Communication method and device based on reference model

Abstract

The invention discloses a communication method and device based on a reference model, and belongs to the technical field of communication, and the method comprises the steps that a first device executes a first operation, and the first operation comprises at least one of the following items: sending first information related to the reference model to a second device; collecting indexes of the reference model; the reference model is a model used by the second equipment, or the reference model is used by the second equipment to determine to use.

Description

Technical Field

[0001] The present application belongs to the field of communication technology, and specifically relates to a communication method and device based on a reference model. Background Art

[0002] Artificial Intelligence (AI) has been widely used in various fields. Integrating AI into wireless communication networks and significantly improving technical indicators such as throughput, latency and user capacity are important tasks for future wireless communication networks.

[0003] At present, the AI ​​models used in communication equipment may have a variety of designs, and their performance, complexity, and size may vary greatly, making it difficult to manage AI models on the network side and increasing the implementation complexity of the communication system. Summary of the invention

[0004] The embodiments of the present application provide a communication method and device based on a reference model, which can solve the problem of high implementation complexity of the communication system.

[0005] In a first aspect, a reference model-based communication method is provided, comprising:

[0006] The first device performs a first operation, where the first operation includes at least one of the following:

[0007] Sending first information related to the reference model to the second device;

[0008] Collect metrics for the reference model;

[0009] The reference model is a model used by the second device, or the reference model is used by the second device to determine a model to use.

[0010] In a second aspect, a reference model-based communication method is provided, comprising:

[0011] The second device performs a third operation, where the third operation includes at least one of the following:

[0012] Receiving first information related to the reference model sent by the first device;

[0013] receiving auxiliary information sent by the third device, and generating second model information based on the auxiliary information;

[0014] generating second model information;

[0015] Sending first model information to the first device;

[0016] Collect metrics for the reference model;

[0017] The reference model is a model used by the second device, or the reference model is used by the second device to determine a model to use.

[0018] In a third aspect, a reference model-based communication device is provided, comprising:

[0019] The processing module is configured to perform a first operation, wherein the first operation includes at least one of the following:

[0020] Sending first information related to the reference model to the second device;

[0021] Collect metrics for the reference model;

[0022] The reference model is a model used by the second device, or the reference model is used by the second device to determine a model to use.

[0023] In a fourth aspect, a reference model-based communication device is provided, comprising:

[0024] The processing module is configured to perform a third operation, wherein the third operation includes at least one of the following:

[0025] Receiving first information related to the reference model sent by the first device;

[0026] receiving auxiliary information sent by the third device, and generating second model information based on the auxiliary information;

[0027] generating second model information;

[0028] Sending first model information to the first device;

[0029] Collect metrics for the reference model;

[0030] The reference model is a model used by the second device, or the reference model is used by the second device to determine a model to use.

[0031] In a fifth aspect, a first device is provided, comprising a processor and a memory, wherein the memory stores a program or instruction that can be executed on the processor, and when the program or instruction is executed by the processor, the steps of the method described in the first aspect are implemented.

[0032] In a sixth aspect, a first device is provided, including a processor and a communication interface, wherein the processor is configured to perform a first operation, and the first operation includes at least one of the following:

[0033] Sending first information related to the reference model to the second device;

[0034] Collect metrics for the reference model;

[0035] The reference model is a model used by the second device, or the reference model is used by the second device to determine a model to use.

[0036] In the seventh aspect, a second device is provided, which includes a processor and a memory, wherein the memory stores a program or instruction that can be executed on the processor, and when the program or instruction is executed by the processor, the steps of the method described in the second aspect are implemented.

[0037] In an eighth aspect, a second device is provided, including a processor and a communication interface, wherein the processor is configured to perform a third operation, and the third operation includes at least one of the following:

[0038] Receiving first information related to the reference model sent by the first device;

[0039] receiving auxiliary information sent by the third device, and generating second model information based on the auxiliary information;

[0040] generating second model information;

[0041] Sending first model information to the first device;

[0042] Collect metrics for the reference model;

[0043] The reference model is a model used by the second device, or the reference model is used by the second device to determine a model to use.

[0044] In a ninth 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 second aspect are implemented.

[0045] In the tenth aspect, a wireless communication system is provided, comprising: a first device and a second device, wherein the first device can be used to execute the steps of the method described in the first aspect, and the second device can be used to execute the steps of the method described in the second aspect.

[0046] In the eleventh 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 instruction to implement the method described in the first aspect, or to implement the method described in the second aspect.

[0047] In the twelfth aspect, a computer program / program product is provided, wherein the computer program / program product is stored in a storage medium, and the program / program product is executed by at least one processor to implement the steps of the reference model-based communication method as described in the first aspect or the second aspect.

[0048] In an embodiment of the present application, a first device performs a first operation, and the first operation includes at least one of the following: sending first information related to a reference model to a second device; collecting indicators of the reference model; the reference model is a model used by the second device, or the reference model is used for the model determined by the second device to use. In the above scheme, performance testing or communication is performed based on the reference model, so that the AI ​​models used by different communication devices have a certain consistency, which can make it easier for the network side to manage the AI ​​models of communication devices, thereby improving the overall performance of the communication network. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a schematic diagram of the architecture of a wireless communication system provided in an embodiment of the present application;

[0050] Figure 2 It is one of the flowcharts of the communication method based on the reference model provided in the embodiment of the present application;

[0051] Figure 3 This is the second flow chart of the reference model-based communication method provided in the embodiment of the present application;

[0052] Figure 4 It is one of the structural schematic diagrams of a communication device based on a reference model provided in an embodiment of the present application;

[0053] Figure 5 This is the second structural diagram of the communication device based on the reference model provided in the embodiment of the present application;

[0054] Figure 6 is a schematic diagram of the structure of a communication device provided in an embodiment of the present application;

[0055] Figure 7 is a schematic diagram of the structure of a network side device in an embodiment of the present application;

[0056] Figure 8 It is a schematic diagram of the structure of the terminal of an embodiment of the present application. DETAILED DESCRIPTION

[0057] 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.

[0058] 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.

[0059] 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.

[0060] 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 thGeneration, 6G) communication system.

[0061] 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 and a network side device 12. 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 an access network device or a core network device, wherein the access network device may also be referred to as a radio access network (Radio Access Network, RAN) device, a radio access network function or a radio access network unit. The access network device may include a base station, a wireless local area network (Wireless Local Area Network, WLAN) access point (Access Point, 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, 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.

[0062] 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.

[0063] First, the nouns and application scenarios involved in the embodiments of the present application are introduced:

[0064] 1. The AI ​​unit / AI model described in the embodiments of the present application may also be referred to as an AI unit, an AI model, a machine learning (ML) model, an ML unit, an AI structure, an AI function, an AI feature, a machine learning model, a neural network, a neural network function, a neural network function, etc., or the AI ​​unit / AI model may also refer to a processing unit capable of implementing specific algorithms, formulas, processing procedures, capabilities, etc. related to AI, or the AI ​​unit / AI model may be a processing method, algorithm, function, module or unit for a specific data set, or the AI ​​unit / AI model may be a processing method, algorithm, function, module or unit running on AI / ML related hardware such as a graphics processing unit (GPU), a neural network processing unit (NPU), a tensor processing unit (TPU), an application specific integrated circuit (ASIC), etc., and the present application does not specifically limit this. Optionally, the specific data set includes the input and / or output of the AI ​​unit / AI model.

[0065] Optionally, the identifier of the AI ​​unit / AI model may be an AI model identifier, an AI structure identifier, an AI algorithm identifier, or an identifier of a specific data set associated with the AI ​​unit / AI model, or an identifier of a specific scenario, environment, channel feature, or device related to the AI / ML, or an identifier of a function, feature, capability, or module related to the AI / ML, which is not specifically limited in the embodiments of the present application.

[0066] 2. There are many ways to implement AI models, such as neural networks, decision trees, support vector machines, Bayesian classifiers, etc. This application uses neural networks as an example for illustration, but does not limit the specific type of AI models.

[0067] Optionally, the parameters of the neural network are optimized using a gradient optimization algorithm. A gradient optimization algorithm is a type of algorithm that minimizes or maximizes an objective function (also called a loss function), and the objective function is often a mathematical combination of model parameters and data. For example, given data X and its corresponding label Y, a neural network model f(.) is constructed. With the model, the predicted output f(x) can be obtained based on the input x, and the difference between the predicted value and the true value (f(x)-Y) can be calculated. (f(x)-Y) is the loss function. The ultimate goal is to find a suitable W,b to minimize the value of the above loss function. The smaller the loss value, the closer the model is to the actual situation.

[0068] At present, the common optimization algorithms are basically based on the BP (error back propagation) algorithm. The basic idea of ​​the BP algorithm is that the learning process consists of two processes: the forward propagation of the signal and the back propagation of the error. During the forward propagation, the input sample is transmitted from the input layer, processed by each hidden layer layer by layer, and then transmitted to the output layer. If the actual output of the output layer does not match the expected output, it will enter the back propagation stage of the error. Error back propagation is to propagate the output error layer by layer through the hidden layer to the input layer in some form, and distribute the error to all units in each layer, so as to obtain the error signal of each layer unit, and this error signal is used as the basis for correcting the weights of each unit. This process of adjusting the weights of each layer of the signal forward propagation and error back propagation is repeated. The process of continuous adjustment of weights is the learning and training process of the network. This process continues until the error of the network output is reduced to an acceptable level, or until the pre-set number of learning times is reached.

[0069] Common optimization algorithms include gradient descent, stochastic gradient descent (SGD), mini-batch gradient descent, momentum method (Momentum), Nesterov (also known as stochastic gradient descent with momentum), adaptive gradient descent (Adagrad), root mean square error prop (RMSprop), adaptive momentum estimation (Adam), etc.

[0070] When the error is back-propagated, these optimization algorithms calculate the derivative / partial derivative of the current neuron based on the error / loss obtained by the loss function, add the influence of the learning rate, the previous gradient / derivative / partial derivative, etc., get the gradient, and pass the gradient to the previous layer.

