WO2024067245A1 - 模型匹配的方法和通信装置 - Google Patents

模型匹配的方法和通信装置 Download PDF

Info

Publication number
WO2024067245A1
WO2024067245A1 PCT/CN2023/119662 CN2023119662W WO2024067245A1 WO 2024067245 A1 WO2024067245 A1 WO 2024067245A1 CN 2023119662 W CN2023119662 W CN 2023119662W WO 2024067245 A1 WO2024067245 A1 WO 2024067245A1
Authority
WO
WIPO (PCT)
Prior art keywords
model
communication device
identifier
models
group
Prior art date
Application number
PCT/CN2023/119662
Other languages
English (en)
French (fr)
Inventor
柴晓萌
孙琰
庞继勇
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2024067245A1 publication Critical patent/WO2024067245A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/70—Services for machine-to-machine communication [M2M] or machine type communication [MTC]

Definitions

  • the present application relates to the field of communication technology, and more specifically, to a model matching method and a communication device.
  • AI artificial intelligence
  • the multiple sub-AI models that constitute the two-end model can be deployed in different communication devices.
  • the sub-models deployed in different communication devices need to be matched so that the sub-AI models deployed in different communication devices can be used to jointly perform some communication tasks.
  • how to match the sub-models deployed in different communication devices is a problem to be solved.
  • the present application provides a model matching method and a communication device, so as to achieve matching of AI models deployed in different communication devices.
  • a model matching method is provided, which can be executed by a communication device, or can also be executed by a chip or circuit used for a communication device.
  • the communication device can be a second communication device, or a third communication device different from the second communication device, which is not limited in the present application.
  • the method may include: receiving an identifier of a first artificial intelligence (AI) model from a first communication device; and determining, based on the identifier of the first AI model, whether a second communication device deploys a second AI model, wherein the second AI model is an AI model that matches the first AI model.
  • AI artificial intelligence
  • communication devices can match AI models by exchanging AI model identifiers. For example, a first communication device sends an identifier of a first AI model to a second communication device, and the second communication device can identify the first AI model based on the identifier of the first AI model, and then determine whether it has deployed an AI model that matches the first AI model.
  • the second communication device can identify the first AI model based on the identifier of the first AI model, and then determine whether it has deployed an AI model that matches the first AI model.
  • the privacy of the AI model can be protected by transmitting the identifier of the AI model instead of the AI model, thereby reducing the signaling overhead caused by transmitting the AI model.
  • determining whether the second communication device deploys the second AI model based on the identifier of the first AI model includes: determining whether the second communication device deploys the second AI model based on the identifier and an association relationship of the first AI model, wherein the association relationship represents the relationship between the identifiers of mutually matching AI models.
  • the identifiers of mutually matching AI models have an associated relationship, so when determining whether to deploy a second AI model matching the first AI model based on the identifier of the first AI model, it is possible to jointly determine whether the second AI model matching the first AI model is deployed in the second communication device based on the associated relationship and the identifier of the first AI model. For example, it is possible to first determine that the AI model matching the first AI model is the second AI model based on the associated relationship, and then determine whether the second AI model is deployed in the second communication device. For another example, it is possible to first determine which AI models are deployed in the second communication device, and then determine whether the AI deployed in the second communication device has an AI model matching the first AI model based on the associated relationship.
  • the method also includes: in a case where it is determined that the second communication device has not deployed the second AI model, sending first feedback information to the first communication device, the first feedback information being used to provide feedback that the second communication device has not deployed the second AI model; or, in a case where it is determined that the second communication device has deployed the second AI model, sending second feedback information to the first communication device, the second feedback information being used to provide feedback that the second communication device has deployed the second AI model.
  • the method when determining that the second communication device deploys the second AI model In this case, the method further includes: sending an identifier of the second AI model to the first communication device, where the identifier of the second AI model is used to determine whether the first AI model in the first communication device matches the second AI model.
  • the identifier of the second AI model can be sent to the first communication device so that the first communication device can again perform the judgment on whether the first AI model matches the second AI model. In this way, it is equivalent to at least two communication devices performing the AI model matching operation, which can reduce the probability of AI model matching errors.
  • the second AI model belongs to a first group of AI models
  • sending the identifier of the second AI model to the first communication device includes: sending the identifier of at least one AI model in the first group of AI models to the first communication device.
  • sending the identifier of the second AI model to the first communication device includes: sending the identifier of the second AI model and the identifier of at least one AI model in the first group of AI models other than the second AI model to the first communication device.
  • the sending the identifier of the second AI model to the first communication device includes: sending the group identifier of the first group of AI models to the first communication device.
  • the group identifier of the first group of AI models is an identifier of the AI model group.
  • the identifiers of each AI model in the AI model group can be obtained.
  • the second AI model belongs to a group of AI models (such as recorded as the first group of AI models)
  • the identifier of at least one AI model in the first group of AI models or the group identifier of the first group of AI models can be sent to the first communication device.
  • the first communication device can determine whether the first AI model matches at least one AI model in the first group of AI models based on the identifier of the at least one AI model or the group identifier of the first group of AI models.
  • each AI model in the first group of AI models satisfies: when the input information of each AI model in the first group of AI models is the same, the output information is the same or the difference in output information is within a preset range.
  • the method when it is determined that the second communication device has not deployed the second AI model, the method further includes: sending an identifier of a fourth AI model to the first communication device, the identifier of the fourth AI model being used to determine whether the fourth AI model in the first communication device matches at least one AI model deployed by the first communication device.
  • the fourth AI model is an AI model deployed in the second communication device.
  • the fourth AI model belongs to a third group of AI models
  • sending the identifier of the fourth AI model to the first communication device includes: sending the identifier of at least one AI model in the third group of AI models to the first communication device.
  • sending the identifier of the fourth AI model to the first communication device includes: sending the identifier of the fourth AI model and the identifier of at least one AI model in the third group of AI models except the fourth AI model to the first communication device.
  • sending the identifier of the fourth AI model to the first communication device includes: sending the group identifier of the third group of AI models to the first communication device.
  • the fourth AI model belongs to a certain group of AI models (such as recorded as the third group of AI models)
  • the identifier of at least one AI model in the third group of AI models can be sent to the first communication device.
  • the first communication device can determine whether the first communication device has deployed an AI model that matches at least one AI model in the third group of AI models based on the identifier of at least one AI model in the third group of AI models.
  • each AI model in the third group of AI models satisfies: when the input information of each AI model in the third group of AI models is the same, the output information is the same or the difference in the output information is within a preset range.
  • the second AI model is an AI model that matches at least one AI model in a second group of AI models
  • the second group of AI models includes the first AI model
  • the receiving of the identifier of the first AI model from the first communication device includes: receiving the identifier of at least one AI model in the second group of AI models from the first communication device; determining whether the second communication device deploys the second AI model based on the identifier of the first AI model includes: determining whether the second communication device deploys the second AI model based on the identifier of at least one AI model in the second group of AI models.
  • the receiving the identifier of the first AI model from the first communication device includes: receiving the identifier of the first AI model from the first communication device and the identifier of at least one AI model in the second group of AI models other than the first AI model.
  • the receiving an identifier of the first AI model from the first communication device includes: receiving an identifier of the first AI model from the first communication device.
  • the group identifier of the second group of AI models of the device is the identifier of the AI model group.
  • the identifiers of each AI model in the AI model group can be obtained.
  • the second AI model may be an AI model that matches at least one AI model in a group of AI models (such as the second group of AI models).
  • a group of AI models such as the second group of AI models.
  • the second group of AI models includes the first AI model and a third AI model, and the priority of the first AI model is higher than the priority of the third AI model, and determining whether the second communication device deploys the second AI model based on the identifier of the first AI model and the identifier of at least one AI model other than the first AI model in the second group of AI models includes: when it is determined that the second communication device has not deployed the second AI model based on the identifier of the first AI model, determining whether the second communication device deploys the second AI model based on the identifier of the third AI model.
  • the priorities of each AI model in the AI model group can be different. In this way, when matching based on the identifiers of each AI model in the AI model group, they can be matched in order according to the priority.
  • each AI model in the second group of AI models satisfies: when the input information of each AI model in the second group of AI models is the same, the output information is the same or the difference in output information is within a preset range.
  • the identifier of the first AI model indicates at least one of the following: the type of the first communication device to which the first AI model belongs, the identifier of the first communication device to which the first AI model belongs, the manufacturer of the first communication device to which the first AI model belongs, the function of the first AI model, the scenario applicable to the first AI model, the data set of the first AI model, the communication parameters applicable to the first AI model, or the version of the first AI model.
  • the type identifier of the first communication device to which the first AI model belongs is used to indicate the type of the communication device that trains and/or uses the first AI model.
  • the identifier of the first communication device to which the first AI model belongs is used to indicate the communication device that trains and/or uses the first AI model.
  • the manufacturer identification of the first communication device to which the first AI model belongs is used to indicate the manufacturer that trained and/or used the first AI model.
  • the function identifier of the first AI model is used to indicate the function of the first AI model.
  • the function of the first AI model can be understood as a problem that the first AI model can solve, or a task that the first AI model can perform.
  • the scenario identifier to which the first AI model is applicable is used to indicate the scenario to which the first AI model is applicable, such as an indoor scene, an outdoor scene, an urban scene, or a suburban scene.
  • the data set identifier of the first AI model is used to indicate which data set the first AI model is trained on.
  • the data set identifier may have a corresponding relationship with the scene identifier.
  • the communication parameter identifier applicable to the first AI model is used to indicate the wireless parameter configuration applicable to the first AI model.
  • the wireless parameter configuration may include at least one of the following configurations: antenna configuration, bandwidth, rank, CSI feedback overhead, reference signal configuration, carrier frequency, or subcarrier spacing, etc.
  • the version identifier of the first AI model is used to indicate the version of the first AI model.
  • the versions of the AI model may be different versions of the same AI model, or may be different AI models.
  • the version identifier may indicate one or more of the complexity or performance of the AI model.
  • the version identifier may include a complexity identifier, or a performance identifier, or a complexity identifier and a performance identifier.
  • the identifier of the first AI model is predefined, or the identifier of the first AI model is configured by the communication device.
  • the identifier of the first AI model is configured by a communication device, including: the identifier of the first AI model is configured by the first communication device; or the identifier of the first AI model is configured by other communication devices.
  • the identifier of the first AI model can be predefined, or can be determined by the first communication device itself, or can be determined by other devices without limitation.
  • the method further includes: obtaining a validity period of an identifier of the first AI model; and determining whether the second communication device deploys a second AI model based on the identifier of the first AI model includes: When the validity period of the identifier of the first AI model determines that the identifier of the first AI model is valid, it is determined whether the second communication device deploys the second AI model based on the identifier of the first AI model.
  • a model matching method is provided, which can be performed by a communication device, or can also be performed by a chip or circuit used for a communication device, and this application does not limit this.
  • a communication device or can also be performed by a chip or circuit used for a communication device, and this application does not limit this.
  • the following is an example of execution by a first communication device.
  • the method may include: a first communication device sends an identifier of a first artificial intelligence (AI) model, and the identifier of the first AI model is used to determine whether a second communication device has deployed a second AI model that matches the first AI model; and the first communication device determines whether the second communication device has deployed the second AI model based on the response.
  • AI artificial intelligence
  • the method also includes: the first communication device receives first feedback information, and the first feedback information is used to feedback that the second communication device has not deployed the second AI model; determining whether the second communication device has deployed the second AI model based on the response of the second communication device includes: determining that the second communication device has not deployed the second AI model based on the first feedback information.
  • the method also includes: the first communication device receives second feedback information, and the second feedback information is used to feedback that the second communication device deploys the second AI model; determining whether the second communication device deploys the second AI model based on the response of the second communication device includes: determining that the second communication device deploys the second AI model based on the second feedback information.
  • the method after the first communication device sends the identifier of the first AI model, the method also includes: the first communication device receives the identifier of the second AI model; and the first communication device determines whether the first AI model matches the second AI model based on the identifier of the second AI model.
  • the second AI model belongs to a first group of AI models
  • the first communication device receives the identifier of the second AI model, including: the first communication device receives the identifier of the second AI model and the identifier of at least one AI model in the first group of AI models other than the second AI model; the first communication device determines whether the first AI model matches the second AI model based on the identifier of the second AI model and the identifier of at least one AI model in the first group of AI models other than the second AI model, including: the first communication device determines whether the first AI model matches at least one AI model in the first group of AI models based on the identifier of the second AI model and the identifier of at least one AI model in the first group of AI models other than the second AI model.
  • each AI model in the first group of AI models satisfies: when the input information of each AI model in the first group of AI models is the same, the output information is the same or the difference in output information is within a preset range.
  • the method after the first communication device sends the identifier of the first AI model, the method also includes: the first communication device receives the identifier of a fourth AI model; and the first communication device determines, based on the identifier of the fourth AI model, whether the first communication device has deployed an AI model that matches the fourth AI model.
  • the fourth AI model belongs to the third group of AI models
  • the first communication device receives the identifier of the fourth AI model, including: the first communication device receives the identifier of the fourth AI model and the identifier of at least one AI model in the third group of AI models other than the fourth AI model; the first communication device determines, based on the identifier of the fourth AI model, whether the first communication device deploys an AI model that matches the fourth AI model, including: the first communication device determines, based on the identifier of the fourth AI model and the identifier of at least one AI model in the third group of AI models other than the fourth AI model, whether the first communication device deploys an AI model that matches the third group of AI models.
  • each AI model in the third group of AI models satisfies: when the input information of each AI model in the third group of AI models is the same, the output information is the same or the difference in the output information is within a preset range.
  • the second AI model is an AI model that matches at least one AI model in a second group of AI models
  • the second group of AI models includes the first AI model
  • the first communication device sends the identifier of the first AI model, including: the first communication device sends the identifier of the first AI model and the identifier of at least one AI model in the second group of AI models other than the first AI model.
  • each AI model in the second group of AI models satisfies: when the input information of each AI model in the second group of AI models is the same, the output information is the same or the difference in output information is within a preset range.
  • the method also includes: the first communication device sends a request message, where the request message is used to request an identifier of the first AI model; and the first communication device receives the identifier of the first AI model.
  • the first communication device can request the identifier of the first AI model from other devices, so that the other devices can
  • the identifiers of the various AI models are uniformly assigned, thereby avoiding duplication of the AI model identifiers when each communication device is configured with its own AI model identifier.
  • the request information includes at least one of the following: the type of the first communication device to which the first AI model belongs, the identifier of the first communication device to which the first AI model belongs, the manufacturer of the first communication device to which the first AI model belongs, the function of the first AI model, the scenario applicable to the first AI model, the data set of the first AI model, the communication parameters applicable to the first AI model, or the version of the first AI model.
  • the method also includes: the first communication device sends registration information, the registration information includes an identifier of the first AI model, and the registration information is used to register the identifier of the first AI model.
  • the first communication device can register the identifier of the first AI model with other devices, so that the identifiers of various AI models can be managed uniformly by other devices.
  • the identifier of the first AI model indicates at least one of the following: the type of the first communication device to which the first AI model belongs, the identifier of the first communication device to which the first AI model belongs, the manufacturer of the first communication device to which the first AI model belongs, the function of the first AI model, the scenario applicable to the first AI model, the data set of the first AI model, the communication parameters applicable to the first AI model, or the version of the first AI model.
  • the method further includes: the first communication device sends the validity period of the first AI model.
  • a model matching method is provided, which can be executed by a communication device, or can also be executed by a chip or circuit used for a communication device.
  • the communication device can be a second communication device, or a third communication device different from the second communication device, which is not limited in the present application.
  • the method may include: receiving request information from a first communication device, the request information being used to request an AI model that matches a first artificial intelligence (AI) model; and determining, in response to the request information, whether the second communication device deploys a second AI model, the second AI model being an AI model that matches at least one AI model in a second group of AI models, the second group of AI models including the first AI model.
  • AI artificial intelligence
  • AI models can be grouped.
  • the first communication device sends a request message to the second communication device, and the request message is used to request an AI model that matches the first AI model; since the first AI model belongs to the second group of AI models, the second communication device can determine whether an AI model that matches at least one AI model in the second group of AI models is deployed in the second communication device. If an AI model that matches at least one AI model in the second group of AI models is deployed in the second communication device (such as recorded as the second AI model), it can be considered that the second AI model matches each AI model in the second group of AI models.
  • the method when determining that the second communication device deploys the second AI model, the method also includes: determining that each AI model in the first group of AI models matches each AI model in the first group of AI models, and the first group of AI models includes the second AI model.
  • each AI model in the first group of AI models satisfies: when the input information of each AI model in the first group of AI models is the same, the output information is the same or the difference in output information is within a preset range.
  • each AI model in the second group of AI models satisfies: when the input information of each AI model in the second group of AI models is the same, the output information is the same or the difference in output information is within a preset range.
  • a communication method is provided.
  • the method may be executed by a communication device, or may be executed by a chip or circuit used for a communication device, and the present application does not limit this.
  • the method may include: sending a request message, wherein the request message requests an identifier of an artificial intelligence (AI) model; and receiving the identifier of the AI model.
  • AI artificial intelligence
  • the first communication device can request the identifier of the first AI model from other devices, so that the identifiers of each AI model can be uniformly allocated by other devices, thereby avoiding the duplication of AI model identifiers when each communication device configures its own AI model identifier.
  • a communication method is provided.
  • the method can be executed by a communication device, or can also be executed by a chip or circuit used for a communication device, and the present application does not limit this.
  • the method may include: receiving request information, the request information requesting an identifier of an artificial intelligence (AI) model; determining the identifier of the AI model in response to the request information; and sending the identifier of the AI model.
  • AI artificial intelligence
  • the request information includes at least one of the following: the type of communication device to which the AI model belongs, the identifier of the communication device to which the AI model belongs, the manufacturer of the communication device to which the AI model belongs, the function of the AI model, the scenario in which the AI model is applicable, the data set of the AI model, the communication parameters applicable to the AI model, or the version of the AI model.
  • a communication method is provided.
  • the method may be executed by a communication device, or may be executed by a chip or circuit used for a communication device, and the present application does not limit this.
  • the method may include: obtaining an identifier of an artificial intelligence (AI) model; sending registration information, wherein the registration information includes the identifier of the AI model, and the registration information is used to register the identifier of the AI model.
  • AI artificial intelligence
  • the first communication device can register the identifier of the first AI model with other devices, so that the identifiers of various AI models can be managed uniformly by other devices.
  • a communication method is provided.
  • the method may be executed by a communication device, or may be executed by a chip or circuit used for a communication device, and the present application does not limit this.
  • the method may include: receiving registration information, the registration information including an identifier of an artificial intelligence (AI) model, and the registration information is used to register the identifier of the AI model; and saving the identifier of the AI model in response to the registration information.
  • AI artificial intelligence
  • a communication device is provided, the device being used to execute the method provided in any one of the first to seventh aspects.
  • the device may include a unit and/or module, such as a processing unit and/or a communication unit, for executing the method provided in any one of the above implementations of any one of the first to seventh aspects.
  • the device is a communication device (such as a terminal device or a network device).
  • the communication unit may be a transceiver or an input/output interface; the processing unit may be at least one processor.
  • the transceiver may be a transceiver circuit.
  • the input/output interface may be an input/output circuit.
  • the device is a chip, a chip system or a circuit used in a communication device.
  • the communication unit may be an input/output interface, an interface circuit, an output circuit, an input circuit, a pin or a related circuit on the chip, the chip system or the circuit;
  • the processing unit may be at least one processor, a processing circuit or a logic circuit.
  • a communication device comprising: a memory for storing programs; and at least one processor for executing computer programs or instructions stored in the memory to execute a method provided in any one of the above-mentioned implementations of any one of the above-mentioned first to seventh aspects.
  • the apparatus is a communication device (such as a terminal device or a network device).
  • the apparatus is a chip, a chip system or a circuit used in a communication device.
  • the present application provides a processor for executing the methods provided in the above aspects.
  • a computer-readable storage medium which stores a program code for execution by a device, and the program code includes a method provided by any one of the above-mentioned implementation methods for executing any one of the above-mentioned first to seventh aspects.
  • a computer program product comprising instructions is provided.
  • the computer program product is run on a computer, the computer is caused to execute a method provided by any one of the above-mentioned implementations of any one of the above-mentioned first to seventh aspects.
  • a chip including a processor and a communication interface, the processor reads instructions stored in a memory through the communication interface, and executes a method provided by any one of the above-mentioned implementation methods of any one of the above-mentioned first to seventh aspects.
  • the chip also includes a memory, in which a computer program or instruction is stored, and the processor is used to execute the computer program or instruction stored in the memory.
  • the processor is used to execute the method provided in any one of the above-mentioned implementation methods of any one of the first to seventh aspects.
  • a communication system comprising the aforementioned first communication device and the second communication device; or, comprising the aforementioned first communication device and the third communication device; or, comprising the aforementioned first communication device, the second communication device, and the third communication device.
  • FIG. 1 is a schematic diagram of a wireless communication system 100 applicable to an embodiment of the present application.
  • FIG. 2 is a schematic diagram of a wireless communication system 200 applicable to an embodiment of the present application.
  • Figure 3 is a schematic diagram of the neuron structure.
  • FIG4 is a schematic diagram of the layer relationship of a neural network.
  • FIG5 is a schematic diagram of a two-terminal model.
  • FIG. 6 is a schematic diagram of a communication method 600 provided in an embodiment of the present application.
  • FIG. 7 is a schematic diagram of matching two groups of AI models applicable to an embodiment of the present application.
  • FIG8 is a schematic diagram of an identification of an AI model applicable to an embodiment of the present application.
  • FIG. 9 is another schematic diagram of the identification of the AI model applicable to the embodiment of the present application.
  • FIG. 10 is another schematic diagram of the identification of the AI model applicable to the embodiment of the present application.
  • FIG. 11 is a schematic diagram of the registration of an identifier of an AI model applicable to an embodiment of the present application.
  • FIG. 12 is a schematic diagram of a notification of an identification of an AI model applicable to an embodiment of the present application.
  • FIG. 13 is a schematic diagram of a communication method 1300 provided according to an embodiment of the present application.
  • FIG. 14 is a schematic diagram of a communication method 1400 provided according to another embodiment of the present application.
  • FIG. 15 is a schematic diagram of a communication method 1500 provided according to another embodiment of the present application.
  • FIG. 16 is a schematic diagram of a communication method 1600 provided according to another embodiment of the present application.
  • FIG. 17 is a schematic block diagram of a communication device 1700 provided in an embodiment of the present application.
  • FIG. 18 is a schematic diagram of another communication device 1800 provided according to an embodiment of the present application.
  • FIG. 19 is a schematic diagram of a chip system 1900 provided according to an embodiment of the present application.
  • the technical solution provided in this application can be applied to various communication systems, such as: the fifth generation (5th generation, 5G) or new radio (new radio, NR) system, long term evolution (long term evolution, LTE) system, LTE frequency division duplex (frequency division duplex, FDD) system, LTE time division duplex (time division duplex, TDD) system, etc.
  • the technical solution provided in this application can also be applied to future communication systems, such as the sixth generation mobile communication system.
  • the technical solution provided in this 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.
  • D2D device to device
  • V2X vehicle to everything
  • M2M machine to machine
  • MTC machine type communication
  • IoT Internet of things
  • the terminal devices in the embodiments of the present application include various devices with wireless communication functions, which can be used to connect people, objects, machines, etc.
  • the terminal devices can be widely used in various scenarios, such as: cellular communication, D2D, V2X, peer to peer (P2P), M2M, MTC, IoT, virtual reality (VR), augmented reality (AR), industrial control, automatic driving, telemedicine, smart grid, smart furniture, smart office, smart wear, smart transportation, smart city drones, robots, remote sensing, passive sensing, positioning, navigation and tracking, autonomous delivery, etc.
  • the terminal device can be a terminal in any of the above scenarios, such as an MTC terminal, an IoT terminal, etc.
  • the terminal device can be a user equipment (UE), terminal, fixed device, mobile station device or mobile device of the third generation partnership project (3GPP) standard, a subscriber unit, a handheld device, a vehicle-mounted device, a wearable device, a cellular phone, a smart phone, a SIP phone, a wireless data card, a personal digital assistant (PDA), a computer, a tablet computer, a notebook computer, a wireless modem, a handheld device (handset), a laptop computer, a computer with wireless transceiver function, a smart book, a vehicle, a satellite, a global positioning system (GPS) device, a target tracking device, an aircraft (such as a drone, a helicopter, a multi-copter, a quadcopter, or an airplane), a ship, a remote control device, a smart home device, an industrial device, or a device built into the above device (for example, a communication module, a modem or a chip in the above device), or other processing devices connected to the wireless
  • the UE can also be used to act as a base station.
  • the UE can act as a scheduling entity in V2X, D2D or provide sidelink signals between UEs in scenarios such as P2P.
  • the device for realizing the function of the terminal device can be the terminal device, or it can be a device that can support the terminal device to realize the function, such as a chip system or a chip, which can be installed in the terminal device.
  • the chip system can be composed of a chip, or it can include a chip and other discrete devices.
  • the network device in the embodiment of the present application may be a device for communicating with a terminal device, and the network device may also be referred to as an access network device or a wireless access network device, such as a base station.
  • the network device in the embodiment of the present application may refer to a wireless access network (RAN) node (or device) that connects a terminal device to a wireless network.
  • RAN wireless access network
  • Base station can broadly cover various names as follows, or be replaced with the following names, such as: NodeB, evolved NodeB (eNB), next generation NodeB (gNB), relay station, access point, transmitting point (TRP), transmitting point (TP), master station, auxiliary station, multi-standard wireless (motor slide retainer, 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), positioning node, etc.
  • NodeB evolved NodeB (eNB), next generation NodeB (gNB), relay station, access point, transmitting point (TRP), transmitting point (TP), master station, auxiliary station, multi-standard wireless (motor slide retainer, MSR) node, home base station, network controller, access node, wireless node, access point (AP), transmission node, transceiver node,
  • 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 may also refer to a communication module, modem or chip used to be set in the aforementioned equipment or device.
  • the base station may also be a mobile switching center and a device that performs the base station function in D2D, V2X, and M2M communications, a network-side device in a 6G network, and a device that performs the base station function in a future communication system.
  • the base station may support networks with the same or different access technologies. The embodiments of the present application do not limit the specific technology and specific device form used by the network equipment.
  • Base stations can be fixed or mobile.
  • a helicopter or drone can be configured to act as a mobile base station, and one or more cells can move based on the location of the mobile base station.
  • a helicopter or drone can be configured to act as a device that communicates with another base station.
  • the device for realizing the function of the network device can be a terminal device, or a device that can support the network device to realize the function, such as a chip system or a chip, which can be installed in the network device.
  • the chip system can be composed of a chip, or it can include a chip and other discrete devices.
  • the network equipment and terminal equipment can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; they can also be deployed on the water surface; they can also be deployed on aircraft, balloons and satellites in the air.
  • the embodiments of the present application do not limit the scenarios in which the network equipment and terminal equipment are located.
  • FIG1 is a schematic diagram of a wireless communication system 100 applicable to an embodiment of the present application.
  • the wireless communication system 100 may include at least one network device, such as the network device 110 shown in FIG1
  • the wireless communication system 100 may also include at least one terminal device, such as the terminal device 120 and the terminal device 130 shown in FIG1 .
  • Both the network device and the terminal device may be configured with multiple antennas, and the network device and the terminal device may communicate using multi-antenna technology.
  • Terminal devices may also communicate with each other. For example, terminal devices may communicate directly with each other. For another example, terminal devices may communicate with each other through other communication devices, such as network devices or other terminal devices.
  • the network device can manage one or more cells, and there can be an integer number of terminal devices in a cell.
  • the network device 110 and the terminal device 120 form a single-cell communication system, and without loss of generality, the cell is referred to as cell #1.
  • the network device 110 can be a network device in cell #1, or the network device 110 can serve a terminal device (such as terminal device 120) in cell #1.
  • a cell can be understood as an area within the coverage of wireless signals of network equipment.
  • FIG2 is a schematic diagram of a wireless communication system 200 applicable to an embodiment of the present application.
  • the wireless communication system 200 may include at least one network device, such as the network device 210 shown in FIG2
  • the wireless communication system 200 may also include at least one terminal device, such as the terminal device 220 and the terminal device 230 shown in FIG2
  • the wireless communication system 200 may also include at least one artificial intelligence (AI) node, such as the AI node 240 shown in FIG2 .
  • AI artificial intelligence
  • the AI node is deployed in any of the following: a network device, a terminal device, or a core network.
  • FIG. 2 is an exemplary illustration, which lists the case where the AI node 240 is deployed separately, such as being deployed in a location other than a network device and a terminal device.
  • the AI node 240 can communicate with the network device 210, and the AI node 240 can also communicate with the terminal device 220 and the terminal device 230 through the network device 210. It is understood that the AI node may also communicate directly with the terminal device, and this is not limited.
  • the AI node is used to perform AI-related operations.
  • the AI-related operations may include: model failure test, model performance test, model training test, data collection, etc.
  • the network device may forward the data related to the AI model reported by the terminal device to the AI node, and the AI node may perform AI-related operations.
  • the network device or the terminal device may forward the data related to the AI model to the AI node, and the AI node may perform AI-related operations.
  • the AI node may send the output of AI-related operations, such as trained neural network models, model evaluations, test results, etc., to the network device and/or the terminal device.
  • the AI node may directly send the output of AI-related operations to the network device and the terminal device.
  • the AI node may send the output of AI-related operations to the terminal device through the network device.
  • the AI node may send the output of AI-related operations to the network device through the terminal device.
  • the present application does not limit the number of AI nodes.
  • the multiple AI nodes can be divided based on functions, such as different AI nodes are responsible for different functions.
  • AI nodes can be independent devices, or they can be integrated into the same device to implement different functions, or they can be network elements in hardware devices, or they can be software functions running on dedicated hardware, or they can be virtualized functions instantiated on a platform (for example, a cloud platform).
  • a platform for example, a cloud platform
  • the communication device (such as the first communication device, the second communication device, and the third communication device) can be a terminal device or a component of a terminal device (such as a chip or a circuit); or it can be a network device or a component of a network device (such as a chip or a circuit); or it can be an AI node or a component of an AI node (such as a chip or a circuit), without limitation.
  • Figures 1 and 2 are simplified schematic diagrams for ease of understanding, and the wireless communication system may also include other network devices, or may also include other terminal devices, or may also include other AI nodes, which are not drawn in Figures 1 and 2.
  • Artificial Intelligence It is to enable machines to learn, accumulate experience, and solve problems that humans can solve through experience, such as natural language understanding, image recognition, and chess. Artificial Intelligence can be understood as the intelligence displayed by machines made by humans. Usually artificial intelligence refers to the technology that presents human intelligence through computer programs. The goals of artificial intelligence include understanding intelligence by building computer programs with symbolic reasoning or inference.
  • Machine learning It is a way to implement artificial intelligence. Machine learning is a method that can give machines the ability to learn, so that machines can complete functions that cannot be completed by direct programming. In a practical sense, machine learning is a method of using data to train a model and then using the model to predict. There are many methods of machine learning, such as neural networks (NN), decision trees, support vector machines, etc. The theory of machine learning is mainly to design and analyze some algorithms that allow computers to learn automatically. Machine learning algorithms are a type of algorithm that automatically analyzes data to obtain rules and uses the rules to predict unknown data.
  • NN neural networks
  • decision trees decision trees
  • support vector machines etc.
  • the theory of machine learning is mainly to design and analyze some algorithms that allow computers to learn automatically.
  • Machine learning algorithms are a type of algorithm that automatically analyzes data to obtain rules and uses the rules to predict unknown data.
  • Neural network It is a specific embodiment of machine learning methods. Neural network is a mathematical model that imitates the behavioral characteristics of animal neural networks and processes information. The idea of neural network comes from the neuron structure of brain tissue. Each neuron can perform a weighted sum operation on its input value, and the result of the weighted sum operation is used to generate an output through an activation function.
