WO2024149288A1 - Ai模型分发、接收方法、终端及网络侧设备 - Google Patents

Ai模型分发、接收方法、终端及网络侧设备 Download PDF

Info

Publication number
WO2024149288A1
WO2024149288A1 PCT/CN2024/071560 CN2024071560W WO2024149288A1 WO 2024149288 A1 WO2024149288 A1 WO 2024149288A1 CN 2024071560 W CN2024071560 W CN 2024071560W WO 2024149288 A1 WO2024149288 A1 WO 2024149288A1
Authority
WO
WIPO (PCT)
Prior art keywords
model
information
terminal
side device
capability
Prior art date
Legal status (The legal status 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 status listed.)
Ceased
Application number
PCT/CN2024/071560
Other languages
English (en)
French (fr)
Inventor
孙晓文
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Vivo Mobile Communication Co Ltd
Original Assignee
Vivo Mobile Communication Co Ltd
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 Vivo Mobile Communication Co Ltd filed Critical Vivo Mobile Communication Co Ltd
Publication of WO2024149288A1 publication Critical patent/WO2024149288A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
    • H04W16/22—Traffic simulation tools or models
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition

Definitions

  • the present application belongs to the field of communication technology, and specifically relates to an AI model distribution, receiving method, terminal and network side equipment.
  • ML machine learning
  • AI artificial intelligence
  • the terminal when the terminal performs image recognition or other media tasks, it can be processed through the AI model/ML model.
  • the processing may involve interaction with network-side devices, such as the network-side device sending the AI model to the terminal. Therefore, how to accurately distribute the AI model to the terminal is a problem that needs to be solved.
  • the embodiments of the present application provide an AI model distribution and receiving method, a terminal, and a network-side device, which can solve the problem of how to accurately distribute the AI model to the terminal.
  • an AI model distribution method comprising:
  • the network side device When triggering AI model distribution, the network side device sends information of the first AI model to each terminal based on the first information of at least one terminal; the first information of the terminal matches the first AI model corresponding to the terminal; the first information includes at least one of the following: capability information of the terminal, requirement information of the AI model and transmission method.
  • an AI model receiving method comprising:
  • the terminal receives information of a first AI model sent by a network side device, where the first AI model matches first information of the terminal, and the first information includes at least one of the following: capability information of the terminal, requirement information of the AI model, and a transmission mode;
  • the terminal performs service processing based on the first AI model.
  • an AI model distribution device including:
  • a sending module is used to send information of a first AI model to each terminal based on first information of at least one terminal when the AI model distribution is triggered; the first information of the terminal matches the first AI model corresponding to the terminal; the first information includes at least one of the following: capability information of the terminal, requirements of the AI model Information and transmission methods.
  • an AI model receiving device comprising:
  • a receiving module configured to receive information of a first AI model sent by a network side device, where the first AI model matches first information of the terminal, and the first information includes at least one of the following: capability information of the terminal, requirement information of the AI model, and a transmission mode;
  • a processing module is used to perform business processing based on the first AI model.
  • a terminal comprising a processor and a memory, wherein the memory stores a program or instruction that can be run on the processor, and when the program or instruction is executed by the processor, the steps of the method described in the second aspect are implemented.
  • a terminal comprising a processor and a communication interface; wherein the communication interface is used to receive information of a first AI model sent by a network side device, the first AI model matches the first information of the terminal, and the first information includes at least one of the following: capability information of the terminal, requirement information of the AI model, and a transmission method; the processor is used to perform business processing based on the first AI model.
  • a network side device which includes a processor and a memory, wherein the memory stores programs or instructions that can be run on the processor, and when the program or instructions are executed by the processor, the steps of the method described in the first aspect are implemented.
  • a network side device comprising a processor and a communication interface; wherein the communication interface is used to send information of a first AI model to each terminal based on first information of at least one terminal when AI model distribution is triggered; the first information of the terminal matches the first AI model corresponding to the terminal; the first information includes at least one of the following: capability information of the terminal, requirement information of the AI model and transmission method.
  • a communication system comprising: a terminal and a network side device, wherein the terminal can be used to execute the steps of the method described in the second aspect, and the network side device can be used to execute the steps of the method described in the first aspect.
  • a readable storage medium on which a program or instruction is stored.
  • the program or instruction is executed by a processor, the steps of the method described in the first aspect are implemented, or the steps of the method described in the second aspect are implemented.
  • a chip comprising a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the method described in the first aspect, or to implement the method described in the second aspect.
  • a computer program/program product is provided, wherein the computer program/program product is stored in a storage medium, and the computer program/program product is executed by at least one processor to implement the steps of the method described in the first aspect, or to implement the steps of the method described in the second aspect.
  • the network side device when the AI model distribution is triggered, sends the information of the first AI model to each terminal based on the first information of at least one terminal; since the first AI model is based on the first information of the terminal, The first information is determined by at least one of the terminal's capability information, the AI model's requirement information, and the transmission method, that is, the first information matches the first AI model corresponding to the terminal, so that the AI model can be distributed according to the actual terminal situation, and the accuracy of the AI model distribution is high.
  • FIG1 is a schematic diagram of a wireless communication system applicable to an embodiment of the present application.
  • FIG2 is a flow chart of an AI model distribution method provided in an embodiment of the present application.
  • FIG3 is a schematic diagram of an AI model distribution scenario provided in an embodiment of the present application.
  • FIG4 is one of the interactive flow diagrams of the AI model distribution method provided in an embodiment of the present application.
  • FIG5 is a second interactive flow diagram of the AI model distribution method provided in an embodiment of the present application.
  • FIG6 is a schematic diagram of a flow chart of an AI model receiving method provided in an embodiment of the present application.
  • FIG7 is a schematic diagram of the structure of an AI model distribution device provided in an embodiment of the present application.
  • FIG8 is a schematic diagram of the structure of an AI model receiving device provided in an embodiment of the present application.
  • FIG9 is a schematic diagram of the structure of a communication device provided in an embodiment of the present application.
  • FIG10 is a schematic diagram of the structure of a terminal provided in an embodiment of the present application.
  • FIG11 is a schematic diagram of a structure of a network side device according to an embodiment of the present application.
  • FIG. 12 is a second schematic diagram of the structure of the network side device according to an embodiment of the present application.
  • first, second, etc. in the specification and claims of the present application are used to distinguish similar objects, but not to describe a specific order or sequence. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by “first” and “second” are generally of the same type, and the number of objects is not limited.
  • the first object can be one or more.
  • “and/or” in the specification and claims represents at least one of the connected objects, and the character “/” generally indicates that the objects associated with each other are in an "or” relationship.
  • the term “indication” in the specification and claims of the present application can be either an explicit indication or an implicit indication.
  • an explicit indication can be understood as the sender explicitly notifying the receiver of the operation or request result to be performed in the indication sent;
  • an implicit indication can be understood as the receiver making a judgment based on the indication sent by the sender and determining the operation or request result to be performed based on the judgment result.
  • LTE Long Term Evolution
  • LTE-A Long Term Evolution
  • CDMA Code Division Multiple Access
  • TDMA Time Division Multiple Access
  • FDMA Frequency Division Multiple Access
  • OFDMA Orthogonal Frequency Division Multiple Access
  • SC-FDMA Single-carrier Frequency Division Multiple Access
  • NR New Radio
  • 6G 6th Generation
  • FIG1 is a schematic diagram of a wireless communication system applicable to an embodiment of the present application.
  • the wireless communication system shown in FIG1 includes a terminal 11 and a network side device 12.
  • the terminal 11 may be a mobile phone, a tablet computer (Tablet Personal Computer), a laptop computer (Laptop Computer) or a notebook computer, a personal digital assistant (Personal Digital Assistant, PDA), a handheld computer, a netbook, an ultra-mobile personal computer (ultra-mobile personal computer, UMPC), a mobile Internet device (Mobile Internet Device, MID), an augmented reality (augmented reality, AR)/virtual reality, or a network side device 12.
  • VR Virtual reality
  • robots wearable devices
  • VUE vehicle-mounted equipment
  • PUE pedestrian terminals
  • smart homes home appliances with wireless communication functions, such as refrigerators, televisions, washing machines or furniture, etc.
  • game consoles personal computers (personal computers, PCs), teller machines or self-service machines and other terminal-side devices
  • wearable devices include: smart watches, smart bracelets, smart headphones, smart glasses, smart jewelry (smart bracelets, smart bracelets, smart rings, smart necklaces, smart anklets, smart anklets, etc.), smart wristbands, smart clothing, etc.
  • terminal 11 can also be a chip in the terminal, such as a modem chip, a system-on-chip (System on Chip, SoC). It should be noted that the specific type of terminal 11 is not limited in the embodiment of the present application.
  • the network side device 12 may include an access network device or a core network device, wherein the access network device may also be referred to as a radio access network device, a radio access network (RAN), a radio access network function or a radio access network unit.
  • the access network device may include a base station, a WLAN access point or a WiFi node, etc.
  • the base station may be referred to as a node B, an evolved node B (eNB), an access point, a base transceiver station (BTS), a radio base station, a radio transceiver, a basic service set (BSS), an extended service set (ESS), a home node B, a home evolved node B, a transmitting and receiving point (TRP) or some other suitable term in the field, as long as the same technical effect is achieved, the base station is not limited to a specific technical vocabulary, it should be noted that in the embodiment of the present application, only the base station in the NR system is used as an example for introduction, and the specific type of the base station is not limited.
  • the core network equipment may include but is not limited to at least one of the following: core network node, core network function, mobility management entity (Mobility Management Entity, MME), access mobility management function (Access and Mobility Management Function, AMF), session management function (Session Management Function, SMF), user plane function (User Plane Function, UPF), policy control function (Policy Control Function, PCF), policy and charging rules function unit (Policy and Charging Rules Function, PCRF), edge application service discovery function (Edge Application Server Discovery Function, EASDF), unified data management (Unified Data Management, UDM), unified data storage (Unified Data Repository, UDR), Home Subscriber Server (HSS), centralized network configuration (CNC), Network Repository Function (NRF), Network Exposure Function (NEF), Local NEF (L-NEF), Binding Support Function (BSF), Application Function (AF), Location Management Function (LMF), Enhanced Serving Mobile Location Centre (E-SMLC), Network data analytics function (NWDAF), etc.
  • MME mobility management entity
  • FIG2 is a flow chart of an AI model distribution method provided in an embodiment of the present application. As shown in FIG2 , the method includes:
  • Step 101 When AI model distribution is triggered, the network side device sends information of the first AI model to each terminal based on the first information of at least one terminal; the first information of the terminal matches the first AI model corresponding to the terminal; the first information includes at least one of the following: capability information of the terminal, requirement information of the AI model, and transmission method.
  • the network side device in the embodiment of the present application determines the first AI model corresponding to the terminal based on the first information of the terminal, and sends the information of the first AI model to the terminal.
  • the AI model is also called an AI unit, which can be an AI model, an AI module, or an operation rule with AI/machine learning (ML) functions.
  • AI unit can be an AI model, an AI module, or an operation rule with AI/machine learning (ML) functions.
  • the terminal capability information includes at least one of the following: computing power, storage capacity, computing power usage, current remaining computing power, currently available computing power, storage capacity usage, current remaining storage capacity, currently available storage capacity, and processable business information.
  • the service information includes, for example, a service identifier, a service type, and the like, such as a computing type and a storage type.
  • the demand information of the AI model includes at least one of the following: service demand information and description information of the AI model.
  • the description information of the AI model includes at least one of the following: identification, functional description information, topology structure, parameter information, and used resource information.
  • the transmission methods include, for example: through the 5G system (5G system, 5GS) network, or directly through the application layer (i.e., 5GS unaware (over 5GS)).
  • the terminal receives information of the first AI model and can perform business processing based on the first AI model, such as image recognition.
  • the transmission of the AI model can be triggered by the communication between the terminal and the network side device, such as the terminal sending a request.
  • the network side sends information of the first AI model to each terminal based on the first information of at least one terminal. Since the first AI model is determined based on the first information of the terminal, such as at least one of the capability information of the terminal, the requirement information of the AI model and the transmission method, that is, the first information matches the first AI model corresponding to the terminal, the AI model can be distributed according to the actual terminal situation, and the accuracy of the AI model distribution is relatively high.
  • step 101 the following steps are further included:
  • the network side device receives a request message from at least one terminal, where the request message is used to request the network side device to send information of a first AI model, and the request message includes: first information of the terminal.
  • information related to the required AI model can be exchanged and negotiated between the network side device and the terminal.
  • the capability information of the terminal including but not limited to computing power, storage capacity, current computing power usage, etc.
  • the demand information of the AI model for example, service demand information, description information of the AI model, service demand information such as the business information and type that the required AI model can process, etc.
  • the above information can be used to select a suitable AI model for the service required by the terminal.
  • the method further includes:
  • the network side device establishes a first session connection with the terminal; the first session connection is used to transmit information of the first AI model.
  • the method further comprises:
  • the network-side device receives training result data sent by at least one terminal, where the training result data is obtained by training the first AI model by the terminal;
  • the network side device performs AI model training based on the training result data sent by at least one terminal.
  • the terminal can train the first AI model based on the training data, and the training data is, for example, the training sample data obtained by the terminal. Further, the training result data obtained by the training can be sent to the network side device, and the network side device can perform AI model training based on the training result data.
  • the training result data can include, for example, description information of the trained AI model, such as: identification, function description information, topology structure, parameter information, used resource information, etc.
  • the method further comprises:
  • the network side device establishes a second session connection with the terminal, and the second session connection is used to transmit the training result data.
