Communication method and device

Abstract

The invention provides a communication method and device which are used for updating an AI model and ensuring the effectiveness of the deployed AI model. The method comprises the following steps: a first node sends a first request, wherein the first request comprises first identification information of a first model to be deployed; the first node receives a first access request, wherein the first access request is related to the first identification information; the first node sends a first model admission request, wherein the first model admission request comprises first model admission test information of the first model trained according to the first admission requirement; the first node receives first model admission indication information, the first model admission indication information is related to a test result of the trained first model, and the test result is obtained by testing the first model based on the first model admission test information.

Description

Technical Field

[0001] This application relates to the field of wireless communication technology, and more specifically, to a communication method and apparatus. Background Technology

[0002] Based on the vision of future intelligent and inclusive communication, intelligence will further evolve at the level of wireless communication network architecture. Artificial intelligence (AI) will be deeply integrated with wireless communication networks, bringing about changes in wireless communication network architecture and communication modes. AI models can be deployed and used as functional modules of wireless communication devices for tasks such as channel state information (CSI) compression feedback, channel prediction, channel estimation, localization, and beam management. Due to changes in the wireless communication environment, deployed AI models may become invalid, compromising the normal operation of the wireless communication network. This necessitates a mechanism to update AI models in a timely manner to trigger deployment upgrades and ensure the effectiveness of AI models. Summary of the Invention

[0003] This application provides a communication method and apparatus for updating AI models to ensure the effectiveness of deployed AI models.

[0004] In a first aspect, embodiments of this application provide a communication method applied to a first node. For example, the method can be executed by the first node, which may be a network-side device, a terminal-side device, or a model-providing network element. The first node may also be a component (e.g., a circuit, processor, chip, or chip system), logic module, or software that implements all or part of the functions of the first node; this application does not limit this. The following description uses a first node as an example. The method includes: the first node sending a first request, the first request including first identification information of a first model to be deployed; the first node receiving a first admission requirement, the first admission requirement being related to the first identification information; the first node sending a first model admission request, the first model admission request including first model admission test information of the first model trained according to the first admission requirement; and the first node receiving first model admission indication information, the first model admission indication information being related to the test results of the trained first model, the test results being obtained by testing the first model based on the first model admission test information.

[0005] Based on the above technical solution, the communication method provided in this application embodiment can train the model according to the received first admission requirements, and the obtained first model admission test information can be used to identify the first model that meets the update requirements, thereby realizing the deployment of the first model.

[0006] In conjunction with the first aspect, in some implementations of the first aspect, the first node sends the first model information of the trained first model, which includes the model structure information and model parameter information of the first model. Thus, when permission to update the model is granted, sending the first model information of the trained first model enables the deployment of the trained first model.

[0007] Secondly, embodiments of this application provide a communication method applied to a second node. For example, this method can be executed by the second node, which can be a network-side device, a terminal-side device, or an independent network element. The second node can also be a component (e.g., a circuit, processor, chip, or chip system), logic module, or software that implements all or part of the functions of the second node; this application does not limit this. The following description uses a second node as an example. The method includes: the second node receiving a first request, the first request including first identification information of a first model to be deployed; the second node sending a first admission requirement, the first admission requirement being related to the first identification information; the second node receiving a first model admission request, the first model admission request including first model admission test information of the first model trained according to the first admission requirement; and the second node sending first model admission indication information, the first model admission indication information being related to the test results of the trained first model, the test results being obtained by testing the first model based on the first model admission test information.

[0008] Based on the above technical solution, the communication method provided in this application embodiment can identify the first model that meets the update requirements by receiving the first model admission test information obtained by training the model according to the first admission requirements, and realize the deployment of the first model.

[0009] In conjunction with the second aspect, in some implementations of the second aspect, the second node receives the first model information of the trained first model, the first model information including the model structure information and the model parameter information of the first model. Thus, when model updates and upgrades are permitted, receiving the first model information of the trained first model enables the deployment of the trained first model.

[0010] Thirdly, embodiments of this application provide a communication method applied to a network-side device. For example, the method can be executed by a network-side device, which is a network device or a component of a network device (e.g., a circuit, processor, chip, or chip system), or it can be a logic module or software capable of implementing all or part of the functions of the network-side device. This application does not limit this. The following description uses a network-side device as an example. The method includes: the network-side device sending a first request to a third node, the first request including first identification information of a first model to be deployed; the network-side device receiving a first admission requirement from the third node, the first admission requirement being related to the first identification information; the network-side device sending first model training information to a terminal-side device, the first model training information including the first admission requirement; the network-side device sending a first model admission request to the third node, the first model admission request including first model admission test information of the first model trained according to the first admission requirement; and the network-side device receiving first model admission indication information, the first model admission indication information being related to the test result of the trained first model, the test result being obtained by testing the first model based on the first model admission test information.

[0011] Based on the above technical solution, the communication method provided in this application embodiment can train the model according to the received first admission requirements, and the obtained first model admission test information can be used to identify the first model that meets the update requirements, thereby realizing the deployment of the first model.

[0012] In conjunction with the third aspect, in some implementations of the third aspect, the network-side device receives a second model admission request from the terminal-side device. The second model admission request includes second model admission test information of the first model trained according to the first model training information; the first model admission test information is related to the second model admission test information. Thus, the network-side device can generate first model admission test information based on the received second model admission test information, which is used to identify the first model that meets the update requirements and to deploy the first model.

[0013] In conjunction with the third aspect, in some implementations of the third aspect, the network-side device sends the first model admission indication information to the terminal-side device. Thus, the terminal-side device can determine whether to deploy the first model based on the received first model admission indication information.

[0014] Fourthly, embodiments of this application provide a communication method applied to a terminal-side device. For example, the method can be executed by the terminal-side device, which is a terminal or a component within a terminal (e.g., a circuit, processor, chip, or chip system). Alternatively, it can be a logic module or software capable of implementing all or part of the terminal's functions; this application does not limit this. The following description uses a terminal-side device as an example. The method includes: the terminal-side device receiving first model training information, the first model training information including first admission requirements; the first model training information being used to train the first model.

[0015] Based on the above technical solution, the communication method provided in this application embodiment can ensure the effectiveness of the training results of the first model by training the first model according to the first admission requirements in the received first model training information.

[0016] In conjunction with the fourth aspect, in some implementations of the fourth aspect, the terminal-side device sends a second model admission request, which includes second model admission test information of the first model trained according to the first model training information. Thus, the terminal-side device can indicate the second model admission test information of the first model to the third node via the network-side device.

[0017] In conjunction with the fourth aspect, in some implementations of the fourth aspect, the terminal-side device receives first model admission indication information. Thus, the terminal-side device can determine whether to deploy the first model based on the received first model admission indication information.

[0018] In conjunction with the first, second, third, or fourth aspect, in some implementations of the first, second, third, or fourth aspect, the first admission requirement includes the input / output interface information of the first model. This ensures the effectiveness of the first model trained according to the first admission requirement.

[0019] In conjunction with the first, second, third, or fourth aspect, in some implementations of the first, second, third, or fourth aspect, if the first model admission indication information indicates that the trained first model is used in the current inference service, then the model update / upgrade flag is activated. Thus, by activating the model update / upgrade flag, the deployment of the first model is indicated.

[0020] In conjunction with the first, second, third, or fourth aspects, and in some implementations of the first, second, third, or fourth aspects, the format of the trained first model is adjusted based on the first admission requirement. Thus, the trained first model can be converted into the required model format for deployment according to the first admission requirement.

[0021] Fifthly, embodiments of this application provide an electronic device. This electronic device is used to perform the methods provided in the first, second, third, or fourth aspects described above. Specifically, the electronic device may include units and / or modules for performing the methods provided in the first aspect or any of the above-described implementations of the first aspect, such as a processing unit and an acquisition unit. Alternatively, the electronic device may include units and / or modules for performing the methods provided in the second aspect or any of the above-described implementations of the second aspect, such as a processing unit and an acquisition unit. Alternatively, the electronic device may include units and / or modules for performing the methods provided in the third aspect or any of the above-described implementations of the third aspect, such as a processing unit and an acquisition unit. Alternatively, the electronic device may include units and / or modules for performing the methods provided in the fourth aspect or any of the above-described implementations of the fourth aspect, such as a processing unit and an acquisition unit.