  • FIG3 is a schematic diagram of a neuron structure.
  • the bias of the weighted sum is b.
  • b can be an integer, a decimal, or a complex number.
  • the form of the activation function can be diversified.
  • the output of the neuron is: As shown in Figure 3.
  • the activation functions of different neurons in a neural network can be the same or different.
  • a neural network generally includes a multi-layer structure, and each layer may include one or more logic judgment units, which may be called neurons.
  • each layer may include one or more logic judgment units, which may be called neurons.
  • the depth of a neural network can be understood as the number of layers the neural network includes, and the number of neurons included in each layer can be called the width of the layer.
  • FIG4 is a schematic diagram of the layer relationship of a neural network.
  • the neural network includes an input layer and an output layer.
  • the input layer of the neural network receives the input through the neural network. After meta-processing, the result is passed to the output layer, and the output result of the neural network is obtained from the output layer.
  • the neural network includes an input layer, a hidden layer and an output layer, as shown in Figure 4.
  • the input layer of the neural network processes the received input through neurons and passes the result to the middle hidden layer.
  • the hidden layer then passes the calculation result to the output layer or the adjacent hidden layer, and finally the output layer obtains the output result of the neural network.
  • a neural network can include one or more hidden layers connected in sequence without restriction.
  • a loss function can be defined.
  • the loss function is used to measure the difference between the predicted value of the model and the true value.
  • the loss function describes the gap or difference between the output value of the neural network and the ideal target value.
  • the training process of a neural network is the process of adjusting the neural network parameters so that the value of the loss function is less than the threshold value or meets the target requirements.
  • the neural network parameters may include at least one of the following: the number of layers, width, weights of neurons, or parameters in the activation function of neurons.
  • AI model It is an algorithm or computer program that can realize AI functions.
  • AI model represents the mapping relationship between the input and output of the model, or it can be said that AI model is a function model that maps input of a certain dimension to output of a certain dimension.
  • the parameters of the function model can be obtained through machine learning training.
  • a and b are the parameters of the AI model, which can be obtained through machine learning training.
  • the implementation of the AI model can be a hardware circuit, or software, or a combination of software and hardware, without limitation.
  • Non-limiting examples of software include: program code, program, subroutine, instruction, instruction set, code, code segment, software module, application, or software application, etc.
  • Dataset Data used for model training, model validation, or model testing in machine learning. The quantity and quality of data will affect the effectiveness of machine learning.
  • Hyperparameters the number of neural network layers, the number of neurons, activation functions, loss functions and other parameters.
  • Model training By selecting a suitable loss function and using the optimization algorithm to train the model parameters, the loss function value is minimized.
  • Model application Use the trained model to solve practical problems.
  • Two-ended model also called two-sided model, collaborative model, dual model, etc.
  • the two-ended model refers to: a model composed of at least two AI models combined together.
  • the at least two AI models can be deployed on at least two nodes, that is, the at least two AI models are not deployed on the same node, and the multiple AI models constituting the two-ended model match each other.
  • AI model #1 and AI model #2 match each other, indicating that AI model #1 can understand the output of AI model #2, and can decode the output of AI model #2 into the expected output.
  • an auto-encoder in which the encoder and decoder are deployed on different nodes is a two-end model.
  • the encoder and decoder of AE match each other, that is, the decoder can understand the output of the encoder and can decode the output of the encoder into the desired output.
  • FIG5 is a schematic diagram of a two-end model.
  • AI model #1 is deployed in the encoder
  • AI model #2 is deployed in the decoder.
  • the input of AI model #1 is V
  • the output of AI model #1 is z.
  • the input of AI model #2 is z
  • the output of AI model #2 is V’, which is the same as V, or V’ can more accurately reflect V.
  • AI model #1 and AI model #2 can be trained on the same node and then deployed on two nodes respectively, or they can be trained on two nodes in a distributed manner.
  • indication may include direct indication, indirect indication, explicit indication, and implicit indication.
  • indication information may include direct indication, indirect indication, explicit indication, and implicit indication.
  • the information indicated by the indication information is referred to as the information to be indicated.
  • the information to be indicated can be directly indicated, such as the information to be indicated itself or the index of the information to be indicated.
  • the information to be indicated can also be indirectly indicated by indicating other information, wherein there is an association relationship between the other information and the information to be indicated. It is also possible to indicate only a part of the information to be indicated, while the other parts of the information to be indicated are known or agreed in advance.
  • the indication of specific information can also be achieved by means of the arrangement order of each information agreed in advance (for example, specified by the protocol), thereby reducing the indication overhead to a certain extent.
  • the information to be indicated can be sent together as a whole, or it can be divided into multiple sub-information and sent separately, and the sending period and/or sending time of these sub-information can be the same or different.
  • a communication device (such as a first communication device, a second communication device, and a third communication device) can be a terminal device or a component of a terminal device (such as a chip or a circuit); or it can also be a network device or a component of a network device (such as a chip or a circuit); or it can also be an AI node or a component of an AI node (such as a chip or a circuit).
  • Fig. 6 is a schematic diagram of a communication method 600 provided in an embodiment of the present application.
  • the method 600 shown in Fig. 6 may include the following steps.
  • the first communication device sends the identifier of the first AI model.
  • the first AI model may be an AI model deployed in the first communication device, that is, the first AI model is an AI model in the first communication device.
  • the number of the first AI models may be at least one, that is, the first communication device may send an identifier of an AI model, or may send identifiers of at least two AI models, without limitation.
  • communication devices can match AI models by exchanging AI model identifiers. For example, a first communication device sends an identifier of a first AI model to a second communication device, and the second communication device can identify the first AI model based on the identifier of the first AI model, and then determine whether it has deployed an AI model that matches the first AI model.
  • the identifier of the AI model is transmitted instead of the AI model, the privacy of the AI model can be protected, and the signaling overhead caused by transmitting the AI model can be reduced.
  • the above steps 610 and 620 may at least include the following implementation methods.
  • the second communication device receives an identifier of a first AI model from the first communication device; in step 620, the second communication device determines whether the second communication device deploys a second AI model based on the identifier of the first AI model.
  • the first communication device directly sends the identifier of the first AI model to the second communication device, and then the second communication device determines whether the second communication device deploys the second AI model based on the identifier of the first AI model.
  • a first communication device sends an identifier of a first AI model to other communication devices, and the other communication devices forward the identifier of the first AI model to a second communication device. Then, the second communication device determines whether the second communication device deploys a second AI model based on the identifier of the first AI model.
  • the third communication device receives an identifier of the first AI model from the first communication device; in step 620, the third communication device determines whether the second communication device deploys the second AI model based on the identifier of the first AI model.
  • the first communication device directly sends the identifier of the first AI model to the third communication device, and then the third communication device determines whether the second communication device deploys the second AI model based on the identifier of the first AI model.
  • a first communication device sends an identifier of a first AI model to other communication devices (such as a second communication device), and the other communication device forwards the identifier of the first AI model to a third communication device. Then, the third communication device determines whether the second communication device deploys the second AI model based on the identifier of the first AI model.
  • the first communication device directly sends the identifier of the first AI model to the third communication device
  • the second communication device directly sends the identifier of at least one AI model to the third communication device.
  • the third communication device determines whether there is an AI model in the at least one AI model that matches the first AI model, that is, determines whether there is a second AI model in the at least one AI model, thereby determining whether the second communication device deploys the second AI model.
  • the first communication device determines whether the second communication device deploys the second AI model based on the response.
  • the first communication device determines whether the second communication device deploys the second AI model based on the received feedback information. Based on this implementation, method 600 also includes: sending feedback information to the first communication device, where the feedback information is used to feedback whether the second communication device deploys the second AI model.
  • Example 1 When it is determined that the second communication device has not deployed the second AI model, feedback information is sent to the first communication device, where the feedback information is used to feedback that the second communication device has not deployed the second AI model.
  • the feedback information used to feedback that the second communication device has not deployed the second AI model is recorded as the first feedback information.
  • the second communication device determines that the second communication device has not deployed the second AI model based on the identifier of the first AI model, it sends the first feedback information to the first communication device.
  • the second communication device sends the first feedback information to other communication devices, and the other communication devices forward the first feedback information to the first communication device.
  • the third communication device determines that the second communication device has not deployed the second AI model based on the identifier of the first AI model, it sends the first feedback information to the first communication device.
  • the third communication device directly sends the first feedback information to the first communication device.
  • the third communication device sends the first feedback information to other communication devices (such as the second communication device), and the other communication devices forward the first feedback information to the first communication device.
  • Example 2 When it is determined that the second communication device deploys the second AI model, feedback information is sent to the first communication device, where the feedback information is used to feedback that the second communication device deploys the second AI model.
  • the feedback information used to feedback that the second communication device deploys the second AI model is recorded as the second feedback information.
  • the second communication device after the second communication device determines that the second communication device deploys the second AI model based on the identifier of the first AI model, it sends the second feedback information to the first communication device.
  • the second communication device directly sends the second feedback information to the first communication device.
  • the second communication device sends the second feedback information to other communication devices, and the other communication devices forward the second feedback information to the first communication device.
  • the third communication device determines that the second communication device deploys the second AI model based on the identifier of the first AI model, it sends the second feedback information to the first communication device.
  • the third communication device directly sends the second feedback information to the first communication device.
  • the third communication device sends the second feedback information to other communication devices (such as the second communication device), and the other communication devices forward the second feedback information to the first communication device.
  • the first communication device determines whether the second communication device deploys the second AI model based on not receiving feedback information.
  • the second communication device does not deploy the second AI model.
  • the second communication device determines that the second communication device has not deployed the second AI model based on the identifier of the first AI model, the second communication device does not send feedback information.
  • the feedback information may be an affirmative response.
  • the first communication device if the first communication device does not receive an affirmative response from the second communication device within a period of time (for distinction, recorded as time period #1), the first communication device assumes that the second communication device has not deployed a second AI model that matches the first AI model.
  • the starting moment of time period #1 may be the moment when the first communication device sends the identifier of the first AI model, and the duration of time period #1 may be predefined, or it may be estimated based on historical circumstances and is not restricted.
  • time period #1 may be implemented by a timer.
  • the second communication device deploys a second AI model.
  • the second communication device determines that the second communication device deploys the second AI model based on the identifier of the first AI model
  • the second communication device does not send feedback information.
  • the feedback information may be a negative response.
  • the first communication device if the first communication device does not receive a negative response from the second communication device within a period of time (for distinction, recorded as time period #2), the first communication device defaults to deploying a second AI model matching the first AI model in the second communication device.
  • the starting moment of time period #2 may be the moment when the first communication device sends the identifier of the first AI model, and the duration of time period #2 may be predefined, or it may be estimated based on historical circumstances and is not restricted.
  • time period #2 may be implemented by a timer.
  • determining whether the second communication device deploys the second AI model based on the identifier of the first AI model includes: determining whether the second communication device deploys the second AI model based on the identifier and an association relationship of the first AI model.
  • the association relationship represents the relationship between the identifiers of the AI models that match each other. Specifically, if two AI models are matched, then the identifiers of the two AI models are also associated, that is, there is an association relationship between the identifiers of the two AI models.
  • the association relationship includes the relationship between the identifier of AI model #1 and the identifier of AI model #2. That is, after the identifier of AI model #1 is determined, the identifier of AI model #2 that matches AI model #1 is also determined.
  • the communication device that deploys AI model #1 and/or the communication device that deploys AI model #2 can maintain the association relationship between the identifier of AI model #1 and the identifier of AI model #2.
  • association relationship may exist in the form of a table, a function, or a string, such as for storage or transmission.
  • Table 1 below is an example of presenting an association relationship in a table form.
  • AI model #1 and AI model #2 are matched AI models
  • AI model #3 and AI model #4 are matched AI models
  • AI model #5 and AI model #6, and AI model #7 are matched AI models
  • the association relationship includes: AI model #1’s label The relationship between the identification and the identification of AI model #2, the relationship between the identification of AI model #3 and the identification of AI model #4, the relationship between the identification of AI model #5 and the identification of AI model #6, and the relationship between the identification of AI model #5 and the identification of AI model #7.
  • the AI model that matches the first AI model is the second AI model based on the association relationship, and then determine whether the second AI model is deployed in the second communication device. For example, if the first AI model is AI model #1, then based on the identifier of AI model #1 and the association relationship shown in Table 1 above, it is determined that the AI model that matches the AI model #1 is AI model #2, and then determine whether the AI model #2 is deployed in the second communication device.
  • Another possible implementation method is to first determine which AI models are deployed in the second communication device, and then determine whether the AI models deployed in the second communication device have an AI model that matches the first AI model based on the association relationship. For example, if the AI models deployed in the second communication device include: AI model #2, AI model #4, and the first AI model is AI model #1, then based on the association relationship shown in Table 1 above, it can be known that AI model #1 matches AI model #2, so it can be determined that the AI model deployed in the second communication device that matches AI model #1 is AI model #2.
  • Table 1 is an exemplary description and is not limited to this.
  • one AI model can match a larger number of AI models.
  • the identifier of the AI model in Table 1 can be replaced with other information that can identify the AI model.
  • the association relationship can exist in the form of text, which is not limited to this.
  • AI models can be grouped. If there is at least one pair of matching AI models between two groups of AI models (such as AI models matched through training), the two groups of AI models are matched. Taking AI model group #A and AI model group #B as an example, if there is at least one pair of matching AI models between the two groups of AI models (such as AI models matched through training), any AI model in AI model group #A can match any AI model in AI model group #B.
  • FIG. 7 is a schematic diagram of matching two groups of AI models applicable to an embodiment of the present application.
  • AI models in AI model group #A are: AI model A1, AI model A2, AI model A3, and the AI models in AI model group #B are: AI model B1, AI model B2, AI model B3. It can be considered that any AI model in AI model group #A can be matched with any AI model in AI model group #B.
  • one AI model may match multiple AI models, as shown in Figure 7.
  • AI model A1 and AI model B1 are AI models that are matched through training
  • AI model A2 and AI model B1 are AI models that are matched through training
  • AI model A2 and AI model B2 are AI models that are matched through training
  • AI model A3 and AI model B2 are AI models that are matched through training.
  • both AI model A1 and AI model A2 match AI model B1, so it can be considered that AI model A1 and AI model A2 are functionally similar, and because AI model A2 and AI model B2 match, it can be considered that AI model A1 and AI model B2 are also matched. Therefore, it can be considered that any AI model in AI model group #A can match any AI model in AI model group #B.
  • the AI model group may be determined when the AI model is registered, but this is not limited.
  • AI models with similar functions can be grouped together. Functional similarity can be understood as when two AI models have the same input, their outputs are also the same, or the difference in output is within a preset range (such as less than a certain threshold).
  • the second AI model is an AI model that matches at least one AI model in the second group of AI models, and the second group of AI models includes the first AI model.
  • receiving the identifier of the first AI model from the first communication device includes: receiving the identifier of at least one AI model in the second group of AI models from the first communication device; in step 620, determining whether the second communication device deploys the second AI model based on the identifier of the first AI model includes: determining whether the second communication device deploys the second AI model based on the identifier of at least one AI model in the second group of AI models.
  • each AI model in the second group of AI models satisfies: when the input information of each AI model in the second group of AI models is the same, the output information is the same or the difference in the output information is within a preset range.
  • step 610 the identifiers of all AI models in the second group of AI models are received from the first communication device, and accordingly, in step 620, based on the identifiers of all AI models in the second group of AI models, it is determined whether the second communication device is partially If an AI model matching at least one AI model in the second group of AI models is deployed in the second communication device, it is determined that an AI model matching each AI model in the second group of AI models is deployed in the second communication device.
  • the second group of AI models includes a first AI model and a third AI model.
  • the identifier of the first AI model and the identifier of the third AI model are received from the first communication device.
  • the identifiers of each AI model can be sent simultaneously or in batches without restriction.
  • the identifiers of each AI model can be carried in the same signaling or in different signaling without restriction.
  • step 610 identifiers of some AI models in the second group of AI models are received from the first communication device, and accordingly, in step 620, it is determined whether the second communication device deploys the second AI model based on the identifiers of some AI models in the second group of AI models. If an AI model matching at least one AI model in the part of AI models is deployed in the second communication device, it is determined that an AI model matching each AI model in the second group of AI models is deployed in the second communication device.
  • the second group of AI models includes a first AI model and a third AI model.
  • an identifier of the first AI model or an identifier of the third AI model is received from the first communication device.
  • the identifiers of each AI model can be sent simultaneously or in batches without restriction.
  • the identifiers of each AI model can be carried in the same signaling or in different signaling without restriction.
  • a group identifier of a second group of AI models is received from the first communication device, and accordingly, in step 620, it is determined whether the second communication device deploys a second AI model based on the group identifier of the second group of AI models.
  • the group identifier is the identifier of the AI model group. As an example, based on the group identifier, the identifiers of each AI model in the AI model group can be obtained.
  • the group identifier and the identifiers of each AI model in the AI model group corresponding to the group identifier can exist in the form of a table, a function, or a string, such as storage or transmission, where an example of storage in table form is given in Table 2 below.
  • the group identifier of the AI model group is known, the identifiers of each AI model in the AI model group can be known.
  • the second group of AI models is AI model group #A
  • the group identifier AI model group #A is received from the first communication device in step 610
  • the identifier of AI model A1, the identifier of AI model A2, and the identifier of AI model A3 can be known based on the AI model group #A and the above Table 2, and then according to the identifier of AI model A1, the identifier of AI model A2, and the identifier of AI model A3, it can be determined whether the second communication device deploys an AI model that matches at least one of AI model A1, AI model A2, and AI model A3.
  • the second communication device deploys an AI model that matches at least one of AI model A1, AI model A2, and AI model A3, it is considered that the AI model matches all the AI models in the second group of AI models.
  • an AI model group may include a larger number of AI models.
  • the association relationship may also be an association relationship between the group identifiers of the AI model group.
  • the group identifier and the identifiers of each AI model in the AI model group corresponding to the group identifier may exist in the form of text, which is not limited to this.
  • the priorities of the AI models in the AI model group are different, and the ones with higher priorities are matched first.
  • the second group of AI models includes the first AI model and the third AI model, and the priority of the first AI model is higher than that of the third AI model.
  • the second communication device If it is determined based on the identifier of the first AI model that the second communication device does not deploy an AI model that matches the first AI model, it is then determined based on the identifier of the third AI model whether the second communication device deploys an AI model that matches the third AI model.
  • the priority of each AI model can be obtained in any of the following ways.
  • the first possible implementation method is to explicitly indicate the priority of each AI model.
  • the first communication device sends the priority of each AI model in the second group of AI models.
  • the first communication device sends the priority of each AI model in the second group of AI models to the second communication device, and the second communication device determines whether there is an AI model that matches each AI model in the second group of AI models in order of priority based on the priority of each AI model.
  • the priority and AI model identifier or group identifier can be carried in the same signaling or in different signaling without restriction.
  • the second possible implementation method is to implicitly indicate the priority of each AI model.
  • the AI models can be sorted in sequence according to their priorities, such as from high to low or from low to high.
  • the first communication device sends the identifiers of each AI model in the second group of AI models to the second communication device in sequence from high to low according to the priority of the AI model; the second communication device starts with the identifier of the first AI model received and determines in sequence whether there is an AI model that matches each AI model in the second group of AI models.
  • the above-mentioned sorting is carried out in sequence according to the priority of the AI model. For example, it can be sorted in sequence by the time of sending the identification of different AI models, or it can be sorted in sequence by different fields in the same signaling, without limitation.
  • the above two implementation methods are illustrative of the priority of each AI model in the AI model group, and are not limited to this.
  • the at least two AI models may also have different priorities
  • the indication method of the priority can refer to the indication method of the priority of each AI model in the aforementioned AI model group, which will not be repeated here.
  • method 600 when it is determined that the second communication device deploys the second AI model, method 600 also includes: sending an identifier of the second AI model to the first communication device, and the identifier of the second AI model is used to determine whether the first AI model in the first communication device matches the second AI model.
  • the identifier of the second AI model is sent to the first communication device, so that the second communication device or the third communication device, as well as the first communication device, all confirm the AI model matching, which can reduce the probability of AI model matching errors. For example, taking the example of the second communication device determining that the second communication device deploys the second AI model, if the second communication device and the first communication device have inconsistent understandings of the identifier of the AI model, or there is a conflict in the identifier of the AI model, etc., it may cause AI model matching errors. By confirming the AI model matching by both the second communication device and the first communication device, the probability of AI model matching errors can be reduced.
  • the first communication device and the third communication device are each configured with an identifier of an AI model
  • the first communication device may feedback the result of the match, that is, the second communication device determines the result of the match of the first communication device based on the response.
  • This implementation is similar to the implementation of the first communication device determining whether the second communication device deploys the second AI model based on the response, and will not be repeated here.
  • sending the identifier of the second AI model to the first communication device includes: sending the identifier of at least one AI model of the first group of AI models to the first communication device.
  • each AI model in the first group of AI models satisfies: when the input information of each AI model in the first group of AI models is the same, the output information is the same or the difference in the output information is within a preset range.
  • the identifiers of all AI models in the first group of AI models are sent to the first communication device. If an AI model matching at least one AI model in the first group of AI models is deployed in the first communication device, it is determined that each AI model in the first group of AI models matches the first AI model, that is, the second AI model matches the first AI model.
  • identifiers of some AI models in the first group of AI models are sent to the first communication device. If an AI model matching at least one AI model in the part of the AI models is deployed in the first communication device, it is determined that each AI model in the first group of AI models matches the first AI model, that is, the second AI model matches the first AI model.
  • a third possible implementation is to send a group identifier in the first group of AI models to the first communication device.
  • An AI model that matches at least one AI model in the first group of AI models determines that each AI model in the first group of AI models matches the first AI model, that is, the second AI model matches the first AI model.
  • the identifier of the AI model refers to an identifier that can identify some attributes of the AI model.
  • a global identifier refers to an identifier that can identify each attribute of the AI model, that is, based on the global identifier, each attribute of the AI model can be known.
  • a local identifier refers to an identifier that can identify a specific attribute of the AI model, that is, based on the local identifier, one or some attributes of the AI model can be known.
  • the global identifier of the AI model when the identifier of the AI model is transmitted between communication devices, the global identifier of the AI model can be transmitted, or a partial identifier of the AI model can be transmitted without limitation.
  • the identifier of the first AI model indicates at least one of the following: the type of the first communication device to which the first AI model belongs, the identifier of the first communication device to which the first AI model belongs, the manufacturer of the first communication device to which the first AI model belongs, the function of the first AI model, the scenario in which the first AI model is applicable, the data set of the first AI model, the communication parameters applicable to the first AI model, or the version of the first AI model.
  • the type identifier of the first communication device to which the first AI model belongs or simply the type identifier: used to indicate the type of the communication device that trains and/or uses the first AI model.
  • the type of the communication device that trains and/or uses the first AI model is a terminal device; for another example, the type of the communication device that trains and/or uses the first AI model is a network device.
  • the type identification is implemented by at least one bit. For example, assume that the type identification is indicated by 1 bit. If the bit is set to "1", it means that the communication device that trains and/or uses the first AI model is a terminal device; if the bit is set to "0", it means that the communication device that trains and/or uses the first AI model is a network device. It should be understood that the above is only an exemplary description and is not limiting.
  • An identifier of the first communication device to which the first AI model belongs or simply referred to as a device identifier: used to indicate the communication device that trains and/or uses the first AI model.
  • device identification can be achieved through any of the following: permanent equipment identifier (PEI), subscription permanent identifier (SUPI), global unique temporary identifier (global unique temporary identifier), cell radio network temporary identifier (C-RNTI), and physical layer cell identity (physical layer cell identity).
  • PEI permanent equipment identifier
  • SUPI subscription permanent identifier
  • C-RNTI cell radio network temporary identifier
  • physical layer cell identity physical layer cell identity
  • the manufacturer identifier of the first communication device to which the first AI model belongs or simply referred to as the manufacturer identifier: used to indicate the manufacturer that trained and/or used the first AI model.
  • the manufacturer identifier may be the manufacturer identifier of the terminal device.
  • the manufacturer identifier may be the manufacturer identifier of the network device.
  • the manufacturer identifier may be the manufacturer identifier of the chip.
  • the manufacturer identifier may be the operator identifier of the system.
  • Functional identifier of the first AI model or simply referred to as functional identifier: used to indicate the function of the first AI model.
  • the function of the first AI model can be understood as a problem that the first AI model can solve, or a task that the first AI model can perform.
  • the functions of the first AI model include at least one of the following: channel status information (CSI) feedback, CSI prediction, positioning, or beam management, etc.
  • CSI channel status information
  • the function identifier is implemented by at least one bit.
  • the function identifier is indicated by 2 bits. If the bit is set to "01”, it means that the function of the first AI model is CSI feedback; if the bit is set to "10”, it means that the function of the first AI model is CSI prediction; if the bit is set to "00”, it means that the function of the first AI model is positioning; if the bit is set to "11”, it means that the function of the first AI model is beam management. It should be understood that the above is only an exemplary description and is not limiting.
  • Scenario identifier for which the first AI model is applicable or simply referred to as scenario identifier: used to indicate the scenario for which the first AI model is applicable, such as indoor scene, outdoor scene, urban scene, or suburban scene.
  • the scene identification is implemented by at least one bit.
  • the bit is set to "00", it means that the scene applicable to the first AI model is an indoor scene; if the bit is set to "01”, it means that the scene applicable to the first AI model is an outdoor scene; if the bit is set to "10”, it means that the scene applicable to the first AI model is an urban scene; if the bit is set to "11”, it means that the scene applicable to the first AI model is a suburban scene.
  • Dataset identifier of the first AI model or simply referred to as dataset identifier: used to indicate which dataset the first AI model is trained based on.
  • dataset identifier can indicate the dataset used by a certain AI model.
  • different datasets can be identified by different identifiers.
  • the dataset identification is implemented by at least one bit. For example, assume that the dataset identification is indicated by 1 bit. If the bit is set to "0", it means that the first AI model is trained based on the first dataset; if the bit is set to "1", it means that the first AI model is trained based on the second dataset. It should be understood that the above is only an exemplary description and is not limiting.
  • the data set identifier can have a corresponding relationship with one or more of the scene identifier or the communication parameter identifier applicable to the first AI model.
  • the data set identifier can replace the scene identifier or one or more of the communication parameter identifiers applicable to the first AI model, and can also coexist with the scene identifier or one or more of the communication parameter identifiers applicable to the first AI model.
  • the correspondence between the data set identifier and the scene identifier can be predefined by the protocol, or, pre-stored, or, pre-configured.
  • Communication parameter identifier applicable to the first AI model or simply referred to as configuration identifier or communication parameter identifier: used to indicate the wireless parameter configuration applicable to the first AI model.
  • the wireless parameter configuration may include, for example, at least one of the following configurations: antenna configuration, bandwidth, rank, CSI feedback overhead, reference signal configuration, carrier frequency, or subcarrier spacing, etc.
  • Version identifier of the first AI model or simply referred to as version identifier: used to indicate the version of the first AI model.
  • the version of the AI model can be different versions of the same AI model, or can also be different AI models, without limitation.
  • the version identifier may indicate one or more of the complexity or performance of the AI model.
  • the version identifier may include a complexity identifier, or a performance identifier, or a complexity identifier and a performance identifier.
  • At least two AI models have the same function or scenario, but the complexity of the at least two AI models is different, or the at least two AI models are used to solve problems of different complexity, so they can be distinguished by version identification.
  • the version identification can also be called a complexity identification.
  • AI model #1 and AI model #2 have the same function or scenario, and the complexity of AI model #1 and AI model #2 is different.
  • AI model #1 has fewer layers, such as AI model #1 includes an input layer and an output layer, and AI model #1 is used to solve problems of lower complexity
  • AI model #2 has more layers, such as AI model #2 includes an input layer, a hidden layer, and an output layer, and AI model #2 is used to solve problems of higher complexity.
  • At least two AI models have the same function or scenario, but the performance of the at least two AI models is different, or when the input of the at least two AI models is the same, the gap between the output value and the ideal target value is different, so they can be distinguished by the version identifier.
  • the version identifier can also be called a performance identifier.
  • AI model #1 has fewer layers, such as AI model #1 includes an input layer, one hidden layer, and an output layer, and the performance of AI model #1 is lower, that is, the gap between the output value and the ideal target value is larger;
  • AI model #2 has more layers, such as AI model #2 includes an input layer, at least two hidden layers, and an output layer, and the performance of AI model #2 is higher, that is, the gap between the output value and the ideal target value is smaller.
  • the version identifier may have a corresponding relationship with the complexity identifier and the performance identifier, that is, when the version identifier is different, at least one of the complexity or performance of the corresponding AI model is different.
  • the corresponding relationship may be predefined, pre-stored, or pre-configured by the protocol.
  • the version identifier is implemented by at least one bit. For example, assume that the version identifier is indicated by 1 bit. If the bit is set to "0", it means that the first AI model is suitable for solving problems with higher complexity; if the bit is set to "1", it means that the first AI model is suitable for solving problems with lower complexity. It should be understood that the above is only an exemplary description and is not limiting.
  • the identification of the AI model has a validity period.
  • the identification of the AI model corresponds to a validity period, which can be implemented by a timer. When the timer times out, the identification of the AI model becomes invalid.
  • the identification of the AI model performs at least one of the operations of updating, activating, or re-registering, the validity period can be updated, such as restarting the timer.
  • the identification of the AI model becomes invalid, other devices can be notified that the identification of the AI model is invalid.
  • the method for obtaining the validity period of the identifier of the AI model is not limited. Take the first communication device and the second communication device as an example. In one possible implementation, when the first communication device sends the identifier of the first AI model to the second communication device, it sends the validity period of the identifier of the first AI model; when the second communication device determines that the identifier of the first AI model is in the validity period, it determines whether to deploy an AI model matching the first AI model based on the identifier of the first AI model.
  • the identifier of the first AI model and the validity period of the identifier of the first AI model can be carried in the same signaling or in different signaling, without limitation.
  • the second communication device itself knows the validity period of the identifier of the first AI model. In this way, after the second communication device receives the identifier of the first AI model, it determines that the identifier of the first AI model is in the validity period, and then determines whether to deploy an AI model matching the first AI model based on the identifier of the first AI model. It can be understood that the above two implementations are exemplary descriptions and are not limited to this. For example, it can also be another possible implementation, when the communication device sends the identifier of the AI model, it sends the identifier of a valid AI model, or the identifier of an AI model with a validity period greater than a certain threshold.
  • the following describes how to transmit the AI model identifier.
  • the first communication device sends the identifier of the first AI model
  • the following two possible implementation methods may be included.
  • the identifier of the first AI model includes at least one of the following: a type identifier of the first communication device to which the first AI model belongs, an identifier of the first communication device to which the first AI model belongs, a manufacturer identifier of the first communication device to which the first AI model belongs, a function identifier of the first AI model, an identifier of a scenario to which the first AI model is applicable, a data set identifier of the first AI model, a communication parameter identifier applicable to the first AI model, or a version identifier of the first AI model.
  • the first communication device can send at least one of the above identifications, so that the at least one of the above identifications can be directly obtained through the identification of the first AI model. It can be understood that the above identifications can be carried in the same signaling or in different signalings without limitation.
  • the identifier of the first AI model is certain information, based on which at least one of the following items can be obtained: a type identifier of the first communication device to which the first AI model belongs, an identifier of the first communication device to which the first AI model belongs, a manufacturer identifier of the first communication device to which the first AI model belongs, a function identifier of the first AI model, an identifier of a scenario applicable to the first AI model, a dataset identifier of the first AI model, an identifier of a communication parameter applicable to the first AI model, or a version identifier of the first AI model. That is, the above information has a corresponding relationship with at least one of the above identifiers.
  • the first communication device may send a message, through which the at least one identification mentioned above may be indirectly obtained.
