WO2024120286A1 - 传输方法、装置、终端及网络侧设备 - Google Patents

传输方法、装置、终端及网络侧设备 Download PDF

Info

Publication number
WO2024120286A1
WO2024120286A1 PCT/CN2023/135262 CN2023135262W WO2024120286A1 WO 2024120286 A1 WO2024120286 A1 WO 2024120286A1 CN 2023135262 W CN2023135262 W CN 2023135262W WO 2024120286 A1 WO2024120286 A1 WO 2024120286A1
Authority
WO
WIPO (PCT)
Prior art keywords
model
user plane
information
terminal
plane function
Prior art date
Application number
PCT/CN2023/135262
Other languages
English (en)
French (fr)
Inventor
崇卫微
Original Assignee
维沃移动通信有限公司
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 维沃移动通信有限公司 filed Critical 维沃移动通信有限公司
Publication of WO2024120286A1 publication Critical patent/WO2024120286A1/zh

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00Machine learning
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W8/00Network data management
    • H04W8/22Processing or transfer of terminal data, e.g. status or physical capabilities
    • H04W8/24Transfer of terminal data

Definitions

  • AI Artificial Intelligence
  • ML Machine Learning
  • the terminal needs to download the latest ML model from the network side.
  • the terminal may also participate in the training process of some models. For example, in federated learning, after the terminal generates a local model through local training, it uploads it to the network side for aggregation, which involves the issue of uploading the terminal model.
  • the communication system only provides a solution for the transmission of ordinary user data to the terminal, and it is still unclear how to realize the transmission of ML models between the terminal and the network side.
  • the embodiments of the present application provide a transmission method, apparatus, terminal, and network-side equipment, which can solve the problem in the related art that the transmission of ML models between the terminal and the network side is still unclear.
  • a transmission method comprising:
  • the terminal sends a first request message to the application function AF, where the first request message is used to request transmission of the machine learning ML model;
  • the terminal receives a first response message sent by the AF, where the first response message includes information about a user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • the AF receives a first request message sent by the terminal, where the first request message is used to request transmission of the ML model;
  • the AF sends a first response message to the terminal, where the first response message includes information about a user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • a third aspect provides a transmission method, including:
  • the model application platform receives a second request message sent by the AF, where the second request message is used to request information about a user plane function, where the user plane function is used to transmit the ML model to the terminal using a user plane session;
  • the model application platform sends the information of the user plane function to the AF.
  • a transmission device comprising:
  • a first sending module configured to send a first request message to the application function AF, wherein the first request message is used to request transmission of a machine learning ML model;
  • the first receiving module is used to receive a first response message sent by the AF, where the first response message includes information about a user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • a transmission device including:
  • a second receiving module configured to receive a first request message sent by a terminal, wherein the first request message is used to request transmission of an ML model
  • the second sending module is used to send a first response message to the terminal, where the first response message includes information about the user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • a transmission device including:
  • a third receiving module is used to receive a second request message sent by the AF, where the second request message is used to request to obtain information about a user plane function, where the user plane function is used to transmit an ML model with the terminal using a user plane session;
  • the third sending module is used to send the information of the user plane function to the AF.
  • a terminal comprising a processor and a memory, wherein the memory stores a program or instruction that can be executed on the processor, and when the program or instruction is executed by the processor, the steps of the transmission method described in the first aspect are implemented.
  • a terminal comprising a processor and a communication interface, wherein the communication interface is used to send a first request message to an application function AF, wherein the first request message is used to request the transmission of a machine learning ML model; and to receive a first response message sent by the AF, wherein the first response message includes information about a user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • 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 transmission method described in the second aspect or the third aspect are implemented.
  • a network side device including a processor and a communication interface, wherein the communication interface is used to receive a first request message sent by a terminal, the first request message is used to request the transmission of an ML model; and is used to send a first response message to the terminal, the first response message includes information about a user plane function corresponding to the ML model requested to be transmitted by the first request message; or,
  • the communication interface is used to receive a second request message sent by the AF, where the second request message is used to request information about a user plane function, where the user plane function is used to transmit an ML model with a terminal using a user plane session; and to send the information about the user plane function to the AF.
  • a communication system comprising: a terminal and a network side device, wherein the terminal can be used to execute the steps of the transmission method described in the first aspect, and the network side device can be used to execute the steps of the transmission method described in the second aspect or the third aspect.
  • a readable storage medium wherein a program or instruction is stored on the readable storage medium.
  • the program or instruction is executed by the processor, the steps of the transmission method as described in the first aspect, the second aspect, or the third 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 instructions to implement the steps of the transmission method described in the first aspect, the second aspect, or the third 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 transmission method as described in the first aspect, the second aspect, or the third aspect.
  • the terminal sends a first request message to the AF requesting the transmission of the ML model, and receives a first response message sent by the AF in response to the first request message, wherein the first response message includes information about the user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • the terminal can use the AF to obtain the transmission address information or storage location information of the ML model on the network side, thereby dynamically downloading the ML model from the user plane function based on the user plane session and/or uploading the ML model by the user plane function, thereby clarifying the transmission method of the ML model between the terminal and the network side.
  • FIG1a is a block diagram of a wireless communication system applicable to an embodiment of the present application.
  • Figure 1b is a flowchart of model training
  • FIG1c is a flow chart of PDU session establishment
  • FIG2 is a flow chart of a transmission method provided in an embodiment of the present application.
  • FIG3 is a second flow chart of a transmission method provided in an embodiment of the present application.
  • FIG4 is a third flow chart of a transmission method provided in an embodiment of the present application.
  • FIG5 is a fourth flow chart of a transmission method provided in an embodiment of the present application.
  • FIG6 is one of the structural diagrams of a transmission device provided in an embodiment of the present application.
  • FIG7 is a second structural diagram of a transmission device provided in an embodiment of the present application.
  • FIG8 is a third structural diagram of a transmission device provided in an embodiment of the present application.
  • FIG10 is a structural diagram of a terminal provided in an embodiment of the present application.
  • first, second and the like in the specification and claims of this application are used to distinguish similar objects. It is not used to describe a specific order or sequence. It should be understood that the terms used in this way are interchangeable where appropriate, so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein, 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 means at least one of the connected objects, and the character “/" generally indicates that the objects associated with each other are in an "or” relationship.
  • LTE Long Term Evolution
  • LTE-Advanced, 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
  • 5G communication system for example purposes, and 5G terms are used in most of the following descriptions, but these technologies can also be applied to applications other than 5G system applications, such as the 6th Generation (6G) communication system.
  • 6G 6th Generation
  • 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 wireless local area network (WLAN) access point or a WiFi node, etc.
  • WLAN wireless local area network
  • 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 B node, a home evolved B node, a transmission reception point (TRP) or other appropriate terms 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 user server (Home Subscriber Server, HSS), centralized network configuration (CNC), network storage function (Network Repository Function, NRF), network exposure function (Network Exposure Function, NEF), local NEF (Local NEF, L-NEF), binding support function (Binding Support Function, BSF), application function (Application Function, AF), etc. It should be noted that in the
  • the model reasoning function is based on the model, takes the reasoning data as input, and obtains the reasoning output, such as air interface indicator prediction.
  • the model reasoning function has lower requirements for hardware performance and computing power than the model training function.
  • NPU Neuro Processing Unit
  • the data that needs to be transmitted includes at least user data and control signaling.
  • the transmission characteristics of user data and control signaling are obviously different. For example, the amount of user data may be very large, while the amount of control signaling data is relatively smaller. Control signaling requires high latency and reliability. Control signaling is visible to the network and terminates in the network, requiring the network to process it, while the user data content is transparent to the network.
  • the user plane is transmitted through the Protocol Data Unit (PDU) session.
  • PDU Protocol Data Unit
  • Each PDU session can contain one or more Quality of Service (QoS) flows.
  • QoS Quality of Service
  • Each QoS flow can provide a quality of service guarantee for the user.
  • Step 1 When the User Equipment (UE) needs to establish a PDU session, the UE sends a non-access stratum (NAS) message to the AMF, which contains the parameters corresponding to the PDU session to be established, for example:
  • NAS non-access stratum
  • DNN Data Network Name
  • DN data network
  • PDU session type such as IPv4, IPv6, IPv4v6;
  • SSC Session and Service Continuity
  • Step 2 AMF selects the session management function (SMF) according to the information provided by the UE in the NAS message, such as DNN, S-NASAI, Request Type, etc., and sends a Nsmf_PDUSession_CreateSMContext Request to SMF, which includes the AMF identity document (ID), UE location information, access type (Access type), Radio Access Technology (RAT) type, DNN carried by the UE, S-NASAI, Request Type, N1 SM container, etc.
  • SMF session management function
  • SMF sends Nsmf_PDUSession_CreateSMContext Response to AMF to inform AMF whether it accepts the request. If not, the subsequent steps can be omitted.
  • Step 3 SMF sends Npcf_SMPolicyControl_Create request to PCF to establish SM policy association (Policy Association) and obtain the establishment policy (Policy) of PDU session from PCF.
  • the request contains information such as DNN, S-NSSAI, RAT Type, PDU Session Type, Request Type, Access Type, UE location information, etc.
  • Step 4 PCF sends Npcf_SMPolicyControl_Create response to SMF, which contains policy information for establishing PDU session, such as available session parameters and QoS parameters.
  • Step 5 SMF selects the user plane function (UPF) and initiates the N4 Session establishment process to the UPF to establish the N4 session between the SMF and UPF, so as to control the UPF and the base station to establish a data transmission channel for user transmission.
  • UPF user plane function
  • UPF sends a response message to SMF.
  • Step 6 SMF sends a Namf_Communication_N1N2MessageTransfer message to AMF, which contains the N1 SM container, and the N1 SM container contains the PDU Session Establishment Accept message.
  • AMF sends a response message to SMF.
  • Step 7. AMF sends N1 SM container to UE via NAS message, which includes PDU Session Establishment Accept message.
  • Figure 2 is one of the flow charts of a transmission method provided in an embodiment of the present application, and the method is applied to a terminal. As shown in Figure 2, the method includes the following steps:
  • Step 201 The terminal sends a first request message to the AF, where the first request message is used to request transmission of an ML model.
  • the ML model may also be an AI model, which is not specifically limited in the embodiments of the present application.
  • the first request message includes at least one of the following:
  • Model identification information of the ML model is used to uniquely identify the requested model instance within a certain range, that is, to identify which specific model is requested;
  • Model function type information of the ML model
  • first indication information where the first indication information is used to indicate that the ML model needs to be transmitted using a user plane session
  • IP Internet Protocol
  • model identifier is used to uniquely identify the requested model instance within a certain range (such as within a public land mobile network (PLMN)), that is, to identify which specific model is requested;
  • model function type information is used to characterize the function or purpose of the ML model, for example, it may include model type (model type), data analysis task type (analytics ID), model functionality identifier (model functionality ID), etc.
  • the IP address information of the terminal includes the IP address information of the terminal corresponding to the user plane session for transmitting the ML model.
  • the user plane session may be a PDU session, and the PDU session may be an existing PDU session or a newly created PDU session.
  • the ML model-related capability information of the terminal includes at least one of the following:
  • Model identification information supported by the terminal such as UE supported model ID, where the model ID is used to uniquely identify a model instance within a certain range;
  • Model function type information supported by the terminal such as UE supported model type, UE supported analytics ID or model functionality ID, used to characterize the functions or uses of the supported models;
  • Second indication information for indicating that the terminal supports model downloading and/or uploading, for example, whether the terminal supports the capability of model downloading and/or uploading.
  • the capability indication information may also be different based on different model identifiers (model ID) or different model types (model type);
  • the terminal transmits capability information of the model through the user plane, such as whether the terminal supports uploading and/or downloading the ML model through the user plane session. Further, the capability information may also be different based on different model IDs or different based on different model types;
  • the model storage space information of the terminal for example, the initial or remaining storage space size that can store the model in the terminal.
  • the requirement information of the ML model includes at least one of the following:
  • the transmission delay requirement information of the model is used to indicate the deadline, maximum delay, and other information of the feedback model to the peer end;
  • Model size requirement information used to indicate the storage space requirement information of the model fed back by the peer end
  • Shareable indication information is used to indicate that the model needs to support sharing between devices from different manufacturers or with different functions
  • Identity qualification information used to limit the identity of the model provider, for example, used to limit the manufacturer that generates the model to one or more specific manufacturers, or limit the model provider to one or more specific devices;
  • Model expression limitation information for example, it is used to limit the model to be expressed in certain specific languages or based on certain specific AI frameworks.
  • Commonly used model languages include Open Neural Network Exchange (ONNX) and PyTorch Neural Network Exchange (PNNX).
  • AI frameworks can be TensorFlow, Pytorch, etc.
  • Model performance requirement information is used to indicate the terminal's minimum and/or maximum requirements for the accuracy of the model; the accuracy of the model can be expressed in terms of the mean absolute error (MAE), minimum mean squared error (MMSE) or other forms of the model's prediction results.
  • MAE mean absolute error
  • MMSE minimum mean squared error
  • the model uses scope-limited information to indicate the model's valid area, applicable DNN, applicable slice, valid time, and other scope information.
  • Step 202 The terminal receives a first response message sent by the AF, where the first response message includes information about a user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • the user plane function supports a function of transmitting the ML model using a user plane session.
  • the information of the user plane function includes at least one of the following:
  • the address information of the user plane function such as IP address, Media Access Control (MAC) address, etc.;
  • a Uniform Resource Locator for indicating the location of the user plane function storage model file, for example, for indicating the location of the user plane function storage model file;
  • Security information related to the user plane tunnel such as security certificates, etc.
  • the terminal sends a first request message to the AF requesting the transmission of the ML model, and receives a first response message sent by the AF in response to the first request message, wherein the first response message includes information about the user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • the terminal can use the AF to obtain the transmission address information or storage location information of the ML model on the network side, thereby dynamically downloading the ML model from the user plane function based on the user plane session and/or uploading the ML model by the user plane function, thereby clarifying the transmission method of the ML model between the terminal and the network side.
  • the first response message also includes third indication information, and the third indication information is used to indicate that the ML model needs to be transmitted using a user plane session, that is, the ML model requested in the first request message is transmitted.
  • the user plane session may be a PDU session.
  • the terminal can also learn based on the third indication information that the ML model needs to be transmitted using a user plane session, thereby clarifying the method of transmitting the ML model between the terminal and the network side.
  • the method further comprises:
  • the terminal determines the user plane session for transmitting the ML model.
  • the terminal may determine the user plane session for transmitting the ML model before sending the first request message to the AF, or the terminal may determine the user plane session for transmitting the ML model after receiving the first response message sent by the AF.
  • the user plane session may be a PDU session, and the PDU session may be a PDU session corresponding to an existing data service, or a newly created PDU session for transmitting the ML model.
  • the terminal may first determine its own capability of supporting user plane transmission, thereby ensuring that ML transmission can be performed based on the user plane session.
  • the method further includes:
  • the terminal transmits the ML model using a user plane session and a user plane function corresponding to the information of the user plane function.
  • the terminal may transmit the ML model using the user plane session and the user plane function corresponding to the information of the user plane function.