[0071] 3. Mobile communication test

[0072] In the process of R&D, production and acceptance of mobile communication equipment, testing technology is an indispensable link throughout the process. The test objects include base station equipment, terminal equipment, chips and other communication equipment. The testing of mobile communication equipment can be divided into certification testing, R&D testing, production testing and other links according to its different life cycle and different test purposes. Among them, certification testing needs to be handed over to organizations with certification qualifications. The general test specifications are based on the test cases and minimum requirements formulated by the 3rd Generation Partnership Project (3GPP), Cellular Telecommunication Internet Association (CTIA), Global Certification Forum (GCF) and other organizations. The requirements for the environment, equipment and passing standards are relatively high, and the main focus is on consistency testing; R&D testing focuses more on whether certain problems occur in the product during the R&D process, and does not necessarily need to be verified according to the test cases of the specification standards; production testing is an automated test performed during the mass production of the product, and pays more attention to the efficiency of the test. Therefore, for wireless communication equipment, the test system often has high costs, complex test systems, strict test passing requirements and test cases, and is also one of the most important links before the product enters the market. Developing a set of standardized test procedures that meet the specifications is also one of the goals that many standard organizations are committed to. At present, 3GPP focuses on RF conduction testing, RF over-the-air (OTA) testing, and corresponding performance testing. The indicators concerned include RF indicators such as total radiated power (Total Radiated Power), equivalent isotropic sensitivity (EIS), error vector magnitude (Error Vector Magnitude, EVM), adjacent channel leakage ratio (Adjacent Channel Leakage Ratio, ACLR) and throughput (Throughput), etc.3GPP has also developed standards for baseband-related Radio Resource Management (RRM) and demodulation tests. For different use cases, 3GPP has also specified many RRM indicators that need to be tested (such as TS 38.133). For example, for positioning scenarios, the terminal needs to pass the test case under certain conditions and meet the specified minimum time requirements for Reference Signal Time Difference (RSTD), minimum time requirements for Rx-Tx time difference, and Reference Signal Received Power (RSRP) / RSRP (PathRSRP, RSRPP) minimum power requirements for a certain path in multipath conditions. For the use case of beam prediction, the RRM measurement requirements for L1-RSRP are specified. 3GPP also specifies the baseband demodulation performance requirements of the equipment (such as TS 38.101-4). For the reporting of Channel State Information (CSI), it specifies the reporting requirements of Channel Quality Indication (CQI), Precoding Matrix Indicator (PMI) and Rank Indicator (RI) under certain test conditions. In the entire communication process, the baseband test needs to meet the corresponding core requirements, such as interruption, delay and other requirements.

[0073] The method of the embodiment of the present application can be used in the test process, that is, for test equipment (TE) and device under test (DUT); the method can also be used in the formal communication process; in this case, TE can be an access network device (such as a base station) and DUT is a terminal; or TE is a terminal and DUT is an access network device.

[0074] The communication 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.

[0075] Please refer to Figure 2 , an embodiment of the present application provides a communication method, the execution subject of this embodiment is a first device, and the method includes:

[0076] Step 101: The first device performs a first operation, where the first operation includes at least one of the following:

[0077] Sending first information related to the reference model to the second device;

[0078] Collect metrics for the reference model;

[0079] The reference model is a model used by the second device, or the reference model is used by the second device to determine the model to use.

[0080] Specifically, the first device may be a TE, a DUT, a terminal, a base station, an access network device, etc.

[0081] The first device can send first information related to the reference model to the second device. For example, the first information can be used to instruct the second device to use the reference model, so that the first device and the second device can align the method of using the reference model to save signaling overhead; or, the first information can be used to instruct the first device to use the reference model. When the second device learns that the first device uses the reference model, it can use the reference model, or a model that matches the reference model, or a model determined based on the reference model.

[0082] The first device may also collect indicators of the reference model, and the indicators of the reference model may include, for example, performance indicators of the model itself, such as the error and similarity between the model output and the expected output, or communication performance indicators determined based on information output by the model, such as the throughput of the communication link, the block error rate, etc. The first device may test the second device based on the collected indicators, or further optimize the performance of the communication network.

[0083] The first device can collect indicators on its own side, and can also receive indicators reported by the second device.

[0084] In the method of this embodiment, the first device performs a first operation, and the first operation includes at least one of the following: sending first information related to a reference model to a second device; collecting indicators of the reference model; the reference model is a model used by the second device, or the reference model is used for the model determined to be used by the second device. In the above scheme, performance testing or communication is performed based on the reference model, so that the AI ​​models used by different communication devices have a certain consistency, which can make it easier for the network side to manage the AI ​​models of communication devices, thereby improving the overall performance of the communication network.

[0085] Optionally, the reference model is predefined by the protocol, including at least one of the following situations:

[0086] All model structures and all model parameters are predefined by the protocol;

[0087] All model structures are predefined by the protocol;

[0088] All model parameters are predefined by the protocol;

[0089] All model structures and some model parameters are predefined by the protocol;

[0090] Some model structures and some model parameters are predefined by the protocol;

[0091] Some model structures and all model parameters are predefined by the protocol.

[0092] Specifically, the protocol predefinition may be predefined in a standard protocol, i.e., certain agreements are made; for example, predefined in a 3GPP standard protocol. Alternatively, the protocol predefinition may also be predefined in a private agreement; for example, the first device developer and the second device developer privately negotiate, and a third party may also participate in the private agreement. Alternatively, the protocol predefinition may also be agreed upon in advance before the first device or the second device uses the reference model or the matching unit of the reference model, without interacting through communication signaling.

[0093] Predefined methods for reference models:

[0094] 1. Predefined type 1: The protocol predefines the entire model, including all model structures and all model parameters;

[0095] 2. Predefined type 2: The protocol predefines all or part of the model structure, and / or all or part of the model parameter; this type does not include "the protocol predefines all model structures and all model parameters", that is, there must be remaining parts that are not predefined by the protocol.

[0096] (1) The remaining part is transmitted from the first device to the second device (e.g., a base station, a terminal), or from the second device to the first device;

[0097] (2) a portion of the remaining portion is transmitted by the first device to the second device, or by the second device to the first device, and another portion of the remaining portion is implemented by the second device itself;

[0098] Example 1: The protocol predefines the model structure, and the rest is the model parameter;

[0099] Example 2: The protocol predefines the model structure and some model parameters, and the rest are the remaining model parameters;

[0100] Example 3: The protocol predefines part of the model structure and part of the model parameters (the predefined part of the model parameters are the parameters of the predefined part of the model structure), and the remaining part is the remaining model structure and the model parameters of the remaining model structure; that is, the model structure is divided into two parts, A and B, the protocol predefines structure A and the model parameters corresponding to structure A, and the remaining part is structure B and the model parameters corresponding to structure B.

[0101] Optionally, the type of reference model includes at least one of the following: a unilateral type, a bilateral type; the unilateral type is used to indicate a type of model that does not require the cooperation of a model of another device; the bilateral type is used to indicate a type of model in which two devices communicating with each other each have a matching model.

[0102] Specifically, the reference model includes a unilateral reference model and a bilateral reference model;

[0103] 1. Single-side type: This means that the reference model is used only on one device and does not require the cooperation of models of other devices. For example, the reference model is the model used by the second device.

[0104] 2. Bilateral type: In normal use, this model contains two parts located in two devices, and the two parts need to be used together. For example, one part is at the sending end of the communication, and the other part is at the receiving end of the communication.

[0105] For example, in a test scenario, when testing the model on the terminal side, the first device needs to use a matching model on the access network device side; when testing the model on the access network device side, the first device needs to use a matching model on the terminal side.

[0106] Optionally, the reference model may refer to a partial model used by the second device, a partial model used by the first device, or the entirety of both parts. For CSI compression, the model on the terminal side is an encoder, which compresses the channel information into PMI; the model on the access network device side is a decoder, which restores PMI to channel information. If the terminal is a second device (such as DUT), the partial model used by the second device is an encoder, and the model used by the first device (such as TE) is a decoder; if the access network device is the second device, the partial model used by the second device is a decoder, and the model used by the first device is an encoder. In other bilateral model situations, the part that runs first on one side is generally called an encoder, and the part that runs later on the other side is called a decoder.

[0107] Exemplarily, the reference model may refer to an encoder, may refer to a decoder, or may refer to the whole of an encoder and a decoder. For example, if the reference model refers to an encoder, it may be called a reference encoder; if the reference model refers to a decoder, it may be called a reference decoder.

[0108] Optionally, when the reference model refers to a model on one side, the matched part on the other side becomes a matching unit.

[0109] For example, if the reference model refers to the encoder, the decoder is the matching unit; if the reference model refers to the decoder, the encoder is the matching unit.

[0110] Optionally, the matching unit may be an AI unit, an AI model, or may be a non-AI unit, a non-AI model.

[0111] For example, the reference model refers to a model used on the second device side. For example, when testing an encoder of a terminal, the reference model refers to an encoder; when testing a decoder on the network side, the reference model refers to a decoder.

[0112] For example, the reference model refers to the model used on the first device side. For example, when testing the encoder of the terminal, the reference model refers to the decoder that matches the encoder; for example, when testing the decoder on the network side, the reference model refers to the encoder that matches the decoder.

[0113] Optionally, the first information includes at least one of the following:

[0114] first indication information, where the first indication information is used to instruct the first device to use the reference model;

[0115] second indication information, where the second indication information is used to instruct the second device to use the reference model, or to use a matching unit of the reference model; the matching unit is a model that matches the reference model when the type of the reference model is a bilateral type;

[0116] Identification of at least one of the environment, scenario, and use case;

[0117] third indication information, where the third indication information is used to indicate that the second device is in at least one of a specific environment, a specific scenario, and a specific use case;

[0118] a model identifier, the model identifier being associated with at least one of a reference model and a matching unit of the reference model;

[0119] first model information, the first model information comprising at least one of the following: a reference model, information in the reference model that is not predefined by the protocol, a matching unit of the reference model, and a part of the matching unit of the reference model;

[0120] fourth indication information, where the fourth indication information is used to instruct transmission of the first model information;

[0121] fifth indication information, where the fifth indication information is used to instruct the second device to perform model training;

[0122] Training dataset.

[0123] Specifically, the first indication information can be used to directly indicate that the second device uses the reference model, that is, "the second device uses the reference model" is a specific behavior, and the first indication information is used to directly indicate the specific behavior.