  • the identifier of the AI model is predefined and/or determined by at least one device. Three possible implementations are described below.
  • the identification of the AI model is predefined, such as a standard predefined.
  • the number of fields in the identifier of the AI model, the order of the fields, and the length of each field can all be predefined. Based on this, each communication device has a consistent understanding of the meaning of each field in the AI model.
  • Figure 8 is a schematic diagram of the identification of an AI model applicable to an embodiment of the present application.
  • the identification of the AI model includes 8 fields, and the 8 fields carry in sequence: type identification, device identification, manufacturer identification, function identification, scenario identification, data set identification, communication parameter identification, and version identification. If the identification of the AI model is predefined, the number of fields, the order of the fields, and the length of each field in the identification of different AI models are the same, the difference is that the values of each field may be different.
  • the identification of the AI model is determined by at least one device.
  • the number of fields in the AI model identifier, the order of the fields, and the length of each field can be customized.
  • the number of fields in the identifiers of different AI models may be different, or the order of the fields may be different, which can be configured according to actual conditions.
  • each communication device understands the meaning of each field in the AI model determined by itself.
  • Figure 9 is another schematic diagram of the identification of the AI model applicable to the embodiment of the present application.
  • the identification of the AI model includes 4 fields, and the 4 fields carry in sequence: manufacturer identification, function identification, scenario identification, and version identification.
  • the identification of the AI model includes 5 fields, and the 5 fields carry in sequence: device identification, function identification, scenario identification, manufacturer identification, and version identification.
  • the first AI model may be determined by the first communication device itself, or may be determined by other communication devices. Three examples are given below.
  • the communication device that deploys the AI model determines the identifier of the AI model.
  • the identifier of the first AI model may be determined by the first communication device itself.
  • the communication device that trains the AI model determines the identifier of the AI model.
  • the identifier of the first AI model may be determined by the first communication device; or, if the first AI model is trained by the first communication device and the second communication device, the identifier of the first AI model may be determined by negotiation between the first communication device and the second communication device.
  • a third party determines the identifier of the AI model.
  • the third party may be, for example, a core network, or an access network device, or an operation administration and maintenance (OAM) system, or a central management node, or a model management node, or an AI node, which can communicate with each communication device.
  • OAM operation administration and maintenance
  • the valid range of the identifier of the AI model may be considered, and the valid range may be within a cell, or within multiple cells, or within the same network (such as a public land mobile network (PLMN)).
  • PLMN public land mobile network
  • the communication device uses the identifier of the AI model in the cell.
  • the same AI model may have different AI model identifiers in different cells.
  • some identifiers of the AI model are predefined, and some identifiers are determined by at least one device.
  • the order and length of some fields in the identification of the AI model can be predefined, and the number, order, and length of the remaining fields can be configured by at least one device based on actual conditions. Based on this, each communication device has a consistent understanding of the meaning of some fields in the AI model.
  • Figure 10 is another schematic diagram of the identification of an AI model applicable to an embodiment of the present application. Assume that the order and length of the first three fields in the identification of the AI model are predefined, and the three fields carry in sequence: type identification, device identification, and manufacturer identification. As shown in (1) in Figure 10, in addition to the above three fields, the identification of the AI model also includes 1 field, and 1 field carries the function identification. As shown in (2) in Figure 10, in addition to the above three fields, the identification of the AI model also includes 2 fields, and the 2 fields carry: function identification and version identification.
  • the format of the AI model identifier is predefined, and the content of the AI model is determined by the communication device.
  • the communication device may determine the identifier of the AI model by itself and register it with other devices; in another implementation, the communication device may request the identifier of the AI model from other communication devices, and the identifier of the AI model is assigned by the other communication devices.
  • the communication device may be the communication device to which the AI model belongs, or another device different from the communication device to which the AI model belongs. In the case where the communication device is another device different from the communication device to which the AI model belongs, the communication device may notify the communication device to which the AI model belongs of the identifier of the AI model.
  • the following takes the first AI model in the first communication device as an example, and introduces the two implementations in combination with Figures 11 and 12. It can be understood that the two implementations can be used in combination with other parts related to the identification of the aforementioned AI model, such as the information indicated by the identification of the AI model, and/or one or more embodiments in the determination method, and can also be decoupled from one or more embodiments in other parts related to the identification of the aforementioned AI model.
  • the first communication device itself determines the identifier of the first AI model and registers it with other devices (such as a third party, a second communication device, or a third communication device).
  • Figure 11 is a schematic diagram of registering the identifier of the AI model applicable to an embodiment of the present application. As shown in Figure 11, after the first communication device determines the identifier of the first AI model, it sends registration information to other devices, and the registration information includes the identifier of the first AI model.
  • the other devices may send a response to the registration information to the first communication device, and the response to the registration information is used to notify whether the registration is successful.
  • the other devices may not send a response to the registration information to the first communication device.
  • the starting time of the period of time may be the time when the first communication device sends the registration information, and the duration of the period of time may be predefined, or it may be estimated based on historical circumstances and is not limited.
  • the period of time may be implemented by a timer.
  • one possible implementation method is that the other device may send a response to the registration information to the first communication device, and the response to the registration information instructs the first communication device to change the version identifier of the model; or, another possible implementation method is that the other device may reallocate a version identifier for the first AI model and send the reallocated version identifier to the first communication device.
  • the registration information may include other information in addition to the identifier of the first AI model.
  • the registration information includes at least one of the following information: the type of the first communication device to which the first AI model belongs, the first communication device to which the first AI model belongs, the manufacturer of the first communication device to which the first AI model belongs, the function of the first AI model, the scenario to which the first AI model is applicable, the data set of the first AI model, the communication parameters applicable to the first AI model, or the version of the first AI model.
  • the registration information includes the version identifier of the first AI model and the version information of the first AI model, based on which the first communication device and other devices can align their understanding of the version identifier and obtain the version information referred to by the version identifier.
  • the identifier of the AI model is used to indicate the manufacturer of the AI model and the function of the AI model. Assuming that there are 8 manufacturers of the AI model, if the manufacturer identifier of the AI model is to be carried, 3 bits are required; assuming that there are 4 functions of the AI model, if the function identifier of the AI model is to be carried, 2 bits are required.
  • the first communication device carries the manufacturer identifier of the AI model and the function identifier of the AI model in the registration information, 5 bits are required. Taking into account that not all manufacturers and functions may appear within the effective range of the AI model, therefore, when registering the identifier of the AI model, an identifier of less than 5 bits can be used to jointly represent the manufacturer of the AI model and the function of the AI model, for example, 3 bits are used to represent the manufacturer of the AI model and the function of the AI model.
  • which identifier is assigned to each specific combination of manufacturer and function can be determined by the entity that assigns the AI model (such as the first communication device, or other communication devices other than the first communication device, which trains the AI model and then deploys the AI model in the first communication device), and there is no restriction on this.
  • the entity that assigns the AI model such as the first communication device, or other communication devices other than the first communication device, which trains the AI model and then deploys the AI model in the first communication device
  • other devices determine the identifier of the first AI model and notify the first communication device of the identifier of the first AI model.
  • Figure 12 is a schematic diagram of a notification of an identification of an AI model applicable to an embodiment of the present application. As shown in Figure 12, after the other devices determine the identification of the first AI model, they send the identification of the first AI model to the first communication device. Optionally, before the other devices determine the identification of the first AI model, the first communication device sends a request message to the other devices, and the request message is used to request the other devices to assign the identification of the first AI model to the first AI model.
  • the request message includes at least one of the following information: the type of the first communication device to which the first AI model belongs, the first communication device to which the first AI model belongs, the manufacturer of the first communication device to which the first AI model belongs, the function of the first AI model, the scenario to which the first AI model is applicable, the data set of the first AI model, the communication parameters applicable to the first AI model, or the version of the first AI model.
  • the other devices assign a data set identifier of the first AI model to the first AI model.
  • the data sets used to train the AI model include two types: a first data set and a second data set, if the data set of the first AI model included in the request information is the first data set, the data set identifier assigned to the first AI model by the other devices is "0"; if the data set of the first AI model included in the request information is the second data set, the data set identifier assigned to the first AI model by the other devices is "1".
  • Other identifiers are similar and will not be elaborated here.
  • the identifier of the AI model assigned by other devices to the first communication device may not have a specific identifier for each type of information, such as a data set identifier, a function identifier, etc., but an identifier of the AI model assigned after comprehensively considering at least one of the above information.
  • the identifier of the AI model can distinguish different at least one of the above information, but there is no explicit field for each type of information.
  • the identifier of the AI model has a corresponding relationship with the aforementioned at least one information.
  • the following takes the interaction between the first communication device and the second communication device as an example, and describes the process applicable to the embodiment of the present application in conjunction with Figures 13 to 15.
  • the content not described in detail below please refer to the description in method 600.
  • the AI model deployed in the communication device is referred to as the AI model in the communication device.
  • FIG13 is a schematic diagram of a communication method 1300 provided according to an embodiment of the present application.
  • the method 1300 may be applicable to a scenario in which a first communication device sends an identifier of at least one AI model to a second communication device, and the second communication device performs AI model matching.
  • the method 1300 shown in FIG13 may include the following steps.
  • the first communication device sends identifiers of T AI models to the second communication device.
  • T is an integer greater than 1 or equal to 1.
  • the T AI models may be AI models available in the first communication device, or may be some AI models among the AI models available in the first communication device.
  • the second communication device determines, based on the identifiers of the T AI models, whether the second communication device has an AI model that matches the T AI models.
  • the second communication device After receiving the identifications of the T AI models, the second communication device determines whether there is a matching AI model in the second communication device. AI model.
  • the second communication device determines whether the second communication device has an AI model that matches the T AI models based on the identifiers and association relationships of the T AI models.
  • the association relationship represents the relationship between the identifiers of the mutually matching AI models.
  • the second communication device itself maintains the association relationship.
  • the first communication device maintains the association relationship, and the first communication device sends the association relationship to the second communication device.
  • other devices maintain the association relationship, and after the second communication device receives the identifier of the first communication device, it requests the association relationship from other devices.
  • the association relationship please refer to the relevant description in method 600, which will not be repeated here.
  • the first communication device determines whether there is an AI model in the second communication device that matches the T AI models.
  • the second communication device does not have an AI model that matches the T AI models.
  • the first communication device may learn that there is no AI model in the second communication device that matches the T AI models in any of the following ways.
  • the second communication device if the second communication device does not have an AI model that matches the T AI models, the second communication device sends feedback information to the first communication device, where the feedback information indicates that the second communication device does not have an AI model that can match the T AI models.
  • the feedback information may be a negative response.
  • the second communication device if the second communication device does not have an AI model that matches the T AI models, the second communication device does not send feedback information to the first communication device.
  • the feedback information may be a positive response.
  • T is equal to 1
  • the second communication device may send feedback information to the first communication device, the feedback information indicating that the second communication device has an AI model that can match the T AI models.
  • the feedback information may be a positive response.
  • the second communication device does not send feedback information to the first communication device.
  • the feedback information may be a negative response.
  • T is greater than 1, and the second communication device has AI models that match some of the T AI models (eg, recorded as T1 AI models).
  • the second communication device may send feedback information to the first communication device, where the feedback information includes information about T1 AI models (such as identifiers of T1 AI models), and the feedback information indicates that the second communication device has an AI model that matches T1 AI models.
  • the first communication device learns that the second communication device has an AI model that matches T1 AI models based on the feedback information from the second communication device.
  • the second communication device may send feedback information to the first communication device, where the feedback information includes information of T2 AI models (such as identifiers of T2 AI models), and the feedback information indicates that the second communication device does not have an AI model that matches T2 AI models, wherein T2 AI models are AI models other than T1 AI models among the T AI models.
  • the first communication device learns that the second communication device does not have an AI model that matches T2 AI models based on the feedback information from the second communication device, and assumes that the second communication device has an AI model that matches the AI models other than T2 AI models among the T AI models (i.e., T1 AI models).
  • Case 3 T is greater than 1, and the second communication device has an AI model that matches all of the T AI models.
  • the second communication device may send feedback information to the first communication device, the feedback information indicating that the second communication device has an AI model that matches the T AI models.
  • the feedback information may be a positive response.
  • Another possible implementation is that if T is greater than 1, and there are AI models in the second communication device that match all of the T AI models. AI model, the second communication device does not send feedback information to the first communication device.
  • the feedback information may be a negative response.
  • Another possible implementation method is that if T is greater than 1, and the second communication device has an AI model that matches all of the T AI models, the second communication device can send feedback information to the first communication device, where the feedback information includes information of the T AI models (such as identifiers of the T AI models), and the feedback information indicates that the second communication device has an AI model that matches the T AI models.
  • FIG14 is a schematic diagram of a communication method 1400 provided according to another embodiment of the present application.
  • the method 1400 may be applicable to a scenario in which a first communication device sends an identifier of at least one AI model to a second communication device, and the second communication device and the first communication device perform AI model matching.
  • the method 1400 shown in FIG14 may include the following steps.
  • the first communication device sends identifiers of T AI models to the second communication device.
  • Step 1410 is similar to step 1310 and is not described in detail here.
  • the second communication device determines, based on the identifiers of the T AI models, that the AI models in the second communication device that match the T AI models are N AI models.
  • the second communication device After receiving the identifiers of the T AI models, the second communication device determines whether there is an AI model matching the T AI models in the second communication device. Assume that there is an AI model matching the T AI models in the second communication device, and the AI models matching the T AI models are N AI models.
  • N is an integer greater than 1 or equal to 1.
  • the N AI models may be AI models available in the second communication device and matching with an AI model among the T AI models, that is, there is an AI model among the T AI models matching with the N AI models.
  • Step 1420 is similar to step 1320 and will not be described in detail here.
  • the second communication device sends identifiers of N AI models to the first communication device.
  • the first communication device determines whether there is an AI model in the first communication device that matches the N AI models.
  • method 1400 also includes step 1450 .
  • the second communication device determines a matching result of the first communication device.
  • the second communication device may determine whether there is an AI model in the first communication device that matches the N AI models based on the response of the first communication device, that is, determine the result of the first communication device matching the N AI models.
  • both the second communication device and the first communication device confirm the AI model matching, which can reduce the probability of AI model matching errors.
  • the relevant description in the above method 600 please refer to the relevant description in the above method 600.
  • the first communication device determines that there is no AI model in the first communication device that matches the N AI models.
  • the first communication device if the first communication device determines that there is no AI model in the first communication device that matches the N AI models, the first communication device sends feedback information to the second communication device, where the feedback information indicates that there is no AI model in the first communication device that matches the N AI models.
  • the feedback information may be a negative response.
  • the first communication device determines that there is no AI model in the first communication device that matches the N AI models, the first communication device does not send feedback information to the second communication device.
  • the feedback information may be a positive response.
  • the above two implementation methods can refer to the relevant description in method 600 about the first communication device determining whether the second communication device deploys the second AI model based on the response, which will not be repeated here.
  • the first communication device determines that there is an AI model in the first communication device that matches the N AI models.
  • the following describes the three scenarios respectively.
  • N is equal to 1
  • the first communication device determines that there is an AI model in the first communication device that matches the N AI models.
  • the first communication device may send feedback information to the second communication device, where the feedback information indicates that there is an AI model in the first communication device that matches the N AI models.
  • the feedback information may be a positive response.
  • the first communication device determines that there is an AI model in the first communication device that matches the N AI models, the first communication device does not send feedback information to the first communication device.
  • the feedback information may be a negative response.
  • Case 2 N is greater than 1, and the first communication device determines that there are AI models in the first communication device that match some of the N AI models (eg, recorded as N1 AI models).
  • the first communication device may send feedback information to the second communication device, where the feedback information includes information of the N1 AI models (such as identifiers of the N1 AI models), and the feedback information indicates that there is an AI model in the first communication device that matches the N1 AI models.
  • Another possible implementation method is that if N is greater than 1, and the first communication device determines that there is an AI model in the first communication device that matches N1 AI models, the first communication device may send feedback information to the second communication device, where the feedback information includes information of N2 AI models (such as identifiers of N2 AI models), and the feedback information indicates that there is no AI model in the first communication device that matches the N2 AI models, wherein the N2 AI models are AI models other than the N1 AI model in the N AI models.
  • N2 AI models such as identifiers of N2 AI models
  • Case 3 N is greater than 1, and the first communication device determines that there is an AI model in the first communication device that matches all of the N AI models.
  • the first communication device may send feedback information to the second communication device, where the feedback information indicates that the first communication device has AI models that match all of the N AI models.
  • the feedback information may be a positive response.
  • the first communication device determines that the first communication device has all the AI models in the N AI models, the first communication device does not send feedback information to the second communication device.
  • the feedback information may be a negative response.
  • the first communication device may send feedback information to the second communication device, where the feedback information includes information of the N AI models (such as identifiers of the N AI models), and the feedback information indicates that the first communication device has an AI model that matches all the N AI models.
  • FIG15 is a schematic diagram of a communication method 1500 provided according to another embodiment of the present application.
  • the method 1500 may be applicable to a scenario in which a first communication device sends an identifier of at least one AI model and/or an identifier of an AI model in the same group as the at least one AI model to a second communication device, and the second communication device performs AI model matching.
  • the method 1500 shown in FIG15 may include the following steps.
  • the first communication device sends the identifiers of T AI models and/or the identifier of at least one AI model in the same group as the T AI models to the second communication device.
  • the first communication device may send the identifiers of the T AI models and/or the identifier of at least one AI model in the same group as the T AI models to the second communication device.
  • the relevant scheme of AI model grouping refer to the relevant description in method 600, which will not be repeated here.
  • the second communication device determines whether there is an AI model in the second communication device that matches at least one of the following AI models: T AI models, and at least one AI model in the same group as the T AI models.
  • the second communication device determines whether there is an AI model in the second communication device that matches at least one of the following AI models: T AI models, at least one AI model in the same group as the T AI models, based on the identifiers of the T AI models and/or the identifier of at least one AI model in the same group as the T AI models.
  • Step 1520 is similar to step 1320, except that in step 1520, the second communication device may determine whether there is an AI model in the second communication device that matches at least one of the following AI models: T AI models, and at least one AI model in the same group as the T AI models.
  • T AI models at least one of the following AI models
  • AI model in the same group as the T AI models For the scheme of grouping AI models, please refer to the relevant description in method 600, which will not be repeated here.
  • the priorities of the various AI models are different.
  • the second communication device may give priority to whether there is an AI model that matches an AI model with a high priority.
  • priority of the AI model reference may be made to the relevant description in method 600, which will not be repeated here.
  • the first communication device determines whether the second communication device has an AI model that matches at least one of the following AI models: T AI models, and at least one AI model in the same group as the T AI models.
  • step 1530 reference may be made to the relevant description in step 1330 and will not be repeated here.
  • FIG. 16 is a schematic diagram of a communication method 1600 provided according to another embodiment of the present application.
  • the method 1600 may be applicable to a first communication
  • the device sends the identification of at least one AI model and/or the identification of an AI model in the same group as the at least one AI model to the second communication device, and the second communication device and the first communication device perform a scenario of AI model matching.
  • the method 1600 shown in Figure 16 may include the following steps.
  • the first communication device sends the identifiers of T AI models and/or the identifier of at least one AI model in the same group as the T AI models to the second communication device.
  • the first communication device may send the identifiers of T AI models and/or the identifier of at least one AI model in the same group as the T AI models to the second communication device.
  • step 1610 reference may be made to the relevant description in step 1510 and will not be repeated here.
  • the second communication device determines that the AI models in the second communication device that match at least one of the following AI models are N AI models: T AI models and at least one AI model in the same group as the T AI models.
  • the second communication device determines that the AI models in the second communication device that match at least one of the following AI models are N AI models based on the identifiers of T AI models and/or the identifier of at least one AI model in the same group as the T AI models: T AI models and at least one AI model in the same group as the T AI models.
  • the N AI models are AI models that match some or all of the T AI models.
  • the first communication device sends the identifier of at least one AI model in the same group as the T AI models to the second communication device, then the N AI models are AI models that match some or all of the at least one AI model.
  • the second communication device sends the identifiers of the N AI models and/or the identifier of at least one AI model in the same group as the N AI models to the first communication device.
  • step 1630 reference may be made to the relevant description in step 1510 and will not be repeated here.
  • the first communication device determines whether there is an AI model in the first communication device that matches at least one of the following AI models: N AI models, and at least one AI model in the same group as the N AI models.
  • step 1640 reference may be made to the relevant description in step 1520 and will not be repeated here.
  • method 1600 also includes step 1650 .
  • the second communication device determines a matching result of the first communication device.
  • Step 1650 is similar to step 1450 and will not be described in detail here.
  • the methods and operations implemented by the encoder can also be implemented by components that can be implemented by the encoder (such as a chip or circuit); in addition, the methods and operations implemented by the decoder can also be implemented by components that can be implemented by the decoder (such as a chip or circuit), without limitation.
  • FIG. 17 is a schematic block diagram of a communication device 1700 provided in an embodiment of the present application.
  • the device 1700 includes a transceiver unit 1710 and a processing unit 1720.
  • the transceiver unit 1710 can be used to implement corresponding communication functions.
  • the transceiver unit 1710 can also be referred to as a communication interface or a communication unit.
  • the processing unit 1720 can be used to perform processing, such as determining whether an AI model matches.
  • the device 1700 may further include a storage unit, which may be used to store instructions and/or data, and the processing unit 1720 may read the instructions and/or data in the storage unit so that the device implements the aforementioned method embodiment.
  • a storage unit which may be used to store instructions and/or data
  • the processing unit 1720 may read the instructions and/or data in the storage unit so that the device implements the aforementioned method embodiment.
  • the device 1700 is used to execute the steps or processes performed by the communication device in the above method embodiment
  • the transceiver unit 1710 is used to execute the transceiver related operations on the communication device side in the above method embodiment
  • the processing unit 1720 is used to execute the processing related operations on the communication device side in the above method embodiment.
  • the device 1700 is used to execute the steps or processes executed by the second communication device or the third communication device in the embodiment shown in FIG6 , or the steps or processes executed by the second communication device in the embodiments shown in FIG13 to FIG16 .
  • the transceiver unit 1710 is used to receive the identifier of the first artificial intelligence AI model from the first communication device; the processing unit 1720 is used to determine whether the second communication device deploys the second AI model based on the identifier of the first AI model, and the second AI model is matched with the first AI model. AI model.
  • the device 1700 is used to execute the steps or processes executed by the first communication device in the embodiment shown in FIG6, or the steps or processes executed by the first communication device in the embodiments shown in FIG13 to FIG16.
  • the transceiver unit 1710 is used to send an identifier of a first artificial intelligence AI model, and the identifier of the first AI model is used to determine whether the second communication device deploys a second AI model that matches the first AI model; the processing unit 1720 is used to determine whether the second communication device deploys the second AI model based on the response.
  • the device 1700 is used to execute the steps or processes executed by the first communication device in the embodiment shown in Figure 11.
  • the processing unit 1720 is used to determine the identifier of the first AI model; the transceiver unit 1710 is used to send registration information to other devices, and the registration information includes the identifier of the first AI model.
  • the device 1700 is used to execute the steps or processes executed by the first communication device in the embodiment shown in Figure 12.
  • the transceiver unit 1710 is used to receive the identifier of the first AI model from other devices.
  • the device 1700 is used to execute the steps or processes executed by other devices in the embodiment shown in Figure 11.
  • the transceiver unit 1710 is used to receive registration information from the first communication device.
  • the device 1700 is used to execute the steps or processes executed by other devices in the embodiment shown in Figure 12.
  • the processing unit 1720 is used to determine the identifier of the first AI model; the transceiver unit 1710 is used to send the identifier of the first AI model to the first communication device.
  • the device 1700 here is embodied in the form of a functional unit.
  • the term "unit” here may refer to an application specific integrated circuit (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 combined logic circuit and/or other suitable components that support the described functions.
  • ASIC application specific integrated circuit
  • the device 1700 can be specifically a communication device in the above-mentioned embodiments (such as a first communication device, a second communication device, and a third communication device), and can be used to execute the various processes and/or steps corresponding to the communication device in the above-mentioned method embodiments. To avoid repetition, it will not be repeated here.
  • the device 1700 of each of the above-mentioned schemes has the function of implementing the corresponding steps performed by the communication device (such as the first communication device, the second communication device, the third communication device, or other devices) in the above-mentioned method.
  • the function can be implemented by hardware, or the corresponding software can be implemented by hardware.
  • 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 (for example, the sending 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 to respectively perform the sending and receiving operations and related processing operations in each method embodiment.
  • the transceiver unit can be replaced by a transceiver (for example, the sending 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 to respectively perform the sending and receiving operations and related processing operations in each method embodiment.
  • transceiver unit 1710 can also be a transceiver circuit (for example, can include a receiving circuit and a sending circuit), and the processing unit can be a processing circuit.
  • the device in FIG. 17 may be a device in the aforementioned embodiment, or may be a chip or a chip system, such as a system on chip (SoC).
  • the transceiver unit may be an input/output circuit or a communication interface; the processing unit may be a processor or a microprocessor or an integrated circuit integrated on the chip. This is not limited here.
  • the device 1800 includes a processor 1810, the processor 1810 is coupled to a memory 1820, the memory 1820 is used to store computer programs or instructions and/or data, and the processor 1810 is used to execute the computer program or instructions stored in the memory 1820, or read the data stored in the memory 1820, so as to execute the methods in the above method embodiments.
  • processors 1810 there are one or more processors 1810 .
  • the memory 1820 is one or more.
  • the memory 1820 is integrated with the processor 1810 or provided separately.
  • the device 1800 further includes a transceiver 1830, and the transceiver 1830 is used for receiving and/or sending signals.
  • the processor 1810 is used to control the transceiver 1830 to receive and/or send signals.
  • the processor 1810 may have the function of the processing unit 1720 shown in FIG. 17
  • the memory 1820 may have the function of a storage unit
  • the transceiver 1830 may have the function of the transceiver unit 1710 shown in FIG. 17 .
  • the device 1800 is used to implement the operations performed by a communication device (such as a first communication device, a second communication device, a third communication device, or other devices, etc.) in the above method embodiments.
  • a communication device such as a first communication device, a second communication device, a third communication device, or other devices, etc.
  • the processor 1810 is used to execute the computer program or instructions stored in the memory 1820 to implement the above various method embodiments.
  • Related operations of a communication device such as a first communication device, a second communication device, a third communication device, or other devices, etc.).
  • processors mentioned in the embodiments of the present application may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
  • the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.
  • the memory mentioned in the embodiments of the present application may be a volatile memory and/or a non-volatile memory.
  • the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory.
  • the volatile memory may be a random access memory (RAM).
  • a RAM may be used as an external cache.
  • RAM includes the following forms: static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).
  • SRAM static RAM
  • DRAM dynamic RAM
  • SDRAM synchronous DRAM
  • DDR SDRAM double data rate SDRAM
  • ESDRAM enhanced SDRAM
  • SLDRAM synchronous link DRAM
  • DR RAM direct rambus RAM
  • the processor is a general-purpose 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 into the processor.
  • memory described herein is intended to include, but is not limited to, these and any other suitable types of memory.
  • FIG19 is a schematic diagram of a chip system 1900 provided in an embodiment of the present application.
  • the chip system 1900 (or also referred to as a processing system) includes a logic circuit 1910 and an input/output interface 1920.
  • the logic circuit 1910 can be a processing circuit in the chip system 1900.
  • the logic circuit 1910 can be coupled to the storage unit and call the instructions in the storage unit so that the chip system 1900 can implement the methods and functions of each embodiment of the present application.
  • the input/output interface 1920 can be an input/output circuit in the chip system 1900, outputting information processed by the chip system 1900, or inputting data or signaling information to be processed into the chip system 1900 for processing.
  • the logic circuit 1910 is coupled to the input/output interface 1920, and the logic circuit 1910 can send the identifier of the first AI model to the decoder through the input/output interface 1920; or the input/output interface 1920 can input the identifier from the second AI model to the logic circuit 1910 for processing.
  • the logic circuit 1910 is coupled to the input/output interface 1920, and the input/output interface 1920 can input the identifier of the first AI model from the first communication device to the logic circuit 1910 for processing.
  • the chip system 1900 is used to implement the operations performed by a communication device (such as a first communication device, a second communication device, a third communication device, or other devices, etc.) in the above method embodiments.
  • a communication device such as a first communication device, a second communication device, a third communication device, or other devices, etc.
  • the logic circuit 1910 is used to implement the processing-related operations performed by the communication device (such as the first communication device, the second communication device, or the third communication device) in the above method embodiments;
  • the input/output interface 1920 is used to implement the sending and/or receiving-related operations performed by the communication device (such as the first communication device, the second communication device, the third communication device, or other devices, etc.) in the above method embodiments.
  • An embodiment of the present application also provides a computer-readable storage medium on which computer instructions are stored for implementing the methods executed by a communication device (such as a first communication device, a second communication device, a third communication device, or other devices, etc.) in the above-mentioned method embodiments.
  • a communication device such as a first communication device, a second communication device, a third communication device, or other devices, etc.
  • the computer when the computer program is executed by a computer, the computer can implement the method performed by a communication device (such as a first communication device, a second communication device, a third communication device, or other devices, etc.) in each embodiment of the above method.
  • a communication device such as a first communication device, a second communication device, a third communication device, or other devices, etc.
  • An embodiment of the present application also provides a computer program product, comprising instructions, which, when executed by a computer, implement the methods performed by a communication device (such as a first communication device, a second communication device, a third communication device, or other devices, etc.) in the above-mentioned method embodiments.
  • a communication device such as a first communication device, a second communication device, a third communication device, or other devices, etc.
  • the present application also provides a communication system, which includes the first communication device and the second communication device in the above embodiments.
  • the system includes the first communication device and the second communication device in the embodiment shown in FIG. 13 to 16 show the first communication device and the second communication device in the embodiments.
  • the present application also provides a communication system, which includes the first communication device and the third communication device in the above embodiments.
  • the system includes the first communication device and the third communication device in the embodiment shown in FIG6 .
  • the present application also provides a communication system, which includes the first communication device and other devices in the above embodiments.
  • the system includes the first communication device and other devices in the embodiments shown in Figures 10 and 11.
  • the disclosed devices and methods can be implemented in other ways.
  • the device embodiments described above are only schematic.
  • the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.
  • the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
  • the computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
  • the computer can be a personal computer, a server, or a network device, etc.
  • 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 a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center.
  • the computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media integrations.
  • the available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid state disk (SSD)).
  • the aforementioned available medium includes, but is not limited to, various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Software Systems (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本申请实施例提供了一种模型匹配的方法和通信装置。该方法包括：接收来自第一通信装置的第一人工智能AI模型的标识；基于第一AI模型的标识，确定第二通信装置是否部署第二AI模型，第二AI模型为与第一AI模型匹配的AI模型。这样，可以通过交互AI模型的标识，实现基于AI模型的标识确定与该AI模型匹配的AI模型。