  • the AI model is trained through the terminal, and the training result data is fed back to the network side device, so that the network side device can further modify the model based on the feedback from the terminal, thereby meeting the needs of the terminal.
  • the network side device when an AI model update is triggered, sends information of an updated second AI model to at least one terminal based on second information; the second information includes at least one of the following: capability update information and model update information of each terminal.
  • the network side device sends information of an updated second model to at least one terminal based on the second information, for example, selecting an updated AI model that matches the terminal based on the terminal's latest capability information, the latest service requirement information, model update information, etc.
  • the model update information is, for example, an AI model trained by a new task (such as a federated learning task) updated by the network side device based on training result data of at least one terminal.
  • the capability update information includes update information of at least one of the following: computing capability, storage capability, usage of computing capability, current remaining computing capability, current available computing capability, usage of storage capability, current remaining storage capability, current available storage capability, and processable business information;
  • the model update information includes update information of at least one of the following: identification, function description information, topology structure, parameter information, and used resource information.
  • Functional description information includes information used to explain the functions of the AI model, such as the names and types of services that can be processed;
  • topological structure refers to the structure of the network model used by the AI model, such as the topological structure of the neural network model, the topological structure of the deep learning model, etc.
  • parameter information includes, for example, the weight information of the AI model, the loss function information used, the activation function, the convolution kernel size, the number of convolution layers, etc.
  • the resource information used includes, for example: the memory size occupied at runtime, the storage space size, etc.
  • the model update information includes updated information of the identification, such as the identification before the update and the identification after the update; or, if the function of the model is updated, the model update information includes updated information of the function description information, such as the updated function description information, topology structure, parameter information, used resource information, and the like.
  • the information of the second AI model includes at least one of the following: identification, function description information, topology structure, parameter information, and used resource information.
  • the second AI model is obtained by training the first AI model based on training result data of at least one terminal; the training result data is obtained by training the first AI model by the terminal.
  • the AI model update is triggered based on third information
  • the third information includes at least one of the following: terminal capability update indication information, network side device update AI model indication information; the terminal capability update indication information is used to indicate at least one of the following: the terminal's computing power utilization rate is greater than or equal to the first threshold, the terminal's computing resource utilization rate is greater than or equal to the second threshold, the terminal's load is greater than or equal to the third threshold, and the terminal's resources are less than or equal to the fourth threshold.
  • the triggering factor for AI model update may be at least one of the following:
  • the current computing power of the terminal changes (for example, the usage of memory, hard disk, etc. changes), and the computing power cannot meet the requirements of the current AI model or far exceeds the requirements of the current AI model;
  • the network side device when the AI model update is triggered, sends the updated second AI model information to each terminal based on the second information of at least one terminal; since the second AI model is determined based on the second information of the terminal, such as at least one of the terminal's capability update information, the AI model's demand update information, and the service demand update information, the AI model can be distributed according to the actual terminal situation, and the accuracy of the AI model distribution is relatively high.
  • the federated learning engine receives part of the training model from the AI model repository, which is passed to the AI model delivery function to be delivered to multiple terminals through 5GS.
  • Training result data from multiple terminals is also received by the federated learning engine through 5GS and then aggregated for continuous training of AI models.
  • Updates to the AI model are transmitted to the terminal during the update process.
  • Updates to the AI model e.g., in topology, weights, and other parameters or other information
  • AI model data is received by the AI model access function through 5GS, and then the data is passed to the AI training engine.
  • the AI training engine in the terminal uses local device data as input training data to train the AI model.
  • the training results (for example, in the form of descriptive information of the updated AI model) are delivered to the network side through the training results delivery function.
  • the network side device may be an access network device and/or a core network device, and the above-mentioned functional modules may be deployed in one or more network side devices.
  • the method includes:
  • Step 1 The terminal APP and the network APP communicate to trigger the transmission of the AI model
  • Step 2 The terminal APP interacts with the network APP and selects the AI model
  • Step 3 The network APP identifies the selected AI model
  • Step 4 The AI model delivery function and the AI model access function establish a session connection for AI model delivery;
  • Step 5 The AI model access function receives the AI model
  • Step 6 The AI model access function transfers the AI model to the AI training engine
  • Step 7 On the terminal, the training data is transmitted to the AI training engine;
  • Step 8 The AI training engine performs model training
  • Step 9 The training result delivery function establishes a session connection with the federated learning engine for training result delivery;
  • Step 10 The training results are transferred to the federated learning engine.
  • the method includes:
  • Step 0 Triggering the AI model update. Triggering factors include but are not limited to the following:
  • the current computing power of the terminal changes (for example, the usage of memory, hard disk, etc. changes);
  • the federated learning engine updates the AI model.
  • Step 1 The terminal APP and the network APP communicate to trigger the update and transmission of the AI model
  • Step 2 The terminal APP interacts with the network APP and selects the updated AI model
  • Step 3 The network APP identifies the updated AI model that has been selected
  • Step 4 The AI model delivery function and the AI model access function establish a session connection for AI model delivery;
  • Step 5 The AI model access function receives the updated AI model
  • Step 6 The AI model access function passes the updated AI model to the AI training engine
  • Step 7 On the terminal, the training input data is passed to the AI training engine
  • Step 8 The AI training engine performs model training
  • Step 9 The training result delivery function establishes a session connection with the federated learning engine for training result delivery;
  • Step 10 The training results are transferred to the federated learning engine.
  • FIG6 is a flow chart of an AI model receiving method provided in an embodiment of the present application. As shown in FIG6 , the method includes:
  • Step 201 The terminal receives information of a first AI model sent by a network-side device.
  • the first AI model matches first information of the terminal.
  • the first information includes at least one of the following: capability information of the terminal, requirement information of the AI model, and a transmission method.
  • Step 202 The terminal performs service processing based on the first AI model.
  • the method further comprises:
  • the terminal trains the first AI model based on the training data to obtain training result data
  • the terminal sends the training result data to the network side device; the training result data is used by the network side device to perform AI model training.
  • the terminal capability information includes at least one of the following: computing power, storage capacity, computing power usage, current remaining computing power, currently available computing power, storage capacity usage, current remaining storage capacity, currently available storage capacity, and processable business information.
  • the demand information of the AI model includes at least one of the following: service demand information, description information of the AI model; the description information of the AI model includes: identification, function description information, topology structure, parameter information, and used resource information.
  • the method further comprises:
  • the terminal receives information of an updated second AI model sent by the network side device, where the second AI model is obtained based on second information; the second information includes at least one of the following: capability update information, service requirement update information, and model update information of each of the terminals.
  • the information of the updated second AI model is sent when an AI model update is triggered, and the AI model update is triggered based on third information
  • the third information includes at least one of the following: capability update indication information of the terminal, indication information of the network side device updating the AI model; the capability update indication information of the terminal is used to indicate at least one of the following: the computing power utilization rate of the terminal is greater than or equal to the first threshold, the computing resource utilization rate of the terminal is greater than or equal to the second threshold, the load of the terminal is greater than or equal to the third threshold, and the resources of the terminal are less than or equal to the fourth threshold.
  • the second AI model is obtained by training the first AI model based on training result data of at least one of the terminals; the training result data is obtained by training the first AI model by the terminal.
  • the information of the second AI model includes at least one of the following: identification, function description information, topology structure, parameter information, and used resource information.
  • the training result data includes description information of the updated AI model.
  • the method before the terminal receives the information of the first AI model sent by the network side device, the method further includes:
  • the terminal sends a request message to the network side device, where the request message is used to request the network side device to send information of the first AI model, and the request message includes: first information of the terminal.
  • the method before the terminal receives the information of the first AI model sent by the network side device, the method further includes:
  • the terminal establishes a first session connection with the network side device; the first session connection is used to transmit information of the first AI model.
  • the method further comprises:
  • the terminal establishes a second session connection with the network side device, where the second session connection is used to transmit training result data.
  • the AI model distribution method provided in the embodiment of the present application may be executed by an AI model distribution device.
  • the AI model receiving method provided in the embodiment of the present application may be executed by an AI model receiving device.
  • the AI model distribution method executed by an AI model distribution device and the AI model receiving method executed by an AI model receiving device are taken as examples to illustrate the AI model distribution device and the AI model receiving device provided in the embodiment of the present application.
  • FIG. 7 is a schematic diagram of the structure of an AI model distribution device provided in an embodiment of the present application. As shown in FIG. 7 , the AI model distribution device is applied to a network side device, including:
  • the sending module 210 is used to send the information of the first AI model to each terminal based on the first information of at least one terminal when the AI model distribution is triggered; the first information of the terminal matches the first AI model corresponding to the terminal; the first information includes at least one of the following: capability information of the terminal, requirement information of the AI model and transmission method.
  • the capability information of the terminal includes at least one of the following: computing capability, storage capability, usage of computing capability, current remaining computing capability, current available computing capability, usage of storage capability, current remaining storage capability, current available storage capability, and processable service information;
  • the demand information of the AI model includes at least one of the following: service demand information and description information of the AI model.
  • the description information of the AI model includes at least one of the following: identification, functional description information, topological structure, parameter information, and used resource information.
  • the device further comprises:
  • a receiving module used to receive training result data sent by at least one of the terminals, where the training result data is obtained by the terminal training the first AI model;
  • a processing module is used to perform AI model training based on training result data sent by at least one of the terminals.
  • the sending module 210 is also used to send information of an updated second AI model to at least one of the terminals based on second information when an AI model update is triggered; the second information includes at least one of the following: capability update information, service requirement update information, and model update information of each of the terminals.
  • the capability update information includes update information of at least one of the following: computing capability, storage capability, The amount of computing power used, the current remaining computing power, the current available computing power, the amount of storage power used, the current remaining storage power, the current available storage power, and the processable business information; and/or,
  • the model update information includes at least one of the following update information: identification, function description information, topology structure, parameter information, and used resource information.
  • the AI model update is triggered based on third information
  • the third information includes at least one of the following: capability update indication information of the terminal, indication information of the network side device updating the AI model; the capability update indication information of the terminal is used to indicate at least one of the following: the computing power utilization rate of the terminal is greater than or equal to the first threshold, the computing resource utilization rate of the terminal is greater than or equal to the second threshold, the load of the terminal is greater than or equal to the third threshold, and the resources of the terminal are less than or equal to the fourth threshold.
  • the second AI model is obtained by training the first AI model based on training result data of at least one of the terminals; the training result data is obtained by training the first AI model by the terminal.
  • the information of the second AI model includes at least one of the following: identification, function description information, topology structure, parameter information, and used resource information.
  • the training result data includes description information of the updated AI model.
  • the receiving module is also used to receive a request message from at least one of the terminals, where the request message is used to request the network side device to send information of the first AI model, and the request message includes: the first information of the terminal.
  • processing module is further configured to:
  • the first session connection is used to transmit information of a first AI model.
  • processing module is further configured to:
  • a second session connection is established with the terminal, where the second session connection is used to transmit training result data.
  • the device of this embodiment can be used to execute the method of any of the embodiments in the aforementioned network side device method embodiments. Its specific implementation process and technical effects are the same as those in the network side device method embodiments. For details, please refer to the detailed introduction in the network side device method embodiments, which will not be repeated here.
  • FIG8 is a schematic diagram of the structure of an AI model receiving device provided in an embodiment of the present application. As shown in FIG8 , the AI model receiving device is applied to a terminal and includes:
  • a receiving module 310 is configured to receive information of a first AI model sent by a network side device, where the first AI model matches first information of the terminal, and the first information includes at least one of the following: capability information of the terminal, requirement information of the AI model, and a transmission mode;
  • the processing module 320 is used to perform business processing based on the first AI model.
  • processing module 320 is further configured to:
  • the device also includes: a sending module, which is used to send the training result data to the network side device;
  • the training result data is used for the network side device to perform AI model training.
  • the terminal capability information includes at least one of the following: computing power, storage capacity, computing power usage, current remaining computing power, currently available computing power, storage capacity usage, current remaining storage capacity, currently available storage capacity, and processable business information.
