A communication method and related apparatus

Abstract

A communication method and related apparatus, in the method, in a process that a first communication device transmits first data based on a first multiple access manner, after the first communication device processes first initial data corresponding to a first node and second initial data corresponding to a second node based on a first artificial intelligence (AI) model to obtain the first data, the first communication device can transmit the first data to the first node and the second node through partially or wholly same resources. In this way, the first communication device can use the same AI model to process initial data of different nodes to obtain and transmit first data through the first multiple access manner, can reuse resources to implement transmission of AI data of different nodes, to improve resource utilization, and at the same time, can reuse the AI model to process data of different nodes, to improve model processing efficiency.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a communication method and related apparatus. Background Technology

[0002] In communication systems, communication nodes typically possess both signal transmission and reception capabilities and computational capabilities. Taking network devices with computational capabilities as an example, the computational capabilities of network devices primarily provide computing power support for signal transmission and reception capabilities, enabling communication between network devices and other communication nodes.

[0003] In addition to processing communication signals within the communication network, communication nodes may also need to handle artificial intelligence (AI) related processing. Therefore, how to integrate AI-related processing with the communication network is a pressing technical problem that needs to be solved. Summary of the Invention

[0004] This application provides a communication method and related apparatus for reusing resources in a communication network to transmit AI data from different nodes, thereby improving resource utilization and enabling the reuse of AI models to process data from different nodes, thus improving model processing efficiency.

[0005] The first aspect of this application provides a communication method applied to a first communication device.

[0006] For example, the first communication device may be a terminal, or it may be a communication module and / or computing module within the terminal, or it may be a circuit or chip within the terminal responsible for communication functions (such as a modem chip, also known as a baseband chip, or a system-on-chip (SoC) chip containing a modem core, or a system-in-package (SIP) chip), or a circuit or chip within the terminal responsible for communication and / or computing functions (such as a graphics processing unit (GPU), artificial intelligence (AI) processor, neural network processing unit (NPU), or application-specific integrated circuit (ASIC)). Alternatively, the method may be executed by a logical node, logical module, or software capable of implementing all or part of the terminal's functions.

[0007] For example, the first communication device may be a network-side device, or it may be a module (e.g., a circuit, chip, or chip system) within a network-side device. Alternatively, the first communication device may be a circuit or chip (e.g., a GPU, AI processor, NPU, or ASIC) within a network-side device responsible for communication and / or computing functions. Alternatively, the method may be executed by a logical node, logical module, or software capable of implementing all or part of the functions of the network-side device. The network-side device may include access network equipment, core network equipment, a server, or other equipment defined in the future network definition.

[0008] In the following description, the method of the first aspect is taken as an example, which is performed by the first communication device.

[0009] In this method, a first communication device sends first information, which is used to indicate a first multiple access mode; the first communication device sends first data based on the first multiple access mode; the first communication device sending the first data based on the first multiple access mode includes: the first communication device sending the first data to a first node and a second node through some or all of the same resources; wherein, the first data is obtained by processing first initial data and second initial data through a first AI model, the first initial data corresponds to the first node, and the second initial data corresponds to the second node.

[0010] Based on the above scheme, the first information sent by the first communication device is used to indicate a first multiple access method, and the first communication device sends first data based on the first multiple access method, enabling the receiver of the first data to receive the first data based on the first multiple access method. Specifically, during the process of the first communication device sending the first data based on the first multiple access method, after the first communication device processes the first initial data corresponding to the first node and the second initial data corresponding to the second node using a first AI model to obtain the first data, the first communication device can send the first data to the first node and the second node using some or all of the same resources. In this way, the first communication device can use the same AI model to process the initial data of different nodes using the first multiple access method to obtain and send the first data, thereby reusing resources to achieve the transmission of AI data from different nodes, improving resource utilization, and also reusing the AI ​​model to process data from different nodes, thus improving model processing efficiency.

[0011] It should be noted that the first communication device can send the first data to the first node and the second node using some or all of the same resources, and such resources can be implemented in various ways. For example, the resources may include one or more of time domain resources, frequency domain resources, spatial domain resources, and code domain resources.

[0012] It should be noted that the first communication device can process the first initial data and the second initial data in various ways to obtain the first data.

[0013] For example, the first communication device can use the first initial data and the second initial data as input to the first AI model, and after processing by the first AI model, obtain the first data.

[0014] For example, the first communication device can use the first initial data as input to the first AI model, and after processing by the first AI model, obtain the first result; the first communication device can use the second initial data as input to the first AI model, and after processing by the first AI model, obtain the second result; thereafter, the first communication device processes (e.g., adds) the first result and the second result to obtain the first data.

[0015] It should be noted that the first initial data corresponds to the first node, which can be understood as: the first node processes the first data to obtain the first processing result, that is, the first processing result is the processing result corresponding to the first initial data; in other words, the destination node of the first initial data is the first node. This first node can deploy a second AI model to process the first data. For example, the first initial data is processed sequentially by the first model and then the second model (this processing can be inference, prediction, etc.) to obtain the first processing result. Accordingly, in the above process, after receiving the first data obtained based on the first initial data, the first node can use this first data as input to the second AI model, and after processing by the second AI model, obtain the first processing result.

[0016] Similarly, the second initial data corresponds to the second node, which can be understood as follows: the second node processes the first data to obtain a second processing result, that is, the second processing result is the processing result corresponding to the second initial data; in other words, the destination node of the second initial data is the second node. This second node can deploy a third AI model to process the first data. For example, the second initial data is processed sequentially by the first model and then the second model (this processing can be inference, prediction, etc.) to obtain the second processing result. Accordingly, in the above process, after receiving the first data obtained based on the second initial data, the second node can use the first data as input to the third AI model, process it, and obtain the second processing result.

[0017] Optionally, in the above implementation process based on the first multiple access method, the first AI model can obtain first data based on the first initial data and the second initial data. Furthermore, the first data can be processed by the second AI model to obtain a first processing result, and the first data can also be processed by the third AI model to obtain a second processing result. In this process, the first AI model can be understood as a common model, while the second and third AI models can be understood as personalized models. This method utilizes the characteristics of low-rank adaptation (LoRA), therefore, this first multiple access method can also be called the LoRA multiple access method, or other names defined in future network definitions.

[0018] It should be understood that the processing of input data by the AI ​​model may include a preprocessing process, or the input data of the AI ​​model may be preprocessed data. In the above implementation based on the first multiple access method, the first data is obtained by the first AI model processing the first initial data and the second initial data, that is, the first AI model supports the processing of the first initial data and the second initial data. Therefore, the processing result of these two initial data after preprocessing conforms to the input requirements of the same AI model (i.e., the first AI model), enabling the first AI model to process the first initial data and the second initial data, reducing or avoiding situations where the first AI model cannot process the data, and improving processing performance. For example, the above input requirements may include requirements for at least one of type, format, dimension, or modality.

[0019] It should be noted that the first data can be obtained by processing at least two initial data points through the first AI model. These at least two initial data points can include the first initial data point corresponding to the first node and the second initial data point corresponding to the second node. Furthermore, these at least two initial data points can also include one or more initial data points corresponding to other nodes. The implementation process of these other nodes and their corresponding initial data points can refer to the implementation process of the first or second node.

[0020] Optionally, in this application, the data sent by the first communication device (such as the first data mentioned above, the second data that may appear later, etc.) can be the input data of the AI ​​model, so that such data can be used as the input of the model, enabling communication between AI models and simplifying the data processing process. For example, the first data includes one or more of the following: one or more tokens, one or more token vectors, one or more patches, one or more embeddings, one or more patch embeddings, or one or more token embeddings, etc.

[0021] In one possible implementation of the first aspect, the method further includes: the first communication device receiving second information; and the first communication device generating the first information based on the second information.

[0022] Based on the above scheme, the first communication device can also receive second information and generate first information based on the second information. The second information may be information associated with the first node and / or the second node. Therefore, the first communication device can instruct the two nodes to communicate using a first multiple access method based on the relevant information of the first node and / or the second node, so that the multiple access method used by these nodes during communication can be adapted to the relevant information of these nodes.

[0023] In one possible implementation of the first aspect, the first information may be generated based on the second information, which is used to indicate the model processing capabilities of the first node and the model processing capabilities of the second node.