[0124] For example, for a bilateral type model, if the second device is a terminal, it means directly instructing the terminal to use a reference encoder; if the second device is a base station, it means directly instructing the base station to use a reference decoder.

[0125] Optionally, the model used by the second device includes at least one of the following: the reference model, a matching unit of the reference model, and a target model; the target model is determined based on second information of the reference model, and the second information includes at least one of the following: function, characteristic, input data, output data, performance, and at least one indicator for lifecycle management.

[0126] Specifically, the second device may not necessarily use the reference model, but may use a model that has similar or identical functions, characteristics, inputs and outputs, or performance as the reference model. In other words, the indicators / characteristics of the lifecycle management of the model used by the second device are the same as the indicators / characteristics of the lifecycle management of the reference model.

[0127] The second indication information can be used to directly indicate that the first device uses the reference model, that is, "the first device uses the reference model" is a specific behavior, and the first indication information is used to directly indicate the specific behavior; or, directly indicate that the second device uses the matching unit of the reference model.

[0128] For example, in a testing scenario, the second device does not necessarily know that it is currently in a testing state, and the first device needs to tell the second device that "the first device uses a reference model"; for example, the second device may have different lifecycle management strategies for different models used by the first device, such as different performance monitoring strategies.

[0129] For example, for the bilateral model, if the second device is a terminal, it means directly instructing the first device to use a reference decoder, and the terminal uses an encoder that matches the reference decoder; if the second device is a base station, it means directly instructing the first device to use a reference encoder, and the base station uses a decoder that matches the reference encoder.

[0130] In one embodiment, the third indication information is used to indicate that the second device is in a certain environment, scenario, or use case (such as a test environment), or the first information includes: an identification of the environment, scenario, or use case in which the second device is located;

[0131] Exemplarily, the second device receives the third indication information and can learn that it is in a certain specific environment, scenario or use case; or, the second device receives an identifier of a specific environment, scenario or use case and determines that a reference model needs to be used.

[0132] Optionally, the first information may include a model identifier, the model identifier being associated with / indicating a reference model (applicable to a unilateral type, a bilateral type), or a matching unit of the reference model (applicable to a bilateral type);

[0133] For example, for a bilateral model, if the second device is a terminal, if the model identifier is associated / indicated with a reference encoder, the terminal uses the reference encoder; if the model identifier is associated / indicated with a reference decoder or an encoder matching the reference decoder, the terminal uses the encoder matching the reference decoder;

[0134] It should be noted that if the model identifier is associated with / indicates a reference decoder, it means that the first device only indicates the model identifier of the reference decoder, and the terminal needs to find the encoder that matches the reference decoder by itself; if the model identifier is associated with / indicates an encoder that matches the reference decoder, it means that the first device indicates the model identifier of the encoder that matches the reference decoder, and the terminal can directly use the encoder associated with / indicated by the model identifier.

[0135] If the second device is a base station, if the model identifier is associated with / indicates a reference decoder, the base station uses the reference decoder; if the model identifier is associated with / indicates a reference encoder or a decoder matching the reference encoder, the base station uses the decoder matching the reference encoder;

[0136] It should be noted that if the model identifier is associated with / indicates a reference encoder, it means that the first device only indicates the model identifier of the reference encoder, and the base station needs to find a decoder that matches the reference encoder by itself; if the model identifier is associated with / indicates a decoder that matches the reference encoder, it means that the first device indicates the model identifier of a decoder that matches the reference encoder, and the base station can directly use the decoder associated with / indicated by the model identifier.

[0137] In one embodiment, the fourth indication information is used to indicate the delivery / transmission of the first model information (model deliver / transfer), and the delivery / transmission refers to the first device delivering / transmitting the reference model or a part of the reference model (applicable to unilateral type, bilateral type, such as information not predefined by the protocol), the matching unit of the reference unit or a part of the matching unit (applicable to bilateral type) to the second device.

[0138] Optionally, the method further comprises:

[0139] The first device receives feedback information from the second device;

[0140] The feedback information is used to indicate whether the second device supports a second operation of the first model information, where the second operation includes at least one of the following: transmission, transfer, inference, execution, storage, compilation, and localization processing.

[0141] Specifically, before and / or after the first model information is passed / transmitted, the second device may also provide feedback on whether the passing / transmission of the first model information is supported; if the first device passes / transmits a reference model or a part of a reference model, the second device may provide feedback on whether a second operation such as passing, transferring, inference, running, storing, compiling or localizing the reference model or a part of the reference model is supported; if the first device passes / transmits a matching unit of a reference unit model or a part of a matching unit, the second device may provide feedback on whether a second operation such as passing, transferring, inference, running, storing, compiling or localizing the matching unit of the reference model unit or a part of a matching unit is supported.

[0142] For example, for the bilateral model, if the second device is a terminal, the first device passes / transmits a reference encoder or an encoder matching the reference decoder to the terminal; if the second device is a base station, the first device passes / transmits a reference decoder or a decoder matching the reference encoder to the base station.

[0143] In one embodiment, the fifth indication information is used to indicate a training model, or the first information may include a training data set, used to instruct the second device to train or update a reference model or a part of a reference model (applicable to a unilateral type, a bilateral type), a matching unit of a reference model unit, or a part of a matching unit (applicable to a bilateral type).

[0144] Optionally, the fourth indication information is used to instruct the second device to perform model training on a first model, where the first model includes at least one of the following:

[0145] The reference model, a part of the reference model, a matching unit of the reference model, a part of the matching unit of the reference model.

[0146] When the type of the reference model is a bilateral type, the first model is trained independently, and the second device is a terminal, the training data set is generated based on a reference decoder, and the training data set includes input data and output data of an encoder; for example, the output of the encoder is the input of the decoder, and the input of the encoder is the expected output of the decoder;

[0147] When the type of the reference model is a bilateral type, the first model is independently trained, and the second device is an access network device, the training data set is generated based on a reference encoder, and the training data set includes input data and output data of a decoder; for example, the output of the encoder is the input of the decoder, and the input of the encoder is the expected output of the reference decoder;

[0148] In the case where the type of the reference model is a bilateral type and the first model is jointly trained, the training data set is generated based on an encoder or a decoder, and the training data set includes at least one of the following:

[0149] Gradients of the encoder’s input layer;

[0150] The gradient of the encoder’s output layer;

[0151] The gradient of the decoder’s input layer;

[0152] The gradient of the decoder’s output layer;

[0153] The output data of the encoder;

[0154] Input data for the encoder;

[0155] Input data to the decoder;

[0156] The output data of the decoder.

[0157] For separate training of the bilateral model, if the second device is a terminal, the training data set is generated based on a decoder, and the training data set includes input data and output data of the encoder.

[0158] If the second device is a base station, the training data set is generated based on a reference encoder, and the training data set includes input data and output data of a decoder.

[0159] For joint training of the bilateral model, the training data set is generated based on at least one of a reference decoder and a reference encoder. For example, the output of the encoder is the input of the decoder, and the input of the encoder is the expected output of the reference decoder.

[0160] In a possible implementation manner, the second device enables a reference model, a matching unit of the reference model, or a target model determined based on the reference model;

[0161] For example, there is a delay between the second device receiving the first information from the first device and the second device enabling the model, and compared with the predefined type one, the predefined type two requires an additional delay.

[0162] For predefined type 2, before the second device enables the reference model, it is also necessary to receive first model information, for example, the first model information includes at least one of the following: information not predefined by the protocol in the reference model, a matching unit of the reference model, and a part of the matching unit of the reference model;

[0163] The second device receives the remaining part (i.e., the remaining part in the reference model that is not predefined by the protocol) transmitted from the first device (e.g., including a base station, a core network node, etc.);

[0164] Optionally, the second device transmits the remaining part (i.e., the remaining part of the reference model that is not predefined by the protocol) to the first device;

[0165] Optionally, the second device obtains the remaining part through a third-party server (eg, an Over The Top (OTT) server), a DUT server, a terminal server, a network (NetWork, NW) server or a core network server.

[0166] Optionally, the second device implements the remaining part by itself, and optionally, auxiliary information from a third device is required to assist model generation, model selection or model training, for example, the auxiliary information includes a data set or configuration parameters, etc.;

[0167] Optionally, the third device includes at least one of the following:

[0168] First equipment, access network equipment, core network equipment, third-party server, DUT server, terminal server, network server, core network server.

[0169] Optionally, the second device implements the remaining part by itself without requiring additional auxiliary information.

[0170] In the above implementation, the second device can use the reference model, the matching unit of the reference model, or the target model determined based on the reference model, which has greater flexibility. In addition, when the protocol predefines part of the reference model information, it can implement the remaining part by itself, or receive the remaining part sent by other devices. The implementation complexity is low, and the signaling overhead is small when it is implemented by itself.

[0171] Optionally, the first device collects the index of the reference model, which can be implemented in the following manner:

[0172] (1) The first device collects indicators of the reference model at the first device;

[0173] (2) The first device receives the indicator of the reference model sent by the second device.

[0174] Specifically, in solution (1), the first device collects the indicators obtained on the first device side;

[0175] In solution (2), the second device collects the indicators obtained at the second device and then reports them to the first device.

[0176] Optionally, the index of the reference model includes at least one of the following:

[0177] Direct indicators and indirect indicators; the direct indicators are used to represent the performance indicators of the reference model, and the indirect indicators are used to represent the overall performance indicators of the communication link obtained after business processing based on the reference model.

[0178] Optionally, the direct indicator includes at least one of the following:

[0179] At least one of cosine similarity and vector correlation between the output data of the reference model and the expected output data;

[0180] At least one of the normalized mean square error NMSE, mean square error MSE, and linear error between the output data of the reference model and the expected output data;

[0181] At least one of beam prediction error and beam prediction accuracy in a beam management scenario;

[0182] At least one of the position error, positioning error and error of intermediate positioning information in the positioning scenario; the intermediate positioning information includes at least one of the following: time of arrival (TOA), time difference of arrival (TDOA), line-of-sight (LOS) delay, angle of arrival (AOA), angle of departure (AOD) and round trip time (RTT);

[0183] The indirect indicator includes at least one of the following:

[0184] The throughput of the communication link, the block error rate (BLER).

