Model information transmission method and apparatus

Abstract

Provided in the present disclosure are a model information transmission method and apparatus, which aim to solve the problem of how to ensure that a UE can use an AI / ML model transferred by a network. The method in the embodiments of the present disclosure comprises: sending model indication information, wherein the model indication information is used for indicating the structure of an artificial intelligence / machine learning (AI / ML) model that is supported by a terminal. By means of model indication information, a network-side device can learn the structure of an AI / ML model that is supported by a terminal, such that the network-side device can subsequently transfer to the terminal a model having the structure of the AI / ML model that is supported by the terminal, thereby enabling the terminal to use the AI / ML model transferred by the network-side device.

Description

Model information transmission method and device

[0001] This disclosure claims priority to the Chinese patent application filed with the China Patent Office on February 1, 2024, with application number 202410145603.4 and application name “Model Information Transmission Method and Device,” the entire contents of which are incorporated herein by reference. Technical Field

[0002] The present disclosure relates to the field of communication technology, and in particular to a model information transmission method and device. Background Art

[0003] With the development of artificial intelligence (AI) and machine learning (ML), AI / ML models are used in related technologies to improve the performance of communication systems. In Rel-18 New Radio (NR), work methods that support AI / ML models or AI / ML functions are studied, and in Rel-19 NR, air interface support for AI / ML is planned to be standardized. In some cases, the network (NW) needs to transfer / deliver the AI / ML model to the user equipment (UE) so that the UE can use the model provided by the NW for reasoning. However, a UE's support for AI / ML models is limited. If the UE cannot support the model structure transmitted by the NW, it cannot use the model transmitted by the NW. Summary of the Invention

[0004] The purpose of the present disclosure is to provide a model information transmission method and device to solve the problem of how to ensure that the UE can use the AI / ML model transmitted by the network.

[0005] In order to achieve the above-mentioned object, the present disclosure provides a model information transmission method applied to a terminal, comprising:

[0006] Send model indication information, where the model indication information is used to indicate the structure of the artificial intelligence or machine learning AI / ML model supported by the terminal.

[0007] In some embodiments, the model indication information includes at least one of the following:

[0008] Model identification;

[0009] Meta information of the model, wherein the meta information is descriptive information of the model and includes model structure information and / or model structure identifier;

[0010] Model structure identification;

[0011] Model structure information;

[0012] AI / ML reference model.

[0013] In some embodiments, before sending the model indication information, the method further includes:

[0014] Obtaining model candidate information sent by a network-side device, the model candidate information including at least one of a model identifier, metadata, a model structure identifier, and model structure information of at least one candidate AI / ML model, and / or including at least one candidate AI / ML reference model;

[0015] In the model candidate information, the model candidate information corresponding to the AI / ML model supported by the terminal is selected as the model indication information.

[0016] In some embodiments, the model structure information indicates at least one of the following:

[0017] Type, number of layers, number of nodes, connection method between layers, and connection method between nodes.

[0018] In some embodiments, the model identifier carries model structure information;

[0019] Alternatively, there is a corresponding relationship between the model identifier and the model structure information.

[0020] In some embodiments, different fields or different domains or different sub-IDs in the model identifier are used to indicate different structural parameters in the model structure information;

[0021] Alternatively, the specific field, specific domain, or specific sub-ID in the model identifier is used to indicate a combination of structural parameters in the model structure information.

[0022] In some embodiments, the model structure identifier carries model structure information;

[0023] Alternatively, there is a corresponding relationship between the model structure identifier and the model structure information.

[0024] In some embodiments, different fields or different domains or different sub-IDs in the model structure identifier are used to indicate different structural parameters in the model structure information;

[0025] Alternatively, the specific field, specific domain, or specific sub-ID in the model structure identifier is used to indicate a combination of structure parameters in the model structure information.

[0026] In some embodiments, after sending the model indication information, the method further includes:

[0027] Obtain a target AI / ML model transmitted by a network-side device, where the structure of the target AI / ML model is the structure of an AI / ML model supported by the terminal.

[0028] The present disclosure also provides a model information transmission method, which is applied to a network-side device and includes:

[0029] Obtaining model indication information, where the model indication information is used to indicate the structure of an artificial intelligence or machine learning AI / ML model supported by the terminal;

[0030] Determine the structure of the AI / ML model supported by the terminal according to the model indication information.

[0031] In some embodiments, the model indication information includes at least one of the following:

[0032] Model identification;

[0033] Meta information of the model, wherein the meta information is descriptive information of the model and includes model structure information and / or model structure identifier;

[0034] Model structure identification;

[0035] Model structure information;

[0036] AI / ML Reference Model:

[0037] In some embodiments, the method of the present disclosure further includes:

[0038] Send model candidate information to the terminal, where the model candidate information includes at least one of a model identifier, metadata, a model structure identifier, and model structure information of at least one candidate AI / ML model, and / or includes at least one candidate AI / ML reference model.

[0039] In some embodiments, the model structure information indicates at least one of the following:

[0040] Type, number of layers, number of nodes, connection method between layers, and connection method between nodes.

[0041] In some embodiments, the model identifier carries model structure information;

[0042] Alternatively, there is a corresponding relationship between the model identifier and the model structure information.

[0043] In some embodiments, different fields or different domains or different sub-IDs in the model identifier are used to indicate different structural parameters in the model structure information;

[0044] Alternatively, a specific field, a specific domain, or a specific sub-ID in the model identifier is used to indicate a combination of structural parameters in the model structure information.

[0045] In some embodiments, the model structure identifier carries model structure information;

[0046] Alternatively, there is a corresponding relationship between the model structure identifier and the model structure information.

[0047] In some embodiments, different fields or different domains or different sub-IDs in the model structure identifier are used to indicate different structural parameters in the model structure information;

[0048] Alternatively, the specific field, specific domain, or specific sub-ID in the model structure identifier is used to indicate a combination of structure parameters in the model structure information.

[0049] In some embodiments, the method of the present disclosure further includes:

[0050] A target AI / ML model is sent to the terminal, where the structure of the target AI / ML model is the structure of the AI / ML model supported by the terminal.

[0051] The embodiment of the present disclosure also provides a model information transmission device, including a memory, a transceiver, and a processor;

[0052] A memory for storing a computer program; a transceiver for transmitting and receiving data under the control of the processor; and a processor for reading the computer program in the memory and performing the following operations:

[0053] Send model indication information, where the model indication information is used to indicate the structure of the artificial intelligence or machine learning AI / ML model supported by the terminal.

[0054] The embodiment of the present disclosure also provides a model information transmission device, including a memory, a transceiver, and a processor;

[0055] A memory for storing a computer program; a transceiver for transmitting and receiving data under the control of the processor; and a processor for reading the computer program in the memory and performing the following operations:

[0056] Obtaining model indication information, where the model indication information is used to indicate the structure of an artificial intelligence or machine learning AI / ML model supported by the terminal;

[0057] Determine the structure of the AI / ML model supported by the terminal according to the model indication information.

[0058] The present disclosure also provides a model information transmission device, including:

[0059] The first sending unit is used to send model indication information, where the model indication information is used to indicate the structure of the artificial intelligence or machine learning AI / ML model supported by the terminal.

[0060] The present disclosure also provides a model information transmission device, including:

[0061] A first acquiring unit is configured to acquire model indication information, where the model indication information is used to indicate a structure of an artificial intelligence or machine learning AI / ML model supported by the terminal;

[0062] A determining unit is configured to determine a structure of an AI / ML model supported by the terminal according to the model indication information.

[0063] An embodiment of the present disclosure further provides a processor-readable storage medium, wherein the processor-readable storage medium stores a computer program, and the computer program is used to enable the processor to execute the steps of the model information transmission method as described above.