  • the demand information of the AI model includes at least one of the following: service demand information, description information of the AI model; the description information of the AI model includes: identification, function description information, topology structure, parameter information, and used resource information.
  • the receiving module 310 is further configured to:
  • the capability update information includes update information of at least one of the following: computing capacity, storage capacity, computing capacity usage, current remaining computing capacity, currently available computing capacity, storage capacity usage, current remaining storage capacity, currently available storage capacity, processable business information; and/or
  • the model update information includes at least one of the following update information: identification, function description information, topology structure, parameter information, and used resource information.
  • the information of the updated second AI model is sent when an AI model update is triggered, and the AI model update is triggered based on third information
  • the third information includes at least one of the following: capability update indication information of the terminal, indication information of the network side device updating the AI model; the capability update indication information of the terminal is used to indicate at least one of the following: the computing power utilization rate of the terminal is greater than or equal to the first threshold, the computing resource utilization rate of the terminal is greater than or equal to the second threshold, the load of the terminal is greater than or equal to the third threshold, and the resources of the terminal are less than or equal to the fourth threshold.
  • the second AI model is obtained by training the first AI model based on training result data of at least one of the terminals; the training result data is obtained by training the first AI model by the terminal.
  • the information of the second AI model includes at least one of the following: identification, function description information, topology structure, parameter information, and used resource information.
  • the training result data includes description information of the updated AI model.
  • the sending module is further used for:
  • a request message is sent to the network side device, where the request message is used to request the network side device to send information of the first AI model, and the request message includes: first information of the terminal.
  • the processing module 320 is also used to establish a first session connection with the network side device; the first session connection is used to transmit information of the first AI model.
  • the processing module 320 is further used to establish a second session connection with the network side device, and the second session connection is used to transmit training result data.
  • the device of this embodiment can be used to execute the method of any of the embodiments in the aforementioned terminal side method embodiments. Its specific implementation process and technical effects are the same as those in the terminal side method embodiments. For details, please refer to the detailed introduction in the terminal side method embodiments, which will not be repeated here.
  • the AI model distribution device and AI model receiving device in the embodiments of the present application may be electronic devices, such as electronic devices with an operating system, or components in electronic devices, such as integrated circuits or chips.
  • the electronic device may be a terminal, or may be other devices other than a terminal.
  • the terminal may include but is not limited to the types of terminal 11 listed above, and other devices may be servers, network attached storage (NAS), etc., which are not specifically limited in the embodiments of the present application.
  • the AI model distribution device and AI model receiving device provided in the embodiments of the present application can implement the various processes implemented by the method embodiments of Figures 2 to 6 and achieve the same technical effects. To avoid repetition, they will not be repeated here.
  • FIG9 is a schematic diagram of the structure of a communication device provided in an embodiment of the present application.
  • the communication device 900 includes a processor 901 and a memory 902.
  • the memory 902 stores a program or instruction that can be run on the processor 901.
  • the program or instruction is executed by the processor 901 to implement the various steps of the above-mentioned AI model receiving method embodiment, and can achieve the same technical effect.
  • the communication device 900 is a network side device
  • the program or instruction is executed by the processor 901 to implement the various steps of the above-mentioned AI model distribution method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.
  • the embodiment of the present application also provides a terminal, including a processor and a communication interface, the communication interface is used to receive information of a first AI model sent by a network side device, the first AI model matches the first information of the terminal, and the first information includes at least one of the following: capability information of the terminal, requirement information of the AI model, and transmission mode; the processor is used to perform business processing based on the first AI model.
  • This terminal embodiment corresponds to the above-mentioned terminal side method embodiment, and each implementation process and implementation method of the above-mentioned method embodiment can be applied to this terminal embodiment and can achieve the same technical effect.
  • Figure 10 is a structural diagram of a terminal provided in an embodiment of the present application.
  • the terminal 1000 includes but is not limited to: a radio frequency unit 1001, a network module 1002, an audio output unit 1003, an input unit 1004, a sensor 1005, a display unit 1006, a user input unit 1007, an interface unit 1008, a memory 1009, and at least some of the components in the processor 1010.
  • the terminal 1000 can also include a power supply (such as a battery) for supplying power to each component, and the power supply can be logically connected to the processor 1010 through a power management system, so as to implement functions such as charging, discharging, and power consumption management through the power management system.
  • a power supply such as a battery
  • the terminal structure shown in FIG10 does not constitute a limitation on the terminal, and the terminal can include more or fewer components than shown in the figure, or combine certain components, or arrange components differently, which will not be described in detail here.
  • the input unit 1004 may include a graphics processing unit (GPU) 10041 and a microphone 10042.
  • the graphics processor 10041 processes the image data of a static picture or video obtained by an image capture device (such as a camera) in a video capture mode or an image capture mode.
  • the display unit 1006 may include a display panel 10061, which may be a liquid crystal display, an organic light emitting diode, or a display panel.
  • the display panel 10061 is configured in the form of a display panel 10061, ...
  • the RF unit 1001 after receiving the downlink data from the network side device, the RF unit 1001 can transmit it to the processor 1010 for processing; in addition, the RF unit 1001 can send the uplink data to the network side device.
  • the RF unit 1001 includes but is not limited to an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, etc.
  • the memory 1009 can be used to store software programs or instructions and various data.
  • the memory 1009 may mainly include a first storage area for storing programs or instructions and a second storage area for storing data, wherein the first storage area may store an operating system, an application program or instruction required for at least one function (such as a sound playback function, an image playback function, etc.), etc.
  • the memory 1009 may include a volatile memory or a non-volatile memory, or the memory 1009 may include a transient and non-transient 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.
  • ROM read-only memory
  • PROM programmable read-only memory
  • EPROM erasable programmable read-only memory
  • EEPROM electrically erasable programmable read-only memory
  • flash memory a flash memory.
  • the volatile memory may be a random access memory (Random Access Memory,
  • the memory 1009 in the embodiment of the present application includes but is not limited to these and any other suitable types of memory.
  • the processor 1010 may include one or more processing units; optionally, the processor 1010 may integrate an application processor and a modem processor, wherein the application processor mainly processes operations related to an operating system, a user interface, and application programs or instructions, and the modem processor mainly processes wireless communication signals, such as a baseband processor. It is understandable that the modem processor may not be integrated into the processor 1010.
  • the radio frequency unit 1001 is configured to receive information of a first AI model sent by a network side device, where the first AI model matches first information of the terminal, and the first information includes at least one of the following: capability information of the terminal, requirement information of the AI model, and a transmission mode;
  • Processor 1010 is used to perform business processing based on the first AI model.
  • processor 1010 is further configured to:
  • the radio frequency unit 1001 is further configured to send the training result data to the network side device; the training result The data is used for AI model training on the network side device.
  • the terminal capability information includes at least one of the following: computing power, storage capacity, computing power usage, current remaining computing power, currently available computing power, storage capacity usage, current remaining storage capacity, currently available storage capacity, and processable business information.
  • the demand information of the AI model includes at least one of the following: service demand information, description information of the AI model; the description information of the AI model includes: identification, function description information, topology structure, parameter information, and used resource information.
  • the radio frequency unit 1001 is further configured to:
  • the capability update information includes update information of at least one of the following: computing capacity, storage capacity, computing capacity usage, current remaining computing capacity, currently available computing capacity, storage capacity usage, current remaining storage capacity, currently available storage capacity, processable business information; and/or
  • the model update information includes at least one of the following update information: identification, function description information, topology structure, parameter information, and used resource information.
  • the information of the updated second AI model is sent when an AI model update is triggered, and the AI model update is triggered based on third information
  • the third information includes at least one of the following: capability update indication information of the terminal, indication information of the network side device updating the AI model; the capability update indication information of the terminal is used to indicate at least one of the following: the computing power utilization rate of the terminal is greater than or equal to the first threshold, the computing resource utilization rate of the terminal is greater than or equal to the second threshold, the load of the terminal is greater than or equal to the third threshold, and the resources of the terminal are less than or equal to the fourth threshold.
  • the second AI model is obtained by training the first AI model based on training result data of at least one of the terminals; the training result data is obtained by training the first AI model by the terminal.
  • the information of the second AI model includes at least one of the following: identification, function description information, topology structure, parameter information, and used resource information.
  • the training result data includes description information of the updated AI model.
  • the radio frequency unit 1001 is further configured to:
  • a request message is sent to the network side device, where the request message is used to request the network side device to send information of the first AI model, and the request message includes: first information of the terminal.
  • the processor 1010 is further used to establish a first session connection with the network side device; the first session connection is used to transmit information of the first AI model.
  • the processor 1010 is further configured to establish a second session connection with the network side device, where the second session connection is used to transmit training result data.
  • the embodiment of the present application also provides a network side device, including a processor and a communication interface, the communication interface is used to send information of a first AI model to each terminal based on the first information of at least one terminal when the AI model distribution is triggered; the first information of the terminal matches the first AI model corresponding to the terminal; the first information includes at least one of the following: capability information of the terminal, requirement information of the AI model and transmission method.
  • This network side device embodiment corresponds to the above-mentioned network side device method embodiment, and each implementation process and implementation method of the above-mentioned method embodiment can be applied to this network side device embodiment, and can achieve the same technical effect.
  • FIG11 is a schematic diagram of the structure of a network side device provided in an embodiment of the present application.
  • the network side device 1100 includes: an antenna 1101, a radio frequency device 1102, a baseband device 1103, a processor 1104, and a memory 1105.
  • the antenna 1101 is connected to the radio frequency device 1102.
  • the radio frequency device 1102 receives information through the antenna 1101 and sends the received information to the baseband device 1103 for processing.
  • the baseband device 1103 processes the information to be sent and sends it to the radio frequency device 1102.
  • the radio frequency device 1102 processes the received information and sends it out through the antenna 1101.
  • the method executed by the network-side device in the above embodiment may be implemented in the baseband device 1103, which includes a baseband processor.
  • the baseband device 1103 may include, for example, at least one baseband board, on which a plurality of chips are arranged, as shown in FIG11 , wherein one of the chips is, for example, a baseband processor, which is connected to the memory 1105 through a bus interface to call the program in the memory 1105 and execute the network device operations shown in the above method embodiment.
  • the network side device may also include a network interface 1106, which is, for example, a common public radio interface (CPRI).
  • a network interface 1106, which is, for example, a common public radio interface (CPRI).
  • CPRI common public radio interface
  • the network side device 1100 of the embodiment of the present application also includes: instructions or programs stored in the memory 1105 and executable on the processor y04.
  • the processor 1104 calls the instructions or programs in the memory 1105 to execute the AI model distribution method as described above and achieve the same technical effect. To avoid repetition, it will not be repeated here.
  • FIG12 is a schematic diagram of the structure of a network side device provided in an embodiment of the present application.
  • the network side device 1200 includes: a processor 1201, a network interface 1202, and a memory 1203.
  • the network interface 1202 is, for example, a common public radio interface (CPRI).
  • CPRI common public radio interface
  • the network side device 1200 of the embodiment of the present application also includes: instructions or programs stored in the memory 1203 and executable on the processor 1201.
  • the processor 1201 calls the instructions or programs in the memory 1203 to execute the method executed by each module shown in Figure 7 and achieves the same technical effect. To avoid repetition, it will not be repeated here.
  • An embodiment of the present application also provides a communication system, including: a terminal and a network side device, wherein the terminal can be used to execute the steps of the AI model receiving method as described above, and the network side device can be used to execute the steps of the AI model distribution method as described above.
  • An embodiment of the present application also provides a readable storage medium, which may be volatile or non-volatile, and stores a program or instruction.
  • a program or instruction When the program or instruction is executed by a processor, the various processes of the above-mentioned AI model distribution and AI model receiving method embodiments are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.
  • the processor is the processor in the terminal described in the above embodiment.
  • the readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.
  • An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned AI model distribution and AI model receiving method embodiments, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.
  • the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.
  • the embodiments of the present application further provide a computer program/program product, which is stored in a storage medium.
  • the computer program/program product is executed by at least one processor to implement the various processes of the above-mentioned AI model distribution and AI model receiving method embodiments, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.
  • the technical solution of the present application can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal (which can be a mobile phone, computer, server, air conditioner, or network equipment, etc.) to execute the methods described in each embodiment of the present application.
  • a storage medium such as ROM/RAM, magnetic disk, optical disk
  • a terminal which can be a mobile phone, computer, server, air conditioner, or network equipment, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本申请公开了一种AI模型分发、接收方法、终端及网络侧设备，属于通信技术领域，本申请实施例的AI模型分发方法包括：在触发AI模型分发的情况下，网络侧设备基于至少一个终端的第一信息，向各个所述终端发送第一AI模型的信息；所述终端的第一信息与所述终端对应的第一AI模型匹配；所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式。