[0185] Optionally, the first device collects the index of the reference model at the first device, including:

[0186] (1-1) When the reference model is of a bilateral type, the indicator is a direct indicator, and the second device is a terminal, the first device collects output data of a decoder;

[0187] The first device collects input data of the encoder, or the first device receives input data of the encoder sent by the second device;

[0188] The first device calculates the indicator based on input data of the encoder and output data of the decoder;

[0189] (1-2) When the reference model is of a bilateral type, the indicator is a direct indicator, and the second device is an access network, the first device receives output data of a decoder sent by the second device;

[0190] The first device collects input data of the encoder;

[0191] The first device calculates the indicator based on input data of the encoder and output data of the decoder.

[0192] Optionally, the first device receiving the indicator of the reference model sent by the second device includes:

[0193] (2-1) When the reference model is of a bilateral type, the indicator is a direct indicator, and the second device is a terminal, the first device collects output data of a decoder and sends the output data of the decoder to the second device;

[0194] The first device receives the indicator sent by the second device, where the indicator is calculated by the second device based on output data of the decoder and input data of the encoder collected by the second device; or

[0195] (2-2) When the reference model is of a bilateral type, the indicator is a direct indicator, and the second device is an access network, the first device collects input data of an encoder and sends the input data of the encoder to the second device;

[0196] The first device receives the indicator sent by the second device, where the indicator is calculated by the second device based on output data of the decoder and input data of the encoder collected by the second device.

[0197] Specifically, for a bilateral type model, if the second device is a terminal, the indicator is a direct indicator, such as cosine similarity, or an associated indicator of cosine similarity (such as the square of cosine similarity);

[0198] (1-1), the first device collects output data of the decoder, the first device collects input data of the encoder or the second device reports input data of the encoder, and then the first device calculates the direct indicator of the reference model;

[0199] (2-1) The first device collects the output data of the decoder and sends it to the second device. The second device collects the input data of the encoder. Then the second device calculates the direct indicator and sends it to the first device.

[0200] If the second device is a base station, the indicator is a direct indicator, such as cosine similarity, or an indicator associated with cosine similarity (such as the square of cosine similarity);

[0201] (1-2) The second device collects the output data of the decoder and reports it to the first device. The first device collects the input data of the encoder, and then the first device calculates the direct indicator;

[0202] (2-2) The second device collects the output data of the decoder, the first device collects the input data of the encoder, and sends it to the second device, and then the second device calculates the direct indicator and sends it to the first device.

[0203] Optionally, when the type of the reference model is a bilateral type and the indicator is an indirect indicator, the indicator is collected by the first device itself.

[0204] Optionally, for indirect indicators, a similar solution to the above direct indicators can also be adopted, which will not be described in detail here.

[0205] In the above implementation manner, the first device can collect the indicators of the reference model in a variety of different ways, which has greater flexibility and lower implementation complexity.

[0206] In summary, the benefits of using reference models are mainly:

[0207] 1. Save signaling overhead:

[0208] For example, reference model related capabilities do not require additional signaling interactions.

[0209] 2. Ensure the stability of the communication system:

[0210] a) The reference model has a classic, simple structure, stable performance, and good robustness, and can achieve good performance in common scenarios / most scenarios;

[0211] b) The reference model needs to undergo RAN testing (such as RAN4 testing), is a model supported by the terminal, and the performance of the reference model is also verified by the RAN.

[0212] Figure 3 This is a second flow chart of a communication method based on a reference model provided in an embodiment of the present application, such as Figure 3 As shown, the method is applied to a second device, comprising:

[0213] Step 201: The second device performs a third operation, where the third operation includes at least one of the following:

[0214] Receiving first information related to the reference model sent by the first device;

[0215] receiving auxiliary information sent by a third device, and generating second model information based on the auxiliary information;

[0216] generating second model information;

[0217] Sending first model information to the first device;

[0218] Collect metrics for the reference model;

[0219] The reference model is a model used by the second device, or the reference model is used by the second device to determine the model to use.

[0220] Optionally, the reference model is predefined by a protocol, including at least one of the following situations:

[0221] All model structures and all model parameters are predefined by the protocol;

[0222] All model structures are predefined by the protocol;

[0223] All model parameters are predefined by the protocol;

[0224] All model structures and some model parameters are predefined by the protocol;

[0225] Some model structures and some model parameters are predefined by the protocol;

[0226] Some model structures and all model parameters are predefined by the protocol.

[0227] Optionally, the type of the reference model includes at least one of the following: a unilateral type, a bilateral type; the unilateral type is used to indicate a type of model that does not require the cooperation of a model of another device; the bilateral type is used to indicate a type of model in which two devices that communicate with each other each have a matching model.

[0228] Optionally, the first information includes at least one of the following:

[0229] first indication information, where the first indication information is used to instruct the first device to use the reference model;

[0230] second indication information, wherein the first indication information is used to instruct the second device to use the reference model, or to use a matching unit of the reference model; the matching unit is a model that matches the reference model when the type of the reference model is a bilateral type;

[0231] Identification of at least one of the environment, scenario, and use case;

[0232] third indication information, where the third indication information is used to indicate that the second device is in at least one of a specific environment, a specific scenario, and a specific use case;

[0233] a model identifier, the model identifier being associated with at least one of a reference model and a matching unit of the reference model;

[0234] first model information, the first model information comprising at least one of the following: the reference model, information in the reference model that is not predefined by the protocol, a matching unit of the reference model, and a part of the matching unit of the reference model;

[0235] fourth indication information, where the fourth indication information is used to instruct transmission of the first model information;

[0236] fifth indication information, where the fifth indication information is used to instruct the second device to perform model training;

[0237] Training dataset.

[0238] Optionally, the second model information includes: information not predefined by the protocol in the reference model, a matching unit of the reference model, and a part of the matching unit of the reference model.

[0239] Optionally, when the first information includes the first model information, the method further includes:

[0240] The second device sends feedback information to the first device;

[0241] The feedback information is used to indicate whether the second device supports a second operation of the first model information, where the second operation includes at least one of the following: transmission, transfer, inference, execution, storage, compilation, and localization processing.

[0242] Optionally, the fifth indication information is used to instruct the second device to perform model training, and the trained model includes at least one of the following:

[0243] The reference model, a part of the reference model, a matching unit of the reference model, a part of the matching unit of the reference model.

[0244] Optionally, when the type of the reference model is a bilateral type and the second device is a terminal, the first model information includes at least one of the following: an encoder and an encoder matched by a decoder;

[0245] When the type of the reference model is a bilateral type and the second device is an access network device, the first model information includes at least one of the following: a decoder, and a decoder matched with an encoder.

[0246] Optionally, when the type of the reference model is a bilateral type, the first model is independently trained, and the second device is a terminal, the training data set is generated based on a decoder, and the training data set includes input data and output data of an encoder;

[0247] When the reference model is of a bilateral type, the first model is trained independently, and the second device is an access network device, the training data set is generated based on an encoder, and the training data set includes input data and output data of a decoder;

[0248] In the case where the type of the reference model is a bilateral type and the first model is jointly trained, the training data set is generated based on an encoder or a decoder, and the training data set includes at least one of the following:

[0249] Gradients of the encoder’s input layer;

[0250] The gradient of the encoder’s output layer;

[0251] The gradient of the decoder’s input layer;

[0252] The gradient of the decoder’s output layer;

[0253] The output data of the encoder;

[0254] Input data for the encoder;

[0255] Input data to the decoder;

[0256] The output data of the decoder.

[0257] Optionally, the model used by the second device includes at least one of the following: the reference model, a matching unit of the reference model, and a target model; the target model is determined based on second information of the reference model, and the second information includes at least one of the following: function, characteristic, input data, output data, performance, and at least one indicator for lifecycle management.

[0258] Optionally, the second device collecting the index of the reference model includes:

[0259] The second device collects the indicators of the reference model at the second device, and sends the collected indicators of the reference model to the first device.

[0260] Optionally, the index of the reference model includes at least one of the following:

[0261] Direct indicators and indirect indicators; the direct indicators are used to represent the performance indicators of the reference model, and the indirect indicators are used to represent the overall performance indicators of the communication link obtained after business processing based on the reference model.

[0262] Optionally, the direct indicator includes at least one of the following:

[0263] At least one of cosine similarity and vector correlation between the output data of the reference model and the expected output data;

[0264] At least one of the normalized mean square error NMSE, mean square error MSE, and linear error between the output data of the reference model and the expected output data;

[0265] At least one of beam prediction error and beam prediction accuracy in a beam management scenario;

[0266] At least one of the position error, positioning error and error of intermediate positioning information in the positioning scenario; the intermediate positioning information includes at least one of the following: arrival time TOA, arrival time difference TDOA, line of sight LOS delay, arrival angle AOA, departure angle AOD and round-trip time RTT;

[0267] The indirect indicator includes at least one of the following:

[0268] The throughput of the communication link, the block error rate BLER.

[0269] Optionally, the method further comprises:

[0270] In a case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is a terminal, the second device collects input data of the encoder;

[0271] The second device sends the input data of the encoder to the first device; the input data of the encoder is used by the first device to calculate the index of the reference model;

[0272] In the case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is an access network, the second device collects output data of a decoder;

[0273] The second device sends the output data of the decoder to the first device; the output data of the decoder is used by the first device to calculate the index of the reference model.

[0274] Optionally, the second device collects the indicator of the reference model at the second device, and sends the collected indicator of the reference model to the second device and the first device, including:

[0275] In the case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is a terminal, the second device receives output data of the decoder sent by the first device;

[0276] The second device calculates the index of the reference model based on the output data of the decoder and the input data of the encoder collected by the second device, and sends the index of the reference model to the first device; or

[0277] In a case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is an access network, the second device receives input data of an encoder sent by the first device;

[0278] The second device calculates the index of the reference model based on the output data of the decoder and the input data of the encoder collected by the second device, and sends the index of the reference model to the first device.

[0279] Optionally, the third device includes at least one of the following:

[0280] The first device, access network device, core network device, third-party server, DUT server, terminal server, network server, and core network server.

[0281] The specific implementation process and technical effects of the method of this embodiment are similar to those in the first device side method embodiment. For details, please refer to the detailed introduction in the first device side method embodiment, and no further details will be given here.