[0022] In conjunction with the fifth aspect, in some implementations of the fifth aspect, the electronic device is a first node, a second node, or a third node. The acquisition unit may include a transceiver, or an input / output interface; the processing unit may include at least one processor. Optionally, the transceiver may include transceiver circuitry. Optionally, the input / output interface may include input / output circuitry.

[0023] In conjunction with the fifth aspect, in some implementations of the fifth aspect, the electronic device is a chip, a chip system, or a circuit. The acquisition unit may include an input / output interface, interface circuit, output circuit, input circuit, or related circuit on the chip, chip system, or circuit; the processing unit may include at least one processor, processing circuit, or logic circuit.

[0024] In conjunction with the fifth aspect, in some implementations of the fifth aspect, the electronic device is a network-side device or a terminal-side device. The acquisition unit may include a transceiver, or an input / output interface; the processing unit may include at least one processor. Optionally, the transceiver may include transceiver circuitry. Optionally, the input / output interface may include input / output circuitry.

[0025] Sixthly, embodiments of this application provide a processor for executing the methods provided in the above aspects.

[0026] In a seventh aspect, embodiments of this application provide a computer-readable storage medium. This computer-readable storage medium stores computer program code, and when the computer program code is executed, the method provided by any one of the first to fourth aspects, and any implementation thereof, is performed.

[0027] Eighthly, embodiments of this application provide a computer program product containing instructions. When these instructions are executed on a computer, the computer performs any one of the first to fourth aspects described above, and the method provided by any implementation of any one of the first to fourth aspects.

[0028] Ninthly, embodiments of this application provide a chip, the chip including a processor and a communication interface, the processor reading computer programs or instructions stored in a memory through the communication interface, executing any one of the first to fourth aspects described above, and the method provided by any implementation of any one of the first to fourth aspects.

[0029] Alternatively, as one implementation, the chip may also include the memory.

[0030] Furthermore, the technical effects of the fifth to ninth aspects mentioned above can be referred to the technical effects of the methods described in the first to fourth aspects mentioned above, and will not be repeated here. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of this application or the background art, the accompanying drawings used in the embodiments of this application or the background art will be described below.

[0032] Figure 1 This is a schematic diagram of an AI module applicable to an embodiment of this application;

[0033] Figure 2 This is a schematic flowchart of a communication method 200 provided in an embodiment of this application;

[0034] Figure 3 This is a schematic flowchart of a communication method 300 provided in an embodiment of this application;

[0035] Figure 4 This is a schematic block diagram of a communication device 400 provided in an embodiment of this application;

[0036] Figure 5 This is a schematic block diagram of an electronic device 500 provided in an embodiment of this application;

[0037] Figure 6 This is a schematic diagram of the structure of an electronic device 600 provided in an embodiment of this application;

[0038] Figure 7 This is a schematic diagram of the structure of an electronic device 700 provided in an embodiment of this application. Detailed Implementation

[0039] The technical solutions in the embodiments of this application will now be described with reference to the accompanying drawings.

[0040] The terminology used in the following embodiments is for the purpose of describing specific embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” and “the” are intended to include expressions such as “one or more,” unless the context clearly indicates otherwise. It should also be understood that in the following embodiments of this application, “at least one,” “at least one,” and “one or more” refer to one, two, or more than two. “First,” “second,” and various numerical designations are merely distinctions for descriptive convenience and are not intended to limit the scope of the embodiments of this application. “And / or” is used to describe the correspondence between corresponding objects, indicating that three relationships can exist. For example, “A and / or B” can represent: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character “ / ” generally indicates that the preceding and following related objects are in an “or” relationship. The order of the process numbers below does not imply the order of execution. The execution order of each process should be determined by its function and internal logic and should not constitute any limitation on the implementation process of the embodiments of this application. For example, in the embodiments of this application, the words "201", "301", "401" etc. are merely identifiers made for the convenience of description and do not limit the order of execution steps.

[0041] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. In this application, the words "exemplary" or "for example" are used to indicate that something is illustrative, exemplary, or descriptive. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the words "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized. In the embodiments of this application, descriptions such as "when," "in the case of," "if," and "if" all refer to the fact that the device will perform a corresponding processing under certain objective circumstances, and are not a limitation on time, nor do they require the device to perform a judgment action during implementation, nor do they imply any other limitations.

[0042] The technical solutions of this application can be applied to various communication systems, such as: Long-Term Evolution (LTE) systems, LTE Frequency Division Duplex (FDD) systems, LTE Time Division Duplex (TDD) systems, Universal Mobile Telecommunication System (UMTS), Worldwide Interoperability for Microwave Access (Wi-MAX) systems, 5th Generation (5G) systems or New Radio (NR) systems, future communication systems (such as 6th Generation (6G) systems), Narrow Band Internet of Things (NB-IoT) systems, Global System for Mobile Communications (GSM), Enhanced Data Rate for GSM Evolution (EDGE) systems, Wideband Code Division Multiple Access (WCDMA) systems, Code Division Multiple Access 2000 (CDMA2000) systems, and Time Division-Synchronization Code Division Multiple Access (TDMA) systems. Non-terrestrial network (NTN) systems include inter-satellite communication (TD-SCDMA), inter-satellite communication, and satellite communication. Satellite communication systems include satellite base stations and terminal devices. The satellite base station provides communication services to the terminal devices. Satellite base stations can also communicate with each other. A satellite can act as a base station or as a terminal device. Satellites can refer to unmanned aerial vehicles (UAVs), hot air balloons, low-Earth orbit (LEO) satellites, medium-Earth orbit (MEO) satellites, high-Earth orbit (HEO) satellites, etc. Satellites can also refer to non-terrestrial base stations or non-terrestrial equipment.

[0043] The embodiments of this application can be applied to terminal-side devices. Terminal-side devices can be wireless or wired terminals. A wireless terminal can be a device that provides voice and / or data connectivity to a user, such as a handheld device, wearable device, computing device, or other processing device connected to a wireless modem with wireless connectivity. The wireless terminal can communicate with one or more core networks via a radio access network (RAN). The wireless terminal can be a mobile terminal, such as a mobile phone (or "cellular" phone), and a computer with a mobile terminal, for example, a portable, pocket-sized, handheld, computer-embedded, or vehicle-mounted mobile device, which exchanges voice and / or data with the radio access network. Examples of wireless terminals include personal communication service (PCS) telephones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, personal digital assistants (PDAs), subscriber units, smartphones, wireless data cards, computers, tablets, wireless modems, handsets, laptop computers, and machine type communication (MTC) terminals. Wireless terminals can also be referred to as systems, subscriber units (SUs), subscriber stations (SSs), mobile stations (MSs), mobile stations, remote stations (RSs), access points (APs), remote terminals (RTs), access terminals (ATs), user terminals (UTs), user agents (UAs), user devices (UDs), or user equipment (UEs).

[0044] In this application embodiment, the device for implementing the terminal's functions can be a terminal itself; or it can be a device capable of supporting the terminal in implementing these functions, such as a chip system. This device can be installed in the terminal or used in conjunction with the terminal. In this application embodiment, the chip system can be composed of chips, or it can include chips and other discrete components.

[0045] The technical solutions in this application embodiment can also be applied to network-side devices. A network-side device can be an access network device or a core network device. An access network device can be a device capable of connecting a terminal to a wireless network. This access network device can also be referred to as a radio access network (RAN) node, radio access network device, network device, or network-side device. For example, the access network device can be a base station. Core network devices can include a mobility management module (MME), access and mobility management function (AMF) network elements, network data analytics function (NWDAF) network elements, session management function (SMF) network elements, or other core network elements.