Description

AI模型分发、接收方法、终端及网络侧设备
相关申请的交叉引用
本申请主张在2023年01月12日在中国提交的申请号为202310084335.5的中国专利的优先权，其全部内容通过引用包含于此。
技术领域
本申请属于通信技术领域，具体涉及一种AI模型分发、接收方法、终端及网络侧设备。
背景技术
随着科学技术的发展，机器学习(machine learning，ML)/人工智能(Artificial Intelligence，AI)得到越来越广泛的应用。目前，ML模型/AI模型在无线通信网络中承担着越来越多的任务。
例如，在终端进行图像识别或其它媒体任务时，可以通过AI模型/ML模型进行处理，但是，由于终端受到处理能力、存储能力等的限制，因此在处理过程中可能涉及和网络侧设备的交互，例如网络侧设备向终端发送AI模型。因此，如何准确地对终端进行AI模型分发是需要解决的问题。
发明内容
本申请实施例提供一种AI模型分发、接收方法、终端及网络侧设备，能够解决如何准确地对终端进行AI模型分发的问题。
第一方面，提供了一种AI模型分发方法，包括：
在触发AI模型分发的情况下，网络侧设备基于至少一个终端的第一信息，向各个所述终端发送第一AI模型的信息；所述终端的第一信息与所述终端对应的第一AI模型匹配；所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式。
第二方面，提供了一种AI模型接收方法，包括：
终端接收网络侧设备发送的第一AI模型的信息，所述第一AI模型与所述终端的第一信息匹配，所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式；
所述终端基于所述第一AI模型进行业务处理。
第三方面，提供了一种AI模型分发装置，包括：
发送模块，用于在触发AI模型分发的情况下，基于至少一个终端的第一信息，向各个所述终端发送第一AI模型的信息；所述终端的第一信息与所述终端对应的第一AI模型匹配；所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求 信息和传输方式。
第四方面，提供了一种AI模型接收装置，该装置包括：
接收模块，用于接收网络侧设备发送的第一AI模型的信息，所述第一AI模型与所述终端的第一信息匹配，所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式；
处理模块，用于基于所述第一AI模型进行业务处理。
第五方面，提供了一种终端，该终端包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如第二方面所述的方法的步骤。
第六方面，提供了一种终端，包括处理器及通信接口；其中，所述通信接口用于接收网络侧设备发送的第一AI模型的信息，所述第一AI模型与所述终端的第一信息匹配，所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式；所述处理器用于基于所述第一AI模型进行业务处理。
第七方面，提供了一种网络侧设备，该网络侧设备包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如第一方面所述的方法的步骤。
第八方面，提供了一种网络侧设备，包括处理器及通信接口；其中，所述通信接口用于在触发AI模型分发的情况下，基于至少一个终端的第一信息，向各个所述终端发送第一AI模型的信息；所述终端的第一信息与所述终端对应的第一AI模型匹配；所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式。
第九方面，提供了一种通信系统，包括：终端及网络侧设备，所述终端可用于执行如第二方面所述的方法的步骤，所述网络侧设备可用于执行如第一方面所述的方法的步骤。
第十方面，提供了一种可读存储介质，所述可读存储介质上存储程序或指令，所述程序或指令被处理器执行时实现如第一方面所述的方法的步骤，或者实现如第二方面所述的方法的步骤。
第十一方面，提供了一种芯片，所述芯片包括处理器和通信接口，所述通信接口和所述处理器耦合，所述处理器用于运行程序或指令，实现如第一方面所述的方法，或实现如第二方面所述的方法。
第十二方面，提供了一种计算机程序/程序产品，所述计算机程序/程序产品被存储在存储介质中，所述计算机程序/程序产品被至少一个处理器执行以实现如第一方面所述的方法的步骤，或者实现如第二方面所述的方法的步骤。
在本申请实施例中，在触发AI模型分发的情况下，网络侧设备基于至少一个终端的第一信息，向各个终端发送第一AI模型的信息；由于第一AI模型为基于终端的第 一信息确定的，例如终端的能力信息、AI模型的需求信息和传输方式中至少一项，即第一信息与终端对应的第一AI模型匹配，使得能够根据实际的终端情况进行AI模型分发，AI模型分发的准确性较高。
附图说明
图1是本申请实施例可应用的无线通信系统的示意图；
图2是本申请实施例提供的AI模型分发方法的流程示意图；
图3是本申请实施例提供的AI模型分发场景示意图；
图4是本申请实施例提供的AI模型分发方法的交互流程示意图之一；
图5是本申请实施例提供的AI模型分发方法的交互流程示意图之二；
图6是本申请实施例提供的AI模型接收方法的流程示意图；
图7是本申请实施例提供的AI模型分发装置的结构示意图；
图8是本申请实施例提供的AI模型接收装置的结构示意图；
图9是本申请实施例提供的通信设备的结构示意图；
图10是本申请实施例提供的终端的结构示意图；
图11是本申请实施例的网络侧设备的结构示意图之一；
图12是本申请实施例的网络侧设备的结构示意图之二。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员所获得的所有其他实施例，都属于本申请保护的范围。
本申请的说明书和权利要求书中的术语“第一”、“第二”等是用于区别类似的对象，而不用于描述特定的顺序或先后次序。应该理解这样使用的术语在适当情况下可以互换，以便本申请的实施例能够以除了在这里图示或描述的那些以外的顺序实施，且“第一”、“第二”所区别的对象通常为一类，并不限定对象的个数，例如第一对象可以是一个，也可以是多个。此外，说明书以及权利要求中“和/或”表示所连接对象的至少其中之一，字符“/”一般表示前后关联对象是一种“或”的关系。本申请的说明书和权利要求书中的术语“指示”既可以是一个明确的指示，也可以是一个隐含的指示。其中，明确的指示可以理解为，发送方在发送的指示中明确告知了接收方需要执行的操作或请求结果；隐含的指示可以理解为，接收方根据发送方发送的指示进行判断，根据判断结果确定需要执行的操作或请求结果。
值得指出的是，本申请实施例所描述的技术不限于长期演进型(Long Term Evolution，LTE)/LTE的演进(LTE-Advanced，LTE-A)系统，还可用于其他无线通信系统，诸如码分多址(Code Division Multiple Access，CDMA)、时分多址(Time Division Multiple Access，TDMA)、频分多址(Frequency Division Multiple Access，FDMA)、正交频分多址(Orthogonal Frequency Division Multiple Access，OFDMA)、 单载波频分多址(Single-carrier Frequency Division Multiple Access，SC-FDMA)和其他系统。本申请实施例中的术语“系统”和“网络”常被可互换地使用，所描述的技术既可用于以上提及的系统和无线电技术，也可用于其他系统和无线电技术。以下描述出于示例目的描述了新空口(New Radio，NR)系统，并且在以下大部分描述中使用NR术语，但是这些技术也可应用于NR系统应用以外的通信系统，如第6代(6th Generation，6G)通信系统。
图1是本申请实施例可应用的无线通信系统的示意图，图1示出的无线通信系统包括终端11和网络侧设备12。其中，终端11可以是手机、平板电脑(Tablet Personal Computer)、膝上型电脑(Laptop Computer)或称为笔记本电脑、个人数字助理(Personal Digital Assistant，PDA)、掌上电脑、上网本、超级移动个人计算机(ultra-mobile personal computer，UMPC)、移动上网装置(Mobile Internet Device，MID)、增强现实(augmented reality，AR)/虚拟现实(virtual reality，VR)设备、机器人、可穿戴式设备(Wearable Device)、车载设备(VUE)、行人终端(PUE)、智能家居(具有无线通信功能的家居设备，如冰箱、电视、洗衣机或者家具等)、游戏机、个人计算机(personal computer，PC)、柜员机或者自助机等终端侧设备，可穿戴式设备包括：智能手表、智能手环、智能耳机、智能眼镜、智能首饰(智能手镯、智能手链、智能戒指、智能项链、智能脚镯、智能脚链等)、智能腕带、智能服装等。除了上述终端设备，也可以是终端内的芯片，例如调制解调器(Modem)芯片，系统级芯片(System on Chip，SoC)。需要说明的是，在本申请实施例并不限定终端11的具体类型。
网络侧设备12可以包括接入网设备或核心网设备，其中，接入网设备也可以称为无线接入网设备、无线接入网(Radio Access Network，RAN)、无线接入网功能或无线接入网单元。接入网设备可以包括基站、WLAN接入点或WiFi节点等，基站可被称为节点B、演进节点B(eNB)、接入点、基收发机站(Base Transceiver Station，BTS)、无线电基站、无线电收发机、基本服务集(Basic Service Set，BSS)、扩展服务集(Extended Service Set，ESS)、家用B节点、家用演进型B节点、发送接收点(Transmitting Receiving Point，TRP)或所述领域中其他某个合适的术语，只要达到相同的技术效果，所述基站不限于特定技术词汇，需要说明的是，在本申请实施例中仅以NR系统中的基站为例进行介绍，并不限定基站的具体类型。核心网设备可以包含但不限于如下至少一项：核心网节点、核心网功能、移动管理实体(Mobility Management Entity，MME)、接入移动管理功能(Access and Mobility Management Function，AMF)、会话管理功能(Session Management Function，SMF)、用户平面功能(User Plane Function，UPF)、策略控制功能(Policy Control Function，PCF)、策略与计费规则功能单元(Policy and Charging Rules Function，PCRF)、边缘应用服务发现功能(Edge Application Server Discovery Function，EASDF)、统一数据管理(Unified Data Management，UDM)，统一数据仓储(Unified Data Repository，UDR)、 归属用户服务器(Home Subscriber Server，HSS)、集中式网络配置(Centralized network configuration，CNC)、网络存储功能(Network Repository Function，NRF)，网络开放功能(Network Exposure Function，NEF)、本地NEF(Local NEF，或L-NEF)、绑定支持功能(Binding Support Function，BSF)、应用功能(Application Function，AF)、位置管理功能(location manage function，LMF)、增强服务移动定位中心(Enhanced Serving Mobile Location Centre，E-SMLC)、网络数据分析功能(network data analytics function，NWDAF)等。需要说明的是，在本申请实施例中仅以NR系统中的核心网设备为例进行介绍，并不限定核心网设备的具体类型。
