Model processing method and device and communication equipment
By confirming the model adaptation between communication devices and using the target model, the problem of model mismatch between different devices is solved, and the task processing effect is improved.
Patent Information
- Application Number
- CN202310506608.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-06
- Publication Date
- 2024-11-08
AI Technical Summary
In mobile communication networks, the model misfit between different communication devices leads to the impact of task processing effects, and the respective selection of models in the prior art leads to poor adaptability.
A communication device receives the model information sent by another communication device through one communication device, and uses the target model after confirming the adaptation, reducing the occurrence of model mismatch.
It improves the task processing effect on different communication equipment sides, ensures model adaptability, and improves communication performance.
Smart Images

Figure CN118921144A_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the field of communication technology, and specifically relates to a model processing method, device and communication equipment. Background Art
[0002] Currently in mobile communication networks, artificial intelligence (AI) can be used to improve network communication performance. For example, network-side devices or terminals can perform AI or machine learning (ML) model training based on training data to obtain a model suitable for a certain AI task, and can infer the input data of the AI task based on the trained AI / ML model to obtain the inference result data corresponding to the AI task.
[0003] In practical applications, different communication devices often need to use associated or matching models for task processing. For example, for the channel state information (CSI) prediction task, the terminal compresses and encodes the CSI based on the coding model to obtain the coding result and sends it to the network side device. The network side device decodes the coding result received from the terminal based on the decoding model matching the above coding model to restore the channel state information. However, in the prior art, different communication devices often directly select their own models for task processing, which easily leads to incompatibility of the models used by different communication devices, affecting the task processing effect. Summary of the invention
[0004] The embodiments of the present application provide a model processing method, apparatus and communication equipment, which can reduce the occurrence of incompatibility of models used by different communication equipment sides, thereby improving the task processing effect.
[0005] In a first aspect, a model processing method is provided, the method comprising:
[0006] The first communication device receives first information sent by the second communication device, where the first information includes model information of the first model;
[0007] The first communication device sends second information to the second communication device, where the second information is used to instruct the second communication device to use a target model, and the first model includes the target model.
[0008] In a second aspect, a model processing device is provided, the device comprising:
[0009] A first receiving module, configured to receive first information sent by a second communication device, wherein the first information includes model information of a first model;
[0010] The first sending module is used to send second information to the second communication device, where the second information is used to instruct the second communication device to use a target model, and the first model includes the target model.
[0011] In a third aspect, a model processing method is provided, the method comprising:
[0012] The second communication device sends first information to the first communication device, where the first information includes model information of the first model;
[0013] The second communication device receives second information or third information sent by the first communication device, the second information is used to instruct the second communication device to use a target model, the first model includes the target model, and the third information is used to indicate that the second communication device is denied to use the first model.
[0014] In a fourth aspect, a model processing device is provided, the device comprising:
[0015] A second sending module, configured to send first information to the first communication device, wherein the first information is used to indicate that the second communication device requests to use the first model;
[0016] The second receiving module is used to receive second information or third information sent by the first communication device, the second information is used to instruct the second communication device to use a target model, the first model includes the target model, and the third information is used to indicate that the second communication device is denied to use the first model.
[0017] In a fifth aspect, a model processing method is provided, the method comprising:
[0018] The third communication device performs a first operation, wherein the first operation is at least one of the following:
[0019] receiving first registration request information sent by a first communication device, and sending first response information to the first communication device, wherein the first registration request information is used to request registration of a second model, and the first response information includes at least one of the following: a model identifier of the second model, and an identifier of a data set associated with the second model;
[0020] Receive second registration request information sent by a fourth communication device, and send second response information to the fourth communication device, wherein the second registration request information is used to request registration of the first model, and the second response information includes at least one of the following: a model identifier of the first model, and a data set identifier associated with the first model.
[0021] In a sixth aspect, a model processing device is provided, the device comprising:
[0022] The first execution module is configured to execute a first operation, wherein the first operation is at least one of the following:
[0023] receiving first registration request information sent by a first communication device, and sending first response information to the first communication device, wherein the first registration request information is used to request registration of a second model, and the first response information includes at least one of the following: a model identifier of the second model, and an identifier of a data set associated with the second model;
[0024] Receive second registration request information sent by a fourth communication device, and send second response information to the fourth communication device, wherein the second registration request information is used to request registration of the first model, and the second response information includes at least one of the following: a model identifier of the first model, and a data set identifier associated with the first model.
[0025] In a seventh aspect, a model processing method is provided, the method comprising:
[0026] The fourth communication device performs a second operation, wherein the second operation is at least one of the following:
[0027] Sending second registration request information to a third communication device, and receiving second response information from the third communication device, wherein the second registration request information is used to request registration of the first model, and the second response information includes at least one of the following: a model identifier of the first model, and an identifier of a data set associated with the first model;
[0028] Sending fourth information to a fourth communication device, where the fourth information includes at least one of the following:
[0029] Model file information of the first model;
[0030] a model identifier of the first model;
[0031] a version identifier of the first model;
[0032] a type identifier of the first model;
[0033] parameter information of the first model;
[0034] usable condition information corresponding to the first model;
[0035] capability information of the first model;
[0036] a model identifier of a model that matches the first model;
[0037] an identifier of a data set associated with a model that matches the first model;
[0038] The storage address of the first model.
[0039] In an eighth aspect, a model processing device is provided, the device comprising:
[0040] The second execution module is configured to execute a second operation, wherein the second operation is at least one of the following:
[0041] Sending second registration request information to a third communication device, and receiving second response information from the third communication device, wherein the second registration request information is used to request registration of the first model, and the second response information includes at least one of the following: a model identifier of the first model, and an identifier of a data set associated with the first model;
[0042] Sending fourth information to a fourth communication device, where the fourth information includes at least one of the following:
[0043] Model file information of the first model;
[0044] a model identifier of the first model;
[0045] a version identifier of the first model;
[0046] a type identifier of the first model;
[0047] parameter information of the first model;
[0048] usable condition information corresponding to the first model;
[0049] capability information of the first model;
[0050] a model identifier of a model that matches the first model;
[0051] an identifier of a data set associated with a model that matches the first model;
[0052] The storage address of the first model.
[0053] In the ninth aspect, a communication device is provided, which includes a processor and a memory, wherein the memory stores a program or instruction that can be run on the processor, and when the program or instruction is executed by the processor, it implements the steps of the method described in the first aspect, or implements the steps of the method described in the third aspect, or implements the steps of the method described in the fifth aspect, or implements the steps of the method described in the seventh aspect.
[0054] In a tenth aspect, a communication device is provided, comprising a processor and a communication interface, wherein the communication interface is used to receive first information sent by a second communication device, the first information comprising model information of a first model; the processor is used to search for a second model matching the target model; the communication interface is further used to send second information to the second communication device, the second information being used to instruct the second communication device to use the target model, the first model comprising the target model;
[0055] or,
[0056] The communication interface is used to send first information to a first communication device, the first information including model information of a first model; receive second information or third information sent by the first communication device, the second information is used to instruct the second communication device to use a target model, the first model includes the target model, and the third information is used to instruct the second communication device to refuse to use the first model;
[0057] or,
[0058] The communication interface is used to perform a first operation, wherein the first operation is at least one of the following: receiving a first registration request message sent by a first communication device, and sending a first response message to the first communication device, wherein the first registration request message is used to request registration of a second model, and the first response message is at least one of the following: a model identifier of the second model, and a data set identifier associated with the second model; receiving a second registration request message sent by a fourth communication device, and sending a second response message to the fourth communication device, wherein the second registration request message is used to request registration of a first model, and the second response message includes at least one of the following: a model identifier of the first model, and a data set identifier associated with the first model;
[0059] or,
[0060] The communication interface is used to perform a second operation, and the second operation includes at least one of the following: sending a second registration request message to a third communication device and receiving a second response message from the third communication device, the second registration request message is used to request registration of a first model, and the second response message includes at least one of the following: a model identifier of the first model and a data set identifier associated with the first model; sending fourth information to a fourth communication device, the fourth information includes at least one of the following: model file information of the first model; a model identifier of the first model; a version identifier of the first model; a type identifier of the first model; parameter information of the first model; usable condition information corresponding to the first model; capability information of the first model; a model identifier of a model matching the first model; a data set identifier associated with a model matching the first model; and a storage address of the first model.
[0061] In the eleventh aspect, a readable storage medium is provided, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the method described in the first aspect are implemented, or the steps of the method described in the third aspect are implemented, or the steps of the method described in the fifth aspect are implemented, or the steps of the method described in the seventh aspect are implemented.
[0062] In the twelfth aspect, a chip is provided, 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 method described in the first aspect, or the steps of the method described in the third aspect, or the steps of the method described in the fifth aspect, or the steps of the method described in the seventh aspect.
[0063] In the thirteenth aspect, a computer program / program product is provided, wherein the computer program / program product is stored in a storage medium, and the computer program / program product is executed by at least one processor to implement the steps of the method described in the first aspect, or the steps of the method described in the third aspect, or the steps of the method described in the fifth aspect, or the steps of the method described in the seventh aspect.
[0064] In an embodiment of the present application, a first communication device receives first information sent by a second communication device, the first information including model information of a first model, and sends second information to the second communication device, the second information being used to instruct the second communication device to use a target model, the first model including the target model, that is, a communication device sends model information of a model to be used to another communication device before the model is needed, and the model is used only after the other communication device confirms that the model is allowed to be used. This can reduce the occurrence of incompatibility of models used on different communication device sides, thereby improving the task processing effect. BRIEF DESCRIPTION OF THE DRAWINGS
[0065] Figure 1 is a block diagram of a wireless communication system to which the embodiments of the present application can be applied;
[0066] Figure 2 is a schematic diagram of a neural network structure provided in an embodiment of the present application;
[0067] Figure 3 is a schematic diagram of a neuron structure provided in an embodiment of the present application;
[0068] Figure 4 is a flow chart of a model processing method provided in an embodiment of the present application;
[0069] Figure 5is a flow chart of another model processing method provided in an embodiment of the present application;
[0070] Figure 6 is a flow chart of another model processing method provided in an embodiment of the present application;
[0071] Figure 7 is a flow chart of another model processing method provided in an embodiment of the present application;
[0072] Figure 8 is a flow chart of another model processing method provided in an embodiment of the present application;
[0073] Figure 9a is a flow chart of another model processing method provided in an embodiment of the present application;
[0074] Figure 9b is a flow chart of another model processing method provided in an embodiment of the present application;
[0075] Fig.10 is a flow chart of another model processing method provided in an embodiment of the present application;
[0076] Fig.11 is a structural diagram of a model processing device provided in an embodiment of the present application;
[0077] Fig.12 is a structural diagram of another model processing device provided in an embodiment of the present application;
[0078] Fig.13 is a structural diagram of another model processing device provided in an embodiment of the present application;
[0079] Fig.14 is a structural diagram of another model processing device provided in an embodiment of the present application;
[0080] Fig.15 is a structural diagram of a communication device provided in an embodiment of the present application;
[0081] Fig.16 It is a structural diagram of the network side device provided in an embodiment of the present application.
[0082] Fig.17 is a structural diagram of a terminal provided in an embodiment of the present application;
[0083] Fig.18 It is a structural diagram of the server provided in an embodiment of the present application. DETAILED DESCRIPTION
[0084] The following will be combined with the drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of this application.
[0085] The terms "first", "second", etc. of the present application are used to distinguish similar objects, and are 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 one type, and the number of objects is not limited, for example, the first object can be one or more. In addition, "or" in the present application represents at least one of the connected objects. For example, "A or B" covers three schemes, namely, Scheme 1: including A but not including B; Scheme 2: including B but not including A; Scheme 3: including both A and B. The character " / " generally indicates that the objects associated with each other are in an "or" relationship.
[0086] The term "indication" in this application can be a direct indication (or explicit indication) or an indirect indication (or implicit indication). A direct indication can be understood as the sender explicitly informing the receiver of specific information, operations to be performed, or request results in the sent indication; an indirect indication can be understood as the receiver determining the corresponding information according to the indication sent by the sender, or making a judgment and determining the operation to be performed or the request result according to the judgment result.
[0087] It is worth noting that the technology described in the embodiments of the present application is not limited to the Long Term Evolution (LTE) / LTE-Advanced (LTE-A) system, but can also be used in other wireless communication systems, such as Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Single-carrier Frequency-Division Multiple Access (SC-FDMA) or other systems. The terms "system" and "network" in the embodiments of the present application are often used interchangeably, and the described technology can be used for the systems and radio technologies mentioned above as well as for other systems and radio technologies. The following description describes a New Radio (NR) system for illustrative purposes, and NR terminology is used in most of the following descriptions, but these technologies can also be applied to systems other than NR systems, such as the 6th generation (6 th Generation, 6G) communication system.
[0088] Figure 1A block diagram of a wireless communication system applicable to an embodiment of the present application is shown. The wireless communication system includes a terminal 11, a network side device 12 and a server 13. Among them, the terminal 11 can be a mobile phone, a tablet computer (Tablet Personal Computer), a laptop computer (Laptop Computer), a notebook computer, a personal digital assistant (Personal Digital Assistant, PDA), a handheld computer, a netbook, an ultra-mobile personal computer (Ultra-mobile Personal Computer, UMPC), a mobile Internet device (Mobile Internet Device, MID), an augmented reality (Augmented Reality, AR), a virtual reality (Virtual Reality, VR) device, a robot, a wearable device (Wearable Device), an aircraft (flight vehicle), a vehicle-mounted device (Vehicle User Equipment, VUE), a ship-mounted device, a pedestrian terminal (Pedestrian User Equipment, PUE), a smart home (home appliances with wireless communication functions, such as refrigerators, televisions, washing machines or furniture, etc.), a game console, a personal computer (Personal Computer, PC), a teller machine or a self-service machine and other terminal side devices. Wearable devices include: smart watches, smart bracelets, smart headphones, smart glasses, smart jewelry (smart bracelets, smart bracelets, smart rings, smart necklaces, smart anklets, smart anklets, etc.), smart wristbands, smart clothing, etc. Among them, the vehicle-mounted device can also be called a vehicle-mounted terminal, a vehicle-mounted controller, a vehicle-mounted module, a vehicle-mounted component, a vehicle-mounted chip or a vehicle-mounted unit, etc. It should be noted that the specific type of the terminal 11 is not limited in the embodiment of the present application. The network side device 12 may include access network equipment and core network equipment, wherein the access network equipment may also be referred to as a radio access network (Radio Access Network, RAN) equipment, a radio access network function or a radio access network unit. The access network equipment may include a base station, a wireless local area network (Wireless Local Area Network, WLAN) access point (AccessPoint, AP) or a wireless fidelity (Wireless Fidelity, WiFi) node, etc.Among them, the base station may be referred to as a Node B (NB), an evolved Node B (eNB), a next generation Node B (gNB), a New Radio Node B (NR Node B), an access point, a Relay Base Station (RBS), a Serving Base Station (SBS), a Base Transceiver Station (BTS), a radio base station, a radio transceiver, a Basic Service Set (BSS), an Extended Service Set (ESS), a Home Node B (HNB), a Home Evolved Node B (home evolved Node B), 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 specific technical terms. It should be noted that in the embodiments 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.