[0046] The base station in this application embodiment can broadly cover various names as follows, or be replaced by the following names, such as: NodeB, evolved NodeB (eNB), gNB in ​​a 5G network, relay station, access point, transceiver point.

[0047] A base station can be a macro base station, micro base station, relay node, donor node, or a combination thereof. It can also refer to a communication module, modem, or chip installed within the aforementioned equipment or apparatus. A base station can also be a network-side device in a future communication network, or a device that performs base station functions in a future communication system. A base station can support networks using the same or different access technologies.

[0048] Base stations can have a separate architecture of central units (CUs) and distributed units (DUs). The RAN can be connected to the core network (e.g., the LTE core network or the 5G core network). CUs and DUs can be understood as a logical functional division of the base station. Physically, CUs and DUs can be separate or deployed together. Multiple DUs can share a single CU. A single DU can also connect to multiple CUs. CUs and DUs can be connected via interfaces, such as F1 interfaces. CUs and DUs can be divided according to the protocol layers of the wireless network. For example, one possible division is: CUs perform functions of the radio resource control (RRC), service data adaptation protocol (SDAP), and packet data convergence protocol (PDCP) layers, while DUs perform functions of the radio link control (RLC), media access control (MAC), and physical layers. This division of CU and DU processing functions according to protocol layers is just one example; other division methods are also possible. For example, a CU or DU can be divided into functions with more protocol layers. Alternatively, a CU or DU can be divided into partial processing functions with protocol layers. In one design, some functions of the RLC layer and the protocol layer functions above the RLC layer are placed in the CU, while the remaining functions of the RLC layer and the protocol layer functions below the RLC layer are placed in the DU. In another design, the functions of the CU or DU can be divided according to service type or other system requirements. For example, based on latency, functions that need to meet latency requirements are placed in the DU, while functions that do not need to meet this latency requirement are placed in the CU. In yet another design, a CU can also have one or more core network functions. One or more CUs can be centrally located or separately located. For example, a CU can be located on the network side for convenient centralized management. A DU can have multiple radio frequency functions, or radio frequency functions can be remotely located.

[0049] The functionality of a CU can be implemented by a single entity or by different entities. For example, the CU's functionality can be further divided, such as separating the control plane (CP) and user plane (UP), i.e., the CU's control plane (CU-CP) and user plane (CU-UP). For instance, CU-CP and CU-UP can be implemented by different functional entities and connected via an E1 interface. CU-CP and CU-UP can be coupled with a DU to jointly complete the base station's functions. The CU's control plane, CU-CP, also includes a further segmented architecture, dividing the existing CU-CP into CU-CP1 and CU-CP2. CU-CP1 includes various radio resource management functions, while CU-CP2 only includes RRC functions and PDCP-C functions (i.e., the basic functions of control plane signaling at the PDCP layer).

[0050] Base stations can be fixed or mobile. For example, a helicopter or drone can be configured to act as a mobile base station, and one or more cells can move depending on the location of the mobile base station. In other examples, a helicopter or drone can be configured as a device to communicate with another base station.

[0051] The technical solutions in this application embodiment can also be applied to Operation Administration and Maintenance (OAM). OAM refers to the division of network management work into three main categories based on the actual needs of operator network operation: operation, administration, and maintenance. OAM can also be referred to as OAM entities or functions. Operation mainly involves the analysis, prediction, planning, and configuration of daily network and services; maintenance mainly involves daily operational activities such as testing and fault management of the network and its services. OAM can detect network operating status, optimize network connectivity and performance, improve network stability, and reduce network maintenance costs.

[0052] The technical solutions in this application embodiment can also be applied to third-party application services, such as over-the-top (OTT) services. OTT refers to providing various application services to users via the Internet. This type of application differs from the communication services currently provided by operators; it only utilizes the operator's network, while the service is provided by a third party outside the operator. Currently, typical OTT services include Internet TV services and app stores.

[0053] AI models, also known as AI algorithms (or AI operators), are a general term for mathematical algorithms built based on the principles of artificial intelligence, and form the foundation for using AI to solve specific problems. This application does not limit the type of AI model. For example, an AI model can be a machine learning model, a deep learning model, a reinforcement learning model, a federated learning model, etc.

[0054] In mobile communication systems, AI models can be used to intelligently collect and analyze data, thereby improving network performance and user experience.

[0055] For example, AI can be applied to CSI feedback enhancement. CSI is the channel attribute of the communication link, which is the channel quality information reported by the terminal to the base station. By reporting downlink channel quality information to the base station, the terminal can enable the base station to select a more suitable modulation and coding scheme (MCS) for the terminal, thus better adapting to changing wireless channels.

[0056] For example, AI can also be applied to beam management (BM). BM is mainly used to discover the most powerful transmit / receive beam pairs. AI-based beam prediction can improve prediction accuracy.

[0057] For example, AI can also be applied to positioning accuracy enhancement. Positioning accuracy is related to the number of total radiated power antennas (TPAs). Generally, the more TPAs ​​there are, the higher the positioning accuracy. AI models can achieve higher positioning accuracy with a smaller number of TPAs. For instance, traditional positioning methods, such as time difference of arrival (TDOA) and round trip time (RTT), rely on collecting line-of-sight (LOS) path information between the terminal and the TPAs. In indoor scenarios, there may not be a sufficient number of LOS paths, and traditional positioning methods cannot work well. Therefore, AI-based positioning can improve positioning accuracy in scenarios with limited LOS paths.

[0058] For example, AI can also be applied to network energy saving. Network energy saving can be achieved through cell activation / deactivation, load reduction, coverage improvement, or other RAN setting adjustments. Optimal energy-saving decisions depend on factors such as the load of different RAN nodes, RAN node capabilities, key performance indicator (KPI) requirements, quality of service (QoS) requirements, number of active users and terminal mobility, and cell utilization. However, improving network energy efficiency is a complex process; incorrect cell shutdowns and incorrect traffic offloading operations can lead to a decline in network performance and even energy efficiency. AI technology can be used to optimize energy-saving decisions by utilizing data collected in the RAN network. AI algorithms can predict the energy efficiency and load status for the next cycle, which can be used to better decide on cell activation / deactivation to save energy. Based on the predicted load, the system can dynamically configure energy-saving strategies to maintain a balance between system performance and energy efficiency, and reduce energy consumption.

[0059] For example, AI can also be applied to load balancing. The purpose of load balancing is to distribute the load evenly between cells and across areas within a cell, or to offload some traffic from congested cells, or to offload terminals on a single cell, carrier, or access standard, thereby improving network performance. This can be achieved by optimizing handover parameters and handover actions. This automated optimization can provide a high-quality user experience while increasing system capacity and minimizing manual intervention in network management and optimization tasks. AI-based solutions can be introduced to improve load balancing performance, such as inputting various user and network node measurements and feedback, historical data, etc., into AI models to enhance load balancing performance, thereby providing a higher-quality user experience and increasing system capacity.

[0060] For example, AI can also be applied to mobility optimization. Mobility management is a solution that ensures service continuity during terminal mobility by minimizing dropped calls, radio link failures (RLF), unnecessary handovers, and ping-pong effects. For future high-frequency networks, as the coverage area of ​​a single node decreases, the frequency of terminal handovers between nodes will become very high, especially for highly mobile terminals. Furthermore, for applications with strict QoS requirements such as reliability and latency, quality of experience (QoE) is highly sensitive to handover performance. AI-based solutions can be used to enhance mobility management in the following aspects: 1) reducing the probability of unexpected events, 2) predicting terminal location / mobility / performance, and 3) traffic redirection.

[0061] Before any AI model can be used to solve a specific technical problem, it needs to be trained. AI model training refers to using a specified initial model to compute on training data, and then adjusting the parameters of the initial model based on the computation results, so that the model gradually learns certain patterns and acquires specific functions. Once trained and possessing stable functionality, the AI ​​model can be used for inference. AI model inference is the process of using the trained AI model to compute on input data and obtain the expected inference result (also known as output data).