[0282] The reference model-based communication method provided in the embodiment of the present application can be executed by a reference model-based communication device. In the embodiment of the present application, a reference model-based communication device executing the reference model-based communication method is taken as an example to illustrate the reference model-based communication device provided in the embodiment of the present application.

[0283] Figure 4 is one of the structural diagrams of the communication device based on the reference model provided in the embodiment of the present application, such as Figure 4 As shown, the reference model-based communication device is applied to a first device and includes:

[0284] The processing module 110 is configured to perform a first operation, where the first operation includes at least one of the following:

[0285] Sending first information related to the reference model to the second device;

[0286] Collect metrics for the reference model;

[0287] The reference model is a model used by the second device, or the reference model is used by the second device to determine a model to use.

[0288] Optionally, the reference model is predefined by a protocol, including at least one of the following situations:

[0289] All model structures and all model parameters are predefined by the protocol;

[0290] All model structures are predefined by the protocol;

[0291] All model parameters are predefined by the protocol;

[0292] All model structures and some model parameters are predefined by the protocol;

[0293] Some model structures and some model parameters are predefined by the protocol;

[0294] Some model structures and all model parameters are predefined by the protocol.

[0295] Optionally, the type of the reference model includes at least one of the following: a unilateral type, a bilateral type; the unilateral type is used to indicate a type of model that does not require the cooperation of a model of another device; the bilateral type is used to indicate a type of model in which two devices that communicate with each other each have a matching model.

[0296] Optionally, the first information includes at least one of the following:

[0297] first indication information, where the first indication information is used to instruct the first device to use the reference model;

[0298] second indication information, wherein the first indication information is used to instruct the second device to use the reference model, or to use a matching unit of the reference model; the matching unit is a model that matches the reference model when the type of the reference model is a bilateral type;

[0299] Identification of at least one of the environment, scenario, and use case;

[0300] third indication information, where the third indication information is used to indicate that the second device is in at least one of a specific environment, a specific scenario, and a specific use case;

[0301] a model identifier, the model identifier being associated with at least one of a reference model and a matching unit of the reference model;

[0302] first model information, the first model information comprising at least one of the following: the reference model, information in the reference model that is not predefined by the protocol, a matching unit of the reference model, and a part of the matching unit of the reference model;

[0303] fourth indication information, where the fourth indication information is used to instruct transmission of the first model information;

[0304] fifth indication information, where the fifth indication information is used to instruct the second device to perform model training;

[0305] Training dataset.

[0306] Optionally, when the first information includes the first model information, the apparatus further includes:

[0307] A receiving module, configured to receive feedback information from the second device;

[0308] The feedback information is used to indicate whether the second device supports a second operation of the first model information, where the second operation includes at least one of the following: transmission, transfer, inference, execution, storage, compilation, and localization processing.

[0309] Optionally, the fifth indication information is used to instruct the second device to perform model training, and the trained model includes at least one of the following:

[0310] The reference model, a part of the reference model, a matching unit of the reference model, a part of the matching unit of the reference model.

[0311] Optionally, when the type of the reference model is a bilateral type and the second device is a terminal, the first model information includes at least one of the following: an encoder and an encoder matched by a decoder;

[0312] When the type of the reference model is a bilateral type and the second device is an access network device, the first model information includes at least one of the following: a decoder, and a decoder matched with an encoder.

[0313] Optionally, when the type of the reference model is a bilateral type, the first model is independently trained, and the second device is a terminal, the training data set is generated based on a decoder, and the training data set includes input data and output data of an encoder;

[0314] When the reference model is of a bilateral type, the first model is trained independently, and the second device is an access network device, the training data set is generated based on an encoder, and the training data set includes input data and output data of a decoder;

[0315] In the case where the type of the reference model is a bilateral type and the first model is jointly trained, the training data set is generated based on an encoder or a decoder, and the training data set includes at least one of the following:

[0316] Gradients of the encoder’s input layer;

[0317] The gradient of the encoder’s output layer;

[0318] The gradient of the decoder’s input layer;

[0319] The gradient of the decoder’s output layer;

[0320] The output data of the encoder;

[0321] Input data for the encoder;

[0322] Input data to the decoder;

[0323] The output data of the decoder.

[0324] Optionally, the model used by the second device includes at least one of the following: the reference model, a matching unit of the reference model, and a target model; the target model is determined based on second information of the reference model, and the second information includes at least one of the following: function, characteristic, input data, output data, performance, and at least one indicator for lifecycle management.

[0325] Optionally, the processing module 110 is specifically configured to perform at least one of the following:

[0326] collecting metrics of the reference model at the first device;

[0327] The control receiving module receives the index of the reference model sent by the second device.

[0328] Optionally, the index of the reference model includes at least one of the following:

[0329] Direct indicators and indirect indicators; the direct indicators are used to represent the performance indicators of the reference model, and the indirect indicators are used to represent the overall performance indicators of the communication link obtained after business processing based on the reference model.

[0330] Optionally, the direct indicator includes at least one of the following:

[0331] At least one of cosine similarity and vector correlation between the output data of the reference model and the expected output data;

[0332] At least one of the normalized mean square error NMSE, mean square error MSE, and linear error between the output data of the reference model and the expected output data;

[0333] At least one of beam prediction error and beam prediction accuracy in a beam management scenario;

[0334] At least one of the position error, positioning error and error of intermediate positioning information in the positioning scenario; the intermediate positioning information includes at least one of the following: arrival time TOA, arrival time difference TDOA, line of sight LOS delay, arrival angle AOA, departure angle AOD and round-trip time RTT;

[0335] The indirect indicator includes at least one of the following:

[0336] The throughput of the communication link, the block error rate BLER.

[0337] Optionally, the processing module 110 is specifically configured to:

[0338] In a case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is a terminal, collecting output data of a decoder;

[0339] Collecting input data of the encoder, or controlling the receiving module to receive input data of the encoder sent by the second device;

[0340] Calculating the indicator based on input data of the encoder and output data of the decoder;

[0341] When the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is an access network, controlling the receiving module to receive output data of the decoder sent by the second device;

[0342] Collect input data for the encoder;

[0343] The indicator is calculated based on input data of the encoder and output data of the decoder.

[0344] Optionally,

[0345] In the case where the reference model is of a bilateral type, the indicator is a direct indicator, and the second device is a terminal, the processing module 110 is further configured to: collect output data of the decoder; and a sending module is configured to send the output data of the decoder to the second device;

[0346] The receiving module is further used to receive the indicator sent by the second device, where the indicator is calculated by the second device based on the output data of the decoder and the input data of the encoder collected by the second device; or

[0347] In the case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is an access network, the processing module 110 is further used to: collect input data of the encoder; a sending module is used to send the input data of the encoder to the second device;

[0348] The receiving module is further used to receive the indicator sent by the second device, where the indicator is calculated by the second device based on the output data of the decoder collected by the second device and the input data of the encoder.

[0349] Optionally, when the type of the reference model is a bilateral type and the indicator is an indirect indicator, the indicator is collected by the first device itself.

[0350] The device of this embodiment can be used to execute the method of any one of the embodiments in the aforementioned first device side method embodiment. Its specific implementation process and technical effects are similar to those in the first device side method embodiment. For details, please refer to the detailed introduction in the first device side method embodiment, which will not be repeated here.

[0351] Figure 5 This is a second structural diagram of a communication device based on a reference model provided in an embodiment of the present application, such as Figure 5 As shown, the reference model-based communication device, applied to a second device, includes:

[0352] The processing module 210 is configured to perform a third operation, where the third operation includes at least one of the following:

[0353] Receiving first information related to the reference model sent by the first device;

[0354] receiving auxiliary information sent by the third device, and generating second model information based on the auxiliary information;

[0355] generating second model information;

[0356] Sending first model information to the first device;

[0357] Collect metrics for the reference model;

[0358] The reference model is a model used by the second device, or the reference model is used by the second device to determine a model to use.

[0359] Optionally, the reference model is predefined by a protocol, including at least one of the following situations:

[0360] All model structures and all model parameters are predefined by the protocol;

[0361] All model structures are predefined by the protocol;

[0362] All model parameters are predefined by the protocol;

[0363] All model structures and some model parameters are predefined by the protocol;

[0364] Some model structures and some model parameters are predefined by the protocol;

[0365] Some model structures and all model parameters are predefined by the protocol.

[0366] Optionally, the type of the reference model includes at least one of the following: a unilateral type, a bilateral type; the unilateral type is used to indicate a type of model that does not require the cooperation of a model of another device; the bilateral type is used to indicate a type of model in which two devices that communicate with each other each have a matching model.

[0367] Optionally, the first information includes at least one of the following:

[0368] first indication information, where the first indication information is used to instruct the first device to use the reference model;

[0369] second indication information, wherein the first indication information is used to instruct the second device to use the reference model, or to use a matching unit of the reference model; the matching unit is a model that matches the reference model when the type of the reference model is a bilateral type;

[0370] Identification of at least one of the environment, scenario, and use case;

[0371] third indication information, where the third indication information is used to indicate that the second device is in at least one of a specific environment, a specific scenario, and a specific use case;

[0372] a model identifier, the model identifier being associated with at least one of a reference model and a matching unit of the reference model;

[0373] first model information, the first model information comprising at least one of the following: the reference model, information in the reference model that is not predefined by the protocol, a matching unit of the reference model, and a part of the matching unit of the reference model;

[0374] fourth indication information, where the fourth indication information is used to instruct transmission of the first model information;

[0375] fifth indication information, where the fifth indication information is used to instruct the second device to perform model training;

[0376] Training dataset.

[0377] Optionally, the second model information includes: information not predefined by the protocol in the reference model, a matching unit of the reference model, and a part of the matching unit of the reference model.

[0378] Optionally, when the first information includes the first model information, the apparatus further includes:

[0379] A sending module, used to send feedback information to the first device;

[0380] The feedback information is used to indicate whether the second device supports a second operation of the first model information, where the second operation includes at least one of the following: transmission, transfer, inference, execution, storage, compilation, and localization processing.

[0381] Optionally, the fifth indication information is used to instruct the second device to perform model training, and the trained model includes at least one of the following:

[0382] The reference model, a part of the reference model, a matching unit of the reference model, a part of the matching unit of the reference model.