[0024] Based on the above scheme, the first information used to indicate the first multiple access method can be generated based on the second information, which indicates the model processing capabilities of the first node and the second node. Therefore, the first communication device can instruct the two nodes to communicate using the first multiple access method based on their model processing capabilities, ensuring that the multiple access method used by these nodes during communication is compatible with their model processing information, thereby improving model processing efficiency.

[0025] Optionally, model processing capabilities may include one or more of computing power, storage capacity, or idle computing power.

[0026] In one possible implementation of the first aspect, the method further includes: a first communication device determining a second AI model corresponding to the first node and a third AI model corresponding to the second node based on the second information; wherein the first AI model has a mapping relationship with the second AI model and the third AI model.

[0027] For example, the first communication device can determine a second AI model adapted to the model processing capability of the first node based on second information (or the model processing capability of the first node indicated by the second information). That is, the first communication device can determine the second AI model that the first node supports for deployment based on the second information (or the model processing capability of the first node indicated by the second information). Similarly, the first communication device can determine a third AI model adapted to the model processing capability of the second node based on the second information (or the model processing capability of the second node indicated by the second information). That is, the first communication device can determine the third AI model that the second node supports for deployment based on the second information (or the model processing capability of the second node indicated by the second information).

[0028] Based on the above scheme, the first communication device can determine the second AI model corresponding to the first node and the third AI model corresponding to the second node based on the second information. Furthermore, the first communication device can determine whether these two AI models have a mapping relationship with the first AI model based on one or more mapping relationships. If the first AI model has a mapping relationship with the second AI model and the third AI model, the first communication device can determine that these three AI models can perform data processing in the above manner; that is, the first communication device generates first information to provide an indication of the first multiple access method.

[0029] Optionally, after the first communication device determines the second AI model corresponding to the first node and the third AI model corresponding to the second node based on the second information, the first communication device can also deploy AI models to the first node and / or the second node. For example, the first communication device sends model information of the second AI model to the first node, and / or the first communication device sends model information of the third AI model to the second node.

[0030] In one possible implementation of the first aspect, the second information is used to indicate the second AI model deployed by the first node and the third AI model deployed by the second node; wherein the first AI model has a mapping relationship with the second AI model and the third AI model.

[0031] Based on the above scheme, the first communication device can determine the second AI model and the third AI model deployed by the first node based on the second information. Furthermore, the first communication device can determine whether these two AI models have a mapping relationship with the first AI model based on one or more mapping relationships. If the first AI model has a mapping relationship with the second AI model and the third AI model, the first communication device can determine that these three AI models can perform data processing in the above manner; that is, the first communication device generates first information to provide an indication of the first multiple access method.

[0032] In one possible implementation of the first aspect, the method further includes: the first communication device transmitting second data based on the first multiple access method; transmitting the second data based on the first multiple access method includes: transmitting the second data to the first node and the second node through some or all of the same resources; wherein the second data is obtained by processing the first test data and the second test data through the first AI model, the first test data corresponding to the first node, and the second test data corresponding to the second node; wherein the first information may be generated based on the second information, the second information being used to determine the processing performance of the first node on the second data and the processing performance of the second node on the second data.

[0033] Based on the above scheme, the first information used to indicate the first multiple access method can be generated based on the second information, which indicates the processing performance of the first node and the second node on the test data. Therefore, the first communication device can instruct the two nodes to communicate using the first multiple access method based on their processing performance on the test data, ensuring that the multiple access method used by these nodes during communication is compatible with their testing performance, thereby improving model processing efficiency.

[0034] Optionally, the second information can be implemented in various ways. For example, the second information can indicate the processing performance of the first node on the second data and the processing performance of the second node on the second data. Alternatively, the second information can indicate the processing result of the first node on the second data and the processing result of the second node on the second data, enabling the recipient of the second information to determine the corresponding performance based on these processing results.

[0035] Optionally, the second information may further indicate the model processing capabilities of the first node and the second node, or the second information may further indicate the second AI model deployed by the first node and the third AI model deployed by the second node. In this way, the first communication device can determine whether the two AI models have a mapping relationship with the first AI model based on one or more mapping relationships. If there is no mapping relationship between the first AI model and at least one of the two AI models (the second AI model and the third AI model), the first communication device can send second data based on the first multiple access method to achieve the aforementioned test verification process through the test process corresponding to the second data.

[0036] In one possible implementation of the first aspect, the second information is further used to determine that there is a mapping relationship between the first AI model and the second AI model deployed by the first node and the third AI model deployed by the second node.

[0037] Based on the above scheme, when the second information is used to determine the processing performance of the first node and the second node on the second data, the first communication device can also determine the mapping relationship between different models based on the second information. For example, the first communication device can determine that the two nodes can communicate through the first multiple access method based on the processing performance of the first node and the second node on the test data. Correspondingly, there is a mapping relationship between the first AI model used to obtain the second data, the second AI model of the first node processing the second data, and the third AI model of the second node processing the second data, enabling the first communication device to maintain or save this mapping relationship. Subsequently, it can instruct the first multiple access method based on one or more maintained or saved mapping relationships, thereby improving the instruction efficiency of the multiple access method and reducing processing complexity.

[0038] In one possible implementation of the first aspect, the first AI model is mapped to the second AI model deployed on the first node and the third AI model deployed on the second node.

[0039] Based on the above scheme, the first communication device can determine one or more mapping relationships through pre-configuration or configuration. One of the mapping relationships is the mapping relationship between the first AI model and the second AI model deployed on the first node and the third AI model deployed on the second node. This enables the first communication device to indicate the first multiple access mode based on the one or more mapping relationships, thereby improving the indication efficiency of the multiple access mode and reducing the processing complexity.

[0040] Optionally, the first AI model has a mapping relationship with the first node, the second AI model deployed on the first node, the second node, and the third AI model deployed on the second node.

[0041] Optionally, the first AI model can be a public model, and the second and third AI models can be personalized models. Therefore, any one of the one or more mapping relationships involved in this application can be understood as a mapping relationship between a public model and one or more personalized models. That is, the public model indicated by the mapping relationship can process one or more initial data corresponding to the one or more personalized models to obtain processing results, and the one or more personalized models can process the processing results to obtain their respective personalized outputs.

[0042] A second aspect of this application provides a communication method applied to a second communication device, which is either a first node or a second node.

[0043] For example, the second communication device can be a terminal, or it can be a communication module and / or computing module in the terminal, or it can be a circuit or chip in the terminal responsible for communication functions (such as a modem chip, also known as a baseband chip, or a SoC chip or SIP chip containing a modem core), or a circuit or chip in the terminal responsible for communication and / or computing functions (such as a GPU, NPU, AI processor, or ASIC). Alternatively, the method can be executed by a logical node, logical module, or software that can implement all or part of the terminal functions through the second communication device.

[0044] For example, the second communication device can be a network-side device, or it can be a module (e.g., a circuit, chip, or chip system) within the network-side device, or it can be a circuit or chip (e.g., a GPU, NPU, AI processor, or ASIC) within the network-side device responsible for communication and / or computing functions. Alternatively, the method can be executed by a logical node, logical module, or software capable of implementing all or part of the functions of the network-side device. The network-side device may include access network equipment, core network equipment, a server, or other equipment defined in the future network.

[0045] In the following description, the method of the second aspect is taken as an example, which is performed by the second communication device.

[0046] In this method, a second communication device receives first information, which is used to indicate a first multiple access mode; the second communication device receives first data based on the first multiple access mode.

[0047] For example, the first communication device can send the first data based on the first multiple access method, that is, the first communication device can send the first data to the first node and the second node through some or all of the same resources; wherein, the first data is obtained by processing the first initial data and the second initial data through the first AI model, the first initial data corresponds to the first node, and the second initial data corresponds to the second node.

[0048] In the above process, the second communication device can be the first node. Accordingly, the second communication device can receive the first data based on the first multiple access method and process the first data based on the first multiple access method to obtain the processing result corresponding to the first initial data. Alternatively, the second communication device can be the second node. Accordingly, the second communication device can receive the first data based on the first multiple access method and process the first data based on the first multiple access method to obtain the processing result corresponding to the second initial data.