Description

模型匹配的方法和通信装置
本申请要求于2022年09月29日提交中国专利局、申请号为202211197193.5、申请名称为“模型匹配的方法和通信装置”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及通信技术领域，并且更具体地，涉及一种模型匹配的方法和通信装置。
背景技术
为了应对未来智能普惠的愿景，智能化将在无线网络架构层面进一步演进，人工智能(artificial intelligence，AI)将与无线网络进一步深度的融合。以双端模型为例，构成双端模型的多个子AI模型，可以分别部署在不同的通信装置。部署在不同的通信装置中的子模型需要进行匹配，以便采用不同的通信装置中部署的子AI模型联合执行一些通信任务。那么，部署在不同的通信装置中的子模型如何匹配，是一个待解决的问题。
发明内容
本申请提供一种模型匹配的方法和通信装置，以期可以实现部署在不同的通信装置中的AI模型进行匹配。
第一方面，提供了一种模型匹配的方法，该方法可以由通信装置执行，或者，也可以由用于通信装置的芯片或电路执行，该通信装置可以为第二通信装置，或，不同于第二通信装置的第三通信装置，本申请对此不作限定。
该方法可以包括：接收来自第一通信装置的第一人工智能AI模型的标识；基于所述第一AI模型的标识，确定第二通信装置是否部署第二AI模型，所述第二AI模型为与所述第一AI模型匹配的AI模型。
基于上述技术方案，通信装置之间可通过交互AI模型的标识，进行AI模型的匹配。举例来说，第一通信装置向第二通信装置发送第一AI模型的标识，第二通信装置可以基于该第一AI模型的标识识别第一AI模型，进而可确定自身是否部署与该第一AI模型匹配的AI模型。这样，不仅可以实现AI模型匹配，还可以因为传输AI模型的标识而不是AI模型，从而保护AI模型的隐私，降低传输AI模型带来的信令开销。
结合第一方面，在第一方面的某些实现方式中，所述基于所述第一AI模型的标识，确定第二通信装置是否部署第二AI模型，包括：基于所述第一AI模型的标识和关联关系，确定所述第二通信装置是否部署所述第二AI模型，所述关联关系表示相互匹配的AI模型的标识之间的关系。
基于上述技术方案，相互匹配的AI模型的标识具有关联关系，这样在基于第一AI模型的标识确定是否部署与该第一AI模型匹配的第二AI模型时，可以基于关联关系和第一AI模型的标识，共同判断第二通信装置中是否部署与第一AI模型匹配的第二AI模型。例如，可以先基于关联关系判断与第一AI模型匹配的AI模型为第二AI模型，进而再判断第二通信装置中是否部署该第二AI模型。再例如，可以先判断第二通信装置中部署的AI模型有哪些，然后再基于关联关系判断第二通信装置中部署的AI是否有与第一AI模型匹配的AI模型。
结合第一方面，在第一方面的某些实现方式中，所述方法还包括：在确定所述第二通信装置未部署所述第二AI模型的情况下，向所述第一通信装置发送第一反馈信息，所述第一反馈信息用于反馈所述第二通信装置未部署所述第二AI模型；或者，在确定所述第二通信装置部署所述第二AI模型的情况下，向所述第一通信装置发送第二反馈信息，所述第二反馈信息用于反馈所述第二通信装置部署所述第二AI模型。
结合第一方面，在第一方面的某些实现方式中，在确定所述第二通信装置部署所述第二AI模型的 情况下，所述方法还包括：向所述第一通信装置发送所述第二AI模型的标识，所述第二AI模型的标识用于所述第一通信装置中所述第一AI模型是否与所述第二AI模型匹配的判断。
基于上述技术方案，若确定第二通信装置部署与第一AI模型匹配的第二AI模型，则可以向第一通信装置发送该第二AI模型的标识，以便第一通信装置可以再次执行第一AI模型与第二AI模型是否匹配的判断。这样，相当于至少两个通信装置执行了AI模型匹配的操作，可以降低AI模型匹配错误发生的概率。
结合第一方面，在第一方面的某些实现方式中，所述第二AI模型属于第一组AI模型，所述向所述第一通信装置发送所述第二AI模型的标识，包括：向所述第一通信装置发送所述第一组AI模型中至少一个AI模型的标识。
一示例，所述向所述第一通信装置发送所述第二AI模型的标识，包括：向所述第一通信装置发送所述第二AI模型的标识和所述第一组AI模型中除所述第二AI模型以外的至少一个AI模型的标识。
另一示例，所述向所述第一通信装置发送所述第二AI模型的标识，包括：向所述第一通信装置发送所述第一组AI模型的组标识。其中，所述第一AI组模型的组标识为AI模型组的标识。作为示例，基于组标识，可获知AI模型组中各个AI模型的标识。
基于上述技术方案，若第二AI模型属于某一组AI模型(如记为第一组AI模型)，则可以向第一通信装置发送第一组AI模型中至少一个AI模型的标识或者第一组AI模型的组标识，这样，第一通信装置可以基于该至少一个AI模型的标识或者第一组AI模型的组标识，判断第一AI模型是否与第一组AI模型中的至少一个AI模型匹配。
结合第一方面，在第一方面的某些实现方式中，所述第一组AI模型中的各个AI模型满足：所述第一组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
结合第一方面，在第一方面的某些实现方式中，在确定所述第二通信装置未部署所述第二AI模型的情况下，所述方法还包括：向所述第一通信装置发送第四AI模型的标识，所述第四AI模型的标识用于所述第一通信装置中所述第四AI模型是否与所述第一通信装置部署的至少一个AI模型匹配的判断。
示例地，第四AI模型为第二通信装置中部署的AI模型。
结合第一方面，在第一方面的某些实现方式中，所述第四AI模型属于第三组AI模型，所述向所述第一通信装置发送第四AI模型的标识，包括：向所述第一通信装置发送所述第三组AI模型中至少一个AI模型的标识。
一示例，所述向所述第一通信装置发送第四AI模型的标识，包括：向所述第一通信装置发送所述第四AI模型的标识和所述第三组AI模型中除所述第四AI模型以外的至少一个AI模型的标识。
另一示例，所述向所述第一通信装置发送第四AI模型的标识，包括：向所述第一通信装置发送所述第三组AI模型的组标识。
基于上述技术方案，若第四AI模型属于某一组AI模型(如记为第三组AI模型)，则可以向第一通信装置发送第三组AI模型中至少一个AI模型的标识，这样，第一通信装置可以基于该第三组AI模型中至少一个AI模型的标识，判断第一通信装置是否部署与第三组AI模型中的至少一个AI模型匹配的AI模型。
结合第一方面，在第一方面的某些实现方式中，所述第三组AI模型中的各个AI模型满足：所述第三组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
结合第一方面，在第一方面的某些实现方式中，所述第二AI模型为与第二组AI模型中至少一个AI模型匹配的AI模型，所述第二组AI模型包括所述第一AI模型，所述接收来自所述第一通信装置的第一AI模型的标识，包括：接收来自所述第一通信装置的所述第二组AI模型中至少一个AI模型的标识；所述基于所述第一AI模型的标识，确定第二通信装置是否部署第二AI模型，包括：基于所述第二组AI模型中至少一个AI模型的标识，确定所述第二通信装置是否部署所述第二AI模型。
一示例，所述接收来自所述第一通信装置的第一AI模型的标识，包括：接收来自所述第一通信装置的所述第一AI模型的标识和所述第二组AI模型中除所述第一AI模型以外的至少一个AI模型的标识。
另一示例，所述接收来自所述第一通信装置的第一AI模型的标识，包括：接收来自所述第一通信 装置的所述第二组AI模型的组标识。其中，所述第二AI组模型的组标识为AI模型组的标识。作为示例，基于组标识，可获知AI模型组中各个AI模型的标识。
基于上述技术方案，第二AI模型可以为与一组AI模型(如记为第二组AI模型)中至少一个AI模型匹配的AI模型。这样，在确定第二通信装置是否部署第二AI模型时，只要确定第二通信装置部署的AI模型与第二组AI模型中至少一个AI模型匹配，则可以认为第二通信装置部署的AI模型与第二组AI模型中各个AI模型都匹配。
结合第一方面，在第一方面的某些实现方式中，所述第二组AI模型包括所述第一AI模型和第三AI模型，且所述第一AI模型的优先级高于所述第三AI模型的优先级，所述基于所述第一AI模型的标识以及所述第二组AI模型中除所述第一AI模型以外的至少一个AI模型的标识，确定所述第二通信装置是否部署所述第二AI模型，包括：在基于所述第一AI模型的标识确定所述第二通信装置未部署所述第二AI模型的情况下，再基于所述第三AI模型的标识，确定所述第二通信装置是否部署所述第二AI模型。
基于上述技术方案，AI模型组内各个AI模型的优先级可以不同，这样在基于AI模型组内的各个AI模型的标识进行匹配时，可以按照优先级高低，依次进行匹配。
结合第一方面，在第一方面的某些实现方式中，所述第二组AI模型中的各个AI模型满足：所述第二组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
结合第一方面，在第一方面的某些实现方式中，所述第一AI模型的标识指示以下至少一项：所述第一AI模型所属的所述第一通信装置的类型、所述第一AI模型所属的所述第一通信装置的标识、所述第一AI模型所属的所述第一通信装置的厂商、所述第一AI模型的功能、所述第一AI模型适用的场景、所述第一AI模型的数据集、所述第一AI模型适用的通信参数、或，所述第一AI模型的版本。
示例地，第一AI模型所属的第一通信装置的类型标识：用于指示训练和/或使用该第一AI模型的通信装置的类型。
示例地，第一AI模型所属的第一通信装置的标识：用于指示训练和/或使用该第一AI模型的通信装置。
示例地，第一AI模型所属的第一通信装置的厂商标识：用于指示训练和/或使用该第一AI模型的厂商。
示例地，第一AI模型的功能标识：用于指示第一AI模型的功能。第一AI模型的功能，例如可以理解为第一AI模型能够解决的问题，或者第一AI模型能够执行的任务。
示例地，第一AI模型适用的场景标识：用于指示第一AI模型适用的场景，如室内场景、室外场景、城区场景、或，郊区场景等。
示例地，第一AI模型的数据集标识：用于指示第一AI模型是根据哪个数据集训练得到的。作为示例，数据集标识可以和场景标识具有对应关系。
示例地，第一AI模型适用的通信参数标识：用于指示第一AI模型适用的无线参数配置。其中，无线参数配置例如可以包括以下至少一项配置：天线配置、带宽、秩、CSI反馈开销、参考信号配置、载频、或，子载波间隔等。
示例地，第一AI模型的版本标识：用于指示第一AI模型的版本。AI模型的版本，可以是针对同一个AI模型的不同版本，或者也可以针对不同的AI模型。一种可能的方式中，版本标识可以指示AI模型的复杂度或性能中的一项或多项。比如，版本标识可以包括复杂度标识，或，性能标识，或，复杂度标识和性能标识。
结合第一方面，在第一方面的某些实现方式中，所述第一AI模型的标识是预定义的，或者，所述第一AI模型的标识是通信装置配置的。
其中，所述第一AI模型的标识是通信装置配置的，包括：所述第一AI模型的标识是所述第一通信装置配置的；或者，所述第一AI模型的标识是其他通信装置配置的。
基于上述技术方案，第一AI模型的标识可以是预定义的，或者也可以是第一通信装置自己确定的，或者也可以是其他装置确定的，不予限制。
结合第一方面，在第一方面的某些实现方式中，所述方法还包括：获取所述第一AI模型的标识的有效期；所述基于所述第一AI模型的标识，确定第二通信装置是否部署第二AI模型，包括：在基于 所述第一AI模型的标识的有效期确定所述第一AI模型的标识有效的情况下，基于所述第一AI模型的标识，确定所述第二通信装置是否部署所述第二AI模型。
第二方面，提供了一种模型匹配的方法，该方法可以由通信装置执行，或者，也可以由用于通信装置的芯片或电路执行，本申请对此不作限定。为了便于描述，下面以由第一通信装置执行为例进行说明。
该方法可以包括：第一通信装置发送第一人工智能AI模型的标识，所述第一AI模型的标识用于第二通信装置是否部署与所述第一AI模型匹配的第二AI模型的确定；所述第一通信装置基于响应，确定所述第二通信装置是否部署所述第二AI模型。
结合第二方面，在第二方面的某些实现方式中，所述方法还包括：所述第一通信装置接收第一反馈信息，所述第一反馈信息用于反馈所述第二通信装置未部署所述第二AI模型；所述基于所述第二通信装置的响应，确定所述第二通信装置是否部署所述第二AI模型，包括：基于所述第一反馈信息，确定所述第二通信装置未部署所述第二AI模型。
结合第二方面，在第二方面的某些实现方式中，所述方法还包括：所述第一通信装置接收第二反馈信息，所述第二反馈信息用于反馈所述第二通信装置部署所述第二AI模型；所述基于所述第二通信装置的响应，确定所述第二通信装置是否部署所述第二AI模型，包括：基于所述第二反馈信息，确定所述第二通信装置部署所述第二AI模型。
结合第二方面，在第二方面的某些实现方式中，在所述第一通信装置发送第一AI模型的标识后，所述方法还包括：所述第一通信装置接收所述第二AI模型的标识；所述第一通信装置基于所述第二AI模型的标识，确定所述第一AI模型是否与所述第二AI模型匹配。
结合第二方面，在第二方面的某些实现方式中，所述第二AI模型属于第一组AI模型，所述第一通信装置接收所述第二AI模型的标识，包括：所述第一通信装置接收所述第二AI模型的标识和所述第一组AI模型中除所述第二AI模型以外的至少一个AI模型的标识；所述第一通信装置基于所述第二AI模型的标识，确定所述第一AI模型是否与所述第二AI模型匹配，包括：所述第一通信装置基于所述第二AI模型的标识和所述第一组AI模型中除所述第二AI模型以外的至少一个AI模型的标识，确定所述第一AI模型是否与所述第一组AI模型中的至少一个AI模型匹配。
结合第二方面，在第二方面的某些实现方式中，所述第一组AI模型中的各个AI模型满足：所述第一组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
结合第二方面，在第二方面的某些实现方式中，在所述第一通信装置发送第一AI模型的标识后，所述方法还包括：所述第一通信装置接收第四AI模型的标识；所述第一通信装置基于所述第四AI模型的标识，确定所述第一通信装置是否部署与所述第四AI模型匹配的AI模型。
结合第二方面，在第二方面的某些实现方式中，所述第四AI模型属于第三组AI模型，所述第一通信装置接收第四AI模型的标识，包括：所述第一通信装置接收所述第四AI模型的标识和所述第三组AI模型中除所述第四AI模型以外的至少一个AI模型的标识；所述第一通信装置基于所述第四AI模型的标识，确定所述第一通信装置是否部署与所述第四AI模型匹配的AI模型，包括：所述第一通信装置基于所述第四AI模型的标识和所述第三组AI模型中除所述第四AI模型以外的至少一个AI模型的标识，确定所述第一通信装置是否部署与所述第三组AI模型中的至少一个AI模型匹配。
结合第二方面，在第二方面的某些实现方式中，所述第三组AI模型中的各个AI模型满足：所述第三组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
结合第二方面，在第二方面的某些实现方式中，所述第二AI模型为与第二组AI模型中至少一个AI模型匹配的AI模型，所述第二组AI模型包括所述第一AI模型，所述第一通信装置发送第一AI模型的标识，包括：所述第一通信装置发送所述第一AI模型的标识以及所述第二组AI模型中除所述第一AI模型以外的至少一个AI模型的标识。
结合第二方面，在第二方面的某些实现方式中，所述第二组AI模型中的各个AI模型满足：所述第二组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
结合第二方面，在第二方面的某些实现方式中，所述方法还包括：所述第一通信装置发送请求信息，所述请求信息用于请求所述第一AI模型的标识；所述第一通信装置接收所述第一AI模型的标识。
基于上述技术方案，第一通信装置可以向其他装置请求第一AI模型的标识，这样可以由其他装置 统一分配各个AI模型的标识，进而可以避免发生各个通信装置各自配置AI模型的标识时AI模型的标识重复。
结合第二方面，在第二方面的某些实现方式中，所述请求信息包括以下至少一项：所述第一AI模型所属的所述第一通信装置的类型、所述第一AI模型所属的所述第一通信装置的标识、所述第一AI模型所属的所述第一通信装置的厂商、所述第一AI模型的功能、所述第一AI模型适用的场景、所述第一AI模型的数据集、所述第一AI模型适用的通信参数、或，所述第一AI模型的版本。
结合第二方面，在第二方面的某些实现方式中，所述方法还包括：所述第一通信装置发送注册信息，所述注册信息包括所述第一AI模型的标识，所述注册信息用于注册所述第一AI模型的标识。
基于上述技术方案，第一通信装置可以向其他装置注册第一AI模型的标识，这样可以由其他装置统一管理各个AI模型的标识。
结合第二方面，在第二方面的某些实现方式中，所述第一AI模型的标识指示以下至少一项：所述第一AI模型所属的所述第一通信装置的类型、所述第一AI模型所属的所述第一通信装置的标识、所述第一AI模型所属的所述第一通信装置的厂商、所述第一AI模型的功能、所述第一AI模型适用的场景、所述第一AI模型的数据集、所述第一AI模型适用的通信参数、或，所述第一AI模型的版本。
结合第二方面，在第二方面的某些实现方式中，所述方法还包括：所述第一通信装置发送所述第一AI模型的有效期。
第二方面及各个可能的设计的有益效果可以参考第一方面相关的描述，在此不予赘述。
第三方面，提供了一种模型匹配的方法，该方法可以由通信装置执行，或者，也可以由用于通信装置的芯片或电路执行，该通信装置可以为第二通信装置，或，不同于第二通信装置的第三通信装置，本申请对此不作限定。
该方法可以包括：接收来自第一通信装置的请求信息，所述请求信息用于请求与第一人工智能AI模型匹配的AI模型；响应于所述请求信息，确定所述第二通信装置是否部署第二AI模型，所述第二AI模型为与第二组AI模型中至少一个AI模型匹配的AI模型，所述第二组AI模型包括所述第一AI模型。
基于上述技术方案，AI模型可以进行分组。举例来说，第一通信装置向第二通信装置发送请求信息，该请求信息用于请求与第一AI模型匹配的AI模型；由于第一AI模型属于第二组AI模型，因此第二通信装置可以判断第二通信装置中是否部署与第二组AI模型中至少一个AI模型匹配的AI模型。若第二通信装置中部署与第二组AI模型中至少一个AI模型匹配的AI模型(如记为第二AI模型)，则可认为该第二AI模型与第二组AI模型中的各个AI模型都匹配。
结合第三方面，在第三方面的某些实现方式中，在确定所述第二通信装置部署所述第二AI模型的情况下，所述方法还包括：确定第一组AI模型中的各个AI模型与所述第一组AI模型中的各个AI模型匹配，所述第一组AI模型包括所述第二AI模型。
结合第三方面，在第三方面的某些实现方式中，所述第一组AI模型中的各个AI模型满足：所述第一组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
结合第三方面，在第三方面的某些实现方式中，所述第二组AI模型中的各个AI模型满足：所述第二组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
第四方面，提供了一种通信方法，该方法可以由通信装置执行，或者，也可以由用于通信装置的芯片或电路执行，本申请对此不作限定。
该方法可以包括：发送请求信息，所述请求信息请求人工智能AI模型的标识；接收所述AI模型的标识。
基于上述技术方案，第一通信装置可以向其他装置请求第一AI模型的标识，这样可以由其他装置统一分配各个AI模型的标识，进而可以避免发生各个通信装置各自配置AI模型的标识时AI模型的标识重复。
第五方面，提供了一种通信方法，该方法可以由通信装置执行，或者，也可以由用于通信装置的芯片或电路执行，本申请对此不作限定。
该方法可以包括：接收请求信息，所述请求信息请求人工智能AI模型的标识；响应于请求信息，确定所述AI模型的标识；发送所述AI模型的标识。
结合第四方面或第五方面，在某些实现方式中，所述请求信息包括以下至少一项：所述AI模型所属的通信装置的类型、所述AI模型所属的通信装置的标识、所述AI模型所属的通信装置的厂商、所述AI模型的功能、所述AI模型适用的场景、所述AI模型的数据集、所述AI模型适用的通信参数、或，所述AI模型的版本。
第五方面的有益效果可以参考第四方面相关的描述，在此不予赘述。
第六方面，提供了一种通信方法，该方法可以由通信装置执行，或者，也可以由用于通信装置的芯片或电路执行，本申请对此不作限定。
该方法可以包括：获取人工智能AI模型的标识；发送注册信息，所述注册信息包括所述AI模型的标识，所述注册信息用于注册所述AI模型的标识。
基于上述技术方案，第一通信装置可以向其他装置注册第一AI模型的标识，这样可以由其他装置统一管理各个AI模型的标识。
第七方面，提供了一种通信方法，该方法可以由通信装置执行，或者，也可以由用于通信装置的芯片或电路执行，本申请对此不作限定。
该方法可以包括：接收注册信息，所述注册信息包括人工智能AI模型的标识，所述注册信息用于注册所述AI模型的标识；响应于所述注册信息，保存所述AI模型的标识。
第七方面的有益效果可以参考第六方面相关的描述，在此不予赘述。
关于第三方面至第七方面，未详细描述的各种的设计，可参考第一方面或第二方面的描述。
第八方面，提供一种通信装置，该装置用于执行上述第一方面至第七方面中任一方面提供的方法。具体地，该装置可以包括用于执行第一方面至第七方面中任一方面的上述任一种实现方式提供的方法的单元和/或模块，如处理单元和/或通信单元。
在一种实现方式中，该装置为通信设备(如终端设备，又如网络设备)。当该装置为通信设备时，通信单元可以是收发器，或，输入/输出接口；处理单元可以是至少一个处理器。可选地，收发器可以为收发电路。可选地，输入/输出接口可以为输入/输出电路。
在另一种实现方式中，该装置为用于通信设备中的芯片、芯片系统或电路。当该装置为用于终端设备中的芯片、芯片系统或电路时，通信单元可以是该芯片、芯片系统或电路上的输入/输出接口、接口电路、输出电路、输入电路、管脚或相关电路等；处理单元可以是至少一个处理器、处理电路或逻辑电路等。
第九方面，提供一种通信装置，该装置包括：存储器，用于存储程序；至少一个处理器，用于执行存储器存储的计算机程序或指令，以执行上述第一方面至第七方面中任一方面的上述任一种实现方式提供的方法。
在一种实现方式中，该装置为通信设备(如终端设备，又如网络设备)。
在另一种实现方式中，该装置为用于通信设备中的芯片、芯片系统或电路。
第十方面，本申请提供一种处理器，用于执行上述各方面提供的方法。
对于处理器所涉及的发送和获取/接收等操作，如果没有特殊说明，或者，如果未与其在相关描述中的实际作用或者内在逻辑相抵触，则可以理解为处理器输出和输入等操作，也可以理解为由射频电路和天线所进行的发送和接收操作，本申请对此不做限定。
第十一方面，提供一种计算机可读存储介质，该计算机可读介质存储用于设备执行的程序代码，该程序代码包括用于执行上述第一方面至第七方面中任一方面的上述任一种实现方式提供的方法。
第十二方面，提供一种包含指令的计算机程序产品，当该计算机程序产品在计算机上运行时，使得计算机执行上述第一方面至第七方面中任一方面的上述任一种实现方式提供的方法。
第十三方面，提供一种芯片，芯片包括处理器与通信接口，处理器通过通信接口读取存储器上存储的指令，执行上述第一方面至第七方面中任一方面的上述任一种实现方式提供的方法。
可选地，作为一种实现方式，芯片还包括存储器，存储器中存储有计算机程序或指令，处理器用于执行存储器上存储的计算机程序或指令，当计算机程序或指令被执行时，处理器用于执行上述第一方面至第七方面中任一方面的上述任意一种实现方式提供的方法。
第十四方面，提供一种通信系统，包括前述的第一通信装置和第二通信装置；或者，包括前述的第一通信装置和第三通信装置；或者，包括前述的第一通信装置、第二通信装置、以及第三通信装置。
附图说明
图1是适用于本申请实施例的无线通信系统100的示意图。
图2是适用于本申请实施例的无线通信系统200的示意图。
图3是神经元结构的示意图。
图4是神经网络的层关系的示意图。
图5是双端模型的示意图。
图6是本申请实施例提供的一种通信方法600的示意图。
图7是适用于本申请实施例的两组AI模型匹配的示意图。
图8是适用于本申请实施例的AI模型的标识的一示意图。
图9是适用于本申请实施例的AI模型的标识的另一示意图。
图10是适用于本申请实施例的AI模型的标识的另一示意图。
图11是适用于本申请实施例的AI模型的标识的注册的示意图。
图12是适用于本申请实施例的AI模型的标识的通知的示意图。
图13是根据本申请一实施例提供的通信方法1300的示意图。
图14是根据本申请另一实施例提供的通信方法1400的示意图。
图15是根据本申请另一实施例提供的通信方法1500的示意图。
图16是根据本申请另一实施例提供的通信方法1600的示意图。
图17是本申请实施例提供的一种通信装置1700的示意性框图。
图18是本申请实施例提供另一种通信装置1800的示意图。
图19是本申请实施例提供一种芯片系统1900的示意图。
具体实施方式
下面将结合附图，对本申请实施例中的技术方案进行描述。
本申请提供的技术方案可以应用于各种通信系统，例如：第五代(5th generation，5G)或新无线(new radio，NR)系统、长期演进(long term evolution，LTE)系统、LTE频分双工(frequency division duplex，FDD)系统、LTE时分双工(time division duplex，TDD)系统等。本申请提供的技术方案还可以应用于未来的通信系统，如第六代移动通信系统。本申请提供的技术方案还可以应用于设备到设备(device to device，D2D)通信，车到万物(vehicle-to-everything，V2X)通信，机器到机器(machine to machine，M2M)通信，机器类型通信(machine type communication，MTC)，以及物联网(internet of things，IoT)通信系统或者其他通信系统。
本申请实施例中的终端设备包括各种具有无线通信功能的设备，其可用于连接人、物、机器等。终端设备可以广泛应用于各种场景，例如：蜂窝通信，D2D，V2X，端到端(peer to peer，P2P)，M2M，MTC，IoT，虚拟现实(virtual reality，VR)，增强现实(augmented reality，AR)，工业控制，自动驾驶，远程医疗，智能电网，智能家具，智能办公，智能穿戴，智能交通，智慧城市无人机，机器人，遥感，被动传感，定位，导航与跟踪，自主交付等场景。终端设备可以是上述任一场景下的终端，如MTC终端、IoT终端等。终端设备可以是第三代合作伙伴项目(3rd generation partnership project，3GPP)标准的用户设备(user equipment，UE)、终端(terminal)、固定设备、移动台(mobile station)设备或者说移动设备、用户单元(subscriber unit)、手持设备、车载设备、可穿戴设备、蜂窝电话(cellular phone)、智能电话(smart phone)、SIP电话、无线数据卡、个人数字助理(personal digital assistant，PDA)、电脑、平板电脑、笔记本电脑、无线调制解调器、手持设备(handset)、膝上型电脑(laptop computer)、具有无线收发功能的计算机、智能书、车辆、卫星、全球定位系统(global positioning system，GPS)设备、目标跟踪设备、飞行器(例如无人机、直升机、多直升机、四直升机、或飞机等)、船只、遥控设备智能家居设备、工业设备，或者内置于上述设备中的装置(例如，上述设备中的通信模块、调制解调器或芯片等)，或者连接到无线调制解调器的其它处理设备。为了描述方便，下文将终端设备以终端或UE为例来描述。
应理解，在某些场景下，UE还可以用于充当基站。例如，UE可以充当调度实体，其在V2X、D2D 或P2P等场景中的UE之间提供侧行链路信号。
本申请实施例中，用于实现终端设备的功能的装置可以是终端设备，也可以是能够支持终端设备实现该功能的装置，例如芯片系统或芯片，该装置可以被安装在终端设备中。本申请实施例中，芯片系统可以由芯片构成，也可以包括芯片和其他分立器件。
本申请实施例中的网络设备可以是用于与终端设备通信的设备，该网络设备也可以称为接入网设备或无线接入网设备，如网络设备可以是基站。本申请实施例中的网络设备可以是指将终端设备接入到无线网络的无线接入网(radio access network，RAN)节点(或设备)。基站可以广义的覆盖如下中的各种名称，或与如下名称进行替换，比如：节点B(NodeB)、演进型基站(evolved NodeB，eNB)、下一代基站(next generation NodeB，gNB)、中继站、接入点、传输点(transmitting and receiving point，TRP)、发射点(transmitting point，TP)、主站、辅站、多制式无线(motor slide retainer，MSR)节点、家庭基站、网络控制器、接入节点、无线节点、接入点(AP)、传输节点、收发节点、基带单元(BBU)、射频拉远单元(remote radio unit，RRU)、有源天线单元(active antenna unit，AAU)、射频头(remote radio head，RRH)、中心单元(central unit，CU)、分布式单元(distributed unit，DU)、定位节点等。基站可以是宏基站、微基站、中继节点、施主节点或类似物，或其组合。基站还可以指用于设置于前述设备或装置内的通信模块、调制解调器或芯片。基站还可以是移动交换中心以及D2D、V2X、M2M通信中承担基站功能的设备、6G网络中的网络侧设备、未来的通信系统中承担基站功能的设备等。基站可以支持相同或不同接入技术的网络。本申请的实施例对网络设备所采用的具体技术和具体设备形态不做限定。
基站可以是固定的，也可以是移动的。例如，直升机或无人机可以被配置成充当移动基站，一个或多个小区可以根据该移动基站的位置移动。在其他示例中，直升机或无人机可以被配置成用作与另一基站通信的设备。
本申请实施例中，用于实现网络设备的功能的装置可以是终端设备，也可以是能够支持网络设备实现该功能的装置，例如芯片系统或芯片，该装置可以被安装在网络设备中。本申请实施例中，芯片系统可以由芯片构成，也可以包括芯片和其他分立器件。
网络设备和终端设备可以部署在陆地上，包括室内或室外、手持或车载；也可以部署在水面上；还可以部署在空中的飞机、气球和卫星上。本申请实施例中对网络设备和终端设备所处的场景不做限定。
首先简单介绍适用于本申请实施例的通信系统，如下。
图1是适用于本申请实施例的无线通信系统100的示意图。如图1所示，该无线通信系统100可以包括至少一个网络设备，例如图1所示的网络设备110，该无线通信系统100还可以包括至少一个终端设备，例如图1所示的终端设备120和终端设备130。网络设备和终端设备均可配置多个天线，网络设备与终端设备可使用多天线技术通信。终端设备与终端设备之间也可以通信。例如，终端设备与终端设备之间可以直接进行通信。再例如，终端设备与终端设备之间可以通过其他通信设备，如网络设备或其他终端设备，进行通信。
其中，网络设备和终端设备通信时，网络设备可以管理一个或多个小区，一个小区中可以有整数个终端设备。可选地，网络设备110和终端设备120组成一个单小区通信系统，不失一般性，将小区称为小区#1。网络设备110可以是小区#1中的网络设备，或者，网络设备110可以为小区#1中的终端设备(例如终端设备120)服务。
需要说明的是，小区可以理解为网络设备的无线信号覆盖范围内的区域。
图2是适用于本申请实施例的无线通信系统200的示意图。如图2所示，该无线通信系统200可以包括至少一个网络设备，例如图2所示的网络设备210，该无线通信系统200还可以包括至少一个终端设备，例如图2所示的终端设备220和终端设备230，该无线通信系统200还可以包括至少一个人工智能(artificial intelligence，AI)节点，例如图2所示的AI节点240。
可选地，AI节点部署于以下任一项：网络设备、终端设备、核心网。图2中为示例性说明，列举了AI节点240单独部署的情况，如部署于网络设备和终端设备之外的位置。AI节点240可以与网络设备210通信，AI节点240还可通过网络设备210与终端设备220和终端设备230通信。可以理解，AI节点也可能直接与终端设备通信，对此不予限制。
可选地，AI节点用于执行与AI相关的操作。作为示例，作为示例，与AI相关的操作例如可以包括：模型失效测试、模型性能测试、模型训练测试、数据采集等。
举例来说，网络设备可将终端设备上报的与AI模型相关的数据转发给AI节点，由AI节点执行与AI相关的操作。再举例来说，网络设备或终端设备可将与AI模型相关的数据转发给AI节点，由AI节点执行与AI相关的操作。再举例来说，AI节点可以将AI相关操作的输出，如训练好的神经网络模型、模型评估、测试结果等，发送给网络设备和/或终端设备。例如，AI节点可直接将AI相关操作的输出发送给网络设备和终端设备。再例如，AI节点可通过网络设备将AI相关操作的输出发送给终端设备。再例如，AI节点可通过终端设备将AI相关操作的输出发送给网络设备。
可以理解，本申请对于AI节点的数量不予限制。例如，当有多个AI节点时，多个AI节点可以基于功能进行划分，如不同的AI节点负责不同的功能。
还可以理解，AI节点可以是各自独立的设备，也可以集成于同一设备中实现不同的功能，或者可以是硬件设备中的网络元件，也可以是在专用硬件上运行的软件功能，或者是平台(例如，云平台)上实例化的虚拟化功能，本申请对于上述AI节点的具体形态不作限定。
还可以理解，在本申请实施例中，通信装置(如第一通信装置，又如第二通信装置，又如第三通信装置)可以是终端设备或终端设备的组成部件(例如芯片或者电路)；或者也可以是网络设备或网络设备的组成部件(例如芯片或者电路)；或者也可以是AI节点或AI节点的组成部件(例如芯片或者电路)，不予限制。
还可以理解，图1和图2是为便于理解而示例的简化示意图，无线通信系统中还可以包括其他网络设备，或者还可以包括其他终端设备，或者还可以包括其他的AI节点，图1和图2中未予以画出。
为了便于理解本申请实施例，下面先对本申请实施例中涉及的术语做简单说明。
1、人工智能：就是让机器具有学习能力，能积累经验，解决人类通过经验可以解决的诸如自然语言理解、图像识别和下棋等问题。人工智能，可以理解为由人制造出来的机器所表现出来的智能。通常人工智能是指通过计算机程序来呈现人类智能的技术。人工智能的目标包括通过构建具有象征意义的推理或推理的计算机程序来理解智能。
2、机器学习(machine learning)：是人工智能的一种实现方式。机器学习是一种能够赋予机器学习的能力，以此让机器完成直接编程无法完成的功能的方法。从实践的意义上来说，机器学习是一种通过利用数据，训练出模型，然后使用模型预测的一种方法。机器学习的方法很多，如神经网络(neural network，NN)、决策树、支持向量机等。机器学习理论主要是设计和分析一些让计算机可以自动学习的算法。机器学习算法是一类从数据中自动分析获得规律，并利用规律对未知数据进行预测的算法。
3、神经网络：是机器学习方法的一种具体体现。神经网络是一种模仿动物神经网络行为特征，进行信息处理的数学模型。神经网络的思想来源于大脑组织的神经元结构。每个神经元可对其输入值做加权求和运算，将加权求和运算的结果通过一个激活函数产生输出。
图3是神经元结构的示意图。如图3所示，假设神经元的输入为x＝[x0,x1,…,xn]，与各输入对应的权值分别为w＝[w,w1,…,wn]，加权求和的偏置为b。其中，b可以为整数，也可以为小数，或者也可以为复数等各种可能的取值。激活函数的形式可以多样化。作为一示例，假设一个神经元的激活函数为：y＝f(z)＝max(0,z)，则该神经元的输出为：作为另一示例，假设一个神经元的激活函数为：y＝f(z)＝z，则该神经元的输出为： 如图3所示。神经网络中不同神经元的激活函数可以相同或不同。