[0383] Optionally, when the type of the reference model is a bilateral type and the second device is a terminal, the first model information includes at least one of the following: an encoder and an encoder matched by a decoder;

[0384] When the type of the reference model is a bilateral type and the second device is an access network device, the first model information includes at least one of the following: a decoder, and a decoder matched with an encoder.

[0385] Optionally, when the type of the reference model is a bilateral type, the first model is independently trained, and the second device is a terminal, the training data set is generated based on a decoder, and the training data set includes input data and output data of an encoder;

[0386] When the reference model is of a bilateral type, the first model is trained independently, and the second device is an access network device, the training data set is generated based on an encoder, and the training data set includes input data and output data of a decoder;

[0387] In the case where the type of the reference model is a bilateral type and the first model is jointly trained, the training data set is generated based on an encoder or a decoder, and the training data set includes at least one of the following:

[0388] Gradients of the encoder’s input layer;

[0389] The gradient of the encoder’s output layer;

[0390] The gradient of the decoder’s input layer;

[0391] The gradient of the decoder’s output layer;

[0392] The output data of the encoder;

[0393] Input data for the encoder;

[0394] Input data to the decoder;

[0395] The output data of the decoder.

[0396] Optionally, the model used by the second device includes at least one of the following: the reference model, a matching unit of the reference model, and a target model; the target model is determined based on second information of the reference model, and the second information includes at least one of the following: function, characteristic, input data, output data, performance, and at least one indicator for lifecycle management.

[0397] Optionally, the processing module 210 is specifically configured to:

[0398] collecting metrics of the reference model at the second device;

[0399] A sending module is used to send the collected indicators of the reference model to the second device and the first device.

[0400] Optionally, the index of the reference model includes at least one of the following:

[0401] Direct indicators and indirect indicators; the direct indicators are used to represent the performance indicators of the reference model, and the indirect indicators are used to represent the overall performance indicators of the communication link obtained after business processing based on the reference model.

[0402] Optionally, the direct indicator includes at least one of the following:

[0403] At least one of cosine similarity and vector correlation between the output data of the reference model and the expected output data;

[0404] At least one of the normalized mean square error NMSE, mean square error MSE, and linear error between the output data of the reference model and the expected output data;

[0405] At least one of beam prediction error and beam prediction accuracy in a beam management scenario;

[0406] At least one of the position error, positioning error and error of intermediate positioning information in the positioning scenario; the intermediate positioning information includes at least one of the following: arrival time TOA, arrival time difference TDOA, line of sight LOS delay, arrival angle AOA, departure angle AOD and round-trip time RTT;

[0407] The indirect indicator includes at least one of the following:

[0408] The throughput of the communication link, the block error rate BLER.

[0409] Optionally, the processing module 210 is further configured to:

[0410] In a case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is a terminal, collecting input data of an encoder;

[0411] In a case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is an access network, collecting output data of a decoder;

[0412] The sending module is further used to send the input data of the encoder to the first device when the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is a terminal; the input data of the encoder is used by the first device to calculate the indicator of the reference model;

[0413] When the reference model is of bilateral type, the indicator is a direct indicator, and the second device is an access network, the output data of the decoder is sent to the first device; the output data of the decoder is used by the first device to calculate the indicator of the reference model.

[0414] Optionally, the device further includes: a receiving module;

[0415] In the case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is a terminal, the receiving module is used to receive output data of the decoder sent by the first device;

[0416] The processing module 210 is specifically configured to: calculate the index of the reference model based on the output data of the decoder and the input data of the encoder collected by the second device;

[0417] The sending module is further configured to send the index of the reference model to the first device; or,

[0418] In the case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is an access network, the receiving module is used to receive input data of the encoder sent by the first device;

[0419] The processing module 210 is specifically configured to: calculate the index of the reference model based on the output data of the decoder and the input data of the encoder collected by the second device;

[0420] The sending module is further used to send the index of the reference model to the first device.

[0421] Optionally, the third device includes at least one of the following:

[0422] The first device, access network device, core network device, third-party server, DUT server, terminal server, network server, and core network server.

[0423] The device of this embodiment can be used to execute the method of any one of the embodiments in the aforementioned second device side method embodiment. Its specific implementation process and technical effects are similar to those in the second device side method embodiment. For details, please refer to the detailed introduction in the second device side method embodiment, which will not be repeated here.

[0424] The reference model-based communication device in the embodiment of the present application may 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 may be a terminal, or may be other devices other than a terminal. Exemplarily, the terminal may include but is not limited to the types of the terminal 11 listed above, and other devices may be servers, network attached storage (NAS), etc., which are not specifically limited in the embodiment of the present application.

[0425] The communication device based on the reference model provided in the embodiment of the present application can realize Figures 2 to 3 The various processes implemented by the method embodiment and achieving the same technical effect are not described here to avoid repetition.

[0426] like Figure 6 As shown, the embodiment of the present application further provides a communication device 600, including a processor 601 and a memory 602, and the memory 602 stores a program or instruction that can be run on the processor 601. For example, when the communication device 600 is a network side device, the program or instruction is executed by the processor 601 to implement the various steps of the above-mentioned communication method embodiment based on the reference model, and can achieve the same technical effect. When the communication device 600 is a terminal, the program or instruction is executed by the processor 601 to implement the various steps of the above-mentioned communication method embodiment based on the reference model, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0427] The embodiment of the present application also provides a network side device, including a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the following Figure 2-Figure 3 The steps of the method embodiment shown. This network side device embodiment corresponds to the above network side device method embodiment, and each implementation process and implementation mode of the above method embodiment can be applied to this network side device embodiment and can achieve the same technical effect.

[0428] Specifically, the embodiment of the present application further provides a network side device. Specifically, the embodiment of the present application further provides a network side device. Figure 7 As shown, the network side device 700 includes: an antenna 71, a radio frequency device 72, a baseband device 73, a processor 74 and a memory 75. The antenna 71 is connected to the radio frequency device 72. In the uplink direction, the radio frequency device 72 receives information through the antenna 71 and sends the received information to the baseband device 73 for processing. In the downlink direction, the baseband device 73 processes the information to be sent and sends it to the radio frequency device 72. The radio frequency device 72 processes the received information and sends it out through the antenna 71.

[0429] The method executed by the network-side device in the above embodiment may be implemented in the baseband device 73, which includes a baseband processor.

[0430] The baseband device 73 may include, for example, at least one baseband board on which a plurality of chips are arranged. Figure 7 As shown, one of the chips is, for example, a baseband processor, which is connected to the memory 75 via a bus interface to call the program in the memory 75 to execute the network device operations shown in the above method embodiment.

[0431] The network side device may further include a network interface 76, which is, for example, a Common Public Radio Interface (CPRI).

[0432] Specifically, the network side device 700 of the embodiment of the present application further includes: instructions or programs stored in the memory 75 and executable on the processor 74, and the processor 74 calls the instructions or programs in the memory 75 to execute. Figure 4 or Figure 5 The methods executed by the modules shown achieve the same technical effects, and therefore will not be described here in detail to avoid repetition.

[0433] The embodiment of the present application also provides a terminal, including a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the following Figure 2-Figure 3 The steps in the method embodiment shown. 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, Figure 8 The hardware structure diagram of a terminal for implementing an embodiment of the present application is shown in FIG. The terminal 800 includes, but is not limited to, at least some of the components of a radio frequency unit 801, a network module 802, an audio output unit 803, an input unit 804, a sensor 805, a display unit 806, a user input unit 807, an interface unit 808, a memory 809, and a processor 810.

[0434] Those skilled in the art will appreciate that the terminal 800 may also include a power source (such as a battery) for supplying power to various components, and the power source may be logically connected to the processor 810 through a power management system, thereby implementing functions such as managing charging, discharging, and power consumption management through the power management system. Figure 8 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.

[0435] It should be understood that in the embodiment of the present application, the input unit 804 may include a graphics processing unit (GPU) 8041 and a microphone 8042, and the graphics processor 8041 processes the image data of a static picture or video obtained by an image capture device (such as a camera) in a video capture mode or an image capture mode. The display unit 806 may include a display panel 8061, and the display panel 8061 may be configured in the form of a liquid crystal display, an organic light emitting diode, etc. The user input unit 807 includes a touch panel 8071 and at least one of other input devices 8072. The touch panel 8071 is also called a touch screen. The touch panel 8071 may include two parts: a touch detection device and a touch controller. Other input devices 8072 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.

[0436] In the embodiment of the present application, after receiving downlink data from the network side device, the radio frequency unit 801 can transmit the data to the processor 810 for processing; in addition, the radio frequency unit 801 can send uplink data to the network side device. Generally, the radio frequency unit 801 includes but is not limited to an antenna, an amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, etc.

[0437] The memory 809 can be used to store software programs or instructions and various data. The memory 809 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 809 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 809 in the embodiment of the present application includes but is not limited to these and any other suitable types of memory.

[0438] The processor 810 may include one or more processing units; optionally, the processor 810 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 810.

[0439] The processor 810 is configured to perform a third operation, where the third operation includes at least one of the following:

[0440] Receiving first information related to the reference model sent by the first device;

[0441] receiving auxiliary information sent by the third device, and generating second model information based on the auxiliary information;

[0442] generating second model information;

[0443] Sending first model information to the first device;

[0444] Collect metrics for the reference model;

[0445] The reference model is a model used by the second device, or the reference model is used by the second device to determine a model to use.

[0446] Optionally, the reference model is predefined by a protocol, including at least one of the following situations:

[0447] All model structures and all model parameters are predefined by the protocol;

[0448] All model structures are predefined by the protocol;

[0449] All model parameters are predefined by the protocol;

[0450] All model structures and some model parameters are predefined by the protocol;

[0451] Some model structures and some model parameters are predefined by the protocol;

[0452] Some model structures and all model parameters are predefined by the protocol.

[0453] Optionally, the type of the reference model includes at least one of the following: a unilateral type, a bilateral type; the unilateral type is used to indicate a type of model that does not require the cooperation of a model of another device; the bilateral type is used to indicate a type of model in which two devices that communicate with each other each have a matching model.