[0089] 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, or L-NEF), binding support function (Binding Support Function, BSF), application function (Application Function, AF), etc. It should be noted that in the embodiments of the present application, only the core network device in the NR system is introduced as an example, and the specific type of the core network device is not limited.
[0090] The server 13 may be a supplier server of the terminal 11 (ie, a UE Vendor Server), a third-party server, or an application server.
[0091] For ease of understanding, some contents involved in the embodiments of the present application are described below:
[0092] 1. Artificial Intelligence (AI)
[0093] In the communication network, AI methods are introduced to improve network communication performance. Network-side devices or terminals can perform AI or machine learning (ML) model training based on training data to obtain a model suitable for a certain AI task, and can send the trained AI / ML model to other terminals so that other terminals can infer the inference input data of the AI task based on the AI / ML model and obtain the inference result data corresponding to the AI task. For example, the terminal predicts the signal strength based on the AI / ML model and sends the predicted signal strength to the base station so that the base station can make a more accurate selection of the switching timing.
[0094] Artificial intelligence has been widely used in various fields. AI models can be implemented based on a variety of algorithms, such as neural networks, decision trees, support vector machines, Bayesian classifiers, etc. The following uses a neural network as an example for AI model description, but does not constitute a limitation on the type of AI module.
[0095] The structure of a neural network can be as follows Figure 2 As shown in the figure, X1, X2…Xn are input values, Y is the output result, a circle represents a neuron, and many neurons form the input layer, hidden layer and output layer, that is, a neural network. Among them, the number of hidden layers and the number of neurons in each layer are the network structure of the neural network.
[0096] The structure of each neuron is as follows: Figure 3 As shown, a1, a2, ... aK (i.e. X1 ... Xn mentioned above) are inputs, w is the weight (multiplicative coefficient), b is the bias (additive coefficient), σ (.) is the activation function, and z is the output value. Common activation functions include Sigmoid function, hyperbolic tangent (tanh) function, Rectified Linear Unit (ReLU) function, etc. The parameter information of all neurons and the algorithm used together is the parameter information of the entire neural network, which is a very important part of the AI model file.
[0097] In actual use, an AI model refers to a file containing elements such as network structure and parameter information. The trained AI model can be directly reused by its framework platform without repeated construction or learning, and can directly perform intelligent functions such as judgment and recognition.
[0098] 2. Channel State Information (CSI)
[0099] When CSI performs compression model training, joint training is required because the decoder is in the base station and the encoder is in the terminal.
[0100] Accurate CSI is crucial to channel capacity, especially for multi-antenna systems. The transmitter can optimize signal transmission based on CSI to better match the channel status. For example, the Channel Quality Indicator (CQI) can be used to select an appropriate Modulation and Coding Scheme (MCS) to achieve link adaptation; the Precoding Matrix Indicator (PMI) can be used to implement Eigen Beamforming to maximize the strength of the received signal, or to suppress interference (such as interference between cells, interference between multiple users, etc.). Therefore, since the introduction of multi-input multi-output (MIMO) technology, CSI acquisition has always been a research hotspot.
[0101] Typically, the base station sends a channel state information reference signal (CSI-RS) on certain time-frequency resources in a certain time slot. The terminal performs channel estimation based on the CSI-RS, calculates the channel information on this slot, and feeds back the PMI to the base station through the codebook. The base station combines the channel information based on the codebook information fed back by the terminal, and uses this to perform data precoding and multi-user scheduling before the next CSI reporting.
[0102] In order to reduce overhead, the base station can precode the CSI-RS in advance and send the encoded CSI-RS to the terminal. The terminal sees the channel corresponding to the encoded CSI-RS. The terminal only needs to select several ports with higher strength from the ports indicated by the network side and report the coefficients corresponding to these ports.
[0103] The CSI compression recovery process is that the terminal estimates CSI-RS, calculates channel information, obtains the encoding result through the calculated channel information or the original estimated channel information through the encoding network, and sends the encoding result to the base station. The base station receives the encoded result and inputs it into the decoding network to recover the channel information.
[0104] Specifically, the CSI compression feedback scheme based on neural network is to compress and encode the channel information at the terminal, send the compressed content to the base station, and decode the compressed content at the base station to restore the channel information. At this time, the decoding network of the base station and the encoding network of the terminal need to be jointly trained to achieve a reasonable match. The input of the encoding model is channel information, and the output is encoding information, that is, channel feature information. The input of the decoding model is encoding information, and the output is restored channel information.
[0105] Exemplarily, the UE side and the network side jointly training the model may include the following steps:
[0106] Step S1: UE performs model training.
[0107] In this step, the UE side trains the CSI generation part (i.e., encoder) and the CSI reconstruction part (i.e., decoder) on the UE side. The CSI reconstruction part on the UE side will not be used for inference.
[0108] Step S2: The UE generates an exchange data set.
[0109] Step S3: The UE sends the exchange data set to the network.
[0110] After the UE side training is completed, the UE side shares a series of information (eg, data set) with the network (Network, NW) side (eg, base station, base station gateway, etc.), and this series of information can be used for the NW side to perform training on the CSI reconstruction part.
[0111] Step S4: The network side trains the CSI reconstruction part.
[0112] The NW side performs training of the CSI reconstruction part based on the received information (such as a data set).
[0113] Step S5: model reasoning.
[0114] The UE side and the NW side use matching models for reasoning.
[0115] The model processing method provided in the embodiment of the present application is described in detail below through some embodiments and their application scenarios in combination with the accompanying drawings.
[0116] See also Figure 4 , Figure 4 is a flowchart of a model processing method provided in an embodiment of the present application, which can be executed by a first communication device, such as Figure 4 As shown, the following steps are included:
[0117] Step 401: A first communication device receives first information sent by a second communication device, where the first information includes model information of a first model.
[0118] In this embodiment, the first communication device may include but is not limited to a first network side device (e.g., access network device, access network gateway, operation administration and maintenance (OAM) function, etc.), a terminal or a server, etc. The second communication device is a peer device of the first communication device or a communication device that needs to use a model that matches the model requested by the first communication device. For example, if the first communication device is a first network side device, the second communication device may be a server or a terminal, etc.; if the first communication device is a server or a terminal, etc., the second communication device may be a first network side device. The server may be a terminal provider server, a third-party server, an application server, etc.
[0119] The model information of the first model may include but is not limited to at least one of the model identifier of the first model, the version identifier of the first model, the dataset identifier associated with the first model, the type identifier of the first model, etc. It should be noted that the first model may include one or more models. In the case where the first model includes multiple models, the model information of the first model may include model information of each of the multiple models.
[0120] Exemplarily, the first information can be used to indicate the first model that the second communication device requests to use, or the first information can be used to indicate the first model that the second communication device needs to use, or the first information can be used to indicate the first model to be used or to be used by the second communication device, or the first information can be used to indicate the first model of the second communication device, or the first information can be used to indicate the first model that the first communication device needs to activate or to be activated, or the first information can be used to inquire the first communication device whether there is a model matching the first model, etc.
[0121] Step 402: The first communication device sends second information to the second communication device, where the second information is used to instruct the second communication device to use a target model, and the first model includes the target model.
[0122] The target model may be any model in the first model.
[0123] Exemplarily, after receiving the first information, the first communication device may determine whether there is a model that matches the model in the first model based on the first information. If it is determined that there is a model that matches the target model in the first model, the second information may be sent to the second communication device, wherein the second information is used to instruct the second communication device to use the target model, and may also be referred to as the second information used to instruct the second communication device to activate the target model, or the second information is used to indicate that the first communication device has a model that matches the target model, etc. The second model that matches the target model may also be referred to as a second model associated with the target model or a second model corresponding to the target model, etc. The target model may be used in pair with the second model, that is, when the second communication device uses the target model, the first communication device needs to use the second model accordingly.
[0124] It should be noted that the words such as association and matching in this application have the same meaning and can be replaced with each other. For example, the second model determined by the first communication device to match the target model and the second model determined by the first communication device to match the target model have the same meaning, which will not be repeated later. The words such as requesting to use a model and needing to use a model in this application have the same meaning and can be replaced with each other. For example, the first information is used to indicate that the second communication device requests to use a first model and the first information is used to indicate that the second communication device needs to use a first model have the same meaning, which will not be repeated later. The words such as activation and use in this application have the same meaning and can be replaced with each other.
[0125] For example, if the above-mentioned first information includes the model identifier of the first model, then based on the model identifier of the first model, it can be searched whether the first communication device has a model identifier identical to the model identifier of the first model stored therein, and the model associated with the model identifier identical to the model identifier of the first model can be used as the model matching the first model (i.e., the second model). For example, the first communication device can store the model identifier of its trained model and the model identifier of the model matching the trained model in an associated manner, so that the model identifier of the model matching the trained model can be quickly determined based on the model identifier of the first model; or, if the above-mentioned first information includes the model identifier of the first model, then the service targeted by the first communication device can be determined based on the model identifier of the first model, and then the second model can be determined based on the targeted service; or, if the above-mentioned first information includes the data set identifier associated with the first model, then the model trained using the data set identified by the data set identifier can be determined based on the data set identifier, and the model can be used as the model matching the first model.
[0126] The model processing method provided in the embodiment of the present application receives, by a first communication device, first information sent by a second communication device, wherein the first information includes model information of a first model, and sends second information to the second communication device, wherein the second information is used to instruct the second communication device to use a target model, and the first model includes the target model, that is, before a communication device needs to use the model, it first sends model information of the model to be used to another communication device, and then uses the model when the other communication device confirms that the model is allowed to be used. This can reduce the occurrence of incompatibility of models used by different communication devices, thereby improving the task processing effect.
[0127] It should be noted that the words "allow" in this application have the same meaning as the words "can", "may", "capable", etc. and can be replaced with each other. For example, the second information used to indicate that the second communication device allows the use of the first model and the second information used to indicate that the second communication device can use the first model have the same meaning and will not be repeated later.
[0128] Optionally, the first information further includes at least one of the following:
[0129] capability information of the first model;
[0130] usable condition information corresponding to the first model;
[0131] First indication information, where the first indication information is used to indicate that the second communication device requests to use the first model.
[0132] In this embodiment, the capability information of the first model may be used to indicate the functions or services supported or corresponding to the first model, or to indicate the service type or function type corresponding to the first model.
[0133] The usable condition information corresponding to the first model may include condition information such as time range and area range, and may be used to indicate at least one of the usable time range and area range corresponding to the first model. The usable condition information corresponding to the first model may also be referred to as the condition information under which the first model can be used.
[0134] The above-mentioned first indication information is used to indicate that the second communication device requests to use the first model, and may also be referred to as the first indication information used to indicate the first model to be used by the second communication device, or the above-mentioned first indication information is used to indicate the first model of the second communication device, or the above-mentioned first indication information is used to indicate the first model that needs to be activated or to be activated by the first communication device, or the above-mentioned first indication information is used to inquire the first communication device whether there is a model matching the first model, etc.
[0135] Optionally, the model information of the first model includes at least one of the following: a model identifier of the first model, a version identifier of the first model, a data set identifier associated with the first model, and a type identifier of the first model.
[0136] The data set identifier associated with the first model may include an identifier of a data set used for training the first model. The data set identifier for training the first model may refer to the first model obtained by training with the data set, that is, the data set used for training the first model. The data set associated with the first model may also refer to a data set generated by the first model, which is related to the data set for training the first model, such as containing part of the data set for training the first model, but also modified or added data, etc. The data set generated by the first model may be a data set specifically required for other communication devices to train a model that matches the model.
[0137] The above type identifier can be used to identify the type of model trained by different communication equipment sides based on the data set identified by the above data set identifier. For example, for CSI prediction, the model type of the model used for the terminal side is the encoder type, and the model type of the model used for the network side device is the decoder type, etc.
[0138] Optionally, the type identifier may also indicate functional information of the model, or explain what task the model is for, etc.
[0139] Optionally, the model information of the first model may also include a model file corresponding to the first model, or the model itself corresponding to the first model.
[0140] In some optional embodiments, the model information of the first model includes a model identifier of the first model (ie, Model ID) and an optional version identifier of the first model (ie, Version ID).
[0141] In some optional embodiments, the model information of the first model includes a data set identifier (ie, DataSet ID) associated with the first model and an optional type identifier (ie, Type ID) of the first model.
[0142] Optionally, the second information includes at least one of the following: second indication information, model information of the target model, model information of the second model;
[0143] The second indication information is used to indicate that the second communication device is allowed to use the target model, and the second model is a model that matches the target model.
[0144] Optionally, the model information of the second model may also include a model file corresponding to the second model, or the model itself corresponding to the second model.
[0145] The second indication information is used to indicate that the second communication device is allowed to use the target model, and may also be referred to as the second indication information being used to instruct the second communication device to activate the target model, or the second indication information being used to indicate that the second communication device has a model that matches the target model, etc.
[0146] The model information of the target model may include at least one of the following: a model identifier of the target model, a version identifier of the target model, an identifier of a data set associated with the target model, and a type identifier of the target model.