  • the terminal can determine the corresponding user plane function according to the information of the user plane function corresponding to the transmission ML model included in the first response message, and then the terminal uses the user plane session, such as an existing PDU session, or a new PDU session, to transmit the ML model to the user plane function corresponding to the user plane function information.
  • the user plane session such as an existing PDU session, or a new PDU session
  • the terminal can transmit the ML model with the user plane function based on the user plane session, thereby clarifying the transmission method of the ML model between the terminal and the network side.
  • transmitting the ML model includes at least one of the following:
  • the terminal uploads information of the ML model to the user plane function.
  • the terminal downloads information of the ML model from the user plane function.
  • some functions of the ML model are trained on the terminal side, and the terminal can upload this part of the trained ML model to the user plane function, and then the network side can combine or continue training based on the trained ML model part uploaded by the terminal with the model part on its side to obtain a complete ML model, and the terminal can further download the complete ML model information from the network side (user plane function).
  • the ML model is trained on the network side, in which case the terminal downloads the ML model information from the user plane function.
  • the information of the ML model includes at least one of the following:
  • Model parameter information of the ML model
  • Model identification information of the ML model
  • Model function type information of the ML model
  • the terminal can transmit the information of the above-mentioned ML model to the user plane function based on the user plane session, for example, the terminal uploads the model structure information, model parameter information, etc. of the ML model to the user plane function, or the terminal can download the model structure information, model identification information, model function type information, etc. of the ML model from the user plane function, which will not be described in detail here.
  • the method further includes:
  • the terminal establishes a secure tunnel with the user plane function based on the user plane session.
  • the secure tunnel may be an encrypted transmission channel, so that the terminal can transmit the ML model to the user plane function based on the encrypted transmission channel, thereby ensuring the security of the ML model transmission.
  • the terminal transmits the ML model using a user plane session and a user plane function corresponding to the information of the user plane function, including:
  • the terminal transmits the ML model using the secure tunnel and the user plane function.
  • the terminal when the terminal establishes a secure tunnel with the user plane function based on the user plane session, the terminal uses the secure tunnel to transmit the ML model to the user plane function, thereby effectively ensuring the security of the ML model transmission.
  • the terminal transmits the ML model using a user plane session or tunnel with the user plane function, including:
  • the terminal determines, according to the information of the user plane function, a destination address of the opposite end corresponding to transmitting the ML model;
  • the terminal uses the user plane session or tunnel to perform model transmission with the opposite end destination address.
  • the terminal determines, according to the information of the user plane function, a destination address of a peer end corresponding to transmitting the ML model, including at least one of the following:
  • the terminal determines the address of the user plane function according to the fully qualified domain name (FQDN) of the user plane function, and the terminal uses the address of the user plane function as the opposite end destination address;
  • FQDN fully qualified domain name
  • the terminal uses the address of the user plane function as the opposite end destination address
  • the terminal stores the uniform resource locator URL of the ML model according to the user plane function, and determines the opposite end destination address.
  • the terminal may directly use the address of the user plane function as the destination address of the opposite end for transmitting the ML model, or the terminal may determine the address of the user plane function based on the FQDN of the user plane function and use the address of the user plane function as the destination address of the opposite end; or the terminal determines the destination address of the opposite end according to the URL of the user plane function storing the ML model. In this way, the terminal can flexibly determine the destination address of the opposite end for transmitting the ML model in different ways.
  • Figure 3 is a second flow chart of a transmission method provided in an embodiment of the present application, and the method is applied to AF. As shown in Figure 3, the method includes the following steps:
  • Step 301 The AF receives a first request message sent by a terminal, where the first request message is used to request transmission of an ML model.
  • Model identification information of the ML model
  • Model function type information of the ML model
  • first indication information where the first indication information is used to indicate that the ML model needs to be transmitted using a user plane session
  • the IP address information of the terminal is the IP address information of the terminal.
  • the IP address information of the terminal includes the IP address information of the terminal corresponding to the user plane session used to transmit the ML model.
  • Second indication information used to indicate that the terminal supports model downloading and/or uploading
  • the terminal transmits capability information of the model via the user plane
  • the requirement information of the ML model includes at least one of the following:
  • Shareable indication information used to indicate that the model needs to support sharing
  • the model's expression method limits information
  • Step 302 The AF sends a first response message to the terminal, where the first response message includes information about a user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • the AF may directly send the first response message to the terminal, wherein the message includes information about the user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • the AF after receiving the first request information sent by the terminal, the AF sends a second request information to the network side, where the second request information is used to request information about the user plane function, and the user plane function is used to transmit the ML model with the terminal using a user plane session.
  • the AF sends the second request information to the network side, which may include at least one of the following methods:
  • the AF sends the second request message to a network exposure function (NEF), and obtains information of a user plane function corresponding to the transmission ML model from the NEF;
  • NEF network exposure function
  • the AF sends the second request message to the model application platform in the network, and obtains information about the user plane function corresponding to the transmission ML model from the model application platform.
  • the AF sends a first response message to the terminal, where the first response message includes information of the user plane function corresponding to the transmission ML model.
  • model application platform may include model database network elements (such as analysis data storage function network element (Analytics Data Repository Function, ADRF)), model training logic function network element (Model Training Logical Function, MTLF), network data analysis function network element (Network Data Analytics Function, NWDAF), etc.
  • ADRF analysis data storage function network element
  • MTLF model training logic function network element
  • NWDAF Network Data Analytics Function
  • the user plane function refers to a network element or functional module that stores a model or can provide a model instance.
  • the user plane function can be an independently deployed database network element (such as ADRF, or a model application platform, a model store, etc.), and the model generated by MTLF or other devices can be stored in the database network element.
  • the user plane function can be a functional module that is integrated with or plug-in to the model training function (MTLF).
  • the information of the user plane function includes at least one of the following:
  • the AF after the AF receives a first request message sent by the terminal to request the transmission of an ML model, the AF sends a first response message to the terminal, where the first response message includes information about the user plane function corresponding to the ML model requested to be transmitted by the first request message, so that the terminal can use the AF to obtain the transmission address information or storage location information of the ML model on the network side, thereby dynamically downloading and/or uploading the ML model from the user plane function based on the user plane session and uploading the ML model to the user plane function, thereby clarifying the transmission method of the ML model between the terminal and the network side.
  • the method further comprises:
  • the AF sends a second request message to the network side, where the second request message is used to request information about the user plane function, where the user plane function is used to transmit the ML model to the terminal using a user plane session.
  • the AF after receiving the first request message sent by the terminal, the AF sends a second request message to the network side to request information about the user plane function used to transmit the ML model between the terminal using the user plane session, and then obtains the information about the user plane function from the network side.
  • the AF sends a second request message to the network side, including any one of the following:
  • the AF sends the second request message to a network exposure function (NEF);
  • NEF network exposure function
  • the AF sends the second request message to the model application platform.
  • the AF may request the NEF or the model application platform to obtain the information of the user plane function.
  • the method further comprises any one of the following:
  • the AF receives the information of the user plane function sent by the NEF;
  • the AF receives the information of the user plane function sent by the preset model application platform.
  • the AF may send a first response message to the terminal, where the first response message includes the information about the user plane function.
  • the method further includes: the AF receiving fourth indication information sent by the NEF or the model application platform, where the fourth indication information is used to indicate that the ML model needs to be transmitted using a user plane session.
  • the AF sends the second request message to the NEF to request information about the user plane function used to transmit the ML model with the terminal using a user plane session.
  • the NEF sends the user plane function information to the AF, and may also send fourth indication information to the AF to indicate that the ML model needs to be transmitted using a user plane session, thereby indicating through the network side that the terminal and the network side need to use a user plane session to achieve transmission of the ML model, thereby clarifying the transmission method of the ML model between the terminal and the network side.
  • the model application platform can also send the fourth indication information to the AF when sending the information of the user plane function to the AF, so as to make it clear that the transmission of the ML model between the terminal and the network side needs to be achieved through a user plane session.
  • Model identification information of the ML model
  • Model function type information of the ML model
  • the fifth indication information is used to indicate that the ML model needs to be transmitted using a user plane session
  • the IP address information of the terminal is the IP address information of the terminal.
  • the message content included in the second request message may be the same as the message content included in the first request message.
  • the AF after receiving the first request message sent by the terminal, the AF generates a second request message according to the message content included in the first request message, and the second request message also includes the message content of the first request message, so that the AF requests the network side to obtain the information of the user plane function.
  • the AF may also forward the first request message to the network side to request the network side to obtain the information of the user plane function.
  • the terminal uploads information of the ML model to the user plane function.
  • the terminal downloads information of the ML model from the user plane function.
  • the information of the ML model includes at least one of the following:
  • Model identification information of the ML model
  • Model function type information of the ML model
  • the terminal uploads the model structure information, model parameter information, etc. of the ML model to the user plane function, or the terminal may download the model structure information, model identification information, model function type information, etc. of the ML model from the user plane function, which will not be described in detail here.
  • Figure 4 is a flowchart of a transmission method provided in an embodiment of the present application, which is applied to a model application platform. As shown in Figure 4, the method includes the following steps:
  • the model application platform receives the second request message sent by the AF in any of the following ways:
  • the model application platform directly receives the second request message sent by the AF;
  • the model application platform receives the second request message from the AF forwarded by the NEF.
  • the second request message includes at least one of the following:
  • Model identification information of the ML model
  • Model function type information of the ML model
  • the fifth indication information is used to indicate that the ML model needs to be transmitted using a user plane session
  • the IP address information of the terminal is the IP address information of the terminal.
  • message content included in the second request message may be the same as the message content included in the first request message in the above embodiment.
  • Step 402 The model application platform sends the information of the user plane function to the AF.
  • the model application platform sends the information of the user plane function to the AF in any of the following ways:
  • the model application platform directly sends the information of the user plane function to the AF;
  • the model application platform sends the information of the user plane function to the AF via the NEF.
  • the information of the user plane function includes at least one of the following:
  • the model application platform when the model application platform receives a second request message sent by the AF for requesting to obtain information about a user plane function, the model application platform responds to the second request message and sends the information about the user plane function to the AF, where the user plane function is used to transmit the ML model with the terminal using a user plane session. Then, the terminal can use AF to obtain the transmission address information or storage location information of the ML model on the network side, so as to dynamically download and/or upload the ML model from the user plane function based on the user plane session, thereby clarifying the transmission method of the ML model between the terminal and the network side.
  • the method further comprises:
  • the model application platform sends fourth indication information to the AF, where the fourth indication information is used to indicate that the ML model needs to be transmitted using a user plane session.
  • the model application platform may send fourth indication information to the AF before sending the information of the user plane function to the AF, so that the AF can be informed that the user plane session needs to be used to transmit the ML model, and then indicate through the network side that the user plane session is needed between the terminal and the network side to realize the transmission of the ML model, thereby clarifying the transmission method of the ML model between the terminal and the network side.
  • the model application platform may also send fourth indication information to the AF after sending the information of the user plane function to the AF.
  • the method further includes:
  • the model application platform determines information about the user plane function.
  • the model application platform can directly determine the information of the user plane function, such as the FDQN of the user plane function, the address of the user plane function, the URL location information of the user plane function storing the ML model, etc., and then send the information of the user plane function to the AF.
  • the information of the user plane function can be determined through the network side, so as to realize the transmission of the ML model between the terminal and the user plane function.
  • the model application platform determines the information of the user plane function, including:
  • the model application platform determines the information of the user plane function when determining to use the user plane session to transmit the ML model.
  • the model application platform when the model application platform receives the second request message sent by the AF, if the model application platform determines to use a user plane session (e.g., a PDU session) to transmit the ML model, the model application platform determines the information of the user plane function and sends the information of the user plane function to the AF. In this way, it is determined through the network side that the user plane session is required to realize the transmission of the ML model between the terminal and the network side, thereby clarifying the transmission method of the ML model between the terminal and the network side.
  • a user plane session e.g., a PDU session
  • the method may further include:
  • the model application platform determines the information of the user plane function according to the second request message.
  • the model application platform may determine the information of the user plane function, such as the user plane function FQDN, the address information of the user plane function, a URL indicating the location of the model file stored in the user plane function, etc., upon receiving the second request message sent by the AF for requesting to obtain the information of the user plane function, and then send the information of the user plane function to the AF.
  • the model application platform can determine the information of the user plane function based on the second request message, and then clarify the transmission method of the ML model between the terminal and the user plane function.
  • the model application platform determines the information of the user plane function according to the second request message, including:
  • the model application platform sends a discovery message for determining the user plane function corresponding to the ML model to the network storage function, wherein the discovery message includes at least one of the following: identification information of the ML model, function information of the ML model, type information of the ML model, and indication information of a user plane transmission model;
  • the model application platform receives the information of the user plane function sent by the network storage function.
  • the model application platform may be a case where it receives a second request message sent by the AF for requesting to obtain information about user plane functions, and sends the discovery message to the network storage function.
  • the discovery message includes the identification information of the ML model and the functional information of the ML model, so that the network storage function can determine which ML model needs to be transmitted and what type of model it is based on the discovery message, and determine the user plane function information based on the discovery message.
  • the user plane function information includes the DNN, S-NSSAI, etc. corresponding to the user plane session for transmitting the ML model determined based on the identification information of the ML model and the functional information of the ML model, and send the user plane function information to the model application platform.
  • the model application platform can send the user plane function information to the AF, so that the AF can clearly define the DNN, S-NSSAI, etc. corresponding to the user plane session for transmitting the ML model, so as to clarify the user plane session for transmitting the ML model between the terminal and the network side, thereby ensuring smooth transmission of the ML model between the terminal and the network side.
  • the model application platform determines the information of the user plane function according to the second request message, including:
  • the model application platform sends a third request message to the model storage network element, where the third request message is used to request the model storage network element to feedback information about the user plane function corresponding to the ML model;