[0049] Based on the above scheme, the first information received by the second communication device is used to indicate the first multiple access method, and the second communication device receives the first data based on the first multiple access method, enabling the second communication device to receive the first data based on the first multiple access method. Specifically, during the process of the first communication device sending the first data based on the first multiple access method, the second communication device can process the first initial data corresponding to the first node and the second initial data corresponding to the second node based on the first AI model to obtain the first data. Then, the first communication device can send the first data to the first node and the second node using some or all of the same resources. In this way, the first communication device can use the same AI model to process the initial data of different nodes using the first multiple access method to obtain and send the first data. This allows for resource reuse to achieve the transmission of AI data from different nodes, improving resource utilization and also enabling the reuse of the AI ​​model to process data from different nodes, thereby improving model processing efficiency.

[0050] Optionally, in this application, the data received by the second communication device (such as the first data mentioned above, the second data that may appear later, etc.) can be the input data of the AI ​​model, so that such data can be used as the input of the model, enabling communication between AI models and simplifying the data processing process. For example, the first data includes one or more of the following: one or more tokens, one or more token vectors, one or more patches, one or more embeddings, one or more patch embeddings, or one or more token embeddings, etc.

[0051] In one possible implementation of the second aspect, the method further includes: the second communication device transmitting partial second information, the second information being used to generate or determine the first information.

[0052] Based on the above scheme, after the second communication device sends the second information, the first communication device can receive the second information and generate first information based on it. The second information can be information associated with the first node and / or the second node. For example, if the second communication device is the first node, part of the second information sent by the second communication device is associated with the first node; or, if the second communication device is the second node, part of the second information sent by the second communication device is associated with the second node. Therefore, the second communication device can instruct the two nodes to communicate using a first multiple access method based on the relevant information of the first node and / or the second node, so that the multiple access method used by these nodes during communication can be adapted to the relevant information of these nodes.

[0053] In one possible implementation of the second aspect, the first information may be generated based on the second information, which is used to indicate the model processing capabilities of the first node and the second node. For example, if the second communication device is the first node, a portion of the second information sent by the second communication device indicates the model processing capabilities of the first node; or, if the second communication device is the second node, a portion of the second information sent by the second communication device indicates the model processing capabilities of the second node.

[0054] Based on the above scheme, the first information used to indicate the first multiple access method can be generated based on the second information, which indicates the model processing capabilities of the first node and the second node. Therefore, the first communication device can instruct the two nodes to communicate using the first multiple access method based on their model processing capabilities, ensuring that the multiple access method used by these nodes during communication is compatible with their model processing information, thereby improving model processing efficiency.

[0055] Optionally, model processing capabilities may include one or more of computing power, storage capacity, or idle computing power.

[0056] For example, upon receiving second information, the first communication device can determine a second AI model adapted to the model processing capabilities of the first node based on the second information (or the model processing capabilities of the first node indicated by the second information). In other words, the first communication device can determine the second AI model that the first node supports for deployment based on the second information (or the model processing capabilities of the first node indicated by the second information). Similarly, the first communication device can determine a third AI model adapted to the model processing capabilities of the second node based on the second information (or the model processing capabilities of the second node indicated by the second information). In other words, the first communication device can determine the third AI model that the second node supports for deployment based on the second information (or the model processing capabilities of the second node indicated by the second information).

[0057] Optionally, after the first communication device determines the second AI model corresponding to the first node and the third AI model corresponding to the second node based on the second information, the first communication device can also deploy the AI ​​models to the first node and / or the second node. For example, if the second communication device is the first node, the first communication device sends the model information of the second AI model to the first node. Similarly, if the second communication device is the second node, the first communication device sends the model information of the third AI model to the second node.

[0058] In one possible implementation of the second aspect, the second information is used to indicate a second AI model deployed by the first node and a third AI model deployed by the second node; wherein the first AI model has a mapping relationship with the second AI model and the third AI model. For example, if the second communication device is the first node, part of the second information sent by the second communication device indicates the second AI model deployed by the first node; or, if the second communication device is the second node, part of the second information sent by the second communication device indicates the third AI model deployed by the second node.

[0059] Based on the above scheme, the first communication device can determine the second AI model and the third AI model deployed by the first node based on the second information. Furthermore, the first communication device can determine whether these two AI models have a mapping relationship with the first AI model based on one or more mapping relationships. If the first AI model has a mapping relationship with the second AI model and the third AI model, the first communication device can determine that these three AI models can perform data processing in the above manner; that is, the first communication device generates first information to provide an indication of the first multiple access method.

[0060] In one possible implementation of the second aspect, the method further includes: the second communication device receiving second data based on the first multiple access method.

[0061] For example, the first communication device sends the second data through the first multiple access method, that is, the first communication device sends the second data to the first node and the second node through some or all of the same resources; wherein, the second data is obtained by processing the first test data and the second test data through the first AI model, the first test data corresponds to the first node, and the second test data corresponds to the second node.

[0062] In the above process, the second communication device can be the first node. Accordingly, the second communication device can receive and process the second data based on the first multiple access method to obtain the processing result corresponding to the first test data. Alternatively, the second communication device can be the second node. Similarly, the second communication device can receive and process the second data based on the first multiple access method to obtain the processing result corresponding to the second test data. The first information can be generated based on the second information, which is used to determine the processing performance of the first node and the second node on the second data.

[0063] Based on the above scheme, the first information used to indicate the first multiple access method can be generated based on the second information, which indicates the processing performance of the first node and the second node on the test data. For example, if the second communication device is the first node, the second information sent by the second communication device indicates the processing performance of the first node on the test data; similarly, if the second communication device is the second node, the second information sent by the second communication device indicates the processing performance of the second node on the test data. Therefore, the second communication device can instruct the two nodes to communicate using the first multiple access method based on the processing performance of the first and second nodes on the test data, ensuring that the multiple access method used by these nodes during communication is compatible with their processing performance on the test data, thereby improving model processing efficiency.

[0064] In one possible implementation of the second aspect, the second information is also used to determine that there is a mapping relationship between the first AI model and the second AI model deployed by the first node and the third AI model deployed by the second node.

[0065] Based on the above scheme, when the second information is used to determine the processing performance of the first node and the second node on the second data, the second communication device can also determine the mapping relationship between different models based on the second information. For example, the second communication device can determine that the two nodes can communicate through the first multiple access method based on the processing performance of the first node and the second node on the test data. Correspondingly, there is a mapping relationship between the first AI model used to obtain the second data, the second AI model of the first node processing the second data, and the third AI model of the second node processing the second data, enabling the second communication device to maintain or save this mapping relationship. Subsequently, it can instruct the first multiple access method based on one or more maintained or saved mapping relationships, thereby improving the instruction efficiency of the multiple access method and reducing processing complexity.

[0066] A third aspect of this application provides a communication device that performs the functions described in the first aspect. For example, the communication device includes modules, units, or means corresponding to the operations involved in the first aspect. These modules, units, or means can be implemented in software, hardware, or a combination of both. For instance, the device includes a communication unit. This communication unit is used to send first information indicating a first multiple access method. The communication unit is also used to send first data based on the first multiple access method. Sending the first data based on the first multiple access method includes: the communication unit sending the first data to a first node and a second node using some or all of the same resources; wherein the first data is obtained by processing first initial data and second initial data using a first artificial intelligence (AI) model, the first initial data corresponding to the first node, and the second initial data corresponding to the second node.

[0067] In one possible implementation of the third aspect, the communication unit is further configured to receive second information; the device further includes a processing unit configured to generate the first information based on the second information.

[0068] In the third aspect of this application, the constituent modules of the communication device can also be used to perform the steps executed in various possible implementations of the first aspect and achieve the corresponding technical effects. For details, please refer to the first aspect, which will not be repeated here.

[0069] A fourth aspect of this application provides a communication device that performs the functions described in the second aspect above. For example, the communication device includes modules, units, or means corresponding to the operations involved in the second aspect. These modules, units, or means can be implemented in software, hardware, or a combination of both. For instance, the device includes a communication unit. This communication unit is used to receive first information indicating a first multiple access mode; the communication unit is also used to receive first data based on the first multiple access mode.

[0070] In the fourth aspect of this application, the constituent modules of the communication device can also be used to perform the steps executed in various possible implementations of the second aspect and achieve the corresponding technical effects. For details, please refer to the second aspect, which will not be repeated here.