[0064] The embodiment of the present disclosure further provides a computer program product, including computer instructions, which implement the steps of the model information transmission method described above when executed by a processor.

[0065] The above technical solution disclosed in the present invention has at least the following beneficial effects:

[0066] In an embodiment of the present disclosure, model indication information is sent to indicate the structure of an artificial intelligence or machine learning AI / ML model supported by the terminal. This model indication information enables the network-side device to learn the structure of the AI / ML model supported by the terminal, so that the network-side device can subsequently transmit the model of the AI / ML model structure supported by the terminal to the terminal, thereby enabling the terminal to use the AI / ML model transmitted by the network-side device. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] FIG1 shows one structural diagram of a network system to which the embodiments of the present disclosure may be applied;

[0068] FIG2 shows a flow chart of a model information transmission method according to an embodiment of the present disclosure;

[0069] FIG3 shows a second structural diagram of a network system applicable to the embodiment of the present disclosure;

[0070] FIG4 shows one of the interactive schematic diagrams of the model information transmission method according to an embodiment of the present disclosure;

[0071] FIG5 shows a second interactive diagram of the model information transmission method according to an embodiment of the present disclosure;

[0072] FIG6 shows a third interactive diagram of the model information transmission method according to an embodiment of the present disclosure;

[0073] FIG7 shows a fourth interactive diagram of the model information transmission method according to an embodiment of the present disclosure;

[0074] FIG8 shows a second flow chart of the model information transmission method according to an embodiment of the present disclosure;

[0075] FIG9 shows one structural block diagram of the model information transmission device according to an embodiment of the present disclosure;

[0076] FIG10 shows a second structural block diagram of the model information transmission device according to an embodiment of the present disclosure;

[0077] FIG11 shows one of the module schematic diagrams of the model information transmission device according to an embodiment of the present disclosure;

[0078] FIG12 shows a second module schematic diagram of the model information transmission device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0079] The following will clearly and completely describe the technical solutions in the embodiments of the present disclosure in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present disclosure without making any creative efforts shall fall within the scope of protection of the present disclosure.

[0080] The terms "first," "second," and the like in the specification and claims of the present disclosure are used to distinguish similar objects and are not necessarily used to describe a particular order or sequential sequence. It should be understood that the terms used in this manner are interchangeable where appropriate so that the embodiments of the present disclosure described herein may be implemented in a sequence other than those illustrated or described herein. In addition, the terms "including" and "having," and any variations thereof, are intended to cover non-exclusive inclusions, e.g., a process, method, system, product, or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such processes, methods, products, or apparatus.

[0081] In the embodiments of the present disclosure, the term "and / or" describes the relationship between associated objects, indicating that three possible relationships exist. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally indicates that the associated objects are in an "or" relationship. In the embodiments of the present disclosure, the term "plurality" refers to two or more, and other quantifiers are similar.

[0082] In the embodiments of the present disclosure, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present disclosure should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0083] Figure 1 shows a block diagram of a wireless communication system to which the embodiments of the present disclosure can be applied. The wireless communication system includes a terminal device 11 and a network-side device (or network device) 12. The terminal device 11 may also be referred to as a terminal or a user equipment (UE). It should be noted that the specific type of the terminal 11 is not limited in the embodiments of the present disclosure. The network-side device 12 may be a base station or a core network. It should be noted that in the embodiments of the present disclosure, only a base station in the NR system is used as an example, but the specific type of the base station is not limited.

[0084] In order to enable those skilled in the art to better understand the embodiments of the present disclosure, the following description is first given.

[0085] In this disclosure, the term "model" generally refers to an AI / ML model, which is a data-driven algorithm that can infer and predict outputs based on certain inputs. AI / ML models may have unilateral models, that is, the model is deployed only on one side, such as the network (NW) side or the terminal (UE) side; or bilateral models, that is, the model is deployed on both the NW and UE sides, and collaboration between both sides is required for complete reasoning. In this disclosure, functionality generally refers to AI / ML functionality. AI / ML functionality often refers to the AI / ML capabilities that the UE side can possess. A UE with AI / ML functionality can perform reasoning, and the NW does not need to identify which AI / ML models the UE has or which models it uses.

[0086] In some cases, the NW needs to transfer / deliver (transfer / deliver, hereinafter mainly uses transfer or "transfer" to refer to "transmission / delivery") the AI / ML model to the UE so that the UE can use the model provided by the NW for reasoning. For example, the NW has trained an AI / ML model that is particularly suitable for use in this cell based on the data collected in this cell, and it is beneficial for the UE to use this model. However, generally speaking, a UE can only support a specific AI / ML model structure, which depends on the capabilities of each UE. Therefore, a UE's support for AI / ML models is limited. If the UE cannot support the model structure transmitted by the NW, it cannot use the model transmitted by the NW.

[0087] Regarding the transmission of AI / ML models from NW to UE, relevant technologies believe that transmitting a model with a "known structure" to the UE is a more feasible solution than transmitting a model with an "unknown structure".

[0088] By definition, a "known structure" refers to a structure that has been recognized by both the NW and the UE, and that the UE has explicitly indicated its support for. Therefore, when the NW transmits a model to the UE, a "known structure" can be understood as a model structure that is known to be supported by the UE and recognized by both the UE and the NW.

[0089] When the NW transmits a model with a known structure to the UE, the structure of the model should be the same as the structure of the model that has been previously recognized by the NW and the UE.

[0090] The concept of "model recognition" is involved in related technologies. The process of model recognition can be understood as enabling multiple parties to align their understanding of the same model through the interaction of AI / ML model meta information.

[0091] The model information transmission method provided by the embodiment of the present disclosure is described in detail below through some embodiments and their application scenarios in conjunction with the accompanying drawings.

[0092] As shown in FIG2 , an embodiment of the present disclosure provides a model information transmission method, which is applied to a terminal. The method includes:

[0093] Step 201: Send model indication information, where the model indication information is used to indicate the structure of the artificial intelligence or machine learning AI / ML model supported by the terminal.

[0094] In an embodiment of the present disclosure, the above-mentioned model indication information may directly indicate the structure of the AI / ML model supported by the terminal, for example, directly indicating the model structure information of the AI / ML model supported by the terminal, or may indirectly indicate the structure of the AI / ML model supported by the terminal, for example, by indicating the AI / ML model supported by the terminal, to indirectly indicate the structure of the AI / ML model supported by the terminal, for example, by indicating the identifier of the AI / ML model supported by the terminal or the identifier of the AI / ML model structure or by indicating the reference AI / ML model to indirectly indicate the structure of the AI / ML model supported by the terminal.

[0095] By sending model indication information, the network-side device can identify the structure of the AI / ML model supported by the terminal, so that the NW can transmit the model of the AI / ML model structure supported by the UE, reducing the complexity of the UE and making it possible for the NW to transmit the AI / ML model to the UE, allowing the UE to use the AI / ML model trained by the NW and adapted to the NW environment, thereby improving the performance of the communication system.

[0096] In an embodiment of the present disclosure, a terminal transmits model indication information, which indicates the structure of an artificial intelligence or machine learning (AI / ML) model supported by the terminal. This model indication information enables a network device to learn the structure of the AI / ML model supported by the terminal, so that the network device can subsequently transmit the model of the AI / ML model structure supported by the terminal to the terminal, thereby enabling the terminal to use the AI / ML model transmitted by the network device.

[0097] In some embodiments, the model indication information includes at least one of the following:

[0098] The first item: Model identification.

[0099] In some embodiments, the model corresponding to the model identifier is a model that has been identified by the network-side device.