[0062] The AI ​​module is a module with AI learning and computing capabilities. In a wireless communication system, the AI ​​module can be located in the OAM (Operational Management Assistant), the gNB (or the CU in a separate architecture), some UEs, or as a standalone network element entity, such as the RAN intelligence controller (RIC). The main function of the AI ​​module in a wireless communication system is to perform a series of AI calculations, including model building, training approximation, and reinforcement learning, based on input data (e.g., network operation data provided by the RAN or monitored by the OAM, such as network load and channel quality). The trained model provided by the AI ​​module has predictive capabilities for changes in the RAN network and can typically be used for load prediction and UE trajectory prediction. Furthermore, the AI ​​module can also perform policy reasoning from the perspectives of network energy saving and mobility optimization based on the predicted RAN network performance results of the trained model, to obtain reasonable and efficient energy-saving strategies and mobility optimization strategies. When the AI ​​module is located in the OAM, its communication with the RAN-side gNB can reuse the current northbound interface; when the AI ​​module is located in the gNB or CU, it can reuse the current F1, Xn, Uu and other interfaces; when the AI ​​module is an independent network entity, it can communicate through communication links to the OAM and RAN sides, such as wired links or wireless links.

[0063] Figure 1 This is a schematic diagram illustrating an AI module 100 applicable to an embodiment of this application. For example... Figure 1 The AI ​​module 100 shown includes a database module 101, a training module 102, a model module 103, and an execution module 104.

[0064] Database module 101 can store training data. The training data can also come from terminal-side devices or network-side devices. For example, training data can come from a base station (e.g., gNB) or functional units that make up the base station (e.g., CU or DU). Alternatively, training data can come from other network-side devices besides the base station, such as gateways, management entities (e.g., MME or other core network equipment).

[0065] Training module 102 analyzes the training data provided by database module 101 to obtain an AI model. Training module 102 can then send the trained AI model to model module 103. After completing AI model training, training module 102 can also update or fine-tune the trained model and send the model parameters used for updating or fine-tuning to model module 103. During the operation of the AI ​​model, model module 103 can also collect some model runtime data and send this data to training module 102. Training module 102 can update the AI ​​model based on this runtime data.

[0066] The model module 103 can determine output data based on the AI ​​model and input data. The output data may include predictions of network operation based on the input data and the AI ​​model. The output data can also be used to adjust the AI ​​model strategy. In some embodiments, the network-side device and / or the terminal-side device can directly send input data to the model module 103. In other embodiments, the database module 101 may also collect data from the network-side device and / or the terminal-side device, determine the input data, and send the input data to the model module 103.

[0067] The execution module 104 can be used to execute the adjustment strategy determined by the model module 103. The execution module 104 can also collect the specific performance of the network after applying the adjustment strategy, such as network performance parameters, and feed this information back to the database module 101. The database module 101 can store this feedback information. This feedback information can be used for subsequent model training or to improve the AI ​​model.

[0068] The communication methods provided in the embodiments of this application (corresponding to methods 200 and 300 below) are described in detail below with reference to the accompanying drawings.

[0069] It should be noted that, in the embodiments of this application, the first node and / or the second node may be located in the terminal-side device, or in different network-side devices (such as access network devices, core network devices, etc.) in the wireless communication system formed with the terminal-side device, or in OAM, or in the application service OTT provided by a third party, or the first node may be an independent network element (such as a server of an equipment vendor), etc., to realize the function of updating the model.

[0070] Furthermore, this application embodiment involves input / output interface information, reference model structure information, model parameter accuracy information, model storage method information, model saving and loading method information, model performance requirements information, and model latency requirements information. The names used for this information are merely for convenience in describing this embodiment and do not affect the scope of protection of this application. In the future development of technology, these information may have other names, but if their function is the same as that described in this application embodiment, they should also be within the scope of protection of this application.

[0071] Figure 2 This is a schematic flowchart illustrating a communication method 200 provided in this application embodiment from the perspective of device interaction. As shown in the figure, the method 200 may include steps S201 to S204. The steps of method 200 are described in detail below.

[0072] S201, the first node sends a first request to the second node to request the deployment of the first model.

[0073] Specifically, the first request includes first identification information of the first model to be deployed. This first identification information indicates the first model.

[0074] Optionally, the model identifier corresponds one-to-one with the model type, and the first identifier information includes the model type of the first model. As shown in Table 1, model identifier A corresponds to a model of type "CSI compressed feedback", model identifier B corresponds to a model of type "beam management", and model identifier C corresponds to a model of type "positioning".

[0075] Table 1

[0076] Model Identifier Model type A CSI compression feedback B Beam management C position … …

[0077] Optionally, the model identifier indicates that the first model is one of multiple models of a certain type. The first identifier information includes the identifier of the first model, and the type of the first model is type 1. Type 1 also includes other models besides the first model, such as the second model. As shown in Table 2, there are multiple models of type "CSI Compressed Feedback": A1, A2, ..., representing "CSI Compressed Feedback" models with different performance indicators. The first identifier information includes the identifier of the first model, such as A1, indicating the first model identified as A1 among the "CSI Compressed Feedback" models. Similarly, there can be multiple models of type "Beam Management": B1, B2, ..., and multiple models of type "Location": C1, C2, ... Besides "CSI Compressed Feedback," "Beam Management," and "Location," there can be more model types.

[0078] Table 2

[0079]

[0080]

[0081] It should be understood that different types of models correspond to different model identifiers.

[0082] S202, the first node receives a first admission request from the second node, which is related to the first identification information.

[0083] Specifically, the second node determines the first admission requirements of the first model based on the first identification information of the first model. The first admission requirements indicate the conditions and criteria that need to be met to deploy the first model.

[0084] Optionally, the first admission requirement includes input / output interface information. This input / output interface information includes, but is not limited to, at least one of input / output interface dimension information, model format information, and input data precision information.

[0085] Optionally, the first admission requirement may further include one or more of the following: reference model structure information, model parameter accuracy information, model storage method information, model saving and loading method information, model performance requirement information, and model latency requirement information. Specifically, the reference model structure information indicates the model structure that the first model can adopt; the model parameter accuracy information indicates the accuracy requirements for the model parameters of the first model; the model storage method information indicates the technical methods used to save the model and describe the model inference operations and control flow; the model performance requirement information indicates the inference performance requirements of the current AI inference service for the first model; and the model latency requirement information indicates the maximum latency that the current AI inference service can support.

[0086] Optionally, the reference model structure information includes at least one of the following: trainable or fine-tunable model parameter information, underlying supported operator information, model structure description information, and reference model structure examples. For example, the model structure description information indicates that the first model's model structure is a transformer structure.

[0087] Optionally, the model parameter accuracy information includes at least one of fixed quantization accuracy, variable quantization accuracy, maximum quantization accuracy, and minimum quantization accuracy.

[0088] Optionally, the model storage method information indicates that the first model is saved as a static graph, which includes the structural information and parameter information of the first model.

[0089] Optionally, the model saving and loading method information includes at least one of the following: the external interaction method of the first model, the storage location of the first model, the compilation method of the first model, and the upgrade method of the first model. For example, the external interaction method of the first model indicates the C++ interface corresponding to the first model or the FPGA interface corresponding to the first model.

[0090] S203, the first node sends a first model admission request to the second node, the first model admission request including the first model admission test information of the first model trained according to the first admission requirements.

[0091] Optionally, the first model admission test information includes the model structure information and model parameter information of the first model, and / or, the first model admission test information includes the local test results of the first model.

[0092] Optionally, if the first model admission test information includes the model structure information and the model parameter information of the first model, the second node can reconstruct the first model based on the received first model admission test information.

[0093] Optionally, if the first model admission test information includes the local test results of the first model, the local test results indicate the performance of the first model.