下面结合附图，通过一些实施例及其应用场景对本申请实施例提供的AI模型分发方法、AI模型接收方法进行详细地说明。
图2是本申请实施例提供的AI模型分发方法的流程示意图，如图2所示，该方法包括：
步骤101、在触发AI模型分发的情况下，网络侧设备基于至少一个终端的第一信息，向各个终端发送第一AI模型的信息；终端的第一信息与终端对应的第一AI模型匹配；第一信息包括以下至少一项：终端的能力信息、AI模型的需求信息和传输方式。
具体地，考虑到终端能力以及移动性问题，可能在不同时刻、不同位置，网络情况、终端的能力(例如算力)都是实时变化的，为了保证分发结果的准确及有效性，本申请实施例中网络侧设备基于终端的第一信息，确定终端对应的第一AI模型，并向该终端发送第一AI模型的信息。
其中，AI模型也称为AI单元，可以是AI模型、AI模块或者具备AI/机器学习(Machine Learning，ML)功能的运算规则等。
可选地，终端的能力信息包括以下至少一项：计算能力、存储能力、计算能力的使用量、当前剩余计算能力、当前可用计算能力、存储能力的使用量、当前剩余存储能力、当前可用存储能力、可处理的业务信息。
其中，业务信息例如包括业务标识、业务类型等，例如计算类型、存储类型。
可选地，AI模型的需求信息包括以下至少一项：服务需求信息、AI模型的描述信息。
可选地，AI模型的描述信息包括以下至少一项：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
其中传输方式例如包括：通过5G系统(5G system，5GS)网络，或直接通过应用层传输(即5GS不感知(over 5GS))。
进一步，终端接收到第一AI模型的信息，可以基于第一AI模型进行业务处理，例如图像识别。
其中，可以通过终端和网络侧设备的通信触发AI模型的传递，例如终端发送请求。
本申请实施例提供的AI模型分发方法中，在触发AI模型分发的情况下，网络侧 设备基于至少一个终端的第一信息，向各个终端发送第一AI模型的信息；由于第一AI模型为基于终端的第一信息确定的，例如终端的能力信息、AI模型的需求信息和传输方式中至少一项，即第一信息与终端对应的第一AI模型匹配，使得能够根据实际的终端情况进行AI模型分发，AI模型分发的准确性较高。
可选地，步骤101之前还包括：
网络侧设备接收至少一个终端的请求消息，请求消息用于请求网络侧设备发送第一AI模型的信息，请求消息包括：终端的第一信息。
具体地，在触发AI模型传递之前，可以在网络侧设备和终端之间交换和协商与所需AI模型相关的信息。例如包括终端的能力信息(包含但不限于计算能力、存储能力、当前计算能力的使用量等)、AI模型的需求信息(例如，服务需求信息、AI模型的描述信息，服务需求信息例如所需AI模型可处理的业务信息、类型等)和传输方式相关的信息。上述信息可以用于为终端所需服务选择合适的AI模型。
可选地，步骤101之前，还包括：
网络侧设备建立与终端的第一会话连接；第一会话连接用于传输第一AI模型的信息。
可选地，该方法还包括：
网络侧设备接收至少一个终端发送的训练结果数据，训练结果数据为终端对第一AI模型进行训练得到的；
网络侧设备基于至少一个终端发送的训练结果数据进行AI模型训练。
具体地，终端接收到第一AI模型的信息之后，可以基于训练数据对第一AI模型进行训练，训练数据例如为终端获取的训练样本数据。进一步，可以将训练得到的训练结果数据发送给网络侧设备，网络侧设备可以基于训练结果数据进行AI模型训练，训练结果数据例如可以包括训练后的AI模型的描述信息，例如包括：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息等。
可选地，该方法还包括：
网络侧设备建立与终端的第二会话连接，第二会话连接用于传输训练结果数据。
上述实施方式中，通过终端对AI模型进行训练，将训练结果数据反馈给网络侧设备，使得网络侧设备可以基于终端的反馈进一步修正模型，进而能够满足终端的需求。
在一实施例中，在触发AI模型更新的情况下，网络侧设备基于第二信息向至少一个终端发送更新后的第二AI模型的信息；第二信息包括以下至少一项：各个终端的能力更新信息、模型更新信息。
具体地，网络侧设备基于第二信息向至少一个终端发送更新后的第二模型的信息，例如基于终端最新的能力信息、最新的服务需求信息、模型更新信息等选择更新的与该终端匹配的AI模型，模型更新信息例如为网络侧设备基于至少一个终端的训练结果数据更新的新的任务(如联邦学习任务)训练得到的AI模型。
可选地，能力更新信息包括以下至少一项的更新信息：计算能力、存储能力、计算能力的使用量、当前剩余计算能力、当前可用计算能力、存储能力的使用量、当前剩余存储能力、当前可用存储能力、可处理的业务信息；
可选地，模型更新信息包括以下至少一项的更新信息：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
功能描述信息例如用于对AI模型的功能进行说明的信息，AI模型的功能例如可以处理的业务名称、业务类型等；拓扑结构是指AI模型采用的网络模型的结构，例如神经网络模型的拓扑结构、深度学习模型的拓扑结构等；参数信息例如包括AI模型的权重信息、采用的损失函数信息、激活函数、卷积核大小、卷积层数量等；使用的资源信息例如包括：运行时占用的内存大小、存储空间大小等。
例如，模型的标识进行了更新，则模型更新信息包括标识的更新信息，例如包括更新前的标识以及更新后的标识，或，模型的功能进行了更新，则模型更新信息包括功能描述信息的更新信息，例如包括更新后的功能描述信息，拓扑结构、参数信息、使用的资源信息等类似。
可选地，第二AI模型的信息包括以下至少一项：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
可选地，第二AI模型为基于至少一个终端的训练结果数据对第一AI模型进行训练得到的；训练结果数据为终端对所述第一AI模型进行训练得到的。
可选地，AI模型更新为基于第三信息触发的，第三信息包括以下至少一项：终端的能力更新指示信息、网络侧设备更新AI模型的指示信息；终端的能力更新指示信息用于指示以下至少一项：终端的计算能力使用率大于或等于第一阈值、终端的计算资源使用率大于或等于第二阈值、终端的负载大于或等于第三阈值、终端的资源小于或等于第四阈值。
具体地，AI模型更新的触发因素可以是以下至少一项：
(1)终端当前的计算能力改变(例如内存、硬盘等使用率发生改变)，计算能力不能满足当前AI模型的需求或远超于当前AI模型的需求；
(2)网络侧设备更新AI模型。
上述实施方式中，在触发AI模型更新的情况下，网络侧设备基于至少一个终端的第二信息，向各个终端发送更新的第二AI模型的信息；由于第二AI模型为基于终端的第二信息确定的，例如终端的能力更新信息、AI模型的需求更新信息、服务需求更新信息中至少一项，使得能够根据实际的终端情况进行AI模型分发，AI模型分发的准确性较高。
示例性地，如图3所示的架构中，对于网络侧来说，联邦学习引擎(Federated learning engine)从AI模型库(AI Model repository)接收部分训练模型，该模型被传递到AI模型交付功能(AI model delivery function)，以通过5GS交付给多个终端。
来自多个终端的训练结果数据(Training Result Data)也通过5GS被联邦学习引擎接收，然后聚合用于AI模型的持续训练。
AI模型的更新(例如在拓扑、权重等参数或其他信息)在更新过程中被传送到终端。
对于终端侧来说，AI模型数据通过5GS由AI模型访问功能(AI model access function)接收，然后将数据传递给AI训练引擎(AI training engine)。
例如，终端中的AI训练引擎使用本地设备数据作为输入的训练数据来训练AI模型。
训练结果(例如以更新的AI模型的描述信息的形式)通过训练结果传递功能(Training results delivery function)传递到网络侧。
需要说明的是，网络侧设备可以是接入网设备和/或核心网设备，上述功能模块可以部署在一个或多个网络侧设备中。
示例性地，如图4所示，该方法包括：
步骤1、终端APP和网络APP通信触发AI模型的传递；
步骤2、终端APP和网络APP交互，并选择AI模型；
步骤3、网络APP标识被选中的AI模型
步骤4、AI模型交付功能与AI模型访问功能建立用于AI模型传递的会话连接；
步骤5、AI模型访问功能接收AI模型；
步骤6、AI模型访问功能将AI模型传递至AI训练引擎；
步骤7、终端上，训练数据传递至AI训练引擎；
步骤8、AI训练引擎执行模型训练；
步骤9、训练结果传递功能与联邦学习引擎建立用于训练结果传递的会话连接；
步骤10、训练结果传递至联邦学习引擎。
示例性地，如图5所示，该方法包括：
步骤0、触发AI模型更新，触发因素例如包含但不限于以下几条：
(1)终端当前的计算能力改变(例如内存、硬盘等使用率发生改变)；
(2)联邦学习引擎更新AI模型。
步骤1、终端APP和网络APP通信触发AI模型的更新传递；
步骤2、终端APP和网络APP交互，并选择更新的AI模型；
步骤3、网络APP标识被选中的更新的AI模型
步骤4、AI模型交付功能与AI模型访问功能建立用于AI模型传递的会话连接；
步骤5、AI模型访问功能接收更新的AI模型；
步骤6、AI模型访问功能将更新的AI模型传递至AI训练引擎；
步骤7、终端上，training input data传递至AI training engine
步骤8、AI训练引擎执行模型训练；
步骤9、训练结果传递功能与联邦学习引擎建立用于训练结果传递的会话连接；
步骤10、训练结果传递至联邦学习引擎。
图6是本申请实施例提供的AI模型接收方法的流程示意图，如图6所示，该方法包括：
步骤201、终端接收网络侧设备发送的第一AI模型的信息，第一AI模型与终端的第一信息匹配，第一信息包括以下至少一项：终端的能力信息、AI模型的需求信息和传输方式；
步骤202、终端基于第一AI模型进行业务处理。
可选地，所述方法还包括：
所述终端基于训练数据对所述第一AI模型进行训练，得到训练结果数据；
所述终端向所述网络侧设备发送所述训练结果数据；所述训练结果数据用于所述网络侧设备进行AI模型训练。