神经网络一般包括多层结构，每层可包括一个或多个逻辑判断单元，这种逻辑判断单元可被称为神经元(neuron)。通过增加神经网络的深度和/或宽度可以提高该神经网络的表达能力，为复杂系统提供更强大的信息提取和抽象建模能力。其中，神经网络的深度可以理解为神经网络包括的层数，每层包括的神经元个数可以称为该层的宽度。
图4是神经网络的层关系的示意图。
一种可能的实现方式，神经网络包括输入层和输出层。神经网络的输入层将接收到的输入经过神 经元处理后，将结果传递给输出层，由输出层得到神经网络的输出结果。
另一种可能的实现方式，神经网络包括输入层、隐藏层和输出层，如图4所示。神经网络的输入层将接收到的输入经过神经元处理后，将结果传递给中间的隐藏层，隐藏层再将计算结果传递给输出层或者相邻的隐藏层，最后由输出层得到神经网络的输出结果。一个神经网络可以包括一层或多层依次连接的隐藏层，不予限制。
在神经网络的训练过程中，可以定义损失函数。损失函数用于衡量模型的预测值和真实值之间的差别。在神经网络的训练过程中，损失函数描述了神经网络的输出值和理想目标值之间的差距或差异。神经网络的训练过程就是通过调整神经网络参数，使得损失函数的值小于阈值门限值或者满足目标需求的过程。其中，神经网络参数可以包括以下至少一项：神经网络的层数、宽度、神经元的权值、或，神经元的激活函数中的参数。
4、AI模型：是能实现AI功能的算法或者计算机程序，AI模型表征了模型的输入和输出之间的映射关系，或者说AI模型是将一定维度的输入映射到一定维度的输出的函数模型，函数模型的参数可通过机器学习训练得到。例如，f(x)＝ax2+b是一个二次函数模型，它可以看做一个AI模型，a和b为该AI模型的参数，a和b可以通过机器学习训练得到。
可以理解，AI模型的实现可以是硬件电路，也可以是软件，或者也可以是软件和硬件结合的方式，不予限制。软件的非限制性示例包括：程序代码、程序、子程序、指令、指令集、代码、代码段、软件模块、应用程序、或软件应用程序等。
5、数据集：机器学习中用于模型训练、模型验证、或模型测试的数据，数据的数量和质量将影响到机器学习的效果。
6、超参数：神经网络的层数，神经元的个数，激活函数，损失函数等参数。
7、模型训练：通过选择合适的损失函数，利用优化算法对模型参数进行训练，使得损失函数值最小化。
8、模型应用：利用训练好的模型去解决实际问题。
9、双端模型：或者叫双边(two-sided)模型、协作模型、对偶模型等。在本申请中，双端模型指的是：由至少两个AI模型组合在一起构成的一个模型。其中，该至少两个AI模型可以部署在至少两个节点，也即该至少两个AI模型没有部署在同一节点，构成双端模型的多个AI模型互相匹配。以两个AI模型为例，为区分，分别记为AI模型#1和AI模型#2，AI模型#1和AI模型#2互相匹配，表示AI模型#1能理解AI模型#2的输出，且能将AI模型#2的输出再解码为期望的输出。
作为示例，编码器和解码器分别部署在不同节点的自编码器(auto-encoder，AE)就是一种双端模型，AE的编码器和解码器互相匹配，也即解码器能理解编码器的输出，且能将编码器的输出再解码为期望的输出。
图5是双端模型的示意图。如图5所示，AI模型#1部署在编码器中，AI模型#2部署在解码器中。对于编码器中部署的AI模型#1来说，AI模型#1的输入为V，AI模型#1的输出为z。对于解码器中部署的AI模型#2，AI模型#2的输入为z，AI模型#2的输出为V’，V’与V相同，或者V’可以较准确地反映V。
通常情况下，双端模型的AI模型为同时训练的，即AI模型是相互匹配的。以AI模型#1和AI模型#2为例，该AI模型#1和AI模型#2可以是由同一节点训练完成再分别部署的两个节点上，或者也可以是两个节点分布式训练完成。
需要说明的是，在本申请中，“指示”可以包括直接指示、间接指示、显示指示、隐式指示。当描述某一指示信息用于指示A时，可以理解为该指示信息携带A、直接指示A，或间接指示A。
本申请中，指示信息所指示的信息，称为待指示信息。在具体实现过程中，对待指示信息进行指示的方式有很多种，例如但不限于，可以直接指示待指示信息，如待指示信息本身或者该待指示信息的索引等。也可以通过指示其他信息来间接指示待指示信息，其中该其他信息与待指示信息之间存在关联关系。还可以仅仅指示待指示信息的一部分，而待指示信息的其他部分则是已知的或者提前约定的。例如，还可以借助预先约定(例如协议规定)的各个信息的排列顺序来实现对特定信息的指示，从而在一定程度上降低指示开销。此外，待指示信息可以作为一个整体一起发送，也可以分成多个子信息分开发送，而且这些子信息的发送周期和/或发送时机可以相同，也可以不同。
下文将结合附图详细说明本申请实施例提供的通信方法。本申请提供的实施例可以应用于上述图1或图2所示的通信系统中，不作限定。在下文实施例中，通信装置(如第一通信装置，又如第二通信装置，又如第三通信装置)可以是终端设备或终端设备的组成部件(例如芯片或者电路)；或者也可以是网络设备或网络设备的组成部件(例如芯片或者电路)；或者也可以是AI节点或AI节点的组成部件(例如芯片或者电路)。
图6是本申请实施例提供的一种通信方法600的示意图。图6所示的方法600可以包括如下步骤。
610，接收来自第一通信装置的第一AI模型的标识。
相应地，第一通信装置发送第一AI模型的标识。
其中，第一AI模型可以是部署在第一通信装置中的AI模型，也即第一AI模型为第一通信装置中的AI模型。第一AI模型的数量可以是至少一个，也就是说，第一通信装置可以发送一个AI模型的标识，或者也可以发送至少两个AI模型的标识，对此不予限制。
关于AI模型的标识，后面详细说明。
620，基于第一AI模型的标识，确定第二通信装置是否部署第二AI模型，第二AI模型为与第一AI模型匹配的AI模型。
在本申请实施例中，通信装置之间可通过交互AI模型的标识，进行AI模型的匹配。举例来说，第一通信装置向第二通信装置发送第一AI模型的标识，第二通信装置可以基于该第一AI模型的标识识别第一AI模型，进而可确定自身是否部署与该第一AI模型匹配的AI模型。这样，不仅可以实现AI模型匹配，还可以因为传输AI模型的标识而不是AI模型，从而保护AI模型的隐私，降低传输AI模型带来的信令开销。
上述步骤610和步骤620至少可以包括如下实现方式。
第一种可能的实现方式，在步骤610中，第二通信装置接收来自第一通信装置的第一AI模型的标识；在步骤620中，第二通信装置基于第一AI模型的标识，确定第二通信装置是否部署第二AI模型。
例如，第一通信装置直接向第二通信装置发送第一AI模型的标识，进而第二通信装置基于第一AI模型的标识，确定第二通信装置是否部署第二AI模型。
再例如，第一通信装置向其他通信装置发送第一AI模型的标识，其他通信装置向第二通信装置转发该第一AI模型的标识，进而第二通信装置基于第一AI模型的标识，确定第二通信装置是否部署第二AI模型。
第二种可能的实现方式，在步骤610中，第三通信装置接收来自第一通信装置的第一AI模型的标识；在步骤620中，第三通信装置基于第一AI模型的标识，确定第二通信装置是否部署第二AI模型。
例如，第一通信装置直接向第三通信装置发送第一AI模型的标识，进而第三通信装置基于第一AI模型的标识，确定第二通信装置是否部署第二AI模型。
再例如，第一通信装置向其他通信装置(如第二通信装置)发送第一AI模型的标识，其他通信装置向第三通信装置转发该第一AI模型的标识，进而第三通信装置基于第一AI模型的标识，确定第二通信装置是否部署第二AI模型。
再例如，第一通信装置直接向第三通信装置发送第一AI模型的标识，第二通信装置直接向第三通信装置发送至少一个AI模型的标识，第三通信装置确定至少一个AI模型中是否有与第一AI模型匹配的AI模型，即确定至少一个AI模型中是否有第二AI模型，从而确定第二通信装置是否部署第二AI模型。
可选地，第一通信装置基于响应，确定第二通信装置是否部署第二AI模型。
一种可能的实现方式，第一通信装置基于收到的反馈信息，确定第二通信装置是否部署第二AI模型。基于该实现方式，方法600还包括：向第一通信装置发送反馈信息，反馈信息用于反馈第二通信装置是否部署第二AI模型。
示例1，在确定第二通信装置未部署第二AI模型的情况下，向第一通信装置发送反馈信息，该反馈信息用于反馈第二通信装置未部署第二AI模型。为区分，将用于反馈第二通信装置未部署第二AI模型的反馈信息记为第一反馈信息。
举例来说，以上述第一种可能的实现方式为例，第二通信装置基于第一AI模型的标识确定第二通信装置未部署第二AI模型后，向第一通信装置发送第一反馈信息。例如，第二通信装置直接向第一通 信装置发送第一反馈信息。再例如，第二通信装置向其他通信装置发送第一反馈信息，其他通信装置向第一通信装置转发该第一反馈信息。
再举例来说，以上述第二种可能的实现方式为例，第三通信装置基于第一AI模型的标识确定第二通信装置未部署第二AI模型后，向第一通信装置发送第一反馈信息。例如，第三通信装置直接向第一通信装置发送第一反馈信息。再例如，第三通信装置向其他通信装置(如第二通信装置)发送第一反馈信息，其他通信装置向第一通信装置转发该第一反馈信息。
示例2，在确定第二通信装置部署第二AI模型的情况下，向第一通信装置发送反馈信息，该反馈信息用于反馈第二通信装置部署第二AI模型。为区分，将用于反馈第二通信装置部署第二AI模型的反馈信息记为第二反馈信息。
举例来说，以上述第一种可能的实现方式为例，第二通信装置基于第一AI模型的标识确定第二通信装置部署第二AI模型后，向第一通信装置发送第二反馈信息。例如，第二通信装置直接向第一通信装置发送第二反馈信息。再例如，第二通信装置向其他通信装置发送第二反馈信息，其他通信装置向第一通信装置转发该第二反馈信息。
再举例来说，以上述第二种可能的实现方式为例，第三通信装置基于第一AI模型的标识确定第二通信装置部署第二AI模型后，向第一通信装置发送第二反馈信息。例如，第三通信装置直接向第一通信装置发送第二反馈信息。再例如，第三通信装置向其他通信装置(如第二通信装置)发送第二反馈信息，其他通信装置向第一通信装置转发该第二反馈信息。
另一种可能的实现方式，第一通信装置基于未收到反馈信息，确定第二通信装置是否部署第二AI模型。
下面结合两种情形分别进行介绍。
第一种可能的情形，第二通信装置未部署第二AI模型。
举例来说，以上述第一种可能的实现方式为例，第二通信装置基于第一AI模型的标识确定第二通信装置未部署第二AI模型后，第二通信装置不发送反馈信息。作为示例，该反馈信息可以为肯定应答。基于该实现方式，第一通信装置在一段时间(为区分，记为时间段#1)内没有收到来自第二通信装置的肯定应答，则第一通信装置默认第二通信装置中没有部署与第一AI模型匹配的第二AI模型。作为示例，时间段#1的起始时刻可以是第一通信装置发送第一AI模型的标识的时刻，时间段#1的时长可以是预定义的，或者也可以是根据历史情况估计的，不予限制。作为示例，时间段#1可通过定时器实现。
第二种可能的情形，第二通信装置部署第二AI模型。
举例来说，以上述第一种可能的实现方式为例，第二通信装置基于第一AI模型的标识确定第二通信装置部署第二AI模型后，第二通信装置不发送反馈信息。作为示例，该反馈信息可以为否定应答。基于该实现方式，第一通信装置在一段时间(为区分，记为时间段#2)内没有收到来自第二通信装置的否定应答，则第一通信装置默认第二通信装置中部署与第一AI模型匹配的第二AI模型。作为示例，时间段#2的起始时刻可以是第一通信装置发送第一AI模型的标识的时刻，时间段#2的时长可以是预定义的，或者也可以是根据历史情况估计的，不予限制。作为示例，时间段#2可通过定时器实现。
可选地，步骤620中，基于第一AI模型的标识确定第二通信装置是否部署第二AI模型，包括：基于第一AI模型的标识和关联关系，确定第二通信装置是否部署第二AI模型。
其中，关联关系表示相互匹配的AI模型的标识之间的关系。具体来说，若两个AI模型是匹配的，那么该两个AI模型的标识也是关联的，也即该两个AI模型的标识之间具有关联关系。
例如，若AI模型#1和AI模型#2为匹配的AI模型，如AI模型#1和AI模型#2构成了双端模型，则关联关系包括AI模型#1的标识和AI模型#2的标识之间的关系。也即在确定了AI模型#1的标识后，也确定了与AI模型#1匹配的AI模型#2的标识。其中，部署AI模型#1的通信装置和/或部署AI模型#2的通信装置，可以维护AI模型#1的标识和AI模型#2的标识之间的关联关系。
作为示例，关联关系可以以表格，函数，或，字符串的形式存在，如存储或传输，如下表1为以表格形式呈现关联关系的示例。
以表1为例，AI模型#1与AI模型#2为匹配的AI模型，AI模型#3与AI模型#4为匹配的AI模型，AI模型#5与AI模型#6、以及AI模型#7为匹配的AI模型，因此关联关系包括：AI模型#1的标 识和AI模型#2的标识之间的关系，AI模型#3的标识和AI模型#4的标识之间的关系，AI模型#5的标识和AI模型#6的标识之间的关系，AI模型#5的标识和AI模型#7的标识之间的关系。
表1
一种可能的实现方式，可以先基于关联关系判断与第一AI模型匹配的AI模型为第二AI模型，进而再判断第二通信装置中是否部署该第二AI模型。举例来说，若第一AI模型为AI模型#1，则基于AI模型#1的标识以及上述表1所示的关联关系，确定与该AI模型#1匹配的AI模型为AI模型#2，进而再判断第二通信装置中是否部署AI模型#2。
另一种可能的实现方式，可以先判断第二通信装置中部署的AI模型有哪些，然后再基于关联关系判断第二通信装置中部署的AI是否有与第一AI模型匹配的AI模型。举例来说，若第二通信装置中部署的AI模型包括：AI模型#2、AI模型#4，且第一AI模型为AI模型#1，那么基于上述表1所示的关联关系，可获知AI模型#1与AI模型#2匹配，因此可确定第二通信装置中部署与AI模型#1匹配的AI模型，即AI模型#2。
可以理解，上述表1为示例性说明，对此不予限制。例如，一个AI模型可以匹配更多数量的AI模型。再例如，表1中的AI模型的标识可以替换为其他能够识别AI模型的信息。再例如，关联关系可以以文本的形式存在，对此不予限制。
可选地，在本申请实施例中，AI模型可以进行分组。如果两组AI模型间存在至少1对匹配的AI模型(如通过训练匹配的AI模型)，则两组AI模型是匹配的。以AI模型组#A和AI模型组#B为例，若两组AI模型间存在至少1对匹配的AI模型(如通过训练匹配的AI模型)，则AI模型组#A中的任意一个AI模型都可以与AI模型组#B中的任意AI模型匹配。
图7是适用于本申请实施例的两组AI模型匹配的示意图。
如图7所示，AI模型组#A中的AI模型有：AI模型A1、AI模型A2、AI模型A3，AI模型组#B中的AI模型有：AI模型B1、AI模型B2、AI模型B3，可以认为AI模型组#A中的任意一个AI模型都可以与AI模型组#B中的任意一个AI模型匹配。
具体来说，一个AI模型可能与多个AI模型匹配，如图7。假设AI模型A1与AI模型B1是通过训练匹配的AI模型，AI模型A2与AI模型B1、以及AI模型A2与AI模型B2是通过训练匹配的AI模型，AI模型A3与AI模型B2、以及AI模型A3与AI模型B3是通过训练匹配的AI模型。由此可知，AI模型A1和AI模型A2都与AI模型B1匹配，因此可以认为AI模型A1和AI模型A2是功能相似的，又因为AI模型A2和AI模型B2是匹配的，因此可以认为AI模型A1和AI模型B2也是匹配的。因此，可以认为AI模型组#A中的任意一个AI模型都可以与AI模型组#B中的任意一个AI模型匹配。
示例地，AI模型组可以是在AI模型注册时确定，对此不予限制。
关于AI模型分组的具体方式不予限制。
一种可能的实现方式，功能相似的AI模型可以看成一组。其中，功能相似可以理解为当两个AI模型输入相同时，输出也相同，或者输出的差异位于预设范围内(如小于某一阈值)。
可选地，第二AI模型为与第二组AI模型中至少一个AI模型匹配的AI模型，第二组AI模型包括第一AI模型，步骤610中，接收来自第一通信装置的第一AI模型的标识，包括：接收来自第一通信装置的第二组AI模型中的至少一个AI模型的标识；步骤620中，基于第一AI模型的标识，确定第二通信装置是否部署第二AI模型，包括：基于第二组AI模型中的至少一个AI模型的标识，确定第二通信装置是否部署第二AI模型。作为示例，第二组AI模型中的各个AI模型满足：第二组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
第一种可能的实现方式，步骤610中，接收来自第一通信装置的第二组AI模型中所有AI模型的标识，相应地，步骤620中，基于第二组AI模型中的所有AI模型的标识，确定第二通信装置是否部 署第二AI模型。若第二通信装置中部署与第二组AI模型中至少一个AI模型匹配的AI模型，则确定第二通信装置中部署与第二组AI模型中各个AI模型匹配的AI模型。
下面举例说明。假设，第二组AI模型包括第一AI模型和第三AI模型。基于第一种可能实现方式，步骤610中接收来自第一通信装置的第一AI模型的标识和第三AI模型的标识，相应地，步骤620中，根据第一AI模型的标识和第三AI模型的标识，确定第二通信装置是否部署与第一AI模型或第三AI模型匹配的AI模型。只要第二通信装置部署有与第一AI模型或第三AI模型匹配的AI模型，则认为该AI模型与第一AI模型和第三AI模型均匹配。
可以理解，各个AI模型的标识可以同时发送或者也可以分批次发送，不予限制。此外，各个AI模型的标识可以携带于同一信令中，也可以携带于不同信令中，不予限制。
第二种可能的实现方式，步骤610中，接收来自第一通信装置的第二组AI模型中部分AI模型的标识，相应地，步骤620中，基于第二组AI模型中的部分AI模型的标识，确定第二通信装置是否部署第二AI模型。若第二通信装置中部署与该部分AI模型中至少一个AI模型匹配的AI模型，则确定第二通信装置中部署与第二组AI模型中各个AI模型匹配的AI模型。
下面举例说明。假设，第二组AI模型包括第一AI模型和第三AI模型。基于第二种可能实现方式，步骤610中接收来自第一通信装置的第一AI模型的标识或第三AI模型的标识，相应地，步骤620中，根据第一AI模型的标识或第三AI模型的标识，确定第二通信装置是否部署与第一AI模型或第三AI模型匹配的AI模型。只要第二通信装置部署有与第一AI模型或第三AI模型匹配的AI模型，则认为该AI模型与第一AI模型和第三AI模型均匹配。
可以理解，各个AI模型的标识可以同时发送或者也可以分批次发送，不予限制。此外，各个AI模型的标识可以携带于同一信令中，也可以携带于不同信令中，不予限制。
第三种可能的实现方式，步骤610中，接收来自第一通信装置的第二组AI模型的组标识，相应地，步骤620中，基于第二组AI模型的组标识，确定第二通信装置是否部署第二AI模型。
其中，组标识为AI模型组的标识。作为示例，基于组标识，可获知AI模型组中各个AI模型的标识。
作为示例，组标识和组标识对应的AI模型组内的各个AI模型的标识可以以表格，函数，或，字符串的形式存在，如存储或传输，其中，如下表2给出了以表格形式存储的一种示例。
表2
基于上述表2可知，若获知AI模型组的组标识，则可获知该AI模型组内各个AI模型的标识。举例来说，若第二组AI模型为AI模型组#A，步骤610中接收来自第一通信装置的组标识AI模型组#A，基于该AI模型组#A以及上述表2可获知AI模型A1的标识、AI模型A2的标识、以及AI模型A3的标识，进而可以根据AI模型A1的标识、AI模型A2的标识、以及AI模型A3的标识，确定第二通信装置是否部署与AI模型A1、AI模型A2、AI模型A3中是至少一个AI模型匹配的AI模型。只要第二通信装置部署有与AI模型A1、AI模型A2、AI模型A3中是至少一个AI模型匹配的AI模型，则认为该AI模型与第二组AI模型中的AI模型均匹配。
可以理解，上述表2为示例性说明，对此不予限制。例如，一个AI模型组中可以包括更多数量的AI模型。再例如，关联关系也可以是AI模型组的组标识之间的关联关系。再例如，组标识和组标识对应的AI模型组内的各个AI模型的标识可以以文本的形式存在，对此不予限制。
可选地，AI模型组内各个AI模型的优先级不同，优先级高的优先匹配。例如，第二组AI模型包括第一AI模型和第三AI模型，且第一AI模型的优先级高于第三AI模型的优先级，在基于第一AI模型的标识以及第三AI模型的标识，确定第二通信装置是否部署第二AI模型时，可以先在基于第一AI模型的标识确定第二通信装置是否部署与第一AI模型匹配的AI模型，在基于第一AI模型的标识确定第二通信装置未部署与第一AI模型匹配的AI模型的情况下，再基于第三AI模型的标识，确定第二通信装置是否部署与第三AI模型匹配的AI模型。
作为示例，可以通过以下任一方式获知各个AI模型的优先级。
第一种可能的实现方式，显式指示各个AI模型的优先级。
以第二组AI模型为例，第一通信装置发送第二组AI模型中各个AI模型的优先级。以第二通信装置确定第二通信装置是否部署第二AI模型为例，第一通信装置向第二通信装置发送第二组AI模型中各个AI模型的优先级，第二通信装置基于该各个AI模型的优先级，按照优先级高低，依次判断是否有与第二组AI模型中各个AI模型匹配的AI模型。
其中，优先级与AI模型的标识或组标识可以携带于同一信令中，也可以携带于不同信令中，不予限制。
第二种可能的实现方式，隐式指示各个AI模型的优先级。
以第二组AI模型为例，一示例，第一通信装置发送第二组AI模型中各个AI模型的标识时，可以根据AI模型的优先级依次排序，如按照优先级从高到低或从低到高依次排序。以第二通信装置确定第二通信装置是否部署第二AI模型为例，例如，第一通信装置根据AI模型的优先级从高到低依次向第二通信装置发送第二组AI模型中各个AI模型的标识；第二通信装置从收到的第一个AI模型的标识开始，依次判断是否有与第二组AI模型中各个AI模型匹配的AI模型。
可以理解，上述根据AI模型的优先级依次排序，例如，可以是通过发送不同AI模型的标识的时间依次排序，或者也可以是通过同一信令中的不同字段依次排序，不予限制。
可以理解，上述两种实现方式以AI模型组内各个AI模型的优先级为例进行示例性说明，对此不予限制。例如，第一通信装置发送至少两个AI模型的标识时(该至少两个AI模型可能不是同一AI组内的AI模型)，该至少两个AI模型也可以具有不同的优先级，优先级的指示方式可以参考前述AI模型组内各个AI模型的优先级的指示方式，在此不予赘述。
可选地，在确定第二通信装置部署第二AI模型的情况下，方法600还包括：向第一通信装置发送第二AI模型的标识，第二AI模型的标识用于第一通信装置中第一AI模型是否与第二AI模型匹配的判断。
基于此，第二通信装置或第三通信装置确定第二通信装置部署第二AI模型的情况下，向第一通信装置发送该第二AI模型的标识，这样第二通信装置或第三通信装置，以及第一通信装置都进行了AI模型匹配的确认，这样可以降低AI模型匹配错误发生的概率。举例来说，以第二通信装置确定第二通信装置部署第二AI模型为例，若第二通信装置和第一通信装置对AI模型的标识的理解不一致，或者AI模型的标识存在冲突等，可能会导致AI模型匹配错误，通过第二通信装置和第一通信装置均进行AI模型匹配的确认，可以降低AI模型匹配错误的概率。例如，若第一通信装置和第三通信装置分别各自配置AI模型的标识，可能会出现第一通信装置和第三通信装置配置的AI模型的标识一样的情况。若第三通信装置中部署的AI模型与第二通信装置中部署的AI模型匹配，且第一通信装置中部署的AI模型与第二通信装置中部署的AI模型不匹配，那么在第二通信装置收到来自第一通信装置的第一AI模型的标识后，可能会认为该第一AI模型的标识为第三通信装置中部署的AI模型的标识，进而出现匹配错误的情况。
进一步可选地，第一通信装置可以反馈匹配的结果，也即第二通信装置基于响应，确定第一通信装置匹配的结果。该实现方式与前面第一通信装置基于响应确定第二通信装置是否部署第二AI模型的实现方式类似，此处不予赘述。
进一步可选地，若第二AI模型属于第一组AI模型，则向第一通信装置发送第二AI模型的标识，包括：向第一通信装置发送第一组AI模型至少一个AI模型的标识。作为示例，第一组AI模型中的各个AI模型满足：第一组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
第一种可能的实现方式，向第一通信装置发送第一组AI模型中所有AI模型的标识。若第一通信装置中部署与第一组AI模型中至少一个AI模型匹配的AI模型，则确定第一组AI模型中的各个AI模型与第一AI模型匹配，也即第二AI模型与第一AI模型是匹配的。
第二种可能的实现方式，向第一通信装置发送第一组AI模型中部分AI模型的标识。若第一通信装置中部署与该部分AI模型中至少一个AI模型匹配的AI模型，则确定第一组AI模型中的各个AI模型与第一AI模型匹配，也即第二AI模型与第一AI模型是匹配的。
第三种可能的实现方式，向第一通信装置发送第一组AI模型中的组标识。若第一通信装置中部署 与该第一组AI模型中至少一个AI模型匹配的AI模型，则确定第一组AI模型中的各个AI模型与第一AI模型匹配，也即第二AI模型与第一AI模型是匹配的。
上述三种实现方式可参考前面关于第二组AI模型的相关描述，此处不予赘述。
下面介绍一下AI模型的标识的相关方案。
AI模型的标识，表示能够识别AI模型的一些属性的标识。举例来说，在本申请实施例中，AI模型的标识可以有两类：全局标识和局部标识。全局标识，表示能够识别AI模型的各个属性的标识，也即基于该全局标识，可获知AI模型的各个属性。局部标识，表示能够识别AI模型的特定属性的标识，也即基于该局部标识，可获知AI模型的某个或某些属性。在本申请实施例中，通信装置之间传输AI模型的标识时，可以传输AI模型的全局标识，也可以传输AI模型的部分标识，不予限制。
以第一AI模型的标识为例，可选地，第一AI模型的标识指示以下至少一项：第一AI模型所属的第一通信装置的类型、第一AI模型所属的第一通信装置的标识、第一AI模型所属的第一通信装置的厂商、第一AI模型的功能、第一AI模型适用的场景、第一AI模型的数据集、第一AI模型适用的通信参数、或，第一AI模型的版本。
1、首先简单介绍一下上述各项标识。
1)第一AI模型所属的第一通信装置的类型标识，或者简称为类型标识：用于指示训练和/或使用该第一AI模型的通信装置的类型。例如，训练和/或使用该第一AI模型的通信装置的类型为终端设备；再例如，训练和/或使用该第一AI模型的通信装置的类型为网络设备。
一种可能的实现方式，类型标识通过至少一个比特来实现。例如，假设通过1比特来指示类型标识。若该比特设置为“1”，则表示训练和/或使用该第一AI模型的通信装置为终端设备；若该比特设置为“0”，则表示训练和/或使用该第一AI模型的通信装置为网络设备。应理解，上述仅是一种示例性说明，不予限制。
2)第一AI模型所属的第一通信装置的标识，或者简称为装置标识：用于指示训练和/或使用该第一AI模型的通信装置。
一种可能的实现方式，装置标识可通过以下任一项实现：永久设备标识(permanent equipment identifier，PEI)、订阅永久标识(subscription permanent identifier，SUPI)、全局唯一临时身份标识(global unique temporary identifier)、小区无线网络临时标识(cell radio network temmporary identify，C-RNTI)、物理层小区标识(physical layer cell identity)。
3)第一AI模型所属的第一通信装置的厂商标识，或者简称为厂商标识：用于指示训练和/或使用该第一AI模型的厂商。
例如，第一通信装置为终端设备，则该厂商标识可以为终端设备的厂商标识。再例如，第一通信装置为网络设备，则该厂商标识可以为网络设备的厂商标识。再例如，第一通信装置为芯片，则该厂商标识可以为芯片的厂商标识。或者，第一通信装置属于第一通信系统，该厂商标识可以为该系统的运营商标识。
4)第一AI模型的功能标识，或者简称为功能标识：用于指示第一AI模型的功能。第一AI模型的功能，例如可以理解为第一AI模型能够解决的问题，或者第一AI模型能够执行的任务。
作为示例，第一AI模型的功能包括以下至少一项：信道状态信息(channel status information，CSI)反馈、CSI预测、定位、或，波束管理等。
一种可能的实现方式，功能标识通过至少一个比特来实现。例如，假设通过2比特来指示功能标识。若该比特设置为“01”，则表示第一AI模型的功能为CSI反馈；若该比特设置为“10”，则表示第一AI模型的功能为CSI预测；若该比特设置为“00”，则表示第一AI模型的功能为定位；若该比特设置为“11”，则表示第一AI模型的功能为波束管理。应理解，上述仅是一种示例性说明，不予限制。
5)第一AI模型适用的场景标识，或者简称为场景标识：用于指示第一AI模型适用的场景，如室内场景、室外场景、城区场景、或，郊区场景等。
一种可能的实现方式，场景标识通过至少一个比特来实现。例如，假设通过2比特来指示场景标识。若该比特设置为“00”，则表示第一AI模型适用的场景为室内场景；若该比特设置为“01”，则表示第一AI模型适用的场景为室外场景；若该比特设置为“10”，则表示第一AI模型适用的场景为城区场景；若该比特设置为“11”，则表示第一AI模型适用的场景为郊区场景。应理解，上述仅是一种示 例性说明，不予限制。
6)第一AI模型的数据集标识，或者简称为数据集标识：用于指示第一AI模型是根据哪个数据集训练得到的。举例来说，某个装置中有多个数据集，该装置根据该多个数据集可训练得到多个AI模型，数据集标识可指示得到某一AI模型所使用的数据集。作为示例，不同的数据集可以通过不同的标识进行标识。
一种可能的实现方式，数据集标识通过至少一个比特来实现。例如，假设通过1比特来指示数据集标识。若该比特设置为“0”，则表示第一AI模型是根据第一数据集训练得到的；若该比特设置为“1”，则表示第一AI模型是根据第二数据集训练得到的。应理解，上述仅是一种示例性说明，不予限制。
可以理解的是，数据集标识可以和场景标识或第一AI模型适用的通信参数标识中的一项或多项具有对应关系。这样，数据集标识可以替换场景标识或第一AI模型适用的通信参数标识中的一项或多项，也可以和场景标识或第一AI模型适用的通信参数标识中的一项或多项共存。以数据集标识和场景标识具有对应关系为例，若数据集标识和场景标识具有对应关系，则基于数据集标识，以及数据集标识与场景标识之间的对应关系，可以获知该数据集标识对应的场景标识；或者，基于场景标识，以及数据集标识与场景标识之间的对应关系，可以获知该场景标识对应的数据集标识。其中，数据集标识和场景标识之间的对应关系可以为协议预定义，或，预存储，或，预先配置。
7)第一AI模型适用的通信参数标识，或者简称为配置标识或通信参数标识：用于指示第一AI模型适用的无线参数配置。其中，无线参数配置例如可以包括以下至少一项配置：天线配置、带宽、秩、CSI反馈开销、参考信号配置、载频、或，子载波间隔等。
8)第一AI模型的版本标识，或者简称为版本标识：用于指示第一AI模型的版本。AI模型的版本，可以是针对同一个AI模型的不同版本，或者也可以针对不同的AI模型，不予限制。
一种可能的方式中，版本标识可以指示AI模型的复杂度或性能中的一项或多项。
比如，版本标识可以包括复杂度标识，或，性能标识，或，复杂度标识和性能标识。
举例来说，至少两个AI模型的功能或场景相同，但是该至少两个AI模型的复杂度不同，或者说该至少两个AI模型用于解决复杂度不同的问题，因此可通过版本标识进行区分。此时，版本标识也可称为复杂度标识。以AI模型#1和AI模型#2为例，AI模型#1和AI模型#2的功能或场景相同，AI模型#1和AI模型#2的复杂度不同，比如AI模型#1的层数较少，如AI模型#1包括输入层和输出层，AI模型#1用于解决复杂度较低的问题；AI模型#2的层数较多，如AI模型#2包括输入层、隐藏层、以及输出层，AI模型#2用于解决复杂度较高的问题。
再举例来说，至少两个AI模型的功能或场景相同，但是该至少两个AI模型的性能不同，或者说该至少两个AI模型输入相同时，输出值和理想目标值之间的差距不同，因此可通过版本标识进行区分。此时，版本标识也可称为性能标识。以AI模型#1和AI模型#2为例，AI模型#1和AI模型#2的功能或场景相同，AI模型#1和AI模型#2的性能不同，比如AI模型#1的层数较少，如AI模型#1包括输入层、一个隐藏层、以及输出层，AI模型#1的性能较低，即输出值和理想目标值之间的差距较大；AI模型#2的层数较多，如AI模型#2包括输入层、至少两个隐藏层、以及输出层，AI模型#2性能较高，即输出值和理想目标值之间的差距较小。
再举例而言，版本标识可以和上述复杂度标识及上述性能标识具有对应关系，即，版本标识不同时，所对应的AI模型的复杂度或性能中的至少一项不同。该对应关系可以为协议预定义，或，预存储，或，预先配置。
上述举例为示例性说明，对此不予限制。
一种可能的实现方式，版本标识通过至少一个比特来实现。例如，假设通过1比特来指示版本标识。若该比特设置为“0”，则表示第一AI模型适用于解决复杂度较高的问题；若该比特设置为“1”，则表示第一AI模型适用于解决复杂度较低的问题。应理解，上述仅是一种示例性说明，不予限制。
上面介绍了AI模型的各个标识。
可选地，AI模型的标识具有有效期。例如，AI模型的标识对应一有效时间，该有效时间可通过定时器实现，当定时器超时后，该AI模型的标识失效。再例如，当AI模型的标识执行更新、激活、或，重注册等操作中的至少一项时，可更新该有效时间，如重新启动定时器。再例如，当AI模型的标识失效时，可通知其他装置该AI模型的标识失效。
在基于AI模型的标识确定是否部署与该AI模型匹配的AI模型时，可以先判断该AI模型的标识是否处于有效期内。在AI模型的标识处于有效期内的情况下，再基于该AI模型的标识确定是否部署与该AI模型匹配的AI模型。
其中，关于AI模型的标识的有效期的获取方式不予限制。以第一通信装置和第二通信装置为例。一种可能的实现方式，第一通信装置向第二通信装置发送第一AI模型的标识时，发送第一AI模型的标识的有效期；第二通信装置判断第一AI模型的标识处于有效期的情况下，再基于第一AI模型的标识确定是否部署与该第一AI模型匹配的AI模型。其中，第一AI模型的标识和第一AI模型的标识的有效期可以携带于同一信令中，也可以携带于不同信令中，不予限制。另一种可能的实现方式，第二通信装置本身知道第一AI模型的标识的有效期，这样，第二通信装置收到第一AI模型的标识后，判断第一AI模型的标识处于有效期的情况下，再基于第一AI模型的标识确定是否部署与该第一AI模型匹配的AI模型。可以理解，上述两种实现方式为示例性说明，对此不予限制。例如，也可以是另一种可能的实现方式，通信装置在发送AI模型的标识时，发送有效的AI模型的标识，或者有效期大于某一阈值的AI模型的标识。
下面介绍一下AI模型的标识的传输方式。
2、AI模型的标识的传输方式。
第一通信装置在发送第一AI模型的标识时，可以包括以下两种可能的实现方式。
一种可能的实现方式，第一AI模型的标识包括以下至少一项：第一AI模型所属的第一通信装置的类型标识、第一AI模型所属的第一通信装置的标识、第一AI模型所属的第一通信装置的厂商标识、第一AI模型的功能标识、第一AI模型适用的场景标识、第一AI模型的数据集标识、第一AI模型适用的通信参数标识、或，第一AI模型的版本标识。
基于该实现方式，第一通信装置可发送上述至少一项标识，这样通过第一AI模型的标识，可直接获取上述至少一项标识。可以理解，上述各项标识可以携带于同一信令中，也可以携带于不同信令中，不予限制。
另一种可能的实现方式，第一AI模型的标识为某信息，基于该信息可获得以下至少一项：第一AI模型所属的第一通信装置的类型标识、第一AI模型所属的第一通信装置的标识、第一AI模型所属的第一通信装置的厂商标识、第一AI模型的功能标识、第一AI模型适用的场景标识、第一AI模型的数据集标识、第一AI模型适用的通信参数标识、或，第一AI模型的版本标识。也即，上述信息与以上标识中的至少一项具有对应关系。
基于该实现方式，第一通信装置可发送一个信息，通过该信息可间接获取上述至少一项标识。
3、AI模型的标识的确定方式。
可选地，AI模型的标识，是预定义的和/或至少一个装置确定的。下面介绍三种可能的实现方式。
第一种可能的实现方式，AI模型的标识，是预定义的，如标准预定义的。
基于该实现方式，AI模型的标识中字段的数量、字段的顺序和每个字段的长度都可以是预定义的。基于此，各个通信装置对AI模型中各个字段的含义的理解是一致的。