[0071] A fifth aspect of this application provides a communication device including an interface circuit and one or more processors. The one or more processors are coupled to a memory. The memory stores part or all of a computer program or instructions necessary for implementing the functions described in the first aspect. The one or more processors are executable to carry out the computer program or instructions, causing the communication device to implement any possible design or implementation method described in the first aspect. The interface circuit is used to implement communication functions within the communication device and / or communication functions between the communication device and other devices or components.

[0072] In one possible design, the processor is used to communicate with other devices or components through the interface circuit.

[0073] In one possible design, the communication device may also include the memory.

[0074] The aforementioned communication device may be a terminal, or a communication and / or computing module within a terminal, or a chip within a terminal responsible for communication functions, such as a modem chip (also known as a baseband chip) or a SoC or SIP chip containing a modem module, or a circuit or chip within a terminal responsible for communication and / or computing functions (such as a GPU, NPU, AI processor, or ASIC), or a logical node or logical module capable of implementing all or part of the terminal's functions. Alternatively, the aforementioned communication device may be a network-side device, or a module within a network-side device (e.g., a circuit, chip, or chip system), or a circuit or chip within a network-side device responsible for communication and / or computing functions (such as a GPU, NPU, AI processor, or ASIC), or a logical node or logical module capable of implementing all or part of the network-side device's functions. The network-side device may include access network equipment, core network equipment, a server, or other equipment defined by the future network.

[0075] A sixth aspect of this application provides a communication device including an interface circuit and one or more processors. The one or more processors are coupled to a memory. The memory stores part or all of a computer program or instructions necessary for implementing the functions described in the second aspect above. The one or more processors are executable to carry out the computer program or instructions, causing the communication device to implement any possible design or implementation method described in the second aspect above. The interface circuit is used to implement communication functions within the communication device and / or communication functions between the communication device and other devices or components.

[0076] In one possible design, the processor is used to communicate with other devices or components through the interface circuit.

[0077] In one possible design, the communication device may also include the memory.

[0078] The aforementioned communication device may be a network-side device, or a module (e.g., circuit, chip, or chip system) within a network-side device, or a circuit or chip (e.g., GPU, NPU, AI processor, or ASIC) within a network-side device responsible for communication and / or computing functions, or a logical node or logical module capable of implementing all or part of the functions of the network-side device. The network-side device may include access network equipment, core network equipment, servers, or other equipment defined in the future network.

[0079] A seventh aspect of this application provides a communication system including the aforementioned first communication device. The communication device may further include a first node and / or a second node, the implementation of which can refer to the second communication device and related implementation process described above.

[0080] An eighth aspect of this application provides a computer-readable storage medium for storing one or more computer-executable instructions that, when executed by a processor, perform the method as described in any possible implementation of the first or second aspect above.

[0081] The ninth aspect of this application provides a computer program product (or computer program) that, when executed by a processor, performs the method described in any possible implementation of the first or second aspect described above.

[0082] The tenth aspect of this application provides a chip or chip system including at least one processor for supporting a communication device in implementing the method described in any possible implementation of the first or second aspect. For example, the chip may be a baseband chip, a modem chip, a system-on-a-chip (SoC) chip containing a modem core, a system-in-package (SIP) chip, or a communication module, etc.

[0083] In one possible design, the chip or chip system may further include a memory for storing program instructions and data necessary for the communication device. The chip system may be composed of chips or may include chips and other discrete devices. Optionally, the chip system may also include interface circuitry that provides program instructions and / or data to the at least one processor.

[0084] The technical effects of any of the design methods in aspects three through ten can be found in the first or second aspects and their different design methods, and will not be repeated here. Attached Figure Description

[0085] Figures 1 to 3A schematic diagram of the communication system provided in this application;

[0086] Figures 4a to 4b A schematic diagram of the communication method provided in this application;

[0087] Figure 5 A schematic diagram illustrating the processing procedure of the AI ​​model provided in this application;

[0088] Figures 6 to 7 A schematic diagram of the communication device provided in this application. Detailed Implementation

[0089] First, some terms used in the embodiments of this application will be explained to facilitate understanding by those skilled in the art.

[0090] (1) The terms "system" and "network" in the embodiments of this application can be used interchangeably. "Multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, or B exists alone, where A and B can be singular or plural. The character " / " generally indicates that the related objects before and after are in an "or" relationship. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, "at least one of A, B and C" includes A, B, C, AB, AC, BC or ABC. And, unless otherwise specified, the ordinal numbers such as "first" and "second" mentioned in the embodiments of this application are used to distinguish multiple objects and are not used to limit the order, sequence, priority or importance of multiple objects.

[0091] (2) In the embodiments of this application, "send" and "receive" indicate the direction of signal transmission. For example, "send information to XX" can be understood as the destination of the information being XX, which may include sending directly through the air interface or sending indirectly through the air interface by other units or modules. "Receive information from YY" can be understood as the source of the information being YY, which may include receiving directly from YY through the air interface or receiving indirectly from YY through the air interface by other units or modules. "Send" can also be understood as the "output" of the chip interface, and "receive" can also be understood as the "input" of the chip interface.

[0092] In other words, sending and receiving can occur between devices, such as between network devices and terminals, or within a device, such as between components, modules, chips, software modules, or hardware modules within the device via buses, wiring, or interfaces.

[0093] It is understandable that information may undergo necessary processing, such as encoding and modulation, between the source and destination, but the destination can understand the valid information from the source. Similar statements in this application can be interpreted in a similar way and will not be elaborated further.

[0094] (3) In the embodiments of this application, "instruction" may include direct instruction and indirect instruction, as well as explicit instruction and implicit instruction. The information indicated by a certain piece of information (as described below, the instruction information) is called the information to be instructed. In the specific implementation process, there are many ways to indicate the information to be instructed, such as, but not limited to, directly indicating the information to be instructed, such as the information to be instructed itself or its index. It can also indirectly indicate the information to be instructed by indicating other information, where there is an association between the other information and the information to be instructed; or it can only indicate a part of the information to be instructed, while the other parts of the information to be instructed are known or pre-agreed upon. For example, the instruction can be implemented by using a pre-agreed (e.g., protocol predefined) arrangement of various information, thereby reducing the instruction overhead to a certain extent. This application does not limit the specific method of instruction. It is understood that for the sender of the instruction information, the instruction information can be used to indicate the information to be instructed, and for the receiver of the instruction information, the instruction information can be used to determine the information to be instructed.

[0095] (4) Multiple Access Technologies. Taking uplink and downlink communication as an example, the purpose of multiple access is to allow multiple terminals to simultaneously access network-side equipment (such as base stations) and enjoy the communication services provided by the network-side equipment, ensuring that the signals between the terminals do not interfere with each other. Generally, common multiple access technologies include frequency division multiple access (FDMA), time division multiple access (TDMA), code division multiple access (CDMA), orthogonal frequency division multiple access (OFDMA), and non-orthogonal multiple access (NOMA), etc.

[0096] (5) User pairing: In a multi-user multiple-input multiple-output (MU-MIMO) system, user pairing can improve the system's spectral efficiency and uplink / downlink capacity. For example, when MU-MIMO is enabled, network-side equipment (such as base stations) can schedule each terminal through pairing strategies and select suitable terminals for pairing and transmission.

[0097] (6) Configuration and pre-configuration: Configuration and pre-configuration may be used in this application. Configuration refers to the network device / server sending configuration information or parameter values ​​to the communication device (such as the terminal device or network-side device mentioned later) via messages or signaling, so that the communication device can determine the communication parameters or resources for transmission based on these values ​​or information. Pre-configuration may refer to the network device / server negotiating parameter information or parameter values ​​with the communication device in advance, or it may refer to the parameter information or parameter values ​​specified by the standard protocol for the communication device, or it may refer to the parameter information or parameter values ​​pre-stored in the communication device. This application does not limit this.

[0098] Furthermore, these values ​​and parameters can be changed or updated.

