Communication method and communication apparatus

Abstract

The present application provides a communication method and a communication apparatus. The method comprises: receiving first information, the first information indicating one or more of the following: performance information of a first model, generalization information of the first model, or related information of a first data set, and the first data set being used for training the first model; and determining a monitoring means of the first model according to the first information. The solution of the embodiments of the present application helps improve the reliability and efficiency of model monitoring.

Description

Method and apparatus for communication

[0001] The present application claims priority to the Chinese patent application No. 202410627977.X, filed on May 20, 2024, entitled "Method and apparatus for communication", the content of which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[0002] The present application relates to the field of communication, and more particularly, to a method and apparatus for communication. BACKGROUND

[0003] In a wireless communication network, such as a mobile communication network, the services supported by the network are increasingly diverse, and thus the requirements to be met are increasingly diverse. In order to support diversified services, artificial intelligence (AI) technology can be introduced into the wireless communication network, thereby realizing intelligentization of the network. For example, an AI model is used for compression feedback of channel information. A user equipment (UE) obtains downlink channel information according to a reference signal, takes the downlink channel information as an input of a UE-side encoder model, and obtains a feedback amount of the channel information. The UE feeds back the feedback amount of the channel information to a base station, and the base station inputs the feedback amount into a base station-side decoder model to recover the downlink channel information.

[0004] However, the performance of the AI model in actual use can not be stable. For example, as the environmental conditions change, the AI model can no longer adapt to the current communication environment, making it difficult to achieve the expected communication function, or making it difficult to guarantee the stability of the communication performance. SUMMARY

[0005] The present application provides a method and apparatus for communication, to facilitate improving the reliability and efficiency of model monitoring.

[0006] In a first aspect, a method for communication is provided, which can be executed by a communication apparatus or a module (e.g., a chip or a circuit, etc.) applied to the communication apparatus. The communication apparatus can be a first device in the method embodiment.

[0007] The first device can be a device at a terminal device side or a device at a network device side. The terminal device side can include at least one of a terminal device or an AI entity at the terminal device side. The AI entity at the terminal device side can be the terminal device itself or an AI entity serving the terminal device, for example, a server such as an over the top (OTT) server or a cloud server. The network device side can include at least one of a network device or an AI entity at the network device side. The AI entity at the network device side can be the network device itself or an AI entity serving the network device, for example, a radio access network (RAN) intelligent controller (RIC), an operation administration and maintenance (OAM), or a server such as an OTT server or a cloud server.

[0008] The method includes: receiving first information, the first information indicating one or more of the following: performance information of the first model, generalization information of the first model, or related information of a first data set used for training of the first model; and determining a monitoring manner of the first model according to the first information.

[0009] In the scheme of the embodiments of the present application, the performance information of the model, the generalization information of the model, or the related information of the data set can reflect the expected performance of the model and / or the expected generalization capability of the first model, which is conducive to determining a suitable monitoring manner for the model, i.e., determining a monitoring manner matching the performance and / or generalization of the model, so as to effectively monitor models with different performance and / or generalization, thereby improving the reliability and efficiency of model monitoring.

[0010] For example, the first model can be partially or entirely deployed in the first device. The first device can perform model monitoring on the first model according to the monitoring manner of the first model.

[0011] Optionally, the method can further include: receiving third information, the third information indicating model monitoring on the first model.

[0012] The first device can perform model monitoring on the first model according to the third information.

[0013] In combination with the first aspect, in some implementations of the first aspect, the method further includes: sending second information, the second information indicating the monitoring manner of the first model.

[0014] Exemplarily, the first model is deployed in the second device partially or entirely. The first device can send the second information to the second device, and the second device can perform model monitoring on the first model according to the monitoring manner indicated by the second information.

[0015] With reference to the first aspect, in some implementations of the first aspect, the first information indicates the performance information of the first model, the generalization information of the first model, or one or more of the related information of the first data set by indicating an identifier of the first model.

[0016] With reference to the first aspect, in some implementations of the first aspect, the performance information of the first model includes a performance level of the first model, the performance level of the first model being related to an expected performance of the first model, and / or the generalization information of the first model includes a generalization level of the first model, the generalization level of the first model being related to an expected generalization capability of the first model.

[0017] Optionally, the performance information of the first model can include an expected performance of the first model or an expected performance range of the first model.

[0018] Optionally, the generalization information of the first model can include an expected generalization capability of the first model or an expected generalization capability range of the first model.

[0019] In the scheme of the embodiments of the present application, the performance information of the first model can be used to reflect the expected performance of the first model, and the monitoring manner of the first model is determined according to the performance information of the first model, which is equivalent to determining the monitoring manner of the first model according to the expected performance of the first model, and is conducive to obtaining a monitoring manner that matches the performance of the first model, thereby facilitating to improve the reliability of model monitoring. The generalization information of the first model can be used to reflect the expected generalization capability of the first model, and the monitoring manner of the first model is determined according to the generalization information of the first model, which is equivalent to determining the monitoring manner of the first model according to the expected generalization capability of the first model, and is conducive to obtaining a monitoring manner that matches the generalization of the first model, thereby facilitating to improve the reliability and efficiency of model monitoring.

[0020] With reference to the first aspect, in some implementations of the first aspect, the first information further indicates a model category of the first model.

[0021] With reference to the first aspect, in some implementations of the first aspect, the model category of the first model includes a basic general model or a cell-specific model.

[0022] In the scheme of the embodiments of the present application, the monitoring manner of the first model is determined according to the model category of the first model, and the relationship between the category of the model and the expected performance and / or expected generalization capability of the model is considered, so as to make the division of the expected performance and / or expected generalization capability more accurate and more fine, and facilitate selection of a monitoring manner matching the performance and / or generalization capability for a model with different generalization capability, thereby facilitating improvement of the reliability of model monitoring.

[0023] With reference to the first aspect, in some implementations of the first aspect, the related information of the first data set includes an identifier of the first data set.

[0024] In the scheme of the embodiments of the present application, the data set can reflect the expected performance and / or expected generalization capability of the model trained by the data set, so as to facilitate matching of the performance and / or generalization capability of the model with the monitoring manner, thereby facilitating effective monitoring of the model with different performance and / or generalization capability and improving the reliability and efficiency of the monitoring.

[0025] With reference to the first aspect, in some implementations of the first aspect, the related information of the first data set includes one or more of the following: performance information corresponding to the first data set or generalization capability information corresponding to the first data set.

[0026] In the scheme of the embodiments of the present application, the performance information corresponding to the first data set can be used to reflect the expected performance of the model trained by the first data set, and the monitoring manner of the first model is determined according to the performance information corresponding to the first data set, which is equivalent to determining the monitoring manner of the first model according to the expected performance of the model trained by the first data set, and facilitates obtaining a monitoring manner matching the performance of the first model, thereby facilitating improvement of the reliability of model monitoring. The generalization capability information corresponding to the first data set can be used to reflect the expected generalization capability of the model trained by the first data set, and the monitoring manner of the first model is determined according to the generalization capability information corresponding to the first data set, which is equivalent to determining the monitoring manner of the first model according to the expected generalization capability of the model trained by the first data set, and facilitates obtaining a monitoring manner matching the generalization capability of the first model, thereby facilitating improvement of the reliability and efficiency of model monitoring.

[0027] With reference to the first aspect, in some implementations of the first aspect, the performance information corresponding to the first data set includes a performance gear of the model trained by the first data set, the performance gear of the model trained by the first data set is related to the expected performance of the model trained by the first data set, and / or the generalization capability information corresponding to the first data set includes a generalization gear of the model trained by the first data set, the generalization gear of the model trained by the first data set is related to the expected generalization capability of the model trained by the first data set.

[0028] With reference to the first aspect, in some implementations of the first aspect, the monitoring parameter employed by the monitoring manner of the first model comprises at least one of a performance threshold of the first model, a monitoring frequency of the first model, a monitoring duration of the first model, a monitoring times of the first model, a monitoring error tolerance of the first model, or a switching threshold of the first model.

[0029] The second aspect provides a method of communication, which can be performed by a communication device or a module (e.g., a chip or a circuit, etc.) applied to the communication device. The communication device can be the second device in the method embodiment.

[0030] The second device can be a device on a terminal device side or a device on a network device side. The terminal device side can include at least one of a terminal device or an AI entity on the terminal device side. The AI entity on the terminal device side can be the terminal device itself or an AI entity serving the terminal device, such as a server, e.g., an OTT server or a cloud server. The network device side can include at least one of a network device or an AI entity on the network device side. The AI entity on the network device side can be the network device itself or an AI entity serving the network device, such as a RIC, an OAM, or a server, e.g., an OTT server or a cloud server.

[0031] The method includes: sending first information, the first information indicating one or more of performance information of the first model, generalization information of the first model, or related information of a first data set used for training of the first model, the first information being used for determination of a monitoring manner of the first model; and receiving second information, the second information indicating the monitoring manner of the first model.

[0032] In the scheme of the embodiments of the present application, the performance information of the model, the generalization information of the model, or the related information of the data set can reflect the expected performance of the model and / or the expected generalization capability of the first model, which is conducive to determining a suitable monitoring manner for the model, i.e., determining a monitoring manner matching the performance and / or generalization of the model, so as to effectively monitor models with different performance and / or generalization, thereby improving the reliability and efficiency of model monitoring.

[0033] With reference to the second aspect, in some implementations of the second aspect, the first information indicates one or more of the performance information of the first model, the generalization information of the first model, or the related information of the first data set by indicating an identifier of the first model.

[0034] With reference to the second aspect, in some implementations of the second aspect, the performance information of the first model comprises a performance level of the first model, the performance level of the first model being related to an expected performance of the first model, and / or the generalization information of the first model comprises a generalization level of the first model, the generalization level of the first model being related to an expected generalization capability of the first model.

[0035] With reference to the second aspect, in some implementations of the second aspect, the first information further indicates a model category of the first model.

[0036] With reference to the second aspect, in some implementations of the second aspect, the model category of the first model comprises a base general model or a cell-specific model.

[0037] With reference to the second aspect, in some implementations of the second aspect, the related information of the first dataset comprises an identity of the first dataset.

[0038] With reference to the second aspect, in some implementations of the second aspect, the related information of the first dataset comprises one or more of the following: performance information corresponding to the first dataset or generalization information corresponding to the first dataset.

[0039] With reference to the second aspect, in some implementations of the second aspect, the performance information corresponding to the first dataset comprises a performance level of a model trained by the first dataset, the performance level of the model trained by the first dataset being related to an expected performance of the model trained by the first dataset, and / or the generalization information corresponding to the first dataset comprises a generalization level of the model trained by the first dataset, the generalization level of the model trained by the first dataset being related to an expected generalization capability of the model trained by the first dataset.

[0040] With reference to the second aspect, in some implementations of the second aspect, the monitoring parameter adopted by the monitoring manner of the first model comprises at least one of the following: a performance threshold of the first model, a monitoring frequency of the first model, a monitoring duration of the first model, a monitoring number of times of the first model, a monitoring error tolerance of the first model, or a switching threshold of the first model.

[0041] The third aspect can be performed by a communication apparatus or a module (e.g., a chip or a circuit, etc.) applied to the communication apparatus, which can be the second device in the method embodiments.

[0042] The second device can be a device at a terminal device side or a device at a network device side. The terminal device side can include at least one of a terminal device or an AI entity at the terminal device side. The AI entity at the terminal device side can be the terminal device itself or an AI entity serving the terminal device, for example, a server such as an OTT server or a cloud server. The network device side can include at least one of a network device or an AI entity at the network device side. The AI entity at the network device side can be the network device itself or an AI entity serving the network device, for example, a RIC, an OAM, or a server such as an OTT server or a cloud server.

[0043] The method includes: sending first information, the first information indicating one or more of the following: performance information of the first model, generalization information of the first model, or related information of a first data set used for training of the first model, the first information being used for determination of a monitoring manner of the first model.

[0044] In the scheme of the embodiments of the present application, the performance information of the model, the generalization information of the model, or the related information of the data set can reflect the expected performance of the model and / or the expected generalization capability of the first model, which is conducive to determining a suitable monitoring manner for the model, that is, determining a monitoring manner matching the performance and / or generalization of the model, so as to effectively monitor models with different performance and / or generalization, thereby improving the reliability and efficiency of model monitoring.

[0045] In combination with the third aspect, in some implementations of the third aspect, the method further includes: sending third information, the third information indicating that the first model is subjected to model monitoring.

[0046] In combination with the third aspect, in some implementations of the third aspect, the first information indicates one or more of the performance information of the first model, the generalization information of the first model, or the related information of the first data set by indicating an identifier of the first model.

[0047] In combination with the third aspect, in some implementations of the third aspect, the performance information of the first model includes a performance gear of the first model, the performance gear of the first model being related to the expected performance of the first model, and / or the generalization information of the first model includes a generalization gear of the first model, the generalization gear of the first model being related to the expected generalization capability of the first model.

[0048] In combination with the third aspect, in some implementations of the third aspect, the first information further indicates a model category of the first model.

[0049] In combination with the third aspect, in some implementations of the third aspect, the model category of the first model includes a basic general model or a cell-specific model.

[0050] In some implementations of the third aspect, in conjunction with the third aspect, the related information of the first data set includes an identification of the first data set.

[0051] In some implementations of the third aspect, in conjunction with the third aspect, the related information of the first data set includes one or more of the following: performance information corresponding to the first data set or generalization information corresponding to the first data set.

[0052] In some implementations of the third aspect, in conjunction with the third aspect, the performance information corresponding to the first data set includes a performance level of a model trained by the first data set, the performance level of the model trained by the first data set being related to an expected performance of the model trained by the first data set, and / or the generalization information corresponding to the first data set includes a generalization level of the model trained by the first data set, the generalization level of the model trained by the first data set being related to an expected generalization capability of the model trained by the first data set.

[0053] In some implementations of the third aspect, in conjunction with the third aspect, the monitoring parameter used by the monitoring manner of the first model includes at least one of the following: a performance threshold of the first model, a monitoring frequency of the first model, a monitoring duration of the first model, a monitoring number of times of the first model, a monitoring error tolerance of the first model, or a switching threshold of the first model.

[0054] In a fourth aspect, a communication apparatus is provided, which can be a terminal device, a device, a module, a circuit, a chip, or the like configured to be arranged in a terminal device, or an apparatus capable of being used in matching with a terminal device. In one design, the communication apparatus can include a module corresponding to each of the methods described in the first aspect. The module can be a hardware circuit, a software module, or a combination of the hardware circuit and the software module. In one design, the communication apparatus can include a processing module and a communication module.

[0055] In one design, the sending module is configured to perform the sending actions in the methods described in the first aspect, the processing module is configured to perform the processing actions in the methods described in any one of the first aspect to the third aspect, and the receiving module is configured to perform the receiving actions in the methods described in any one of the first aspect to the third aspect.

[0056] In a fifth aspect, a communication apparatus is provided, which can be a network device, or a device, module, circuit or chip configured to be deployed in a network device, or a device capable of being used in association with a network device. In one design, the communication apparatus can include a module corresponding to each of the methods described above with respect to the second aspect, which can be implemented in hardware, software, or combination thereof. In one design, the communication apparatus can include a processing module and a communication module.

[0057] The receiving module is configured to perform the receiving actions in the methods described above with respect to the second aspect, the processing module is configured to perform the processing actions in the methods described above with respect to any of the first aspect to the third aspect, and the sending module is configured to perform the sending actions in the methods described above with respect to any of the first aspect to the third aspect.

[0058] In a sixth aspect, a communication apparatus is provided, which includes one or more processors coupled with one or more storage media storing instructions that, when executed by the one or more processors, cause the methods described above with respect to the first aspect or any possible implementation of the first aspect to be implemented, the methods described above with respect to the second aspect or any possible implementation of the second aspect to be implemented, or the methods described above with respect to the third aspect or any possible implementation of the third aspect to be implemented.