[0094] S204, the first node receives the first model admission instruction information from the second node.

[0095] Specifically, the first model admission indication information is related to the test results of the trained first model, which are obtained by testing the first model based on the first model admission test information.

[0096] The first model admission indication information indicates whether the deployment of the first model is permitted. For example, the indication field occupied by the first model admission indication information includes 1 bit. If the value of the indication field is 1, it indicates that the deployment of the first model is permitted; if the value of the indication field is 0, it indicates that the deployment of the first model is not permitted. Alternatively, if the value of the indication field is 0, it indicates that the deployment of the first model is permitted; if the value of the indication field is 1, it indicates that the deployment of the first model is not permitted.

[0097] The test results of the first model are obtained by testing the first model based on the first model admission test information. When the test results of the first model meet the requirements, the first model admission indication information indicates that the deployment of the first model is permitted; when the test results of the first model fail to meet the requirements, the first model admission indication information indicates that the deployment of the first model is not permitted. For example, when the first model is used for CSI compressed feedback, the test results of the first model can be squared generalized cosine similarity (SGCS). This SGCS measures the similarity between the model output and the output validation label, indicating the accuracy of the CSI compressed feedback. The value of the SGCS is between 0 and 1, with a value closer to 1 indicating a higher similarity between the model output and the output validation label. When both are completely identical, the value reaches its maximum value of 1. When the pass rate of the model test results set by the second node is 0.99, if the test results of the first model are greater than or equal to 0.99, it means that the test results of the first model meet the requirements and the first model can be deployed. Conversely, if the test result of the first model is lower than 0.99, it means that the test result of the first model cannot meet the requirements and the first model cannot be deployed.

[0098] Optionally, when the first model admission test information includes the model structure information and the model parameter information of the first model, the second node reconstructs the first model according to the first model admission test information and tests the first model to obtain the test results of the first model.

[0099] Optionally, when the admission test information of the first model includes the local test results of the first model, the test results of the first model are the local test results of the first model, or the test results of the first model are a transformation of the local test results of the first model. For example, when the local test results of the first model and the test results of the first model are measured using squared generalized cosine similarity and generalized cosine similarity (GCS) respectively, if the local test results of the first model are 0.9, the test results of the first model obtained by transforming the local test results are 0.81.

[0100] Optionally, after step S204, the method 200 further includes the following steps S205 and S206:

[0101] S205, the second node receives the first model information from the first node.

[0102] Specifically, the first model information includes the model structure information and the model parameter information of the first model. When the first model admission indication information indicates that the deployment of the first model is permitted on the second node, and the first model admission request sent by the first node to the second node does not contain the model structure information and the model parameter information of the first model, the first node sends the first model information to the second node.

[0103] S206, Activation model update and upgrade identifier.

[0104] Specifically, if the first model admission instruction information indicates that the trained first model is used in the current inference service, then the model update and upgrade flag is activated.

[0105] When the first model is deployed on the first node, the update and upgrade flag is activated on the first node, indicating that the first model is deployed on the first node; when the first model is deployed on the second node, the update and upgrade flag is activated on the second node, indicating that the first model is deployed on the second node.

[0106] Optionally, before step S203, the method 200 further includes the following steps S207 and S208:

[0107] S207, Train the first model according to the first admission requirement.

[0108] Specifically, a model training method for the first model is determined based on a first admission requirement, and this method is used to train the first model. The model training method includes determining the model structure and / or determining a model training strategy.

[0109] Optionally, the model structure of the first model can be a model structure indicated by the reference model structure information included in the first admission requirement, or it can be a custom model structure.

[0110] Alternatively, the model training strategy can be either fine-tuning the model or training the model.

[0111] S208, adjust the format of the trained first model according to the first admission requirements.

[0112] Specifically, based on the first admission requirements, the format of the trained first model is adjusted to the required deployment format.

[0113] Optionally, the trained first model can be converted into low-level code, enabling the first or second node that has deployed the first model to call the first model through the low-level code.

[0114] Optionally, the first node or the second node deploys the first model via an FPGA interface.

[0115] Figure 3This is a schematic flowchart illustrating a communication method 300 provided in this application embodiment from the perspective of device interaction. As shown in the figure, the method 300 may include steps S301 to S305. The steps of method 300 are described in detail below.

[0116] S301, the network-side device sends a first request to the third node to request the deployment of the first model.

[0117] Specifically, the first request includes first identification information of a first model to be deployed. This first identification information indicates the first model. The network-side device informs the third node of the information about the first model to be deployed by sending the first identification information.

[0118] The third node can be located in an application service OTT provided by a third party, or it can be an independent network element (such as a server of an equipment vendor) to realize the function of updating the model.

[0119] For a description of the first identifier information, please refer to the relevant description above.

[0120] S302, the network-side device receives a first access request from a third node, which is related to first identification information.

[0121] Specifically, the third node determines the first admission requirements of the first model based on the first identification information of the first model. The first admission requirements indicate the conditions and criteria that need to be met to deploy the first model.

[0122] For a description of the first admission requirements, please refer to the relevant description above.

[0123] S303, the network-side device sends first model training information to the terminal-side device, the first model training information including first admission requirements.

[0124] Optionally, the first model training information also includes data acquisition instruction information, which instructs the terminal device to acquire first training data for training the first model.

[0125] S304, the network-side device sends a first model admission request to the third node. The first model admission request includes first model admission test information of the first model trained according to the first admission requirements.

[0126] For a description of the first model admission test information, please refer to the previous text.

[0127] S305, the network-side device receives the first model admission instruction information from the third node.

[0128] Please refer to the previous text for a description of the first model admission instruction information.

[0129] Optionally, before step S304, the method 300 further includes the following step S306:

[0130] S306, the network-side device receives a second model admission request from the terminal-side device.

[0131] Specifically, the second model admission request includes the second model admission test information of the first model trained by the terminal device according to the first model training information, and the first model admission test information and the second model admission test information are related.

[0132] Optionally, the second model access test information includes the model structure information and model parameter information of the first model to be deployed on the terminal device, and / or, the second model access test information includes the local test results of the first model obtained by the terminal device.

[0133] Optionally, the first model access test information includes the second model access test information. For example, the second model access test information includes the model structure information and model parameter information of the first model to be deployed on the terminal-side device, and the first model access test information includes the second model access test information and the model structure information and model parameter information of the first model to be deployed on the network-side device.

[0134] Optionally, the first model admission test information is the same as the second model admission test information. For example, the second model admission test information includes the local test results of the first model obtained by the terminal-side device, and the network-side device directly sends the local test results to the third node through the first model admission test information.

[0135] Optionally, the first model admission test information is a transformation of the second model admission test information. For example, the network-side device and the terminal-side device respectively use squared generalized cosine similarity and generalized cosine similarity to measure the test result and local test result of the first model. When the second model admission test information indicates that the local test result of the first model obtained by the terminal-side device is 0.9, according to the transformation relationship between squared generalized cosine similarity and generalized cosine similarity, the first model admission test information sent by the network-side device to the third node indicates that the test result of the first model is 0.81.

[0136] Optionally, the terminal device sends the second model admission request to the network device via an AMF non-access stratum (NAS) message.

[0137] Optionally, after step S305, the method 300 further includes the following steps S307 and S308:

[0138] S307, the network-side device sends the first model admission instruction information to the terminal-side device.

[0139] Specifically, the first model admission indication information indicates whether the deployment of the first model is permitted. For example, the indication field occupied by the first model admission indication information includes 1 bit. If the value of the indication field is 1, it indicates that the deployment of the first model is permitted; if the value of the indication field is 0, it indicates that the deployment of the first model is not permitted. Alternatively, if the value of the indication field is 0, it indicates that the deployment of the first model is permitted; if the value of the indication field is 1, it indicates that the deployment of the first model is not permitted.

[0140] Optionally, the network-side device may send the first model admission instruction information to the terminal-side device via the RAN system information block (SIB) or via the AMF NAS message.