[0099] In this application, unless otherwise specified, the same or similar parts between the various embodiments can be referred to each other. In the various embodiments of this application, and the various methods / designs / implementations within each embodiment, unless otherwise specified or logically conflicting, the terminology and / or descriptions between different embodiments and between the various methods / designs / implementations within each embodiment are consistent and can be mutually referenced. The technical features in different embodiments and the various methods / designs / implementations within each embodiment can be combined to form new embodiments, methods, or implementations based on their inherent logical relationships. The following descriptions of the embodiments of this application do not constitute a limitation on the scope of protection of this application.

[0100] This application can be applied to long-term evolution (LTE) systems, new radio (NR) systems, or future communication systems. These communication systems include at least one network device and / or at least one terminal.

[0101] Please see Figure 1 This is a schematic diagram of the architecture of the communication system 10 used in an embodiment of this application. Figure 1 As shown, the communication system 10 includes a radio access network (RAN) 100 and a core network (CN) 200. RAN 100 includes at least one RAN node (e.g., ...). Figure 1 110a and 110b (collectively referred to as 110) and at least one terminal (such as Figure 1 RAN 100, denoted as RAN 120a-120j, is collectively referred to as RAN 120. RAN 100 may also include other RAN nodes, such as wireless relay equipment and / or wireless backhaul equipment. Figure 1 (Not shown in the image). Terminal 120 is connected to RAN node 110 wirelessly. RAN node 110 is connected to core network 200 wirelessly or via wired connection. The core network equipment in core network 200 and RAN node 110 in RAN 100 can be different physical devices, or they can be the same physical device integrating core network logical functions and radio access network logical functions.

[0102] RAN 100 can be a cellular system related to the 3rd Generation Partnership Project (3GPP), such as 4G, 5G mobile communication systems, or future-oriented evolution systems. RAN 100 can also be an open RAN (O-RAN or ORAN), a cloud RAN (CRAN), a virtualized RAN (vRAN), an artificial intelligence radio access network (AI RAN), or a wireless fidelity (WiFi) system. RAN 100 can also be a communication system that integrates two or more of the above systems.

[0103] RAN node 110, sometimes also referred to as access network equipment, RAN entity, or access node, constitutes part of the communication system and is used to help terminals achieve wireless access. Multiple RAN nodes 110 in communication system 10 can be of the same type or different types. In some scenarios, the roles of RAN node 110 and terminal 120 are relative, for example... Figure 1 Network element 120i can be a helicopter or a drone, and it can be configured as a mobile base station. For terminals 120j that access RAN 100 through network element 120i, network element 120i is a base station; however, for base station 110a, network element 120i is a terminal. RAN node 110 and terminal 120 are sometimes referred to as communication devices, for example... Figure 1 Network elements 110a and 110b can be understood as communication devices with base station functions, while network elements 120a-120j can be understood as communication devices with terminal functions.

[0104] In one possible scenario, a RAN node can be a base station, an evolved NodeB (eNodeB), an access point (AP), a transmission reception point (TRP), a next-generation NodeB (gNB), a base station in a future mobile communication system, or an access node in a WiFi system, etc. Figure 1 110a), micro base stations or indoor stations (such as Figure 1 The RAN node can be a relay node or donor node (as described in section 110b), or a wireless controller in a CRAN scenario. Optionally, the RAN node can also be a server, wearable device, vehicle, or in-vehicle equipment. For example, the access network equipment in vehicle-to-everything (V2X) technology can be a roadside unit (RSU). All or part of the functions of the RAN node in this application can also be implemented through software functions running on hardware, or through virtualization functions instantiated on a platform (e.g., a cloud platform). The RAN node can also be equipped with communication modules, circuits, or chips that perform corresponding communication functions. The RAN node can also be configured with program instructions for performing corresponding communication functions and corresponding program instructions. The RAN node in this application can also be a logical node, logical module, or software that can implement all or part of the access node functions, or a circuit or chip (such as a GPU, AI processor, NPU, or ASIC) responsible for communication and / or computing functions in the access node.

[0105] In another possible scenario, multiple RAN nodes collaborate to assist terminals in achieving wireless access, with different RAN nodes each implementing some of the base station's functions. For example, RAN nodes can be central units (CUs), distributed units (DUs), CU-control plane (CPs), CU-user plane (UPs), or radio units (RUs), etc. CUs and DUs can be set up separately or included in the same network element, such as a baseband unit (BBU). RUs can be included in radio frequency equipment or radio frequency units, such as remote radio units (RRUs), active antenna units (AAUs), or remote radio heads (RRHs). Furthermore, RAN nodes can also be computing units, providing computing power for tasks (such as model inference and / or model training), and can also be used to implement one or more of the following: task partitioning, scheduling, and orchestration. The functionality of a computing unit can be implemented by a separate module independent of other units (e.g., CU, DU, RU), or by one or more other units (e.g., one or more of CU, DU, RU).

[0106] In different systems, CU (or CU-CP and CU-UP), DU, computing unit, or RU may have different names, but those skilled in the art will understand their meaning. For example, in an ORAN system, CU can also be called O-CU (open CU), DU can also be called O-DU, CU-CP can also be called O-CU-CP, CU-UP can also be called O-CU-UP, and RU can also be called O-RU. For ease of description, this application uses CU, CU-CP, CU-UP, DU, computing unit, and RU as examples. Any of the units among CU (or CU-CP, CU-UP), DU, computing unit, and RU in this application can be implemented through software modules, hardware modules, or a combination of software modules and hardware modules.

[0107] A terminal can be a device or module that accesses the aforementioned communication system and has corresponding communication functions. A terminal can also be called a terminal device, user equipment (UE), mobile station, mobile terminal, etc. Terminals can be widely used in various scenarios, such as device-to-device (D2D), vehicle-to-everything (V2X) communication, machine-type communication (MTC), Internet of Things (IoT), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grids, smart furniture, smart offices, smart wearables, smart transportation, smart cities, etc. Terminals can be mobile phones, tablets, computers with wireless transceiver capabilities, wearable devices, vehicles, drones, helicopters, airplanes, ships, robots, robotic arms, smart home devices, transportation vehicles with wireless communication capabilities, communication modules, etc. The embodiments of this application do not limit the device form of the terminal. The terminal typically contains communication modules, circuits, or chips that perform corresponding communication functions. Furthermore, it may also contain modules, circuits, or chips (such as GPUs, AI processors, NPUs, or ASICs) that perform corresponding communication and / or computing functions. The terminal can also be configured with program instructions for performing these communication and / or computing functions.

[0108] To support AI technology in wireless networks, AI nodes may also be introduced into the network.

[0109] AI nodes can be deployed in one or more of the following locations within the communication system: access network nodes (RAN nodes), terminal devices, or core network devices. Alternatively, AI nodes can be deployed independently, for example, in a location other than any of the aforementioned devices, such as in the host or cloud server of an over-the-top (OTT) system. AI nodes can communicate with other devices in the communication system, which can be one or more of the following: network devices, terminal devices, or core network elements.

[0110] It is understood that this application does not limit the number of AI nodes. For example, when there are multiple AI nodes, these nodes can be divided based on function, such as different AI nodes being responsible for different functions.

[0111] It can also be understood that AI nodes can be independent devices, or they can be integrated into the same device to achieve different functions. Alternatively, they can be network elements in hardware devices, software functions running on dedicated hardware, or virtualization functions instantiated on a platform (e.g., a cloud platform). This application does not limit the specific form of the aforementioned AI nodes.

[0112] Optionally, the AI ​​node can be an AI network element or an AI module.

[0113] Please see Figure 2 This is a schematic diagram of a possible application framework in a communication system. For example... Figure 2 As shown, network elements in a communication system are connected via interfaces (e.g., NG, Xn) or air interfaces. These network element nodes, such as core network equipment, access network nodes (RAN nodes), terminals, or one or more devices in operations administration and maintenance (OAM), are equipped with one or more AI modules (for clarity, ...). Figure 2 (Only one is shown in the image). An access network node can be a single RAN node or can include multiple RAN nodes, such as a CU and a DU. The CU and / or DU can also be equipped with one or more AI modules. The CU can also be split into CU-CP and CU-UP, and one or more AI modules can be set in the CU-CP and / or CU-UP.