  • the model application platform receives information about user plane functions corresponding to the ML model sent by the model storage network element, where the information about user plane functions includes a URL where the model storage network element stores the ML model.
  • the model application platform may be a case where, upon receiving a second request message sent by the AF for requesting information about user plane functions, the model application platform sends a third request message to the model storage network element for requesting the model storage network element to feedback information about the user plane function corresponding to the ML model, wherein the model application platform may determine the model storage network element based on information of a preset model storage network element, or the model application platform may also determine the model storage network element based on historical ML model storage records.
  • the model storage network element If the model storage network element is the user plane function, the model storage network element sends the URL of the ML model stored in it to the model application platform, and further, the model application platform sends the URL of the ML model to the AF, so that the AF can clearly identify the location where the model storage network element stores the ML model, so as to ensure smooth transmission of the ML model between the terminal and the network side.
  • FIG. 5 is a fourth flow chart of a transmission method provided in an embodiment of the present application. As shown in FIG. 5 , the method includes the following steps:
  • Step 0a UE selects an existing PDU session
  • Step 0b The UE establishes a new PDU session; wherein, step 0a and step 0b may be performed alternatively;
  • Step 0y AF sends a model storage request to the model application platform (model store, such as ADRF), which includes model information; the model application platform feeds back the model storage address to AF;
  • model application platform model store, such as ADRF
  • steps 0a to 0y are optional steps
  • Step 1 The UE sends a model request (i.e., the first request message mentioned above) to the AF.
  • the request includes the model ID, model Model function ID, UE model-related capability information, etc.
  • step 2a AF sends a model acquisition request to the model application platform, where the request includes the terminal IP address;
  • step 2b The model application platform feeds back a model acquisition response to the AF, where the response includes information about the user plane function corresponding to the model;
  • Step 3 The AF sends a model response message (that is, the first response message mentioned above) to the terminal, which includes information about the user plane function corresponding to the model;
  • step 4a The UE selects an existing PDU session; or, step 4b.
  • the UE creates a new PDU session; wherein, step 4a or 4b and the above step 0a or 0b may be performed in a selected manner;
  • step 5 establishing a secure channel with the user plane function
  • Step 6 Based on the terminal IP address, the user plane function interacts with the UE with model-related information, including model ID, model information, such as model structure information and/or model parameter information;
  • model-related information including model ID, model information, such as model structure information and/or model parameter information;
  • Step 7 Based on the information of the user plane function, the UE exchanges model-related information, including model ID, model information, etc., such as model structure information and/or model parameter information, with the user plane function through a PDU session.
  • model-related information including model ID, model information, etc., such as model structure information and/or model parameter information
  • the transmission method provided in the embodiment of the present application can be executed by a transmission device.
  • the transmission device provided in the embodiment of the present application is described by taking the transmission method executed by the transmission device as an example.
  • the transmission device 600 includes:
  • a first sending module 601 is used to send a first request message to the application function AF, where the first request message is used to request the transmission of a machine learning ML model;
  • the first receiving module 602 is configured to receive a first response message sent by the AF, where the first response message includes information about a user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • the device further comprises:
  • the first transmission module is used to transmit the ML model using a user plane session and a user plane function corresponding to the information of the user plane function.
  • transmitting the ML model includes at least one of the following:
  • the information of the ML model includes at least one of the following:
  • Model parameter information of the ML model
  • Model identification information of the ML model
  • Model function type information of the ML model
  • the device further comprises:
  • the first determining module is configured to determine the user plane session for transmitting the ML model.
  • the first request message includes at least one of the following:
  • Model identification information of the ML model
  • Model function type information of the ML model
  • first indication information where the first indication information is used to indicate that the ML model needs to be transmitted using a user plane session
  • the IP address information of the device is the IP address information of the device.
  • the IP address information of the device includes IP address information of the device corresponding to a user plane session used to transmit the ML model.
  • the ML model-related capability information of the device includes at least one of the following:
  • second indication information used to indicate that the device supports model downloading and/or uploading
  • the device transmits capability information of the model via a user plane
  • the model of the device stores spatial information.
  • the requirement information of the ML model includes at least one of the following:
  • Shareable indication information used to indicate that the model needs to support sharing
  • the model's expression method limits information
  • the device further comprises:
  • An establishing module is used to establish a secure tunnel with the user plane function based on the user plane session.
  • the first transmission module is further used for:
  • the ML model is transmitted using the secure tunnel and the user plane function.
  • the first response message further includes third indication information, where the third indication information is used to indicate that the ML model needs to be transmitted using a user plane session.
  • the information of the user plane function includes at least one of the following:
  • a uniform resource locator URL for indicating the location of the user plane function storage model file
  • the device sends a first request message to the AF requesting the transmission of the ML model, and receives a first response message sent by the AF in response to the first request message, wherein the first response message includes information about the user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • the device can use the AF to obtain the transmission address information or storage location information of the ML model on the network side, thereby dynamically downloading the ML model from the user plane function and/or uploading the ML model by the user plane function based on the user plane session, thereby clarifying the transmission method of the ML model between the network side.
  • the transmission device 600 in the embodiment of the present application can be an electronic device, such as an electronic device with an operating system, or a component in an electronic device, such as an integrated circuit or a chip.
  • the electronic device can be a terminal, or it can be other devices other than a terminal.
  • the terminal can include but is not limited to the types of terminals 11 listed above, and other devices can be servers, network attached storage (NAS), etc., which are not specifically limited in the embodiment of the present application.
  • the transmission device 600 provided in the embodiment of the present application can implement each process implemented by the method embodiment described in Figure 2 and achieve the same technical effect. To avoid repetition, it will not be repeated here.
  • FIG. 7 is a second structural diagram of a transmission device provided in an embodiment of the present application.
  • the transmission device 700 includes:
  • the second receiving module 701 is used to receive a first request message sent by a terminal, where the first request message is used to request transmission of an ML model;
  • the second sending module 702 is configured to send a first response message to the terminal, where the first response message includes information about a user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • the second sending module 702 is further used to:
  • a second request message is sent to the network side, where the second request message is used to request information about the user plane function, where the user plane function is used to transmit the ML model to the terminal using a user plane session.
  • transmitting the ML model includes at least one of the following:
  • the second sending module 702 is further configured to perform any one of the following:
  • the second receiving module 701 is further configured to perform any one of the following:
  • the device sends the second request message to the NEF, receiving information about the user plane function sent by the NEF;
  • the device When the device sends the second request message to the model application platform, the device receives the information of the user plane function sent by the preset model application platform.
  • the second receiving module 701 is further used for:
  • the model application platform includes at least one of the following: a model database network element, a model training logic function network element, and a network data analysis function network element.
  • the second request message includes at least one of the following:
  • Model identification information of the ML model
  • Model function type information of the ML model
  • the fifth indication information is used to indicate that the ML model needs to be transmitted using a user plane session
  • the IP address information of the terminal is the IP address information of the terminal.
  • the information of the user plane function includes at least one of the following:
  • the device after the device receives a first request message sent by a terminal for requesting the transmission of an ML model, the device sends a first response message to the terminal, where the first response message includes information about a user plane function corresponding to the ML model requested to be transmitted by the first request message, so that the terminal can obtain the transmission address information or storage location information of the ML model on the network side, thereby dynamically downloading and/or uploading the ML model from the user plane function based on the user plane session and uploading the ML model to the user plane function, thereby clarifying the transmission method of the ML model between the terminal and the network side.
  • the transmission device 700 provided in the embodiment of the present application can implement each process implemented by the method embodiment described in Figure 3 and achieve the same technical effect. To avoid repetition, it will not be repeated here.
  • FIG. 8 is a third structural diagram of a transmission device provided in an embodiment of the present application.
  • the transmission device 800 includes:
  • the third receiving module 801 is used to receive a second request message sent by the AF, where the second request message is used to request to obtain information about a user plane function, where the user plane function is used to transmit an ML model with the terminal using a user plane session;
  • the third sending module 802 is configured to send the information of the user plane function to the AF.
  • the second request message includes at least one of the following:
  • Model identification information of the ML model
  • Model function type information of the ML model
  • the fifth indication information is used to indicate that the ML model needs to be transmitted using a user plane session
  • the IP address information of the terminal is the IP address information of the terminal.
  • the third sending module 802 is further used to:
  • the device further comprises:
  • the second determining module is used to determine the information of the user plane function.
  • the second determining module is further used for:
  • the device further comprises:
  • the third determining module is used to determine the information of the user plane function according to the second request message.
  • the third determining module is further used to:
  • the discovery message including at least one of the following: identification information of the ML model, function information of the ML model, type information of the ML model, and indication information of a user plane transmission model;
  • the third determining module is further used to:
  • the device when the device receives the second request message sent by the AF for requesting to obtain the information of the user plane function, the device responds to the second request message and sends the information of the user plane function to the AF.
  • the user plane function is used to transmit the ML model with the terminal using the user plane session, so that the terminal can use the AF to obtain the transmission address information or storage location information of the ML model on the network side, thereby dynamically downloading and/or uploading the ML model from the user plane function based on the user plane session, thereby clarifying the connection between the terminal and the network side. How to transfer ML models between different platforms
  • the transmission device 700 provided in the embodiment of the present application can implement each process implemented by the method embodiment described in Figure 3 and achieve the same technical effect. To avoid repetition, it will not be repeated here.
  • the embodiment of the present application further provides a communication device 900, including a processor 901 and a memory 902, wherein the memory 902 stores a program or instruction that can be run on the processor 901.
  • the communication device 900 is a terminal
  • the program or instruction is executed by the processor 901 to implement the various steps of the transmission method embodiment described in FIG2 above, 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 transmission method embodiment described in FIG3 or FIG4 above, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.
  • An embodiment of the present application also provides a terminal, including a processor and a communication interface, the communication interface is used to send a first request message to an application function AF, the first request message is used to request the transmission of a machine learning ML model; and is used to receive a first response message sent by the AF, the first response message includes information about the user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • 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 the terminal embodiment, and can achieve the same technical effect.
  • Figure 10 is a schematic diagram of the hardware structure of a terminal that implements 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 of a processor 1010.
  • the terminal 1000 may also include a power source (such as a battery) for supplying power to each component, and the power source may be logically connected to the processor 1010 through a power management system, so as to implement functions such as managing charging, discharging, and power consumption management through the power management system.
  • a power source such as a battery
  • the terminal structure shown in FIG10 does not constitute a limitation on the terminal, and the terminal may 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, and the graphics processor 10041 processes the image data of the static picture or video obtained by the image capture device (such as a camera) in the video capture mode or the image capture mode.
  • the display unit 1006 may include a display panel 10061, and the display panel 10061 may be configured in the form of a liquid crystal display, an organic light emitting diode, etc.
  • the user input unit 1007 includes a touch panel 10071 and at least one of other input devices 10072.
  • the touch panel 10071 is also called a touch screen.
  • the touch panel 10071 may include two parts: a touch detection device and a touch controller.
  • Other input devices 10072 may include, but are not limited to, a physical keyboard, function keys (such as a volume control key, a switch key, etc.), a trackball, a mouse, and a joystick, which will not be repeated here.
  • the RF unit 1001 can transmit the data to the processor 1010 for processing; in addition, the RF unit 1001 can send uplink data to the network side device.
  • the RF unit 1001 includes but is not limited to an antenna, an 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 can mainly include storage A first storage area for 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 instructions 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 both volatile and non-volatile memories.
  • 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 static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate synchronous dynamic random access memory (DDRSDRAM), an enhanced synchronous dynamic random access memory (ESDRAM), a synchronous link dynamic random access memory (SLDRAM) and a direct memory bus random access memory (DRRAM).
  • 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 integrates 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, 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 used to send a first request message to the application function AF, where the first request message is used to request transmission of a machine learning ML model;
  • the AF receiving a first response message sent by the AF, wherein the first response message includes information about the user plane function corresponding to the ML model requested to be transmitted by the first request message
  • the terminal sends a first request message to the AF requesting the transmission of the ML model, and receives a first response message sent by the AF in response to the first request message, wherein the first response message includes information about the user plane function corresponding to the ML model requested to be transmitted by the first request message.
  • the terminal can use the AF to obtain the transmission address information or storage location information of the ML model on the network side, thereby dynamically downloading the ML model from the user plane function based on the user plane session and/or uploading the ML model by the user plane function, thereby clarifying the transmission method of the ML model between the terminal and the network side.
  • the embodiment of the present application further provides a network side device.
  • the network side device 1100 includes: a processor 1101, a network interface 1102, and a memory 1103.
  • the network interface 1102 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 1103 and executable on the processor 1101.
  • the processor 1101 calls the instructions or programs in the memory 1103 to execute the method executed by each module shown in Figure 7 or Figure 8, and achieves the same technical effect. To avoid repetition, it will not be repeated here.
  • the embodiment of the present application further provides a readable storage medium, on which a program or instruction is stored.
  • a program or instruction is stored.
  • the program or instruction is executed by a processor, each process of the transmission method embodiment described in FIG. 2 or FIG. 3 or FIG. 4 is implemented. And the same technical effect can be achieved, so in order 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 transmission method embodiments described in Figures 2 or 3 or 4 above, and can achieve the same technical effect. To avoid repetition, they 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, and is executed by at least one processor to implement the various processes of the transmission method embodiments described in Figures 2, 3, or 4 above, and can achieve 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 transmission method described in Figure 2 above, and the network side device can be used to execute the steps of the transmission method described in Figure 3 or Figure 4 above.
  • 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, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present application.
  • a storage medium such as ROM/RAM, a magnetic disk, or an optical disk
  • a terminal which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Medical Informatics (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Databases & Information Systems (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本申请公开了一种传输方法、装置、终端及网络侧设备,属于通信技术领域,本申请实施例的传输方法包括:终端向应用功能AF发送第一请求消息,所述第一请求消息用于请求传输机器学习ML模型;所述终端接收所述AF发送的第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。