[0059] In a seventh aspect, a communication apparatus is provided, which includes one or more processors configured to process data and / or information to cause the methods described above with respect to the first aspect or any possible implementation of the first aspect to be implemented, the methods described above with respect to the second aspect or any possible implementation of the second aspect to be implemented, or the methods described above with respect to the third aspect or any possible implementation of the third aspect to be implemented.

[0060] Optionally, the communication apparatus can further include a communication interface configured to receive data and / or information and transmit the received data and / or information to the processor. Optionally, the communication interface is further configured to output the data and / or information processed by the processor.

[0061] In an eighth aspect, a chip is provided, which includes a processor configured to execute programs or instructions to cause the methods described above with respect to the first aspect or any possible implementation of the first aspect to be implemented, the methods described above with respect to the second aspect or any possible implementation of the second aspect to be implemented, or the methods described above with respect to the third aspect or any possible implementation of the third aspect to be implemented.

[0062] Optionally, the chip can further comprise a memory for storing programs or instructions. Optionally, the chip can further comprise the transceiver.

[0063] Optionally, the chip is an application specific integrated circuit (ASIC) or a system on chip (SoC).

[0064] In a ninth aspect, a computer-readable storage medium is provided, which comprises instructions, when the instructions are run by a processor, cause the method in the first aspect or any possible implementation of the first aspect to be implemented, cause the method in the second aspect or any possible implementation of the second aspect to be implemented, or cause the method in the third aspect or any possible implementation of the third aspect to be implemented.

[0065] In a tenth aspect, a computer program product is provided, which comprises computer program codes or instructions, when the computer program codes or instructions are run, cause the method in the first aspect or any possible implementation of the first aspect to be implemented, cause the method in the second aspect or any possible implementation of the second aspect to be implemented, or cause the method in the third aspect or any possible implementation of the third aspect to be implemented.

[0066] In an eleventh aspect, a communication system is provided, which comprises one or more combinations of the following apparatuses: the communication apparatus performing the method in the first aspect or any possible implementation of the first aspect, the communication apparatus performing the method in the second aspect or any possible implementation of the second aspect, and the communication apparatus performing the method in the third aspect or any possible implementation of the third aspect. For example, the communication system can comprise the communication apparatus provided in the fourth aspect, and / or the communication apparatus provided in the fifth aspect. BRIEF DESCRIPTION OF DRAWINGS

[0067] FIG. 1 is a schematic diagram of a communication system suitable for embodiments of the application;

[0068] FIG. 2 is a schematic diagram of another communication system suitable for embodiments of the application;

[0069] FIG. 3 is a schematic diagram of yet another communication system suitable for embodiments of the application;

[0070] FIG. 4 is a schematic diagram of an application framework in a communication system suitable for embodiments of the application;

[0071] FIG. 5 is a schematic flow diagram of a method of communication provided by embodiments of the application;

[0072] FIG. 6 is a schematic flow chart of another method of communication provided by embodiments of the present application;

[0073] FIG. 7 is a schematic flow chart of yet another method of communication provided by embodiments of the present application;

[0074] FIG. 8 is a schematic flow chart of yet another method of communication provided by embodiments of the present application;

[0075] FIG. 9 is a schematic flow chart of yet another method of communication provided by embodiments of the present application;

[0076] FIG. 10 is a schematic flow chart of yet another method of communication provided by embodiments of the present application;

[0077] FIG. 11 is a schematic block diagram of an apparatus of communication provided by embodiments of the present application;

[0078] FIG. 12 is a schematic block diagram of another apparatus of communication provided by embodiments of the present application. DETAILED DESCRIPTION

[0079] The technical solutions in the present application will be described below with reference to the accompanying drawings.

[0080] The technical solutions provided by the present application can be applied to various communication systems, for example: a 5th generation (5G) or new radio (NR) system, a long term evolution (LTE) system, an LTE frequency division duplex (FDD) system, an LTE time division duplex (TDD) system, a wireless local area network (WLAN) system, a satellite communication system, a future communication system such as a future mobile communication system, or a fusion system of multiple systems, etc. The technical solutions provided by the present application can also be applied to device to device (D2D) communication, vehicle-to-everything (V2X) communication, machine to machine (M2M) communication, machine type communication (MTC), and internet of things (IoT) communication system or other communication systems.

[0081] A device in a communication system can send or receive signals to or from another device. The signals can include information, signaling, or data, etc. The device can also be replaced by an entity, network entity, network element, communication device, communication module, node, communication node, etc. The device is taken as an example for description in the disclosure. For example, the communication system can include at least one terminal device and at least one network device. In the communication system, the network device can send a downlink signal to the terminal device, the terminal device can send an uplink signal to the network device, the network device can send a signal to another network device, and the terminal device can send a sidelink signal to another terminal device.

[0082] In the embodiments of the present application, the terminal device can also be referred to as a user equipment (UE), an access terminal, a user unit, a user station, a mobile station, a mobile station, a remote station, a remote terminal, a mobile device, a user terminal, a terminal, a wireless communication device, a user agent, or a user apparatus.

[0083] The terminal device can be a device providing voice / data, such as a handheld device with wireless connection function, a vehicle-mounted device, etc. At present, some examples of the terminal are: a mobile phone, a tablet computer, a notebook computer, a palm computer, a mobile internet device (MID), a wearable device, a virtual reality (VR) device, an augmented reality (AR) device, a wireless terminal in industrial control, a wireless terminal in self driving, a wireless terminal in remote medical treatment, a wireless terminal in smart grid, a wireless terminal in transportation safety, a wireless terminal in smart city, a wireless terminal in smart home, a cellular phone, a cordless phone, a session initiation protocol (SIP) phone, a wireless local loop (WLL) station, a personal digital assistant (PDA), a handheld device with wireless communication function, a computing device or other processing device connected to a wireless modem, a wearable device, a terminal device in a 5G network, or a terminal device in a future evolved public land mobile network (PLMN), etc. The embodiments of the present application are not limited thereto.

[0084] By way of example and not limitation, in embodiments of the present application, the terminal device can also be a wearable device. The wearable device can also be referred to as a wearable smart device, which is a general term for devices that are designed and developed by applying wearable technology to daily wear, such as glasses, gloves, watches, clothing, and shoes. The wearable device is a portable device that is directly worn on the body or integrated into the clothes or accessories of the user. The wearable device is not only a hardware device, but also has strong functions through software support and data interaction and cloud interaction. The general wearable smart device includes a full function, a large size, and can realize complete or partial functions without relying on a smart phone, such as a smart watch or smart glasses, and focuses on a certain type of application function and needs to be used in cooperation with other devices, such as a smart phone, such as various smart wristbands and smart jewelry for monitoring vital signs.

[0085] In embodiments of the present application, the device for implementing the function of the terminal device can be a terminal device, or a device capable of supporting the terminal device to implement the function, such as a chip system, which can be installed in the terminal device or used in matching with the terminal device. In embodiments of the present application, the chip system can be composed of a chip, or can include a chip and other discrete devices. In embodiments of the present application, only the device for implementing the function of the terminal device is taken as an example for description, and the present application is not limited to the scheme.

[0086] The network device in the embodiments of the present application can include a device for communicating with a terminal device, for example, the network device can include an access network device or a radio access network device, for example, the access network device can be a base station. The radio access network device in the embodiments of the present application can refer to a RAN node (or device) for accessing a terminal device to a wireless network. The base station can broadly cover various names in the following or be replaced by the following names, such as: Node B (NodeB), evolved Node B (eNB), future base station (next generation NodeB, gNB), relay station, access point, transmitting and receiving point (TRP), transmitting point (TP), primary station, secondary station, multi-standard radio (MSR) node, home base station, network controller, access node, wireless node, access point (AP), transmission node, transceiver node, baseband unit (BBU), remote radio unit (RRU), active antenna unit (AAU), remote radio head (RRH), central unit (CU), distributed unit (DU), radio unit (RU), positioning node, etc. The base station can be a macro base station, a micro base station, a relay node, a donor node or the like, or a combination thereof. The base station can also refer to a communication module, modem or chip used in the aforementioned device or apparatus. The base station can also be a mobile switching center and a device assuming a base station function in D2D, V2X, M2M communication, a network side device in future communication network, a device assuming a base station function in future communication system, etc. The base station can support networks of the same or different access technologies. Optionally, the RAN node can also be a server, a wearable device, a vehicle or a vehicle-mounted device, etc. For example, the access network device in V2X technology can be a road side unit (RSU). The embodiments of the present application do not limit the specific technology and specific device form of the network device.

[0087] The base station can be fixed or mobile. For example, a helicopter or a drone can be configured to act as a mobile base station, and one or more cells can move according to the location of the mobile base station. In other examples, the helicopter or the drone can be configured to serve as a device communicating with another base station.

[0088] In some deployments, the network device mentioned by embodiments of the present application can be a device including a CU, or a DU, or a device including a CU and a DU, or a control plane CU node (central unit-control plane (CU-CP)) and a user plane CU node (central unit-user plane (CU-UP)) and a DU node. For example, the network device can include a gNB-CU-CP, a gNB-CU-UP and a gNB-DU.

[0089] In some deployments, wireless access by a terminal is assisted by cooperation of multiple RAN nodes, and different RAN nodes respectively implement part of the functions of a base station. For example, the RAN node can be a CU, a DU, a CU-CP, a CU-UP, or an RU, etc. The CU and the DU can be separately arranged, or can also be included in the same network element, such as a BBU. The RU can be included in a radio frequency device or a radio frequency unit, such as an RRU, an AAU or an RRH.

[0090] The RAN node can support one or more types of front-haul interfaces, and different front-haul interfaces respectively correspond to DUs and RUs with different functions.

[0091] If the front-haul interface between the DU and the RU is a common public radio interface (CPRI), the DU is configured to implement one or more of the baseband functions, and the RU is configured to implement one or more of the radio frequency functions.

[0092] If the front-haul interface between the DU and the RU is another interface, compared with the CPRI, part of the baseband functions of the downlink and / or uplink, such as one or more of precoding, digital beamforming (BF), or inverse fast fourier transform (IFFT) / adding a cyclic prefix (CP) for the downlink, or one or more of digital beamforming (BF), or fast fourier transform (FFT) / removing a cyclic prefix (CP) for the uplink, are moved from the DU to the RU for implementation.

[0093] One possible implementation, the interface can be an enhanced common public radio interface (eCPRI). Under the eCPRI architecture, the split between the DU and the RU is different, corresponding to different categories (Cat) of eCPRI, such as eCPRI Cat A, B, C, D, E, F.

[0094] Taking eCPRI Cat A as an example, for downlink transmission, the split is layer mapping, the DU is configured to implement layer mapping and one or more functions before layer mapping (i.e., one or more of encoding, rate matching, scrambling, modulation, layer mapping), and other functions after layer mapping (e.g., one or more of resource element (RE) mapping, digital beamforming (BF), or inverse fast Fourier transform (IFFT) / adding cyclic prefix (CP)) are implemented in the RU. For uplink transmission, the split is RE demapping, the DU is configured to implement demapping and one or more functions before demapping (i.e., one or more of decoding, de-rate matching, de-scrambling, de-modulation, inverse discrete Fourier transform (IDFT), channel equalization, RE demapping), and other functions after demapping (e.g., one or more of digital BF or fast Fourier transform (FFT) / CP removal) are implemented in the RU. It can be understood that the description of the functions of the DU and the RU corresponding to various types of eCPRI can refer to the eCPRI protocol, which is not described here.

[0095] In one possible design, the processing unit in the BBU for implementing baseband functions is referred to as a base band high (BBH) unit, and the processing unit in the RRU / AAU / RRH for implementing baseband functions is referred to as a base band low (BBL) unit.

[0096] The CU (or CU-CP and CU-UP), DU or RU can also have different names in different systems, but those skilled in the art can understand their meanings. For example, in an open RAN (ORAN) system, the CU can also be referred to as an O-CU (open CU), the DU can also be referred to as an O-DU, the CU-CP can also be referred to as an O-CU-CP, the CU-UP can also be referred to as an O-CU-UP, and the RU can also be referred to as an O-RU. Any of the CUs (or CU-CPs, CU-UPs), DUs and RUs in this application can be implemented by a software module, a hardware module, or a combination of a software module and a hardware module.

[0097] In the embodiments of the present application, the apparatus for implementing the function of the network device can be a network device; or can be an apparatus capable of supporting the network device to implement the function, such as a chip system, a hardware circuit, a software module, or a hardware circuit plus a software module. The apparatus can be installed in the network device or used in matching with the network device. In the embodiments of the present application, only the apparatus for implementing the function of the network device is taken as an example for illustration, and the scheme of the embodiments of the present application is not limited.

[0098] The network device and / or the terminal device can be deployed on land, including indoors, outdoors, handheld, and / or vehicle-mounted; can also be deployed on the water surface (such as a ship, etc.); and can also be deployed in the air (such as an airplane, a balloon, and / or a satellite). The scenarios in which the network device and the terminal device are located are not limited in the embodiments of the present application.

[0099] In addition, the terminal device and the network device can be hardware devices, or software functions running on special hardware, software functions running on general hardware, such as virtualized functions instantiated on a platform (for example, a cloud platform), or entities including special or general hardware devices and software functions. The specific forms of the terminal device and the network device are not limited in the present application.

[0100] In a wireless communication network, e.g., in a mobile communication network, the services supported by the network are increasingly diverse, and thus the requirements to be met are increasingly diverse. For example, the network needs to be able to support ultra-high rates, ultra-low latency, and / or ultra-large connections. This feature makes network planning, network configuration, and / or resource scheduling increasingly complex. In addition, as the functions of the network become increasingly powerful, e.g., support increasingly high frequency spectrums, support high-order multiple input multiple output (MIMO) technology, support beamforming, and / or support new technologies such as beam management, etc., network energy saving has become a hot research topic. These new requirements, new scenarios, and new features bring unprecedented challenges to network planning, operation and maintenance, and efficient operation. In order to meet this challenge, artificial intelligence technology can be introduced into the wireless communication network, thereby realizing network intelligentization.

[0101] In order to support artificial intelligence (AI) technology in the wireless network, AI nodes (which can also be referred to as AI entities) can also be introduced into the network.

[0102] Optionally, the AI entity can be deployed in one or more of the following positions in the communication system: an access network device, a terminal device, or a core network device, etc., or the AI entity can also be deployed separately, e.g., in a position other than any of the above devices, such as a host or a cloud server of an OTT system. The AI entity can communicate with other devices in the communication system, which can be one or more of the following: a network device, a terminal device, or a network element of a core network, etc. Based on the object served by the AI entity, the AI entity can include a network device-side AI entity, a terminal device-side AI entity, or a core network-side AI entity.

[0103] It can be understood that the present application does not limit the number of AI entities. For example, when there are multiple AI entities, the multiple AI entities can be divided based on functions, e.g., different AI entities are responsible for different functions.

[0104] It can also be understood that the AI entity can be a device independent of each other, can be integrated into the same device to realize different functions, or can be a network element in a hardware device, or can be a software function running on a dedicated hardware, or a virtualized function instantiated on a platform (e.g., a cloud platform), and the present application does not limit the specific form of the AI entity.

[0105] An AI entity can be an AI network element or an AI module. An AI entity is configured to implement a corresponding AI function. AI modules deployed in different network elements can be the same or different. AI models in an AI entity can implement different functions according to different parameter configurations. An AI model in an AI entity can be configured based on one or more of the following parameters: a structural parameter (e.g., at least one of a number of neural network layers, a width of a neural network, a connection relationship between layers, a weight of a neuron, an activation function of a neuron, or a bias in an activation function), an input parameter (e.g., a type of an input parameter and / or a dimension of an input parameter), or an output parameter (e.g., a type of an output parameter and / or a dimension of an output parameter). The bias in an activation function can also be referred to as a bias of a neural network.

[0106] One AI entity can have one or more models. One model can infer an output including one parameter or multiple parameters. Learning processes, training processes, or inference processes of different models can be deployed in different entities or devices, or can be deployed in the same entity or device.