[0114] AI modules are used to implement corresponding AI functions. AI modules deployed in different network elements can be the same or different. The models of AI modules can achieve different functions depending on the parameter configurations. The models of AI modules can be configured based on one or more of the following parameters: structural parameters (e.g., at least one of the following: number of neural network layers, neural network width, inter-layer connections, neuron weights, neuron activation function, or biases in the activation function), input parameters (e.g., the type and / or dimension of the input parameters), or output parameters (e.g., the type and / or dimension of the output parameters). The biases in the activation function can also be referred to as the biases of the neural network.

[0115] In one example, the neural network mentioned above can be a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), or a generative adversarial network (GAN).

[0116] Deep Neural Networks (DNNs) are artificial neural network architectures with multiple layers of nonlinear transformation units stacked in a hierarchical structure to form deep computational models. Compared to shallow neural networks, deep neural networks have more hidden layers, allowing the network model to capture more complex data structures and higher-level abstract features.

[0117] A CNN is a deep neural network with a convolutional structure. A CNN contains a feature extractor consisting of convolutional layers and subsampling layers. This feature extractor can be viewed as a filter, and the convolution process can be seen as performing convolution between a trainable filter and an input image or a convolutional feature map.

[0118] RNN is a type of recursive neural network that takes sequence data as input, recursively moves along the direction of sequence evolution, and connects all nodes (recurrent units) in a chain-like manner.

[0119] GAN is a deep learning model. It consists of a generator and a discriminator, and is trained through adversarial learning. Its purpose is to estimate the potential distribution of data samples and generate new data samples.

[0120] For example, an AI module may have one or more models. A model can infer an output that includes one or more parameters. The learning, training, or inference processes of different models may be deployed on different nodes or devices, or they may be deployed on the same node or device.

[0121] Optionally, the AI ​​processing involved in this application may include LoRA. LoRA is a low-rank fine-tuning method whose core idea is to freeze the weights of the pre-trained model and then inject the trainable low-rank factorization matrix into one or more layers of the Transformer architecture, thereby significantly reducing the number of trainable parameters for downstream tasks. Furthermore, by fine-tuning the same pre-trained model using different LoRA parameters, the resulting fine-tuned model can be used for different downstream tasks.

[0122] Please see Figure 3 This is a schematic diagram of a possible application framework in a communication system. For example... Figure 3 As shown, the communication system includes a RAN intelligent controller (RIC). For example, the RIC could be... Figure 2The AI ​​module shown is used to implement AI-related functions. RICs include near-real-time RICs (near-RT RICs) and non-real-time RICs (non-RT RICs). Non-real-time RICs primarily process non-real-time information, such as data that is not sensitive to latency, with latency in the order of seconds. Real-time RICs primarily process near-real-time information, such as data that is relatively sensitive to latency, with latency in the order of tens of milliseconds.

[0123] Near real-time (NRT) RICs are used for model training and inference. For example, they are used to train AI models and then use those models for inference. NRT RICs can obtain network-side and / or terminal-side information from RAN nodes (e.g., CUs, CU-CPs, CU-UPs, DUs, compute nodes, and / or RUs) and / or terminals. This information can be used as training data or inference data. NRT RICs can deliver inference results to RAN nodes and / or terminals. Inference results can be exchanged between CUs and DUs, and / or between DUs and RUs. For example, a NRT RIC delivers an inference result to a DU, which then forwards it to an RU.

[0124] Non-real-time RICs are also used for model training and inference. For example, they are used to train AI models and then use those models for inference. Non-real-time RICs can obtain network-side and / or terminal-side information from RAN nodes (e.g., CUs, CU-CPs, CU-UPs, DUs, compute nodes, and / or RUs) and / or terminals. This information can be used as training data or inference data, and the inference results can be delivered to RAN nodes and / or terminals. Inference results can be exchanged between CUs and DUs, and / or between DUs and RUs; for example, a non-real-time RIC delivers inference results to a DU, which then forwards them to an RU.

[0125] Near real-time RICs and non-real-time RICs can also be configured as separate network elements. Near real-time RICs and non-real-time RICs can also be part of other devices. For example, near real-time RICs can be set in RAN nodes (e.g., CU, DU, compute nodes), while non-real-time RICs can be set in OAM, cloud servers, core network devices, or other network devices.

[0126] As described above, in a communication network, communication nodes not only need to process communication signals but may also need to handle AI-related processing. Therefore, how to integrate AI-related processing with the communication network is a pressing technical problem that needs to be solved.

[0127] As an example, multiple access methods (such as FDMA and TDMA) can improve communication performance in communication networks. Channel state information is a crucial factor influencing the communication performance between different communication devices. Therefore, MU-MIMO pairing primarily achieves multi-user pairing through measured channel state information. Subsequent network devices can then schedule or instruct multiple access methods (such as FDMA and TDMA) based on the pairing results to improve communication performance. For instance, network devices can determine the channel orthogonality of different users using channel state information. By selecting users with high channel orthogonality for pairing, the paired users communicate via multiple access, leveraging spatial separation to reduce interference between different users. Similarly, network devices can determine system performance using channel state information, pairing users who contribute significantly to system gain. The paired users then communicate via multiple access to improve system performance. However, in future communication networks, communication nodes may involve AI processing, for example… Figure 2 The terminals and access network nodes shown can be deployed with AI modules for AI processing; the AI ​​processing performance of these AI modules may also be related to factors other than channel state information, which may lead to a decrease in the performance of the above multiple access methods when AI processing is involved.

[0128] To address the aforementioned problems, this application provides a communication method and related apparatus, which will be described in detail below with reference to the accompanying drawings.

[0129] It should be understood that the following description uses different communication devices as examples to illustrate the method, but this application does not limit the subject of the interaction.

[0130] For example, the first communication device may be a terminal, or a communication and / or computing module within the terminal, or a circuit or chip within the terminal responsible for communication and / or computing functions (such as a modem chip (also known as a baseband chip), or a SoC chip / SIP chip containing a modem core, or a GPU / NPU / AI processor / ASIC). Alternatively, the method involving the first communication device can be executed by a logical node, logical module, or software capable of implementing all or part of the terminal's functions. Or, the first communication device may be a network-side device, or a circuit or chip within the network-side device responsible for communication and / or computing functions (such as a GPU, NPU, AI processor, or ASIC). Alternatively, the method involving the first communication device can be executed by a logical node, logical module, or software capable of implementing all or part of the network-side device's functions.

[0131] For example, the second communication device can be a terminal, or a communication and / or computing module within the terminal, or a circuit or chip within the terminal responsible for communication and / or computing functions (such as a modem chip (also known as a baseband chip), or a SoC chip / SIP chip containing a modem core, or a GPU / NPU / AI processor / ASIC). Alternatively, the method involving the second communication device can be implemented through a logical node, logical module, or software capable of implementing all or part of the terminal's functions. Or, the second communication device can be a network-side device, or a circuit or chip within the network-side device responsible for communication and / or computing functions (such as a GPU, NPU, AI processor, or ASIC), or the method involving the second communication device can be implemented through a logical node, logical module, or software capable of implementing all or part of the network-side device's functions.

[0132] Optionally, the aforementioned network-side equipment may include access network equipment, core network equipment, servers, or other equipment defined in the future network.

[0133] Please see Figure 4a This is a schematic diagram of an implementation of the communication method provided in this application, which includes the following steps.

[0134] S401. The first communication device sends first information, and correspondingly, the second communication device receives the first information. The first information is used to indicate a first multiple access mode.

[0135] S402. The first communication device transmits first data based on a first multiple access method, and correspondingly, the second communication device receives the first data based on the first multiple access method. The first data is obtained by processing first initial data and second initial data through a first AI model; the first initial data corresponds to a first node, and the second initial data corresponds to a second node.

[0136] In step S402, the first communication device sending the first data based on the first multiple access method includes: the first communication device sending the first data to the first node and the second node through some or all of the same resources.

[0137] It should be noted that the first communication device can send the first data to the first node and the second node using some or all of the same resources, and such resources can be implemented in various ways. For example, the resources may include one or more of time domain resources, frequency domain resources, spatial domain resources, and code domain resources.

[0138] It should be noted that the first communication device can process the first initial data and the second initial data in various ways to obtain the first data.

[0139] For example, the first communication device can use the first initial data and the second initial data as input to the first AI model, and after processing by the first AI model, obtain the first data.