[0147] Optionally, the model information of the second model may include at least one of the following: a model identifier of the second model, a version identifier of the second model, a data set identifier associated with the second model, and a type identifier of the second model.
[0148] It should be noted that, for the model information of the first model and the model information of the target model in this embodiment, reference may be made to the relevant description of the model information of the first model in the aforementioned embodiment, which will not be elaborated here.
[0149] In some optional embodiments, if the model information of the target model includes the model identifier of the target model and optionally the version identifier of the target model, the model information of the second model may also include the model identifier of the second model and optionally the version identifier of the second model; if the model information of the target model includes the dataset identifier associated with the target model and the type identifier of the target model, the model information of the second model may also include the dataset identifier associated with the second model and the type identifier of the second model.
[0150] It should be noted that the model version identifier is optional information and does not necessarily appear at the same time as the model identifier.
[0151] Optionally, before the first communication device sends the second information to the second communication device, the method further includes:
[0152] The first communications device determines a second model that matches the target model.
[0153] In this embodiment, after the first communication device receives the first information, it can determine whether there is a second model that matches the target model based on the model information of the target model in the first information. Among them, the second model that matches the target model can also be referred to as a second model associated with the target model or a second model corresponding to the target model, etc. The determination by the first communication device whether there is a second model that matches the target model can also be referred to as the first communication device aligning the target model and the second model or the first communication device associating the target model and the second model. The above process can also be that the first communication device determines the second model that matches the target model, and the current environment or the state of the second communication device meets the usable conditions of the target model or the second model, and the first communication device uses or prepares to use the second model. The first communication performs operations such as alignment, matching or associating the target model and the second model, which can be understood as the first communication device determining that the first communication device itself has a second model that matches it after knowing that the second communication device uses the target model, and then uses the second model to better interact with the second communication device, etc.
[0154] For example, if the above-mentioned first information includes the model identifier of the target model, then based on the model identifier of the target model, it is possible to search whether the first communication device has a model identifier identical to the model identifier of the target model, and the model associated with the model identifier identical to the model identifier of the target model can be used as a model matching the target model (i.e., the second model). For example, the first communication device can store the model identifier of its trained model and the model identifier of the model matching the trained model in an associated manner, so that the model identifier of the model matching the trained model can be quickly determined based on the model identifier of the target model; or, if the above-mentioned first information includes the model identifier of the target model, then the service targeted by the first communication device can be determined based on the model identifier of the target model, and then the second model can be determined based on the targeted service. For example, the capability information of the target model can be searched based on the model identifier of the target model, and then the service targeted by the first communication device can be determined based on the capability information of the target model; or, if the above-mentioned first information includes a data set identifier associated with the target model, then the model trained using the data set identified by the data set identifier can be determined based on the data set identifier, and the model can be used as a model matching the target model.
[0155] It should be noted that the terms business, function, task, data analysis, etc. in this application have the same meaning and can be replaced with each other.
[0156] Optionally, when the first communication device determines the second model that matches the target model, it can be determined through an association relationship or mapping relationship stored locally in the first communication device, or through an acquired association relationship or mapping relationship, or perform model alignment, etc.
[0157] Exemplarily, if the first communication device stores an association relationship or mapping relationship between model IDs, model alignment can be performed based on the mapping relationship or association relationship. For example, model ID2 can be found based on model ID1 and the above association relationship or mapping relationship. If the first communication device does not store an association relationship or mapping relationship between model IDs, the above association relationship or mapping relationship can be obtained from a third communication device, and model alignment can be performed based on the mapping relationship or association relationship.
[0158] Optionally, the first communication device sending second information to the second communication device includes:
[0159] When determining that there is a second model matching the target model, the first communication device sends second information to the second communication device.
[0160] In this embodiment, when the first communication device determines that there is a second model that matches the target model, the second information can be sent to the second communication device, so that the second communication device can perform data processing based on the target model. Correspondingly, the first communication device can perform data processing based on the second model. For example, for the CSI prediction task, the terminal encodes the channel information to be transmitted through the target model (i.e., the encoding model) to obtain the encoding result, and sends the encoding result to the network side device. The network side device decodes the encoding result transmitted by the UE through the second model (i.e., the decoding model) to obtain the recovered channel information.
[0161] It should be noted that, when the first communication device determines that there is a second model matching the target model, it may also determine whether to send the second information to the second communication device in combination with some other conditions.
[0162] Optionally, the second communication device sends the second information, including:
[0163] The first communication device sends second information to the second communication device when it is determined that there is a second model matching the target model and the second communication device meets the usable condition corresponding to the target model.
[0164] The usable conditions corresponding to the above-mentioned target model may also be referred to as the validity conditions (i.e., validity conditions) of the target model, or may also be referred to as the use conditions, which may include but are not limited to time range, area range, and other conditions. For example, the target model may be used in a first time period or a first area. The above-mentioned second communication device satisfies the usable conditions corresponding to the target model. For example, if the second communication device is located within the usable location range corresponding to the target model, it is determined that the second communication device satisfies the usable conditions corresponding to the target model, or if the time when the second communication device currently requests to use the target model is within the usable time range corresponding to the target model, it is determined that the second communication device satisfies the usable conditions corresponding to the target model, etc. It may also be that multiple usable conditions are met at the same time, etc. Among them, area and location have similar meanings and can be used interchangeably.
[0165] In this embodiment, when the first communication device determines that there is a second model matching the target model and the second communication device meets the usable conditions corresponding to the target model, the first communication device sends second information to the second communication device. This can ensure that the first communication device and the second communication device can more accurately align the models used.
[0166] Optionally, the method further includes:
[0167] When it is determined that there is no model matching the first model or the second communication device does not meet the usable condition corresponding to the first model, the first communication device sends third information to the second communication device, where the third information is used to indicate that the second communication device is denied to use the first model.
[0168] The above-mentioned absence of a model matching the first model can be understood as the absence of a model matching any model in the first model. The above-mentioned failure to meet the usable conditions corresponding to the first model can be understood as failure to meet the usable conditions corresponding to any model in the first model, or failure to meet the usable conditions corresponding to the model with a matching model in the first model. For example, the first communication device has a model matching the third model, and the second communication device does not meet the usable conditions corresponding to the three models, and the first model includes the third model.
[0169] The third information is used to indicate that the second communication device is denied to use the first model. It can also be said that the third information is used to indicate that the second communication device does not use or activate the first model. Alternatively, the third information is used to indicate that the second communication device does not have a model that matches the first model.
[0170] Exemplarily, the first communication device may send third information to the second communication device if any of the following conditions is met: there is no model matching any of the first models; there is a model matching the third model but does not meet the usable condition corresponding to the third model.
[0171] In this embodiment, when the first communication device determines that there is no second model matching the first model or the second communication device does not meet the usable conditions corresponding to the first model, the first communication device can send third information to the second communication device. This can avoid the second communication device using the first model when the first communication device cannot use the second model or the second model does not exist.
[0172] Optionally, the third information includes at least one of the following: third indication information, model information of the first model, and fourth indication information;
[0173] The third indication information is used to indicate that the second communication device is denied from using the first model, and the fourth indication information is used to indicate a reason value for denying the second communication device from using the first model.
[0174] The third indication information is used to indicate that the second communication device is denied to use the first model. It can also be called that the third indication information is used to indicate that the second communication device does not use or activate the first model. Alternatively, the third indication information is used to indicate that the second communication device does not have a model that matches the first model, etc.
[0175] The fourth indication information is used to indicate the reason value for refusing the second communication device to use the first model, for example, there is no model matching the first model, the model information of the first model has not been received, the conditions for the use of the first model are not met, etc.
[0176] Optionally, before the first communication device receives the first information sent by the second communication device, the method further includes:
[0177] The first communication device sends first registration request information to the third communication device, where the first registration request information is used to request registration of the second model;
[0178] The first communication device receives a first response message sent by the third communication device, and the first response message includes a model identifier of the second model and / or a data set identifier associated with the second model. In this embodiment, the third communication device can be any communication device that can perform operations such as model storage and identifier allocation. Exemplarily, the third communication device can be a second network side device, such as a core network element such as NEF, NRF or Analytics Data Repository Function (ADRF). The data set identifier associated with the second model can be used to identify a data set used to train the second model.
[0179] Specifically, the first communication device may send the first registration request information to the third communication device when the second model is acquired. For example, the first communication device may send the first registration request information to the third communication device when the second model is received or the second model is obtained through training. The first registration request information may include relevant information of the second model, such as model file information of the second model, model description information of the second model, etc., so that the third communication device can store the relevant information of the second model, assign a model identifier or an associated data set identifier to the second model, and send it to the first communication device.
[0180] In some optional embodiments, the first response information may also include a version identifier of the second model or a type identifier of the second model, that is, the first response information includes the model identifier of the second model and / or the version identifier of the second model, or the first response information includes the data set identifier associated with the second model and / or the type identifier of the second model.
[0181] Optionally, the first response information may further include: successful storage indication information, used to indicate that the third communication device successfully receives the registration information and successfully stores the registration information.
[0182] In actual applications, different communication devices can register each acquired model to a third communication device, and receive the model identifier or associated data set identifier assigned to each model by the third communication device. This not only facilitates the unified management of models, but also facilitates the alignment of models between different communication devices.
[0183] Optionally, the first registration request information includes at least one of the following:
[0184] Model file information of the second model;
[0185] Model description information of the second model.
[0186] In this embodiment, the above-mentioned model file information may also be referred to as a model file, which may include but is not limited to information such as the architecture or network structure of the model, the parameters of the model, and the values of the parameters of the model.
[0187] Optionally, the model description information of the second model includes at least one of the following:
[0188] a version identifier of the second model;
[0189] a type identifier of the second model;
[0190] parameter information of the second model;
[0191] usable condition information corresponding to the second model;
[0192] capability information of the second model;
[0193] Producer information of the second model;
[0194] consumer information of the second model;
[0195] a model identifier of a model that matches the second model;
[0196] The identifier of the dataset associated with the model that matches the second model.
[0197] The parameter information of the second model may include a parameter list of the second model, or a parameter type list corresponding to the model, which is used to indicate the input and output parameter types involved in the model. The usable condition information corresponding to the second model may include condition information such as time range and area range, which may be used to indicate at least one of the time and area where the second model can be used. The capability information of the second model may be used to indicate the functions or services supported or corresponding to the second model, or to indicate the service type or function type corresponding to the second model.
[0198] The producer information of the second model can be used to indicate the device that generates the second model, for example, a UE supplier server. The producer information of the second model can also include a producer identifier, or a producer manufacturer identifier, etc. The consumer information of the second model can be used to indicate the device that uses the second model. For example, the consumer information of the second model can include at least one of the UE identification information and the supplier's identification information. It should be noted that when the consumer information is the supplier's identification information, it means that all devices belonging to the supplier can use the second model.
[0199] In some optional embodiments, when a model matching the second model is already known before the first communication device acquires the second model (for example, the first communication device trains the second model), the model description information of the second model may also include model information of the model matching the second model, such as a model identifier or a data set identifier, etc., which can be used to inform a third communication device of the association between the second model and other models.
[0200] Optionally, before the first communication device receives the first information sent by the second communication device, the method further includes:
[0201] The first communication device trains a second model based on the first data set;
[0202] The first communication device determines a second data set;
[0203] The first communication device sends first exchange data, where the first exchange data includes the second data set, and the second data set is used to train the target model.
[0204] The first data set may be a data set for training the second model, and may be determined based on data obtained by the first communication device, etc. The first data set may be reasonably set according to the function or service corresponding to the model to be trained, for example, for model training for CSI prediction, the first data set may be historical CSI data, and the second model may be an Encoder model for CSI encoding or a Decoder model for CSI decoding.
[0205] Specifically, the first communication device may first train the second model according to the first data set, and then send the second data set to the second communication device or the server corresponding to the second communication device, so that the second communication device or the server corresponding to the second communication device can train the first model based on the second data set. The second data set is a data set related to the training of the second model. For example, the second data set may include at least one of input data, intermediate data, and output data in the training process of the second model.
[0206] Optionally, the first communication device may send the first exchange data through an existing solution (such as a non-access stratum (NAS) message, a radio resource control (RRC) message, etc.), or may transmit the first data through a data collection element (DCAF), or transmit the first data based on an event exposure (EVEX) architecture, etc.
[0207] Taking the first communication device as a base station, the second communication device as a UE, and CSI prediction model training as an example, the base station can jointly train the encoder model and decoder model for the base station side based on the collected historical CSI data, and determine the data set for training the encoder model on the UE side. For example, the output data of the encoder model for the base station side is used as the label data for training the encoder model for the UE side, and the input data of the encoder model for the base station side or the output data of the decoder model for the base station side is used as the input data for training the encoder model for the UE side, and the determined data set is sent to the UE or the server corresponding to the UE. The UE or the server corresponding to the UE can train the encoder model for the UE side based on the received data set.
[0208] In this embodiment, the data sets used for training the first model and the second model are associated, which can improve the accuracy of the first model and the second model when used in pairs.
[0209] Optionally, the first exchange data further includes at least one of the following:
[0210] model information of the second model;
[0211] The usable condition information corresponding to the second model.
[0212] The model information of the second model and the usable condition information of the second model in this embodiment can be found in the relevant description of the aforementioned embodiment, and will not be elaborated here.
[0213] Optionally, the usable condition information includes at least one of the following: usable time information, usable location information.
[0214] Among them, the above-mentioned usable time information is used to indicate the time range in which the model can be used, and the above-mentioned usable location information can be used to indicate the location range in which the model can be used, for example, one or more tracking areas (TA), cells, or longitude and latitude ranges (such as by limiting longitude and latitude to enclose an area), etc.
[0215] Exemplarily, the usable condition information corresponding to the first model may include at least one of the following: usable time information corresponding to the first model, and usable location information corresponding to the first model. The usable condition information corresponding to the second model may include at least one of the following: usable time information corresponding to the second model, and usable location information corresponding to the second model.
[0216] Optionally, before the first communication device receives the first information sent by the second communication device, the method further includes:
[0217] The first communication device receives second exchange data, where the second exchange data includes a third data set, and the third data set is a data set related to the target model training;
[0218] The first communication device trains a second model according to the third data set.