[0454] Optionally, the first information includes at least one of the following:

[0455] first indication information, where the first indication information is used to instruct the first device to use the reference model;

[0456] second indication information, wherein the first indication information is used to instruct the second device to use the reference model, or to use a matching unit of the reference model; the matching unit is a model that matches the reference model when the type of the reference model is a bilateral type;

[0457] Identification of at least one of the environment, scenario, and use case;

[0458] third indication information, where the third indication information is used to indicate that the second device is in at least one of a specific environment, a specific scenario, and a specific use case;

[0459] a model identifier, the model identifier being associated with at least one of a reference model and a matching unit of the reference model;

[0460] first model information, the first model information comprising at least one of the following: the reference model, information in the reference model that is not predefined by the protocol, a matching unit of the reference model, and a part of the matching unit of the reference model;

[0461] fourth indication information, where the fourth indication information is used to instruct transmission of the first model information;

[0462] fifth indication information, where the fifth indication information is used to instruct the second device to perform model training;

[0463] Training dataset.

[0464] Optionally, the second model information includes: information not predefined by the protocol in the reference model, a matching unit of the reference model, and a part of the matching unit of the reference model.

[0465] Optionally, when the first information includes the first model information, the radio frequency unit 801 is configured to send feedback information to the first device;

[0466] The feedback information is used to indicate whether the second device supports a second operation of the first model information, where the second operation includes at least one of the following: transmission, transfer, inference, execution, storage, compilation, and localization processing.

[0467] Optionally, the fifth indication information is used to instruct the second device to perform model training, and the trained model includes at least one of the following:

[0468] The reference model, a part of the reference model, a matching unit of the reference model, a part of the matching unit of the reference model.

[0469] Optionally, when the type of the reference model is a bilateral type and the second device is a terminal, the first model information includes at least one of the following: an encoder and an encoder matched by a decoder;

[0470] When the type of the reference model is a bilateral type and the second device is an access network device, the first model information includes at least one of the following: a decoder, and a decoder matched with an encoder.

[0471] Optionally, when the type of the reference model is a bilateral type, the first model is independently trained, and the second device is a terminal, the training data set is generated based on a decoder, and the training data set includes input data and output data of an encoder;

[0472] When the reference model is of a bilateral type, the first model is trained independently, and the second device is an access network device, the training data set is generated based on an encoder, and the training data set includes input data and output data of a decoder;

[0473] In the case where the type of the reference model is a bilateral type and the first model is jointly trained, the training data set is generated based on an encoder or a decoder, and the training data set includes at least one of the following:

[0474] Gradients of the encoder’s input layer;

[0475] The gradient of the encoder’s output layer;

[0476] The gradient of the decoder’s input layer;

[0477] The gradient of the decoder’s output layer;

[0478] The output data of the encoder;

[0479] Input data for the encoder;

[0480] Input data to the decoder;

[0481] The output data of the decoder.

[0482] Optionally, the model used by the second device includes at least one of the following: the reference model, a matching unit of the reference model, and a target model; the target model is determined based on second information of the reference model, and the second information includes at least one of the following: function, characteristic, input data, output data, performance, and at least one indicator for lifecycle management.

[0483] Optionally, the processor 810 is specifically configured to:

[0484] collecting metrics of the reference model at the second device;

[0485] The radio frequency unit 801 is used to send the collected indicators of the reference model to the second device and the first device.

[0486] Optionally, the index of the reference model includes at least one of the following:

[0487] Direct indicators and indirect indicators; the direct indicators are used to represent the performance indicators of the reference model, and the indirect indicators are used to represent the overall performance indicators of the communication link obtained after business processing based on the reference model.

[0488] Optionally, the direct indicator includes at least one of the following:

[0489] At least one of cosine similarity and vector correlation between the output data of the reference model and the expected output data;

[0490] At least one of the normalized mean square error NMSE, mean square error MSE, and linear error between the output data of the reference model and the expected output data;

[0491] At least one of beam prediction error and beam prediction accuracy in a beam management scenario;

[0492] At least one of the position error, positioning error and error of intermediate positioning information in the positioning scenario; the intermediate positioning information includes at least one of the following: arrival time TOA, arrival time difference TDOA, line of sight LOS delay, arrival angle AOA, departure angle AOD and round-trip time RTT;

[0493] The indirect indicator includes at least one of the following:

[0494] The throughput of the communication link, the block error rate BLER.

[0495] Optionally, the processor 810 is further configured to:

[0496] In a case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is a terminal, collecting input data of an encoder;

[0497] In a case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is an access network, collecting output data of a decoder;

[0498] The radio frequency unit 801 is further configured to, when the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is a terminal, send the input data of the encoder to the first device; the input data of the encoder is used by the first device to calculate the indicator of the reference model;

[0499] When the reference model is of bilateral type, the indicator is a direct indicator, and the second device is an access network, the output data of the decoder is sent to the first device; the output data of the decoder is used by the first device to calculate the indicator of the reference model.

[0500] Optionally, when the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is a terminal, the radio frequency unit 801 is further used to receive output data of the decoder sent by the first device;

[0501] The processor 810 is specifically configured to: calculate an index of the reference model based on the output data of the decoder and the input data of the encoder collected by the second device;

[0502] The radio frequency unit 801 is further configured to send the index of the reference model to the first device; or,

[0503] When the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is an access network, the radio frequency unit 801 is further used to receive input data of the encoder sent by the first device;

[0504] The processor 810 is specifically configured to: calculate an index of the reference model based on the output data of the decoder and the input data of the encoder collected by the second device;

[0505] The radio frequency unit 801 is further configured to send the index of the reference model to the first device.

[0506] Optionally, the third device includes at least one of the following:

[0507] The first device, access network device, core network device, third-party server, DUT server, terminal server, network server, and core network server.

[0508] The above description is made using the solution on the second device side as an example. The terminal can also execute the solution on the first device side, which will not be repeated here.

[0509] It can be understood that the implementation process of each implementation method mentioned in this embodiment can refer to Figure 2 or Figure 3 The relevant descriptions of the method embodiments shown in the figure achieve the same or corresponding technical effects, and will not be repeated here to avoid repetition.

[0510] 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 reference model-based communication method embodiment are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0511] 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.

[0512] 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 reference model-based communication method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0513] 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.

[0514] The embodiment of the present application further provides a computer program / program product, which is stored in a storage medium, and is executed by at least one processor to implement the various processes of the above-mentioned reference model-based communication method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0515] An embodiment of the present application also provides a communication system, including: a first device and a second device, wherein the first device can be used to execute the steps of the communication method based on the reference model as described above, and the second device can be used to execute the steps of the communication method based on the reference model as described above.

[0516] 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.

[0517] 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.

[0518] 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 communication method based on a reference model, characterized in that: include: The first device performs a first operation, where the first operation includes at least one of the following: Sending first information related to the reference model to the second device; Collect metrics for the reference model; The reference model is a model used by the second device, or the reference model is used by the second device to determine a model to use.

2. The method according to claim 1, characterized in that The reference model is predefined by the protocol and includes at least one of the following situations: All model structures and all model parameters are predefined by the protocol; All model structures are predefined by the protocol; All model parameters are predefined by the protocol; All model structures and some model parameters are predefined by the protocol; Some model structures and some model parameters are predefined by the protocol; Some model structures and all model parameters are predefined by the protocol.

3. The method according to claim 1 or 2, characterized in that: The type of the reference model includes at least one of the following: a unilateral type and a bilateral type; the unilateral type is used to indicate a type of model that does not require the cooperation of a model of another device; the bilateral type is used to indicate a type of model that matches each other between two devices that communicate with each other.

4. The method according to any one of claims 1 to 3, characterized in that: The first information includes at least one of the following: first indication information, where the first indication information is used to instruct the first device to use the reference model; second indication information, where the second indication information is used to instruct the second device to use the reference model, or to use a matching unit of the reference model; the matching unit is a model that matches the reference model when the type of the reference model is a bilateral type; Identification of at least one of the environment, scenario, and use case; third indication information, where the third indication information is used to indicate that the second device is in at least one of a specific environment, a specific scenario, and a specific use case; a model identifier, the model identifier being associated with at least one of a reference model and a matching unit of the reference model; first model information, the first model information comprising at least one of the following: the reference model, information in the reference model that is not predefined by the protocol, a matching unit of the reference model, and a part of the matching unit of the reference model; fourth indication information, where the fourth indication information is used to instruct transmission of the first model information; fifth indication information, where the fifth indication information is used to instruct the second device to perform model training; Training dataset.

5. The method according to claim 4, characterized in that In the case where the first information includes the first model information, the method further includes: The first device receives feedback information from the second device; The feedback information is used to indicate whether the second device supports a second operation of the first model information, where the second operation includes at least one of the following: transmission, transfer, inference, execution, storage, compilation, and localization processing.

6. The method according to claim 4, characterized in that The fifth indication information is used to instruct the second device to perform model training, where the trained model includes at least one of the following: The reference model, a part of the reference model, a matching unit of the reference model, a part of the matching unit of the reference model.

7. The method according to any one of claims 4 to 6, characterized in that: In the case where the type of the reference model is a bilateral type and the second device is a terminal, the first model information includes at least one of the following: an encoder matched by an encoder and a decoder; When the type of the reference model is a bilateral type and the second device is an access network device, the first model information includes at least one of the following: a decoder, and a decoder matched with an encoder.

8. The method according to any one of claims 4 to 6, characterized in that: When the reference model is of a bilateral type, the first model is trained independently, and the second device is a terminal, the training data set is generated based on a decoder, and the training data set includes input data and output data of an encoder; When the reference model is of a bilateral type, the first model is trained independently, and the second device is an access network device, the training data set is generated based on an encoder, and the training data set includes input data and output data of a decoder; In the case where the type of the reference model is a bilateral type and the first model is jointly trained, the training data set is generated based on an encoder or a decoder, and the training data set includes at least one of the following: Gradients of the encoder’s input layer; The gradient of the encoder’s output layer; The gradient of the decoder’s input layer; The gradient of the decoder’s output layer; The output data of the encoder; Input data for the encoder; Input data to the decoder; Output data of the decoder.