[0140] For example, the first communication device can use the first initial data as input to the first AI model, and after processing by the first AI model, obtain the first result; the first communication device can use the second initial data as input to the first AI model, and after processing by the first AI model, obtain the second result; thereafter, the first communication device processes (e.g., adds) the first result and the second result to obtain the first data.

[0141] Optionally, the first data can be obtained by processing at least two initial data sets through a first AI model. These at least two initial data sets may include the first initial data corresponding to the first node and the second initial data corresponding to the second node. Furthermore, these at least two initial data sets may also include one or more initial data sets corresponding to other nodes. The implementation process of these other nodes and their corresponding initial data can refer to the implementation process of the first or second node. In other words, the number of nodes receiving the first data can be two or more. The following example uses two nodes receiving the first data (these two nodes are the aforementioned first and second nodes) as an example. Figure 4b The process shown is illustrated by way of example.

[0142] Optionally, in this application, the data sent by the first communication device (such as the first data mentioned above, the second data that may appear later, etc.) can be the input data of the AI ​​model, so that such data can be used as the input of the model, enabling communication between AI models and simplifying the data processing process. For example, the first data includes one or more of the following: one or more tokens, one or more token vectors, one or more patches, one or more embeddings, one or more patch embeddings, or one or more token embeddings, etc.

[0143] like Figure 4b As shown, in Figure 4a In the process shown, the second communication device can be either the first node or the second node mentioned above. The implementation process of these two nodes can refer to the implementation process of steps S401 and S402. Specifically, in... Figure 4a In step S401 shown, the first communication device can transmit the first information through either method one or method two.

[0144] Method 1: The first communication device transmits the first information in a single transmission process.

[0145] For example, the first communication device sends first information in step S401a, enabling the first node and the second node to receive the first information in step S401a. This reduces overhead and improves communication efficiency.

[0146] In Method 1, the way the first communication device sends the first information can be understood as broadcasting, multicasting, etc.

[0147] Method 2: The first communication device transmits the first information through at least two transmission processes.

[0148] For example, in step S401b, the first communication device sends first information, enabling the first node to receive the first information in step S401b; and in step S401c, the first communication device sends first information, enabling the second node to receive the first information in step S401c. In this way, the first communication device can flexibly instruct the multiple access method for each node.

[0149] In one possible implementation, the aforementioned first initial data and the first node may be associated. For example, in the above process, the first initial data corresponding to the first node can be understood as follows: the first node processes the first data to obtain a first processing result, i.e., the first processing result is the processing result corresponding to the first initial data; in other words, the destination node of the first initial data is the first node. This first node may deploy a second AI model to process the first data. For example, the first initial data is processed sequentially by the first model and then by the second model (this processing can be inference, prediction, etc.) to obtain the first processing result. Correspondingly, in the above process, after receiving the first data obtained based on the first initial data, the first node can use the first data as input to the second AI model, and after processing by the second AI model, obtain the first processing result.

[0150] In one possible implementation, the aforementioned second initial data and the second node can be associated. For example, in the above process, the second initial data corresponding to the second node can be understood as follows: the second node processes the first data to obtain a second processing result, i.e., the second processing result is the processing result corresponding to the second initial data; in other words, the destination node of the second initial data is the second node. This second node can deploy a third AI model to process the first data. For example, the second initial data is processed sequentially by the first model and then the second model (this processing can be inference, prediction, etc.) to obtain the second processing result. Correspondingly, in the above process, after receiving the first data obtained based on the second initial data, the second node can use the first data as input to the third AI model, and after processing by the third AI model, obtain the second processing result.

[0151] like Figure 5 The example shown still uses the case where the number of nodes receiving the first data is 2 (these 2 nodes are the first node and the second node mentioned above). The implementation process of each device / node is as follows.

[0152] A first communication device can deploy a first AI model. The input to the first AI model includes first initial data and second initial data. After processing by the first AI model, the output includes the first data. Subsequently, the first communication device transmits the first data based on a first multiple access method.

[0153] The first node can deploy a second AI model. Specifically, the first node can receive first data (or an estimate of the first data) based on a first multiple access method, and then process the first data based on the second AI model to obtain a first processing result.

[0154] The second node can deploy a third AI model. Specifically, the second node can receive the first data (or an estimate of the first data) based on a first multiple access method. Subsequently, the second node can process the first data based on the third AI model to obtain a second processing result.

[0155] Optionally, in the above implementation process based on the first multiple access method, the first AI model can obtain first data based on first initial data and second initial data. Furthermore, this first data can be processed by the second AI model to obtain a first processing result, and the first data can also be processed by the third AI model to obtain a second processing result. In this process, the first AI model can be understood as a common model, while the second and third AI models can be understood as personalized models. This method utilizes the characteristics of LoRA; therefore, this first multiple access method can also be called the LoRA multiple access method, or other names defined in the future network definition.

[0156] Optionally, the AI ​​model's processing of the input data may include a preprocessing process, or the input data of the AI ​​model may be preprocessed data. In the above implementation based on the first multiple access method, the first data is obtained by the first AI model processing the first initial data and the second initial data, that is, the first AI model supports processing the first initial data and the second initial data. Therefore, the processing result of these two initial data after preprocessing conforms to the input requirements of the same AI model (i.e., the first AI model), enabling the first AI model to process the first initial data and the second initial data, reducing or avoiding situations where the first AI model cannot process the data, and improving processing performance. For example, the above input requirements may include requirements for at least one of type, format, dimension, or modality.

[0157] Based on the above scheme, the first information sent by the first communication device in step S401 is used to indicate the first multiple access method. Furthermore, the first communication device can send first data based on the first multiple access method in step S402, enabling the receiver of the first data to receive the first data based on the first multiple access method. Specifically, during the process of the first communication device sending the first data based on the first multiple access method, after processing the first initial data corresponding to the first node and the second initial data corresponding to the second node using the first AI model to obtain the first data, the first communication device can send the first data to the first node and the second node using some or all of the same resources. In this way, the first communication device can use the same AI model to process the initial data of different nodes using the first multiple access method to obtain and send the first data. This allows for resource reuse to achieve the transmission of AI data from different nodes, improving resource utilization and also enabling the reuse of the AI ​​model to process data from different nodes, thereby improving model processing efficiency.

[0158] In one possible implementation, the aforementioned first information can be obtained in multiple ways, and some possible implementation examples will be provided below.

[0159] In Example 1, the first piece of information is determined based on one or more mapping relationships. For example, a mapping relationship can indicate the mapping relationship between multiple AI models. These multiple AI models with mapping relationships include the AI ​​model deployed on the first communication device, as well as the AI ​​models deployed on other nodes.

[0160] In Example 1, the first AI model is mapped to the second AI model deployed on the first node and the third AI model deployed on the second node. Therefore, the first communication device can determine one or more mapping relationships through pre-configuration or configuration. One type of mapping relationship is between the first AI model and the second AI model deployed on the first node and the third AI model deployed on the second node. This allows the first communication device to instruct the first multiple access mode based on these one or more mapping relationships, thereby improving the efficiency of multiple access mode instruction and reducing processing complexity.

[0161] For example, the first AI model mentioned above can be a public model, and the second and third AI models can be personalized models. Therefore, any one of the one or more mapping relationships involved in this application can be understood as a mapping relationship between a public model and one or more personalized models. That is, the public model indicated by any mapping relationship can process one or more initial data corresponding to the one or more personalized models to obtain processing results, and the one or more personalized models can process the processing results to obtain their respective personalized outputs. The process shown in Table 1 will be illustrated below.

[0162] Table 1

[0163] Public Model Index Personalized Model Index Personalized Model Index ... 1 4 7 2 5 6

[0164] In Table 1, a public model index can indicate a public model (such as the first AI model), and a personalized model index can indicate a personalized model (such as the second AI model or the third AI model). Personalized model indexes located in the same row as the public model index indicate that there is a mapping relationship. AI models with a mapping relationship can communicate with each other through the first multiple access method.

[0165] For example, in Table 1, the public model with a public model index of 1 can communicate with the two personalized models with personalized model indices of 4 and 7 (and other personalized models may also exist) through the first multiple access method. For example, the personalized model index of the second AI model described above can be 4, and the personalized model index of the third AI model described above can be 7.