[0219] In this embodiment, the second communication device or the server corresponding to the second communication device can first train the target model, determine the third data set and send it to the first communication device, and the first communication device trains the second model based on the received third data set.
[0220] Taking the first communication device as a base station, the second communication device as a UE, and CSI prediction model training as an example, the UE or the server corresponding to the UE can jointly train the Encoder model and the Decoder model for the UE side based on the collected historical CSI data, and determine the data set for training the Decoder model on the base station side, for example, using the output data of the Encoder model for the UE side as the input data for training the Decoder model for the base station side, using the input data of the Encoder model for the UE side as the label data for training the Decoder model for the base station side, and sending the determined data set to the base station. The base station can train the Decoder model for the base station side based on the received data set.
[0221] In this embodiment, the data sets used for training the target model and the second model are associated, which can improve the accuracy of the target model and the second model when used in pairs.
[0222] Optionally, the second exchange data further includes at least one of the following:
[0223] Model information of the target model;
[0224] The usable condition information corresponding to the target model.
[0225] The model information of the target model and the usable condition information of the target model in this embodiment can be found in the relevant description of the aforementioned embodiment, and will not be elaborated here.
[0226] Optionally, the first communication device is a first network side device, and the second communication device is a terminal.
[0227] The first network side device mentioned above may be an access network device (eg, a base station), an access network gateway (eg, a base station gateway) or an OAM, etc.
[0228] See also Figure 5 , Figure 5 is a flow chart of a model processing method provided in an embodiment of the present application, which can be executed by a second communication device, such as Figure 5 As shown, the following steps are included:
[0229] Step 501: A second communication device sends first information to a first communication device, where the first information includes model information of a first model;
[0230] Step 502: The second communication device receives second information or third information sent by the first communication device, wherein the second information is used to instruct the second communication device to use a target model, the first model includes the target model, and the third information is used to instruct the second communication device to deny use of the first model.
[0231] Optionally, the first information further includes at least one of the following:
[0232] capability information of the first model;
[0233] usable condition information corresponding to the first model;
[0234] First indication information, where the first indication information is used to indicate that the second communication device requests to use the first model.
[0235] Optionally, the model information of the first model includes at least one of the following: a model identifier of the first model, a version identifier of the first model, a data set identifier associated with the first model, and a type identifier of the first model.
[0236] Optionally, the second information includes at least one of the following: second indication information, model information of the target model, model information of the second model;
[0237] The second indication information is used to indicate that the second communication device is allowed to use the target model, and the second model is a model that matches the target model.
[0238] Optionally, the third information includes at least one of the following: third indication information, model information of the first model, and fourth indication information;
[0239] The third indication information is used to indicate that the second communication device is denied from using the first model, and the fourth indication information is used to indicate a reason value for denying the second communication device from using the first model.
[0240] Optionally, the model information of the second model includes at least one of the following: a model identifier of the second model, a version identifier of the second model, a data set identifier associated with the second model, and a type identifier of the second model.
[0241] Optionally, before the second communication device sends the first information to the first communication device, the method further includes:
[0242] The second communication device sends second registration request information to the third communication device, where the second registration request information is used to request registration of the first model;
[0243] The second communication device receives second response information sent by the third communication device, where the second response information includes at least one of the following: a model identifier of the first model, and a data set identifier associated with the first model.
[0244] In this embodiment, when the first model is obtained by training by the second communication device, the second communication device sends a second registration request message to the third communication device, and the second registration request message is used to request registration of the first model. The third communication device can then store relevant information of the first model, assign a model identifier or an associated data set identifier to the first model, and send it to the second communication device.
[0245] Optionally, the second registration request information includes at least one of the following:
[0246] Model file information of the first model;
[0247] Model description information of the first model.
[0248] Optionally, the model description information of the first model includes at least one of the following:
[0249] the version number of the first model;
[0250] a type identifier of the first model;
[0251] parameter information of the first model;
[0252] usable condition information corresponding to the first model;
[0253] capability information of the first model;
[0254] Producer information of the first model;
[0255] consumer information of the first model;
[0256] a model identifier of a model that matches the first model;
[0257] The identifier of the data set associated with the model that matches the first model.
[0258] Optionally, before the second communication device sends the first information to the first communication device, the method further includes:
[0259] The second communication device receives fourth information;
[0260] The fourth information includes at least one of the following:
[0261] Model file information of the first model;
[0262] a model identifier of the first model;
[0263] a version identifier of the first model;
[0264] a type identifier of the first model;
[0265] parameter information of the first model;
[0266] usable condition information corresponding to the first model;
[0267] capability information of the first model;
[0268] a model identifier of a model that matches the first model;
[0269] an identifier of a data set associated with a model that matches the first model;
[0270] The storage address of the first model.
[0271] The storage address of the first model is, for example, an IP address, a node address, etc. The second communication device can obtain the model file information of the first model based on the storage address of the first model.
[0272] Specifically, the second communication device may receive the fourth information from the third communication device, or may receive the fourth information from the fourth communication device. Exemplarily, the second communication device may send a request message to the third communication device or the fourth communication device to request to obtain the first model, and then the third communication device or the fourth communication device may send the fourth information to the second communication device. The second communication device may be a terminal. The third communication device may be any communication device that can perform operations such as model storage and identifier allocation. For example, the third communication device may be a core network element such as NEF, NRF or ADRF. The fourth communication device may be a server, such as a terminal supplier server, a third-party server or an application server.
[0273] Optionally, the usable condition information includes at least one of the following: usable time information, usable location information.
[0274] Optionally, the second communication device receives fourth information including at least one of the following:
[0275] The second communication device receives fourth information from a fourth communication device in an Over The Top (OTT) manner, where the fourth communication device is a server corresponding to the second communication device;
[0276] The second communication device receives fourth information from the fourth communication device through the second network side device;
[0277] The second communication device receives fourth information from a third communication device, the third communication device having the first model registered therein.
[0278] In this embodiment, the second communication device may be a terminal. The third communication device may be any communication device that can perform operations such as model storage and identity allocation. For example, the third communication device may be a core network element such as NEF, NRF or ADRF. The fourth communication device may be a server, such as a terminal supplier server, a third-party server or an application server.
[0279] The above-mentioned receiving the fourth information from the fourth communication device in an OTT manner can be understood as receiving the fourth information from the fourth communication device in an application layer manner, or without passing through the core network.
[0280] The second communication device receives the fourth information from the fourth communication device through the second network side device. For example, the second communication device receives the fourth information from the fourth communication device through a core network element (for example, a data collection application function (DCAF)).
[0281] In some optional embodiments, the second communication device may periodically request whether there is new model file information from the third communication device or the fourth communication device. When the third communication device or the fourth communication device has new model file information, the second communication device is notified, and the second communication device obtains the new model file information from the third communication device or the fourth communication device according to the notification.
[0282] Optionally, the first communication device is a first network side device, and the second communication device is a terminal.
[0283] It should be noted that the implementation method of this embodiment can be found in Figure 4 The relevant description of the illustrated embodiment is not repeated here.
[0284] See also Figure 6 , Figure 6 is a flow chart of a model processing method provided in an embodiment of the present application, which can be executed by a second communication device, such as Figure 6 As shown, the following steps are included:
[0285] Step 601: The third communication device performs a first operation, wherein the first operation is at least one of the following:
[0286] receiving first registration request information sent by a first communication device, and sending first response information to the first communication device, wherein the first registration request information is used to request registration of a second model, and the first response information includes at least one of the following: a model identifier of the second model, and an identifier of a data set associated with the second model;
[0287] Receive second registration request information sent by a fourth communication device, and send second response information to the fourth communication device, wherein the second registration request information is used to request registration of the first model, and the second response information includes at least one of the following: a model identifier of the first model, and a data set identifier associated with the first model.
[0288] Specifically, the fourth communication device can send a second registration request message to the third communication device after training the second model, wherein the second registration request message may include relevant information of the first model, for example, model file information of the first model, model description information of the first model, etc., so that the third communication device can store the relevant information of the first model, assign a model identifier or an associated data set identifier to the first model and send it to the first communication device.
[0289] In some optional embodiments, the second response information may also include a version identifier of the first model or a type identifier of the first model, that is, the second response information includes the model identifier of the first model and / or the version identifier of the first model, or the second response information includes the data set identifier associated with the first model and / or the type identifier of the first model.
[0290] Optionally, the first registration request information includes at least one of the following:
[0291] Model file information of the second model;
[0292] Model description information of the second model.
[0293] Optionally, the model description information of the second model includes at least one of the following:
[0294] a version identifier of the second model;
[0295] a type identifier of the second model;
[0296] parameter information of the second model;
[0297] usable condition information corresponding to the second model;
[0298] capability information of the second model;
[0299] Producer information of the second model;
[0300] consumer information of the second model;
[0301] a model identifier of a model that matches the second model;
[0302] The identifier of the dataset associated with the model that matches the second model.
[0303] Optionally, the second registration request information includes at least one of the following:
[0304] Model file information of the first model;
[0305] Model description information of the first model.
[0306] Optionally, the model description information of the first model includes at least one of the following:
[0307] a version identifier of the first model;
[0308] a type identifier of the first model;
[0309] parameter information of the first model;
[0310] usable condition information corresponding to the first model;
[0311] capability information of the first model;
[0312] Producer information of the first model;
[0313] consumer information of the first model;
[0314] a model identifier of a model that matches the first model;
[0315] The identifier of the data set associated with the model that matches the first model.
[0316] Optionally, the method further includes:
[0317] The third communication device sends fourth information to the second communication device;
[0318] The fourth information includes at least one of the following:
[0319] Model file information of the first model;
[0320] a model identifier of the first model;
[0321] a version identifier of the first model;
[0322] a type identifier of the first model;
[0323] parameter information of the first model;
[0324] usable condition information corresponding to the first model;
[0325] capability information of the first model;
[0326] a model identifier of a model that matches the first model;
[0327] an identifier of a data set associated with a model that matches the first model;
[0328] The storage address of the first model.
[0329] Optionally, the usable condition information includes at least one of the following: usable time information, usable location information.
[0330] It should be noted that the implementation method of this embodiment can be found in Figure 4 and Figure 5 The relevant description of the illustrated embodiment is not repeated here.
[0331] See also Figure 7 , Figure 7 is a flow chart of a model processing method provided in an embodiment of the present application, which can be executed by a second communication device, such as Figure 7 As shown, the following steps are included:
[0332] Step 701: The fourth communication device performs a second operation, wherein the second operation is at least one of the following:
[0333] Sending second registration request information to a third communication device, and receiving second response information from the third communication device, wherein the second registration request information is used to request registration of the first model, and the second response information includes at least one of the following: a model identifier of the first model, and an identifier of a data set associated with the first model;
[0334] Sending fourth information to a fourth communication device, where the fourth information includes at least one of the following:
[0335] Model file information of the first model;
[0336] a model identifier of the first model;
[0337] a version identifier of the first model;
[0338] a type identifier of the first model;
[0339] parameter information of the first model;
[0340] usable condition information corresponding to the first model;
[0341] capability information of the first model;
[0342] a model identifier of a model that matches the first model;
[0343] an identifier of a data set associated with a model that matches the first model;
[0344] The storage address of the first model.
[0345] Optionally, the second registration request information includes at least one of the following:
[0346] Model file information of the first model;
[0347] Model description information of the first model.
[0348] Optionally, the model description information of the first model includes at least one of the following:
[0349] a version identifier of the first model;
[0350] a type identifier of the first model;
[0351] parameter information of the first model;
[0352] usable condition information corresponding to the first model;
[0353] capability information of the first model;
[0354] Producer information of the first model;
[0355] consumer information of the first model;
[0356] a model identifier of a model that matches the first model;
[0357] The identifier of the data set associated with the model that matches the first model.
[0358] Optionally, before the fourth communication device performs the second operation, the method further includes:
[0359] The fourth communication device receives first exchange data, where the first exchange data includes a second data set, and the second data set is a data set related to second model training;
[0360] The fourth communication device trains a first model according to the second data set.
[0361] Optionally, the first exchange data further includes at least one of the following:
[0362] model information of the second model;
[0363] The usable condition information corresponding to the second model.
[0364] Optionally, before the fourth communication device performs the second operation, the method further includes:
[0365] The fourth communication device trains the first model according to the fourth data set;
[0366] The fourth communication device determines second exchange data, where the second exchange data includes a third data set, and the third data set is a data set related to the target model training;
[0367] The fourth communication device sends the second exchange data.
[0368] Optionally, the second exchange data further includes at least one of the following:
[0369] Model information of the target model;
[0370] The usable condition information corresponding to the target model.
[0371] Optionally, the usable condition information includes at least one of the following: usable time information, usable location information.
[0372] It should be noted that the implementation method of this embodiment can be seen in FIG. Figures 4 to 6 The relevant description of the illustrated embodiment is not repeated here.
[0373] The following examples illustrate the embodiments of the present application:
[0374] Example 1: The server first trains the model and identifies different models by model ID.
[0375] like Figure 8 As shown, the model processing method provided in the embodiment of the present application includes the following steps:
[0376] Step 1. The server collects data.
[0377] The server may be a UE vendor server (i.e., a UE vendor server), a chip vendor server, or other third-party servers. Specifically, the server may collect UE data through DCAF, and the collected data may include historical air interface measurement data, service data, quality of experience (QOE) or user experience data, etc.
[0378] Step 2a. The server performs training based on the collected data to obtain a model (ie, the target model mentioned above).
[0379] Exemplarily, the UE side can jointly train the CSI encoder (ie, Encoder model) and the CSI decoder (ie, Decoder model) for the UE side through the server.
[0380] Step 2b. The server generates exchange data (ie, the second exchange data mentioned above).
[0381] Exemplarily, after the server has trained one side model (i.e., the Encoder model or the Decoder model), it can generate a large number of samples for model training (i.e., the third data set mentioned above) and send them to the network side device for the network side device to train the other side model (i.e., the Decoder model or the Encoder model). For example, the data set may include input data and / or label data for Decoder model training, wherein the input data for Decoder model training may be the output data of the Encoder model, and the label data for Decoder model training may be the input data of the Encoder model.