Description

传输方法、装置、终端及网络侧设备
相关申请的交叉引用
本申请主张在2022年12月07日在中国提交的中国专利申请No.202211567252.3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,具体涉及一种传输方法、装置、终端及网络侧设备。
背景技术
人工智能(Artificial Intelligence,AI)目前已广泛地应用于各个领域,将AI融入通信***,以显著提升吞吐量、时延及用户容量等技术指标是未来通信网络的重要优化方向。而机器学习(Machine Learning,ML)是一种实现人工智能的方式,终端为了达到最优的通信效果,需要从网络侧下载最新的ML模型。另外,在一些场景下,终端也可能会参与部分模型的训练过程,例如联邦学习中,终端在本地训练生成局部模型后,上传给网络侧进行聚合,这就涉及终端模型的上传问题。
目前,通信***仅为终端提供了普通用户数据传输的解决方法,对于如何实现终端与网络侧之间ML模型的传输尚不明确。
发明内容
本申请实施例提供一种传输方法、装置、终端及网络侧设备,能够解决相关技术中终端与网络侧之间ML模型的传输尚不明确的问题。
第一方面,提供了一种传输方法,包括:
终端向应用功能AF发送第一请求消息,所述第一请求消息用于请求传输机器学习ML模型;
所述终端接收所述AF发送的第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。
第二方面,提供了一种传输方法,包括:
AF接收终端发送的第一请求消息,所述第一请求消息用于请求传输ML模型;
所述AF向所述终端发送第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。
第三方面,提供了一种传输方法,包括:
模型应用平台接收AF发送的第二请求消息,所述第二请求消息用于请求获取用户面功能的信息,所述用户面功能用于与终端之间利用用户面会话传输ML模型;
所述模型应用平台向所述AF发送所述用户面功能的信息。
第四方面,提供了一种传输装置,包括:
第一发送模块,用于向应用功能AF发送第一请求消息,所述第一请求消息用于请求传输机器学习ML模型;
第一接收模块,用于接收所述AF发送的第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。
第五方面,提供了一种传输装置,包括:
第二接收模块,用于接收终端发送的第一请求消息,所述第一请求消息用于请求传输ML模型;
第二发送模块,用于向所述终端发送第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。
第六方面,提供了一种传输装置,包括:
第三接收模块,用于接收AF发送的第二请求消息,所述第二请求消息用于请求获取用户面功能的信息,所述用户面功能用于与终端之间利用用户面会话传输ML模型;
第三发送模块,用于向所述AF发送所述用户面功能的信息。
第七方面,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一方面所述的传输方法的步骤。
第八方面,提供了一种终端,包括处理器及通信接口,其中,所述通信接口用于向应用功能AF发送第一请求消息,所述第一请求消息用于请求传输机器学习ML模型;以及用于接收所述AF发送的第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。
第九方面,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二方面或第三方面所述的传输方法的步骤。
第十方面,提供了一种网络侧设备,包括处理器及通信接口,其中,所述通信接口用于接收终端发送的第一请求消息,所述第一请求消息用于请求传输ML模型;以及用于向所述终端发送第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息;或者,
所述通信接口用于接收AF发送的第二请求消息,所述第二请求消息用于请求获取用户面功能的信息,所述用户面功能用于与终端之间利用用户面会话传输ML模型;以及用于向所述AF发送所述用户面功能的信息。
第十一方面,提供了一种通信***,包括:终端及网络侧设备,所述终端可用于执行如第一方面所述的传输方法的步骤,所述网络侧设备可用于执行如第二方面或第三方面所述的传输方法的步骤。
第十二方面,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或第二方面或第三方面所述的传输方法的步骤。
第十三方面,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,实现如第一方面或第二方面或第三方面所述的传输方法的步骤。
第十四方面,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或第二方面或第三方面所述的传输方法的步骤。
在本申请实施例中,终端通过向AF发送请求传输ML模型的第一请求消息,并接收AF响应于所述第一请求消息发送的第一响应消息,该第一响应消息中包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息,进而终端能够利用AF获得网络侧对于ML模型的传输地址信息或存储位置信息,从而基于用户面会话从用户面功能动态地下载和/或所述用户面功能上传ML模型,进而也就明确了终端与网络侧之间ML模型的传输方式。
附图说明
图1a是本申请实施例可应用的一种无线通信***的框图;
图1b是模型训练的流程图;
图1c是PDU会话建立的流程图;
图2是本申请实施例提供的一种传输方法的流程图之一;
图3是本申请实施例提供的一种传输方法的流程图之二;
图4是本申请实施例提供的一种传输方法的流程图之三;
图5是本申请实施例提供的一种传输方法的流程图之四;
图6是本申请实施例提供的一种传输装置的结构图之一;
图7是本申请实施例提供的一种传输装置的结构图之二;
图8是本申请实施例提供的一种传输装置的结构图之三;
图9是本申请实施例提供的一种通信设备的结构图;
图10是本申请实施例提供的一种终端的结构图;
图11是本申请实施例提供的一种网络侧设备的结构图。
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。
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)***,还可用于其他无线通信***,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***。本申请实施例中的术语“***”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***和无线电技术,也可用于其他***和无线电技术。以下描述出于示例目的描述了5G通信***,并且在以下大部分描述中使用5G术语,但是这些技术也可应用于5G***应用以外的应用,如第6代(6th Generation,6G)通信***。
图1a示出本申请实施例可应用的一种无线通信***的框图。无线通信***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(Vehicle User Equipment,VUE)、行人终端(Pedestrian User Equipment,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。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、无线局域网(Wireless Local Area Network,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(Transmission Reception 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***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,L-NEF)、绑定支持功能(Binding Support Function,BSF)、应用功能(Application Function,AF)等。需要说明的是,在本申请实施例中仅以NR***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。
为更好地理解,以下对本申请实施例中可能涉及的相关概念及原理进行解释说明。
机器学习(Machine Learning,ML)目前在各个领域得到了广泛的应用,将机器学习融入通信网络,显著提升吞吐量、时延以及用户容量等技术指标是通信网络的重要优化方向。基于ML的通信网络优化整体流程如图1b所示。其中,模型训练功能和模型推理功能是关键功能。模型训练功能基于训练数据生成模型,完成模型有效性测试后,将模型部署至模型推理功能;在生成模型的过程中,模型训练功能需要获取并分析大量数据,对硬件性能和算力的要求很高,主要部署在网络侧设备上,如运营商服务器,或第三方服务器。模型推理功能基于所述模型,以推理数据为输入,获得推理输出,如空口指标预测。模型推理功能对硬件性能和算力要求相比模型训练功能更低。随着终端能力不断提升,甚至增加了专用芯片-神经网络处理器(Neural Processing Unit,NPU),模型驱动效率也不断提升,使得在终端侧进行模型推理,甚至部分模型训练成为可能。
本申请实施例中,一种可能的场景是,模型训练在通信网络侧(例如模型训练逻辑功能(Model Training Logical Function,MTLF),或应用功能(Application Function,AF))完成,模型推理在终端侧,此种场景下,终端需要从网络侧下载模型。另一种可能的场景是,部分模型训练在终端侧完成,模型推理在网络侧获其他端侧,终端需要将所训练的模型传递给网络侧。
在移动通信***中,需要传输的数据至少包括用户数据和控制信令,用户数据和控制信令的传输特征明显不同,比如,用户数据可能数据量很大,控制信令数据量相对更小,控制信令要求的时延和可靠性高,控制信令对网络可见并终结在网络,需要网络进行处理,而用户数据内容对网络透明。
5G通信***中,通过协议数据单元(Protocol Data Unit,PDU)会话(session)进行用户面传输的,每个PDU session可以包含一个或多个服务质量(Quality of Service,QoS)流(flow),每个QoS flow可以为用户提供一种服务质量的保证。用户在使用PDU session 进行数据传输前,需要通过控制信令进行PDU session建立过程来建立PDU session,通过PDU session修改过程来增加或删除QoS flow,通过PDU session释放过程删除PDU session。
请参照图1c,PDU session建立过程主要包括以下步骤:
步骤1.当用户设备(User Equipment,UE)需要建立PDU session时,UE向AMF发送非接入层(Non-access stratum,NAS)消息,在该消息中包含待建立的PDU session对应的参数,例如:
数据网名称(Data Network Name,DNN),用于指示该PDU session接入的数据网络(Data Network,DN);
单个网络切片选择辅助信息(Single Network Slice Selection Assistance information,S-NSSAI),用于指示该PDU session对应的网络切片;
请求类型(Request Type),用于指示该PDU session建立过程是初始请求(Initial request),或者紧急PDU会话(emergency PDU session)。
在NAS消息中,包含N1会话管理(Session Management,SM)容器(container),在N1 SM container中,包含PDU session建立请求(establishment request)。在PDU session establishment request中,包含如下信息:
PDU session类型(Type),如IPv4,IPv6,IPv4v6;
会话和业务连续性(Session and Service Continuity,SSC)模式(mode)等。
步骤2.AMF根据UE在NAS消息中提供的信息,例如DNN、S-NASAI、Request Type等信息选择会话管理功能(Session Management Function,SMF),并向SMF发送Nsmf_PDUSession_CreateSMContext Request,在该请求中包含AMF身份标识(Identity document,ID)、UE位置信息、接入类型(Access type)、无线电接入技术(Radio Access Technology,RAT)type、UE携带的DNN、S-NASAI、Request Type、N1 SM container等内容。
SMF向AMF发送Nsmf_PDUSession_CreateSMContext Response,用于通知AMF其是否接受该请求。若不接受,则后续步骤可省略。
步骤3.SMF向PCF发送Npcf_SMPolicyControl_Create请求,用于建立SM策略关联(Policy Association),并从PCF获取PDU session的建立策略(Policy)。在该请求中包含DNN、S-NSSAI、RAT Type、PDU Session Type、Request Type、Access Type、UE位置信息等信息。
步骤4.PCF向SMF发送Npcf_SMPolicyControl_Create response,在该消息中包含建立PDU session的策略信息,例如可使用的session参数和QoS参数等。
步骤5.SMF选择用户面功能(User Plane Function,UPF),并向该UPF发起N4 Session建立(establishment)流程,用于建立SMF与UPF间的N4session,以便控制UPF与基站建立用于为用户传输就的数据传输通道。
UPF向SMF回复响应消息。
步骤6.SMF向AMF发送Namf_Communication_N1N2MessageTransfer消息,在该消息中包含N1 SM container,在N1 SM container中包含PDU Session Establishment接受(Accept)消息。
AMF向SMF发送响应消息。
步骤7.AMF通过NAS消息向UE发送N1 SM container,在N1 SM container中包含PDU Session Establishment Accept消息。
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传输方法进行详细地说明。
请参照图2,图2是本申请实施例提供的一种传输方法的流程图之一,该方法应用于终端。如图2所示,所述方法包括以下步骤:
步骤201、终端向AF发送第一请求消息,所述第一请求消息用于请求传输ML模型。
需要说明地,在一些场景下,所述ML模型也可以是AI模型,本申请实施例对此不做具体限定。
可选地,所述第一请求消息包括以下至少一项:
所述终端的ML模型相关能力信息;
所述ML模型的模型标识信息;该模型标识用于在一定范围内唯一地标识所请求的模型实例,也即标识所请求的模型具体是哪个;
所述ML模型的模型功能类型信息;
所述ML模型的要求信息;
第一指示信息,所述第一指示信息用于指示所述ML模型需要利用用户面会话进行传输;
所述终端的因特网协议(Internet Protocol,IP)地址信息。
需要说明地,所述模型标识(ID)用于在一定范围内(如公共陆地移动网络(Public Land Mobile Network,PLMN)内)唯一地标识所请求的模型实例,也即标识所请求的模型具体是哪个;所述模型功能类型信息用于表征所述ML模型的功能或用途,例如可以是包括模型类型(model type)、数据分析任务类型(analytics ID)、模型功能标识(model functionality ID)等。
可选地,所述终端的IP地址信息包括用于传输所述ML模型的用户面会话对应的所述终端的IP地址信息。其中,所述用户面会话可以是PDU会话,所述PDU会话可以是已有的PDU会话,或者是新建的PDU会话。
可选地,所述终端的ML模型相关能力信息包括以下至少一项:
所述终端支持的模型标识信息,例如UE supported model ID,该model ID用于在一定范围内唯一地标识模型实例;
所述终端支持的模型功能类型信息,例如UE supported model type、UE supported analytics ID或model functionality ID,用于表征所支持模型的功能或用途;
用于指示所述终端支持模型下载和/或上传的第二指示信息,例如终端是否支持模型下载和/或上传的能力,进一步地,该能力指示信息还可以基于模型标识(model ID)的不同而不同,或者是基于模型类型(model type)的不同而不同;
所述终端通过用户面传输模型的能力信息,例如终端是否支持通过用户面会话来上传和/或下载ML模型,进一步地,该能力信息还可以基于model ID的不同而不同,或者是基于model type的不同而不同;
所述终端的模型存储空间信息,例如,终端中初始或剩余的可存储模型的存储空间大小。
可选地,所述ML模型的要求信息包括以下至少一项:
模型的传输时延要求信息,用于指示对端反馈模型的截止时间、最大时延等信息;
模型的大小要求信息,用于指示对端反馈的模型对存储空间的要求信息;
可共享指示信息,用于指示模型需支持在不同厂商或不同功能设备之间共享;
用于限定模型提供者身份的身份限定信息,例如用于限定产生模型的厂商为某个或某些特定厂商,或限定模型提供者为某个或某些特定的设备;
模型表述方式限定信息,例如用于限定模型用某些特定的语言或基于某些特定AI框架(framework)来表述,如常用的模型语言有开放神经网络交换(Open Neural Network Exchange,ONNX)、PyTorch神经网络交换(PyTorch Neural Network Exchange,PNNX)等,AI framework可以是TensorFlow、Pytorch等;
模型性能要求信息,用于指示终端对模型的准确度的最低和/或最高要求;其中模型的准确度可以用模型预测结果的平均绝对误差(Mean Absolute Error,MAE)、最小均方误差(Minimum Mean Squared Error,MMSE)或其他形式表示。
模型使用范围限定信息,用于指示模型的有效区域、适用DNN、适用切片、有效时间等范围信息。