[0107] FIG. 1 is a schematic diagram of a communication system applicable to a communication method according to an embodiment of the present application. As shown in FIG. 1, the communication system 100 can include at least one network device, such as the network device 110 shown in FIG. 1, and can include at least one terminal device, such as the terminal device 120 and the terminal device 130 shown in FIG. 1. The network device 110 and the terminal devices (e.g., the terminal device 120 and the terminal device 130) can communicate with each other through wireless links. The communication devices in the communication system, such as the network device 110 and the terminal device 120, can communicate with each other through multi-antenna technology.

[0108] FIG. 2 is a schematic diagram of another communication system applicable to a communication method according to an embodiment of the present application. Compared with the communication system 100 shown in FIG. 1, the communication system 200 shown in FIG. 2 further includes an AI network element 140. The AI network element 140 is configured to perform AI-related operations, such as constructing a training data set or training an AI model.

[0109] In a possible implementation, the network device 110 can send data related to training of the AI model to the AI network element 140, and the AI network element 140 can construct a training data set and train the AI model. For example, the data related to training of the AI model can include data reported by the terminal device. The AI network element 140 can send a result of an operation related to the AI model to the network device 110, and forward the result to the terminal device through the network device 110. For example, the result of the operation related to the AI model can include at least one of the following: a trained AI model, an evaluation result or a test result of the model, and the like. For example, part of the trained AI model can be deployed on the network device 110, and the other part can be deployed on the terminal device. Alternatively, the trained AI model can be deployed on the network device 110. Alternatively, the trained AI model can be deployed on the terminal device.

[0110] It should be understood that FIG. 2 is only used as an example to illustrate that the AI network element 140 is directly connected to the network device 110, and in other scenarios, the AI network element 140 can also be connected to the terminal device. Alternatively, the AI network element 140 can be connected to both the network device 110 and the terminal device. Alternatively, the AI network element 140 can also be connected to the network device 110 through a third-party network element. The embodiments of the present application do not limit the connection relationship between the AI network element and other network elements.

[0111] The AI network element 140 can also be arranged as a module in the network device and / or the terminal device, for example, in the network device 110 or the terminal device shown in FIG. 1. One or more AI modules can be deployed in the network device 110. One or more AI modules can be deployed in the terminal device.

[0112] It should be noted that FIGS. 1 and 2 are only simplified schematic diagrams for illustration, for example, the communication system can further include other devices, such as a wireless relay device and / or a wireless backhaul device, which are not shown in FIGS. 1 and 2. In actual applications, the communication system can include multiple network devices and / or multiple terminal devices. The embodiments of the present application do not limit the number of network devices and terminal devices included in the communication system.

[0113] FIG. 3 is a schematic diagram of a possible application framework of a communication system according to an embodiment of the present application. As shown in FIG. 3, the network elements in the communication system are connected through interfaces (e.g., NG, Xn) or air interfaces. One or more AI modules (only one is shown in FIG. 3 for clarity) are deployed in one or more of the network element nodes, such as a core network device, an access network node (RAN node), a terminal device, or an operation administration and maintenance (OAM) device. The access network node can be a single RAN node or can include multiple RAN nodes, such as a CU and a DU. The CU and / or the DU can also be provided with one or more AI modules. Optionally, the CU can be further split into a CU-CP and a CU-UP. One or more AI modules are deployed in the CU-CP and / or the CU-UP. For example, the CU and the DU are connected through an F1 interface. The CUs are connected through an Xn interface.

[0114] The network device can be a network device provided with one or more AI modules. The network device can be one or more of the core network device, the access network node (RAN node), or the OAM device shown in FIG. 3. For example, the AI module can be a RIC, such as a near-real-time RIC or a non-real-time RIC, as shown in FIG. 4. For example, the near-real-time RIC is deployed in the RAN node (e.g., the CU, the CU-CP, the CU-UP, the DU, and / or the RU), and the non-real-time RIC is deployed in the OAM, the cloud server, the core network device, or another network device. The RIC can obtain a subset of data from multiple terminal devices from the RAN node (e.g., the CU, the CU-CP, the CU-UP, the DU, and / or the RU), reorganize the subset of data into a training data set #2, and train based on the training data set #2. For example, the near-real-time RIC and the non-real-time RIC can be separately deployed as a network element, and the network device can be the near-real-time RIC or the non-real-time RIC.

[0115] FIG. 4 is a schematic diagram of a possible application framework in a communication system. As shown in FIG. 4, the communication system includes a RAN intelligent controller (RIC). For example, the RIC can be an AI module shown in FIG. 3, which is used to implement AI-related functions. The RIC includes a near-real-time RIC (near-RT RIC) and a non-real-time RIC (Non-RT RIC). The non-real-time RIC mainly processes non-real-time information, such as data that is not sensitive to latency, which can be in the order of seconds. The near-RT RIC mainly processes near-real-time information, such as data that is relatively sensitive to latency, which can be in the order of tens of milliseconds.

[0116] Near real-time RIC is used for model training and inference. For example, it is used for training an AI model, and inference with the AI model. Near real-time RIC can obtain network side and / or terminal side information from RAN nodes (e.g., CU, CU-CP, CU-UP, DU, and / or RU) and / or terminals. The information can be used as training data or inference data. Optionally, near real-time RIC can deliver inference results to RAN nodes and / or terminals. Optionally, inference results can be exchanged between CU and DU, and / or between DU and RU. For example, near real-time RIC delivers inference results to DU, which delivers them to RU.

[0117] Non-real-time RIC is also used for model training and inference. For example, it is used for training an AI model, and inference with the AI model. Non-real-time RIC can obtain network side and / or terminal side information from RAN nodes (e.g., CU, CU-CP, CU-UP, DU, and / or RU) and / or terminals. The information can be used as training data or inference data, and inference results can be delivered to RAN nodes and / or terminals. Optionally, inference results can be exchanged between CU and DU, and / or between DU and RU. For example, non-real-time RIC delivers inference results to DU, which delivers them to RU.

[0118] Near real-time RIC and non-real-time RIC can also be separately configured as a network element. Optionally, near real-time RIC and non-real-time RIC can also be part of other devices. For example, near real-time RIC can be configured in RAN nodes (e.g., CU, DU), and non-real-time RIC can be configured in OAM, cloud server, core network device, or other network devices.

[0119] To facilitate understanding of the scheme of the embodiments of the present application, the following explains the terms that can be involved in the embodiments of the present application.

[0120] (1) Neural network (NN):

[0121] Neural network is a specific implementation form of AI or machine learning (ML). According to the universal approximation theorem, neural network can theoretically approximate any continuous function, so that neural network has the ability to learn any mapping.

[0122] Taking the type of the AI model as a neural network as an example, the AI model involved in the present disclosure can be a deep neural network (DNN). A conventional communication system needs to be designed with the help of rich expert knowledge of a communication module, while a deep learning communication system based on a deep neural network (DNN) can automatically discover an implicit pattern structure from a large amount of data sets, establish a mapping relationship between data, and obtain a performance better than a conventional modeling method.

[0123] A neural network can be composed of neurons, each of which performs a weighted summation operation on its input values, and the weighted summation result generates an output through a nonlinear function. A DNN generally has a multi-layer structure, and each layer of the DNN can include multiple neurons. The input layer transmits the received values to the intermediate hidden layer after processing by the neurons. Similarly, the hidden layer transmits the calculation results to the final output layer to generate the final output of the DNN.

[0124] A DNN generally has more than one hidden layer, and the hidden layer often directly affects the ability to extract information and fit a function. Increasing the number of hidden layers of the DNN or expanding the width of each layer can improve the function fitting ability of the DNN. The weighted value in each neuron is the parameter of the DNN network model. The model parameters are optimized through a training process, so that the DNN network has the ability to extract data features and express mapping relationships. A DNN generally uses a supervised learning or unsupervised learning strategy to optimize the model parameters.

[0125] According to the construction method of the network, the DNN can include a feedforward neural network (FNN), a convolutional neural network (CNN), and a recurrent neural network (RNN), etc.

[0126] A CNN is a neural network specially designed to process data with a similar grid structure. For example, time series data (time axis discrete sampling) and image data (two-dimensional discrete sampling) can be considered as data with a similar grid structure. A CNN does not use all input information for operation at one time, but uses a fixed-size window to extract part of the information for convolution operation, which greatly reduces the calculation amount of model parameters. In addition, according to the different types of information extracted by the window (such as people and objects in the same image), each window can use different convolution kernel operations, which enables the CNN to better extract the features of the input data.

[0127] RNN is a kind of DNN network using feedback time series information. The input of RNN includes the new input value at the current time and the output value of itself at the previous time. RNN is suitable for obtaining sequence features with correlation in time, and is particularly suitable for applications such as speech recognition and channel coding and decoding.

[0128] The characteristic of FNN network is that the neurons of adjacent layers are completely connected with each other, which makes FNN usually need a large amount of storage space and leads to high computational complexity.

[0129] The above FNN, CNN and RNN are all constructed based on neurons. As described above, each neuron performs weighted summation operation on its input value, and the weighted summation result generates output through a nonlinear function. The weights of the weighted summation operation of the neurons in the neural network and the nonlinear function are called the parameters of the neural network. The parameters of all neurons of a neural network constitute the parameters of the neural network.

[0130] (2) Two-side model:

[0131] The two-side model can also be called a double-sided model, a collaborative model or a dual model. The two-side model refers to a model composed of multiple sub-models. The multiple sub-models constituting the model need to be matched with each other. The multiple sub-models can be deployed in different nodes.

[0132] Taking the encoder for compressing channel information and the decoder for restoring channel information in the CSI feedback process as an example, the encoder and the decoder are matched for use, and it can be understood that the encoder and the decoder are matched AI models. An encoder can include one or more AI models, and the decoder matched with the encoder also includes one or more AI models. The number of AI models included in the matched encoder and decoder is the same and one-to-one correspondence. The encoder can also include a quantization module, which can be used for quantization processing of the output of the AI model in the encoder. The decoder can include a dequantization module, which can be used for dequantization processing of the received feedback information of the channel information, to obtain the input of the AI model in the decoder. The dequantization processing can also be called the dequantization processing.

[0133] In a possible design, an encoder and a decoder used in a matching manner can be two parts of the same auto-encoder (AE). The AE model in which the encoder and the decoder are respectively deployed on different nodes is a typical bilateral model. The encoder and the decoder of the AE model are usually jointly trained and used in a matching manner. The auto-encoder is a neural network of unsupervised learning, and is characterized in that input data is used as label data, and therefore the auto-encoder can also be understood as a neural network of self-supervised learning. The auto-encoder can be used for data compression and recovery. For example, the encoder in the auto-encoder can compress (encode) data A to obtain data B; the decoder in the auto-encoder can decompress (decode) data B to recover data A. Alternatively, it can be understood that the decoder is the inverse operation of the encoder.

[0134] The AI model in the embodiment of the application can include an encoder deployed on a terminal device side and a decoder deployed on a network device side, or an encoder deployed on a terminal device side and a decoder deployed on another terminal device side, or an encoder deployed on a network device side and a decoder deployed on another network device side.

[0135] With the development of AI technology, AI models are increasingly applied in communication systems. For example, an AI model is used for compression feedback of channel information. A UE obtains downlink channel information according to a reference signal, takes the downlink channel information as input of a UE-side encoder model, and obtains a feedback amount of the channel information. The UE feeds back the feedback amount of the channel information to a base station, and the base station inputs the feedback amount into a base station-side decoder model to recover the downlink channel information.

[0136] Through training of an AI model, the AI model can be enabled to complete an expected communication function. One method of enabling a model to complete an expected communication function is to determine a training data set of the model, so as to train the model to achieve a function matching the expected function.

[0137] The performance of an AI model in actual use can not be stable. For example, changes in the use environment of the AI model can affect the performance of the AI model. In the use process of the AI model, the AI model can be monitored, and the AI model can be adjusted according to the monitoring result, so as to ensure the performance of the network. However, different models have different performance or generalization, and the monitoring process of the model can not match the performance of the model, resulting in invalid monitoring results.

[0138] Therefore, the method and apparatus for communication are provided in the present application, and the model monitoring is performed by using the appropriate monitoring parameter for different models, so as to improve the reliability and efficiency of the model monitoring. For example, for the models with different performance and / or different generalization, the model monitoring is performed by using the monitoring parameter matching the performance and / or generalization of the models, so as to improve the reliability and efficiency of the model monitoring.

[0139] It should be understood that, in the present application, the indication includes direct indication (also referred to as explicit indication) and implicit indication. Among them, the direct indication of information A means to include the information A; the implicit indication of information A means to indicate the information A by the correspondence between the information A and information B and the direct indication of information B. The correspondence between the information A and information B can be predefined, pre-stored, pre-burned, or pre-configured.

[0140] It should be understood that, in the present application, the determination of information D based on information C includes that the information D is determined based on the information C only, and that the information D is determined based on the information C and other information. In addition, the determination of information D based on information C can also include the case of indirect determination, such as the case that the information D is determined based on information E, and the information E is determined based on information C.

[0141] In addition, in the embodiments of the present application, "the network element A sends information A to the network element B" can be understood as that the destination of the information A or the intermediate network element in the transmission path between the destination is the network element B, which can include direct or indirect sending of the information to the network element B. "The network element B receives the information A from the network element A" can be understood as that the source of the information A or the intermediate network element in the transmission path between the source is the network element A, which can include direct or indirect receiving of the information from the network element A. The information can be processed as necessary between the source and the destination of the information sending, for example, format change, etc., but the destination can understand the valid information from the source. Similar expressions in the present application can be understood similarly, which will not be described here.

[0142] FIG. 5 is a schematic flowchart of a method of communication provided in the present application.

[0143] As shown in FIG. 5, the method 500 can include the following steps.

[0144] 510, the first device acquires one or more of the performance information of the first model, the generalization information of the first model, or the related information of the first data set.

[0145] 520, the first device determines the first monitoring mode according to one or more of the performance information of the first model, the generalization information of the first model, or the related information of the first data set.

[0146] The first data set is used for training of the first model.

[0147] The first monitoring manner can be used for model monitoring of the first model. The first monitoring manner is a monitoring manner of the first model.

[0148] The model monitoring of the first model can also be replaced by performance monitoring of the first model.

[0149] The first model can be an AI model or a non-AI model.

[0150] The first model can be a single-sided model or a double-sided model, or a sub-model in a double-sided model.

[0151] For example, the first model can be a double-sided model including a first sub-model and a second sub-model. In this case, the model monitoring of the first model can include model monitoring of the first sub-model and / or the second sub-model. Taking an example in which the first sub-model and the second sub-model are respectively deployed on a first device and a second device, the first device monitoring the first model can include the first device monitoring the first sub-model, or the first device and the second device jointly monitoring the first sub-model and the second sub-model. The second device monitoring the first model can include the second device monitoring the second sub-model, or the first device and the second device jointly monitoring the first sub-model and the second sub-model.

[0152] For ease of description, AI models are mainly taken as examples for description in the embodiments of the present application, which does not constitute a limitation on the schemes of the embodiments of the present application.

[0153] Optionally, the first device can further obtain a category of the first model. In this case, the first device can further determine the first monitoring manner according to the category of the first model.

[0154] The category of the model can also be referred to as the type of the model.

[0155] Optionally, the method 500 can further include step 530.

[0156] 530. The first device sends second information to the second device. The second information is used to indicate the first monitoring manner.

[0157] Optionally, the method 500 can further include step 540.

[0158] 540. The second device sends third information to the first device. The third information is used to indicate the model monitoring of the first model.

[0159] In the embodiments of the present application, the device A sends information to the device B, which can be directly sent by the device A to the device B, or sent by the device A to the device B through forwarding of other devices. The device A receives information from the device B, which can be directly received by the device A from the device B, or received by the device A from the device B through forwarding of other devices.