[0100] In some embodiments, there is a correspondence between the model identifier and the model structure.

[0101] As an implementation method, the model structure corresponding to the above-mentioned model identifier is predefined by the protocol, or is determined by other means (such as indicated by the network side device or determined by negotiation between manufacturers), and this disclosure does not make specific limitations on this.

[0102] The second item: meta information of the model, which is the description information of the model, and includes model structure information and / or model structure identifier.

[0103] In some embodiments, the meta information may include other information in addition to the model structure information and / or the model structure identifier.

[0104] In the disclosed embodiments, a terminal may initiate a model recognition process, wherein the model identified by the model recognition process is a model supported by the terminal. Specifically, when performing model recognition, the terminal reports descriptive information related to the model, i.e., the model's metadata, to the network-side device, so that the network-side device can obtain the structure of the AI / ML model supported by the terminal based on the metadata.

[0105] Item 3: Model structure identification.

[0106] In some embodiments, there is a corresponding relationship between the model structure identifier and the model structure.

[0107] In some embodiments, the model structure corresponding to the model structure identifier is predefined by a protocol, or determined by an agreement between manufacturers, or indicated by a network-side device.

[0108] Item 4: Model structure information.

[0109] The model structure information indicates at least one of the following:

[0110] Type, number of layers, number of nodes, connection method between layers, and connection method between nodes.

[0111] The above-mentioned model structure information may indicate at least one of the above-mentioned items in an explicit or implicit manner.

[0112] Item 5: AI / ML reference model.

[0113] In some embodiments, the model structure of the AI / ML reference model is a model structure supported by the terminal, or the AI / ML reference model is a model supported by the terminal.

[0114] By transmitting the AI / ML reference model, the network-side device can determine the structure of the AI / ML reference model supported by the terminal based on the AI / ML reference model.

[0115] In some embodiments, before sending the model indication information, the method further includes:

[0116] Obtaining model candidate information sent by a network-side device, the model candidate information including at least one of a model identifier, metadata, a model structure identifier, and model structure information of at least one candidate AI / ML model, and / or including at least one candidate AI / ML reference model;

[0117] In the model candidate information, the model candidate information corresponding to the AI / ML model supported by the terminal is selected as the model indication information.

[0118] In some embodiments, the above-mentioned model structure identification, model structure information and / or meta-information can also be described as AI / ML model structure candidate information.

[0119] In the embodiment of the present disclosure, the terminal can directly send model indication information to the network side device, or it can first obtain model candidate information sent by the network side device, determine the model indication information based on the model candidate information, and then send the model indication information to the network side device.

[0120] In some embodiments, the model identifier carries model structure information;

[0121] Alternatively, there is a corresponding relationship between the model identifier and the model structure information.

[0122] As an implementation manner, different fields or different domains or different sub-identifications (ID) in the model identification are used to indicate different structural parameters in the model structure information; in some embodiments, there is no overlap between different fields;

[0123] Alternatively, a specific field, a specific domain, or a specific sub-ID in the model identifier is used to indicate a combination of structural parameters in the model structure information.

[0124] For example, field 1 of the model ID indicates the connection mode of the model, field 2 of the model ID indicates the number of layers, and fields 3, 4, 5, ... of the model ID indicate the number of nodes in the 1st, 2nd, 3rd, ... layers.

[0125] For example, the above structural parameter combination can be {Convolutional Neural Network (CNN), convolution kernel size, number of layers, number of nodes per layer, pooling method, ...}, or {Fully Connected Neural Network (FCNN), number of layers, number of nodes in the first layer, number of nodes in the second layer, number of nodes in the third layer, ...}.

[0126] In some embodiments, the model structure identifier carries model structure information;

[0127] Alternatively, there is a corresponding relationship between the model structure identifier and the model structure information.

[0128] As an implementation manner, different fields or different domains or different sub-IDs in the model structure identifier are used to indicate different structural parameters in the model structure information;

[0129] Alternatively, the specific field, specific domain, or specific sub-ID in the model structure identifier is used to indicate a combination of structure parameters in the model structure information.

[0130] Exemplarily, the field X of the model structure identification (ID) indicates the connection mode between layers, and the field X+Y of the model structure ID indicates the connection mode between nodes, where X and Y are integers respectively.

[0131] Exemplarily, the above structural parameter combination can be {CNN, convolution kernel size, number of layers, number of nodes per layer, pooling method, ...}, or, {FCNN, number of layers, number of nodes in the first layer, number of nodes in the second layer, number of nodes in the third layer, ...}.

[0132] In some embodiments, after sending the model indication information, the method further includes:

[0133] Obtain a target AI / ML model transmitted by a network-side device, where the structure of the target AI / ML model is the structure of an AI / ML model supported by the terminal.

[0134] Here, the network-side device transmits the above-mentioned target AI / ML model to the terminal, allowing the UE to use the AI / ML model trained by the NW and adapted to the NW environment, thereby improving the performance of the communication system.

[0135] The model information transmission method disclosed in the present invention is described below with reference to embodiments.

[0136] The present disclosure is mainly applied to the fifth-generation mobile communication technology (5th-Generation, 5G) NR system, which includes network-side equipment and terminal equipment; the present disclosure can also be applied to other systems, as long as the system requires the NW to determine the AI / ML model structure supported by the UE, the solution of the present disclosure can be applied.

[0137] FIG3 illustrates an applicable scenario of the present disclosure. Multiple UEs, including UE1 and UE2, apply for wireless network connection services; network-side devices provide wireless services for them. The network-side devices may be next-generation Node B (gNB), transmission-reception point (TRP), location management function (LMF), etc. The network-side devices and UE1 and UE2 exchange and transmit data through wireless communication. For example, the gNB provides AI / ML-related services to UE1 and UE2, including transmitting AI / ML models to UE1 and UE2. In addition, a third-party server and the gNB can be connected via a wired connection and exchange information, or interact with the gNB through the UE. The third-party server may be an over-the-top (OTT) server or an AI / ML model provider.

[0138] Example 1:

[0139] As shown in FIG4 , the method includes:

[0140] Step 1: The UE reports the identifier of the AI / ML model it supports or reports the meta-information of the AI / ML model it supports to the NW.

[0141] In some embodiments, the UE may report the identifier of at least one AI / ML model supported by it to the NW, or report the meta information of at least one AI / ML model supported by it.

[0142] Implementation method 1: The UE reports the identifier of the supported model (model ID), and the model is the model recognized by the NW.

[0143] In some embodiments, the model ID is an index value that identifies the model. For an identified model, the AI / ML model corresponding to the same model ID is the same.

[0144] An identified AI / ML model may be: a model previously initiated and completed by the UE; a model previously initiated and completed by other UEs; a model previously initiated and completed by a third-party server, etc.

[0145] Implementation method 2: The UE initiates model identification, and the identified model is a model supported by the UE; when performing model identification, the UE reports relevant description information of the model, that is, meta information of the model, to the NW.

[0146] In this implementation, the model identified by the UE initiating model identification is an AI / ML model supported by the UE, so the UE also supports the model structure of the AI / ML model;

[0147] The model's meta-information includes information related to the model structure, see Method 1 in Step 2 for details.

[0148] Step 2: The NW determines the structure of the AI / ML model supported by the UE.

[0149] Specifically, the structure of the AI / ML model supported by the UE is determined by at least one of the following methods:

[0150] Method 1: The meta information of the AI / ML model includes model structure information of the AI / ML model;

[0151] The model structure information is used to determine the structure of the AI / ML model and includes at least one of the following information (partial or complete):

[0152] The type of model structure; for example, fully connected structure, convolutional neural network structure, residual neural network structure, etc.