图8是适用于本申请实施例的AI模型的标识的一示意图。如图8所示，AI模型的标识包括8个字段，且8个字段依次承载：类型标识、装置标识、厂商标识、功能标识、场景标识、数据集标识、通信参数标识、版本标识。若AI模型的标识是预定义的，则不同AI模型的标识中，字段的数量、字段的顺序和每个字段的长度都是相同的，不同之处在于，各个字段的取值可能不同。
第二种可能的实现方式，AI模型的标识是至少一个装置确定的。
基于该实现方式，AI模型的标识中字段的数量、字段的顺序和每个字段的长度都可以是自定义的。举例来说，不同AI模型的标识中字段的数量可能是不同的，或者字段的顺序可能是不同的，具体可以实际情况进行配置。基于此，各个通信装置理解自身确定的AI模型中各个字段的含义。
图9是适用于本申请实施例的AI模型的标识的另一示意图。如图9中的(1)所示，该AI模型的标识包括4个字段，且4个字段依次承载：厂商标识、功能标识、场景标识、版本标识。如图9中的(2)所示，该AI模型的标识包括5个字段，且5个字段依次承载：装置标识、功能标识、场景标识、厂商标识、版本标识。
以第一AI模型为例，第一AI模型可以是第一通信装置自身确定的，或者也可以是其他通信装置 确定的，不予限制。下面介绍三个示例。
第一个示例，部署AI模型的通信装置确定该AI模型的标识。以第一AI模型为例，第一AI模型的标识可以是第一通信装置自身确定的。
第二个示例，训练AI模型的通信装置确定该AI模型的标识。以第一AI模型为例，若第一AI模型是第一通信装置训练得到的，则该第一AI模型的标识可以是第一通信装置确定的；或者，若第一AI模型是第一通信装置和第二通信装置共同训练得到的，则该第一AI模型的标识可以是第一通信装置和第二通信装置协商确定的。
第三个示例，第三方确定AI模型的标识。其中，第三方例如可以为核心网，或者可以为接入网设备，或者可以为操作维护管理(operation administration and maintenance，OAM)系统，或者可以为中心管理节点，或者可以为模型管理节点，或者可以为AI节点，该AI节点可与各个通信装置通信。在该示例下，可选地，第三方分配AI模型的标识时，可以考虑AI模型的标识的有效范围，该有效范围可以是小区内，或多个小区内，或者同一个网络内(如公共陆地移动网络(public land mobile network，PLMN)内)。举例来说，如果AI模型的标识的有效范围为某一个小区，则通信设备在该小区使用该AI模型的标识。作为示例，同一个AI模型在不同小区可能有不同的AI模型的标识。
第三种可能的实现方式，AI模型的标识中部分标识是预定义的，部分标识是至少一个装置确定的。
基于该实现方式，AI模型的标识中部分字段的顺序和字段的长度都可以是预定义的，其余字段的数量、顺序、以及长度等可以是至少一个装置基于实际情况进行配置。基于此，各个通信装置对AI模型中部分字段的含义的理解是一致的。
图10是适用于本申请实施例的AI模型的标识的另一示意图。假设AI模型的标识中前3个字段的顺序和长度是预定义的，且该3个字段依次承载：类型标识、装置标识、厂商标识。如图10中的(1)所示，该AI模型的标识除了包括上述3个字段，还包括1个字段，且1个字段承载功能标识。如图10中的(2)所示，该AI模型的标识除了包括上述3个字段，还包括2个字段，且2个字段承载：功能标识和版本标识。
可以理解，上述几种可能的实现方式为示例性说明，对此不予限制。例如，AI模型的标识的格式是预定义的，AI模型的内容是通信装置确定的。
在一种实现方式中，通信装置可以自己确定AI模型的标识，并向其他装置注册；在另一种实现方式中，通信装置可以向其他通信装置请求AI模型的标识，由其他通信装置分配AI模型的标识。该通信装置可以是AI模型所属的通信装置，或者是不同于AI模型所属的通信装置的另一装置。在该通信装置是不同于AI模型所属的通信装置的另一装置的情况下，该通信装置可以向AI模型所属的通信装置通知AI模型的标识。
下面第一通信装置中的第一AI模型为例，结合图11和12分别介绍这两种实现方式。可以理解的是，这两种实现方式可以与前述AI模型的标识相关的其他部分，如AI模型的标识所指示的信息，和/或，确定方式，中的一个或多个实施例，结合使用，也可以与前述AI模型的标识相关的其他部分中的一个或多个实施例解耦。
第一种可能的实现方式，第一通信装置自身确定第一AI模型的标识，并向其他装置(如第三方，又如第二通信装置，又如第三通信装置)注册。
图11是适用于本申请实施例的AI模型的标识的注册的示意图。如图11所示，第一通信装置自身确定第一AI模型的标识后，向其它装置发送注册信息，该注册信息包括第一AI模型的标识。
可选地，其他装置可向第一通信装置发送注册信息的响应，该注册信息的响应用于通知是否注册成功。或者，其他装置在注册成功的情况下，也可不向第一通信装置发送注册信息的响应，第一通信装置在一段时间内未收到该注册信息的响应后，可默认注册成功。作为示例，该段时间的起始时刻可以是第一通信装置发送注册信息的时刻，段时间的时长可以是预定义的，或者也可以是根据历史情况估计的，不予限制。作为示例，段时间可通过定时器实现。
举例来说，例如当第一通信装置请求注册的第一AI模型的标识中某个标识(如版本标识)已经被占用，则一种可能的实现方式，其他装置可向第一通信装置发送注册信息的响应，该注册信息的响应指示第一通信装置更换模型的版本标识；或者，另一种可能的实现方式，其他装置可为第一AI模重新分配一个版本标识，并向第一通信装置发送重新分配的版本标识。
可选地，该注册信息中除了包括第一AI模型的标识外，还可以包括其他信息。可选地，该注册信息包括以下至少一项信息：第一AI模型所属的第一通信装置的类型、第一AI模型所属的第一通信装置、第一AI模型所属的第一通信装置的厂商、第一AI模型的功能、第一AI模型适用的场景、第一AI模型的数据集、第一AI模型适用的通信参数、或，第一AI模型的版本。举例来说，注册信息包括第一AI模型的版本标识以及第一AI模型的版本信息，基于此，第一通信装置和其他装置可对齐该版本标识的理解，以及获知该版本标识所指的版本信息。
可选地，第一通信装置在注册信息中携带上述至少一项标识时，可以对部分标识进行压缩，这样可以减少发送各项标识所占的比特数，降低信令开销。举例来说，AI模型的标识用于指示AI模型的厂商和AI模型的功能，假设AI模型的厂商共8个，若要携带AI模型的厂商标识，则要3比特；假设AI模型的功能共4个，若要携带AI模型的功能标识，则要2比特。因此，如果第一通信装置在注册信息中携带AI模型的厂商标识和AI模型的功能标识，则要5比特。考虑到在AI模型的有效范围内，可能并不是所有的厂商和功能都会出现，因此，在AI模型的标识注册时，可以用小于5比特的标识来联合表示AI模型的厂商和AI模型的功能，例如使用3比特表示AI模型的厂商和AI模型的功能。其中，具体每个厂商和功能的组合分配哪个标识，可以由分配AI模型的实体(如第一通信装置，或者，除第一通信装置外的其他通信装置，该其他通信装置训练该AI模型后将该AI模型部署在第一通信装置中)决定，对此不予限制。
第二种可能的实现方式，其他装置(如第三方，又如第二通信装置，又如第三通信装置)确定第一AI模型的标识，向第一通信装置通知第一AI模型的标识。
图12是适用于本申请实施例的AI模型的标识的通知的示意图。如图12所示，其他装置确定第一AI模型的标识后，向第一通信装置发送第一AI模型的标识。可选地，其他装置确定第一AI模型的标识之前，第一通信装置向其他装置发送请求信息，该请求信息用于请求其他装置为第一AI模型分配第一AI模型的标识。可选地，该请求信息包括以下至少一项信息：第一AI模型所属的第一通信装置的类型、第一AI模型所属的第一通信装置、第一AI模型所属的第一通信装置的厂商、第一AI模型的功能、第一AI模型适用的场景、第一AI模型的数据集、第一AI模型适用的通信参数、或，第一AI模型的版本。
举例来说，若请求信息包括第一AI模型的数据集，则其他装置为第一AI模型分配该第一AI模型的数据集标识。例如假设用于训练得到AI模型的数据集包括两种：第一数据集和第二数据集，若请求信息中包括的第一AI模型的数据集为第一数据集，则其他装置为该第一AI模型分配的数据集标识为“0”；若请求信息中包括的第一AI模型的数据集为第二数据集，则其他装置为该第一AI模型分配的数据集标识为“1”。关于其他的标识类似，此处不予赘述。
示例地，其他装置给第一通信装置分配的AI模型的标识中也可以没有每种信息的具体标识，如数据集标识、功能标识等，而是综合考虑上述至少一项信息后分配的AI模型的标识，该AI模型的标识可以区分不同的上述至少一项信息，但没有每种信息的显式字段。也就是说，该AI模型的标识与前述至少一项信息具有对应关系。
为便于理解，下面以第一通信装置和第二通信装置之间的交互为例，结合图13至图15介绍适用于本申请实施例的流程。下面未详细描述的内容可以参考方法600中的描述。此外，下文为简洁，将通信装置中部署的AI模型简称为通信装置中的AI模型。
图13是根据本申请一实施例提供的通信方法1300的示意图。该方法1300可以适用于第一通信装置向第二通信装置发送至少一个AI模型的标识，第二通信装置执行AI模型匹配的场景。图13所示的方法1300可以包括如下步骤。
1310，第一通信装置向第二通信装置发送T个AI模型的标识。
其中，T为大于1或等于1的整数。T个AI模型可以是第一通信装置中可用的AI模型，或者也可以是第一通信装置中可用的AI模型中的部分AI模型。
关于AI模型的标识，可以参考方法600中的相关描述，此处不予赘述。
1320，第二通信装置根据T个AI模型的标识，判断第二通信装置是否有与该T个AI模型匹配的AI模型。
第二通信装置收到T个AI模型的标识后，判断第二通信装置中是否有与该T个AI模型匹配的 AI模型。
一种可能的实现方式，第二通信装置根据T个AI模型的标识以及关联关系，判断第二通信装置是否有与该T个AI模型匹配的AI模型。其中，关联关系表示相互匹配的AI模型的标识之间的关系。例如，第二通信装置中本身维护关联关系。再例如，第一通信装置中维护关联关系，并且第一通信装置向第二通信装置发送该关联关系。再例如，其它装置维护关联关系，第二通信装置收到第一通信装置的标识后，向其他装置请求关联关系。关于关联关系，可以参考方法600中的相关描述，此处不予赘述。
1330，基于第二通信装置的响应，第一通信装置确定第二通信装置中是否有与该T个AI模型匹配的AI模型。
下面结合两种情形分别进行介绍。
第一种可能的情形，第二通信装置中没有与T个AI模型匹配的AI模型。
在该情形下，第一通信装置可以通过以下任一方式，获知第二通信装置中没有与T个AI模型匹配的AI模型。
一种可能的实现方式，若第二通信装置中没有与T个AI模型匹配的AI模型，则第二通信装置向第一通信装置发送反馈信息，该反馈信息指示第二通信装置没有可以与T个AI模型匹配的AI模型。作为示例，在该实现方式中，反馈信息可以为否定应答。
另一种可能的实现方式，若第二通信装置中没有与T个AI模型匹配的AI模型，则第二通信装置不向第一通信装置发送反馈信息。作为示例，在该实现方式中，反馈信息可以为肯定应答。
关于上述两种实现方式，可以参考方法600中关于第一通信装置基于响应，确定第二通信装置是否部署第二AI模型的相关描述，此处不予赘述。
第二种可能的情形，第二通信装置中有与T个AI模型匹配的AI模型。下面结合三种情况，分别进行说明。
情况1，T等于1，且第二通信装置中有与T个AI模型匹配的AI模型。
一种可能的实现方式，若T等于1，且第二通信装置中有与T个AI模型匹配的AI模型，则第二通信装置可以向第一通信装置发送反馈信息，该反馈信息指示第二通信装置有可以与T个AI模型匹配的AI模型。作为示例，该反馈信息可以为肯定应答。
另一种可能的实现方式，若T等于1，且第二通信装置中有与T个AI模型匹配的AI模型，则第二通信装置不向第一通信装置发送反馈信息。作为示例，该反馈信息可以为否定应答。
关于上述两种实现方式，可以参考方法600中关于第一通信装置基于响应，确定第二通信装置是否部署第二AI模型的相关描述，此处不予赘述。
情况2，T大于1，且第二通信装置中有与T个AI模型中部分AI模型(如记为T1个AI模型)匹配的AI模型。
一种可能的实现方式，若T大于1，且第二通信装置中有与T1个AI模型匹配的AI模型，则第二通信装置可以向第一通信装置发送反馈信息，该反馈信息包括T1个AI模型的信息(如T1个AI模型的标识)，该反馈信息指示第二通信装置有与T1个AI模型匹配的AI模型。基于该实现方式，第一通信装置基于来自第二通信装置的反馈信息获知第二通信装置中有与T1个AI模型匹配的AI模型。
另一种可能的实现方式，若T大于1，且第二通信装置中有与T1个AI模型匹配的AI模型，则第二通信装置可以向第一通信装置发送反馈信息，该反馈信息包括T2个AI模型的信息(如T2个AI模型的标识)，该反馈信息指示第二通信装置没有与T2个AI模型匹配的AI模型，其中，T2个AI模型为T个AI模型中除T1个AI模型以外的AI模型。基于该实现方式，第一通信装置基于来自第二通信装置的反馈信息获知第二通信装置中没有与T2个AI模型匹配的AI模型，且默认第二通信装置中有与T个AI模型中除T2个AI模型以外的AI模型(也即T1个AI模型)匹配的AI模型。
情况3，T大于1，且第二通信装置中有与T个AI模型中所有AI模型匹配的AI模型。
一种可能的实现方式，若T大于1，且第二通信装置中有与T个AI模型中所有AI模型匹配的AI模型，则第二通信装置可以向第一通信装置发送反馈信息，该反馈信息指示第二通信装置有与T个AI模型匹配的AI模型。作为示例，该反馈信息可以为肯定应答。
另一种可能的实现方式，若T大于1，且第二通信装置中有与T个AI模型中所有AI模型匹配的 AI模型，则第二通信装置不向第一通信装置发送反馈信息。作为示例，该反馈信息可以为否定应答。
关于上述两种实现方式，可以参考方法600中关于第一通信装置基于响应，确定第二通信装置是否部署第二AI模型的相关描述，此处不予赘述。
另一种可能的实现方式，若T大于1，且第二通信装置中有与T个AI模型中所有AI模型匹配的AI模型，则第二通信装置可以向第一通信装置发送反馈信息，该反馈信息包括T个AI模型的信息(如T个AI模型的标识)，该反馈信息指示第二通信装置有与T个AI模型匹配的AI模型。
图14是根据本申请另一实施例提供的通信方法1400的示意图。该方法1400可以适用于第一通信装置向第二通信装置发送至少一个AI模型的标识，第二通信装置和第一通信装置执行AI模型匹配的场景。图14所示的方法1400可以包括如下步骤。
1410，第一通信装置向第二通信装置发送T个AI模型的标识。
步骤1410与步骤1310类似，此处不予赘述。
1420，第二通信装置根据T个AI模型的标识，判断第二通信装置中与该T个AI模型匹配的AI模型为N个AI模型。
第二通信装置收到T个AI模型的标识后，判断第二通信装置中是否有与该T个AI模型匹配的AI模型。假设第二通信装置中有与该T个AI模型匹配的AI模型，且与该T个AI模型匹配的AI模型为N个AI模型。
其中，N为大于1或等于1的整数。N个AI模型可以是第二通信装置中可用的，且与T个AI模型中的AI模型匹配的AI模型，也即T个AI模型中有与N个AI模型匹配的AI模型。
步骤1420与步骤1320类似，此处不予赘述。
1430，第二通信装置向第一通信装置发送N个AI模型的标识。
1440，第一通信装置判断第一通信装置中是否有与N个AI模型匹配的AI模型。
可选地，方法1400还包括步骤1450。
1450，基于第一通信装置的响应，第二通信装置确定第一通信装置的匹配结果。
具体来说，第二通信装置可基于第一通信装置的响应，确定第一通信装置中是否有与N个AI模型匹配的AI模型，也即确定第一通信装置对N个AI模型匹配的结果。
在该方案中，第二通信装置和第一通信装置都进行了AI模型匹配的确认，这样可以降低AI模型匹配错误发生的概率。具体的可以参考前面方法600中的相关描述。
下面结合两种情形分别进行介绍。
第一种可能的情形，第一通信装置确定第一通信装置中没有与N个AI模型匹配的AI模型。
一种可能的实现方式，若第一通信装置确定第一通信装置中没有与N个AI模型匹配的AI模型，则第一通信装置向第二通信装置发送反馈信息，该反馈信息指示第一通信装置中没有与N个AI模型匹配的AI模型。作为示例，在该实现方式中，反馈信息可以为否定应答。
另一种可能的实现方式，若第一通信装置确定第一通信装置中没有与N个AI模型匹配的AI模型，则第一通信装置不向第二通信装置发送反馈信息。作为示例，在该实现方式中，反馈信息可以为肯定应答。
上述两种实现方式可以参考方法600中关于第一通信装置基于响应，确定第二通信装置是否部署第二AI模型的相关描述，此处不予赘述。
第二种可能的情形，第一通信装置确定第一通信装置中有与N个AI模型匹配的AI模型。下面结合三种情况，分别进行说明。
情况1，N等于1，且第一通信装置确定第一通信装置中有与N个AI模型匹配的AI模型。
一种可能的实现方式，若N等于1，且第一通信装置确定第一通信装置中有与N个AI模型匹配的AI模型，则第一通信装置可以向第二通信装置发送反馈信息，该反馈信息指示第一通信装置中有与N个AI模型匹配的AI模型。作为示例，该反馈信息可以为肯定应答。
另一种可能的实现方式，若N等于1，且第一通信装置确定第一通信装置中有与N个AI模型匹配的AI模型，则第一通信装置不向第一通信装置发送反馈信息。作为示例，该反馈信息可以为否定应答。
关于上述两种实现方式，可以参考方法600中关于第一通信装置基于响应，确定第二通信装置是否部署第二AI模型的相关描述，此处不予赘述。
情况2，N大于1，且第一通信装置确定第一通信装置中有与N个AI模型中部分AI模型(如记为N1个AI模型)匹配的AI模型。
一种可能的实现方式，若N大于1，且第一通信装置确定第一通信装置中有与N1个AI模型匹配的AI模型，则第一通信装置可以向第二通信装置发送反馈信息，该反馈信息包括N1个AI模型的信息(如N1个AI模型的标识)，该反馈信息指示第一通信装置中有与N1个AI模型匹配的AI模型。
另一种可能的实现方式，若N大于1，且第一通信装置确定第一通信装置中有与N1个AI模型匹配的AI模型，则第一通信装置可以向第二通信装置发送反馈信息，该反馈信息包括N2个AI模型的信息(如N2个AI模型的标识)，该反馈信息指示第一通信装置中没有与N2个AI模型匹配的AI模型，其中，N2个AI模型为N个AI模型中除N1个AI模型以外的AI模型。
关于上述两种实现方式，可以参考方法1300中第二种可能的情形中的情况2的相关描述，此处不予赘述。
情况3，N大于1，且第一通信装置确定第一通信装置中有与N个AI模型中所有AI模型匹配的AI模型。
一种可能的实现方式，若N大于1，且第一通信装置确定第一通信装置中有与N个AI模型中所有AI模型，则第一通信装置可以向第二通信装置发送反馈信息，该反馈信息指示第一通信装置中有与N个AI模型中所有AI模型匹配的AI模型。作为示例，该反馈信息可以为肯定应答。
另一种可能的实现方式，若N大于1，且第一通信装置确定第一通信装置中有与N个AI模型中所有AI模型，则第一通信装置不向第二通信装置发送反馈信息。作为示例，该反馈信息可以为否定应答。
另一种可能的实现方式，若N大于1，且第一通信装置确定第一通信装置中有与N个AI模型中所有AI模型，则第一通信装置可以向第二通信装置发送反馈信息，该反馈信息包括N个AI模型的信息(如N个AI模型的标识)，该反馈信息指示第一通信装置中有与N个AI模型中所有AI模型匹配的AI模型。
关于上述各种实现方式，可以参考方法1300中第二种可能的情形中的情况3的相关描述，此处不予赘述。
图15是根据本申请另一实施例提供的通信方法1500的示意图。该方法1500可以适用于第一通信装置向第二通信装置发送至少一个AI模型的标识和/或与该至少一个AI模型同组的AI模型的标识，第二通信装置执行AI模型匹配的场景。图15所示的方法1500可以包括如下步骤。
1510，第一通信装置向第二通信装置发送T个AI模型的标识和/或与T个AI模型同组的至少一个AI模型的标识。
可以看出，不同于步骤1310，在步骤1510中，第一通信装置可以向第二通信装置发送T个AI模型的标识和/或与T个AI模型同组的至少一个AI模型的标识。关于AI模型分组的相关方案，参考方法600中的相关描述，此处不予赘述。
1520，第二通信装置判断第二通信装置中是否有与以下至少一个AI模型匹配的AI模型：T个AI模型、与T个AI模型同组的至少一个AI模型。
具体来说，在步骤1520中，第二通信装置基于T个AI模型的标识和/或与T个AI模型同组的至少一个AI模型的标识，判断第二通信装置中是否有与以下至少一个AI模型匹配的AI模型：T个AI模型、与T个AI模型同组的至少一个AI模型。
步骤1520与步骤1320类似，不同于之处在于，在步骤1520中，第二通信装置可以判断第二通信装置中是否有与以下至少一个AI模型匹配的AI模型：T个AI模型、与T个AI模型同组的至少一个AI模型。关于AI模型分组的方案，可以参考方法600中的相关描述，此处不再赘述。
可选地，同一组AI模型中，各个AI模型的优先级不同。第二通信装置在进行模型匹配时，可以优先考虑是否有与优先级高的AI模型匹配的AI模型。关于AI模型的优先级，可以参考方法600中的相关描述，此处不予赘述。
1530，基于第二通信装置的响应，第一通信装置确定第二通信装置中是否有与以下至少一个AI模型匹配的AI模型：T个AI模型、与T个AI模型同组的至少一个AI模型。
步骤1530可参考步骤1330中的相关描述，此处不予赘述。
图16是根据本申请另一实施例提供的通信方法1600的示意图。该方法1600可以适用于第一通信 装置向第二通信装置发送至少一个AI模型的标识和/或与该至少一个AI模型同组的AI模型的标识，第二通信装置和第一通信装置执行AI模型匹配的场景。图16所示的方法1600可以包括如下步骤。
1610，第一通信装置向第二通信装置发送T个AI模型的标识和/或与T个AI模型同组的至少一个AI模型的标识。
可以看出，不同于步骤1510，在步骤1610中，第一通信装置可以向第二通信装置发送T个AI模型的标识和/或与T个AI模型同组的至少一个AI模型的标识。
步骤1610可参考步骤1510中的相关描述，此处不予赘述。
1620，第二通信装置判断第二通信装置中与以下至少一个AI模型匹配的AI模型匹配的AI模型为N个AI模型：T个AI模型、与T个AI模型同组的至少一个AI模型。
具体来说，在步骤1620中，第二通信装置基于T个AI模型的标识和/或与T个AI模型同组的至少一个AI模型的标识，确定第二通信装置中与以下至少一个AI模型匹配的AI模型匹配的AI模型为N个AI模型：T个AI模型、与T个AI模型同组的至少一个AI模型。
例如，若在步骤1610中，第一通信装置向第二通信装置发送T个AI模型的标识，则N个AI模型为与该T个AI模型中部分或全部AI模型匹配的AI模型。
再例如，若在步骤1610中，第一通信装置向第二通信装置发送与T个AI模型同组的至少一个AI模型的标识，则N个AI模型为与该至少一个AI模型中部分或全部AI模型匹配的AI模型。
1630，第二通信装置向第一通信装置发送N个AI模型的标识和/或与N个AI模型同组的至少一个AI模型的标识。
步骤1630可参考步骤1510中的相关描述，此处不予赘述。
1640，第一通信装置判断第一通信装置中是否有与以下至少一个AI模型匹配的AI模型：N个AI模型、与N个AI模型同组的至少一个AI模型。
步骤1640可参考步骤1520中的相关描述，此处不予赘述。
可选地，方法1600还包括步骤1650。
1650，基于第一通信装置的响应，第二通信装置确定第一通信装置的匹配结果。
步骤1650与步骤1450类似，此处不予赘述。
可以理解，本申请的各实施例中的一些可选的特征，在某些场景下，可以不依赖于其他特征，也可以在某些场景下，与其他特征进行结合，不作限定。
还可以理解，本申请的各实施例中的方案可以进行合理的组合使用，并且实施例中出现的各个术语的解释或说明可以在各个实施例中互相参考或解释，对此不作限定。
还可以理解，上述各个方法实施例中，由编码器实现的方法和操作，也可以由可由编码器的组成部件(例如芯片或者电路)来实现；此外，由解码器实现的方法和操作，也可以由可由解码器的组成部件(例如芯片或者电路)来实现，不作限定。
以上，结合图6至图16详细说明了本申请实施例提供的方法。以下，结合图11至图13详细说明本申请实施例提供的装置。应理解，装置实施例的描述与方法实施例的描述相互对应，因此，未详细描述的内容可以参见上文方法实施例，为了简洁，这里不予赘述。
图17是本申请实施例提供的一种通信装置1700的示意性框图。该装置1700包括收发单元1710和处理单元1720。收发单元1710可以用于实现相应的通信功能。收发单元1710还可以称为通信接口或通信单元。处理单元1720可以用于进行处理，如确定AI模型是否匹配。
可选地，该装置1700还可以包括存储单元，该存储单元可以用于存储指令和/或数据，处理单元1720可以读取存储单元中的指令和/或数据，以使得装置实现前述方法实施例。
作为一种设计，该装置1700用于执行上文方法实施例中通信装置执行的步骤或者流程，收发单元1710用于执行上文方法实施例中通信装置侧的收发相关的操作，处理单元1720用于执行上文方法实施例中通信装置侧的处理相关的操作。
一种可能的实现方式，该装置1700用于执行如图6所示实施例中第二通信装置或第三通信装置执行的步骤或者流程，或者图13至图16所示实施例中第二通信装置执行的步骤或者流程。可选地，收发单元1710，用于接收来自第一通信装置的第一人工智能AI模型的标识；处理单元1720，用于基于第一AI模型的标识，确定第二通信装置是否部署第二AI模型，第二AI模型为与第一AI模型匹配的 AI模型。
另一种可能的实现方式，该装置1700用于执行如图6所示实施例中第一通信装置执行的步骤或者流程，或者图13至图16所示实施例中第一通信装置执行的步骤或者流程。可选地，收发单元1710，用于发送第一人工智能AI模型的标识，第一AI模型的标识用于确定第二通信装置是否部署与第一AI模型匹配的第二AI模型；处理单元1720，用于基于响应，确定第二通信装置是否部署第二AI模型。
另一种可能的实现方式，该装置1700用于执行如图11所示实施例中第一通信装置执行的步骤或者流程。可选地，处理单元1720，用于确定第一AI模型的标识；收发单元1710，用于向其他装置发送注册信息，该注册信息包括该第一AI模型的标识。或者，该装置1700用于执行如图12所示实施例中第一通信装置执行的步骤或者流程。可选地，收发单元1710，用于从其他装置接收第一AI模型的标识。
另一种可能的实现方式，该装置1700用于执行如图11所示实施例中其他装置执行的步骤或者流程。可选地，收发单元1710，用于接收来自第一通信装置的注册信息。或者，该装置1700用于执行如图12所示实施例中其他装置执行的步骤或者流程。可选地，处理单元1720，用于确定第一AI模型的标识；收发单元1710，用于向第一通信装置发送该第一AI模型的标识。
应理解，各单元执行上述相应步骤的具体过程在上述方法实施例中已经详细说明，为了简洁，在此不再赘述。
还应理解，这里的装置1700以功能单元的形式体现。这里的术语“单元”可以指应用特有集成电路(application specific integrated circuit，ASIC)、电子电路、用于执行一个或多个软件或固件程序的处理器(例如共享处理器、专有处理器或组处理器等)和存储器、合并逻辑电路和/或其它支持所描述的功能的合适组件。在一个可选例子中，本领域技术人员可以理解，装置1700可以具体为上述实施例中的通信装置(如第一通信装置，又如第二通信装置，又如第三通信装置)，可以用于执行上述各方法实施例中与通信装置对应的各个流程和/或步骤，为避免重复，在此不再赘述。
上述各个方案的装置1700具有实现上述方法中通信装置(如第一通信装置，又如第二通信装置，又如第三通信装置，或，其他装置等)所执行的相应步骤的功能。所述功能可以通过硬件实现，也可以通过硬件执行相应的软件实现。所述硬件或软件包括一个或多个与上述功能相对应的模块；例如收发单元可以由收发机替代(例如，收发单元中的发送单元可以由发送机替代，收发单元中的接收单元可以由接收机替代)，其它单元，如处理单元等可以由处理器替代，分别执行各个方法实施例中的收发操作以及相关的处理操作。
此外，上述收发单元1710还可以是收发电路(例如可以包括接收电路和发送电路)，处理单元可以是处理电路。
需要指出的是，图17中的装置可以是前述实施例中的设备，也可以是芯片或者芯片系统，例如：片上系统(system on chip，SoC)。其中，收发单元可以是输入输出电路、通信接口；处理单元为该芯片上集成的处理器或者微处理器或者集成电路。在此不做限定。
图18是本申请实施例提供另一种通信装置1800的示意图。该装置1800包括处理器1810，处理器1810与存储器1820耦合，存储器1820用于存储计算机程序或指令和/或数据，处理器1810用于执行存储器1820存储的计算机程序或指令，或读取存储器1820存储的数据，以执行上文各方法实施例中的方法。
可选地，处理器1810为一个或多个。
可选地，存储器1820为一个或多个。
可选地，该存储器1820与该处理器1810集成在一起，或者分离设置。
可选地，如图18所示，该装置1800还包括收发器1830，收发器1830用于信号的接收和/或发送。例如，处理器1810用于控制收发器1830进行信号的接收和/或发送。
作为示例，处理器1810可以具有图17中所示的处理单元1720的功能，存储器1820可以具有存储单元的功能，收发器1830可以具有图17中所示的收发单元1710的功能。
作为一种方案，该装置1800用于实现上文各个方法实施例中由通信装置(如第一通信装置，又如第二通信装置，又如第三通信装置，或，其他装置等)执行的操作。
例如，处理器1810用于执行存储器1820存储的计算机程序或指令，以实现上文各个方法实施例 中通信装置(如第一通信装置，又如第二通信装置，又如第三通信装置，或，其他装置等)的相关操作。
应理解，本申请实施例中提及的处理器可以是中央处理单元(central processing unit，CPU)，还可以是其他通用处理器、数字信号处理器(digital signal processor，DSP)、专用集成电路(application specific integrated circuit，ASIC)、现成可编程门阵列(field programmable gate array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
还应理解，本申请实施例中提及的存储器可以是易失性存储器和/或非易失性存储器。其中，非易失性存储器可以是只读存储器(read-only memory，ROM)、可编程只读存储器(programmable ROM，PROM)、可擦除可编程只读存储器(erasable PROM，EPROM)、电可擦除可编程只读存储器(electrically EPROM，EEPROM)或闪存。易失性存储器可以是随机存取存储器(random access memory，RAM)。例如，RAM可以用作外部高速缓存。作为示例而非限定，RAM包括如下多种形式：静态随机存取存储器(static RAM，SRAM)、动态随机存取存储器(dynamic RAM，DRAM)、同步动态随机存取存储器(synchronous DRAM，SDRAM)、双倍数据速率同步动态随机存取存储器(double data rate SDRAM，DDR SDRAM)、增强型同步动态随机存取存储器(enhanced SDRAM，ESDRAM)、同步连接动态随机存取存储器(synchlink DRAM，SLDRAM)和直接内存总线随机存取存储器(direct rambus RAM，DR RAM)。
需要说明的是，当处理器为通用处理器、DSP、ASIC、FPGA或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件时，存储器(存储模块)可以集成在处理器中。
还需要说明的是，本文描述的存储器旨在包括但不限于这些和任意其它适合类型的存储器。
图19是本申请实施例提供一种芯片系统1900的示意图。该芯片系统1900(或者也可以称为处理系统)包括逻辑电路1910以及输入/输出接口(input/output interface)1920。
其中，逻辑电路1910可以为芯片系统1900中的处理电路。逻辑电路1910可以耦合连接存储单元，调用存储单元中的指令，使得芯片系统1900可以实现本申请各实施例的方法和功能。输入/输出接口1920，可以为芯片系统1900中的输入输出电路，将芯片系统1900处理好的信息输出，或将待处理的数据或信令信息输入芯片系统1900进行处理。
具体地，例如，若第一通信装置安装了该芯片系统1900，逻辑电路1910与输入/输出接口1920耦合，逻辑电路1910可通过输入/输出接口1920向解码器发送第一AI模型的标识；或者输入/输出接口1920可将来自第二AI模型的标识输入至逻辑电路1910进行处理。又如，若第二通信装置安装了该芯片系统1900，逻辑电路1910与输入/输出接口1920耦合，输入/输出接口1920可将来自第一通信装置的第一AI模型的标识输入至逻辑电路1910进行处理。
作为一种方案，该芯片系统1900用于实现上文各个方法实施例中由通信装置(如第一通信装置，又如第二通信装置，又如第三通信装置，或，其他装置等)执行的操作。
例如，逻辑电路1910用于实现上文方法实施例中由通信装置(如第一通信装置，又如第二通信装置，又如第三通信装置)执行的处理相关的操作；输入/输出接口1920用于实现上文方法实施例中由通信装置(如第一通信装置，又如第二通信装置，又如第三通信装置，或，其他装置等)执行的发送和/或接收相关的操作。
本申请实施例还提供一种计算机可读存储介质，其上存储有用于实现上述各方法实施例中由通信装置(如第一通信装置，又如第二通信装置，又如第三通信装置，或，其他装置等)执行的方法的计算机指令。
例如，该计算机程序被计算机执行时，使得该计算机可以实现上述方法各实施例中由通信装置(如第一通信装置，又如第二通信装置，又如第三通信装置，或，其他装置等)执行的方法。
本申请实施例还提供一种计算机程序产品，包含指令，该指令被计算机执行时以实现上述各方法实施例中由通信装置(如第一通信装置，又如第二通信装置，又如第三通信装置，或，其他装置等)执行的方法。
本申请实施例还提供一种通信系统，该通信系统包括上文各实施例中的第一通信装置和第二通信装置。例如，该系统包含图6所示实施例中的第一通信装置和第二通信装置。再例如，该系统包含图 13至图16所示实施例中的第一通信装置和第二通信装置。
本申请实施例还提供一种通信系统，该通信系统包括上文各实施例中的第一通信装置和第三通信装置。例如，该系统包含图6所示实施例中的第一通信装置和第三通信装置。
本申请实施例还提供一种通信系统，该通信系统包括上文各实施例中的第一通信装置和其他装置。例如，该系统包含图10至图11所示实施例中的第一通信装置和其他装置。
上述提供的任一种装置中相关内容的解释及有益效果均可参考上文提供的对应的方法实施例，此处不再赘述。
在本申请所提供的几个实施例中，应该理解到，所揭露的装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。此外，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机程序指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。例如，所述计算机可以是个人计算机，服务器，或者网络设备等。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。所述可用介质可以是磁性介质(例如，软盘、硬盘、磁带)、光介质(例如，DVD)、或者半导体介质(例如固态硬盘(solid state disk，SSD)等。例如，前述的可用介质包括但不限于：U盘、移动硬盘、只读存储器(read-only memory，ROM)、随机存取存储器(random access memory，RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (53)