[0160] In the scheme of the embodiments of the present application, the performance information of the model, the generalization information of the model or the related information of the data set can reflect the expected performance of the model and / or the expected generalization ability of the first model, which is conducive to determining a suitable monitoring manner for the model, i.e., determining a monitoring manner matching the performance and / or generalization of the model, so as to effectively monitor models with different performance and / or generalization, thereby improving the reliability and efficiency of model monitoring.

[0161] In a possible implementation, the second device can be an AI entity. Part or all of the first model can be deployed in the second device. For example, the first model can be a single-sided model, which can be deployed in the second device. For another example, the first model can be a double-sided model, including a first sub-model and a second sub-model, the first sub-model and the second sub-model are respectively deployed in the second device and the first device, the first sub-model can be an encoder or a decoder, and the second sub-model can be a decoder or an encoder.

[0162] In this case, the method 500 can include step 530. After determining the first monitoring manner, the first device can notify the second device, and the second device can perform model monitoring on the first model according to the monitoring manner indicated by the first device.

[0163] As an example, the second device can be an AI entity on the side of a terminal device. The devices on the side of the terminal device include terminal devices, or other devices in communication with the terminal devices, such as devices controlled by or serving the terminal devices. The AI entity can be the terminal device itself, or an AI entity in communication with the terminal device. For example, the second device can be a server, such as an OTT server or a cloud server.

[0164] For example, the first device can be a server, and the second device can be a terminal device.

[0165] Alternatively, the first device can be a device at a network device side. The device at the network device side includes the network device, or other devices in communication with the network device, such as devices controlled by or serving the network device. For example, the first device can be an AI entity at the network device side. The AI entity can be the network device itself, or an AI entity in communication with the network device. For example, the first device can be a RIC, an OAM or a server, such as an OTT server or a cloud server. A near-real-time RIC is disposed in a RAN node, for example, in a CU / DU.

[0166] As another example, the second device can be an AI entity at the network device side.

[0167] Exemplarily, the first device can be a device at a terminal device side. For example, the first device can be an AI entity at the terminal device side.

[0168] In another possible implementation, the first device can be an AI entity, and the first model can be deployed in the first device in its entirety. For example, the first model can be a one-sided model, which can be deployed in the first device.

[0169] In this case, the method 500 can not include the step 530. After determining the monitoring manner of the first model, the first device can perform model monitoring on the first model according to the monitoring manner.

[0170] The first device can autonomously perform model monitoring. Alternatively, the method 500 can include the step 540, and the first device can perform model monitoring according to the third information sent by the second device.

[0171] Exemplarily, the first device can be an AI entity at a terminal device side.

[0172] Alternatively, the first device can be an AI entity at a network device side.

[0173] The determination manner of the first monitoring manner is described below.

[0174] The monitoring manner of the model can be represented by a monitoring parameter in the monitoring manner. Accordingly, the monitoring manner in the embodiments of the present application can also be replaced by the monitoring parameter.

[0175] The monitoring parameter can include at least one of the following types: a performance index, a performance threshold, a monitoring frequency, a monitoring period, a monitoring duration, a monitoring number of times, a monitoring error tolerance or a switching threshold. Exemplarily, the performance threshold is a threshold of the performance index.

[0176] Optionally, the second information can be used to indicate a first monitoring parameter in the first monitoring manner, i.e., a monitoring parameter of the first model. The first monitoring parameter can include at least one of the following: a first performance threshold, a first monitoring frequency, a first monitoring period, a first monitoring duration, a first monitoring number of times, a first monitoring error tolerance, or a first switching threshold.

[0177] The performance threshold is used for performance monitoring of the model, e.g., to determine whether the model fails or the performance of the model meets a standard. For example, when the model performance does not meet the standard, any of the following operations can be triggered: model switching, model fine-tuning, or accumulation of the number of failures, etc.

[0178] The performance of the model can be reflected by one or more performance indicators. For example, the performance of the model can be reflected by one performance indicator, which can be squared generalized cosine similarity (SGCS). For another example, the performance of the model can be reflected by two performance indicators, which can include SGCS and normalized mean square error (NMSE). Correspondingly, the performance threshold of the model can include the performance threshold corresponding to the one or more performance indicators. That is, the performance threshold of one model can be one or multiple. For the purpose of description, only one performance indicator is taken as an example for illustration in the embodiments of the present application, which does not limit the schemes of the embodiments of the present application. The performance indicator can also be replaced by a performance parameter.

[0179] For example, for some performance indicators, the greater the value of the performance indicator, the better the performance of the model. For this type of performance indicator, if the value of the performance indicator of the model is less than or equal to the corresponding performance threshold, the model is determined to fail. Alternatively, if the value of the performance indicator of the model is less than the corresponding performance threshold, the model is determined to fail. Alternatively, for some performance indicators, the smaller the value of the performance indicator, the better the performance of the model. For this type of performance indicator, if the value of the performance indicator of the model is greater than or equal to the corresponding performance threshold, the model is determined to fail. Alternatively, if the value of the performance indicator of the model is greater than the corresponding performance threshold, the model is determined to fail. For the purpose of description, only the case where the greater the value of the performance indicator, the better the performance of the model is taken as an example for illustration in the embodiments of the present application, which does not limit the schemes of the embodiments of the present application.

[0180] The inference is performed by using the model, and the value of the performance indicator is calculated according to the result of the inference, which can be regarded as monitoring the model once. Taking channel compression feedback as an example, the network device issues a reference signal for monitoring, the terminal device determines channel information according to the reference signal, inputs the channel information into the encoder, and calculates the value of the performance indicator according to the output of the encoder and the channel information input into the encoder. The above process can be regarded as monitoring the encoder once.

[0181] The monitoring frequency is used to represent the frequency of monitoring the model. The monitoring period is used to represent the period of monitoring the model. The monitoring period T means that the model is monitored at intervals of T. That is, after each time interval T, the performance of the model is monitored, for example, to determine whether the model fails. T is a positive number.

[0182] The monitoring duration refers to the number of times the model is repeatedly monitored, or the number of times the value of the performance indicator is calculated.

[0183] The monitoring duration refers to the duration of monitoring the model. During this duration, the model may be monitored once or multiple times.

[0184] Monitoring the model N times can obtain N values of the performance indicator. N is a positive integer.

[0185] The monitoring error tolerance refers to the number of values of the N performance indicators that are lower than the performance threshold, that is, the number of times the performance of the model does not meet the standard in N monitoring, or the proportion. For example, if the number of values of the N performance indicators that are lower than the performance threshold is within the range of the monitoring error tolerance, the model can not be processed. For example, the monitoring error tolerance is 3, and N is 10, that is, the number of values of the 10 performance indicators that are lower than the performance threshold is 3. If the number of values of the 10 performance indicators that are lower than the performance threshold is less than or equal to 3, the model is not processed.

[0186] The switching threshold refers to a threshold for determining model switching.

[0187] The threshold can be a performance-related threshold.

[0188] For example, in a case where a statistical value of a plurality of performance indicators does not meet the threshold, model switching is performed. For example, the statistical value can be at least one of an average value, a maximum value, or a minimum value.

[0189] The threshold can also be a monitoring frequency-related threshold.

[0190] Exemplarily, the model switching is performed in a case where the performance of the continuous n models is substandard, where n is a switching threshold. n is a positive integer.

[0191] In a case where multiple items in the performance information of the first model, the generalization information of the first model, the related information of the first data set, or the category of the first model are used to determine multiple monitoring parameters in the first monitoring manner, the multiple items can be respectively used to determine different monitoring parameters in the multiple monitoring parameters, or the multiple items can also be used to jointly determine the multiple monitoring parameters.

[0192] For example, the performance information of the first model can be used to determine the first performance threshold, and the generalization information of the first model can be used to determine the first monitoring frequency, the first monitoring duration number, and the first switching threshold.

[0193] For another example, the related information of the first data set can be used to determine the first performance threshold, and the generalization information of the first model can be used to determine the first monitoring frequency, the first monitoring duration number, and the first switching threshold.

[0194] For another example, the performance information of the first model can be used to determine the first performance threshold, and the related information of the first data set can be used to determine the first monitoring frequency, the first monitoring duration number, and the first switching threshold.

[0195] For another example, the performance information of the first model and the category of the first model can be used to determine the first performance threshold, and the generalization information of the first model and the category of the first model can be used to determine the first monitoring frequency, the first monitoring duration number, and the first switching threshold.

[0196] The specific description can refer to the examples in the following.

[0197] As a possible implementation manner, the monitoring manner of a model is related to the expected performance of the model, that is, the monitoring manner of the model can be determined according to the expected performance of the model.

[0198] The expected performance of a model is the performance that the model is expected to be able to achieve when the model is working normally.

[0199] In the embodiments of the present application, the performance information of the model can be used to reflect the expected performance of the model, in which case, the performance information of the model can also be understood as the expected performance information of the model. The monitoring manner of the model is determined according to the performance information of the model, which is equivalent to that the monitoring manner of the model is determined according to the expected performance of the model.

[0200] The expected performance of a model is related to a dataset used to train the model. In other words, the expected performance of a model is determined by a dataset used to train the model. Taking the expected performance of a model as an example of channel state information (CSI) compression, the model is a CSI compression model, and the dataset is used to train the CSI compression model. The dataset can include channel input, CSI compression information corresponding to the channel input, and reconstructed channel information corresponding to the channel input. The compression performance of the CSI compression model is limited by the channel input in the dataset and the CSI compression information corresponding to the channel input. For example, the more the number of training samples in the dataset, the better the expected performance of the model trained by the dataset.

[0201] Correspondingly, the dataset can also be used to reflect the expected performance of the model trained by the dataset. The monitoring manner of the model is determined according to the related information of the dataset, which is equivalent to determining the monitoring manner of the model according to the expected performance of the model.

[0202] The expected performance of the model can be used to determine the performance threshold of the model.

[0203] For example, the expected performance of the model and the performance threshold of the model can be in a positive correlation relationship. That is, the better the expected performance of a model, the higher the performance threshold of the model, and the worse the expected performance of a model, the lower the performance threshold of the model. For example, the performance threshold of model A is threshold A, the performance threshold of model B is threshold B, the expected performance of model A is higher than that of model B, and threshold A is greater than threshold B. The expected performance of model A is better, and model A needs to meet a higher performance threshold (i.e., threshold A) to be considered to normally operate in the actual use process, and model B only needs to meet a lower performance threshold (i.e., threshold B) to be considered to normally operate in the actual use process.

[0204] In the scheme of the embodiments of the present application, the performance threshold of the model is related to the expected performance of the model, which is beneficial to guarantee that the performance threshold of the model matches the performance of the model, thereby being beneficial to improve the reliability of the model monitoring.

[0205] As a possible implementation manner, the monitoring manner of the model is related to the expected generalization ability of the model, that is, the monitoring manner of the model can be determined according to the expected generalization ability of the model.

[0206] The expected generalization ability of a model is the generalization that the model is expected to achieve when normally working.

[0207] In the embodiments of the present application, the generalization information of the model can be used to reflect the expected generalization capability of the model, and in this case, the generalization information of the model can also be understood as the expected generalization capability information of the model. The monitoring manner of the model is determined according to the generalization information of the model, which is equivalent to determining the monitoring manner of the model according to the expected generalization capability of the model.

[0208] The expected generalization capability of a model is related to the data set used to train the model. In other words, the expected generalization capability of a model is determined by the data set used to train the model. For example, the CSI compression model is trained by using a data set. The generalization of the CSI compression model is limited by the channel input in the data set and the CSI compression information corresponding to the channel input. For example, the more diverse the training samples in the data set are, the better the expected performance of the model trained by the data set can be.

[0209] Correspondingly, the data set can also be used to reflect the expected generalization capability of the model trained by the data set. The monitoring manner of the model is determined according to the related information of the data set, which is equivalent to determining the monitoring manner of the model according to the expected generalization capability of the model.

[0210] The expected generalization capability of the model can be used to determine at least one of the following: monitoring frequency, monitoring period, monitoring duration, monitoring duration, monitoring error tolerance or switching threshold.

[0211] For example, the expected generalization capability of the model and the monitoring frequency of the model can be negatively correlated. That is, the higher the expected generalization capability of a model is, the lower the monitoring frequency of the model can be, and the lower the expected generalization capability of a model is, the higher the monitoring frequency of the model can be.

[0212] In the scheme of the embodiments of the present application, the monitoring frequency of the model is related to the expected generalization capability of the model, which is beneficial to ensure that the monitoring frequency of the model matches the generalization of the model, thereby improving the reliability and efficiency of model monitoring. For example, for a model with high expected generalization capability, its adaptability can be stronger, and it can perform well in different environments. A lower monitoring frequency can be used to monitor such a model, which is beneficial to improve the efficiency of model monitoring. For a model with low expected generalization capability, its adaptability can be weaker, and its performance difference can be larger when the environment changes. A higher monitoring frequency can be used to monitor such a model, which is beneficial to ensure the reliability of model monitoring.

[0213] For example, the expected generalization capability of the model and the monitoring duration of the model can be negatively correlated. That is, the higher the expected generalization capability of a model is, the lower the monitoring duration of the model can be, and the lower the expected generalization capability of a model is, the higher the monitoring duration of the model can be.

[0214] For example, the expected generalization capability of a model and the switching threshold of the model are in a positive correlation. That is, the higher the expected generalization capability of a model is, the higher the switching threshold of the model is, and the lower the expected generalization capability of a model is, the lower the switching threshold of the model is.

[0215] Specific examples of the monitoring parameter can be referred to the description in the scheme #1 and the scheme #2.

[0216] In the scheme of the embodiments of the present application, the monitoring manner of a model is related to the expected performance of the model and / or the expected generalization capability of the model, which is beneficial to provide a suitable monitoring manner for models with different performances and / or different generalization capabilities, so that the monitoring manner of a model matches the performance and / or the generalization capability of the model, thereby being beneficial to improve the reliability and efficiency of model monitoring.

[0217] The performance information of a model can be represented in multiple forms.

[0218] Optionally, the performance information of the first model can include the expected performance of the first model or the expected performance range of the first model.

[0219] Optionally, the performance information of the first model can include the performance gear of the first model, and the performance gear of the first model is related to the expected performance of the first model. Alternatively, the performance gear of the first model is the gear corresponding to the expected performance of the first model.

[0220] The expected performance of a model can be divided into multiple performance gears, and the performance gear of a model can be used to reflect the expected performance of the model. Different performance gears correspond to different expected performances. For example, the performance gear can be represented by a numerical value. The lower the performance gear is, the better the expected performance is. Alternatively, the higher the performance gear is, the better the expected performance is. For ease of description, only this example is described in the embodiments of the present application.

[0221] In the embodiments of the present application, the performance gear can be replaced by a performance level or a performance grade, etc.

[0222] The generalization capability information of a model can be represented in multiple forms.

[0223] Optionally, the generalization capability information of the first model can include the expected generalization capability of the first model or the expected generalization capability range of the first model. For example, the expected generalization capability of a model can be the number of scenarios that the model is expected to be applicable to. For another example, the expected generalization capability range of a model can be a range of the number of scenarios that the model is expected to be applicable to. For another example, the expected generalization capability range of a model can include scenarios that the model is expected to be applicable to.

[0224] Optionally, the generalization information of the first model can comprise a generalization gear of the first model, the generalization gear of the first model being related to an expected generalization capability of the first model. In other words, the generalization gear of the first model is a gear corresponding to the expected generalization capability of the first model.

[0225] The expected generalization capability of a model can be divided into multiple generalization gears, and the generalization gear of the model can be used to reflect the expected generalization capability of the model. Different generalization gears can correspond to different expected generalization capabilities. For example, the generalization gear can be represented by a numerical value. The lower the generalization gear, the better the expected generalization capability. Alternatively, the higher the generalization gear, the better the expected generalization capability. For ease of description, only this example is described in the present application embodiment.

[0226] In the present application embodiment, the generalization gear can also be replaced by a generalization level or a generalization grade.

[0227] The related information of the first data set can be represented in various forms.