[0153] Number of layers; for example, the number of layers is 2, 3, 4... layers;

[0154] The number of nodes; for example, the number of nodes common to each layer; or a list of node numbers, the length of the list can be the same as the number of layers, and each item in the list indicates the number of nodes in each layer;

[0155] The connection method between layers; for example, is the Nth layer connected to the N+1th layer, the N+1th layer connected to the N+2th layer, or is the Nth layer connected to the N+1 and N+2th layers, the N+1th layer connected to the N+2 and N+3th layers, etc., where N is an integer;

[0156] The way the nodes are connected; for example, every node in layer N is connected to every node in layer N+m, or the node with index k in layer N is connected to the node with index f in layer N+m. N (k) nodes are connected, where f N (k) is a formula predefined or indicated by NW that can uniquely determine the set of node indices connected to the k-th node in the N-th layer, where m is an integer.

[0157] Method 2: The model ID of the AI / ML model; wherein the model ID partially or fully carries information that can represent the model structure of the AI / ML model;

[0158] Option 1: Different fields / domains / sub-IDs in the ID indicate different model structure parameters; for example:

[0159] Field 1 of the model ID indicates how the model is connected;

[0160] Field 2 of the model ID indicates the number of layers;

[0161] Fields 3, 4, 5, ... of the model ID indicate the number of nodes in the 1st, 2nd, 3rd, ... layers;

[0162] The field X of the model ID indicates the connection method between layers. X is an integer.

[0163] The X+Y field of the model ID indicates the connection method between nodes, where X and Y are both integers;

[0164] Option 2: A specific field / domain / sub-ID in the ID indicates a structural parameter combination used to determine the model structure; an example of a structural parameter combination may be a parameter combination of {structure type, parameter 1, parameter 2, ...}, where the relevant parameters are different for different structure types, for example: {CNN, convolution kernel size, number of layers, number of nodes per layer, pooling method, ...}, or {FCNN, number of layers, number of nodes in the first layer, number of nodes in the second layer, number of nodes in the third layer, ...};

[0165] The model structure corresponding to the above model ID is predefined by the protocol or determined by other means (such as negotiation between manufacturers);

[0166] Method 3: The UE transmits a reference model to the NW, where the model structure of the reference model is a model structure supported by the UE, or the reference model is a model that the UE can support;

[0167] The reference model is used to allow the NW to determine the model structure that the UE can support;

[0168] In some embodiments, the reference model transmitted by the UE to the NW may also meet one or more of the following characteristics:

[0169] Feature 1: The model is an "empty model", for example, it only has hyperparameters but no specific parameter values;

[0170] It should be noted that hyper parameters are parameters whose values ​​are set before starting the learning process, such as the number of layers, number of nodes, learning rate, etc. of the model, while parameters are data obtained through training, such as the weight coefficients multiplied when transferring data between nodes.

[0171] Feature 2: Meta information of the model is also transmitted, but the meta information does not include any of the following indication information, or the meta information includes the following indication information but some or all of the values ​​of these indication information are null, or the indications of these indication information are all or partly invalid / ineffective: performance indication information, usage indication information, applicable condition indication information, etc.;

[0172] Feature 3: Model format information is also transmitted. This information indicates the format of the reference model description, allowing the recipient of the reference model to interpret and understand the reference model. Exemplarily, this format can be a public format or a private format known to the recipient of the reference model. For example, a public format can be ".pb," ".h5," or "ONNX." Model format version information can also be included.

[0173] Step 3: The NW transmits the AI / ML model (i.e., the target AI / ML model) to the UE. The model structure of the AI / ML model is a model structure that can be supported by the UE.

[0174] In some embodiments, before the NW transmits the AI / ML model to the UE, it also initiates model identification from the NW to the UE, where the model identification from the NW to the UE at least indicates the applicable conditions of the AI / ML model transmitted by the NW to the UE.

[0175] Example 2:

[0176] As shown in FIG5 , the method includes:

[0177] Step 1: NW sends AI / ML model candidate information to UE.

[0178] Implementation method 1: The candidate AI / ML model information is meta-information of the candidate AI / ML model, and the meta-information of the model includes model structure information of the AI / ML model;

[0179] The metadata's indication of the model structure is detailed in Example 1.

[0180] Implementation method 2: The model candidate information is the model ID; wherein the model ID partially or fully carries information that can represent the model structure of the AI / ML model;

[0181] The specific indication of the model structure by the model ID is shown in Example 1.

[0182] Implementation 3: The model candidate information is a candidate AI / ML reference model, and the candidate AI / ML reference model is used to allow the UE to determine which models it supports and which model structure candidates it supports.

[0183] For a detailed introduction to the reference model, see Example 1.

[0184] Step 2: The UE feeds back the model candidate information that it can support from the model candidate information sent by the NW, thereby indicating the model structure supported by the UE.

[0185] For example, the model ID in the model candidate information supported by the feedback, or the corresponding index of the model supported by the feedback in the candidate model list;

[0186] Here, it is possible that the UE does not support the model structure corresponding to any model candidate information. In this case, the UE feedbacks that it does not support any model candidate information, thereby indicating that it does not support the model structure corresponding to all model candidate information. In this case, the NW does not perform the following step 3.

[0187] Step 3: The NW transmits the AI / ML model (i.e., the target AI / ML model) to the UE. The model structure of the AI / ML model is a model structure that can be supported by the UE.

[0188] In some embodiments, before the NW transmits the AI / ML model to the UE, it also initiates model identification from the NW to the UE, where the model identification from the NW to the UE at least indicates the applicable conditions of the AI / ML model transmitted by the NW to the UE.

[0189] Example 3:

[0190] As shown in FIG6 , the method includes:

[0191] Step 1: The UE sends AI / ML model structure indication information to the NW, indicating the AI / ML model structure it supports;

[0192] The AI / ML model structure indication information indicates at least one AI / ML model structure supported by the terminal.

[0193] Implementation method 1: The model structure indication information indicates the structural information of the model, at least indicating the model type, the number of layers / nodes, and / or the connection method between layers, and / or the connection method between nodes, for details, refer to embodiment 1);

[0194] Implementation method 2: The model structure indication information is the ID of the AI / ML model structure. The model structure corresponding to the ID is predefined by the protocol, or determined by negotiation between manufacturers, or indicated by the NW. The method for the model structure ID to indicate the model structure is similar to the method for indicating the model structure by "partial field / domain / sub-ID of the model ID" in Example 1, which will not be repeated here.

[0195] Step 2: The NW determines the AI / ML model structure supported by the UE;

[0196] In this embodiment, the NW can determine the model structure supported by the UE according to the model structure indication information in step 1.

[0197] Step 3: The NW transmits the AI / ML model (i.e., the target AI / ML model) to the UE. The model structure of the AI / ML model is a model structure that can be supported by the UE.

[0198] In some embodiments, before the NW transmits the AI / ML model to the UE, it also initiates model identification from the NW to the UE, where the model identification from the NW to the UE at least indicates the applicable conditions of the AI / ML model transmitted by the NW to the UE.

[0199] It should be noted that this embodiment is similar to embodiment 1, except that the information reported by the UE only includes the indication information of the model structure, which reduces the reporting overhead of redundant information; if the meta information of a model includes model structure indication information, it can be considered that the UE does not need to report other information in the meta information except the model structure indication information, or only reports it optionally rather than compulsorily.