步骤202、所述终端接收所述AF发送的第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。
其中,所述用户面功能支持利用用户面会话传输所述ML模型的功能。
本申请实施例中,所述用户面功能的信息包括以下至少一项:
所述用户面功能的全量域名(Fully Qualified Domain Name,FDQN);
所述用户面功能的地址信息,例如IP地址、媒体接入控制(Medium Access Control,MAC)地址等;
用于指示所述用户面功能存储模型文件位置的统一资源定位符(Uniform Resource Locator,URL),例如用于指示用户面功能存储模型文件的位置;
用于指示传输所述ML模型的用户面会话对应的DNN;
用于指示传输所述ML模型的用户面会话对应的S-NSSAI;
用于指示传输所述ML模型的用户面会话对应的RAT类型;
用于指示传输所述ML模型的用户面会话对应的接入类型(Access Type);
用户面隧道相关的安全信息,例如包括安全证书等。
本申请实施例中,终端通过向AF发送请求传输ML模型的第一请求消息,并接收AF响应于所述第一请求消息发送的第一响应消息,该第一响应消息中包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息,进而终端能够利用AF获得网络侧对于ML模型的传输地址信息或存储位置信息,从而基于用户面会话从用户面功能动态地下载和/或所述用户面功能上传ML模型,进而也就明确了终端与网络侧之间ML模型的传输方式。
可选地,所述第一响应消息还包括第三指示信息,所述第三指示信息用于指示需要利用用户面会话传输所述ML模型,也即传输所述第一请求消息中所请求的ML模型。其中,所述用户面会话可以是PDU会话。进而,所述终端也就能够基于所述第三指示信息获知需要利用用户面会话传输所述ML模型,从而也就明确了终端与网络侧之间传输ML模型的方式。
可选地,所述方法还包括:
所述终端确定用于传输所述ML模型的所述用户面会话。
需要说明地,终端可以是在向AF发送第一请求消息之前,确定用于传输ML模型的用户面会话,或者,终端可以是在接收到AF发送的第一响应消息之后,确定用户传输ML模型的用户面会话。其中,所述用户面会话可以是PDU会话,该PDU会话可以是选择已有数据业务对应的PDU会话,或者是为传输ML模型而新建的PDU会话。
另外,终端在确定用于传输所述ML模型的用户面会话之前,可以先确定自身支持用户面传输的能力,进而以确保能够基于用户面会话进行ML传输。
本申请实施例中,所述方法还包括:
所述终端利用用户面会话,与所述用户面功能的信息所对应的用户面功能传输所述ML模型。
需要说明地,终端可以是在确定用于传输所述ML模型的用户面会话后,利用所述用户面会话,与所述用户面功能的信息所对应的用户面功能传输所述ML模型。
或者,终端也可以是在接收到AF发送的第一响应消息后,根据所述第一响应消息中包括的传输ML模型对应的用户面功能的信息,也就能够确定所对应的用户面功能,进而终端利用用户面会话,例如利用已有的PDU会话,或者是新建一个PDU会话,与所述用户面功能的信息所对应的用户面功能传输所述ML模型。
这样,终端也就能够基于用户面会话,实现与用户面功能之间传输所述ML模型,从而也就明确了终端与网络侧之间ML模型的传输方式。
可选地,传输所述ML模型,包括如下至少一项:
所述终端向所述用户面功能上传所述ML模型的信息。
所述终端从所述用户面功能下载所述ML模型的信息。
例如,在一些场景下,所述ML模型的部分功能是在终端侧进行训练的,则所述终端能够将这部分训练的ML模型上传至用户面功能,进而网络侧能够基于终端上传的训练的ML模型部分与自身侧的模型部分结合或继续训练,以得到完整的ML模型,则终端进一步可以从网络侧(用户面功能)下载完整的ML模型的信息。或者,在另一些场景下,所述ML模型是在网络侧完成训练的,这种情况下终端从用户面功能下载所述ML模型的信息。
可选地,所述ML模型的信息包括以下至少一项:
所述ML模型的模型结构信息;
所述ML模型的模型参数信息;
所述ML模型的模型标识信息;
所述ML模型的模型功能类型信息;
所述ML模型的要求信息。
也就是说,所述终端能够基于用户面会话,与所述用户面功能之间传输上述ML模型的信息,例如终端向用户面功能上传所述ML模型的模型结构信息、模型参数信息等,或者也可以是终端从用户面功能下载所述ML模型的模型结构信息、模型标识信息、模型功能类型信息等,此处不做具体赘述。
需要说明地,所述ML模型的要求信息的具体内容可参照前述描述,此处不再赘述。
本申请实施例中,所述方法还包括:
所述终端在所述用户面会话的基础上,建立与所述用户面功能之间的安全隧道。
示例性地,所述安全隧道可以是加密的传输通道,进而终端能够基于所述加密的传输通道与用户面功能之间传输所述ML模型,进而以保证所述ML模型传输的安全性。
可选地,所述终端利用用户面会话,与所述用户面功能的信息所对应的用户面功能传输所述ML模型,包括:
所述终端利用所述安全隧道与所述用户面功能传输所述ML模型。
可以理解地,在终端基于用户面会话,建立与所述用户面功能之间的安全隧道的情况下,则所述终端利用所述安全隧道与所述用户面功能之间传输所述ML模型,进而也就能够有效确保所述ML模型传输的安全性。
本申请实施例中,所述终端利用用户面会话或者隧道与所述用户面功能传输所述ML模型,包括:
所述终端根据所述用户面功能的信息,确定传输所述ML模型对应的对端目的地址;
所述终端利用所述用户面会话或者隧道与所述对端目的地址之间进行模型传输。
进一步可选地,所述终端根据所述用户面功能的信息,确定传输所述ML模型对应的对端目的地址,包括以下至少一项:
所述终端根据所述用户面功能的全量域名(Fully qualified domain name,FQDN),确定所述用户面功能的地址,所述终端将所述用户面功能的地址作为所述对端目的地址;
所述终端将所述用户面功能的地址作为所述对端目的地址;
所述终端根据所述用户面功能存储所述ML模型的统一资源定位符URL,确定所述对端目的地址。
也就是说,终端可以是直接将用户面功能的地址作为传输所述ML模型的对端目的地址,或者,终端也可以是基于用户面功能的FQDN确定用户面功能的地址后,将所述用户面功能的地址作为对端目的地址;或者,终端根据用户面功能存储所述ML模型的URL来确定对端目的地址。这样,也就使得终端能够通过不同方式、灵活地来确定传输ML模型的对端目的地址。
请参照图3,图3是本申请实施例提供的一种传输方法的流程图之二,该方法应用于AF。如图3所示,所述方法包括以下步骤:
步骤301、AF接收终端发送的第一请求消息,所述第一请求消息用于请求传输ML模型。
可选地,所述第一请求消息包括以下至少一项:
所述终端的ML模型相关能力信息;
所述ML模型的模型标识信息;
所述ML模型的模型功能类型信息;
所述ML模型的要求信息;
第一指示信息,所述第一指示信息用于指示所述ML模型需要利用用户面会话进行传输;
所述终端的IP地址信息。
其中,所述终端的IP地址信息包括用于传输所述ML模型的用户面会话对应的所述终端的IP地址信息。
可选地,所述终端的ML模型相关能力信息包括以下至少一项:
所述终端支持的模型标识信息;
所述终端支持的模型功能类型信息;
用于指示所述终端支持模型下载和/或上传的第二指示信息;
所述终端通过用户面传输模型的能力信息;
所述终端的模型存储空间信息。
可选地,所述ML模型的要求信息包括以下至少一项:
模型的传输时延要求信息;
模型的大小要求信息;
用于指示模型需支持共享的可共享指示信息;
用于限定模型提供者身份的身份限定信息;
模型表述方式限定信息;
模型性能要求信息;
模型使用范围限定信息。
需要说明地,本申请实施例中涉及的相关概念可以是参照上述实施例中的描述,为避免重复,此处不再赘述。
步骤302、所述AF向所述终端发送第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。
可选地,所述AF在接收到终端发送的第一请求消息后,可以是直接向终端发送所述第一响应消息,其中包括第一请求消息所请求传输的ML模型对应的用户面功能的信息。
或者,AF在接收到所述终端发送的第一请求信息后,所述AF向网络侧发送第二请求信息,该第二请求信息用于请求获取所述用户面功能的信息,所述用户面功能用于与所述终端之间利用用户面会话传输ML模型。
其中,AF向网络侧发送第二请求信息,可以包括以下至少一种方式:
所述AF向网络开放功能(Network Exposure Function,NEF)发送所述第二请求消息,并从NEF获取传输ML模型对应的用户面功能的信息;
所述AF向网络中的模型应用平台发送所述第二请求消息,并从所述模型应用平台获取传输ML模型对应的用户面功能的信息。
进一步地,AF在从NEF或模型应用平台获取到所述用户面功能的信息后,所述AF向所述终端发送第一响应消息,所述第一响应消息中包括传输ML模型对应的用户面功能的信息。
需要说明地,模型应用平台可以包括模型数据库网元(如分析数据存储功能网元(Analytics Data Repository Function,ADRF))、模型训练逻辑功能网元(Model Training Logical Function,MTLF)、网络数据分析功能网元(Network Data Analytics Function,NWDAF)等形式。
此处,用户面功能是指存储模型或者可以提供模型实例的网元或功能模块,在实际部署中,该用户面功能可以是独立部署的一个数据库网元(如ADRF,或者模型应用平台,模型商店model store等),MTLF或者其他设备产生的模型可以存储在该数据库网元中。或者,另一种部署方式中,该用户面功能可以是与模型训练功能(MTLF)合设或者外挂的一个功能模块。本申请实施例中,所述用户面功能的信息包括以下至少一项:
所述用户面功能的FDQN;
所述用户面功能的地址信息;
用于指示所述用户面功能存储模型文件位置的URL;
用于指示传输所述ML模型的用户面会话对应的DNN;
用于指示传输所述ML模型的用户面会话对应的S-NSSAI;
用于指示传输所述ML模型的用户面会话对应的RAT类型;
用于指示传输所述ML模型的用户面会话对应的接入类型;
用户面隧道相关的安全信息。
本申请实施例中,AF接收终端发送的用于请求传输ML模型的第一请求消息后,向所述终端发送第一响应消息,该第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息,进而终端能够利用AF获得网络侧对于ML模型的传输地址信息或存储位置信息,从而基于用户面会话从用户面功能动态地下载和/或向所述用户面功能上传ML模型,进而也就明确了终端与网络侧之间ML模型的传输方式。
可选地,所述方法还包括:
所述AF向网络侧发送第二请求消息,所述第二请求消息用于请求获取所述用户面功能的信息,所述用户面功能用于与所述终端之间利用用户面会话传输所述ML模型。
例如,所述AF在接收到终端发送的第一请求消息后,则向网络侧发送第二请求消息,以请求获取用于与所述终端之间利用用户面会话传输所述ML模型的用户面功能的信息,进而以从网络侧获取所述用户面功能的信息。
可选地,所述AF向网络侧发送第二请求消息,包括以下任意一项:
所述AF向网络开放功能(Network Exposure Function,NEF)发送所述第二请求消息;
所述AF向模型应用平台发送所述第二请求消息。
也就是说,所述AF可以是向NEF或模型应用平台请求获取所述用户面功能的信息。
进一步地,所述方法还包括以下任意一项:
在所述AF向所述NEF发送所述第二请求消息的情况下,所述AF接收所述NEF发送的所述用户面功能的信息;
在所述AF向所述模型应用平台发送所述第二请求消息的情况下,所述AF接收所述预设模型应用平台发送的所述用户面功能的信息。
需要说明地,所述AF在从NEF或模型应用平台获取到所述用户面功能的信息后,则所述AF可以是向所述终端发送第一响应消息,所述第一响应消息中包括所述用户面功能的信息。
可选地,所述方法还包括:所述AF接收所述NEF或所述模型应用平台发送的第四指示信息,所述第四指示信息用于指示需要利用用户面会话传输所述ML模型。
例如,所述AF向所述NEF发送所述第二请求消息,以请求获取用于与所述终端之间利用用户面会话传输所述ML模型的用户面功能的信息,所述NEF向所述AF发送所述用户面功能的信息,并还可以向所述AF发送第四指示信息,以指示需要利用用户面会话传输所述ML模型,进而也就通过网络侧来指示所述终端与网络侧之间需要通过用户面会话来实现ML模型的传输,从而以明确了终端与网络侧之间关于ML模型的传输方式。
可以理解地,若所述第二请求消息为所述AF向模型应用平台发送,则模型应用平台在向所述AF发送所述用户面功能的信息的情况下,还可以向所述AF发送所述第四指示信息,进而以明确终端与网络侧之间需要通过用户面会话来实现ML模型的传输。
可选地,所述模型应用平台包括以下至少一项:模型数据库网元(如如分析数据存储功能网元(Analytics Data Repository Function,ADRF))、模型训练逻辑功能网元(Model  Training Logical Function,MTLF)、网络数据分析功能网元(Network Data Analytics Function,NWDAF)。
可选地,所述第二请求消息包括以下至少一项:
所述终端的ML模型相关能力信息;
所述ML模型的模型标识信息;
所述ML模型的模型功能类型信息;
所述ML模型的要求信息;
第五指示信息,所述第五指示信息用于指示所述ML模型需要利用用户面会话进行传输;
所述终端的IP地址信息。
需要说明地,所述第二请求消息所包括的消息内容可以是与上述第一请求消息所包括的消息内容相同。例如,所述AF在接收到所述终端发送的第一请求消息后,根据所述第一请求消息所包括的消息内容生成第二请求消息,该第二请求消息同样包括所述第一请求消息的消息内容,进而以使得AF向网络侧请求获取所述用户面功能的信息。
或者,在一些场景下,所述AF在接收到所述终端发送的第一请求消息后,也可以是将所述第一请求消息转发给网络侧,以向网络侧请求获取所述用户面功能的信息。
本申请实施例中,传输所述ML模型,包括如下至少一项:
所述终端向所述用户面功能上传所述ML模型的信息。
所述终端从所述用户面功能下载所述ML模型的信息。
例如,在一些场景下,所述ML模型的部分功能是在终端侧进行训练的,则所述终端能够将这部分训练的ML模型上传至用户面功能,进而网络侧能够基于终端上传的训练的ML模型部分与自身侧的模型部分结合或继续训练,以得到完整的ML模型,则终端进一步可以从网络侧(用户面功能)下载完整的ML模型的信息。或者,在另一些场景下,所述ML模型是在网络侧完成训练的,这种情况下终端从用户面功能下载所述ML模型的信息。
其中,所述ML模型的信息包括以下至少一项:
所述ML模型的模型结构信息;
所述ML模型的模型参数信息;
所述ML模型的模型标识信息;
所述ML模型的模型功能类型信息;
所述ML模型的要求信息。
例如,终端向用户面功能上传所述ML模型的模型结构信息、模型参数信息等,或者也可以是终端从用户面功能下载所述ML模型的模型结构信息、模型标识信息、模型功能类型信息等,此处不做具体赘述。
需要说明地,本申请实施例中涉及的相关概念可以是参照上述应用于终端的传输方法 实施例中的描述,为避免重复,此处不再赘述。
请参照图4,图4是本申请实施例提供的一种传输方法的流程图之三,该方法应用于模型应用平台。如图4所示,所述方法包括以下步骤:
步骤401、模型应用平台接收AF发送的第二请求消息,所述第二请求消息用于请求获取用户面功能的信息,所述用户面功能用于与所述终端之间利用用户面会话传输ML模型。
其中,模型应用平台接收AF发送的第二请求消息包括以下任意一种方式:
模型应用平台直接接收AF发送的第二请求消息;
模型应用平台接收NEF转发的来自AF的第二请求消息。
可选地,所述第二请求消息包括以下至少一项:
所述终端的ML模型相关能力信息;
所述ML模型的模型标识信息;
所述ML模型的模型功能类型信息;
所述ML模型的要求信息;
第五指示信息,所述第五指示信息用于指示所述ML模型需要利用用户面会话进行传输;
所述终端的IP地址信息。
需要说明地,所述第二请求消息所包括的消息内容可以是与上述实施例中第一请求消息所包括的消息内容相同。
步骤402、所述模型应用平台向所述AF发送所述用户面功能的信息。
其中,模型应用平台向所述AF发送所述用户面功能的信息,包括以下任意一种方式:
模型应用平台直接向所述AF发送所述用户面功能的信息;
模型应用平台经过NEF向所述AF发送所述用户面功能的信息。
可选地,所述用户面功能的信息包括以下至少一项:
所述用户面功能的FDQN;
所述用户面功能的地址信息;
用于指示所述用户面功能存储模型文件位置的URL;
用于指示传输所述ML模型的用户面会话对应的DNN;
用于指示传输所述ML模型的用户面会话对应的S-NSSAI;
用于指示传输所述ML模型的用户面会话对应的RAT类型;
用于指示传输所述ML模型的用户面会话对应的接入类型;
用户面隧道相关的安全信息。
本申请实施例中,模型应用平台在接收到AF发送的用于请求获取用户面功能的信息的第二请求消息的情况下,所述模型应用平台响应该第二请求消息,向所述AF发送所述用户面功能的信息,所述用户面功能用于与所述终端之间利用用户面会话传输ML模型, 进而终端能够利用AF获得网络侧对于ML模型的传输地址信息或存储位置信息,从而基于用户面会话从用户面功能动态地下载和/或向所述用户面功能上传ML模型,进而也就明确了终端与网络侧之间ML模型的传输方式。
可选地,所述方法还包括:
所述模型应用平台向所述AF发送第四指示信息,所述第四指示信息用于指示需要利用用户面会话传输所述ML模型。
例如,所述模型应用平台可以是在向所述AF发送所述用户面功能的信息之前,向所述AF发送第四指示信息,也就使得所述AF能够获知需要利用用户面会话来传输ML模型,进而也就通过网络侧来指示所述终端与网络侧之间需要通过用户面会话来实现ML模型的传输,从而以明确了终端与网络侧之间关于ML模型的传输方式。
可选地,所述模型应用平台也可以是在向所述AF发送所述用户面功能的信息之后,向所述AF发送第四指示信息。
本申请实施例中,所述方法还包括:
所述模型应用平台确定所述用户面功能的信息。
可以理解地,所述模型应用平台可以是直接确定所述用户面功能的信息,例如所述用户面功能的FDQN、所述用户面功能的地址、所述用户面功能存储ML模型的URL位置信息等,然后向所述AF发送所述用户面功能的信息。这样,也就能够通过网络侧来确定所述用户面功能的信息,进而以明确终端与用户面功能之间实现ML模型的传输。
可选地,所述模型应用平台确定所述用户面功能的信息,包括:
所述模型应用平台在确定使用用户面会话传输所述ML模型的情况下,确定所述用户面功能的信息。
示例性地,所述模型应用平台在接收到AF发送的所述第二请求消息的情况下,若所述模型应用平台确定使用用户面会话(例如PDU会话)来传输ML模型,则所述模型应用平台确定所述用户面功能的信息,并向所述AF发送所述用户面功能的信息。这样,也就通过网络侧来确定所述终端与网络侧之间需要通过用户面会话来实现ML模型的传输,从而以明确了终端与网络侧之间关于ML模型的传输方式。
