the AI algorithm model has changed.

21. The method according to claim 19 or 20, characterized in that The fourth information is: a short message of DCI scheduled using the model identifier and / or paging identifier corresponding to the AI algorithm model, and / or a MAC CE scheduled using the model identifier corresponding to the AI algorithm model.

22. The method according to any one of claims 13 to 21, characterized in that Before acquiring the first broadcast information, the method further includes: Fifth information is sent, where the fifth information is used to request the model data.

23. The method according to claim 22, characterized in that The fifth information is an SI request.

24. The method according to any one of claims 13 to 23, characterized in that The first information is broadcast multiple times. After obtaining the model data from the first information, the method further includes: The model data obtained from the first information broadcasted multiple times are selectively merged to obtain the complete model data of the AI algorithm model.

25. The method according to any one of claims 13 to 23, characterized in that The terminal device is in connected state, inactive state or idle state.

26. An electronic device, characterized in that: include: processor and memory; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the electronic device performs the method according to any one of claims 1 to 12, or the electronic device performs the method according to any one of claims 13 to 25.

27. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 12 is implemented; or when the computer program is executed by a processor, the method according to any one of claims 13 to 25 is implemented.

28. A chip system, characterized in that: The method comprises at least one processor and a communication interface, wherein the communication interface and the at least one processor are interconnected via a line, and the at least one processor is used to run a computer program or instruction to execute the method according to any one of claims 1 to 12, or to execute the method according to any one of claims 13 to 25.

29. A computer program product, characterized in that The invention comprises a computer program, which, when being executed, causes a computer to execute the method according to any one of claims 1 to 12, or causes a computer to execute the method according to any one of claims 13 to 25.

Images