[0200] Example 4:

[0201] As shown in FIG7 , the method includes:

[0202] Step 1: NW sends AI / ML model structure candidate information to UE;

[0203] Implementation method 1: The AI / ML model structure candidate information indicates the structure of the AI / ML model, wherein at least the model type, and / or the number of layers / nodes, and / or the connection method between layers, and / or the connection method between nodes are indicated;

[0204] The method for indicating the structure of the AI / ML model by the AI / ML model structure candidate information is specifically referred to the method for indicating the structure of the AI / ML model by the model structure information in the first embodiment;

[0205] Implementation method 2: The model structure candidate information is the ID of the AI / ML model structure. The model structure corresponding to the ID of the model structure is predefined by the protocol, or determined by negotiation between manufacturers, or indicated by the NW; the method for the ID of the model structure to indicate the model structure is similar to the method for indicating the model structure by "field / domain / sub-ID of the model ID" in Example 1, which will not be repeated here.

[0206] In some embodiments, the model structure candidate information includes an AI / ML model structure list.

[0207] Step 2: The UE feeds back the model structure that it can support among the model structures indicated by the candidate model structure information.

[0208] For example, the ID of the candidate model structure supported by the feedback, or the index corresponding to the model structure supported by the feedback in the candidate model structure list.

[0209] It should be noted that it is possible that the UE does not support any candidate model structure. In this case, the UE reports that it does not support any candidate model structure. In this case, the NW does not perform the following step 3.

[0210] Step 3: The NW transmits the AI / ML model (i.e., the target AI / ML model) to the UE. The model structure of the AI / ML model is a model structure that can be supported by the UE.

[0211] In some embodiments, before the NW transmits the AI / ML model to the UE, it also initiates model identification from the NW to the UE, where the model identification from the NW to the UE at least indicates the applicable conditions of the AI / ML model transmitted by the NW to the UE.

[0212] The above-mentioned solution of the embodiment of the present disclosure enables the network-side device to identify the structure of the AI / ML model supported by the terminal by sending model indication information, so that the NW can transmit the model of the AI / ML model structure supported by the UE, reducing the complexity of the UE and making it possible for the NW to transmit the AI / ML model to the UE, allowing the UE to use the AI / ML model trained by the NW and adapted to the NW environment, thereby improving the performance of the communication system.

[0213] As shown in FIG8 , an embodiment of the present disclosure further provides a model information transmission method, which is applied to a network-side device and includes:

[0214] Step 801: Obtain model indication information, where the model indication information is used to indicate the structure of an artificial intelligence or machine learning AI / ML model supported by the terminal;

[0215] In an embodiment of the present disclosure, the above-mentioned model indication information may directly indicate the structure of the AI / ML model supported by the terminal, for example, directly indicating the model structure information of the AI / ML model supported by the terminal, or may indirectly indicate the structure of the AI / ML model supported by the terminal, for example, by indicating the AI / ML model supported by the terminal, to indirectly indicate the structure of the AI / ML model supported by the terminal, for example, by indicating the identifier of the AI / ML model supported by the terminal or the identifier of the AI / ML model structure or by indicating the reference AI / ML model to indirectly indicate the structure of the AI / ML model supported by the terminal.

[0216] Step 802: Determine the structure of the AI / ML model supported by the terminal according to the model indication information.

[0217] Through model indication information, network-side devices can identify the structure of the AI / ML model supported by the terminal, so that the NW can transmit the model of the AI / ML model structure supported by the UE, reducing the complexity of the UE and making it possible for the NW to transmit the AI / ML model to the UE. This allows the UE to use the AI / ML model trained by the NW and adapted to the NW environment, thereby improving the performance of the communication system.

[0218] In some embodiments, the model indication information includes at least one of the following:

[0219] Model identification;

[0220] Meta information of the model, wherein the meta information is descriptive information of the model and includes model structure information and / or model structure identifier;

[0221] Model structure identification;

[0222] Model structure information;

[0223] AI / ML reference model.

[0224] The model indication information has been described in detail in the method embodiment on the terminal side and will not be repeated here.

[0225] In some embodiments, the method of the present disclosure further includes:

[0226] Send model candidate information to the terminal, where the model candidate information includes at least one of a model identifier, metadata, a model structure identifier, and model structure information of at least one candidate AI / ML model, and / or includes at least one candidate AI / ML reference model.

[0227] In an embodiment of the present disclosure, a network-side device sends model candidate information to a terminal, and the terminal selects model candidate information corresponding to the AI / ML model supported by the terminal from the model candidate information as the model indication information, thereby achieving the purpose of enabling the network-side device to obtain the structure of the AI / ML model supported by the terminal through the model indication information.

[0228] In some embodiments, the model structure information indicates at least one of the following:

[0229] Type, number of layers, number of nodes, connection method between layers, and connection method between nodes.

[0230] In some embodiments, the model identifier carries model structure information;

[0231] Alternatively, there is a corresponding relationship between the model identifier and the model structure information.

[0232] In some embodiments, different fields or different domains or different sub-IDs in the model identifier are used to indicate different structural parameters in the model structure information;

[0233] Alternatively, a specific field, a specific domain, or a specific sub-ID in the model identifier is used to indicate a combination of structural parameters in the model structure information.

[0234] In some embodiments, the model structure identifier carries model structure information;

[0235] Alternatively, there is a corresponding relationship between the model structure identifier and the model structure information.

[0236] In some embodiments, different fields or different domains or different sub-IDs in the model structure identifier are used to indicate different structural parameters in the model structure information;

[0237] Alternatively, the specific field, specific domain, or specific sub-ID in the model structure identifier is used to indicate a combination of structure parameters in the model structure information.

[0238] The above-mentioned model identification or model structure identification has been described in detail in the method embodiment on the terminal side and will not be repeated here.

[0239] In some embodiments, the method of the present disclosure further includes:

[0240] A target AI / ML model is sent to the terminal, where the structure of the target AI / ML model is the structure of the AI / ML model supported by the terminal.

[0241] Here, the network-side device transmits the above-mentioned target AI / ML model to the terminal, allowing the UE to use the AI / ML model trained by the NW and adapted to the NW environment, thereby improving the performance of the communication system.

[0242] It should be noted that the method executed by the network side device is a method corresponding to the method executed by the above-mentioned terminal. The specific interaction process between the two has been described in detail in the embodiment of the terminal side and will not be repeated here.

[0243] As shown in FIG9 , an embodiment of the present disclosure provides a model information transmission device, which is applied to a terminal and includes a memory 920 , a transceiver 900 , and a processor 910 ;

[0244] The memory 920 is used to store computer programs; the transceiver 900 is used to send and receive data under the control of the processor 910; the processor 910 is used to read the computer program in the memory 920 and perform the following operations:

[0245] Send model indication information, where the model indication information is used to indicate the structure of the artificial intelligence or machine learning AI / ML model supported by the terminal.

[0246] In FIG9 , the bus architecture may include any number of interconnected buses and bridges, specifically various circuits connected together by one or more processors represented by processor 910 and memory represented by memory 920. The bus architecture may also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides an interface. The transceiver 900 may be a plurality of components, including a transmitter and a receiver, providing a unit for communicating with various other devices on a transmission medium, such as a wireless channel, a wired channel, an optical cable, and the like. For different user devices, the user interface 930 may also be an interface capable of connecting external or internal devices as required, and the connected devices include but are not limited to a keypad, a display, a speaker, a microphone, a joystick, and the like.

[0247] The processor 910 is responsible for managing the bus architecture and general processing, and the memory 920 can store data used by the processor 910 when performing operations.

[0248] In some embodiments, the processor 910 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or a complex programmable logic device (CPLD). The processor may also adopt a multi-core architecture.

[0249] The processor calls the computer program stored in the memory to execute any of the methods provided by the embodiments of the present disclosure according to the obtained executable instructions. The processor and the memory can also be arranged physically separately.