可选地,所述方法还可以包括:
所述模型应用平台根据所述第二请求消息确定所述用户面功能的信息。
示例性地,所述模型应用平台可以是在接收到AF发送的用于请求获取用户面功能的信息的第二请求消息的情况下,确定所述用户面功能的信息,例如所述用户面功能FQDN、所述用户面功能的地址信息、用于指示所述用户面功能存储模型文件位置的URL等,然后向所述AF发送所述用户面功能的信息。这样,模型应用平台也就能够基于所述第二请求消息来确定所述用户面功能的信息,进而以明确终端与用户面功能之间实现ML模型的传输方式。
可选地,所述模型应用平台根据所述第二请求消息确定所述用户面功能的信息,包括:
所述模型应用平台向网络存储功能发送用于确定所述ML模型对应的所述用户面功能的发现消息,所述发现消息包括如下至少一项:所述ML模型的标识信息、所述ML模型的功能信息、所述ML模型的类型信息、用户面传输模型的指示信息;
所述模型应用平台接收所述网络存储功能发送的所述用户面功能的信息。
示例性地,所述模型应用平台可以是在接收到AF发送的用于请求获取用户面功能的信息的第二请求消息的情况下,向网络存储功能发送所述发现消息,例如所述发现消息中包括所述ML模型的标识信息、所述ML模型的功能信息,进而以使得所述网络存储功能基于所述发现消息也就能够确定需要传输的ML模型是哪一个、是什么功能类型的模型,并基于所述发现消息来确定用户面功能的信息,例如所述用户面功能的信息包括基于所述ML模型的标识信息、所述ML模型的功能信息确定的用于传输所述ML模型的用户面会话对应的DNN、S-NSSAI等,并将所述用户面功能的信息发送给所述模型应用平台,进一步模型应用平台能够将所述用户面功能的信息发送给AF,进而以使得AF能够明确用于传输所述ML模型的用户面会话对应的DNN、S-NSSAI等,以明确终端与网络侧之间传输所述ML模型的用户面会话,从而确保所述终端与网络侧之间ML模型传输的顺畅。
可选地,所述模型应用平台根据所述第二请求消息确定所述用户面功能的信息,包括:
所述模型应用平台向模型存储网元发送第三请求消息,所述第三请求消息用于请求所述模型存储网元反馈所述ML模型对应的用户面功能的信息;
所述模型应用平台接收所述模型存储网元发送的所述ML模型对应的用户面功能的信息,所述用户面功能的信息包括所述模型存储网元存储所述ML模型的URL。
示例性地,所述模型应用平台可以是在接收到AF发送的用于请求获取用户面功能的信息的第二请求消息的情况下,所述模型应用平台向模型存储网元发送用于请求所述模型存储网元反馈所述ML模型对应的用户面功能的信息的第三请求消息,其中,模型应用平台可以根据预设的模型存储网元的信息确定模型存储网元,或者,模型应用平台也可以根据历史ML模型存储记录确定模型存储网元。若所述模型存储网元为所述用户面功能,则所述模型存储网元向所述模型应用平台发送其存储的所述ML模型的URL,进一步地所述模型应用平台将所述ML模型的URL发送给AF,进而以使得AF能够明确所述模型存储网元存储所述ML模型的位置,以确保所述终端和网络侧之间ML模型传输的顺畅。
请参照图5,图5是本申请实施例提供的一种传输方法的流程图之四,如图5所示,所述方法包括以下步骤:
步骤0a.UE选择已有PDU会话;
步骤0b.UE新建PDU会话;其中,步骤0a和步骤0b可以是二选一进行;
步骤0y.AF向模型应用平台(model store,如ADRF)发送模型存储请求,该请求中包括模型信息;模型应用平台向AF反馈模型存储地址;
上述步骤0a~0y为可选步骤;
步骤1.UE向AF发送模型请求(也即上述第一请求消息),该请求中包括模型ID、模 型功能ID、UE的模型相关能力信息等;
可选地,步骤2a.AF向模型应用平台发送模型获取请求,该请求中包括终端IP地址;
可选地,步骤2b.模型应用平台向AF反馈模型获取响应,该响应中包括模型对应的用户面功能的信息;
步骤3.AF向终端发送模型响应消息(也即上述第一响应消息),该消息中包括模型对应的用户面功能的信息;
可选地,步骤4a.UE选择已有PDU会话;或者,步骤4b.UE新建PDU会话;其中,步骤4a或4b与上述步骤0a或0b可以是选其一进行;
可选地,步骤5.建立与所述用户面功能之间的安全通道;
步骤6.基于终端IP地址,用户面功能向UE交互模型相关信息,包括模型ID、模型信息等,如模型结构信息和/或模型参数信息;
步骤7.基于用户面功能的信息,UE通过PDU会话向用户面功能交互模型相关信息,包括模型ID、模型信息等,如模型结构信息和/或模型参数信息。
需要说明地,本申请实施例中的传输方法具体实现流程及相关概念可具体参照上述实施例中的描述,此处不再赘述。
本申请实施例提供的传输方法,执行主体可以为传输装置。本申请实施例中以传输装置执行传输方法为例,说明本申请实施例提供的传输装置。
请参照图6,图6是本申请实施例提供的一种传输装置的结构图之一,如图6所示,所述传输装置600包括:
第一发送模块601,用于向应用功能AF发送第一请求消息,所述第一请求消息用于请求传输机器学习ML模型;
第一接收模块602,用于接收所述AF发送的第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。
可选地,所述装置还包括:
第一传输模块,用于利用用户面会话,与所述用户面功能的信息所对应的用户面功能传输所述ML模型。
可选地,传输所述ML模型,包括如下至少一项:
向所述用户面功能上传所述ML模型的信息。
从所述用户面功能下载所述ML模型的信息。
可选地,所述ML模型的信息包括以下至少一项:
所述ML模型的模型结构信息;
所述ML模型的模型参数信息;
所述ML模型的模型标识信息;
所述ML模型的模型功能类型信息;
所述ML模型的要求信息。
可选地,所述装置还包括:
第一确定模块,用于确定用于传输所述ML模型的所述用户面会话。
可选地,所述第一请求消息包括以下至少一项:
所述装置的ML模型相关能力信息;
所述ML模型的模型标识信息;
所述ML模型的模型功能类型信息;
所述ML模型的要求信息;
第一指示信息,所述第一指示信息用于指示所述ML模型需要利用用户面会话进行传输;
所述装置的IP地址信息。
可选地,所述装置的IP地址信息包括用于传输所述ML模型的用户面会话对应的所述装置的IP地址信息。
可选地,所述装置的ML模型相关能力信息包括以下至少一项:
所述装置支持的模型标识信息;
所述装置支持的模型功能类型信息;
用于指示所述装置支持模型下载和/或上传的第二指示信息;
所述装置通过用户面传输模型的能力信息;
所述装置的模型存储空间信息。
可选地,所述ML模型的要求信息包括以下至少一项:
模型的传输时延要求信息;
模型的大小要求信息;
用于指示模型需支持共享的可共享指示信息;
用于限定模型提供者身份的身份限定信息;
模型表述方式限定信息;
模型性能要求信息;
模型使用范围限定信息。
可选地,所述装置还包括:
建立模块,用于在所述用户面会话的基础上,建立与所述用户面功能之间的安全隧道。
可选地,所述第一传输模块还用于:
利用所述安全隧道与所述用户面功能传输所述ML模型。
可选地,所述第一响应消息还包括第三指示信息,所述第三指示信息用于指示需要利用用户面会话传输所述ML模型。
可选地,所述用户面功能的信息包括以下至少一项:
所述用户面功能的全量域名FDQN;
所述用户面功能的地址信息;
用于指示所述用户面功能存储模型文件位置的统一资源定位符URL;
用于指示传输所述ML模型的用户面会话对应的数据网络名DNN;
用于指示传输所述ML模型的用户面会话对应的单一网络切片选择辅助信息S-NSSAI;
用于指示传输所述ML模型的用户面会话对应的无线接入技术RAT类型;
用于指示传输所述ML模型的用户面会话对应的接入类型;
用户面隧道相关的安全信息。
本申请实施例中,所述装置通过向AF发送请求传输ML模型的第一请求消息,并接收AF响应于所述第一请求消息发送的第一响应消息,该第一响应消息中包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息,进而所述装置能够利用AF获得网络侧对于ML模型的传输地址信息或存储位置信息,从而基于用户面会话从用户面功能动态地下载和/或所述用户面功能上传ML模型,进而也就明确了与网络侧之间ML模型的传输方式。
本申请实施例中的传输装置600可以是电子设备,例如具有操作***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传输装置600能够实现图2所述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。
请参照图7,图7是本申请实施例提供的一种传输装置的结构图之二,如图7所示,所述传输装置700包括:
第二接收模块701,用于接收终端发送的第一请求消息,所述第一请求消息用于请求传输ML模型;
第二发送模块702,用于向所述终端发送第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。
可选地,所述第二发送模块702还用于:
向网络侧发送第二请求消息,所述第二请求消息用于请求获取所述用户面功能的信息,所述用户面功能用于与所述终端之间利用用户面会话传输所述ML模型。
可选地,传输所述ML模型,包括如下至少一项:
向所述用户面功能上传所述ML模型的信息。
从所述用户面功能下载所述ML模型的信息。
可选地,所述第二发送模块702还用于执行以下任意一项:
向网络开放功能NEF发送所述第二请求消息;
向模型应用平台发送所述第二请求消息。
可选地,所述第二接收模块701还用于执行以下任意一项:
在所述装置向所述NEF发送所述第二请求消息的情况下,接收所述NEF发送的所述用户面功能的信息;
在所述装置向所述模型应用平台发送所述第二请求消息的情况下,接收所述预设模型应用平台发送的所述用户面功能的信息。
可选地,所述第二接收模块701还用于:
接收所述NEF或所述模型应用平台发送的第四指示信息,所述第四指示信息用于指示需要利用用户面会话传输所述ML模型。
可选地,所述模型应用平台包括以下至少一项:模型数据库网元、模型训练逻辑功能网元和网络数据分析功能网元。
可选地,所述第二请求消息包括以下至少一项:
所述终端的ML模型相关能力信息;
所述ML模型的模型标识信息;
所述ML模型的模型功能类型信息;
所述ML模型的要求信息;
第五指示信息,所述第五指示信息用于指示所述ML模型需要利用用户面会话进行传输;
所述终端的IP地址信息。
可选地,所述用户面功能的信息包括以下至少一项:
所述用户面功能的FDQN;
所述用户面功能的地址信息;
用于指示所述用户面功能存储模型文件位置的URL;
用于指示传输所述ML模型的用户面会话对应的DNN;
用于指示传输所述ML模型的用户面会话对应的S-NSSAI;
用于指示传输所述ML模型的用户面会话对应的RAT类型;
用于指示传输所述ML模型的用户面会话对应的接入类型;
用户面隧道相关的安全信息。
本申请实施例中,所述装置接收终端发送的用于请求传输ML模型的第一请求消息后,向所述终端发送第一响应消息,该第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息,进而终端能够获得网络侧对于ML模型的传输地址信息或存储位置信息,从而基于用户面会话从用户面功能动态地下载和/或向所述用户面功能上传ML模型,进而也就明确了终端与网络侧之间ML模型的传输方式。
本申请实施例提供的传输装置700能够实现图3所述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。
请参照图8,图8是本申请实施例提供的一种传输装置的结构图之三,如图8所示,所述传输装置800包括:
第三接收模块801,用于接收AF发送的第二请求消息,所述第二请求消息用于请求获取用户面功能的信息,所述用户面功能用于与终端之间利用用户面会话传输ML模型;
第三发送模块802,用于向所述AF发送所述用户面功能的信息。
可选地,所述第二请求消息包括以下至少一项:
所述终端的ML模型相关能力信息;
所述ML模型的模型标识信息;
所述ML模型的模型功能类型信息;
所述ML模型的要求信息;
第五指示信息,所述第五指示信息用于指示所述ML模型需要利用用户面会话进行传输;
所述终端的IP地址信息。
可选地,所述第三发送模块802还用于:
向所述AF发送第四指示信息,所述第四指示信息用于指示需要利用用户面会话传输所述ML模型。
可选地,所述装置还包括:
第二确定模块,用于确定所述用户面功能的信息。
可选地,所述第二确定模块还用于:
在确定使用用户面会话传输所述ML模型的情况下,确定所述用户面功能的信息。
可选地,所述装置还包括:
第三确定模块,用于根据所述第二请求消息确定所述用户面功能的信息。
可选地,所述第三确定模块还用于:
向网络存储功能发送用于确定所述ML模型对应的所述用户面功能的发现消息,所述发现消息包括如下至少一项:所述ML模型的标识信息、所述ML模型的功能信息、所述ML模型的类型信息、用户面传输模型的指示信息;
接收所述网络存储功能发送的所述用户面功能的信息。
可选地,所述第三确定模块还用于:
向模型存储网元发送第三请求消息,所述第三请求消息用于请求所述模型存储网元反馈所述ML模型对应的用户面功能的信息;
接收所述模型存储网元发送的所述ML模型对应的用户面功能的信息,所述用户面功能的信息包括所述模型存储网元存储所述ML模型的URL。
本申请实施例中,所述装置在接收到AF发送的用于请求获取用户面功能的信息的第二请求消息的情况下,响应该第二请求消息,向所述AF发送所述用户面功能的信息,所述用户面功能用于与所述终端之间利用用户面会话传输ML模型,进而终端能够利用AF获得网络侧对于ML模型的传输地址信息或存储位置信息,从而基于用户面会话从用户面功能动态地下载和/或向所述用户面功能上传ML模型,进而也就明确了终端与网络侧之 间ML模型的传输方式
本申请实施例提供的传输装置700能够实现图3所述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。
可选的,如图9所示,本申请实施例还提供一种通信设备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执行时实现上述图2所述传输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执行时实现上述图3或图4所述传输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,所述通信接口用于向应用功能AF发送第一请求消息,所述第一请求消息用于请求传输机器学习ML模型;以及用于接收所述AF发送的第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。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。具体地,图10为实现本申请实施例的一种终端的硬件结构示意图。
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***与处理器1010逻辑相连,从而通过电源管理***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以及其他输入设备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。其他输入设备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接收来自网络侧设备的下行数据后,可以传输给处理器1010进行处理;另外,射频单元1001可以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、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***、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可以包括易失性和非易失性存储器两者。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***、用户界面和应用程序等的操作,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,用于向应用功能AF发送第一请求消息,所述第一请求消息用于请求传输机器学习ML模型;以及,
接收所述AF发送的第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息
本申请实施例中,终端通过向AF发送请求传输ML模型的第一请求消息,并接收AF响应于所述第一请求消息发送的第一响应消息,该第一响应消息中包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息,进而终端能够利用AF获得网络侧对于ML模型的传输地址信息或存储位置信息,从而基于用户面会话从用户面功能动态地下载和/或所述用户面功能上传ML模型,进而也就明确了终端与网络侧之间ML模型的传输方式。
具体地,本申请实施例还提供了一种网络侧设备。如图11所示,该网络侧设备1100包括:处理器1101、网络接口1102和存储器1103。其中,网络接口1102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1103上并可在处理器1101上运行的指令或程序,处理器1101调用存储器1103中的指令或程序执行图7或图8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一种可读存储介质,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图2或图3或图4所述传输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,实现上述图2或图3或图4所述传输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***级芯片,***芯片,芯片***或片上***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图2或图3或图4所述传输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。
本申请实施例还提供了一种通信***,包括:终端及网络侧设备,所述终端可用于执行如上图2所述的传输方法的步骤,所述网络侧设备可用于执行如上图3或图4所述的传输方法的步骤。
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现有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 (36)