[0141] S308, Activation Model Update and Upgrade Indicator.

[0142] Specifically, if the first model admission instruction information indicates that the trained first model is used in the current inference service, then the model update and upgrade flag is activated.

[0143] When the first model is updated and deployed on both the network-side device and the terminal-side device, the update and upgrade identifier is activated on both the network-side device and the terminal-side device; when the first model is updated and deployed only on the network-side device, the update and upgrade identifier is activated only on the network-side device; when the first model is updated and deployed only on the terminal-side device, the update and upgrade identifier is activated only on the terminal-side device.

[0144] Optionally, after step S303, the method 300 further includes the following steps S309 and S310:

[0145] S309, Train the first model according to the first admission requirement.

[0146] Specifically, a model training method for the first model is determined based on a first admission requirement, and this method is used to train the first model. The model training method includes determining the model structure and / or determining a model training strategy.

[0147] The model training method for the first model is determined based on the first admission requirement and used to train the first model, including:

[0148] The network-side device determines the model training method for the first model on both sides or the first model on the network-side device side based on the first admission requirement;

[0149] or,

[0150] The terminal device determines the model training method for the first model on both sides or the first model of the terminal device based on the first admission requirements included in the first model training information;

[0151] or,

[0152] The network-side device and the terminal-side device jointly determine the training method for the first model based on the first admission requirements.

[0153] Optionally, the model structure of the first model can be a model structure indicated by the reference model structure information included in the first admission requirement, or it can be a custom model structure.

[0154] Alternatively, the model training strategy can be either fine-tuning the model or training the model.

[0155] S310, adjust the format of the trained first model according to the first admission requirements.

[0156] Specifically, the network-side device and / or the terminal-side device adjust the format of the trained first model on the network side and / or the format of the first model on the terminal side to the required deployment format according to the first admission requirements.

[0157] Optionally, the trained first model is converted into low-level code, enabling network-side devices and / or terminal-side devices that have deployed the first model to call the first model through the low-level code.

[0158] Optionally, the network-side device and / or the terminal-side device deploy the first model via an FPGA interface.

[0159] Figure 4 This is a schematic block diagram of a communication device 400 provided in an embodiment of this application. The communication device 400 includes a receiving module 401, which can be used to implement corresponding receiving functions. The receiving module 401 can also be referred to as a receiving unit.

[0160] The communication device 400 also includes a processing module 402, which can be used to implement corresponding processing functions.

[0161] The communication device 400 also includes a transmitting module 403, which can be used to implement the corresponding transmitting function. The transmitting module 403 can also be called a transmitting unit.

[0162] The communication device 400 can be used to execute the actions performed by the first node, second node, third node, network-side device, or terminal-side device in the various method embodiments described above. In this case, the communication device 400 can be a component of the first node, second node, third node, network-side device, or terminal-side device. The receiving module 401 is used to execute the receiving-related operations of the first node, second node, third node, network-side device, or terminal-side device in the method embodiments described above. The processing module 402 is used to execute the processing-related operations of the first node, second node, third node, network-side device, or terminal-side device in the method embodiments described above. The sending module 403 is used to execute the sending-related operations of the first node, second node, third node, network-side device, or terminal-side device in the method embodiments described above.

[0163] As a design feature, the communication device 400 is used to perform the actions performed by any device in the various method embodiments described above (method 200 or method 300). In one embodiment, the communication device 400 can be used to perform the aforementioned... Figure 2 The operation of the first node. For example:

[0164] Sending module 403 is used to send a first request, which is used to request the deployment of the first model.

[0165] The receiving module 401 is used to receive a first admission requirement, which is related to the first identification information included in the first request.

[0166] Processing module 402 is used to train the first model according to the first admission requirements.

[0167] It should be understood that the specific process of each module performing the above-mentioned steps has been described in detail in the above method embodiments, and will not be repeated here for the sake of brevity.

[0168] In addition, the receiving module 401, processing module 402 and sending module 403 in the communication device 400 can also implement other operations or functions of the first node in the above method, which will not be described in detail here.

[0169] Optionally, the communication device 400 may include a first node. Alternatively, the communication device 400 may be a component configured in the first node, such as a chip in the first node. In this case, the receiving module 401 and the transmitting module 403 may be interface circuits, etc. Specifically, the interface circuit may include input circuits and output circuits, wherein the receiving module 401 may include input circuits, the transmitting module 403 may include output circuits, and the processing module 402 may include processing circuits.

[0170] In another embodiment, the communication device 400 can be used to perform the above. Figure 2 The operation of the second node. For example:

[0171] The receiving module 401 is used to receive a first request, which is used to request the deployment of a first model.

[0172] Processing module 402 is used to determine the first access requirement based on the first identification information included in the first request.

[0173] The sending module 403 is used to send the first admission requirement.

[0174] In addition, the receiving module 401, processing module 402 and transmitting module 403 in the communication device 400 can also implement other operations or functions of the second node in the above method, which will not be described in detail here.

[0175] Optionally, the communication device 400 may include a second node. Alternatively, the communication device 400 may be a component configured in the second node, such as a chip in the second node. In this case, the receiving module 401 and the transmitting module 403 may be interface circuits, etc. Specifically, the interface circuit may include input circuits and output circuits, wherein the receiving module 401 may include input circuits, the transmitting module 403 may include output circuits, and the processing module 402 may include processing circuits.

[0176] In another embodiment, the communication device 400 can be used to perform the above. Figure 3 Operation of network-side devices. For example:

[0177] Sending module 403 is used to send a first request, which is used to request the deployment of the first model.

[0178] The receiving module 401 is used to receive a first admission requirement, which is related to the first identification information included in the first request.

[0179] Processing module 402 is used to train the first model according to the first admission requirements.

[0180] In addition, the receiving module 401, processing module 402 and transmitting module 403 in the communication device 400 can also implement other operations or functions of the network-side device in the above method, which will not be described in detail here.

[0181] Optionally, the communication device 400 may include a network-side device. Alternatively, the communication device 400 may be a component configured in the network-side device, such as a chip in the network-side device. In this case, the receiving module 401 and the transmitting module 403 may be interface circuits, etc. Specifically, the interface circuit may include input circuits and output circuits, wherein the receiving module 401 may include input circuits, the transmitting module 403 may include output circuits, and the processing module 402 may include processing circuits.

[0182] In another embodiment, the communication device 400 can be used to perform the above. Figure 3 Operation of the terminal-side device. For example:

[0183] The receiving module 401 is used to receive first model training information, which includes first admission requirements.

[0184] Processing module 402 is used to train the first model according to the first admission requirements.

[0185] The sending module 403 is used to send the second model admission request.

[0186] In addition, the receiving module 401, processing module 402 and transmitting module 403 in the communication device 400 can also implement other operations or functions of the terminal side device in the above method, which will not be described in detail here.

[0187] Optionally, the communication device 400 may include a terminal-side device. Alternatively, the communication device 400 may be a component configured in the terminal-side device, such as a chip in the terminal-side device. In this case, the receiving module 401 and the transmitting module 403 may be interface circuits, etc. Specifically, the interface circuit may include input circuits and output circuits, wherein the receiving module 401 may include input circuits, the transmitting module 403 may include output circuits, and the processing module 402 may include processing circuits.

[0188] For details on how each module performs the corresponding steps described above, please refer to the above method implementation examples.

[0189] Figure 5 This is a schematic structural diagram of another electronic device 500 provided in an embodiment of this application. The electronic device 500 includes one or more processors 501, which are single-core or multi-core processors. Optionally, the electronic device 500 may further include at least one memory 502 for storing computer programs or instructions and / or data. The memory 502 is coupled to the processor 501, and the processor 501 is used to execute the computer programs or instructions and / or data stored in the memory 502, causing the methods (methods 200 or 300) in the above method embodiments to be executed. The coupling in the embodiments of this application is an indirect coupling or communication connection between devices, units, or modules, which can be electrical, mechanical, or other forms, for information interaction between devices, units, or modules.