可选地，所述终端的能力信息包括以下至少一项：计算能力、存储能力、计算能力的使用量、当前剩余计算能力、当前可用计算能力、存储能力的使用量、当前剩余存储能力、当前可用存储能力、可处理的业务信息。
可选地，所述AI模型的需求信息包括以下至少一项：服务需求信息、AI模型的描述信息；所述AI模型的描述信息包括：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
可选地，所述方法还包括：
所述终端接收所述网络侧设备发送的更新后的第二AI模型的信息，所述第二AI模型为基于第二信息得到的；所述第二信息包括以下至少一项：各个所述终端的能力更新信息、服务需求更新信息、模型更新信息。
可选地，所述更新后的第二AI模型的信息为触发AI模型更新的情况下发送的，所述AI模型更新为基于第三信息触发的，所述第三信息包括以下至少一项：所述终端的能力更新指示信息、所述网络侧设备更新AI模型的指示信息；所述终端的能力更新指示信息用于指示以下至少一项：所述终端的计算能力使用率大于或等于第一阈值、所述终端的计算资源使用率大于或等于第二阈值、所述终端的负载大于或等于第三阈值、所述终端的资源小于或等于第四阈值。
可选地，所述第二AI模型为基于至少一个所述终端的训练结果数据对所述第一AI模型进行训练得到的；所述训练结果数据为所述终端对所述第一AI模型进行训练得到的。
可选地，所述第二AI模型的信息包括以下至少一项：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
可选地，所述训练结果数据包括更新后的AI模型的描述信息。
可选地，所述终端接收网络侧设备发送的第一AI模型的信息之前，还包括：
所述终端向所述网络侧设备发送请求消息，所述请求消息用于请求所述网络侧设备发送所述第一AI模型的信息，所述请求消息包括：所述终端的第一信息。
可选地，所述终端接收网络侧设备发送的第一AI模型的信息之前，还包括：
所述终端建立与所述网络侧设备的第一会话连接；所述第一会话连接用于传输第一AI模型的信息。
可选地，所述方法还包括：
所述终端建立与所述网络侧设备的第二会话连接，所述第二会话连接用于传输训练结果数据。
本实施例的方法，其具体实现过程与技术效果与终端侧方法实施例中相同，具体可以参见终端侧方法实施例中的详细介绍，此处不再赘述。
本申请实施例提供的AI模型分发方法，执行主体可以为AI模型分发装置。本申请实施例提供的AI模型接收方法，执行主体可以为AI模型接收装置。本申请实施例中以AI模型分发装置执行AI模型分发方法、AI模型接收装置执行AI模型接收方法为例，说明本申请实施例提供的AI模型分发装置、AI模型接收装置。
图7是本申请实施例提供的AI模型分发装置的结构示意图，如图7所示，该AI模型分发装置，应用于网络侧设备，包括：
发送模块210，用于在触发AI模型分发的情况下，基于至少一个终端的第一信息，向各个所述终端发送第一AI模型的信息；所述终端的第一信息与所述终端对应的第一AI模型匹配；所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式。
可选地，所述终端的能力信息包括以下至少一项：计算能力、存储能力、计算能力的使用量、当前剩余计算能力、当前可用计算能力、存储能力的使用量、当前剩余存储能力、当前可用存储能力、可处理的业务信息；
可选地，所述AI模型的需求信息包括以下至少一项：服务需求信息、AI模型的描述信息。
可选地，所述AI模型的描述信息包括以下至少一项：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
可选地，该装置还包括：
接收模块，用于接收至少一个所述终端发送的训练结果数据，所述训练结果数据为所述终端对所述第一AI模型进行训练得到的；
处理模块，用于基于至少一个所述终端发送的训练结果数据进行AI模型训练。
可选地，发送模块210，还用于在触发AI模型更新的情况下，基于第二信息向至少一个所述终端发送更新后的第二AI模型的信息；所述第二信息包括以下至少一项：各个所述终端的能力更新信息、服务需求更新信息、模型更新信息。
可选地，所述能力更新信息包括以下至少一项的更新信息：计算能力、存储能力、 计算能力的使用量、当前剩余计算能力、当前可用计算能力、存储能力的使用量、当前剩余存储能力、当前可用存储能力、可处理的业务信息；和/或，
所述模型更新信息包括以下至少一项的更新信息：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
可选地，所述AI模型更新为基于第三信息触发的，所述第三信息包括以下至少一项：终端的能力更新指示信息、所述网络侧设备更新AI模型的指示信息；所述终端的能力更新指示信息用于指示以下至少一项：所述终端的计算能力使用率大于或等于第一阈值、所述终端的计算资源使用率大于或等于第二阈值、所述终端的负载大于或等于第三阈值、所述终端的资源小于或等于第四阈值。
可选地，所述第二AI模型为基于至少一个所述终端的训练结果数据对所述第一AI模型进行训练得到的；所述训练结果数据为所述终端对所述第一AI模型进行训练得到的。
可选地，所述第二AI模型的信息包括以下至少一项：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
可选地，所述训练结果数据包括更新后的AI模型的描述信息。
可选地，接收模块，还用于接收至少一个所述终端的请求消息，所述请求消息用于请求所述网络侧设备发送所述第一AI模型的信息，所述请求消息包括：所述终端的第一信息。
可选地，处理模块，还用于：
建立与所述终端的第一会话连接；所述第一会话连接用于传输第一AI模型的信息。
可选地，处理模块，还用于：
建立与所述终端的第二会话连接，所述第二会话连接用于传输训练结果数据。
本实施例的装置，可以用于执行前述网络侧设备方法实施例中任一实施例的方法，其具体实现过程与技术效果与网络侧设备方法实施例中相同，具体可以参见网络侧设备方法实施例中的详细介绍，此处不再赘述。
图8是本申请实施例提供的AI模型接收装置的结构示意图，如图8所示，该AI模型接收装置，应用于终端，包括：
接收模块310，用于接收网络侧设备发送的第一AI模型的信息，所述第一AI模型与所述终端的第一信息匹配，所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式；
处理模块320，用于基于所述第一AI模型进行业务处理。
可选地，所述处理模块320，还用于：
基于训练数据对所述第一AI模型进行训练，得到训练结果数据；
该装置还包括：发送模块，用于向所述网络侧设备发送所述训练结果数据；所述 训练结果数据用于所述网络侧设备进行AI模型训练。
可选地，所述终端的能力信息包括以下至少一项：计算能力、存储能力、计算能力的使用量、当前剩余计算能力、当前可用计算能力、存储能力的使用量、当前剩余存储能力、当前可用存储能力、可处理的业务信息。
可选地，所述AI模型的需求信息包括以下至少一项：服务需求信息、AI模型的描述信息；所述AI模型的描述信息包括：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
可选地，接收模块310，还用于：
接收所述网络侧设备发送的更新后的第二AI模型的信息，所述第二AI模型为基于第二信息得到的；所述第二信息包括以下至少一项：各个所述终端的能力更新信息、服务需求更新信息、模型更新信息。
可选地，所述能力更新信息包括以下至少一项的更新信息：计算能力、存储能力、计算能力的使用量、当前剩余计算能力、当前可用计算能力、存储能力的使用量、当前剩余存储能力、当前可用存储能力、可处理的业务信息；和/或
所述模型更新信息包括以下至少一项的更新信息：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
可选地，所述更新后的第二AI模型的信息为触发AI模型更新的情况下发送的，所述AI模型更新为基于第三信息触发的，所述第三信息包括以下至少一项：所述终端的能力更新指示信息、所述网络侧设备更新AI模型的指示信息；所述终端的能力更新指示信息用于指示以下至少一项：所述终端的计算能力使用率大于或等于第一阈值、所述终端的计算资源使用率大于或等于第二阈值、所述终端的负载大于或等于第三阈值、所述终端的资源小于或等于第四阈值。
可选地，所述第二AI模型为基于至少一个所述终端的训练结果数据对所述第一AI模型进行训练得到的；所述训练结果数据为所述终端对所述第一AI模型进行训练得到的。
可选地，所述第二AI模型的信息包括以下至少一项：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
可选地，所述训练结果数据包括更新后的AI模型的描述信息。
可选地，发送模块，还用于：
向所述网络侧设备发送请求消息，所述请求消息用于请求所述网络侧设备发送所述第一AI模型的信息，所述请求消息包括：所述终端的第一信息。
可选地，处理模块320，还用于建立与所述网络侧设备的第一会话连接；所述第一会话连接用于传输第一AI模型的信息。
可选地，处理模块320，还用于建立与所述网络侧设备的第二会话连接，所述第二会话连接用于传输训练结果数据。
本实施例的装置，可以用于执行前述终端侧方法实施例中任一实施例的方法，其具体实现过程与技术效果与终端侧方法实施例中相同，具体可以参见终端侧方法实施例中的详细介绍，此处不再赘述。
本申请实施例中的AI模型分发装置、AI模型接收装置可以是电子设备，例如具有操作系统的电子设备，也可以是电子设备中的部件，例如集成电路或芯片。该电子设备可以是终端，也可以为除终端之外的其他设备。示例性的，终端可以包括但不限于上述所列举的终端11的类型，其他设备可以为服务器、网络附属存储器(Network Attached Storage，NAS)等，本申请实施例不作具体限定。
本申请实施例提供的AI模型分发装置、AI模型接收装置能够实现图2至图6的方法实施例实现的各个过程，并达到相同的技术效果，为避免重复，这里不再赘述。
图9是本申请实施例提供的通信设备的结构示意图，如图9所示，该通信设备900，包括处理器901和存储器902，存储器902上存储有可在所述处理器901上运行的程序或指令，例如，该通信设备900为终端时，该程序或指令被处理器901执行时实现上述AI模型接收方法实施例的各个步骤，且能达到相同的技术效果。该通信设备900为网络侧设备时，该程序或指令被处理器901执行时实现上述AI模型分发方法实施例的各个步骤，且能达到相同的技术效果，为避免重复，这里不再赘述。
本申请实施例还提供一种终端，包括处理器和通信接口，通信接口用于接收网络侧设备发送的第一AI模型的信息，所述第一AI模型与所述终端的第一信息匹配，所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式；处理器用于基于所述第一AI模型进行业务处理。该终端实施例与上述终端侧方法实施例对应，上述方法实施例的各个实施过程和实现方式均可适用于该终端实施例中，且能达到相同的技术效果。
图10是本申请实施例提供的终端的结构示意图，如图10所示，该终端1000包括但不限于：射频单元1001、网络模块1002、音频输出单元1003、输入单元1004、传感器1005、显示单元1006、用户输入单元1007、接口单元1008、存储器1009、以及处理器1010等中的至少部分部件。