[0228] For example, the related information of the first data set can be used to indicate the first data set.

[0229] Optionally, the related information of the first data set can comprise an identification (ID) of the first data set. The identification of the first data set can also be replaced by an index of the first data set.

[0230] Alternatively, the related information of the data set can also comprise other contents, as long as different data sets can be distinguished. For example, the related information of the first data set can comprise any one or more of the following: the number of training samples in the first data set, the data format of the training samples in the first data set, or the identification of the provider of the first data set, etc.

[0231] Optionally, the related information of the first data set can comprise performance information corresponding to the first data set.

[0232] The performance information corresponding to a data set can be understood as the performance information of a model trained by the data set, and can be used to reflect the expected performance of the model trained by the data set. For example, the first data set is used for training the first model, and the performance information corresponding to the first data set can be used to reflect the expected performance of the first model.

[0233] Optionally, the performance information corresponding to the first data set can comprise an expected performance or an expected performance range corresponding to the first data set.

[0234] The expected performance or the expected performance range corresponding to a data set can be understood as the expected performance or the expected performance range of a model trained by the data set.

[0235] Optionally, the performance information corresponding to the first data set can comprise a performance level corresponding to the first data set, the performance level corresponding to the first data set being related to the expected performance of the model trained by the first data set. In other words, the performance level corresponding to the first data set is the level corresponding to the expected performance of the model trained by the first data set.

[0236] The expected performance of the model can be divided into a plurality of performance levels, and the performance level corresponding to the data set can be used to reflect the expected performance of the model trained by the data set. Different performance levels correspond to different expected performances. For example, the performance level can be represented by a numerical value. The lower the performance level, the better the expected performance. Alternatively, the higher the performance level, the better the expected performance. For ease of description, only this example is described in the embodiments of the present application.

[0237] Optionally, the related information of the first data set can comprise generalization information corresponding to the first data set.

[0238] The generalization information corresponding to a data set can be understood as the generalization information of the model trained by the data set, and can be used to reflect the expected generalization ability of the model trained by the data set. For example, the first data set is used for training the first model, and the generalization information corresponding to the first data set can be used to reflect the expected generalization ability of the first model.

[0239] Optionally, the generalization information corresponding to the first data set can comprise an expected generalization ability or an expected generalization ability range corresponding to the first data set.

[0240] The expected generalization ability or the expected generalization ability range corresponding to a data set can be understood as the expected generalization ability or the expected generalization ability range of the model trained by the data set.

[0241] Optionally, the generalization information corresponding to the first data set can comprise a generalization level corresponding to the first data set, the generalization level corresponding to the first data set being related to the expected generalization ability of the model trained by the first data set. In other words, the generalization level corresponding to the first data set is the level corresponding to the expected generalization ability of the model trained by the first data set.

[0242] The expected generalization ability of the model can be divided into a plurality of generalization levels, and the generalization level corresponding to the data set can be used to reflect the expected generalization ability of the model trained by the training data set. Different generalization levels can correspond to different expected generalization abilities. For example, the generalization level can be represented by a numerical value. The lower the generalization level, the better the expected generalization ability. Alternatively, the higher the generalization level, the better the expected generalization ability. For ease of description, only this example is described in the embodiments of the present application.

[0243] Optionally, the categories of the models can include a base common model and a cell-specific model.

[0244] That is, the category of the first model can be a base common model or a cell-specific model.

[0245] The base common model is also referred to as a base model. The base common model can be applicable to multiple scenarios or multiple cells. The model is not associated with a specific scenario, that is, any cell can be configured in any scenario.

[0246] The cell-specific model can also be referred to as a dedicated model. The cell-specific model is only applicable to a specific cell or a specific scenario, or in other words, the cell-specific model is only matched / associated with a specific cell or a specific scenario, that is, the user accesses a specific cell or enters a specific scenario, and the model can be configured. The performance of the cell-specific model in the matched cell or scenario can be better than the performance of the base model applied to the cell or scenario, but the performance of the cell-specific model in the unmatched cell or scenario can be worse than the performance of the base model applied to the cell or scenario, that is, the performance degradation caused by the mismatch of the cell or scenario is more severe.

[0247] The following describes the determination manner of the monitoring parameter in several examples (manner #1 to manner #6).

[0248] Manner #1:

[0249] As an example, the first network element can determine the first monitoring manner according to the performance information of the first model.

[0250] Exemplarily, the first network element can determine the first monitoring parameter in the first monitoring manner according to the performance information of the first model. The first monitoring parameter can include a first monitoring index, for example, the first monitoring parameter can include a first performance threshold.

[0251] That is, there is a corresponding relationship between the performance information of the first model and the first monitoring parameter. In the embodiments of the present application, the corresponding relationship can also be replaced by an association relationship. The first monitoring parameter is a monitoring parameter related to the performance information of the first model.

[0252] In other words, the performance information of the first model and the first monitoring parameter have a first corresponding relationship, and the first network element can determine the first monitoring parameter according to the corresponding relationship between the performance information and the monitoring parameter, and the corresponding relationship includes the first corresponding relationship.

[0253] Table 1 shows an example of the corresponding relationship between the performance information and the monitoring parameter.

[0254] Table 1

[0255] For example, the performance information of the first model is performance information #A, and according to the correspondence relationship shown in Table 1, it can be determined that the first monitoring parameter includes parameter value #A.

[0256] The following is an example of a channel information feedback scenario. The first model is a channel compression feedback model.

[0257] Table 2 shows an example of the correspondence relationship between performance levels, expected performance ranges, and performance thresholds.

[0258] Table 2

[0259] For example, as shown in Table 2, the expected performance of the channel compression feedback model can be divided into three performance levels. Different performance levels correspond to different expected performance ranges. Different expected performance corresponds to different monitoring parameters.

[0260] In Table 2, the performance of the channel compression feedback model is represented by the channel recovery accuracy, such as the squared generalized cosine similarity (SGCS). The monitoring parameter is the performance threshold.

[0261] In Table 2, the performance level is positively correlated with the expected performance. That is, the higher the performance level of a model, the better the expected performance of the model. Accordingly, the performance level and the performance threshold can be positively correlated.

[0262] The following mainly takes performance level 1 as an example to explain Table 2.

[0263] In Table 2, the expected performance range of the channel compression feedback model in performance level 1 when working normally is SGCS 0.6-0.7, and the associated performance threshold is 0.6-Δ1. For the channel compression feedback model in performance level 1, if the SGCS of the model is less than 0.6-Δ1 during model monitoring, it can be determined that the model has failed once. As shown in Table 2, the higher the performance level, the better the associated expected performance, and accordingly, the higher the performance threshold.

[0264] Taking Table 2 as an example, for example, the performance information of the first model can include performance level 1 or SGCS: [0.6-0.7]. According to the correspondence relationship indicated by Table 2, the associated performance threshold (i.e., the first performance threshold) is 0.6-Δ1.

[0265] The association manner is beneficial to matching the performance of the model and the monitoring manner, i.e., beneficial to selecting the monitoring manner matching the performance for the model with different performance, thereby beneficial to improving the reliability of model monitoring.

[0266] Δ1, Δ2, Δ3 represent deviation values of the performance threshold values associated with different expected performances. The deviation values can be the same or different for different expected performances, i.e., Δ1, Δ2, Δ3 can be the same or different.

[0267] The deviation values of the performance threshold values can be determined by the first device, or determined by the second device, or predefined.

[0268] It should be understood that Table 1 and Table 2 are only examples and do not limit the scheme of the embodiments of the present application.

[0269] In the scheme of the embodiments of the present application, the performance information of the first model can be used to reflect the expected performance of the first model, and the first monitoring manner is determined according to the performance information of the first model, which is equivalent to determining the first monitoring manner according to the expected performance of the first model, and is beneficial to obtaining the monitoring manner matching the performance of the first model, thereby beneficial to improving the reliability of model monitoring.

[0270] Manner #2:

[0271] As an example, the first network element can determine the first monitoring manner according to the generalization information of the first model.

[0272] For example, the first monitoring parameter can include at least one of the following: a first monitoring frequency, a first monitoring period, a first monitoring duration, a first monitoring number of times, a first monitoring error tolerance, or a first switching threshold.

[0273] That is, there is a corresponding relationship between the generalization information of the first model and the first monitoring parameter. The first monitoring parameter is a monitoring parameter related to the performance information of the first model.

[0274] Alternatively, the generalization information of the first model and the first monitoring parameter have a second corresponding relationship, and the first network element can determine the first monitoring parameter according to the corresponding relationship between the generalization information and the monitoring parameter, and the corresponding relationship includes the second corresponding relationship.

[0275] Table 3 shows an example of the corresponding relationship between the generalization information and the monitoring parameter.

[0276] Table 3

[0277] For example, the generalization information of the first model is generalization information #A, and according to the correspondence relationship shown in Table 3, it can be determined that the first monitoring parameter includes parameter value #D.

[0278] The following is an example of a channel information feedback scenario. The first model is a channel compression feedback model.

[0279] Table 4 shows an example of the association between the generalization gear and the monitoring frequency, the monitoring duration, and the switching threshold.

[0280] Table 4

[0281] For example, as shown in Table 4, the expected generalization ability of the channel compression feedback model can be divided into three generalization gears. Different generalization gears correspond to different expected generalization abilities. Different expected generalization abilities correspond to different monitoring parameters. In Table 4, the monitoring parameters include the monitoring frequency, the monitoring duration, and the switching threshold.

[0282] In Table 4, the generalization gear and the expected generalization ability are positively correlated. That is, the higher the generalization gear of a model, the better the expected generalization ability of the model.

[0283] As shown in Table 4, the generalization gear and the monitoring frequency can be negatively correlated.

[0284] The model with higher generalization ability has higher performance robustness, and can be monitored with lower monitoring frequency. As shown in Table 4, during the model monitoring process, for the channel compression feedback model in generalization gear 3, performance monitoring can be performed every 60 minutes (min), and for the channel compression feedback model in generalization gear 1, performance monitoring needs to be performed every 1 min.

[0285] As shown in Table 4, the generalization gear and the monitoring duration can be negatively correlated.

[0286] The model with higher generalization ability has higher performance robustness, and can be monitored with fewer times of performance monitoring, i.e., with fewer monitoring durations. As shown in Table 4, during the model monitoring process, for the channel compression feedback model in generalization gear 3, if the model meets the performance standard for two times, it is considered to be in a normal working state, and for the channel compression feedback model in generalization gear 1, if the model meets the performance standard for 10 times, it is considered to be in a normal working state.

[0287] As shown in Table 4, the generalization gear and the switching threshold can be positively correlated.

[0288] The model with higher generalization capability has higher performance robustness and certain tolerance to performance monitoring fluctuations, and a larger switching threshold can be used. As shown in Table 4, during model monitoring, if the channel compression feedback model in the generalization gear 3 continuously fails to meet the performance for 4 times, the model switching is performed, and if the channel compression feedback model in the generalization gear 1 continuously fails to meet the performance for 2 times, the model switching is performed.

[0289] For example, the generalization information of the first model can include the generalization gear 1. According to the corresponding relationship indicated by Table 4, it can be determined that the generalization gear 1 is associated with a monitoring frequency <1 min, a monitoring duration number of 10, and a switching threshold of 2.

[0290] In this way, the association manner is beneficial to match the generalization of the model and the monitoring manner, that is, beneficial to select the monitoring manner matched with the model with different generalization, thereby beneficial to improve the efficiency of model monitoring while ensuring the reliability of model monitoring, and beneficial to avoid resource waste.

[0291] It should be understood that Table 3 and Table 4 are only examples and do not limit the scheme of the embodiments of the present application.

[0292] In the scheme of the embodiments of the present application, the generalization information of the first model can be used to reflect the expected generalization capability of the first model, and the first monitoring manner is determined according to the generalization information of the first model, which is equivalent to determining the first monitoring manner according to the expected generalization capability of the first model, and is beneficial to obtain the monitoring manner matched with the generalization of the first model, thereby beneficial to improve the reliability and efficiency of model monitoring.

[0293] Method #3:

[0294] As an example, the first network element can determine the first monitoring manner according to the identification of the first data set.

[0295] For example, the first monitoring parameter can include at least one of the following: a first performance threshold, a first monitoring frequency, a first monitoring period, a first monitoring duration, a first monitoring duration number, a first monitoring error tolerance, or a first switching threshold.

[0296] That is, there is a corresponding relationship between the identification of the first data set and the first monitoring parameter. The first monitoring parameter is a monitoring parameter related to the identification of the first data set.

[0297] Or, the first data set identifier has a third correspondence relationship with the first monitoring parameter, and the first network element can determine the first monitoring parameter according to the correspondence relationship between the data set identifier and the monitoring parameter, and the correspondence relationship includes the third correspondence relationship.

[0298] Table 5 shows an example of the correspondence relationship between the data set identifier and the monitoring parameter.

[0299] Table 5

[0300] For example, the identifier of the first data set is identifier #A, and according to the correspondence relationship shown in Table 5, it can be determined that the first monitoring parameter includes parameter value #G.

[0301] The parameter values of the monitoring parameters associated with different data set identifiers can be the same or different. That is, parameter value #G, parameter value #H and parameter value #I can be the same or different.

[0302] The following is an example of a channel information feedback scenario. The first data set is used to train the channel compression feedback model.

[0303] Table 6 shows an example of the correspondence relationship between the data set identifier and the performance threshold and the monitoring frequency.

[0304] Table 6

[0305] As shown in Table 6, the data set identifier is associated with the monitoring parameter. According to the data set identifier, the relevant monitoring parameter can be determined.

[0306] Taking data set 1 as an example, the performance threshold associated with data set 1 is 0.6-Δ1, and the monitoring frequency associated with data set 1 is <60min. For the model trained by data set 1, if the SGCS of the model is less than 0.6-Δ1 during model monitoring, it can be judged that the model fails once, and the performance monitoring can be carried out on it with 60min as the period.

[0307] Taking Table 6 as an example, for example, the identifier of the first data set can be data set 1. According to the correspondence relationship indicated by Table 6, it can be determined that the performance threshold associated with data set 1 is 0.6-Δ1, and the monitoring frequency associated with data set 1 is <60min.

[0308] It should be understood that Table 5 and Table 6 are only examples and do not limit the scheme of the embodiments of the present application.

[0309] The data set can reflect the expected performance and / or expected generalization ability of the model trained by the data set. When establishing the correlation between the data set and the monitoring parameter, the expected performance and / or expected generalization ability corresponding to the data set can be considered, which is beneficial to match the performance and / or generalization of the model with the monitoring mode, thereby facilitating effective monitoring of the model for different performance and / or generalization, and improving the reliability and efficiency of the monitoring.

[0310] Way #4:

[0311] As an example, the first network element can determine the first monitoring mode according to the performance information corresponding to the first data set.

[0312] Exemplarily, the first network element can determine the first monitoring parameter in the first monitoring mode according to the performance information corresponding to the first data set. For example, the first monitoring parameter can include a first performance threshold.

[0313] That is, there is a corresponding relationship between the performance information corresponding to the first data set and the first monitoring parameter. The first monitoring parameter is a monitoring parameter related to the performance information corresponding to the first data set.

[0314] In other words, the performance information corresponding to the first data set and the first monitoring parameter have a fourth corresponding relationship, and the first network element can determine the first monitoring parameter according to the corresponding relationship between the performance information and the monitoring parameter, which includes the fourth corresponding relationship.

[0315] Examples of the corresponding relationship between the performance information corresponding to the data set and the monitoring parameter can refer to Table 1 and Table 2 in Way #1, which only need to replace the performance information of the model with the performance information corresponding to the data set.

[0316] For example, the performance information corresponding to the first data set is performance information #A, and according to the corresponding relationship shown in Table 1, it can be determined that the first monitoring parameter includes parameter value #A.