[0382] It is understandable that the third data set may include multiple sets of input data and / or label data corresponding to each set of input data. In addition, since the Encoder model and the Decoder model are symmetrical, the method for generating a large number of samples for training the Encoder model can be the same as the method for generating a large number of samples for training the Decoder model.
[0383] Step 3. The server registers the trained model.
[0384] In this step, the server can register the trained model (i.e., the target model mentioned above) to the storage node (for example, core network elements such as NEF, NRF or ADRF), and receive the identification information assigned by the storage node (i.e., model ID and / or version ID). For example, for the above-mentioned target model, the model ID assigned by the storage node is model ID1, and the version ID assigned by the storage node is version ID1.
[0385] Specifically, the server can send relevant information of the trained model to the storage node. The relevant information of the above model may include at least one of the following: model file information, model description information. It should be noted that the model file information and model description information in this step can refer to the relevant description of the model file information and model description information of the target model in the above embodiment, which will not be repeated here. The storage node assigns identification information such as model ID and / or version ID to the model, and sends the above assigned identification information to the server.
[0386] Step 4. The server sends the exchange data to the network side device so that the network side device can perform corresponding model training.
[0387] Specifically, the server sends the exchange data generated in step 2b (i.e., the second exchange data described above) to the network side device, wherein the exchange data in this step can refer to the relevant description of the second exchange data in the aforementioned embodiment, which will not be repeated here. It should be noted that when the server sends the exchange data to the network side device, it will carry the model ID and / or the version ID of the model.
[0388] The network side device in this example may be an access network device (eg, a base station), an access network gateway (eg, a base station gateway), a core network device or an OAM device, etc.
[0389] Step 5. The network side device trains the model for the network side device (ie, the second model mentioned above) based on the exchanged data.
[0390] The network side device trains a model (e.g., a decoder model) for the network side device based on the exchange data received in step 4. A specific training step may be to use the input data in the data set included in the exchange data as input data for training, and to use the label data in the exchange data set as label data for training.
[0391] Step 6. The network-side device registers the trained model to the storage node.
[0392] In this step, the network side device can register the model trained in step 5 (i.e., the second model mentioned above) to the storage node, and receive the model ID, version ID and other information assigned by the storage node. For example, for the second model mentioned above, the model ID assigned by the storage node is model ID2, and the version ID assigned by the storage node is version ID2.
[0393] Specifically, the network side device can send relevant information of the trained model (i.e., the above-mentioned second model) to the storage node. Among them, the relevant information of the above-mentioned model may include at least one of the following: model file information, model description information. It should be noted that, in addition to the information of the model trained by the network side device, the above-mentioned model description information may also include information of models that match or are associated with the model trained by the network side device (for example, model ID1, version ID1, etc. of the target model), and the information of the associated model can be used to inform the storage node that the model is associated with another model. Among them, the model file information and model description information in this step can refer to the relevant description of the model file information and model description information of the second model in the aforementioned embodiment, which will not be repeated here.
[0394] After step 6, this example can also include the following optional steps:
[0395] Step 7a. The server sends the trained model (ie, the target model mentioned above) to the UE.
[0396] In this step, the server may send the model to the UE via an OTT method, or the server may send the model to the UE via a control plane path or a user plane path of a communication network.
[0397] It should be noted that, in step 7a, the server may send model ID1 to the UE. Optionally, the server may also send version ID1 to the UE.
[0398] Step 7b. The storage node sends the model trained by the server (ie, the target model mentioned above) to the UE.
[0399] It should be noted that, in step 7b, the storage node may send the model ID 1 to the UE. Optionally, the storage node may also send the version ID 1 to the UE.
[0400] In this step, the UE may send a request message to the storage node to request a model that can complete a certain function. The storage node may send fourth information to the UE in response to the request message, wherein the fourth information may refer to the relevant description of the above embodiment and will not be described in detail here.
[0401] Exemplarily, for the above step 7a or step 7b, the UE may periodically request whether there is a new model file from the server or storage node; when the server or storage node has a new model file, the UE is notified, and the UE obtains the new model file from the server or storage node according to the notification.
[0402] It can be understood that only one of the above steps 7a and 7b can be performed, or both can be performed.
[0403] Step 8. The UE sends a model capability indication to the network side device.
[0404] The above-mentioned model capability indication may include information such as the model ID and / or version ID of the model requested to be used by the UE or the model possessed by the UE, for example, model ID1 and / or version ID1.
[0405] In an optional implementation, when the UE is ready to use a certain model (ie, the target model mentioned above), it may inform the network-side device of the model information (eg, model ID and version ID, etc.) to be used.
[0406] In an optional implementation, when the UE is ready to obtain one or more models (i.e., the target models mentioned above), it can inform the network side device of its model information (e.g., model ID and version ID, etc.). That is, the UE sends the model information of one or more models to the network side device. Subsequently, the network side device selects to activate a target model in the UE and a corresponding second model in the network side device according to the actual network environment in which the UE is located.
[0407] Optionally, the information sent by the UE to the network side device (i.e., the first information mentioned above) may also carry first indication information, used to indicate that the model will be used, or to ask the network side device whether the model can be used, i.e., whether the network side device has a matching model, etc.
[0408] Step 9a. The network side device determines whether there is a model matching the model information received in step 8, and performs association alignment of the models.
[0409] Exemplarily, the network side device may match the model information (e.g., the model ID and version ID of the target model) received in step 8 with the model information (e.g., the model ID) corresponding to the data set received in step 4 to determine whether the model information has been received. If it is determined that the model information has been received, it can be determined whether a model for the network side device is trained based on the exchange data corresponding to the model information.
[0410] Step 9b. The network-side device informs the UE to activate the target model.
[0411] In an optional implementation, if the network side device obtains model information (e.g., model ID and version ID, etc.) of multiple models from the UE in step 8, then when the target model (corresponding to model ID1 and version ID1) and the associated second model (corresponding to model ID2 and version ID2) can be activated according to the current network environment, the UE can be informed that the target model can be used or activated. For example, the network side device can send the following information to the UE: indication information of the model that can be used, model ID1, version ID1, etc.
[0412] In another optional implementation, if the network side device obtains a model information (e.g., model ID and version ID of the target model) to be used from the UE, the current network environment determines whether the target model and the corresponding second model can be activated. If yes, the UE is notified to activate the target model;
[0413] It should be noted that the above step 9a can be performed before the above step 9b or after the above step 9b, and this example does not limit this.
[0414] Step 10. The network side device runs the model corresponding to model ID2, or the network side device runs the model corresponding to model ID2 and version ID2.
[0415] It should be noted that before step 10, it is also necessary to map or associate the model ID2 corresponding to the model ID1, or to map or associate the model ID2 and version ID2 corresponding to the model ID1 and version ID1. The mapping or association relationship may exist locally on the network side device, or be obtained from the storage node. In the latter case, the network side device also needs to obtain the model ID2 corresponding to the model ID1 from the storage node, or to obtain the model ID2 and version ID2 corresponding to the model ID1 and version ID1 from the storage node.
[0416] Step 11. Inference data exchange.
[0417] Step 12: The UE runs the model corresponding to model ID1, or the UE runs the model corresponding to model ID1 and version ID1.
[0418] For the above steps 10 to 12, different data are exchanged between the UE and the network side device for different tasks, thereby completing actions such as model reasoning. For example, for the CSI task, the UE obtains the encoding result of the information to be transmitted (such as the estimated channel information) through the encoding model (i.e., the model on the UE side), and sends the encoding result to the network side device. The network side device inputs the information transmitted by the UE through the decoding model (i.e., the model of the network side device) to obtain the recovered channel information, etc.
[0419] Example 2: The server first trains the model and associates different models through the dataset ID.
[0420] like Figure 9a As shown, the model processing method provided in the embodiment of the present application includes the following steps:
[0421] Step 1. The server collects data.
[0422] Step 2a. The server performs training based on the collected data to obtain a model (ie, the target model mentioned above).
[0423] Exemplarily, the UE side can jointly train the CSI encoder (ie, Encoder model) and the CSI decoder (ie, Decoder model) for the UE side through the server.
[0424] Step 2b. The server generates exchange data (ie, the second exchange data mentioned above).
[0425] Step 3. The server registers the trained model.
[0426] In this step, the server can register the trained model (i.e., the target model mentioned above) to a storage node (e.g., a core network element such as NEF, NRF, or ADRF), and receive identification information assigned by the storage node to the target model, such as a data set ID and / or model type 1.
[0427] Step 4. The server sends the exchange data to the network side device so that the network side device can perform corresponding model training.
[0428] Step 5. The network side device trains the model for the network side device (ie, the second model mentioned above) based on the exchanged data.
[0429] Step 6. The network-side device registers the trained model to the storage node.
[0430] In this step, the network side device can register the model trained in step 5 (i.e., the second model mentioned above) to the storage node, and receive identification information assigned by the storage node to the second model, such as a data set ID and / or model type 2.
[0431] Specifically, the network side device can send relevant information of the trained model (i.e., the above-mentioned second model) to the storage node. The relevant information of the above-mentioned model may include at least one of the following: model file information, model description information. It should be noted that, in addition to the information of the model trained by the network side device, the above-mentioned model description information may also include information of a model that matches or is associated with the model trained by the network side device (for example, a data set ID associated with the associated model, a model type, etc.), and the information of the associated model can be used to inform the storage node that the model is associated with another model.
[0432] After step 6, this example can also include the following optional steps:
[0433] Step 7a. The server sends the trained model (ie, the target model mentioned above) to the UE.
[0434] Step 7b. The storage node sends the model trained by the server (ie, the target model mentioned above) to the UE.
[0435] Step 8. The UE sends a model capability indication to the network side device.
[0436] The above-mentioned model capability indication may include information such as the model requested by the UE or the data set ID and / or model type associated with the model possessed, for example, the data set ID and / or model type 1.
[0437] Step 9a. The network side device determines whether there is a model that matches the model information received in step 8, and performs association alignment of the models.
[0438] Exemplarily, the network side device can compare the data set ID received in step 8 with the data set ID received in step 4. If the two are the same, it can be determined whether a model for the network side device has been trained based on the data set corresponding to the data set ID. If so, it is determined that there is a model that matches the model requested by the UE.
[0439] Step 9b. The network-side device informs the UE to activate the target model.
[0440] In an optional implementation, if the network side device obtains model information (e.g., data set ID and model type, etc.) of multiple models from the UE in step 8, then when the target model (corresponding to the data set ID and model type 1) and the associated second model (corresponding to the data set ID and model type 2) can be activated according to the current network environment, the UE can be informed that the target model can be used or activated. For example, the network side device can send the following information to the UE: indication information of the model that can be used, data set ID, model type 1, etc.
[0441] In another optional implementation, if the network side device obtains a model information (e.g., data set ID and model type of the target model) to be used from the UE, the current network environment determines whether the target model and the corresponding second model can be activated. If yes, the UE is notified to activate the target model;
[0442] It should be noted that the above step 9a can be performed before the above step 9b or after the above step 9b, and this example does not limit this.
[0443] Step 10. The network-side device runs the model corresponding to the dataset ID.
[0444] It can be understood that the model run by the network-side device is a model corresponding to the data set ID and the model type is model type 2.
[0445] It should be noted that before step 10, it is also necessary to map or associate the dataset ID and the dataset ID corresponding to model type 1 and model type 2. The mapping or association relationship may exist locally on the network side device or be obtained from a storage node. In the latter case, the network side device also needs to obtain the dataset ID and the dataset ID corresponding to model type 1 and model type 2 from the storage node.
[0446] Step 11. Inference data exchange.
[0447] Step 12. UE runs the model corresponding to the dataset ID.
[0448] It can be understood that the model run by the UE is a model corresponding to the data set ID and the model type is model type 1.
[0449] It should be noted that steps 1 to 12 in this example can refer to the relevant descriptions of steps 1 to 12 in example 1, and will not be repeated here. The difference between example 2 and example 1 is that in the process of model registration, data set transfer, model distribution, and model activation, the data set ID and / or model type are not used, but the model ID and / or version ID.
[0450] Example 3: The server first trains the model and determines the dataset ID.
[0451] like Figure 9b As shown, the model processing method provided in the embodiment of the present application includes the following steps:
[0452] Step 1. The server collects data.
[0453] Step 2a. The server performs training based on the collected data to obtain a model (ie, the target model mentioned above).
[0454] Exemplarily, the UE side can jointly train the CSI encoder (ie, Encoder model) and the CSI decoder (ie, Decoder model) for the UE side through the server.
[0455] Step 2b. The server generates exchange data (ie, the second exchange data mentioned above).
[0456] Step 3. The server registers the trained model.
[0457] In this step, the server can register the trained model (i.e., the target model mentioned above) to a storage node (for example, a core network element such as NEF, NRF, or ADRF), wherein the information sent by the server to the storage node may include model file information and / or data set ID, and receive identification information assigned by the storage node to the target model, for example, model ID1, version ID1, model type 1, and at least one of a storage success indication.
[0458] Step 4. The server sends the exchange data to the network side device so that the network side device can perform corresponding model training.
[0459] The exchange data may include at least one of a data set, a data set ID associated with the target model, a model type of the target model, and a valid condition corresponding to the target model.
[0460] Step 5. The network side device trains the model for the network side device (ie, the second model mentioned above) based on the exchanged data.
[0461] Step 6. The network-side device registers the trained model to the storage node.
[0462] In this step, the network side device can register the model trained in step 5 (i.e., the second model mentioned above) to the storage node, wherein the information sent by the network side device to the storage node may include at least one of the model file information, the data set ID associated with the target model, the model ID of the target model (i.e., model ID1) and the version identifier of the target model (i.e., version ID1), and receive the identification information assigned by the storage node to the target model, for example, at least one of model ID2, version ID2, model type 2, and a storage success indication.
[0463] It should be noted that the storage node can associate or map the model identifiers of models with the same associated data set ID and store the resulting association relationship or mapping relationship, or can associate or map the model identifier and version identifier of models with the same associated data set ID and store the resulting association relationship or mapping relationship.
[0464] After step 6, this example can also include the following optional steps:
[0465] Step 7a. The server sends the trained model (ie, the target model mentioned above) to the UE.