  1. 一种传输方法,包括:
    终端向应用功能AF发送第一请求消息,所述第一请求消息用于请求传输机器学习ML模型;
    所述终端接收所述AF发送的第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。
  2. 根据权利要求1所述的方法,所述方法还包括:
    所述终端利用用户面会话,与所述用户面功能的信息所对应的用户面功能传输所述ML模型。
  3. 根据权利要求1或2所述的方法,其中,传输所述ML模型,包括如下至少一项:
    所述终端向所述用户面功能上传所述ML模型的信息;
    所述终端从所述用户面功能下载所述ML模型的信息。
  4. 根据权利要求3所述的方法,其中,所述ML模型的信息包括以下至少一项:
    所述ML模型的模型结构信息;
    所述ML模型的模型参数信息;
    所述ML模型的模型标识信息;
    所述ML模型的模型功能类型信息;
    所述ML模型的要求信息。
  5. 根据权利要求2所述的方法,所述方法还包括:
    所述终端确定用于传输所述ML模型的所述用户面会话。
  6. 根据权利要求1所述的方法,其中,所述第一请求消息包括以下至少一项:
    所述终端的ML模型相关能力信息;
    所述ML模型的模型标识信息;
    所述ML模型的模型功能类型信息;
    所述ML模型的要求信息;
    第一指示信息,所述第一指示信息用于指示所述ML模型需要利用用户面会话进行传输;
    所述终端的因特网协议IP地址信息。
  7. 根据权利要求6所述的方法,其中,所述终端的IP地址信息包括用于传输所述ML模型的用户面会话对应的所述终端的IP地址信息。
  8. 根据权利要求6所述的方法,其中,所述终端的ML模型相关能力信息包括以下至少一项:
    所述终端支持的模型标识信息;
    所述终端支持的模型功能类型信息;
    用于指示所述终端支持模型下载和/或上传的第二指示信息;
    所述终端通过用户面传输模型的能力信息;
    所述终端的模型存储空间信息。
  9. 根据权利要求6所述的方法,其中,所述ML模型的要求信息包括以下至少一项:
    模型的传输时延要求信息;
    模型的大小要求信息;
    用于指示模型需支持共享的可共享指示信息;
    用于限定模型提供者身份的身份限定信息;
    模型表述方式限定信息;
    模型性能要求信息;
    模型使用范围限定信息。
  10. 根据权利要求2中任一项所述的方法,所述方法还包括:
    所述终端在所述用户面会话的基础上,建立与所述用户面功能之间的安全隧道。
  11. 根据权利要求10所述的方法,其中,所述终端利用用户面会话,与所述用户面功能的信息所对应的用户面功能传输所述ML模型,包括:
    所述终端利用所述安全隧道与所述用户面功能传输所述ML模型。
  12. 根据权利要求1所述的方法,其中,所述第一响应消息还包括第三指示信息,所述第三指示信息用于指示需要利用用户面会话传输所述ML模型。
  13. 根据权利要求1或2或5-12中任意一项所述的方法,其中,所述用户面功能的信息包括以下至少一项:
    所述用户面功能的全量域名FDQN;
    所述用户面功能的地址信息;
    用于指示所述用户面功能存储模型文件位置的统一资源定位符URL;
    用于指示传输所述ML模型的用户面会话对应的数据网络名DNN;
    用于指示传输所述ML模型的用户面会话对应的单一网络切片选择辅助信息S-NSSAI;
    用于指示传输所述ML模型的用户面会话对应的无线接入技术RAT类型;
    用于指示传输所述ML模型的用户面会话对应的接入类型;
    用户面隧道相关的安全信息。
  14. 一种传输方法,包括:
    AF接收终端发送的第一请求消息,所述第一请求消息用于请求传输ML模型;
    所述AF向所述终端发送第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。
  15. 根据权利要求14所述的方法,所述方法还包括:
    所述AF向网络侧发送第二请求消息,所述第二请求消息用于请求获取所述用户面功能的信息,所述用户面功能用于与所述终端之间利用用户面会话传输所述ML模型。
  16. 根据权利要求14或15所述的方法,其中,传输所述ML模型,包括如下至少一项:
    所述终端向所述用户面功能上传所述ML模型的信息;
    所述终端从所述用户面功能下载所述ML模型的信息。
  17. 根据权利要求15所述的方法,其中,所述AF向网络侧发送第二请求消息包括以下任意一项:
    所述AF向网络开放功能NEF发送所述第二请求消息;
    所述AF向模型应用平台发送所述第二请求消息。
  18. 根据权利要求17所述的方法,其中,所述方法还包括以下任意一项:
    在所述AF向所述NEF发送所述第二请求消息的情况下,所述AF接收所述NEF发送的所述用户面功能的信息;
    在所述AF向所述模型应用平台发送所述第二请求消息的情况下,所述AF接收所述模型应用平台发送的所述用户面功能的信息。
  19. 根据权利要求17所述的方法,所述方法还包括:
    所述AF接收所述NEF或所述模型应用平台发送的第四指示信息,所述第四指示信息用于指示需要利用用户面会话传输所述ML模型。
  20. 根据权利要求16所述的方法,其中,所述模型应用平台包括以下至少一项:模型数据库网元、模型训练逻辑功能网元和网络数据分析功能网元。
  21. 根据权利要求15或17或18所述的方法,其中,所述第二请求消息包括以下至少一项:
    所述终端的ML模型相关能力信息;
    所述ML模型的模型标识信息;
    所述ML模型的模型功能类型信息;
    所述ML模型的要求信息;
    第五指示信息,所述第五指示信息用于指示所述ML模型需要利用用户面会话进行传输;
    所述终端的IP地址信息。
  22. 根据权利要求14所述的方法,其中,所述用户面功能的信息包括以下至少一项:
    所述用户面功能的FDQN;
    所述用户面功能的地址信息;
    用于指示所述用户面功能存储模型文件位置的URL;
    用于指示传输所述ML模型的用户面会话对应的DNN;
    用于指示传输所述ML模型的用户面会话对应的S-NSSAI;
    用于指示传输所述ML模型的用户面会话对应的RAT类型;
    用于指示传输所述ML模型的用户面会话对应的接入类型;
    用户面隧道相关的安全信息。
  23. 一种传输方法,包括:
    模型应用平台接收AF发送的第二请求消息,所述第二请求消息用于请求获取用户面功能的信息,所述用户面功能用于与终端之间利用用户面会话传输ML模型;
    所述模型应用平台向所述AF发送所述用户面功能的信息。
  24. 根据权利要求23所述的方法,其中,所述第二请求消息包括以下至少一项:
    所述终端的ML模型相关能力信息;
    所述ML模型的模型标识信息;
    所述ML模型的模型功能类型信息;
    所述ML模型的要求信息;
    第五指示信息,所述第五指示信息用于指示所述ML模型需要利用用户面会话进行传输;
    所述终端的IP地址信息。
  25. 根据权利要求23所述的方法,所述方法还包括:
    所述模型应用平台向所述AF发送第四指示信息,所述第四指示信息用于指示需要利用用户面会话传输所述ML模型。
  26. 根据权利要求23所述的方法,所述方法还包括:
    所述模型应用平台确定所述用户面功能的信息。
  27. 根据权利要求26所述的方法,其中,所述模型应用平台确定所述用户面功能的信息,包括:
    所述模型应用平台在确定使用用户面会话传输所述ML模型的情况下,确定所述用户面功能的信息。
  28. 根据权利要求23所述的方法,所述方法还包括:
    所述模型应用平台根据所述第二请求消息确定所述用户面功能的信息。
  29. 根据权利要求28所述的方法,其中,所述模型应用平台根据所述第二请求消息确定所述用户面功能的信息,包括:
    所述模型应用平台向网络存储功能发送用于确定所述ML模型对应的所述用户面功能的发现消息,所述发现消息包括如下至少一项:所述ML模型的标识信息、所述ML模型的功能信息、所述ML模型的类型信息、用户面传输模型的指示信息;
    所述模型应用平台接收所述网络存储功能发送的所述用户面功能的信息。
  30. 根据权利要求28或29所述的方法,其中,所述模型应用平台根据所述第二请求消息确定所述用户面功能的信息,包括:
    所述模型应用平台向模型存储网元发送第三请求消息,所述第三请求消息用于请求所述模型存储网元反馈所述ML模型对应的用户面功能的信息;
    所述模型应用平台接收所述模型存储网元发送的所述ML模型对应的用户面功能的 信息,所述用户面功能的信息包括所述模型存储网元存储所述ML模型的URL。
  31. 一种传输装置,包括:
    第一发送模块,用于向应用功能AF发送第一请求消息,所述第一请求消息用于请求传输机器学习ML模型;
    第一接收模块,用于接收所述AF发送的第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。
  32. 一种传输装置,包括:
    第二接收模块,用于接收终端发送的第一请求消息,所述第一请求消息用于请求传输ML模型;
    第二发送模块,用于向所述终端发送第一响应消息,所述第一响应消息包括所述第一请求消息所请求传输的ML模型对应的用户面功能的信息。
  33. 一种传输装置,包括:
    第三接收模块,用于接收AF发送的第二请求消息,所述第二请求消息用于请求获取用户面功能的信息,所述用户面功能用于与终端之间利用用户面会话传输ML模型;
    第三发送模块,用于向所述AF发送所述用户面功能的信息。
  34. 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1-13中任一项所述的传输方法的步骤。
  35. 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14-22中任一项所述的传输方法的步骤,或实现如权利要求23-30中任一项所述的传输方法的步骤。
  36. 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1-30中任一项所述的传输方法的步骤。
PCT/CN2023/135262 2022-12-07 2023-11-30 传输方法、装置、终端及网络侧设备 WO2024120286A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211567252.3 2022-12-07
CN202211567252.3A CN118158655A (zh) 2022-12-07 2022-12-07 传输方法、装置、终端及网络侧设备