[0317] Taking Table 2 as an example, the performance gear corresponding to the data set is performance gear 1, that is, the expected performance range of the model trained by the data set when working normally is SGCS 0.6-0.7, and the associated performance threshold is 0.6-Δ1. For the model trained by the data set, if the SGCS of the model is less than 0.6-Δ1 during the model monitoring process, the model can be judged to fail once. As shown in Table 2, the higher the performance gear, the better the associated expected performance, and accordingly, the higher the performance threshold.

[0318] Taking Table 2 as an example, for example, the performance information corresponding to the first data set can include performance gear 1 or SGCS: [0.6-0.7]. According to the corresponding relationship indicated by Table 2, it can be determined that the associated performance threshold (i.e. the first performance threshold) is 0.6-Δ1.

[0319] In the scheme of the embodiments of the present application, the performance information corresponding to the first data set can be used to reflect the expected performance of the model trained by the first data set, and the first monitoring manner is determined according to the performance information corresponding to the first data set, which is equivalent to determining the first monitoring manner according to the expected performance of the model trained by the first data set, which is beneficial to obtain a monitoring manner matching the performance of the first model, thereby improving the reliability of model monitoring.

[0320] Way # 5:

[0321] As an example, the first network element can determine the first monitoring manner according to the generalization information corresponding to the first data set.

[0322] For example, the first monitoring parameter can include at least one of the following: a first monitoring frequency, a first monitoring period, a first monitoring duration, a first monitoring number of times, a first monitoring error tolerance, or a first switching threshold.

[0323] That is, there is a corresponding relationship between the generalization information corresponding to the first data set and the first monitoring parameter. The first monitoring parameter is a monitoring parameter related to the generalization information corresponding to the first data set.

[0324] In other words, the generalization information corresponding to the first data set and the first monitoring parameter have a fifth corresponding relationship, and the first network element can determine the first monitoring parameter according to the corresponding relationship between the generalization information and the monitoring parameter, which includes the fifth corresponding relationship.

[0325] Examples of the corresponding relationship between the generalization information corresponding to the data set and the monitoring parameter can refer to Table 3 and Table 4 in Way # 2, and only the generalization information of the model is replaced by the generalization information corresponding to the data set.

[0326] For example, the generalization information corresponding to the first data set is generalization information # A, and according to the corresponding relationship shown in Table 3, it can be determined that the first monitoring parameter includes parameter value # D.

[0327] For example, the generalization information corresponding to the first data set is generalization information # A, and according to the corresponding relationship shown in Table 3, it can be determined that the first monitoring parameter includes parameter value # D.

[0328] For example, the generalization information corresponding to the first data set is generalization information # A, and according to the corresponding relationship shown in Table 3, it can be determined that the first monitoring parameter includes parameter value # D.For example, the generalization information corresponding to the first data set can include a generalization gear 1 according to Table 4. According to the correspondence relationship indicated in Table 4, it can be determined that the generalization gear 1 is associated with a monitoring frequency < 1 min, an associated monitoring duration of 10, and an associated switching threshold of 2.

[0329] In the scheme of the embodiments of the present application, the generalization information corresponding to the first data set can be used to reflect the expected generalization capability of the model trained by the first data set. The first monitoring manner is determined according to the generalization information corresponding to the first data set, which is equivalent to determining the first monitoring manner according to the expected generalization capability of the model trained by the first data set. This is conducive to obtaining a monitoring manner that matches the generalization of the first model, thereby improving the reliability and efficiency of model monitoring.

[0330] Mode #6:

[0331] As an example, the first network element can also determine the first monitoring manner according to the category of the first model.

[0332] That is, any one of the above-mentioned mode #1 to mode #5 can be used in combination with the category of the first model.

[0333] For example, the first network element can determine the first monitoring parameter in the first monitoring manner according to the performance information of the first model and the category of the first model. For example, the first monitoring parameter can include a first performance threshold.

[0334] That is, there is a corresponding relationship between the performance information of the first model, the category of the first model, and the first monitoring parameter. The first monitoring parameter is a monitoring parameter related to the performance information of the first model and the category of the first model.

[0335] In other words, the performance information of the first model, the category of the first model, and the first monitoring parameter have a sixth corresponding relationship. The first network element can determine the first monitoring parameter according to the corresponding relationship between the performance information, the category, and the monitoring parameter, and the corresponding relationship includes the sixth corresponding relationship.

[0336] Table 7 shows an example of the corresponding relationship between the performance information, the category, and the monitoring parameter.

[0337] Table 7

[0338] For example, the performance information of the first model is performance information #A, the category of the first model is category #A, and according to the corresponding relationship shown in Table 1, it can be determined that the first monitoring parameter includes parameter value #A1.

[0339] The following is an example of a channel information feedback scenario. The first model is a channel compression feedback model.

[0340] Table 8 shows one example of the correspondence between the category, the performance level, the expected performance range, and the performance threshold.

[0341] Table 8

[0342] For example, as shown in Table 8, the expected performance of the model is related to the category of the model and the performance level of the model, and accordingly, the performance threshold can be jointly determined by the performance level and the category.

[0343] For example, taking Table 8 as an example, the performance information of the first model can include the performance level 1, and the category of the first model can be the basic general model. According to the correspondence indicated by Table 8, the associated performance threshold (i.e., the first performance threshold) can be determined to be 0.6-Δ1.

[0344] In this way, the association manner is beneficial to further ensure the matching between the performance of the model and the monitoring manner, that is, the relationship between the category of the model and the expected performance of the model is considered, so as to make the division of the expected performance more accurate and more fine, which is beneficial to selecting the monitoring manner matched with the performance for the model with different performance, thereby being beneficial to improving the reliability of the model monitoring.

[0345] δ1 and δ2 represent the deviation values of the performance thresholds associated with different performance levels of the cell-specific model. For different performance levels, the deviation values can be the same or different, that is, δ1 and δ2 can be the same or different.

[0346] The deviation values of the performance thresholds can be determined by the first device, can be determined by the second device, or can be predefined.

[0347] It should be understood that for different categories, the number of divided performance levels can be the same or different. For example, in Table 8, the number of performance levels corresponding to the basic general model and the cell-specific model is the same, and both correspond to two performance levels. For another example, in some scenarios, the expected performance span of the basic general model can be large, which can be divided into three performance levels, that is, the basic general model corresponds to three performance levels, and the expected performance span of the cell-specific model can be small, which can be divided into two performance levels, that is, the cell-specific model corresponds to two performance levels.

[0348] For example, taking the combination of the manner #2 and the category of the first model as an example, the first network element can determine the first monitoring parameter in the first monitoring manner according to the generalization information of the first model and the category of the first model. For example, the first monitoring parameter can include the first performance threshold.

[0349] The generalization information of the first model, the category of the first model, and the first monitoring parameter have a seventh corresponding relationship. The first network element can determine the first monitoring parameter according to the corresponding relationship between the generalization information, the category, and the monitoring parameter. The corresponding relationship includes an eighth corresponding relationship.

[0350] In other words, the generalization information of the first model, the category of the first model, and the first monitoring parameter have a seventh corresponding relationship. The first network element can determine the first monitoring parameter according to the corresponding relationship between the generalization information, the category, and the monitoring parameter. The corresponding relationship includes an eighth corresponding relationship.

[0351] Table 9 shows an example of the corresponding relationship between the generalization information, the category, and the monitoring parameter.

[0352] Table 9

[0353] For example, the performance information of the first model is generalization information #A, the category of the first model is category #A, and according to the corresponding relationship shown in Table 1, it can be determined that the first monitoring parameter includes parameter value #A2.

[0354] The following is an example of a channel information feedback scenario. The first model is a channel compression feedback model.

[0355] Table 10 shows an example of the association relationship between the category, the generalization gear, and the monitoring frequency, the monitoring duration, and the switching threshold.

[0356] Table 10

[0357] For example, as shown in Table 10, the expected generalization capability of the model is related to the category of the model and the generalization gear of the model. Accordingly, the monitoring parameter can be determined by the generalization gear and the category.

[0358] For example, according to Table 10, the generalization information of the first model can include generalization gear 1, the category of the first model is a basic general model, the associated monitoring frequency is <1 min, the associated monitoring duration is 5, and the associated switching threshold is 2.

[0359] Such an association manner is beneficial to further ensure the matching between the generalization of the model and the monitoring manner, that is, the relationship between the category of the model and the expected generalization capability of the model is considered, so as to make the division of the expected generalization capability more accurate and more fine, which is beneficial to selecting a monitoring manner matched with the generalization for a model with different generalization, thereby improving the reliability of model monitoring.

[0360] It should be understood that the number of generalization levels for different categories can be the same or different. For example, in Table 10, the number of generalization levels corresponding to Category 1 is different from the number of generalization levels corresponding to Category 2, the base general model corresponds to 3 generalization levels, and the cell-specific model corresponds to 1 generalization level. In other implementations, the number of generalization levels corresponding to the base general model and the cell-specific model can also be the same.

[0361] The combination of Mode #3 to Mode #5 and Mode #6 can refer to the foregoing, which will not be repeated here.

[0362] In addition, part or all of the above-mentioned Mode #1 to Mode #6 can be used in combination. That is, the first monitoring mode can be determined based on multiple items in the performance information of the first model, the generalization information of the first model, the identification of the first data set, the performance information corresponding to the first data set, the generalization information corresponding to the first data set, or the category of the first model. The multiple items can be used together to determine the same type of monitoring parameter, or can be used respectively to determine different types of monitoring parameters.

[0363] Taking the combination of Mode #1 and Mode #2 as an example, the first network element can determine the first monitoring mode according to the performance information of the first model and the generalization information of the first model. For example, the performance information of the first model and the generalization information of the first model can be used to determine different types of monitoring parameters in the first monitoring mode.

[0364] Taking the combination of Mode #1, Mode #2, and Mode #6 as an example, the first network element can determine the first monitoring mode according to the performance information of the first model and the generalization information of the first model. For example, the performance information of the first model and the category of the first model can be used to determine the first performance threshold. The category of the first model and the generalization information of the first model can be used to determine at least one of the following: the first monitoring frequency, the first monitoring period, the first monitoring duration, the first monitoring number of times, the first monitoring error tolerance, or the first switching threshold.

[0365] Taking the combination of Mode #4 and Mode #5 as an example, the first network element can determine the first monitoring mode according to the performance information corresponding to the first data set and the generalization information corresponding to the first data set. For example, the performance information corresponding to the first data set and the generalization information corresponding to the first data set can be used to determine different types of monitoring parameters in the first monitoring mode.

[0366] The following describes step 510.

[0367] The first device can obtain one or more of the performance information of the first model, the generalization information of the first model, or the related information of the first data set in various ways.

[0368] As a possible implementation, step 510 can include that the first device receives the first information. The first information indicates one or more of the following: the performance information of the first model, the generalization information of the first model, or the related information of the first data set.

[0369] FIG. 6 shows a schematic flowchart of a method of communication according to an embodiment of the present application. The method shown in FIG. 6 can be regarded as a specific implementation of the method 500 shown in FIG. 5.

[0370] As shown in FIG. 6, step 510 can include that the first device receives the first information from the second device.

[0371] Before the second device sends the first information to the first device, the method 500 can further include that the second device obtains the first information or the second device determines the first information. In other words, the second device obtains or determines the content indicated by the first information.

[0372] The first information can indicate the performance information of the first model in various forms.

[0373] Optionally, the first information can include an identification of the first model. The identification of the first model can also be replaced by an index of the first model.

[0374] That is, the performance information of the first model is indirectly indicated by the identification of the first model.

[0375] The first device can determine the performance information of the first model according to the identification of the first model. There is a corresponding relationship between the identification of the first model and the performance information of the first model. The performance information of the first model is the performance information associated with the identification of the first model.

[0376] Table 11 shows an example of the corresponding relationship between the identification of the model and the performance level.

[0377] Table 11

[0378] For example, the model ID included in the first information is model 1. The first device determines that the performance level associated with model 1 (i.e., the performance level of the first model) is performance level 1 according to the corresponding relationship shown in Table 11.

[0379] The corresponding relationship between the identification of the model and the performance information, for example, Table 11, can be determined by the first device itself, or can be received by the first device from other devices, for example, from the providing node of the first model, for example, from the second device, or can also be predefined.

[0380] Optionally, the first information can comprise performance information of the first model, i.e., the first information can directly indicate the performance information of the first model. For example, the first information can comprise a performance level of the first model.

[0381] The second device can determine the performance information of the first model in various manners. Illustratively, the second device can determine the performance information of the first model according to a correspondence between the identity of the model and the performance information, e.g., Table 11. The correspondence can be determined by the second device itself, or can be received by the second device from another device, e.g., from the first device, or from a providing node of the first model, or can be predefined.

[0382] The first information can indicate the generalization information of the first model in various manners.

[0383] Optionally, the first information can comprise an identity of the first model.

[0384] i.e., the generalization information of the first model is indirectly indicated by the identity of the first model.

[0385] The first device can determine the generalization information of the first model according to the identity of the first model. There is a correspondence between the identity of the first model and the generalization information of the first model. The generalization information of the first model is the generalization information associated with the identity of the first model.

[0386] Table 11 also shows one example of the correspondence between the identity of the model and the generalization level.

[0387] For example, the first information comprises a model ID of Model 1. The first device determines that the generalization level associated with Model 1 (i.e., the generalization level of the first model) is generalization level 3 according to the correspondence shown in Table 11.

[0388] The correspondence between the identity of the model and the generalization information, e.g., Table 11, can be determined by the first device itself, or can be received by the first device from another device, e.g., from a providing node of the first model, or from the second device, or can be predefined.

[0389] Optionally, the first information can comprise generalization information of the first model, i.e., the first information can directly indicate the generalization information of the first model. For example, the first information can comprise a generalization level of the first model.

[0390] The second device can determine the generalization information of the first model in various manners. For a detailed description, refer to the determination of the performance information of the first model in the foregoing, and replace the performance information of the first model with the generalization information of the first model.

[0391] The first information can indicate the performance information corresponding to the first data set in various forms.

[0392] Optionally, the first information can include an identifier of the first data set.

[0393] That is, the performance information corresponding to the first data set is indirectly indicated through the identifier of the first data set.

[0394] The first device can determine the performance information corresponding to the first data set according to the identifier of the first data set. There is a corresponding relationship between the identifier of the first data set and the performance information corresponding to the first data set. The performance information corresponding to the first data set is the performance information associated with the identifier of the first data set.

[0395] Table 12 shows an example of the corresponding relationship between the identifier of the data set and the performance gear.

[0396] Table 12

[0397] For example, the data set ID included in the first information is data set 1. The first device determines, according to the corresponding relationship shown in Table 12, that the performance gear associated with data set 1 (that is, the performance information corresponding to the first data set) is performance gear 1.

[0398] The corresponding relationship between the identifier of the data set and the performance information, for example, Table 12, can be determined by the first device itself, or can be received by the first device from another device, for example, received from a node providing the first data set, for another example, received from the second device, or can also be predefined.

[0399] Optionally, the first information can include the performance information corresponding to the first data set, that is, the first information can directly indicate the performance information corresponding to the first data set. For example, the first information can include the performance gear corresponding to the first data set.

[0400] Optionally, the first information can include an identifier of the first model.

[0401] That is, the performance information corresponding to the first data set is indicated through the identifier of the first model.

[0402] The first device can determine the performance information corresponding to the first data set according to the identifier of the first model. There is a corresponding relationship between the identifier of the first model and the performance information corresponding to the first data set. The performance information corresponding to the first data set is the performance information corresponding to the data set associated with the identifier of the first model.

[0403] The second device can determine the performance information corresponding to the first data set in various manners. The specific description can refer to the foregoing description of the performance information of the first model, and the performance information of the first model is replaced by the performance information corresponding to the first data set.

[0404] The first information can indicate the generalization information corresponding to the first data set in various manners.

[0405] Optionally, the first information can include an identifier of the first data set.

[0406] That is, the generalization information corresponding to the first data set is indirectly indicated by the identifier of the first data set.