[0466] It should be noted that, in step 7a, the server may send model ID1 to the UE. Optionally, the server may also send version ID1 to the UE.
[0467] Step 7b. The storage node sends the model trained by the server (ie, the target model mentioned above) to the UE.
[0468] It should be noted that, in step 7b, the storage node may send the model ID 1 to the UE. Optionally, the storage node may also send the version ID 1 to the UE.
[0469] Step 8. The UE sends a model capability indication to the network side device.
[0470] The above-mentioned model capability indication may include a model requested by the UE to be used or a model ID associated with the model possessed, for example, model ID1. Optionally, the above-mentioned model capability indication may also include a model version, for example, version ID1.
[0471] Step 9a. The network side device determines whether there is a model that matches the model information received in step 8, and performs association alignment of the models.
[0472] Exemplarily, if the network side device stores an association relationship or mapping relationship between model IDs, model alignment can be performed based on the mapping relationship or association relationship. For example, model ID2 can be found based on model ID1 and the above association relationship or mapping relationship. If the network side device does not store an association relationship or mapping relationship between model IDs, the above association relationship or mapping relationship can be obtained from the storage node, and model alignment can be performed based on the mapping relationship or association relationship.
[0473] The network side device can compare the data set ID received in step 8 with the data set ID received in step 4. If the two are the same, it can be determined whether a model for the network side device has been trained based on the data set corresponding to the data set ID. If so, it is determined that there is a model that matches the model requested by the UE.
[0474] Step 9b. The network-side device informs the UE to activate the target model.
[0475] In an optional implementation, if the network side device obtains model information (e.g., model ID and version ID, etc.) of multiple models from the UE in step 8, then when the target model (corresponding to model ID1 and version ID1) and the associated second model (corresponding to model ID2 and version ID2) can be activated according to the current network environment, the UE can be informed that the target model can be used or activated. For example, the network side device can send the following information to the UE: indication information of the model that can be used, model ID1, version ID1, etc.
[0476] In another optional implementation, if the network side device obtains a model information (e.g., model ID and version ID of the target model) to be used from the UE, the current network environment determines whether the target model and the corresponding second model can be activated. If yes, the UE is notified to activate the target model;
[0477] It should be noted that the above step 9a can be performed before the above step 9b or after the above step 9b, and this example does not limit this.
[0478] Step 10. The network side device runs the model corresponding to model ID2, or the network side device runs the model corresponding to model ID2 and version ID2.
[0479] Step 11. Inference data exchange.
[0480] Step 12: The UE runs the model corresponding to model ID1, or the UE runs the model corresponding to model ID1 and version ID1.
[0481] It can be understood that the model run by the UE is a model corresponding to the data set ID and the model type is model type 1.
[0482] It should be noted that, in this example, the above model type can be a new information element, or can be a multiplexed version of other information elements, such as a model ID, or a model ID and a version ID.
[0483] Example 4: The network-side device first trains the model and associates different models through the model ID.
[0484] like Fig.10 As shown, the model processing method provided in the embodiment of the present application includes the following steps:
[0485] Step 1a. The network-side device performs model training based on the collected data.
[0486] Step 1b. The network side device generates exchange data (ie, the first exchange data mentioned above).
[0487] Step 2. The network-side device registers the trained model (ie, the second model mentioned above) into the storage node.
[0488] Step 3. The network side device sends the exchange data to the server so that the server can perform corresponding model training.
[0489] Step 4. The server trains a model for the UE (ie, the target model mentioned above) based on the exchanged data.
[0490] The server trains a model (eg, Encoder model) for the UE based on the exchange data received in step 3 .
[0491] Step 5. The server registers the trained model (i.e., the target model mentioned above) into the storage node.
[0492] In this step, the storage node may assign a model ID and a version ID to the model trained by the server, for example, model ID1 and version ID1.
[0493] After step 5, this example can also include the following optional steps:
[0494] Step 6a. The server sends the trained model (ie, the target model mentioned above) to the UE.
[0495] In this step, the server may send the model to the UE via an OTT method, or the server may send the model to the UE via a control plane path or a user plane path of a communication network.
[0496] Step 6b. The storage node sends the model trained by the server (ie, the target model mentioned above) to the UE.
[0497] It can be understood that only one of the above steps 6a and 6b can be performed, or both can be performed.
[0498] Step 7. The UE sends a model capability indication to the network side device.
[0499] The model capability indication may include information such as the model ID and version ID of the model requested for use, for example, model ID2 and version ID2. Optionally, the model capability indication may also include model information of a model associated with the model requested for use, for example, model ID1 and version ID1.
[0500] Step 8a. The network side device determines whether there is a model that matches the model information received in step 7, and performs association alignment of the models.
[0501] Step 8b. The network-side device informs the UE to activate the target model.
[0502] Step 9. The network side device runs the model corresponding to model ID1, or the network side device runs the model corresponding to model ID1 and version ID1.
[0503] Step 10. Inference data exchange.
[0504] Step 11. The UE runs the model corresponding to model ID2 and version ID2, or the UE runs the model corresponding to model ID2.
[0505] It should be noted that steps 1a to 11 of this example can be referred to in the relevant description of steps 1 to 12 of example 1, and will not be repeated here. The difference between example 3 and example 1 is that in example 3, the network side device first performs model training, generates exchange data and sends it to the server. In addition, since the network side device informs the UE side of the exchange data and model ID, the network side device does not know the model ID trained by the server using the exchange data, and cannot perform model association alignment. Therefore, when the server registers the trained model to the storage node and sends the model, it is also necessary to carry the model ID, version ID and valid condition corresponding to the exchange data, etc., to inform the storage node that the model trained by the server is associated with the model trained by the network side device. It should also be noted that in step 7 of this example, when the UE requests the network side device to activate the model, it is necessary to carry the model ID, version ID, etc. corresponding to the exchange data, so as to help the network side device to associate the model. In step 8a, when the network side device associates the model, it can judge whether the model association can be performed, whether the model activation can be performed, etc. by whether there is model information corresponding to the exchange data.
[0506] It should be noted that the model processing method provided in the embodiment of the present application can be executed by a model processing device, or a control module in the model processing device for executing the model processing method. In the embodiment of the present application, the model processing device provided in the embodiment of the present application is described by taking the execution of the model processing method by the model processing device as an example.
[0507] See also Fig.11 , Fig.11 is a structural diagram of a model processing device provided in an embodiment of the present application, such as Fig.11 As shown, the model processing device 1100 includes:
[0508] A first receiving module 1101 is configured to receive first information sent by a second communication device, where the first information includes model information of a first model;
[0509] The first sending module 1102 is used to send second information to the second communication device, where the second information is used to instruct the second communication device to use a target model, and the first model includes the target model.
[0510] Optionally, the first information further includes at least one of the following:
[0511] capability information of the first model;
[0512] usable condition information corresponding to the first model;
[0513] First indication information, where the first indication information is used to indicate that the second communication device requests to use the first model.
[0514] Optionally, the model information of the first model includes at least one of the following: a model identifier of the first model, a version identifier of the first model, a data set identifier associated with the first model, and a type identifier of the first model.
[0515] Optionally, the second information includes at least one of the following: second indication information, model information of the target model, model information of the second model;
[0516] The second indication information is used to indicate that the second communication device is allowed to use the target model, and the second model is a model that matches the target model.
[0517] Optionally, the device further comprises:
[0518] The first determination module is used to determine a second model matching the target model before sending the second information to the second communication device.
[0519] Optionally, the model information of the second model includes at least one of the following: a model identifier of the second model, a version identifier of the second model, a data set identifier associated with the second model, and a type identifier of the second model.
[0520] Optionally, the first sending module is specifically used for:
[0521] In a case where it is determined that there is a second model matching the target model, second information is sent to the second communication device.
[0522] Optionally, the first sending module is specifically used for:
[0523] When it is determined that there is a second model matching the target model and the second communication device meets the usable condition corresponding to the target model, second information is sent to the second communication device.
[0524] Optionally, the first sending module is further used for:
[0525] When it is determined that there is no second model matching the first model or the second communication device does not meet the usable condition corresponding to the first model, third information is sent to the second communication device, where the third information is used to indicate that the second communication device is denied to use the first model.
[0526] Optionally, the third information includes at least one of the following: third indication information, model information of the first model, and fourth indication information;
[0527] The third indication information is used to indicate that the second communication device is denied from using the first model, and the fourth indication information is used to indicate a reason value for denying the second communication device from using the first model.
[0528] Optionally, the first sending module is further configured to send first registration request information to a third communication device before the first communication device receives the first information sent by the second communication device, where the first registration request information is used to request registration of the second model;
[0529] The first receiving module is further used to receive first response information sent by the third communication device, where the first response information includes a model identifier of the second model or a data set identifier associated with the second model.
[0530] Optionally, the first registration request information includes at least one of the following:
[0531] Model file information of the second model;
[0532] Model description information of the second model.
[0533] Optionally, the model description information of the second model includes at least one of the following:
[0534] a version identifier of the second model;
[0535] a type identifier of the second model;
[0536] parameter information of the second model;
[0537] usable condition information corresponding to the second model;
[0538] capability information of the second model;
[0539] Producer information of the second model;
[0540] consumer information of the second model;
[0541] a model identifier of a model that matches the second model;
[0542] The identifier of the dataset associated with the model that matches the second model.
[0543] Optionally, the device further comprises:
[0544] A first training module, used for training a second model according to a first data set before receiving the first information sent by the second communication device;
[0545] A second determination module, used to determine a second data set, where the second data set is a data set related to the second model training;
[0546] The first sending module is further used to send first exchange data, where the first exchange data includes the second data set, and the second data set is used to train the target model.
[0547] Optionally, the first exchange data further includes at least one of the following:
[0548] model information of the second model;
[0549] The usable condition information corresponding to the second model.
[0550] Optionally, the first receiving module is further configured to receive second exchange data before receiving the first information sent by the second communication device, where the second exchange data includes a third data set, and the third data set is a data set related to the target model training;
[0551] The device also includes a second training module, used for training a second model according to the third data set.
[0552] Optionally, the second exchange data further includes at least one of the following:
[0553] Model information of the target model;
[0554] The usable condition information corresponding to the target model.
[0555] Optionally, the usable condition information includes at least one of the following: usable time information, usable location information.
[0556] Optionally, the second communication device is a terminal.
[0557] The model processing device 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 network side device, or it can be other devices other than the network side device. Exemplarily, the network side device can include but is not limited to the types of network side devices 12 listed above, and other devices can be servers, network attached storage (Network Attached Storage, NAS), etc., which are not specifically limited in the embodiment of the present application.
[0558] The model processing device provided in the embodiment of the present application can achieve Figure 4 The various processes implemented by the method embodiment and achieving the same technical effect are not described here to avoid repetition.
[0559] See also Fig.12 , Fig.12 is a structural diagram of a model processing device provided in an embodiment of the present application, such as Fig.12 As shown, the model processing device 1200 includes:
[0560] The second sending module 1201 is used to send first information to the first communication device, where the first information is used to indicate that the second communication device requests to use the first model;
[0561] The second receiving module 1202 is used to receive second information or third information sent by the first communication device, the second information is used to instruct the second communication device to use the target model, the first model includes the target model, and the third information is used to indicate that the second communication device is denied to use the first model.
[0562] Optionally, the first information further includes at least one of the following:
[0563] capability information of the first model;
[0564] usable condition information corresponding to the first model;
[0565] First indication information, where the first indication information is used to indicate that the second communication device requests to use the first model.
[0566] Optionally, the model information of the first model includes at least one of the following: a model identifier of the first model, a version identifier of the first model, a data set identifier associated with the first model, and a type identifier of the first model.
[0567] Optionally, the second information includes at least one of the following: second indication information, model information of the target model, model information of the second model;
[0568] The second indication information is used to indicate that the second communication device is allowed to use the target model, and the second model is a model that matches the target model.
[0569] Optionally, the third information includes at least one of the following: third indication information, model information of the first model, and fourth indication information;
[0570] The third indication information is used to indicate that the second communication device is denied from using the first model, and the fourth indication information is used to indicate a reason value for denying the second communication device from using the first model.
[0571] Optionally, the second receiving module is further used for:
[0572] Before sending the first information to the first communication device, receiving fourth information;
[0573] The fourth information includes at least one of the following:
[0574] Model file information of the first model;
[0575] a model identifier of the first model;
[0576] a version identifier of the first model;
[0577] a type identifier of the first model;
[0578] parameter information of the first model;
[0579] usable condition information corresponding to the first model;
[0580] capability information of the first model;
[0581] a model identifier of a model that matches the first model;
[0582] an identifier of a data set associated with a model that matches the first model;
[0583] The storage address of the first model.
[0584] Optionally, the usable condition information includes at least one of the following: usable time information, usable location information.
[0585] Optionally, the second receiving module is specifically used for at least one of the following:
[0586] Receiving fourth information from a fourth communication device in an OTT manner on an upper layer of the network, the fourth communication device being a server corresponding to the second communication device;
[0587] receiving fourth information from a fourth communication device through the second network side device;
[0588] Fourth information is received from a third communication device, the third communication device having the first model registered therein.
[0589] Optionally, the first communication device is a first network side device, and the second communication device is a terminal.
[0590] The model processing device 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. Exemplarily, 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.
[0591] The model processing device provided in the embodiment of the present application can achieve Figure 5 The various processes implemented by the method embodiment and achieving the same technical effect are not described here to avoid repetition.
[0592] See also Fig.13 , Fig.13 is a structural diagram of a model processing device provided in an embodiment of the present application, such as Fig.13 As shown, the model processing device 1300 includes:
[0593] The first execution module 1301 is configured to execute a first operation, wherein the first operation is at least one of the following:
[0594] receiving first registration request information sent by a first communication device, and sending first response information to the first communication device, wherein the first registration request information is used to request registration of a second model, and the first response information includes at least one of the following: a model identifier of the second model, and an identifier of a data set associated with the second model;
[0595] Receive second registration request information sent by a fourth communication device, and send second response information to the fourth communication device, wherein the second registration request information is used to request registration of the first model, and the second response information includes at least one of the following: a model identifier of the first model, and a data set identifier associated with the first model.