  1. 一种模型匹配的方法，其特征在于，包括：
    接收来自第一通信装置的第一人工智能AI模型的标识；
    基于所述第一AI模型的标识，确定第二通信装置是否部署第二AI模型，所述第二AI模型为与所述第一AI模型匹配的AI模型。
  2. 根据权利要求1所述的方法，其特征在于，所述基于所述第一AI模型的标识，确定第二通信装置是否部署第二AI模型，包括：
    基于所述第一AI模型的标识和关联关系，确定所述第二通信装置是否部署所述第二AI模型，所述关联关系表示相互匹配的AI模型的标识之间的关系。
  3. 根据权利要求1或2所述的方法，其特征在于，所述方法还包括：
    在确定所述第二通信装置未部署所述第二AI模型的情况下，向所述第一通信装置发送第一反馈信息，所述第一反馈信息用于反馈所述第二通信装置未部署所述第二AI模型；或者，
    在确定所述第二通信装置部署所述第二AI模型的情况下，向所述第一通信装置发送第二反馈信息，所述第二反馈信息用于反馈所述第二通信装置部署所述第二AI模型。
  4. 根据权利要求1至3中任一项所述的方法，其特征在于，在确定所述第二通信装置部署所述第二AI模型的情况下，所述方法还包括：
    向所述第一通信装置发送所述第二AI模型的标识，所述第二AI模型的标识用于所述第一通信装置中所述第一AI模型是否与所述第二AI模型匹配的判断。
  5. 根据权利要求4所述的方法，其特征在于，所述第二AI模型属于第一组AI模型，
    所述向所述第一通信装置发送所述第二AI模型的标识，包括：
    向所述第一通信装置发送所述第二AI模型的标识和所述第一组AI模型中除所述第二AI模型以外的至少一个AI模型的标识。
  6. 根据权利要求5所述的方法，其特征在于，所述第一组AI模型中的各个AI模型满足：所述第一组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
  7. 根据权利要求1至6中任一项所述的方法，其特征在于，所述第二AI模型为与第二组AI模型中至少一个AI模型匹配的AI模型，所述第二组AI模型包括所述第一AI模型，
    所述接收来自所述第一通信装置的第一AI模型的标识，包括：
    接收来自所述第一通信装置的所述第一AI模型的标识以及所述第二组AI模型中除所述第一AI模型以外的至少一个AI模型的标识；
    所述基于所述第一AI模型的标识，确定第二通信装置是否部署第二AI模型，包括：
    基于所述第一AI模型的标识以及所述第二组AI模型中除所述第一AI模型以外的至少一个AI模型的标识，确定所述第二通信装置是否部署所述第二AI模型。
  8. 根据权利要求7所述的方法，其特征在于，所述第二组AI模型包括所述第一AI模型和第三AI模型，且所述第一AI模型的优先级高于所述第三AI模型的优先级，
    所述基于所述第一AI模型的标识以及所述第二组AI模型中除所述第一AI模型以外的至少一个AI模型的标识，确定所述第二通信装置是否部署所述第二AI模型，包括：
    在基于所述第一AI模型的标识确定所述第二通信装置未部署所述第二AI模型的情况下，再基于所述第三AI模型的标识，确定所述第二通信装置是否部署所述第二AI模型。
  9. 根据权利要求7或8所述的方法，其特征在于，所述第二组AI模型中的各个AI模型满足：所述第二组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
  10. 根据权利要求1至9中任一项所述的方法，其特征在于，所述第一AI模型的标识指示以下至少一项：
    所述第一AI模型所属的所述第一通信装置的类型、所述第一AI模型所属的所述第一通信装置的标识、所述第一AI模型所属的所述第一通信装置的厂商、所述第一AI模型的功能、所述第一AI模型 适用的场景、所述第一AI模型的数据集、所述第一AI模型适用的通信参数、或，所述第一AI模型的版本。
  11. 根据权利要求1至10中任一项所述的方法，其特征在于，所述第一AI模型的标识是预定义的，或者，第三通信装置配置的。
  12. 根据权利要求1至11中任一项所述的方法，其特征在于，所述方法还包括：
    获取所述第一AI模型的标识的有效期；
    所述基于所述第一AI模型的标识，确定第二通信装置是否部署第二AI模型，包括：
    在基于所述第一AI模型的标识的有效期确定所述第一AI模型的标识有效的情况下，基于所述第一AI模型的标识，确定所述第二通信装置是否部署所述第二AI模型。
  13. 一种模型匹配的方法，其特征在于，所述方法由第一通信装置或被配置用于所述第一通信装置的芯片或电路执行，所述方法包括：
    所述第一通信装置发送第一人工智能AI模型的标识，所述第一AI模型的标识用于第二通信装置是否部署与所述第一AI模型匹配的第二AI模型的确定；
    所述第一通信装置基于响应，确定所述第二通信装置是否部署所述第二AI模型。
  14. 根据权利要求13所述的方法，其特征在于，所述方法还包括：
    所述第一通信装置接收第一反馈信息，所述第一反馈信息用于反馈所述第二通信装置未部署所述第二AI模型；
    所述基于所述第二通信装置的响应，确定所述第二通信装置是否部署所述第二AI模型，包括：
    基于所述第一反馈信息，确定所述第二通信装置未部署所述第二AI模型。
  15. 根据权利要求13所述的方法，其特征在于，所述方法还包括：
    所述第一通信装置接收第二反馈信息，所述第二反馈信息用于反馈所述第二通信装置部署所述第二AI模型；
    所述基于所述第二通信装置的响应，确定所述第二通信装置是否部署所述第二AI模型，包括：
    基于所述第二反馈信息，确定所述第二通信装置部署所述第二AI模型。
  16. 根据权利要求13或15所述的方法，其特征在于，在所述第一通信装置发送第一AI模型的标识后，所述方法还包括：
    所述第一通信装置接收所述第二AI模型的标识；
    所述第一通信装置基于所述第二AI模型的标识，确定所述第一AI模型是否与所述第二AI模型匹配。
  17. 根据权利要求16所述的方法，其特征在于，所述第二AI模型属于第一组AI模型，
    所述第一通信装置接收所述第二AI模型的标识，包括：
    所述第一通信装置接收所述第二AI模型的标识和所述第一组AI模型中除所述第二AI模型以外的至少一个AI模型的标识；
    所述第一通信装置基于所述第二AI模型的标识，确定所述第一AI模型是否与所述第二AI模型匹配，包括：
    所述第一通信装置基于所述第二AI模型的标识和所述第一组AI模型中除所述第二AI模型以外的至少一个AI模型的标识，确定所述第一AI模型是否与所述第一组AI模型中的至少一个AI模型匹配。
  18. 根据权利要求17所述的方法，其特征在于，所述第一组AI模型中的各个AI模型满足：所述第一组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
  19. 根据权利要求13至18中任一项所述的方法，其特征在于，所述第二AI模型为与第二组AI模型中至少一个AI模型匹配的AI模型，所述第二组AI模型包括所述第一AI模型，
    所述第一通信装置发送第一AI模型的标识，所述第一AI模型的标识用于第二通信装置是否部署与所述第一AI模型匹配的第二AI模型的确定，包括：
    所述第一通信装置发送所述第一AI模型的标识以及所述第二组AI模型中除所述第一AI模型以外的至少一个AI模型的标识，所述第一AI模型的标识以及所述第二组AI模型中除所述第一AI模型以外的至少一个AI模型的标识用于所述第二通信装置是否部署与所述第一AI模型匹配的第二AI模型的确定。
  20. 根据权利要求19所述的方法，其特征在于，所述第二组AI模型中的各个AI模型满足：所述第二组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
  21. 根据权利要求13至20中任一项所述的方法，其特征在于，所述方法还包括：
    所述第一通信装置发送请求信息，所述请求信息用于请求所述第一AI模型的标识；
    所述第一通信装置接收所述第一AI模型的标识。
  22. 根据权利要求21所述的方法，其特征在于，所述请求信息包括以下至少一项：
    所述第一AI模型所属的所述第一通信装置的类型、所述第一AI模型所属的所述第一通信装置的标识、所述第一AI模型所属的所述第一通信装置的厂商、所述第一AI模型的功能、所述第一AI模型适用的场景、所述第一AI模型的数据集、所述第一AI模型适用的通信参数、或，所述第一AI模型的版本。
  23. 根据权利要求13至20中任一项所述的方法，其特征在于，所述方法还包括：
    所述第一通信装置发送注册信息，所述注册信息包括所述第一AI模型的标识，所述注册信息用于注册所述第一AI模型的标识。
  24. 根据权利要求13至23中任一项所述的方法，其特征在于，所述第一AI模型的标识指示以下至少一项：
    所述第一AI模型所属的所述第一通信装置的类型、所述第一AI模型所属的所述第一通信装置的标识、所述第一AI模型所属的所述第一通信装置的厂商、所述第一AI模型的功能、所述第一AI模型适用的场景、所述第一AI模型的数据集、所述第一AI模型适用的通信参数、或，所述第一AI模型的版本。
  25. 根据权利要求13至24中任一项所述的方法，其特征在于，所述方法还包括：
    所述第一通信装置发送所述第一AI模型的有效期。
  26. 一种模型匹配的方法，其特征在于，包括：
    接收来自第一通信装置的请求信息，所述请求信息用于请求与第一人工智能AI模型匹配的AI模型；
    响应于所述请求信息，确定所述第二通信装置是否部署第二AI模型，所述第二AI模型为与第二组AI模型中至少一个AI模型匹配的AI模型，所述第二组AI模型包括所述第一AI模型。
  27. 根据权利要求26所述的方法，其特征在于，在确定所述第二通信装置部署所述第二AI模型的情况下，所述方法还包括：
    确定第一组AI模型中的各个AI模型与所述第一组AI模型中的各个AI模型匹配，所述第一组AI模型包括所述第二AI模型。
  28. 根据权利要求26或27所述的方法，其特征在于，所述第一组AI模型中的各个AI模型满足：所述第一组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
  29. 根据权利要求26至28中任一项所述的方法，其特征在于，所述第二组AI模型中的各个AI模型满足：所述第二组AI模型中的各个AI模型输入信息相同时，输出信息相同或者输出信息的差异位于预设范围内。
  30. 一种通信方法，其特征在于，包括：
    发送请求信息，所述请求信息请求人工智能AI模型的标识；
    接收所述AI模型的标识。
  31. 根据权利要求30所述的方法，其特征在于，所述请求信息包括以下至少一项：所述AI模型所属的通信装置的类型、所述AI模型所属的通信装置的标识、所述AI模型所属的通信装置的厂商、所述AI模型的功能、所述AI模型适用的场景、所述AI模型的数据集、所述AI模型适用的通信参数、或，所述AI模型的版本。
  32. 一种通信方法，其特征在于，包括：
    接收请求信息，所述请求信息请求人工智能AI模型的标识；
    响应于请求信息，确定所述AI模型的标识；
    发送所述AI模型的标识。
  33. 根据权利要求32所述的方法，其特征在于，所述请求信息包括以下至少一项：所述AI模型所属的通信装置的类型、所述AI模型所属的通信装置的标识、所述AI模型所属的通信装置的厂商、所述AI模型的功能、所述AI模型适用的场景、所述AI模型的数据集、所述AI模型适用的通信参数、或，所述AI模型的版本。
  34. 一种通信方法，其特征在于，包括：
    获取人工智能AI模型的标识；
    发送注册信息，所述注册信息包括所述AI模型的标识，所述注册信息用于注册所述AI模型的标识。
  35. 根据权利要求34所述的方法，其特征在于，所述AI模型的标识指示以下至少一项：
    所述AI模型所属的通信装置的类型、所述AI模型所属的通信装置的标识、所述AI模型所属的通信装置的厂商、所述AI模型的功能、所述AI模型适用的场景、所述AI模型的数据集、所述AI模型适用的通信参数、或，所述AI模型的版本。
  36. 一种通信方法，其特征在于，包括：
    接收注册信息，所述注册信息包括人工智能AI模型的标识，所述注册信息用于注册所述AI模型的标识；
    响应于所述注册信息，保存所述AI模型的标识。
  37. 根据权利要求36所述的方法，其特征在于，所述AI模型的标识指示以下至少一项：
    所述AI模型所属的通信装置的类型、所述AI模型所属的通信装置的标识、所述AI模型所属的通信装置的厂商、所述AI模型的功能、所述AI模型适用的场景、所述AI模型的数据集、所述AI模型适用的通信参数、或，所述AI模型的版本。
  38. 一种通信装置，其特征在于，包括用于执行权利要求1至12中任一项所述的方法的模块或单元。
  39. 一种通信装置，其特征在于，包括用于执行权利要求13至25中任一项所述的方法的模块或单元。
  40. 一种通信装置，其特征在于，包括用于执行权利要求26至29中任一项所述的方法的模块或单元。
  41. 一种通信装置，其特征在于，包括用于执行权利要求30或31所述的方法的模块或单元。
  42. 一种通信装置，其特征在于，包括用于执行权利要求32或33所述的方法的模块或单元。
  43. 一种通信装置，其特征在于，包括用于执行权利要求34或35所述的方法的模块或单元。
  44. 一种通信装置，其特征在于，包括用于执行权利要求36或37所述的方法的模块或单元。
  45. 一种通信装置，其特征在于，包括处理器，所述处理器，用于执行存储器中存储的计算机程序或指令，以使得所述装置执行权利要求1至12中任一项所述的方法，或者，以使得所述装置执行权利要求13至25中任一项所述的方法，或者，以使得所述装置执行权利要求26至29中任一项所述的方法，或者，以使得所述装置执行权利要求30或31所述的方法，或者，以使得所述装置执行权利要求32或33所述的方法，或者，以使得所述装置执行权利要求34或35所述的方法，或者，以使得所述装置执行权利要求36或37所述的方法。
  46. 根据权利要求45所述的装置，其特征在于，所述装置还包括所述存储器和/或通信接口，所述通信接口与所述处理器耦合，
    所述通信接口，用于输入和/或输出信息。
  47. 根据权利要求38至46中任一项所述的装置，其特征在于，所述装置为以下任一项：芯片、芯片系统、或电路。
  48. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质上存储有计算机程序或指令，当所述计算机程序或指令在通信装置上运行时，使得所述通信装置执行如权利要求1至12中任一项所述的方法，或者，使得所述装置执行权利要求13至25中任一项所述的方法，或者，使得所述装置执行权利要求26至29中任一项所述的方法，或者，使得所述装置执行权利要求30或31所述的方法，或者，使得所述装置执行权利要求32或33所述的方法，或者，使得所述装置执行权利要求34或35所述的方法，或者，使得所述装置执行权利要求36或37所述的方法。
  49. 一种计算机程序产品，其特征在于，所述计算机程序产品包括用于执行如权利要求1至12中任一项所述的方法的计算机程序或指令，或者，所述计算机程序产品包括用于执行如权利要求13至25中任一项所述的方法的计算机程序或指令，或者，所述计算机程序产品包括用于执行如权利要求26至29中任一项所述的方法的计算机程序或指令，或者，所述计算机程序产品包括用于执行如权利要求30或31所述的方法的计算机程序或指令，或者，所述计算机程序产品包括用于执行如权利要求32或33所述的方法的计算机程序或指令，或者，所述计算机程序产品包括用于执行如权利要求34或35所述的方法的计算机程序或指令，或者，所述计算机程序产品包括用于执行如权利要求36或37所述的方法的计算机程序或指令。
  50. 一种芯片，其特征在于，所述芯片与存储器耦合，用于读取并执行所述存储器中存储的程序指令，以实现如权利要求1至12中任一项所述的方法，或者，以实现如权利要求13至25中任一项所述的方法，或者，以实现如权利要求26至29中任一项所述的方法，或者，以实现如权利要求30或31所述的方法，或者，以实现如权利要求32或33所述的方法，或者，以实现如权利要求34或35所述的方法，或者，以实现如权利要求36或37所述的方法。
  51. 一种通信系统，其特征在于，包括：如权利要求38所述的通信装置和如权利要求39所述的通信装置。
  52. 一种通信系统，其特征在于，包括：如权利要求41所述的通信装置和如权利要求42所述的通信装置。
  53. 一种通信系统，其特征在于，包括：如权利要求43所述的通信装置和如权利要求44所述的通信装置。
PCT/CN2023/119662 2022-09-29 2023-09-19 模型匹配的方法和通信装置 WO2024067245A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211197193.5 2022-09-29
CN202211197193.5A CN117786418A (zh) 2022-09-29 2022-09-29 模型匹配的方法和通信装置