[0407] The first device can determine the generalization information corresponding to the first data set according to the identifier of the first data set. There is a corresponding relationship between the identifier of the first data set and the generalization information corresponding to the first data set. The generalization information corresponding to the first data set is the generalization information associated with the identifier of the first data set.

[0408] Table 12 also shows an example of the corresponding relationship between the identifier of the data set and the generalization gear.

[0409] For example, the data set ID included in the first information is data set 1. The first device determines, according to the corresponding relationship shown in Table 12, that the generalization gear associated with data set 1 (that is, the generalization gear corresponding to the first data set) is generalization gear 3.

[0410] The corresponding relationship between the identifier of the data set and the generalization information, for example, Table 12, can be determined by the first device itself, can be received by the first device from other devices, for example, received from the providing node of the first data set, for example, received from the second device, or can be predefined.

[0411] Optionally, the first information can include the generalization information corresponding to the first data set. For example, the first information can include the generalization gear corresponding to the first data set.

[0412] Optionally, the first information can include an identifier of the first model.

[0413] That is, the generalization information corresponding to the first data set is indicated by the identifier of the first model.

[0414] The first device can determine the generalization information corresponding to the first data set according to the identifier of the first model. There is a corresponding relationship between the identifier of the first model and the generalization information corresponding to the first data set. The performance information corresponding to the first data set is the generalization information corresponding to the data set associated with the identifier of the first model.

[0415] The second device can determine the generalization information corresponding to the first data set in multiple ways. The specific description can refer to the foregoing, and the performance information corresponding to the first data set is replaced with the generalization information corresponding to the first data set.

[0416] Further, the first information can also be used to indicate the category of the first model.

[0417] The first information can indicate the category of the first model in multiple forms.

[0418] Optionally, the first information can include the identification of the first model.

[0419] That is, the category of the first model is indirectly indicated through the identification of the first model.

[0420] The first device can determine the category of the first model according to the identification of the first model. There is a corresponding relationship between the identification of the first model and the category of the first model. The category of the first model is the category associated with the identification of the first model.

[0421] Table 13 shows an example of the corresponding relationship between the identification of the model and the category of the model.

[0422] Table 13

[0423] For example, the model ID included in the first information is model 1. The first device determines that the category of the model associated with model 1 (i.e., the category of the first model) is the basic general model according to the corresponding relationship shown in Table 3.

[0424] The corresponding relationship between the identification of the model and the category of the model, for example, Table 13, can be determined by the first device itself, can be received by the first device from other devices, for example, received from the providing node of the first model, for example, received from the second device, or can also be predefined.

[0425] Optionally, the first information can include the identification of the first data set.

[0426] That is, the category of the first model is indirectly indicated through the identification of the first data set.

[0427] The first device can determine the category of the model trained by the first data set according to the identification of the first data set, which can be regarded as the category of the first model. There is a corresponding relationship between the identification of the first data set and the category of the model trained by the first data set. The category of the model trained by the first data set is the category associated with the identification of the first data set.

[0428] The correspondence between the identity of the data set and the category of the model can be determined by the first device itself, or can be received by the first device from another device, for example, from a node providing the first data set.

[0429] Optionally, the first information can comprise an identity of the category of the first model. That is, the first information can directly indicate the category of the first model.

[0430] The second device can determine the category of the first model in various manners. Illustratively, the second device can determine the category of the first model according to a correspondence between the identity of the model and the category of the model. The correspondence can be determined by the second device itself, or can be received by the second device from another device, for example, from the first device, or from a node providing the first model. Alternatively, the second device can receive the category of the first model from another device, for example, from a node providing the first model. Alternatively, the second device can determine the category of the model trained by the first data set according to the identity of the first data set, which can be regarded as the category of the first model. The correspondence between the identity of the data set and the category of the model can be determined by the second device itself, or can be received by the second device from another device, for example, from a node providing the first data set. Alternatively, the second device can receive the category associated with the first data set from another device.

[0431] Further, as mentioned above, the first model can be partially or entirely deployed in the second device, and the first information can also come from a third device other than the second device. That is, in step 510, the first device can receive the first information from the third device. The third device can be a node providing the first model and / or the first data set.

[0432] Alternatively, the category of the first model can also be indicated by other information other than the first information.

[0433] As another possible implementation, step 510 can comprise that the first device determines one or more of the following by itself: the performance information of the first model, the generalization information of the first model, or the related information of the first data set.

[0434] For example, the first device can determine the performance information of the first model according to a correspondence between the identity of the model and the performance information.

[0435] The specific determination manners can refer to the descriptions of the above-mentioned items, and thus will not be described here again to avoid repetition.

[0436] It should be understood that the manner of obtaining the performance information of the first model, the generalization information of the first model, or the related information of the first data set can be the same or different. Illustratively, each of the above can be indicated by the first information. Alternatively, part of the above can be indicated by the first information, and the rest can be obtained in other manners. For example, the rest can be indicated by other information. For another example, the rest can be determined by the first device itself. Alternatively, each of the above can be determined by the first device itself.

[0437] The method 500 is described below in two scenarios (scenario #1 and scenario #2).

[0438] Scenario #1:

[0439] In a possible scenario, a model can be obtained by obtaining a data set. That is, a model capable of achieving a related function is obtained by obtaining a data set. Obtaining a model can be replaced by deploying a model.

[0440] Illustratively, taking an AI model as an example, a model can be trained based on a data set, so as to make the model achieve an expected function or complete function matching. The trained model can be deployed on a corresponding node to achieve a corresponding function. Specifically, after obtaining a data set, a device can complete model training based on the data set to achieve an expected function.

[0441] Taking a model on a terminal device side as an example, a terminal side device, for example, a terminal device or a cloud server of a terminal device, etc., can obtain a data set and train a model according to the data set, so as to make the model achieve an expected function.

[0442] In scenario #1, different data sets can be used to distinguish different models. The identification of a data set can serve as the identification of a model. The data set ID can be replaced by an associated ID.

[0443] The method 500 under scenario #1 is described below taking a bilateral model (example #1) and a unilateral model (example #2) as examples.

[0444] Example #1:

[0445] In example #1, the bilateral model can include a first sub-model and a second sub-model, which are respectively deployed on a first device and a second device.

[0446] FIG. 7 shows a schematic flowchart of a method of communication according to an embodiment of the present application. The method 700 shown in FIG. 7 is a specific implementation of the method 500, and the specific description can be referred to the method 500. To avoid repetition, part of the description is appropriately omitted when the method 700 is described.

[0447] In the method 700, the CSI feedback scenario is mainly taken as an example, i.e., the second sub-model is an encoder for compressing channel information, and the first sub-model is a decoder for restoring channel information. For example, the first device can be an AI entity on the network device side, and the second device can be an AI entity on the terminal device side. Alternatively, the first device can be an AI entity on the terminal device side, and the second device can be an AI entity on the network device side. In the method 700, only the first device is taken as the network device and the second device is taken as the terminal device as an example, which does not constitute a limitation to the scheme of the embodiments of the present application.

[0448] As shown in FIG. 7, the method 700 includes the following steps.

[0449] 710, the terminal device obtains a first data set.

[0450] The terminal device can train a first encoder based on the first data set. The first encoder can be regarded as a first model.

[0451] Alternatively, the terminal device can train a first encoder and a first decoder based on the first data set, i.e., the encoder and the decoder of the bilateral model are trained together. The first encoder and the first decoder can be regarded as a first model.

[0452] For example, the terminal device can obtain a plurality of data sets and train a plurality of models based on the plurality of data sets respectively. For example, the plurality of data sets can be obtained from the network device. In this way, in the network device, the plurality of models can be distinguished by the IDs of the plurality of data sets.

[0453] 720, the terminal device sends first information to the network device, the first information indicating related information of the first data set.

[0454] The training data set of the first encoder is the first data set. If the terminal device wants to perform the CSI feedback process through the first encoder, i.e., compress the channel information, the network device can be instructed with the related information of the first data set.

[0455] For example, the first information can include an ID of the first data set.

[0456] For another example, the first information can include performance information corresponding to the first data set and / or generalization information corresponding to the first data set.

[0457] 730, the network device determines a first monitoring mode according to the related information of the first data set.

[0458] For example, the first information includes the ID of the first data set, the network device can determine the first monitoring manner according to the manner #3 after receiving the ID of the first data set. For example, the network device can determine the monitoring parameter associated with the ID of the first data set according to the correspondence between the ID of the data set and the monitoring parameter (such as Table 5 or Table 6).

[0459] Alternatively, the network device can determine the first monitoring manner according to the manner #4 and / or the manner #5 after receiving the ID of the first data set. For example, the network device can determine the performance information and / or the generalization information associated with the ID of the first data set according to the correspondence between the ID of the data set and the performance information and / or the generalization information (such as Table 12), and then determine the relevant monitoring parameter according to the correspondence between the performance information and / or the generalization information and the monitoring parameter (such as any one or more of Table 1 to Table 4).

[0460] For example, the first information includes the performance information corresponding to the first data set and / or the generalization information corresponding to the first data set, and the network device can determine the first monitoring manner according to the manner #4 and / or the manner #5. For example, the first information can include the performance gear 2 and the generalization gear 2 according to Table 2 and Table 4. The network device determines the value of the monitoring parameter associated with the performance gear 2 as follows: the performance threshold is 0.7-Δ2, and determines the value of the monitoring parameter associated with the generalization gear 2 as follows: the monitoring frequency is less than 10 min, the monitoring duration is 4, and the switching threshold is 3 according to Table 4. The first monitoring parameter can include any one or more of the above parameter values.

[0461] 740, the network device sends the second information to the terminal device, and the second information indicates the first monitoring manner.

[0462] 750, the terminal device performs model monitoring according to the first monitoring manner.

[0463] The terminal device can perform model monitoring on the first encoder according to the first monitoring manner.

[0464] For example, the second information can indicate the first performance threshold and the first monitoring frequency. The terminal device can perform monitoring according to the first monitoring threshold. If the performance of the first encoder on the terminal device side is lower than the first performance threshold, model switching can be triggered. The terminal device can perform model monitoring according to the first monitoring frequency.

[0465] Alternatively, the terminal device and the network device can jointly perform model monitoring on the first encoder and the first decoder according to the first monitoring manner.

[0466] In addition, in steps 720 to 750, the terminal device can be replaced by the network device, and the network device can be replaced by the terminal device.

[0467] It should be understood that the above only takes the CSI feedback scenario as an example for illustration, and the method 700 can also be applied to monitoring of a bilateral model in other scenarios.

[0468] Example #2:

[0469] In Example #2, the first model is a unilateral model, and is deployed on the first device.

[0470] FIG. 8 shows a schematic flowchart of a method of communication according to an embodiment of the application. The method 800 shown in FIG. 8 is a specific implementation of the method 500, and the specific description can refer to the method 500. To avoid repetition, part of the description is appropriately omitted when describing the method 800.

[0471] In the method 800, only the first device is taken as a terminal device, and the second device is taken as a network device as an example for illustration, without limiting the scheme of the embodiments of the application.

[0472] As shown in FIG. 8, the method 800 includes the following steps.

[0473] 810, the terminal device acquires a first data set.

[0474] The terminal device can train the first model based on the first data set.

[0475] Exemplarily, the first data set can come from the network device. Alternatively, the first data set can also come from other devices, for example, a cloud server.

[0476] 820, the network device sends first information to the terminal device, the first information indicating related information of the first data set.

[0477] 830, the terminal device determines a first monitoring manner according to the related information of the first data set.

[0478] 840, the terminal device performs model monitoring according to the first monitoring manner.

[0479] Optionally, before step 840, the method 800 can further include step 850.

[0480] 850, the network device sends third information to the terminal device, the third information being used to instruct the terminal device to perform model monitoring.

[0481] In this case, the model monitoring operation of the terminal device can be triggered by the instruction information sent by the network device.

[0482] The method 800 can also not include step 850, and after determining the first monitoring manner, the terminal device can autonomously perform the monitoring operation without the instruction information of the network device to trigger.

[0483] The specific description can refer to Example #1, which will not be repeated here.

[0484] It should be understood that the above only takes the first device deploying a one-sided model as an example. In other scenarios, a one-sided model can also be deployed on the second device. For example, the first device can be a device on the network device side, and the second device can be a device on the terminal device side. For another example, the first device can be a device on the terminal device side, and the second device can be a device on the network device side. Other possible implementations can refer to the description of the foregoing method 500, which will not be repeated here.

[0485] Scenario #2:

[0486] In a possible scenario, the model can be obtained by receiving the model. Obtaining the model can also be replaced by deploying the model.

[0487] Taking a model on the terminal device side as an example, a terminal device, for example, a terminal device, etc., can receive a model sent by a network device or a third-party device, such as receiving model parameters of the model.

[0488] In scenario #2, different models can be distinguished by the identification (ID) of the model.

[0489] The method 500 under scenario #2 will be described below taking a bilateral model (Example #3) and a one-sided model (Example #4) as examples.

[0490] Example #3:

[0491] In Example #3, the bilateral model can include a first sub-model and a second sub-model, which are deployed on the first device and the second device, respectively.

[0492] FIG. 9 shows a schematic flowchart of a method of communication according to an embodiment of the present application. The method 900 shown in FIG. 9 is a specific implementation of the method 500, and the specific description can refer to the method 500. To avoid repetition, part of the description will be appropriately omitted when describing the method 900.

[0493] In the method 900, a CSI feedback scenario is mainly taken as an example for description, that is, the second sub-model is an encoder for compressing channel information, and the first sub-model is a decoder for restoring channel information. The main difference between the method 900 and the method 700 is that the terminal device obtains the first model in different ways, and the content indicated by the first information is different. When describing the method 900, the main difference will be mainly described, and other descriptions can refer to the method 700.

[0494] As shown in FIG. 9, the method 900 includes the following steps.

[0495] 910, the terminal device obtains the first model.

[0496] Exemplarily, the terminal device can receive model parameters of the first model from other devices.

[0497] For example, the terminal device can receive model parameters of a first encoder, which can be regarded as the first model.

[0498] For another example, the terminal device can receive model parameters of a first encoder and a first decoder, which can be regarded as the first model.

[0499] Exemplarily, the first model can be obtained from a network device. In the network device, a plurality of models can be distinguished by IDs of the plurality of models.

[0500] 920, the terminal device sends first information to the network device, the first information indicating performance information of the first model and / or generalization information of the first model.

[0501] For example, the first information can include an ID of the first model.

[0502] For another example, the first information can include performance information of the first model and / or generalization information of the first model.

[0503] 930, the network device determines a first monitoring manner according to the first information.

[0504] Taking an example that the first information includes an ID of the first model, the network device can determine the first monitoring manner according to the manner #1 and / or the manner #2 after receiving the ID of the first model. For example, the network device can determine performance information and / or generalization information associated with the ID of the first model according to a correspondence between the ID of the model and the performance information and / or the generalization information (such as Table 11), and then determine relevant monitoring parameters according to a correspondence between the performance information and / or the generalization information and the monitoring parameters (such as any one or more of Table 1 to Table 4).

[0505] Taking an example that the first information includes performance information of the first model and / or generalization information of the first model, the network device can determine the first monitoring manner according to the manner #1 and / or the manner #2. Taking Table 2 and Table 4 as an example, for example, the first information can include performance gear 2 and generalization gear 2. The network device determines the value of the monitoring parameter associated with the performance gear 2 according to Table 2 as follows: the performance threshold is 0.7-Δ2, and determines the parameter value of the monitoring parameter associated with the generalization gear 2 according to Table 4 as follows: the monitoring frequency is less than 10 min, the monitoring duration number is 4, and the switching threshold is 3. The first monitoring parameter can include any one or more of the above parameter values.

[0506] Further, the first information can also indicate a category of the first model.

[0507] In this case, the first monitoring manner is also related to the category of the first model. Illustratively, the first device can determine the first monitoring manner according to the performance information of the first model, the generalization information of the first model, and the category of the first model. For example, the first device can determine the relevant monitoring parameters according to Tables 7 to 10.