[0250] In some embodiments, the model indication information includes at least one of the following:

[0251] Model identification;

[0252] Meta information of the model, wherein the meta information is descriptive information of the model and includes model structure information and / or model structure identifier;

[0253] Model structure identification;

[0254] Model structure information;

[0255] AI / ML reference model.

[0256] In some embodiments, the processor further implements the following steps:

[0257] Obtaining model candidate information sent by a network-side device, the model candidate information including at least one of a model identifier, metadata, a model structure identifier, and model structure information of at least one candidate AI / ML model, and / or including at least one candidate AI / ML reference model;

[0258] In the model candidate information, the model candidate information corresponding to the AI / ML model supported by the terminal is selected as the model indication information.

[0259] In some embodiments, the model structure information indicates at least one of the following:

[0260] Type, number of layers, number of nodes, connection method between layers, and connection method between nodes.

[0261] In some embodiments, the model identifier carries model structure information;

[0262] Alternatively, there is a corresponding relationship between the model identifier and the model structure information.

[0263] In some embodiments, different fields or different domains or different sub-IDs in the model identifier are used to indicate different structural parameters in the model structure information;

[0264] Alternatively, the specific field, specific domain, or specific sub-ID in the model identifier is used to indicate a combination of structural parameters in the model structure information.

[0265] In some embodiments, the model structure identifier carries model structure information;

[0266] Alternatively, there is a corresponding relationship between the model structure identifier and the model structure information.

[0267] In some embodiments, different fields or different domains or different sub-IDs in the model structure identifier are used to indicate different structural parameters in the model structure information;

[0268] Alternatively, the specific field, specific domain, or specific sub-ID in the model structure identifier is used to indicate a combination of structure parameters in the model structure information.

[0269] In some embodiments, the processor further implements the following steps:

[0270] Obtain a target AI / ML model transmitted by a network-side device, where the structure of the target AI / ML model is the structure of an AI / ML model supported by the terminal.

[0271] It should be noted here that the above-mentioned device provided in the embodiment of the present disclosure can implement all the method steps implemented in the above-mentioned method embodiment applied to the terminal, and can achieve the same technical effect. The parts and beneficial effects of this embodiment that are the same as the method embodiment will not be described in detail here.

[0272] As shown in FIG10 , an embodiment of the present disclosure further provides a model information transmission device, which is applied to a network-side device. The device includes a memory 1020 , a transceiver 1000 , and a processor 1010 ;

[0273] The memory 1020 is used to store computer programs; the transceiver 1000 is used to send and receive data under the control of the processor; the processor 1010 is used to read the computer program in the memory and perform the following operations:

[0274] Obtaining model indication information, where the model indication information is used to indicate the structure of an artificial intelligence or machine learning AI / ML model supported by the terminal;

[0275] Determine the structure of the AI / ML model supported by the terminal according to the model indication information.

[0276] In FIG10 , the bus architecture may include any number of interconnected buses and bridges, specifically linking together various circuits of one or more processors represented by processor 1010 and memory represented by memory 1020. The bus architecture may also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides an interface. The transceiver 1000 may be a plurality of components, namely, a transmitter and a receiver, providing a unit for communicating with various other devices over a transmission medium, such as a wireless channel, a wired channel, an optical cable, or the like. The processor 1010 is responsible for managing the bus architecture and general processing, and the memory 1020 may store data used by the processor 1010 when performing operations.

[0277] The processor 1010 may be a CPU, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or a complex programmable logic device (CPLD). The processor may also adopt a multi-core architecture.

[0278] In some embodiments, the model indication information includes at least one of the following:

[0279] Model identification;

[0280] Meta information of the model, wherein the meta information is descriptive information of the model and includes model structure information and / or model structure identifier;

[0281] Model structure identification;

[0282] Model structure information;

[0283] AI / ML Reference Model:

[0284] In some embodiments, the processor further implements the following steps:

[0285] Send model candidate information to the terminal, where the model candidate information includes at least one of a model identifier, metadata, a model structure identifier, and model structure information of at least one candidate AI / ML model, and / or includes at least one candidate AI / ML reference model.

[0286] In some embodiments, the model structure information indicates at least one of the following:

[0287] Type, number of layers, number of nodes, connection method between layers, and connection method between nodes.

[0288] In some embodiments, the model identifier carries model structure information;

[0289] Alternatively, there is a corresponding relationship between the model identifier and the model structure information.

[0290] In some embodiments, different fields or different domains or different sub-IDs in the model identifier are used to indicate different structural parameters in the model structure information;

[0291] Alternatively, a specific field, a specific domain, or a specific sub-ID in the model identifier is used to indicate a combination of structural parameters in the model structure information.

[0292] In some embodiments, the model structure identifier carries model structure information;

[0293] Alternatively, there is a corresponding relationship between the model structure identifier and the model structure information.

[0294] In some embodiments, different fields or different domains or different sub-IDs in the model structure identifier are used to indicate different structural parameters in the model structure information;

[0295] Alternatively, the specific field, specific domain, or specific sub-ID in the model structure identifier is used to indicate a combination of structure parameters in the model structure information.

[0296] In some embodiments, the processor further implements the following steps:

[0297] A target AI / ML model is sent to the terminal, where the structure of the target AI / ML model is the structure of the AI / ML model supported by the terminal.

[0298] It should be noted here that the above-mentioned device provided in the embodiment of the present disclosure can implement all the method steps implemented in the above-mentioned method embodiment applied to the network side device, and can achieve the same technical effect. The parts and beneficial effects of this embodiment that are the same as the method embodiment will not be described in detail here.

[0299] As shown in FIG11 , an embodiment of the present disclosure further provides a model information transmission device, which is applied to a terminal and includes:

[0300] The first sending unit 1101 is used to send model indication information, where the model indication information is used to indicate the structure of the artificial intelligence or machine learning AI / ML model supported by the terminal.

[0301] In some embodiments, the model indication information includes at least one of the following:

[0302] Model identification;

[0303] Meta information of the model, wherein the meta information is descriptive information of the model and includes model structure information and / or model structure identifier;

[0304] Model structure identification;

[0305] Model structure information;

[0306] AI / ML reference model.

[0307] In some embodiments, the apparatus of the present disclosure further includes:

[0308] a second acquiring unit, configured to acquire, before the first sending unit sends the model indication information, model candidate information sent by the network-side device, the model candidate information including at least one of a model identifier, metadata, a model structure identifier, and model structure information of at least one candidate AI / ML model, and / or including at least one candidate AI / ML reference model;

[0309] A selection unit is configured to select, from the model candidate information, model candidate information corresponding to an AI / ML model supported by the terminal as the model indication information.

[0310] In some embodiments, the model structure information indicates at least one of the following:

[0311] Type, number of layers, number of nodes, connection method between layers, and connection method between nodes.

[0312] In some embodiments, the model identifier carries model structure information;

[0313] Alternatively, there is a corresponding relationship between the model identifier and the model structure information.

[0314] In some embodiments, different fields or different domains or different sub-IDs in the model identifier are used to indicate different structural parameters in the model structure information;

[0315] Alternatively, the specific field, specific domain, or specific sub-ID in the model identifier is used to indicate a combination of structural parameters in the model structure information.

[0316] In some embodiments, the model structure identifier carries model structure information;

[0317] Alternatively, there is a corresponding relationship between the model structure identifier and the model structure information.

[0318] In some embodiments, different fields or different domains or different sub-IDs in the model structure identifier are used to indicate different structural parameters in the model structure information;

[0319] Alternatively, the specific field, specific domain, or specific sub-ID in the model structure identifier is used to indicate a combination of structure parameters in the model structure information.