Publications (1)

Publication Number Publication Date
WO2024120286A1 true WO2024120286A1 (zh) 2024-06-13

Family

ID=91283940

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/135262 WO2024120286A1 (zh) 2022-12-07 2023-11-30 传输方法、装置、终端及网络侧设备

Country Status (2)

Country Link
CN (1) CN118158655A (zh)
WO (1) WO2024120286A1 (zh)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2022115011A1 (en) * 2020-11-26 2022-06-02 Telefonaktiebolaget Lm Ericsson (Publ) Providing distributed ai models in communication networks and related nodes/devices
US20220342713A1 (en) * 2020-01-14 2022-10-27 Guangdong Oppo Mobile Telecommunications Corp., Ltd. Information reporting method, apparatus and device, and storage medium

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220342713A1 (en) * 2020-01-14 2022-10-27 Guangdong Oppo Mobile Telecommunications Corp., Ltd. Information reporting method, apparatus and device, and storage medium
WO2022115011A1 (en) * 2020-11-26 2022-06-02 Telefonaktiebolaget Lm Ericsson (Publ) Providing distributed ai models in communication networks and related nodes/devices

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
JUN CHEN, HUAWEI, HISILICON: "Discussion on AIML methods", 3GPP DRAFT; R2-2212226; TYPE DISCUSSION; FS_NR_AIML_AIR, 3RD GENERATION PARTNERSHIP PROJECT (3GPP), MOBILE COMPETENCE CENTRE ; 650, ROUTE DES LUCIOLES ; F-06921 SOPHIA-ANTIPOLIS CEDEX ; FRANCE, vol. 3GPP RAN 2, no. Toulouse, FR; 20221114 - 20221118, 4 November 2022 (2022-11-04), Mobile Competence Centre ; 650, route des Lucioles ; F-06921 Sophia-Antipolis Cedex ; France, XP052216310 *
SHABNAM SULTANA, ERICSSON: "KI#2, Sol#30: Update: AI/ML Translator (AI/ML-T) for 5GC Information Exposure to UE.", 3GPP DRAFT; S2-2208240; TYPE PCR; FS_AIMLSYS, 3RD GENERATION PARTNERSHIP PROJECT (3GPP), MOBILE COMPETENCE CENTRE ; 650, ROUTE DES LUCIOLES ; F-06921 SOPHIA-ANTIPOLIS CEDEX ; FRANCE, vol. 3GPP SA 2, no. Online; 20221010 - 20221017, 30 September 2022 (2022-09-30), Mobile Competence Centre ; 650, route des Lucioles ; F-06921 Sophia-Antipolis Cedex ; France, XP052208005 *

Also Published As

Publication number Publication date
CN118158655A (zh) 2024-06-07

Similar Documents

Publication Publication Date Title
WO2023093609A1 (zh) 物联网设备的会话建立方法及装置
WO2024120286A1 (zh) 传输方法、装置、终端及网络侧设备
WO2024120285A1 (zh) 信息传输方法、装置、终端及网络侧设备
WO2024120282A1 (zh) 信息发送、信息接收方法、装置、终端及网络侧设备
WO2023143423A1 (zh) 信息获取与存储、上报方法、装置、终端及网络功能
WO2023179595A1 (zh) 非3gpp设备的会话通道建立方法、装置及设备
WO2024140570A1 (zh) 策略配置方法、装置、终端、网络侧设备及可读存储介质
WO2023185810A1 (zh) 终端路由选择策略ursp确定方法、终端及网络侧设备
WO2024051663A1 (zh) 数据通道建立方法、第一终端及第二终端
WO2023179571A1 (zh) 非公共网络接入方法、装置及终端
WO2024045970A1 (zh) 路由选择策略执行结果处理方法、装置及设备
WO2024067439A1 (zh) 数据集生成方法、信息发送方法、装置及相关设备
WO2024140712A1 (zh) 模型提供、模型获取、设备查询方法、装置和通信设备
WO2024067567A1 (zh) 分析结果的获取方法及通信设备
WO2024022370A1 (zh) 信息获取和发送方法、服务器访问和会话建立方法及设备
WO2024022182A1 (zh) 信息查询方法、装置、终端及网络侧设备
WO2023185850A1 (zh) Ursp规则优化方法、装置、终端、网络侧设备及介质
WO2024061145A1 (zh) 网关信息的使用方法、装置、终端及网络侧设备
WO2023179553A1 (zh) 终端不可用时期的协商方法、终端及网络侧设备
WO2024027578A1 (zh) 流量路由方法、装置及设备
WO2023165480A1 (zh) 数据传输方法、装置、终端、设备及存储介质
WO2023143414A1 (zh) 数据传输、配置方法、装置、终端及网络侧设备
WO2024078400A1 (zh) 模型请求方法、装置、通信设备及可读存储介质
WO2024088195A1 (zh) 执行结果上报方法、接收方法、终端及网络侧设备
WO2023217089A1 (zh) 数据传输方法、装置、设备、***及存储介质