[0190] Optionally, the electronic device 500 may include one or more processors 501.

[0191] Alternatively, the memory 502 can be integrated with the processor 501, or it can be set separately.

[0192] The electronic device 500 may further include a transceiver 503 for communicating with other devices via a transmission medium, thereby enabling the device to communicate with other devices. Optionally, the transceiver 503 may be an interface, a bus, a circuit, or a device capable of transmitting and receiving functions.

[0193] Alternatively, the device in transceiver 503 used to implement the receiving function can be regarded as a receiving module, and the device in transceiver 503 used to implement the transmitting function can be regarded as a transmitting module. That is, transceiver 503 includes a receiver and a transmitter.

[0194] This application embodiment does not limit the specific connection medium between the processor 501, memory 502, and transceiver 503. This application embodiment... Figure 5 The processor 501, memory 502, and transceiver 503 are connected via a bus, and the bus is in... Figure 5 The connections between other components are indicated by thick lines and are for illustrative purposes only, not as limiting information. The bus can be divided into address bus, data bus, control bus, etc.

[0195] For ease of representation, Figure 5 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0196] Optionally, such as Figure 5 As shown, the electronic device 500 may also include a transceiver 503 and / or a communication interface, which are used for receiving and / or transmitting signals. For example, the processor 501 is used to control the transceiver 503 and / or the communication interface to receive and / or transmit data.

[0197] A transceiver is sometimes also called a transceiver unit, transceiver module, or transceiver circuit. A receiver is sometimes also called a receiver unit, receiver module, or receiver circuit. A transmitter is sometimes also called a transmitter, transmitter module, or transmitter circuit.

[0198] For example, in one embodiment, processor 501 is configured to implement other operations or functions of the first node. Transceiver 503 is used to enable communication between electronic device 500 and the second node.

[0199] In another embodiment, processor 501 is configured to implement other operations or functions of the second node. Transceiver 503 is used to enable communication between electronic device 500 and the first node.

[0200] In another embodiment, processor 501 is configured to implement other operations or functions of the network-side device. Transceiver 503 is used to enable communication between electronic device 500 and a third node or terminal-side device.

[0201] In another embodiment, processor 501 is configured to implement other operations or functions of the terminal-side device. Transceiver 503 is used to enable communication between electronic device 500 and network-side device.

[0202] One or more of the above modules or units can be implemented by software, hardware, or a combination of both. When any of the above modules or units is implemented by software, the software exists as computer program instructions and is stored in memory. The processor can be used to execute the program instructions and implement the above method flow. The processor can include, but is not limited to, at least one of the following: a central processing unit (CPU), a microprocessor, a digital signal processor (DSP), a microcontroller unit (MCU), or an artificial intelligence processor, etc., and various computing devices that run software. Each computing device may include one or more cores for executing software instructions to perform calculations or processing. The processor can be built into a system-on-chip (SoC) or an application-specific integrated circuit (ASIC), or it can be a separate semiconductor chip. In addition to the cores for executing software instructions to perform calculations or processing, the processor may further include necessary hardware accelerators, such as field-programmable gate arrays (FPGAs), programmable logic devices (PLDs), or logic circuits that implement dedicated logic operations.

[0203] When the above modules or units are implemented in hardware, the hardware can be any one or any combination of CPU, microprocessor, DSP, MCU, artificial intelligence processor, ASIC, SoC, FPGA, PLD, special purpose digital circuit, hardware accelerator or non-integrated discrete device, which can run the necessary software or perform the above method flow independently of software.

[0204] When the above modules or units are implemented using software, they can be implemented in whole or in part as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0205] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above description is only a specific embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solution of this application should be included within the scope of protection of this application.

[0206] This application provides an electronic device 600, which can be a first node / second node / network-side device or a chip. The electronic device 600 can be used to perform the operations executed by the first node / second node / network-side device in the above-described method embodiments (method 200 or method 300).

[0207] When the electronic device 600 is a first node / second node / network side device Figure 6A simplified structural diagram of a first node / second node / network side device is shown. The first node / second node / network side device includes parts 610 and 620. Part 610 includes an antenna and radio frequency (RF) circuitry. The antenna is mainly used for transmitting and receiving RF signals, and the RF circuitry is mainly used for converting RF signals to baseband signals. Part 620 includes a memory and a processor, mainly used for baseband processing and controlling model management network elements. Part 610 is commonly referred to as a transceiver unit, transceiver, transceiver circuit, or transceiver. Part 620 is typically the control center of the first node / second node / network side device, often referred to as a processing unit, used to control the first node / second node / network side device to perform the processing operations described in the above method embodiments.

[0208] Optionally, the devices used to implement the receiving function in part 610 can be regarded as receiving units, and the devices used to implement the transmitting function can be regarded as transmitting units. That is, part 610 includes receiving units and transmitting units. The receiving unit can also be called a receiver, receiver circuit, etc., and the transmitting unit can be called a transmitter, transmitter, or transmitting circuit, etc.

[0209] When data needs to be transmitted, the processor performs baseband processing on the data to be transmitted and outputs the baseband signal to the radio frequency (RF) circuit. The RF circuit then processes the baseband signal and transmits it outward as an electromagnetic wave through the antenna. When data is sent to the model management network element, the RF circuit receives the RF signal through the antenna, converts it into a baseband signal, and outputs the baseband signal to the processor. The processor then converts the baseband signal back into data and processes it.

[0210] The 620 section may include one or more single boards, and each single board may include one or more processors and one or more memories. For ease of illustration, Figure 6 Only one memory and processor are shown. The processor is used to read and execute programs in the memory to implement baseband processing functions and control the model management network elements. If multiple boards exist, they can be interconnected to enhance processing capabilities. As an optional implementation, multiple boards can share one or more processors, or multiple boards can share one or more memories.

[0211] It should be understood that Figure 6 This is merely an example and not a limitation; the first node / second node / network-side device, including the transceiver unit and processing unit, described above may not depend on... Figure 6 The structure shown.

[0212] When the device 600 is a chip, the chip includes a transceiver unit and a processing unit. The transceiver unit can be an input / output circuit or a communication interface; the processing unit is a processor, microprocessor, or integrated circuit integrated on the chip.

[0213] This application also provides another electronic device 700, which may be a first node / second node / terminal-side device or a chip. This electronic device 700 can be used to perform the operations executed by the first node / second node / terminal-side device in the above-described method embodiments (method 200 or method 300).

[0214] When the electronic device 700 is a first node / second node / terminal side device Figure 7 A simplified structural diagram of a first node / second node / terminal side device is shown. Figure 7 As shown, the first node / second node / terminal-side device includes a processor, memory, radio frequency circuitry, antenna, and input / output devices. The processor is primarily used for processing communication protocols and data, controlling the first node / second node / terminal-side device, executing software programs, and processing software program data. The memory is mainly used for storing software programs and data.

[0215] Radio frequency (RF) circuits are primarily used for converting baseband signals to RF signals and processing RF signals. Antennas are mainly used for transmitting and receiving RF signals in the form of electromagnetic waves. Input / output devices, such as touchscreens, displays, and keyboards, are mainly used for receiving user input data and outputting data to the user. It should be noted that some types of first-node / second-node / terminal-side devices may not have input / output devices.

[0216] When data needs to be transmitted, the processor performs baseband processing on the data to be transmitted and outputs a baseband signal to the radio frequency (RF) circuit. The RF circuit then processes the baseband signal and transmits it outward as an electromagnetic wave through the antenna. When data is sent to the first node / second node / terminal device, the RF circuit receives the RF signal through the antenna, converts it into a baseband signal, and outputs the baseband signal to the processor. The processor converts the baseband signal back into data and processes it. For ease of explanation, Figure 7 Only one memory and processor are shown in the illustration. In actual terminal-side device products, there may be one or more processors and one or more memories. Memory can also be called storage medium or storage device, etc. Memory can be set up independently of the processor or integrated with the processor; this application does not limit this.