[0596] Optionally, the first registration request information includes at least one of the following:
[0597] Model file information of the second model;
[0598] Model description information of the second model.
[0599] Optionally, the model description information of the second model includes at least one of the following:
[0600] a version identifier of the second model;
[0601] a type identifier of the second model;
[0602] parameter information of the second model;
[0603] usable condition information corresponding to the second model;
[0604] capability information of the second model;
[0605] Producer information of the second model;
[0606] consumer information of the second model;
[0607] a model identifier of a model that matches the second model;
[0608] The identifier of the dataset associated with the model that matches the second model.
[0609] Optionally, the second registration request information includes at least one of the following:
[0610] Model file information of the first model;
[0611] Model description information of the first model.
[0612] Optionally, the model description information of the first model includes at least one of the following:
[0613] a version identifier of the first model;
[0614] a type identifier of the first model;
[0615] parameter information of the first model;
[0616] usable condition information corresponding to the first model;
[0617] capability information of the first model;
[0618] Producer information of the first model;
[0619] consumer information of the first model;
[0620] a model identifier of a model that matches the first model;
[0621] The identifier of the data set associated with the model that matches the first model.
[0622] Optionally, the device further comprises:
[0623] A third sending module, used to send fourth information to the second communication device;
[0624] The fourth information includes at least one of the following:
[0625] Model file information of the first model;
[0626] a model identifier of the first model;
[0627] a version identifier of the first model;
[0628] a type identifier of the first model;
[0629] parameter information of the first model;
[0630] usable condition information corresponding to the first model;
[0631] capability information of the first model;
[0632] a model identifier of a model that matches the first model;
[0633] an identifier of a data set associated with a model that matches the first model;
[0634] The storage address of the first model.
[0635] Optionally, the usable condition information includes at least one of the following: usable time information, usable location information.
[0636] The model processing device 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 network side device, or it can be other devices other than the network side device. Exemplarily, the network side device can include but is not limited to the types of network side devices 12 listed above, and other devices can be servers, network attached storage (Network Attached Storage, NAS), etc., which are not specifically limited in the embodiment of the present application.
[0637] The model processing device provided in the embodiment of the present application can achieve Figure 6 The various processes implemented by the method embodiment and achieving the same technical effect are not described here to avoid repetition.
[0638] See also Fig.14 , Fig.14 is a structural diagram of a model processing device provided in an embodiment of the present application, such as Fig.14 As shown, the model processing device 1400 includes:
[0639] The second execution module 1401 is configured to execute a second operation, wherein the second operation is at least one of the following:
[0640] Sending second registration request information to a third communication device, and receiving second response information from the third communication device, wherein the second registration request information is used to request registration of the first model, and the second response information includes at least one of the following: a model identifier of the first model, and an identifier of a data set associated with the first model;
[0641] Sending fourth information to a fourth communication device, where the fourth information includes at least one of the following:
[0642] Model file information of the first model;
[0643] a model identifier of the first model;
[0644] a version identifier of the first model;
[0645] a type identifier of the first model;
[0646] parameter information of the first model;
[0647] usable condition information corresponding to the first model;
[0648] capability information of the first model;
[0649] a model identifier of a model that matches the first model;
[0650] an identifier of a data set associated with a model that matches the first model;
[0651] The storage address of the first model.
[0652] Optionally, the second registration request information includes at least one of the following:
[0653] Model file information of the first model;
[0654] Model description information of the first model.
[0655] Optionally, the model description information of the first model includes at least one of the following:
[0656] a version identifier of the first model;
[0657] a type identifier of the first model;
[0658] parameter information of the first model;
[0659] usable condition information corresponding to the first model;
[0660] capability information of the first model;
[0661] Producer information of the first model;
[0662] consumer information of the first model;
[0663] a model identifier of a model that matches the first model;
[0664] The identifier of the data set associated with the model that matches the first model.
[0665] Optionally, the device further comprises:
[0666] A third receiving module is used to receive first exchange data before performing the second operation, where the first exchange data includes a second data set, and the second data set is a data set related to the second model training;
[0667] The third training module is used to train the first model according to the second data set.
[0668] Optionally, the first exchange data further includes at least one of the following:
[0669] model information of the second model;
[0670] The usable condition information corresponding to the second model.
[0671] Optionally, the device further comprises:
[0672] a fourth training module, used for training the first model according to a fourth data set before performing the second operation;
[0673] A third determination module is used to determine second exchange data, where the second exchange data includes a third data set, and the third data set is a data set related to the target model training;
[0674] The fourth sending module is used to send the second exchange data.
[0675] Optionally, the second exchange data further includes at least one of the following:
[0676] Model information of the target model;
[0677] The usable condition information corresponding to the target model.
[0678] Optionally, the usable condition information includes at least one of the following: usable time information, usable location information.
[0679] The model processing device 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 server, or it can be other devices other than a server. Exemplarily, the server can include but is not limited to the types of servers 13 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.
[0680] The model processing device provided in the embodiment of the present application can achieve Figure 7 The various processes implemented by the method embodiment and achieving the same technical effect are not described here to avoid repetition.
[0681] Optional, such as Fig.15 As shown, an embodiment of the present application also provides a communication device 1500, including a processor 1501 and a memory 1502, and the memory 1502 stores a program or instruction that can be executed on the processor 1501. When the program or instruction is executed by the processor 1501, the various steps of the above-mentioned model processing method embodiment are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.
[0682] The embodiment of the present application also provides a network side device, including a processor and a communication interface, wherein the communication interface is used to receive first information sent by a second communication device, the first information including model information of a first model; the processor is used to find a second model matching the target model; the communication interface is also used to send second information to the second communication device, the second information is used to indicate to the second communication device to use the target model, the first model including the target model; or, the communication interface is used to perform a first operation, the first operation is at least one of the following: receiving a first registration request message sent by a first communication device, and sending a first response message to the first communication device, the first registration request message is used to request registration of a second model, the first response message is at least one of the following: a model identifier of the second model, a data set identifier associated with the second model; receiving a second registration request message sent by a fourth communication device, and sending a second response message to the fourth communication device, the second registration request message is used to request registration of a first model, the second response message includes at least one of the following: a model identifier of the first model, a data set identifier associated with the first model. This network side device embodiment corresponds to the above network side device method embodiment, and each implementation process and implementation method of the above method embodiment can be applied to this network side device embodiment, and can achieve the same technical effect.
[0683] Specifically, the embodiment of the present application also provides a network side device. Fig.16 As shown, the network side device 1600 includes: an antenna 1601, a radio frequency device 1602, a baseband device 1603, a processor 1604 and a memory 1605. The antenna 1601 is connected to the radio frequency device 1602. In the uplink direction, the radio frequency device 1602 receives information through the antenna 1601 and sends the received information to the baseband device 1603 for processing. In the downlink direction, the baseband device 1603 processes the information to be sent and sends it to the radio frequency device 1602. The radio frequency device 1602 processes the received information and sends it out through the antenna 1601.
[0684] The method executed by the network-side device in the above embodiment may be implemented in the baseband device 1603, which includes a baseband processor.
[0685] The baseband device 1603 may include, for example, at least one baseband board on which a plurality of chips are arranged. Fig.16 As shown, one of the chips is, for example, a baseband processor, which is connected to the memory 1605 through a bus interface to call the program in the memory 1605 to execute the network device operations shown in the above method embodiment.
[0686] The network side device may further include a network interface 1606, which is, for example, a Common Public Radio Interface (CPRI).
[0687] Specifically, the network side device 1600 of the embodiment of the present application further includes: instructions or programs stored in the memory 1605 and executable on the processor 1604, and the processor 1604 calls the instructions or programs in the memory 1605 to execute Fig.11 or Fig.13 The methods executed by the modules shown achieve the same technical effects, and therefore will not be described here in detail to avoid repetition.
[0688] An embodiment of the present application also provides a terminal, including a processor and a communication interface, the communication interface is used to send first information to a first communication device, the first information includes model information of a first model; receive second information or third information sent by the first communication device, the second information is used to instruct the second communication device to use a target model, the first model includes the target model, and the third information is used to indicate that the second communication device is refused to use the first model. This terminal embodiment corresponds to the above-mentioned terminal side method embodiment, and each implementation process and implementation method of the above-mentioned method embodiment can be applied to this terminal embodiment, and can achieve the same technical effect. Specifically, Fig.17A schematic diagram of the hardware structure of a terminal for implementing an embodiment of the present application.
[0689] The terminal 1700 includes but is not limited to: a radio frequency unit 1701, a network module 1702, an audio output unit 1703, an input unit 1704, a sensor 1705, a display unit 1706, a user input unit 1707, an interface unit 1708, a memory 1709 and at least some of the components of the processor 1710.
[0690] Those skilled in the art will appreciate that the terminal 1700 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 1710 through a power management system, thereby implementing functions such as managing charging, discharging, and power consumption management through the power management system. Fig.17 The terminal structure shown in the figure does not constitute a limitation on the terminal. The terminal may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently, which will not be described in detail here.
[0691] It should be understood that in the embodiment of the present application, the input unit 1704 may include a graphics processing unit (GPU) 17041 and a microphone 17042, and the graphics processor 17041 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 1706 may include a display panel 17061, and the display panel 17061 may be configured in the form of a liquid crystal display, an organic light emitting diode, etc. The user input unit 1707 includes a touch panel 17071 and at least one of other input devices 17072. The touch panel 17071 is also called a touch screen. The touch panel 17071 may include two parts: a touch detection device and a touch controller. Other input devices 17072 may include, but are not limited to, a physical keyboard, function keys (such as volume control keys, switch keys, etc.), a trackball, a mouse, and a joystick, which will not be repeated here.
[0692] In the embodiment of the present application, after receiving downlink data from the network side device, the RF unit 1701 can transmit the data to the processor 1710 for processing; in addition, the RF unit 1701 can send uplink data to the network side device. Generally, the RF unit 1701 includes but is not limited to an antenna, an amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, etc.
[0693] The memory 1709 can be used to store software programs or instructions and various data. The memory 1709 may mainly include a first storage area for storing programs or instructions and a second storage area for storing data, wherein the first storage area may store an operating system, an application program or instruction required for at least one function (such as a sound playback function, an image playback function, etc.), etc. In addition, the memory 1709 may include a volatile memory or a non-volatile memory. Among them, 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 1709 in the embodiment of the present application includes but is not limited to these and any other suitable types of memory.
[0694] The processor 1710 may include one or more processing units; optionally, the processor 1710 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 1710.
[0695] Among them, the radio frequency unit 1701 is used to send first information to a first communication device, the first information includes model information of a first model; receive second information or third information sent by the first communication device, the second information is used to instruct the second communication device to use a target model, the first model includes the target model, and the third information is used to indicate that the second communication device is denied use of the first model.
[0696] In an embodiment of the present application, a first communication device receives first information sent by a second communication device, wherein the first information includes model information of a first model, and sends second information to the second communication device, wherein the second information is used to instruct the second communication device to use a target model, and the first model includes the target model. That is, a communication device first requests another communication device before using a model, and then uses the model after the other communication device confirms that the model is allowed to be used. This can reduce the occurrence of incompatibility of models used by different communication devices, thereby improving the task processing effect.
[0697] It can be understood that the implementation process of each implementation method mentioned in this embodiment can refer to the relevant description of the aforementioned method embodiment, and achieve the same or corresponding technical effect. To avoid repetition, it will not be repeated here.
[0698] An embodiment of the present application also provides a server, comprising a processor and a communication interface, wherein the communication interface is used to perform a second operation, the second operation comprising at least one of the following: sending a second registration request message to a third communication device, and receiving a second response message from the third communication device, the second registration request message being used to request registration of a first model, the second response message comprising at least one of the following: a model identifier of the first model, a data set identifier associated with the first model; sending fourth information to a fourth communication device, the fourth information comprising at least one of the following: model file information of the first model; a model identifier of the first model; a version identifier of the first model; a type identifier of the first model; parameter information of the first model; usable condition information corresponding to the first model; capability information of the first model; a model identifier of a model matching the first model; a data set identifier associated with a model matching the first model; and a storage address of the first model.
[0699] Specifically, the embodiment of the present application also provides a server. Fig.18As shown, the server 1800 includes: a processor 1801, a memory 1802, and a program or instruction stored in the memory 1802 and executable on the processor. The various components in the server 1800 are coupled together via a bus interface 1803. When the program or instruction is executed by the processor 1801, the following steps are implemented: executing a second operation, the second operation includes at least one of the following: sending a second registration request message to a third communication device, and receiving a second response message from the third communication device, the second registration request message is used to request registration of a first model, the second response message includes at least one of the following: a model identifier of the first model, a data set identifier associated with the first model; sending fourth information to a fourth communication device, the fourth information includes at least one of the following: model file information of the first model; a model identifier of the first model; a version identifier of the first model; a type identifier of the first model; parameter information of the first model; usable condition information corresponding to the first model; capability information of the first model; a model identifier of a model matching the first model; a data set identifier associated with a model matching the first model; and a storage address of the first model.
[0700] It should be understood that in the embodiment of the present invention, when the program or instruction is executed by the processor 1801, it can implement the various processes in the above-mentioned server-side model processing method embodiment. To avoid repetition, it will not be repeated here.
[0701] An embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the various processes of the above-mentioned model processing method embodiment are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.
[0702] 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. In some examples, the readable storage medium may be a non-transient readable storage medium.
[0703] An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned model processing method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.
[0704] It should be understood that 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.
[0705] The embodiment of the present application further provides a computer program / program product, which is stored in a storage medium. The computer program / program product is executed by at least one processor to implement the various processes of the above-mentioned model processing method embodiment and can achieve the same technical effect. To avoid repetition, it will not be repeated here.
[0706] The embodiment of the present application also provides a model processing system, including: a first communication device, a second communication device, a third communication device and a fourth communication device, wherein the first communication device is used to execute Figure 4 And each process of each of the above method embodiments, the second communication device is used to perform the following Figure 5 And each process of each of the above method embodiments, the third communication device is used to perform the following Figure 6 And each process of each of the above method embodiments, the fourth communication device is used to perform the following Figure 7 And the various processes of the above-mentioned method embodiments can achieve the same technical effect. To avoid repetition, they will not be described here.