[0320] In some embodiments, the apparatus of the present disclosure further includes:

[0321] The third acquisition unit is used to obtain the target AI / ML model transmitted by the network side device after the first sending unit sends the model indication information, where the structure of the target AI / ML model is the structure of the AI / ML model supported by the terminal.

[0322] It should be noted here that the above-mentioned device provided in the embodiment of the present disclosure can implement all the method steps implemented in the above-mentioned method embodiment applied to the terminal, and can achieve the same technical effect. The parts and beneficial effects of this embodiment that are the same as the method embodiment will not be described in detail here.

[0323] As shown in FIG12 , an embodiment of the present disclosure further provides a model information transmission device, which is applied to a network-side device. The device includes:

[0324] A first acquiring unit 1201 is configured to acquire model indication information, where the model indication information is used to indicate a structure of an artificial intelligence or machine learning AI / ML model supported by the terminal;

[0325] The determining unit 1202 is configured to determine the structure of the AI / ML model supported by the terminal according to the model indication information.

[0326] In some embodiments, the model indication information includes at least one of the following:

[0327] Model identification;

[0328] Meta information of the model, wherein the meta information is descriptive information of the model and includes model structure information and / or model structure identifier;

[0329] Model structure identification;

[0330] Model structure information;

[0331] AI / ML Reference Model:

[0332] In some embodiments, the apparatus of the present disclosure further includes:

[0333] The second sending unit is used to send model candidate information to the terminal, where the model candidate information includes at least one of a model identifier, metadata, a model structure identifier, and model structure information of at least one candidate AI / ML model, and / or includes at least one candidate AI / ML reference model.

[0334] In some embodiments, the model structure information indicates at least one of the following:

[0335] Type, number of layers, number of nodes, connection method between layers, and connection method between nodes.

[0336] In some embodiments, the model identifier carries model structure information;

[0337] Alternatively, there is a corresponding relationship between the model identifier and the model structure information.

[0338] In some embodiments, different fields or different domains or different sub-IDs in the model identifier are used to indicate different structural parameters in the model structure information;

[0339] Alternatively, the specific field, specific domain, or specific sub-ID in the model identifier is used to indicate a combination of structural parameters in the model structure information.

[0340] In some embodiments, the model structure identifier carries model structure information;

[0341] Alternatively, there is a corresponding relationship between the model structure identifier and the model structure information.

[0342] In some embodiments, different fields or different domains or different sub-IDs in the model structure identifier are used to indicate different structural parameters in the model structure information;

[0343] Alternatively, the specific field, specific domain, or specific sub-ID in the model structure identifier is used to indicate a combination of structure parameters in the model structure information.

[0344] In some embodiments, the apparatus of the present disclosure further includes:

[0345] The third sending unit is configured to send a target AI / ML model to the terminal, where the structure of the target AI / ML model is the structure of an AI / ML model supported by the terminal.

[0346] It should be noted here that the above-mentioned device provided in the embodiment of the present disclosure can implement all the method steps implemented in the above-mentioned method embodiment applied to the network side device, and can achieve the same technical effect. The parts and beneficial effects of this embodiment that are the same as the method embodiment will not be described in detail here.

[0347] It should be noted that the division of units in the embodiments of the present disclosure is schematic and is merely a logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional units in the various embodiments of the present disclosure may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0348] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a processor-readable storage medium. Based on this understanding, the technical solution of the present disclosure is essentially or the part that contributes to the relevant technology or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present disclosure. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0349] In some embodiments of the present disclosure, a processor-readable storage medium is also provided, which stores program instructions. The program instructions are used to enable the processor to execute all the steps implemented by the method embodiment for implementing the above-mentioned terminal execution or all the steps implemented by the method embodiment for implementing the network side device, and can achieve the same technical effect. The parts and beneficial effects of this embodiment that are the same as the method embodiment will not be described in detail here.

[0350] The embodiment of the present disclosure also provides a computer program product, including computer instructions. When the computer instructions are executed by a processor, the various processes of the method embodiment shown in Figure 2 or Figure 8 above are implemented, and the same technical effect can be achieved. To avoid repetition, they will not be repeated here.

[0351] The terminal device involved in the embodiments of the present disclosure may be a device that provides voice and / or data connectivity to a user, a handheld device with wireless connection function, or other processing devices connected to a wireless modem. In different systems, the name of the terminal device may also be different. For example, in the fifth-generation mobile communication technology (5th-Generation, 5G) system, the terminal device may be called User Equipment (UE). A wireless terminal device can communicate with one or more core networks (CN) via a radio access network (RAN). The wireless terminal device can be a mobile terminal device, such as a mobile phone (or "cellular" phone) and a computer with a mobile terminal device. For example, it can be a portable, pocket-sized, handheld, computer-built-in or vehicle-mounted mobile device that exchanges voice and / or data with the radio access network. For example, personal communication service (PCS) phones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, personal digital assistants (PDAs), and other devices. The wireless terminal device may also be referred to as a system, a subscriber unit, a subscriber station, a mobile station, a mobile station, a remote station, an access point, a remote terminal device, an access terminal device, a user terminal device, a user agent, or a user device, but is not limited in the embodiments of the present disclosure.

[0352] The network device (or network-side device) involved in the embodiments of the present disclosure may be a base station, which may include multiple cells providing services to terminals. Depending on the specific application scenario, the base station may also be called an access point, or may be a device in an access network that communicates with a wireless terminal device through one or more sectors on an air interface, or may be called another name. The network device may be used to interchange received air frames with Internet Protocol (IP) packets, acting as a router between the wireless terminal device and the rest of the access network, wherein the rest of the access network may include an Internet Protocol (IP) communication network. The network device may also coordinate attribute management of the air interface. For example, the network device involved in the embodiments of the present disclosure may be a base transceiver station (BTS) in the Global System for Mobile communications (GSM) or code division multiple access (CDMA), a network device (NodeB) in wide-band code division multiple access (WCDMA), an evolutionary Node B (eNB or e-NodeB) in the Long Term Evolution (LTE) system, a 5G base station (gNB) in the 5G network architecture (next generation system), a home evolved Node B (HeNB), a relay node, a femto, a pico, etc., and is not limited in the embodiments of the present disclosure. In some network structures, the network device may include a centralized unit (CU) node and a distributed unit (DU) node, and the centralized unit and the distributed unit may also be geographically separated.

[0353] Network devices and terminal devices can each use one or more antennas for Multiple Input Multiple Output (MIMO) transmission. MIMO transmission can be single-user MIMO (SU-MIMO) or multi-user MIMO (MU-MIMO). Depending on the configuration and number of antenna combinations, MIMO transmission can be two-dimensional MIMO (2D-MIMO), three-dimensional MIMO (3D-MIMO), full-dimensional MIMO (FD-MIMO), or massive MIMO. It can also use diversity transmission, precoding, or beamforming.

[0354] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, systems, or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage and optical storage, etc.) containing computer-usable program code.

[0355] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer-executable instructions. These computer-executable instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for implementing the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0356] These processor-executable instructions may also be stored in a processor-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the processor-readable memory produce a product including an instruction device that implements the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0357] These processor-executable instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0358] In addition, it should be noted that, in the apparatus and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. Moreover, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order, and some steps can be performed in parallel or independently of each other. For those of ordinary skill in the art, it will be understood that all or any steps or components of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or a network of computing devices in hardware, firmware, software or a combination thereof, which can be achieved by those of ordinary skill in the art using their basic programming skills after reading the description of the present invention.