[0217] In the embodiments of this application, the antenna and radio frequency circuit with transceiver function can be regarded as the transceiver unit of the first node / second node / terminal side device, and the processor with processing function can be regarded as the processing unit of the first node / second node / terminal side device.

[0218] like Figure 7 As shown, the first node / second node / terminal side device includes a transceiver unit 10 and a processing unit 20. The transceiver unit 10 can also be referred to as a transceiver, transceiver device, transceiver circuit, etc. The processing unit 20 can also be referred to as a processor, processing board, processing module, processing device, etc.

[0219] Optionally, the devices in transceiver unit 10 used to implement the receiving function can be regarded as receiving units, and the devices in transceiver unit 10 used to implement the transmitting function can be regarded as transmitting units. That is, transceiver unit 10 includes receiving units and transmitting units. The receiving unit may also be called a receiver, receiver device, receiving circuit, etc. The transmitting unit may also be called a transmitter, transmitter, transmitting device, transmitting circuit, etc.

[0220] It should be understood that Figure 7 This is merely an example and not a limitation; the first node / second node / terminal-side device mentioned above, including the transceiver unit and the processing unit, may not depend on... Figure 7 The structure shown.

[0221] When the device 700 includes a chip 30, the chip 30 includes a processing unit 20. The processing unit can be a processor, microprocessor, or integrated circuit integrated on the chip.

[0222] Optionally, chip 30 also includes a storage unit.

[0223] According to the method provided in the embodiments of this application, this application also provides a computer program product, which includes: computer program code, which, when run on a computer, causes the computer to execute the method of the first node in the aforementioned method embodiments.

[0224] According to the method provided in the embodiments of this application, this application also provides a computer program product, which includes: computer program code, which, when run on a computer, causes the computer to execute the method of the second node in the aforementioned method embodiments.

[0225] According to the method provided in the embodiments of this application, this application also provides a computer program product, which includes: computer program code, which, when run on a computer, causes the computer to perform the method of the network-side device in the foregoing method embodiments.

[0226] According to the method provided in the embodiments of this application, this application also provides a computer program product, which includes: computer program code, which, when run on a computer, causes the computer to execute the method of the terminal-side device in the foregoing method embodiments.

[0227] According to the method provided in the embodiments of this application, this application also provides a computer-readable medium storing program code, which, when run on a computer, causes the computer to execute the method of the first node in the foregoing method embodiments.

[0228] According to the method provided in the embodiments of this application, this application also provides a computer-readable medium storing program code, which, when run on a computer, causes the computer to execute the method of the second node in the foregoing method embodiments.

[0229] According to the method provided in the embodiments of this application, this application also provides a computer-readable medium storing program code, which, when run on a computer, causes the computer to perform the method of the network-side device in the foregoing method embodiments.

[0230] According to the method provided in the embodiments of this application, this application also provides a computer-readable medium storing program code, which, when run on a computer, causes the computer to perform the method of the terminal-side device in the foregoing method embodiments.

[0231] This application also provides a processing device, including a processor and an interface; the processor is used to execute the communication method in any of the above method embodiments.

[0232] This application also provides a communication system, which includes a first node and a second node as described in the above embodiments.

[0233] This application also provides a communication system, which includes a third node, a network-side device, and a terminal-side device as described in the above embodiments.

[0234] As used in this specification, the terms "component," "module," "system," etc., are used to refer to computer-related entities, hardware, firmware, combinations of hardware and software, software, or software in execution. For example, a component can be, but is not limited to, a process running on a processor, a processor, an object, an executable file, an execution thread, a program, and / or a computer. As illustrated, applications running on computing devices and computing devices can both be components. One or more components may reside in a process and / or an execution thread, and components may be located on a single computer and / or distributed among two or more computers. Furthermore, these components can be executed from various computer-readable media on which various data structures are stored. Components can communicate, for example, via local and / or remote processes based on signals having one or more data packets (e.g., data from two components interacting with another component between a local system, a distributed system, and / or a network, such as the Internet interacting with other systems via signals).

[0235] Those skilled in the art will recognize that the various illustrative logical blocks and steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0236] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0237] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0238] It should be understood that "at least one" in the embodiments of this application refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, and c can represent: a, or, b, or, c, or, a and b, or, a and c, or, b and c, or, a, b, and c. Here, a, b, and c can be single or multiple.

[0239] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0240] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0241] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

Claims

1. A communication method, characterized in that, include: Send a first request, the first request including the first identification information of the first model to be deployed; Receive a first admission request, which is related to the first identification information; Send a first model admission request, the first model admission request including first model admission test information of the first model trained according to the first admission requirements; Receive first model admission indication information, which is related to the test results of the trained first model. The test results are obtained by testing the first model based on the first model admission test information.

2. The method according to claim 1, characterized in that, The method further includes: Send the first model information of the trained first model, which includes the model structure information and the model parameter information of the first model.

3. A communication method, characterized in that, include: Receive a first request, the first request including first identification information of a first model to be deployed; Send a first admission request, which is related to the first identification information; Receive a first model admission request, the first model admission request including first model admission test information of the first model trained according to the first admission requirements; Send first model admission indication information, which is related to the test results of the trained first model. The test results are obtained by testing the first model based on the first model admission test information.

4. The method according to claim 3, characterized in that, The method further includes: Receive the first model information of the trained first model, the first model information including the model structure information and the model parameter information of the first model.

5. A communication method, characterized in that, Applied to a network-side device, the method includes: Send a first request to the third node, the first request including the first identification information of the first model to be deployed; Receive a first admission request from the third node, the first admission request being related to the first identification information; Send first model training information to the terminal device, wherein the first model training information includes the first admission requirement; Send a first model admission request to the third node. The first model admission request includes the first model admission test information of the first model trained according to the first admission requirements. Receive first model admission indication information, which is related to the test results of the trained first model. The test results are obtained by testing the first model based on the first model admission test information.

6. The method according to claim 5, characterized in that, Before sending the first model admission request to the third node, the method further includes: A second model admission request is received from a terminal device. The second model admission request includes second model admission test information of the first model trained according to the first model training information. The first model admission test information is related to the second model admission test information.

7. The method according to claim 5 or 6, characterized in that, The method further includes: The first model admission instruction information is sent to the terminal device.

8. A communication method, characterized in that, Applied to a terminal-side device, the method includes: Receive first model training information, the first model training information including first admission requirements; the first model training information is used to train the first model.

9. The method according to claim 8, characterized in that, The method further includes: Send a second model admission request, which includes the second model admission test information of the first model trained according to the first model training information.

10. The method according to claim 8 or 9, characterized in that, The method further includes: Receive the first model admission instruction information.

11. The method according to any one of claims 1-10, characterized in that, The first admission requirement includes: The input / output interface information of the first model.

12. The method according to any one of claims 1-7, 10-11, characterized in that, If the first model admission indication information indicates that the trained first model is used in the current inference service, then the model update and upgrade flag is activated.

13. The method according to any one of claims 1-12, characterized in that, The method further includes: Based on the first admission requirement, the format of the trained first model is adjusted.

14. An electronic device, characterized in that, The device includes a processor coupled to a memory storing instructions that, when executed by the processor, cause the electronic device to perform the method as described in any one of claims 1-13.

15. An electronic device, characterized in that, The device includes logic circuitry and an input / output interface, the logic circuitry being coupled to the input / output interface for transmitting data through the input / output interface, and the electronic device being used to perform the method as described in any one of claims 1-13.

16. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a computer program or instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1-13.

17. A computer program product, characterized in that, The computer program product includes: a computer program or instructions that, when run on a computer, implement the method as described in any one of claims 1-13.

18. A chip system, characterized in that, It includes a processor and a memory, the memory being used to store a computer program, and the processor being used to invoke and run the computer program stored in the memory to perform the method as described in any one of claims 1-13.

Images