本领域技术人员可以理解，终端1000还可以包括给各个部件供电的电源(比如电池)，电源可以通过电源管理系统与处理器1010逻辑相连，从而通过电源管理系统实现管理充电、放电、以及功耗管理等功能。图10中示出的终端结构并不构成对终端的限定，终端可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置，在此不再赘述。
应理解的是，本申请实施例中，输入单元1004可以包括图形处理单元(Graphics Processing Unit，GPU)10041和麦克风10042，图形处理器10041对在视频捕获模式或图像捕获模式中由图像捕获装置(如摄像头)获得的静态图片或视频的图像数据进行处理。显示单元1006可包括显示面板10061，可以采用液晶显示器、有机发光二极 管等形式来配置显示面板10061。用户输入单元1007包括触控面板10071以及其它输入设备10072中的至少一种。触控面板10071，也称为触摸屏。触控面板10071可包括触摸检测装置和触摸控制器两个部分。其它输入设备10072可以包括但不限于物理键盘、功能键(比如音量控制按键、开关按键等)、轨迹球、鼠标、操作杆，在此不再赘述。
本申请实施例中，射频单元1001将接收来自网络侧设备的下行数据接收后，可以传输给处理器1010进行处理；另外，射频单元1001可以将上行的数据发送给向网络侧设备发送上行数据。通常，射频单元1001包括但不限于天线、至少一个放大器、收发信机、耦合器、低噪声放大器、双工器等。
存储器1009可用于存储软件程序或指令以及各种数据。存储器1009可主要包括存储程序或指令的第一存储区和存储数据的第二存储区，其中，第一存储区可存储操作系统、至少一个功能所需的应用程序或指令(比如声音播放功能、图像播放功能等)等。此外，存储器1009可以包括易失性存储器或非易失性存储器，或者，存储器1009可以包括瞬态和非瞬态存储器。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)，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，DDRSDRAM)、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(Synch link DRAM，SLDRAM)和直接内存总线随机存取存储器(Direct Rambus RAM，DRRAM)。本申请实施例中的存储器1009包括但不限于这些和任意其它适合类型的存储器。
处理器1010可包括一个或多个处理单元；可选的，处理器1010可集成应用处理器和调制解调处理器，其中，应用处理器主要处理涉及操作系统、用户界面和应用程序或指令等的操作，调制解调处理器主要处理无线通信信号，如基带处理器。可以理解的是，上述调制解调处理器也可以不集成到处理器1010中。
其中，射频单元1001，用于接收网络侧设备发送的第一AI模型的信息，所述第一AI模型与所述终端的第一信息匹配，所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式；
处理器1010，用于基于所述第一AI模型进行业务处理。
可选地，所述处理器1010，还用于：
基于训练数据对所述第一AI模型进行训练，得到训练结果数据；
射频单元1001，还用于向所述网络侧设备发送所述训练结果数据；所述训练结果 数据用于所述网络侧设备进行AI模型训练。
可选地，所述终端的能力信息包括以下至少一项：计算能力、存储能力、计算能力的使用量、当前剩余计算能力、当前可用计算能力、存储能力的使用量、当前剩余存储能力、当前可用存储能力、可处理的业务信息。
可选地，所述AI模型的需求信息包括以下至少一项：服务需求信息、AI模型的描述信息；所述AI模型的描述信息包括：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
可选地，射频单元1001，还用于：
接收所述网络侧设备发送的更新后的第二AI模型的信息，所述第二AI模型为基于第二信息得到的；所述第二信息包括以下至少一项：各个所述终端的能力更新信息、服务需求更新信息、模型更新信息。
可选地，所述能力更新信息包括以下至少一项的更新信息：计算能力、存储能力、计算能力的使用量、当前剩余计算能力、当前可用计算能力、存储能力的使用量、当前剩余存储能力、当前可用存储能力、可处理的业务信息；和/或
所述模型更新信息包括以下至少一项的更新信息：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
可选地，所述更新后的第二AI模型的信息为触发AI模型更新的情况下发送的，所述AI模型更新为基于第三信息触发的，所述第三信息包括以下至少一项：所述终端的能力更新指示信息、所述网络侧设备更新AI模型的指示信息；所述终端的能力更新指示信息用于指示以下至少一项：所述终端的计算能力使用率大于或等于第一阈值、所述终端的计算资源使用率大于或等于第二阈值、所述终端的负载大于或等于第三阈值、所述终端的资源小于或等于第四阈值。
可选地，所述第二AI模型为基于至少一个所述终端的训练结果数据对所述第一AI模型进行训练得到的；所述训练结果数据为所述终端对所述第一AI模型进行训练得到的。
可选地，所述第二AI模型的信息包括以下至少一项：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
可选地，所述训练结果数据包括更新后的AI模型的描述信息。
可选地，射频单元1001，还用于：
向所述网络侧设备发送请求消息，所述请求消息用于请求所述网络侧设备发送所述第一AI模型的信息，所述请求消息包括：所述终端的第一信息。
可选地，处理器1010，还用于建立与所述网络侧设备的第一会话连接；所述第一会话连接用于传输第一AI模型的信息。
可选地，处理器1010，还用于建立与所述网络侧设备的第二会话连接，所述第二会话连接用于传输训练结果数据。
本申请实施例还提供一种网络侧设备，包括处理器和通信接口，通信接口用于在触发AI模型分发的情况下，基于至少一个终端的第一信息，向各个所述终端发送第一AI模型的信息；所述终端的第一信息与所述终端对应的第一AI模型匹配；所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式。该网络侧设备实施例与上述网络侧设备方法实施例对应，上述方法实施例的各个实施过程和实现方式均可适用于该网络侧设备实施例中，且能达到相同的技术效果。
图11是本申请实施例提供的网络侧设备的结构示意图，如图11所示，该网络侧设备1100包括：天线1101、射频装置1102、基带装置1103、处理器1104和存储器1105。天线1101与射频装置1102连接。在上行方向上，射频装置1102通过天线1101接收信息，将接收的信息发送给基带装置1103进行处理。在下行方向上，基带装置1103对要发送的信息进行处理，并发送给射频装置1102，射频装置1102对收到的信息进行处理后经过天线1101发送出去。
以上实施例中网络侧设备执行的方法可以在基带装置1103中实现，该基带装置1103包括基带处理器。
基带装置1103例如可以包括至少一个基带板，该基带板上设置有多个芯片，如图11所示，其中一个芯片例如为基带处理器，通过总线接口与存储器1105连接，以调用存储器1105中的程序，执行以上方法实施例中所示的网络设备操作。
该网络侧设备还可以包括网络接口1106，该接口例如为通用公共无线接口(common public radio interface，CPRI)。
具体地，本申请实施例的网络侧设备1100还包括：存储在存储器1105上并可在处理器y04上运行的指令或程序，处理器1104调用存储器1105中的指令或程序执行如上所述的AI模型的分发方法，并达到相同的技术效果，为避免重复，故不在此赘述。
图12是本申请实施例提供的网络侧设备的结构示意图，如图12所示，该网络侧设备1200包括：处理器1201、网络接口1202和存储器1203。其中，网络接口1202例如为通用公共无线接口(common public radio interface，CPRI)。
具体地，本申请实施例的网络侧设备1200还包括：存储在存储器1203上并可在处理器1201上运行的指令或程序，处理器1201调用存储器1203中的指令或程序执行图7所示各模块执行的方法，并达到相同的技术效果，为避免重复，故不在此赘述。
本申请实施例还提供了一种通信系统，包括：终端及网络侧设备，所述终端可用于执行如上所述的AI模型接收方法的步骤，所述网络侧设备可用于执行如上所述的AI模型分发方法的步骤。
本申请实施例还提供一种可读存储介质，所述可读存储介质可以是以易失性的，也可以是非易失性的，所述可读存储介质上存储有程序或指令，该程序或指令被处理器执行时实现上述AI模型分发、AI模型接收方法实施例的各个过程，且能达到相同的技术效果，为避免重复，这里不再赘述。
其中，所述处理器为上述实施例中所述的终端中的处理器。所述可读存储介质，包括计算机可读存储介质，如计算机只读存储器ROM、随机存取存储器RAM、磁碟或者光盘等。
本申请实施例另提供了一种芯片，所述芯片包括处理器和通信接口，所述通信接口和所述处理器耦合，所述处理器用于运行程序或指令，实现上述AI模型分发、AI模型接收方法实施例的各个过程，且能达到相同的技术效果，为避免重复，这里不再赘述。
应理解，本申请实施例提到的芯片还可以称为系统级芯片，系统芯片，芯片系统或片上系统芯片等。
本申请实施例另提供了一种计算机程序/程序产品，所述计算机程序/程序产品被存储在存储介质中，所述计算机程序/程序产品被至少一个处理器执行以实现上述AI模型分发、AI模型接收方法实施例的各个过程，且能达到相同的技术效果，为避免重复，这里不再赘述。
需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者装置不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者装置所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者装置中还存在另外的相同要素。此外，需要指出的是，本申请实施方式中的方法和装置的范围不限按示出或讨论的顺序来执行功能，还可包括根据所涉及的功能按基本同时的方式或按相反的顺序来执行功能，例如，可以按不同于所描述的次序来执行所描述的方法，并且还可以添加、省去、或组合各种步骤。另外，参照某些示例所描述的特征可在其他示例中被组合。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对相关技术做出贡献的部分可以以计算机软件产品的形式体现出来，该计算机软件产品存储在一个存储介质(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端(可以是手机，计算机，服务器，空调器，或者网络设备等)执行本申请各个实施例所述的方法。
上面结合附图对本申请的实施例进行了描述，但是本申请并不局限于上述的具体实施方式，上述的具体实施方式仅仅是示意性的，而不是限制性的，本领域的普通技术人员在本申请的启示下，在不脱离本申请宗旨和权利要求所保护的范围情况下，还可做出很多形式，均属于本申请的保护之内。