[0707] It should be noted that, in this article, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises one..." does not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the method and device in the embodiment of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted or combined. In addition, the features described with reference to certain examples may be combined in other examples.
[0708] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of a computer software product plus a necessary general hardware platform, and of course, can also be implemented by hardware. The computer software product is stored in a storage medium (such as ROM, RAM, disk, CD, etc.), including several instructions to enable a terminal or a network-side device to execute the methods described in each embodiment of the present application.
[0709] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms of implementation methods without departing from the purpose of the present application and the scope of protection of the claims, and these implementation methods are all within the protection of the present application.
Claims
1. A model processing method, It is characterized in that include: The first communication device receives first information sent by the second communication device, where the first information includes model information of the first model; The first communication device sends second information to the second communication device, where the second information is used to instruct the second communication device to use a target model, and the first model includes the target model.
2. The method according to claim 1, It is characterized in that The first information also includes at least one of the following: capability information of the first model; usable condition information corresponding to the first model; First indication information, where the first indication information is used to indicate that the second communication device requests to use the first model.
3. The method according to claim 1 or 2, It is characterized in that The model information of the first model includes at least one of the following: a model identifier of the first model, a version identifier of the first model, a data set identifier associated with the first model, and a type identifier of the first model.
4. The method according to any one of claims 1 to 3, It is characterized in that The second information includes at least one of the following: second indication information, model information of the target model, and model information of the second model; The second indication information is used to indicate that the second communication device is allowed to use the target model, and the second model is a model that matches the target model.
5. The method according to claim 4, It is characterized in that The model information of the second model includes at least one of the following: a model identifier of the second model, a version identifier of the second model, a data set identifier associated with the second model, and a type identifier of the second model.
6. The method according to any one of claims 1 to 5, It is characterized in that Before the first communication device sends the second information to the second communication device, the method further includes: The first communications device determines a second model that matches the target model.
7. The method according to any one of claims 1 to 6, It is characterized in that The second communication device sends second information, including: The first communication device sends second information to the second communication device when it is determined that there is a second model matching the target model and the second communication device meets the usable condition corresponding to the target model.
8. The method according to any one of claims 1 to 7, It is characterized in that The method further comprises: When it is determined that there is no model matching the first model or the second communication device does not meet the usable condition corresponding to the first model, the first communication device sends third information to the second communication device, where the third information is used to indicate that the second communication device is denied to use the first model.
9. The method according to claim 8, It is characterized in that The third information includes at least one of the following: third indication information, model information of the first model, and fourth indication information; The third indication information is used to indicate that the second communication device is denied from using the first model, and the fourth indication information is used to indicate a reason value for denying the second communication device from using the first model.
10. The method according to any one of claims 1 to 9, It is characterized in that Before the first communication device receives the first information sent by the second communication device, the method further includes: The first communication device sends first registration request information to the third communication device, where the first registration request information is used to request registration of the second model; The first communication device receives first response information sent by the third communication device, where the first response information includes a model identifier of the second model or a data set identifier associated with the second model.
11. The method according to claim 10, It is characterized in that The first registration request information includes at least one of the following: Model file information of the second model; Model description information of the second model.
12. The method according to claim 11, It is characterized in that The model description information of the second model includes at least one of the following: a version identifier of the second model; a type identifier of the second model; parameter information of the second model; usable condition information corresponding to the second model; capability information of the second model; Producer information of the second model; consumer information of the second model; a model identifier of a model that matches the second model; The identifier of the dataset associated with the model that matches the second model.
13. The method according to any one of claims 1 to 12, It is characterized in that Before the first communication device receives the first information sent by the second communication device, the method further includes: The first communication device trains a second model based on the first data set; The first communication device determines a second data set; The first communication device sends first exchange data, where the first exchange data includes the second data set, and the second data set is used to train the target model.
14. The method according to claim 13, It is characterized in that The first exchange data further includes at least one of the following: model information of the second model; The usable condition information corresponding to the second model.
15. The method according to any one of claims 1 to 12, It is characterized in that Before the first communication device receives the first information sent by the second communication device, the method further includes: The first communication device receives second exchange data, where the second exchange data includes a third data set, and the third data set is a data set related to the target model training; The first communication device trains a second model according to the third data set.
16. The method according to claim 15, It is characterized in that The second exchange data further includes at least one of the following: Model information of the target model; The usable condition information corresponding to the target model.
17. The method according to claim 2, 12, 14 or 16, It is characterized in that The usable condition information includes at least one of the following: usable time information and usable location information.
18. The method according to any one of claims 1 to 17, It is characterized in that The first communication device is a first network side device, and the second communication device is a terminal.
19. A model processing method, It is characterized in that include: The second communication device sends first information to the first communication device, where the first information includes model information of the first model; The second communication device receives second information or third information sent by the first communication device, the second information is used to instruct the second communication device to use a target model, the first model includes the target model, and the third information is used to indicate that the second communication device is denied to use the first model.
20. The method according to claim 19, It is characterized in that The first information also includes at least one of the following: capability information of the first model; usable condition information corresponding to the first model; First indication information, where the first indication information is used to indicate that the second communication device requests to use the first model.
21. The method according to claim 19 or 20, It is characterized in that The model information of the first model includes at least one of the following: a model identifier of the first model, a version identifier of the first model, a data set identifier associated with the first model, and a type identifier of the first model.
22. The method according to any one of claims 19 to 21, It is characterized in that The second information includes at least one of the following: second indication information, model information of the target model, and model information of the second model; The second indication information is used to indicate that the second communication device is allowed to use the target model, and the second model is a model that matches the target model.
23. The method according to claim 22, It is characterized in that The model information of the second model includes at least one of the following: a model identifier of the second model, a version identifier of the second model, a data set identifier associated with the second model, and a type identifier of the second model.
24. The method according to any one of claims 19 to 21, It is characterized in that The third information includes at least one of the following: third indication information, model information of the first model, and fourth indication information; The third indication information is used to indicate that the second communication device is denied from using the first model, and the fourth indication information is used to indicate a reason value for denying the second communication device from using the first model.
25. The method according to any one of claims 19 to 24, It is characterized in that Before the second communication device sends the first information to the first communication device, the method further includes: The second communication device receives fourth information; The fourth information includes at least one of the following: Model file information of the first model; a model identifier of the first model; a version identifier of the first model; a type identifier of the first model; parameter information of the first model; usable condition information corresponding to the first model; capability information of the first model; a model identifier of a model that matches the first model; an identifier of a data set associated with a model that matches the first model; The storage address of the first model.
26. The method according to claim 25, It is characterized in that The second communication device receives fourth information, including at least one of the following: The second communication device receives fourth information from a fourth communication device in an OTT manner on an upper layer of a network, and the fourth communication device is a server corresponding to the second communication device; The second communication device receives fourth information from the fourth communication device through the second network side device; The second communication device receives fourth information from a third communication device, the third communication device having the first model registered therein.
27. The method according to claim 25, It is characterized in that The usable condition information includes at least one of the following: usable time information and usable location information.
28. The method according to any one of claims 19 to 27, It is characterized in that The first communication device is a first network side device, and the second communication device is a terminal.
29. A model processing method, It is characterized in that include: The third communication device performs a first operation, wherein the first operation is at least one of the following: receiving first registration request information sent by a first communication device, and sending first response information to the first communication device, wherein the first registration request information is used to request registration of a second model, and the first response information includes at least one of the following: a model identifier of the second model, and an identifier of a data set associated with the second model; Receive second registration request information sent by a fourth communication device, and send second response information to the fourth communication device, wherein the second registration request information is used to request registration of the first model, and the second response information includes at least one of the following: a model identifier of the first model, and a data set identifier associated with the first model.
30. The method according to claim 29, It is characterized in that The first registration request information includes at least one of the following: Model file information of the second model; Model description information of the second model.
31. The method according to claim 30, It is characterized in that The model description information of the second model includes at least one of the following: a version identifier of the second model; a type identifier of the second model; parameter information of the second model; usable condition information corresponding to the second model; capability information of the second model; Producer information of the second model; consumer information of the second model; a model identifier of a model that matches the second model; The identifier of the dataset associated with the model that matches the second model.
32. The method according to claim 29, It is characterized in that The second registration request information includes at least one of the following: Model file information of the first model; Model description information of the first model.
33. The method according to claim 32, It is characterized in that The model description information of the first model includes at least one of the following: a version identifier of the first model; a type identifier of the first model; parameter information of the first model; usable condition information corresponding to the first model; capability information of the first model; Producer information of the first model; consumer information of the first model; a model identifier of a model that matches the first model; The identifier of the data set associated with the model that matches the first model.
34. A method according to any one of claims 29 to 33, It is characterized in that The method further comprises: The third communication device sends fourth information to the second communication device; The fourth information includes at least one of the following: Model file information of the first model; a model identifier of the first model; a version identifier of the first model; a type identifier of the first model; parameter information of the first model; usable condition information corresponding to the first model; capability information of the first model; a model identifier of a model that matches the first model; an identifier of a data set associated with a model that matches the first model; The storage address of the first model.
35. The method according to claim 34, It is characterized in that The usable condition information includes at least one of the following: usable time information and usable location information.
36. A model processing method, It is characterized in that include: The fourth communication device performs a second operation, wherein the second operation is at least one of the following: Sending second registration request information to a third communication device, and receiving second response information from the third communication device, wherein the second registration request information is used to request registration of the first model, and the second response information includes at least one of the following: a model identifier of the first model, and an identifier of a data set associated with the first model; Sending fourth information to a fourth communication device, where the fourth information includes at least one of the following: Model file information of the first model; a model identifier of the first model; a version identifier of the first model; a type identifier of the first model; parameter information of the first model; usable condition information corresponding to the first model; capability information of the first model; a model identifier of a model that matches the first model; an identifier of a data set associated with a model that matches the first model; The storage address of the first model.
37. The method according to claim 36, It is characterized in that The second registration request information includes at least one of the following: Model file information of the first model; Model description information of the first model.
38. The method according to claim 37, It is characterized in that The model description information of the first model includes at least one of the following: a version identifier of the first model; a type identifier of the first model; parameter information of the first model; usable condition information corresponding to the first model; capability information of the first model; Producer information of the first model; consumer information of the first model; a model identifier of a model that matches the first model; The identifier of the data set associated with the model that matches the first model.
39. The method according to any one of claims 36 to 38, It is characterized in that Before the fourth communication device performs the second operation, the method further includes: The fourth communication device receives first exchange data, where the first exchange data includes a second data set, and the second data set is a data set related to second model training; The fourth communication device trains a target model according to the second data set, and the first model includes the target model.
40. The method according to claim 39, It is characterized in that The first exchange data further includes at least one of the following: model information of the second model; The usable condition information corresponding to the second model.
41. The method according to any one of claims 36 to 38, It is characterized in that Before the fourth communication device performs the second operation, the method further includes: The fourth communication device trains a target model according to a fourth data set, wherein the first model includes the target model; The fourth communication device determines second exchange data, where the second exchange data includes a third data set, and the third data set is a data set related to the target model training; The fourth communication device sends the second exchange data.
42. The method according to claim 41, It is characterized in that The second exchange data further includes at least one of the following: Model information of the target model; The usable condition information corresponding to the target model.
43. The method according to claim 40 or 42, It is characterized in that The usable condition information includes at least one of the following: usable time information and usable location information.
44. A model processing device, It is characterized in that include: A first receiving module, configured to receive first information sent by a second communication device, wherein the first information includes model information of a first model; The first sending module is used to send second information to the second communication device, where the second information is used to instruct the second communication device to use a target model, and the first model includes the target model.
45. A model processing device, It is characterized in that include: A second sending module, configured to send first information to the first communication device, wherein the first information is used to indicate that the second communication device requests to use the first model; The second receiving module is used to receive second information or third information sent by the first communication device, the second information is used to instruct the second communication device to use a target model, the first model includes the target model, and the third information is used to indicate that the second communication device is denied to use the first model.
46. A model processing device, It is characterized in that include: The first execution module is configured to execute a first operation, wherein the first operation is at least one of the following: receiving first registration request information sent by a first communication device, and sending first response information to the first communication device, wherein the first registration request information is used to request registration of a second model, and the first response information includes at least one of the following: a model identifier of the second model, and an identifier of a data set associated with the second model; Receive second registration request information sent by a fourth communication device, and send second response information to the fourth communication device, wherein the second registration request information is used to request registration of the first model, and the second response information includes at least one of the following: a model identifier of the first model, and a data set identifier associated with the first model.
47. A model processing device, It is characterized in that include: The second execution module is configured to execute a second operation, wherein the second operation is at least one of the following: Sending second registration request information to a third communication device, and receiving second response information from the third communication device, wherein the second registration request information is used to request registration of the first model, and the second response information includes at least one of the following: a model identifier of the first model, and an identifier of a data set associated with the first model; Sending fourth information to a fourth communication device, where the fourth information includes at least one of the following: Model file information of the first model; a model identifier of the first model; a version identifier of the first model; a type identifier of the first model; parameter information of the first model; usable condition information corresponding to the first model; capability information of the first model; a model identifier of a model that matches the first model; an identifier of a data set associated with a model that matches the first model; The storage address of the first model.
48. A communication device, It is characterized in that The method comprises a processor and a memory, wherein the memory stores a program or instruction that can be run on the processor, and when the program or instruction is executed by the processor, the method implements the steps of the model processing method as described in any one of claims 1 to 18, or implements the steps of the model processing method as described in any one of claims 19 to 28, or implements the steps of the model processing method as described in any one of claims 29 to 35, or implements the steps of the model processing method as described in any one of claims 36 to 43.
49. A readable storage medium, It is characterized in that The readable storage medium stores programs or instructions, and when the programs or instructions are executed by the processor, the steps of the model processing method as described in any one of claims 1 to 18 are implemented, or the steps of the model processing method as described in any one of claims 19 to 28 are implemented, or the steps of the model processing method as described in any one of claims 29 to 35 are implemented, or the steps of the model processing method as described in any one of claims 36 to 43 are implemented.
Citation Information
Cited By
Communication method, apparatus, medium, and computer program product
WO2026179763A1