[0508] 940. The network device sends second information to the terminal device, where the second information indicates the first monitoring manner.

[0509] 950. The terminal device performs model monitoring according to the first monitoring manner.

[0510] Example #4:

[0511] In Example #4, the first model is a one-sided model, and is deployed on the first device.

[0512] FIG. 10 shows a schematic flowchart of a method of communication according to an embodiment of the present application. The method 1000 shown in FIG. 10 is a specific implementation of the method 500, and the specific description can refer to the method 500. To avoid repetition, part of the description is appropriately omitted when describing the method 1000.

[0513] In the method 1000, only the first device is taken as the terminal device, and the second device is taken as the network device as an example for illustration, without limiting the scheme of the embodiments of the present application. The main difference between the method 1000 and the method 800 is that the terminal device obtains the first model in different ways, and the content indicated by the first information is different. When describing the method 1000, the main difference is described, and other descriptions can refer to the method 800.

[0514] As shown in FIG. 10, the method 1000 includes the following steps.

[0515] 1010. The terminal device obtains the first model.

[0516] 1020. The network device sends first information to the terminal device, where the first information indicates the performance information of the first model and / or the generalization information of the first model.

[0517] 1030. The terminal device determines the first monitoring manner according to the first information.

[0518] 1040. The terminal device performs model monitoring according to the first monitoring manner.

[0519] Optionally, before step 1040, the method 1000 can further include step 1050.

[0520] 1050. The network device sends third information to the terminal device, where the third information can be used to instruct the terminal device to perform model monitoring.

[0521] It can be understood that, in some of the above embodiments, the information names involved are only examples, and do not limit the protection scope of the embodiments of the present application.

[0522] It can also be understood that the formulas involved in the various embodiments of the present application are only exemplary and do not limit the protection scope of the embodiments of the present application. In the process of calculating the above-mentioned various involved parameters, the above-mentioned formulas can also be used for calculation, or calculation based on the deformation of the above-mentioned formulas, or other ways can be used for calculation to meet the results of formula calculation.

[0523] It can also be understood that some optional features in the embodiments of the present application can not depend on other features in some scenarios, or can be combined with other features in some scenarios, without limitation.

[0524] It can also be understood that the solutions in the embodiments of the present application can be reasonably combined, and the explanations or descriptions of various terms appearing in the embodiments can be mutually referenced or explained in various embodiments, without limitation.

[0525] It can also be understood that the sizes of various serial numbers in the embodiments of the present application do not mean the order of execution, but are only distinguished for convenience of description, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0526] It can also be understood that the methods and operations implemented by the device in the above various method embodiments can also be implemented by the constituent components of the device, such as chips or circuits.

[0527] Corresponding to the methods given in the above various method embodiments, the embodiments of the present application also provide corresponding devices, which include modules for executing the corresponding modules of the above various method embodiments. The module can be software, hardware, or a combination of software and hardware. It can be understood that the technical features described in the above various method embodiments are also applicable to the following device embodiments.

[0528] FIG. 11 is a schematic diagram of a communication device 1900 provided by an embodiment of the present application. The device 1900 includes a transceiver unit 1910 and a processing unit 1920. The transceiver unit 1910 can be used to implement the corresponding communication function. The transceiver unit 1910 can also be referred to as a communication interface or a communication unit, etc. The processing unit 1920 can be used to implement the corresponding processing or control function, such as configuring resources.

[0529] Optionally, the device 1900 further includes a storage unit, which can be used to store instructions and / or data. The processing unit 1920 can read the instructions and / or data in the storage unit, so that the device implements the actions of the device or network element in the above various method embodiments.

[0530] The apparatus 1900 can be the second device, or can be applied to or matched with the second device, and can implement the communication method performed by the second device. Alternatively, the apparatus 1900 can be the first device, or can be applied to or matched with the first device, and can implement the communication method performed by the first device.

[0531] When the apparatus 1900 is applied to the second device, the apparatus 1900 can implement the steps or processes performed by the second device in the above method embodiments. The transceiver unit 1910 can be configured to perform the transceiving related operations of the second device in the above method embodiments, and the processing unit 1920 can be configured to perform the processing related operations of the second device in the above method embodiments.

[0532] When the apparatus 1900 is applied to the first device, the apparatus 1900 can implement the steps or processes performed by the first device in the above method embodiments. The transceiver unit 1910 can be configured to perform the transceiving related operations of the first device in the above method embodiments, and the processing unit 1920 can be configured to perform the processing related operations of the first device in the above method embodiments.

[0533] It should be understood that the specific process of each unit performing the above corresponding steps has been described in detail in the above method embodiments, and thus will not be described here again for the sake of brevity.

[0534] It should also be understood that the apparatus 1900 here is in the form of functional units. The term "unit" here can refer to an ASIC, an electronic circuit, a processor (such as a shared processor, a dedicated processor, or a group processor, etc.) and a memory for executing one or more software or firmware programs, a combination logic circuit, and / or other suitable components supporting the described functions. In an alternative example, those skilled in the art can understand that the apparatus 1900 can be specifically the first device in the above embodiments, and can be configured to perform the processes and / or steps corresponding to the first device in the above method embodiments. Alternatively, the apparatus 1900 can be specifically the second device in the above embodiments, and can be configured to perform the processes and / or steps corresponding to the second device in the above method embodiments. To avoid repetition, these will not be described here again.

[0535] The apparatus 1900 of each of the above-mentioned solutions has the function of implementing the corresponding steps performed by the device (e.g., the first device, or the second device) in the above-mentioned methods. The function can be implemented by hardware, or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above-mentioned functions; for example, the transceiver unit can be replaced by a transceiver (e.g., the transmitting unit in the transceiver unit can be replaced by a transmitter, and the receiving unit in the transceiver unit can be replaced by a receiver), and other units, such as the processing unit, can be replaced by a processor, which respectively performs the transceiving operation and the related processing operation in each method embodiment.

[0536] In addition, the transceiver unit 1910 can also be a transceiver circuit (e.g., which can include a receiving circuit and a transmitting circuit), and the processing unit 1920 can be a processing circuit. The processing circuit can include one or more processors, or a circuit for processing or control functions in one or more processors, etc.

[0537] It should be noted that the apparatus in FIG. 11 can be a network element or device in the foregoing embodiments, or a chip or chip system, such as an SoC. The transceiver unit can be an input / output circuit, a communication interface; and the processing unit can be a processor or microprocessor integrated on the chip or an integrated circuit. This is not limited here.

[0538] FIG. 12 is a schematic diagram of another apparatus 2000 for communication provided by an embodiment of the present application. The apparatus 2000 includes a processor 2010, which is configured to execute computer programs or instructions stored in a memory 2020, or read data / signaling stored in the memory 2020, to perform the methods in the above-mentioned method embodiments. Optionally, the processor 2010 is one or more.

[0539] Optionally, as shown in FIG. 12, the apparatus 2000 further includes the memory 2020, which is configured to store computer programs or instructions and / or data. The memory 2020 can be integrated with the processor 2010, or can be separately arranged. Optionally, the memory 2020 is one or more.

[0540] Optionally, as shown in FIG. 12, the apparatus 2000 further includes a transceiver circuit 2030, which is configured to receive and / or transmit signals. For example, the processor 2010 is configured to control the transceiver circuit 2030 to receive and / or transmit signals. The processor 2010 can also be replaced by a processing circuit.

[0541] The apparatus 2000 can be a network element or a device in the foregoing embodiments, or can be a chip or a chip system. When the apparatus 2000 is a network element or a device in the foregoing embodiments, the transceiver circuit 2030 can be a transceiver. When the apparatus 2000 is a chip or a chip system, the transceiver circuit 2030 can be an interface circuit or an input / output interface.

[0542] As an option, the apparatus 2000 can be applied to the second device, and specifically, the apparatus 2000 can be the second device or can be an apparatus capable of supporting the second device and implementing the functions of the second device in any of the examples described above. The apparatus 2000 is configured to implement the operations performed by the second device in each of the method embodiments described above.

[0543] For example, the processor 2010 is configured to execute the computer programs or instructions stored in the memory 2020 to implement the related operations of the second device in each of the method embodiments described above.

[0544] As another option, the apparatus 2000 can be applied to the first device, and specifically, the apparatus 2000 can be the first device or can be an apparatus capable of supporting the first device and implementing the functions of the first device in any of the examples described above. The apparatus 2000 is configured to implement the operations performed by the first device in each of the method embodiments described above.

[0545] For example, the processor 2010 is configured to execute the computer programs or instructions stored in the memory 2020 to implement the related operations of the first device in each of the method embodiments described above.

[0546] It should be understood that the processor mentioned in the embodiments of the present application can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSPs), ASICs, field programmable gate arrays (FPGAs) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0547] It should also be understood that the memory referred to in the embodiments of the present application can be a volatile memory and / or a non-volatile memory. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically EPROM (EEPROM) or a flash memory. The volatile memory can be a random access memory (RAM). For example, the RAM can be used as an external cache. As an example but not limitation, the RAM includes the following various forms: static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchlink DRAM (SLDRAM) and direct rambus RAM (DR RAM).

[0548] It should be noted that when the processor is a general processor, DSP, ASIC, FPGA or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, the memory (storage module) can be integrated in the processor.

[0549] It should also be noted that the memory described herein is intended to include, but not limited to, these and any other suitable types of memory.

[0550] The embodiments of the present application also provide a computer readable storage medium having stored thereon computer instructions for implementing the method executed by the communication device in each of the above method embodiments.

[0551] For example, the computer program is executed by the computer, so that the computer can implement the method executed by the first device in each of the above method embodiments.

[0552] For another example, the computer program is executed by the computer, so that the computer can implement the method executed by the second device in each of the above method embodiments.

[0553] The embodiments of the present application further provide a computer program product comprising instructions which, when executed by a computer, implement the method performed by the device (e.g., the first device, or the second device) in any of the above method embodiments.

[0554] The embodiments of the present application further provide a system for communication, comprising the first device and the second device as described above. The first device and the second device can implement the method for communication as shown in any of the above examples.

[0555] Optionally, the system further comprises a device in communication with the first device and / or the second device as described above.

[0556] The explanations and advantages of the related contents in any of the above devices can refer to the corresponding method embodiments provided above, and will not be repeated here.

[0557] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0558] In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable apparatus. For example, the computer can be a personal computer, a server, a network device, or the like. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. integrated with one or more available media sets. The available media can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (SSD), etc. For example, the foregoing available media includes but is not limited to: a variety of media that can store program codes such as a U disk, a mobile hard disk, a ROM, a RAM, a magnetic disk or an optical disk, etc.

[0559] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method of communication, comprising: The method comprises: receiving first information, the first information indicating one or more of: performance information of a first model, generalization information of the first model, or information related to a first data set used for training of the first model; determining a monitoring manner of the first model according to the first information.

2. The method of claim 1, wherein, The method further comprises: sending second information, the second information indicating the monitoring manner of the first model.

3. The method according to claim 1 or 2, characterized in that, The first information indicates one or more of the performance information of the first model, the generalization information of the first model, or the information related to the first data set by indicating an identity of the first model.

4. The method according to any one of claims 1 to 3, characterized in that, The performance information of the first model comprises a performance level of the first model, the performance level of the first model being related to an expected performance of the first model, and / or The generalization information of the first model comprises a generalization level of the first model, the generalization level of the first model being related to an expected generalization capability of the first model.

5. The method according to any one of claims 1 to 4, characterized in that, The first information further indicates a model category of the first model.

6. The method of claim 5, wherein, The model category of the first model comprises a base general model or a cell-specific model.

7. The method according to any one of claims 1 to 6, characterized in that, The information related to the first data set comprises an identity of the first data set.

8. The method according to any one of claims 1 to 7, characterized in that, The information related to the first data set comprises one or more of: performance information corresponding to the first data set or generalization information corresponding to the first data set.

9. The method of claim 8, wherein, The performance information corresponding to the first data set comprises a performance level of a model trained by the first data set, the performance level of the model trained by the first data set being related to an expected performance of the model trained by the first data set, and / or The generalization information corresponding to the first data set comprises a generalization level of the model trained by the first data set, the generalization level of the model trained by the first data set being related to an expected generalization capability of the model trained by the first data set.

10. The method according to any one of claims 1 to 9, characterized in that, The monitoring parameter adopted by the monitoring manner of the first model comprises at least one of: a performance threshold of the first model, a monitoring frequency of the first model, a monitoring duration of the first model, a monitoring number of times of the first model, a monitoring error tolerance of the first model, or a switching threshold of the first model.

11. A method of communication, comprising: The method comprises: sending first information, the first information indicating one or more of: performance information of a first model, generalization information of the first model, or information related to a first data set used for training of the first model, the first information being used for determination of a monitoring manner of the first model.

12. The method of claim 11, wherein, The method further comprises: receiving second information, the second information indicating the monitoring manner of the first model.

13. The method of claim 11, wherein, The method further comprises: sending third information, the third information indicating model monitoring on the first model.

14. The method according to any one of claims 11 to 13, characterized in that, The first information indicates one or more of the performance information of the first model, the generalization information of the first model, or the information related to the first data set by indicating an identity of the first model.

15. The method according to any one of claims 11 to 14, characterized in that, The performance information of the first model comprises a performance level of the first model, the performance level of the first model being related to an expected performance of the first model, and / or The generalization information of the first model comprises a generalization level of the first model, the generalization level of the first model being related to an expected generalization capability of the first model.

16. The method according to any one of claims 11 to 15, characterized in that, The first information further indicates a model category of the first model.

17. The method of claim 16, wherein, The model category of the first model comprises a base general model or a cell-specific model.

18. The method according to any one of claims 11 to 17, characterized in that, The related information of the first data set comprises an identity of the first data set.

19. The method according to any one of claims 11 to 18, characterized in that, The related information of the first data set comprises one or more of the following: performance information corresponding to the first data set or generalization information corresponding to the first data set.

20. The method of claim 19, wherein, The performance information corresponding to the first data set comprises a performance level of a model trained by the first data set, the performance level of the model trained by the first data set being related to an expected performance of the model trained by the first data set, and / or The generalization information corresponding to the first data set comprises a generalization level of a model trained by the first data set, the generalization level of the model trained by the first data set being related to an expected generalization capability of the model trained by the first data set.

21. The method of any one of claims 11 to 20, wherein, The monitoring manner of the first model adopts a monitoring parameter comprising at least one of the following: a performance threshold of the first model, a monitoring frequency of the first model, a monitoring duration of the first model, a monitoring number of times of the first model, a monitoring error tolerance of the first model, or a switching threshold of the first model.

22. A computer-readable storage medium, characterized in that, The computer readable storage medium comprises instructions which, when executed by a processor, cause the method of any one of claims 1 to 10 to be implemented, or cause the method of any one of claims 11 to 21 to be implemented.

23. A communications device, characterized by The communication apparatus comprises a processor coupled to a storage medium, the storage medium storing instructions which, when executed by the processor, cause the communication apparatus to perform the method of any one of claims 1 to 10, or perform the method of any one of claims 11 to 21.

24. A communications device, characterized by The apparatus comprises a module to perform the method of any one of claims 1 to 10.

25. A communications device, characterized by The apparatus comprises a module to perform the method of any one of claims 11 to 21.

26. A communications device, characterized by The apparatus comprises one or more processors to process data and / or information to cause the method of any one of claims 1 to 10 to be implemented, or cause the method of any one of claims 11 to 21 to be implemented.

27. A chip, characterized by The apparatus comprises a processor to execute a program or instructions to cause the method of any one of claims 1 to 10 to be implemented, or cause the method of any one of claims 11 to 21 to be implemented.

28. A computer program product, characterised in that, The apparatus comprises computer program code or instructions which, when executed, cause the method of any one of claims 1 to 10 to be implemented, or cause the method of any one of claims 11 to 21 to be implemented.

29. A communication system, characterized by The communication device according to claim 24, and / or the communication device according to claim 25.

Images