[0359] It should be noted that it should be understood that the division of the above modules is merely a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. Moreover, these modules can all be implemented in the form of software called by a processing element; or they can all be implemented in the form of hardware; or some modules can be implemented in the form of software called by a processing element, and some modules can be implemented in the form of hardware. For example, a module can be a separately established processing element, or it can be integrated into a chip of the above-mentioned device. In addition, it can also be stored in the memory of the above-mentioned device in the form of program code, and called by a processing element of the above-mentioned device to perform the functions of the above-mentioned module. The implementation of other modules is similar. In addition, these modules can all or partly be integrated together, or they can be implemented independently. The processing element described here can be an integrated circuit with signal processing capabilities. During implementation, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in the processor element or instructions in the form of software.

[0360] For example, each module, unit, sub-unit or sub-module can be one or more integrated circuits configured to implement the above method, such as one or more application-specific integrated circuits (ASICs), one or more digital signal processors (DSPs), or one or more field programmable gate arrays (FPGAs). For another example, when a module is implemented by scheduling program code through a processing element, the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processor that can call program code. For another example, these modules can be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0361] The terms "first," "second," and the like in the specification and claims of the present disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of the present disclosure described herein may be implemented in a sequence other than that illustrated or described herein. In addition, the terms "including" and "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or device that includes a series of steps or units need not be limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to these processes, methods, products, or devices. In addition, the use of "and / or" in the specification and claims to indicate at least one of the connected objects, for example, A and / or B and / or C, means that seven situations are included: A alone, B alone, C alone, both A and B present, both B and C present, both A and C present, and all A, B, and C present. Similarly, the use of "at least one of A and B" in the specification and claims should be understood to mean "A alone, B alone, or both A and B present."

[0362] Obviously, those skilled in the art may make various changes and modifications to the present disclosure without departing from the spirit and scope of the present disclosure. Thus, if these modifications and variations of the present disclosure fall within the scope of the claims of the present disclosure and their equivalents, the present disclosure is intended to include these modifications and variations.

Claims

1. A model information transmission method, applied to a terminal, comprising: Send model indication information, where the model indication information is used to indicate the structure of the artificial intelligence or machine learning AI / ML model supported by the terminal.

2. The method according to claim 1, wherein The model indication information includes at least one of the following: Model identification; Meta information of the model, wherein the meta information is descriptive information of the model and includes model structure information and / or model structure identifier; Model structure identification; Model structure information; AI / ML reference model.

3. The method according to claim 1, wherein Before sending the model indication information, the method further includes: Obtaining model candidate information sent by a network-side device, the model candidate information including at least one of a model identifier, metadata, a model structure identifier, and model structure information of at least one candidate AI / ML model, and / or including at least one candidate AI / ML reference model; In the model candidate information, the model candidate information corresponding to the AI / ML model supported by the terminal is selected as the model indication information.

4. The method according to claim 2 or 3, wherein: The model structure information indicates at least one of the following: Type, number of layers, number of nodes, connection method between layers, and connection method between nodes.

5. The method according to claim 2 or 3, wherein: The model identifier carries model structure information; Alternatively, there is a corresponding relationship between the model identifier and the model structure information.

6. The method according to claim 5, wherein: Different fields or domains or sub-IDs in the model identifier are used to indicate different structural parameters in the model structure information; Alternatively, a specific field, a specific domain, or a specific sub-ID in the model identifier is used to indicate a combination of structural parameters in the model structure information.

7. The method according to claim 2 or 3, wherein: The model structure identifier carries model structure information; Alternatively, there is a corresponding relationship between the model structure identifier and the model structure information.

8. The method according to claim 7, wherein: Different fields or domains or sub-IDs in the model structure identifier are used to indicate different structural parameters in the model structure information; Alternatively, the specific field, specific domain, or specific sub-ID in the model structure identifier is used to indicate a combination of structure parameters in the model structure information.

9. The method according to claim 1, wherein After sending the model indication information, the method further includes: Obtain a target AI / ML model transmitted by a network-side device, where the structure of the target AI / ML model is the structure of an AI / ML model supported by the terminal.

10. A model information transmission method, applied to a network-side device, comprising: Obtaining model indication information, where the model indication information is used to indicate the structure of an artificial intelligence or machine learning AI / ML model supported by the terminal; Determine the structure of the AI / ML model supported by the terminal according to the model indication information.

11. The method according to claim 10, wherein: The model indication information includes at least one of the following: Model identification; Meta information of the model, wherein the meta information is descriptive information of the model and includes model structure information and / or model structure identifier; Model structure identification; Model structure information; AI / ML reference model.

12. The method according to claim 10, further comprising: Send model candidate information to the terminal, where the model candidate information includes at least one of a model identifier, metadata, a model structure identifier, and model structure information of at least one candidate AI / ML model, and / or includes at least one candidate AI / ML reference model.

13. The method according to claim 11 or 12, wherein: The model structure information indicates at least one of the following: Type, number of layers, number of nodes, connection method between layers, and connection method between nodes.

14. The method according to claim 11 or 12, wherein: The model identifier carries model structure information; Alternatively, there is a corresponding relationship between the model identifier and the model structure information.

15. The method according to claim 14, wherein Different fields or domains or sub-IDs in the model identifier are used to indicate different structural parameters in the model structure information; Alternatively, a specific field, a specific domain, or a specific sub-ID in the model identifier is used to indicate a combination of structural parameters in the model structure information.

16. The method according to claim 11 or 12, wherein: The model structure identifier carries model structure information; Alternatively, there is a corresponding relationship between the model structure identifier and the model structure information.

17. The method according to claim 16, wherein Different fields or domains or sub-IDs in the model structure identifier are used to indicate different structural parameters in the model structure information; Alternatively, the specific field, specific domain, or specific sub-ID in the model structure identifier is used to indicate a combination of structure parameters in the model structure information.

18. The method according to claim 10, further comprising: A target AI / ML model is sent to the terminal, where the structure of the target AI / ML model is the structure of the AI / ML model supported by the terminal.

19. A model information transmission device comprising a memory, a transceiver, and a processor; Memory for storing computer programs; a transceiver, configured to transmit and receive data under the control of the processor; A processor is configured to read the computer program in the memory and perform the following operations: Send model indication information, where the model indication information is used to indicate the structure of the artificial intelligence or machine learning AI / ML model supported by the terminal.

20. A model information transmission device, comprising a memory, a transceiver, and a processor; Memory for storing computer programs; a transceiver, configured to transmit and receive data under the control of the processor; A processor is configured to read the computer program in the memory and perform the following operations: Obtaining model indication information, where the model indication information is used to indicate the structure of an artificial intelligence or machine learning AI / ML model supported by the terminal; Determine the structure of the AI / ML model supported by the terminal according to the model indication information.

21. A model information transmission device, wherein: The device comprises: The first sending unit is used to send model indication information, where the model indication information is used to indicate the structure of the artificial intelligence or machine learning AI / ML model supported by the terminal.

22. A model information transmission device, wherein: The device comprises: A first acquiring unit is configured to acquire model indication information, where the model indication information is used to indicate a structure of an artificial intelligence or machine learning AI / ML model supported by the terminal; A determining unit is configured to determine a structure of an AI / ML model supported by the terminal according to the model indication information.

23. A processor-readable storage medium, wherein the processor-readable storage medium stores a computer program, wherein the computer program is used to enable the processor to execute the steps of the model information transmission method according to any one of claims 1 to 9, or to execute the steps of the model information transmission method according to any one of claims 10 to 18.

24. A computer program product, comprising computer instructions, which, when executed by a processor, implement the steps of the model information transmission method according to any one of claims 1 to 9, or execute the steps of the model information transmission method according to any one of claims 10 to 18.

Images