Claims (32)

  1. 一种人工智能AI模型分发方法，包括：
    在触发AI模型分发的情况下，网络侧设备基于至少一个终端的第一信息，向各个所述终端发送第一AI模型的信息；所述终端的第一信息与所述终端对应的第一AI模型匹配；所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式。
  2. 根据权利要求1所述的AI模型分发方法，其中，
    所述终端的能力信息包括以下至少一项：计算能力、存储能力、计算能力的使用量、当前剩余计算能力、当前可用计算能力、存储能力的使用量、当前剩余存储能力、当前可用存储能力、可处理的业务信息。
  3. 根据权利要求1所述的AI模型分发方法，其中，
    所述AI模型的需求信息包括以下至少一项：服务需求信息、AI模型的描述信息。
  4. 根据权利要求3所述的AI模型分发方法，其中，
    所述AI模型的描述信息包括以下至少一项：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
  5. 根据权利要求1或2所述的AI模型分发方法，其中，所述方法还包括：
    所述网络侧设备接收至少一个所述终端发送的训练结果数据，所述训练结果数据为所述终端对所述第一AI模型进行训练得到的；
    所述网络侧设备基于至少一个所述终端发送的训练结果数据进行AI模型训练。
  6. 根据权利要求1或2所述的AI模型分发方法，其中，所述方法还包括：
    在触发AI模型更新的情况下，所述网络侧设备基于第二信息向至少一个所述终端发送更新后的第二AI模型的信息；所述第二信息包括以下至少一项：各个所述终端的能力更新信息、服务需求更新信息、模型更新信息。
  7. 根据权利要求6所述的AI模型分发方法，其中，
    所述能力更新信息包括以下至少一项的更新信息：计算能力、存储能力、计算能力的使用量、当前剩余计算能力、当前可用计算能力、存储能力的使用量、当前剩余存储能力、当前可用存储能力、可处理的业务信息；和/或，
    所述模型更新信息包括以下至少一项的更新信息：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
  8. 根据权利要求6所述的AI模型分发方法，其中，
    所述AI模型更新为基于第三信息触发的，所述第三信息包括以下至少一项：终端的能力更新指示信息、所述网络侧设备更新AI模型的指示信息；所述终端的能力更新指示信息用于指示以下至少一项：所述终端的计算能力使用率大于或等于第一阈值、所述终端的计算资源使用率大于或等于第二阈值、所述终端的负载大于或等于第三阈值、所述终端的资源小于或等于第四阈值。
  9. 根据权利要求6所述的AI模型分发方法，其中，
    所述第二AI模型为基于至少一个所述终端的训练结果数据对所述第一AI模型进行训练得到的；所述训练结果数据为所述终端对所述第一AI模型进行训练得到的。
  10. 根据权利要求6所述的AI模型分发方法，其中，
    所述第二AI模型的信息包括以下至少一项：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
  11. 根据权利要求5或9所述的AI模型分发方法，其中，
    所述训练结果数据包括更新后的AI模型的描述信息。
  12. 根据权利要求1或2所述的AI模型分发方法，其中，所述网络侧设备基于至少一个终端的第一信息，向各个所述终端发送第一AI模型的信息之前，还包括：
    所述网络侧设备接收至少一个所述终端的请求消息，所述请求消息用于请求所述网络侧设备发送所述第一AI模型的信息，所述请求消息包括：所述终端的第一信息。
  13. 根据权利要求1或2所述的AI模型分发方法，其中，所述网络侧设备基于至少一个终端的第一信息，向各个所述终端发送第一AI模型的信息之前，还包括：
    所述网络侧设备建立与所述终端的第一会话连接；所述第一会话连接用于传输所述第一AI模型的信息。
  14. 根据权利要求5所述的AI模型分发方法，其中，所述方法还包括：
    所述网络侧设备建立与所述终端的第二会话连接，所述第二会话连接用于传输所述训练结果数据。
  15. 一种人工智能AI模型接收方法，包括：
    终端接收网络侧设备发送的第一AI模型的信息，所述第一AI模型与所述终端的第一信息匹配，所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式；
    所述终端基于所述第一AI模型进行业务处理。
  16. 根据权利要求15所述的AI模型接收方法，其中，所述方法还包括：
    所述终端基于训练数据对所述第一AI模型进行训练，得到训练结果数据；
    所述终端向所述网络侧设备发送所述训练结果数据；所述训练结果数据用于所述网络侧设备进行AI模型训练。
  17. 根据权利要求15或16所述的AI模型接收方法，其中，
    所述终端的能力信息包括以下至少一项：计算能力、存储能力、计算能力的使用量、当前剩余计算能力、当前可用计算能力、存储能力的使用量、当前剩余存储能力、当前可用存储能力、可处理的业务信息。
  18. 根据权利要求15或16所述的AI模型接收方法，其中，
    所述AI模型的需求信息包括以下至少一项：服务需求信息、AI模型的描述信息；所述AI模型的描述信息包括：标识、功能描述信息、拓扑结构、参数信息、使用的资 源信息。
  19. 根据权利要求15或16所述的AI模型接收方法，其中，所述方法还包括：
    所述终端接收所述网络侧设备发送的更新后的第二AI模型的信息，所述第二AI模型为基于第二信息得到的；所述第二信息包括以下至少一项：各个所述终端的能力更新信息、服务需求更新信息、模型更新信息。
  20. 根据权利要求19所述的AI模型接收方法，其中，
    所述能力更新信息包括以下至少一项的更新信息：计算能力、存储能力、计算能力的使用量、当前剩余计算能力、当前可用计算能力、存储能力的使用量、当前剩余存储能力、当前可用存储能力、可处理的业务信息；和/或
    所述模型更新信息包括以下至少一项的更新信息：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
  21. 根据权利要求19所述的AI模型接收方法，其中，
    所述更新后的第二AI模型的信息为触发AI模型更新的情况下发送的，所述AI模型更新为基于第三信息触发的，所述第三信息包括以下至少一项：所述终端的能力更新指示信息、所述网络侧设备更新AI模型的指示信息；所述终端的能力更新指示信息用于指示以下至少一项：所述终端的计算能力使用率大于或等于第一阈值、所述终端的计算资源使用率大于或等于第二阈值、所述终端的负载大于或等于第三阈值、所述终端的资源小于或等于第四阈值。
  22. 根据权利要求19所述的AI模型接收方法，其中，
    所述第二AI模型为基于至少一个所述终端的训练结果数据对所述第一AI模型进行训练得到的；所述训练结果数据为所述终端对所述第一AI模型进行训练得到的。
  23. 根据权利要求19-22任一项所述的AI模型接收方法，其中，
    所述第二AI模型的信息包括以下至少一项：标识、功能描述信息、拓扑结构、参数信息、使用的资源信息。
  24. 根据权利要求16或22任一项所述的AI模型接收方法，其中，
    所述训练结果数据包括更新后的AI模型的描述信息。
  25. 根据权利要求15或16所述的AI模型接收方法，其中，所述终端接收网络侧设备发送的第一AI模型的信息之前，还包括：
    所述终端向所述网络侧设备发送请求消息，所述请求消息用于请求所述网络侧设备发送所述第一AI模型的信息，所述请求消息包括：所述终端的第一信息。
  26. 根据权利要求15或16所述的AI模型接收方法，其中，所述终端接收网络侧设备发送的第一AI模型的信息之前，还包括：
    所述终端建立与所述网络侧设备的第一会话连接；所述第一会话连接用于传输所述第一AI模型的信息。
  27. 根据权利要求16所述的AI模型接收方法，其中，所述方法还包括：
    所述终端建立与所述网络侧设备的第二会话连接，所述第二会话连接用于传输所述训练结果数据。
  28. 一种AI模型分发装置，包括：
    发送模块，用于在触发AI模型分发的情况下，基于至少一个终端的第一信息，向各个所述终端发送第一AI模型的信息；所述终端的第一信息与所述终端对应的第一AI模型匹配；所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式。
  29. 一种AI模型接收装置，包括：
    接收模块，用于接收网络侧设备发送的第一AI模型的信息，所述第一AI模型与终端的第一信息匹配，所述第一信息包括以下至少一项：所述终端的能力信息、AI模型的需求信息和传输方式；
    处理模块，用于基于所述第一AI模型进行业务处理。
  30. 一种网络侧设备，包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如权利要求1至14任一项所述的AI模型分发方法的步骤。
  31. 一种终端，包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如权利要求15至27任一项所述的AI模型接收方法的步骤。
  32. 一种可读存储介质，所述可读存储介质上存储程序或指令，所述程序或指令被处理器执行时实现如权利要求1至14任一项所述的AI模型分发方法，或者实现如权利要求15至27任一项所述的AI模型接收方法的步骤。
PCT/CN2024/071560 2023-01-12 2024-01-10 Ai模型分发、接收方法、终端及网络侧设备 Ceased WO2024149288A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202310084335.5A CN118338304A (zh) 2023-01-12 2023-01-12 Ai模型分发、接收方法、终端及网络侧设备
CN202310084335.5 2023-01-12

Publications (1)

Publication Number Publication Date
WO2024149288A1 true WO2024149288A1 (zh) 2024-07-18

Family

ID=91765179

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2024/071560 Ceased WO2024149288A1 (zh) 2023-01-12 2024-01-10 Ai模型分发、接收方法、终端及网络侧设备

Country Status (2)

Country Link
CN (1) CN118338304A (zh)
WO (1) WO2024149288A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021142609A1 (zh) * 2020-01-14 2021-07-22 Oppo广东移动通信有限公司 信息上报方法、装置、设备和存储介质
CN114091679A (zh) * 2020-08-24 2022-02-25 华为技术有限公司 一种更新机器学习模型的方法及通信装置
CN114745712A (zh) * 2021-01-07 2022-07-12 中国移动通信有限公司研究院 基于终端能力的处理方法、装置、终端及网络设备
WO2022161230A1 (zh) * 2021-02-01 2022-08-04 大唐移动通信设备有限公司 通信系统中的模型更新方法、装置及存储介质

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021142609A1 (zh) * 2020-01-14 2021-07-22 Oppo广东移动通信有限公司 信息上报方法、装置、设备和存储介质
CN114091679A (zh) * 2020-08-24 2022-02-25 华为技术有限公司 一种更新机器学习模型的方法及通信装置
CN114745712A (zh) * 2021-01-07 2022-07-12 中国移动通信有限公司研究院 基于终端能力的处理方法、装置、终端及网络设备
WO2022161230A1 (zh) * 2021-02-01 2022-08-04 大唐移动通信设备有限公司 通信系统中的模型更新方法、装置及存储介质

Also Published As

Publication number Publication date
CN118338304A (zh) 2024-07-12

Similar Documents

Publication Publication Date Title
US20240188047A1 (en) Computing session update method and apparatus, and communication device
US20240188046A1 (en) Computing session update method and apparatus, terminal, and network side device
WO2023246584A1 (zh) 算力处理方法、装置及通信设备
WO2024022267A1 (zh) 算力任务迁移方法及通信设备
US20250299110A1 (en) Model training method, terminal, and network-side device
WO2023246756A1 (zh) 算力服务方法、装置、终端及核心网设备
WO2024140725A1 (zh) 传输方法及装置
US20250240345A1 (en) Model request method and apparatus, communication device, and readable storage medium
WO2023125932A1 (zh) Ai网络信息传输方法、装置及通信设备
WO2024149288A1 (zh) Ai模型分发、接收方法、终端及网络侧设备
CN117676642A (zh) 小区配置方法、装置、中继设备、终端及网络设备
WO2023165393A1 (zh) 传输方法、装置、终端和存储介质
WO2023125933A1 (zh) Ai网络信息传输方法、装置及通信设备
CN116847330A (zh) 终端不可用时期的协商方法、终端及网络侧设备
CN116567781B (zh) 传输方法、装置、终端、网络侧设备及可读存储介质
JP7819408B2 (ja) ネットワーク選択方法および端末
CN116567588B (zh) 通知方法、第二终端及第一网络功能
WO2024140570A1 (zh) 策略配置方法、装置、终端、网络侧设备及可读存储介质
WO2024022398A1 (zh) 托管网络的选网信息的获取方法、终端及网络侧设备
WO2024067437A1 (zh) 模型去激活方法、信息发送方法、装置及设备
WO2025036206A1 (zh) 模型的识别方法、装置、终端及网络侧设备
CN117119429A (zh) 终端路由选择策略信息管理方法及相关设备
CN119011044A (zh) 模型处理方法、装置、终端及网络侧设备
CN117641571A (zh) 定位消息传输方法、终端及网络侧设备
WO2024093712A1 (zh) 中继通信链路处理方法、中继通信链路配置方法、中继终端处理方法及相关设备

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: 24741285

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 24741285

Country of ref document: EP

Kind code of ref document: A1