9. The method according to any one of claims 1 to 8, characterized in that: The model used by the second device includes at least one of the following: the reference model, the matching unit of the reference model, and the target model; the target model is determined based on second information of the reference model, and the second information includes at least one of the following: function, characteristic, input data, output data, performance, and at least one indicator for lifecycle management.

10. The method according to any one of claims 1 to 9, characterized in that: The first device collects the reference model index including at least one of the following: The first device collects, at the first device, indicators of the reference model; The first device receives the indicator of the reference model sent by the second device.

11. The method according to any one of claims 1 to 10, characterized in that: The index of the reference model includes at least one of the following: Direct indicators and indirect indicators; the direct indicators are used to represent the performance indicators of the reference model, and the indirect indicators are used to represent the overall performance indicators of the communication link obtained after business processing based on the reference model.

12. The method according to claim 11, characterized in that The direct indicators include at least one of the following: At least one of cosine similarity and vector correlation between the output data of the reference model and the expected output data; At least one of the normalized mean square error NMSE, mean square error MSE, and linear error between the output data of the reference model and the expected output data; At least one of beam prediction error and beam prediction accuracy in a beam management scenario; At least one of the position error, positioning error and error of intermediate positioning information in the positioning scenario; the intermediate positioning information includes at least one of the following: arrival time TOA, arrival time difference TDOA, line of sight LOS delay, arrival angle AOA, departure angle AOD and round-trip time RTT; The indirect indicator includes at least one of the following: The throughput of the communication link, the block error rate BLER.

13. The method according to claim 10, characterized in that The first device collects the index of the reference model at the first device, including: In the case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is a terminal, the first device collects output data of a decoder; The first device collects input data of the encoder, or the first device receives input data of the encoder sent by the second device; The first device calculates the indicator based on input data of the encoder and output data of the decoder; In the case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is an access network, the first device receives output data of a decoder sent by the second device; The first device collects input data of the encoder; The first device calculates the indicator based on input data of the encoder and output data of the decoder.

14. The method according to claim 10, characterized in that The first device receiving the indicator of the reference model sent by the second device includes: In the case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is a terminal, the first device collects output data of a decoder and sends the output data of the decoder to the second device; The first device receives the indicator sent by the second device, where the indicator is calculated by the second device based on output data of the decoder and input data of the encoder collected by the second device; or In a case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is an access network, the first device collects input data of an encoder and sends the input data of the encoder to the second device; The first device receives the indicator sent by the second device, where the indicator is calculated by the second device based on output data of the decoder and input data of the encoder collected by the second device.

15. The method according to claim 10, characterized in that When the type of the reference model is a bilateral type and the indicator is an indirect indicator, the indicator is collected by the first device itself.

16. A communication method based on a reference model, characterized in that: include: The second device performs a third operation, where the third operation includes at least one of the following: Receiving first information related to the reference model sent by the first device; receiving auxiliary information sent by a third device, and generating second model information based on the auxiliary information; generating second model information; Sending first model information to the first device; Collect metrics for the reference model; The reference model is a model used by the second device, or the reference model is used by the second device to determine a model to use.

17. The method according to claim 16, characterized in that The reference model is predefined by the protocol and includes at least one of the following situations: All model structures and all model parameters are predefined by the protocol; All model structures are predefined by the protocol; All model parameters are predefined by the protocol; All model structures and some model parameters are predefined by the protocol; Some model structures and some model parameters are predefined by the protocol; Some model structures and all model parameters are predefined by the protocol.

18. The method according to claim 16 or 17, characterized in that The type of the reference model includes at least one of the following: a unilateral type and a bilateral type; the unilateral type is used to indicate a type of model that does not require the cooperation of a model of another device; the bilateral type is used to indicate a type of model that matches each other between two devices that communicate with each other.

19. The method according to any one of claims 16 to 18, characterized in that: The first information includes at least one of the following: first indication information, where the first indication information is used to instruct the first device to use the reference model; second indication information, where the second indication information is used to instruct the second device to use the reference model, or to use a matching unit of the reference model; the matching unit is a model that matches the reference model when the type of the reference model is a bilateral type; Identification of at least one of the environment, scenario, and use case; third indication information, where the third indication information is used to indicate that the second device is in at least one of a specific environment, a specific scenario, and a specific use case; a model identifier, the model identifier being associated with at least one of a reference model and a matching unit of the reference model; first model information, the first model information comprising at least one of the following: the reference model, information in the reference model that is not predefined by the protocol, a matching unit of the reference model, and a part of the matching unit of the reference model; fourth indication information, where the fourth indication information is used to instruct transmission of the first model information; fifth indication information, where the fifth indication information is used to instruct the second device to perform model training; Training dataset.

20. The method according to claim 19, characterized in that The second model information includes: information not predefined by the protocol in the reference model, a matching unit of the reference model, and a part of the matching unit of the reference model.

21. The method according to claim 19, characterized in that In the case where the first information includes the first model information, the method further includes: The second device sends feedback information to the first device; The feedback information is used to indicate whether the second device supports a second operation of the first model information, where the second operation includes at least one of the following: transmission, transfer, inference, execution, storage, compilation, and localization processing.

22. The method according to claim 19, characterized in that The fifth indication information is used to instruct the second device to perform model training, where the trained model includes at least one of the following: The reference model, a part of the reference model, a matching unit of the reference model, a part of the matching unit of the reference model.

23. The method according to any one of claims 16 to 22, characterized in that: In the case where the type of the reference model is a bilateral type and the second device is a terminal, the first model information includes at least one of the following: an encoder matched by an encoder and a decoder; When the type of the reference model is a bilateral type and the second device is an access network device, the first model information includes at least one of the following: a decoder, and a decoder matched with an encoder.

24. The method according to any one of claims 19 to 22, characterized in that: When the reference model is of a bilateral type, the first model is trained independently, and the second device is a terminal, the training data set is generated based on a decoder, and the training data set includes input data and output data of an encoder; When the reference model is of a bilateral type, the first model is trained independently, and the second device is an access network device, the training data set is generated based on an encoder, and the training data set includes input data and output data of a decoder; In the case where the type of the reference model is a bilateral type and the first model is jointly trained, the training data set is generated based on an encoder or a decoder, and the training data set includes at least one of the following: Gradients of the encoder’s input layer; The gradient of the encoder’s output layer; The gradient of the decoder’s input layer; The gradient of the decoder’s output layer; The output data of the encoder; Input data for the encoder; Input data to the decoder; Output data of the decoder.

25. The method according to any one of claims 16 to 24, characterized in that: The model used by the second device includes at least one of the following: the reference model, the matching unit of the reference model, and the target model; the target model is determined based on second information of the reference model, and the second information includes at least one of the following: function, characteristic, input data, output data, performance, and at least one indicator for lifecycle management.

26. The method according to any one of claims 16 to 25, characterized in that: The second device collects the reference model indicators including: The second device collects the indicators of the reference model at the second device, and sends the collected indicators of the reference model to the first device.

27. The method according to any one of claims 16 to 26, characterized in that: The index of the reference model includes at least one of the following: Direct indicators and indirect indicators; the direct indicators are used to represent the performance indicators of the reference model, and the indirect indicators are used to represent the overall performance indicators of the communication link obtained after business processing based on the reference model.

28. The method according to claim 27, characterized in that The direct indicators include at least one of the following: At least one of cosine similarity and vector correlation between the output data of the reference model and the expected output data; At least one of the normalized mean square error NMSE, mean square error MSE, and linear error between the output data of the reference model and the expected output data; At least one of beam prediction error and beam prediction accuracy in a beam management scenario; At least one of the position error, positioning error and error of intermediate positioning information in the positioning scenario; the intermediate positioning information includes at least one of the following: arrival time TOA, arrival time difference TDOA, line of sight LOS delay, arrival angle AOA, departure angle AOD and round-trip time RTT; The indirect indicator includes at least one of the following: The throughput of the communication link, the block error rate BLER.

29. The method according to claim 26, characterized in that The method further comprises: In a case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is a terminal, the second device collects input data of the encoder; The second device sends the input data of the encoder to the first device; the input data of the encoder is used by the first device to calculate the index of the reference model; In the case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is an access network, the second device collects output data of a decoder; The second device sends the output data of the decoder to the first device; the output data of the decoder is used by the first device to calculate the index of the reference model.

30. The method according to claim 26, characterized in that The second device collects the index of the reference model at the second device, and sends the collected index of the reference model to the second device and the first device, including: In the case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is a terminal, the second device receives output data of the decoder sent by the first device; The second device calculates the index of the reference model based on the output data of the decoder and the input data of the encoder collected by the second device, and sends the index of the reference model to the first device; or In a case where the type of the reference model is a bilateral type, the indicator is a direct indicator, and the second device is an access network, the second device receives input data of an encoder sent by the first device; The second device calculates the index of the reference model based on the output data of the decoder and the input data of the encoder collected by the second device, and sends the index of the reference model to the first device.

31. The method according to any one of claims 16 to 30, characterized in that: The third device includes at least one of the following: The first device, access network device, core network device, third-party server, DUT server, terminal server, network server, and core network server.

32. A communication device based on a reference model, characterized in that: include: The processing module is configured to perform a first operation, wherein the first operation includes at least one of the following: Sending first information related to the reference model to the second device; Collect metrics for the reference model; The reference model is a model used by the second device, or the reference model is used by the second device to determine a model to use.

33. A communication device based on a reference model, characterized in that: include: The processing module is configured to perform a third operation, wherein the third operation includes at least one of the following: Receiving first information related to the reference model sent by the first device; receiving auxiliary information sent by a third device, and generating second model information based on the auxiliary information; generating second model information; Sending first model information to the first device; Collect metrics for the reference model; The reference model is a model used by the second device, or the reference model is used by the second device to determine a model to use.

34. A first device, 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 steps of the reference model-based communication method according to any one of claims 1 to 15 are implemented.

35. A second device, characterized in that: It 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 steps of the reference model-based communication method as described in any one of claims 16 to 31 are implemented.

36. A readable storage medium, characterized in that: The readable storage medium stores a program or instruction, and when the program or instruction is executed by the processor, it implements the reference model-based communication method as described in any one of claims 1 to 15, or implements the steps of the reference model-based communication method as described in any one of claims 16 to 31.

Images