Publications (1)

Publication Number Publication Date
WO2024067245A1 true WO2024067245A1 (zh) 2024-04-04

Family

ID=90393301

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/119662 WO2024067245A1 (zh) 2022-09-29 2023-09-19 模型匹配的方法和通信装置

Country Status (2)

Country Link
CN (1) CN117786418A (zh)
WO (1) WO2024067245A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111026436A (zh) * 2019-12-09 2020-04-17 支付宝(杭州)信息技术有限公司 模型联合训练方法及装置
CN111083722A (zh) * 2019-04-15 2020-04-28 中兴通讯股份有限公司 模型的推送、模型的请求方法及装置、存储介质
WO2021179165A1 (zh) * 2020-03-10 2021-09-16 Oppo广东移动通信有限公司 模型协调方法及装置
CN114978382A (zh) * 2021-02-22 2022-08-30 华为技术有限公司 基于空间划分的数据处理方法和通信装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111083722A (zh) * 2019-04-15 2020-04-28 中兴通讯股份有限公司 模型的推送、模型的请求方法及装置、存储介质
CN111026436A (zh) * 2019-12-09 2020-04-17 支付宝(杭州)信息技术有限公司 模型联合训练方法及装置
CN112925558A (zh) * 2019-12-09 2021-06-08 支付宝(杭州)信息技术有限公司 模型联合训练方法及装置
WO2021179165A1 (zh) * 2020-03-10 2021-09-16 Oppo广东移动通信有限公司 模型协调方法及装置
CN114978382A (zh) * 2021-02-22 2022-08-30 华为技术有限公司 基于空间划分的数据处理方法和通信装置

Also Published As

Publication number Publication date
CN117786418A (zh) 2024-03-29

Similar Documents

Publication Publication Date Title
US20230209390A1 (en) Intelligent Radio Access Network
US11765677B2 (en) Efficient PLMN selection upon authentication failure for each network slice in roaming network
US11206610B2 (en) Method and apparatus for mutually exclusive access to network slices in wireless communication system
WO2023036268A1 (zh) 一种通信方法及装置
US20230284119A1 (en) Access to snpn by using credential owned by entity separated from snpn, and support for f1 interface therefor
US20230262478A1 (en) Model configuration method and apparatus
WO2021197087A1 (zh) 通信优先级确定方法及装置
WO2023040971A1 (zh) 一种人工智能模型下载方法、装置及系统
WO2023104169A1 (zh) 一种无线网络中的人工智能ai模型训练方法及装置
WO2024067245A1 (zh) 模型匹配的方法和通信装置
WO2022206513A1 (zh) 模型处理的方法、通信装置和系统
KR20230139297A (ko) 연합 학습에 관련된 통신
KR102416342B1 (ko) 3차원 셀룰러 네트워크에서 간헐적 인식 연합 학습 방법
US20240073732A1 (en) Method and device for adjusting split point in wireless communication system
US20240054351A1 (en) Device and method for signal transmission in wireless communication system
WO2023044904A1 (zh) 通信方法、终端设备及网络设备
WO2024082274A1 (zh) Ai任务指示的方法、通信装置和系统
US20240107594A1 (en) User equipment pairing and cooperative machine learning inference result sharing
US20240179603A1 (en) Communication method and apparatus
WO2024067248A1 (zh) 一种获取训练数据集的方法和装置
WO2022135288A1 (zh) 信息处理的方法和装置
CN116886475B (zh) 一种信道估计方法、设备及系统
WO2023036280A1 (zh) 一种模型测试方法及装置
WO2023066346A1 (zh) 一种通信方法及装置
CN114342541B (zh) 随机接入的方法和装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23870477

Country of ref document: EP

Kind code of ref document: A1