[0166] For example, in Table 1, the public model with a public model index of 2 can communicate with the two personalized models with personalized model indices of 5 and 6 (and other personalized models may also exist) through the first multiple access method. For instance, the personalized model index of the second AI model described above can be 5, and the personalized model index of the third AI model described above can be 6.

[0167] Alternatively, in Table 1, a public model index can indicate a public model (such as the first AI model), and a personalized model index can indicate a set of personalized models, which can contain one or more personalized models. Personalized model indices located in the same row as the public model index indicate that there is a mapping relationship, and AI models with a mapping relationship can communicate through a first multiple access method.

[0168] For example, in Table 1, the public model with a public model index of 1 can communicate with the two personalized model sets with indexes of 4 and 7 (and other personalized models may also exist) via the first multiple access method. For instance, the personalized model set to which the second AI model described earlier belongs may have an index of 4, and the personalized model set to which the third AI model described earlier belongs may have an index of 7. For example, the second and third AI models described earlier may be contained in the same personalized model set, and the index of this same personalized model set may be 4 or 7.

[0169] For example, in Table 1, the public model with a public model index of 2 can communicate with the two personalized models with personalized model indices of 5 and 6 (and other personalized models may also exist) through the first multiple access method. For instance, the personalized model set to which the second AI model described earlier belongs can have an index of 5, and the personalized model set to which the third AI model described earlier belongs can have an index of 6. For example, the second and third AI models described earlier can be contained in the same personalized model set, and the index of this same personalized model set can be 5 or 6.

[0170] Alternatively, Table 1 can also be implemented in other ways, as shown in Table 2.

[0171] Table 2

[0172] Mapping Relationship Index Public Model Index Personalized Model Index Personalized Model Index ... 1 1 4 7 2 2 5 6

[0173] As shown in Table 2, compared to Table 1, a new column called "Mapping Relationship Index" can be added to represent the rows in Table 1. That is, each row of parameters in Table 1 can correspond to a mapping relationship index, and different rows can correspond to different mapping relationship indices.

[0174] Optionally, the first AI model has a mapping relationship with the first node, the second AI model deployed on the first node, the second node, and the third AI model deployed on the second node. In other words, the above mapping relationship can also indicate the nodes, and correspondingly, each mapping relationship can be implemented through Table 3.

[0175] Table 3

[0176] Public Model Index Node Index & Personalized Model Index Node Index & Personalized Model Index ... 1 1&4 2&7 2 1&5 2&6

[0177] In Table 3, a public model index can indicate a public model (such as the first AI model), and a node index & personalized model index can indicate a node and the personalized model deployed on that node (such as the first node and the second AI model deployed on that first node, or the second node and the third AI model deployed on that second node). Personalized model indices located in the same row as the public model index indicate that there is a mapping relationship. AI models with a mapping relationship can communicate through the first multiple access method.

[0178] For example, in Table 3, a public model with a public model index of 1 can communicate with a personalized model with a node index of 1 and a personalized model index of 4, and a personalized model with a node index of 2 and a personalized model index of 7 (other personalized models may also exist) via a first multiple access method. For example, the index of the first node described above is 1, the personalized model index of the second AI model can be 4, the index of the second node described above is 2, and the personalized model index of the third AI model can be 7.

[0179] For example, in Table 3, the public model with a public model index of 2 can communicate with the personalized model with a node index of 1 and a personalized model index of 5, as well as the personalized model with a node index of 2 and a personalized model index of 6 (and other personalized models may also exist) through the first multiple access method. For instance, if the index of the first node described above is 1, the index of the personalized model set to which the second AI model belongs can be 4; if the index of the second node described above is 2, the index of the personalized model set to which the third AI model belongs can be 7.

[0180] Alternatively, Table 3 can also be implemented in other ways, as shown in Table 4.

[0181] Table 4

[0182]

[0183] As shown in Table 4, compared to Table 3, a new column called "Mapping Relationship Index" can be added to represent the rows in Table 3. That is, each row of parameters in Table 3 can correspond to a mapping relationship index, and different rows can correspond to different mapping relationship indices.

[0184] It should be noted that in Table 1 or Table 2, the mapping relationship corresponding to one row can be communicated through the first multiple access method. Different rows correspond to different mapping relationships, and the data transmission corresponding to different mapping relationships can be carried on different resources to reduce interference between different data transmission processes. Alternatively, the data transmission corresponding to different mapping relationships can be transmitted through the second multiple access method (such as one or more of FDMA, TDMA, CDMA, and OFDMA), which can reduce interference and improve resource utilization.

[0185] Optionally, the first communication device can determine the mapping relationship in one or more mapping relationships based on the relevant information of the first node and the relevant information of the second node, and generate first information. The relevant information of the first node can be used to determine one or more AI models corresponding to the first node, and the relevant information of the second node can be used to determine one or more AI models corresponding to the second node. This allows the first communication device to use the AI ​​models corresponding to these two nodes as personalized model indexes, determine the public model indexes that have a mapping relationship with the personalized model indexes in Table 1 or Table 2, determine the nodes that communicate based on the first multiple access method based on the mapping relationship, and generate the first information to provide multiple access indication for these nodes.

[0186] In addition, the AI ​​model corresponding to a node can be implemented in the following ways.

[0187] For example, if no AI model is deployed on a node, the relevant information of the first node can indicate its model processing capability, enabling the first communication device to determine or predict the AI ​​model supported by the first node based on its model processing capability. Similarly, the relevant information of the second node can indicate its model processing capability, enabling the first communication device to determine or predict the AI ​​model supported by the second node based on its model processing capability. In other words, the AI ​​model corresponding to a node can be the AI ​​model supported by that node, determined based on its model processing capability. For a specific implementation example, please refer to Implementation Example 2 below.

[0188] For example, if an AI model has already been deployed on a node, the relevant information of the first node can indicate the index or identifier of the AI ​​model deployed on that node, and the relevant information of the second node can indicate the index or identifier of the AI ​​model deployed on that node. That is, the AI ​​model corresponding to a node can be based on the AI ​​model already deployed on that node. For a specific implementation example, please refer to Implementation Example 3 below.

[0189] For example, if an AI model has been deployed on a node, the relevant information of the first node can indicate its performance in processing test data, and the relevant information of the second node can indicate its performance in processing test data. That is, the AI ​​model corresponding to a node can be an AI model determined based on that node's performance in processing test data. For a specific implementation example, please refer to Implementation Example 4 below.

[0190] In Example 2, the first information is obtained based on the second information, which is used to indicate the model processing capabilities of one or more nodes.

[0191] For example, in Figure 4b In the scenario shown, the second information indicates at least the model processing capabilities of the first node and the model processing capabilities of the second node.

[0192] In Example 2, the first communication device can instruct some or all nodes to use the first multiple access method to communicate based on the model processing capabilities of one or more nodes. For example, the first communication device can instruct nodes with the same or similar model processing capabilities to use the first multiple access method to communicate, so that the multiple access method used by these nodes during communication can be adapted to the model processing information of these nodes, thereby improving model processing efficiency.

[0193] Optionally, model processing capabilities may include one or more of computing power, storage capacity, or idle computing power.

[0194] For example, computing power can be characterized by trillions of operations per second (TOPS). Correspondingly, nodes with similar model processing capabilities are those whose TOPS capabilities are in the same range (e.g., 2 TOPS to 4 TOPS, or more than 10 TOPS).

[0195] For example, storage capacity can be characterized by memory size or video memory size. Correspondingly, nodes with similar model processing capabilities are those whose memory size or video memory size is in the same range (e.g., 1 terabyte (TB) to 4 TB, or greater than 10 TB).

[0196] For example, idle computing power can be characterized by the utilization rate of processors (such as CPU, GPU, NPU, etc.). Accordingly, taking CPU as an example, nodes with similar model processing capabilities are identified as nodes whose CPU utilization is in the same range (e.g., CPU utilization less than 50%, less than 80%, etc.).

[0197] In one possible implementation of Example 2, the method further includes: the first communication device determining the second AI model corresponding to the first node and the third AI model corresponding to the second node based on the second information; wherein the first AI model has a mapping relationship with the second AI model and the third AI model.

[0198] For example, the first communication device can determine a second AI model adapted to the model processing capability of the first node based on second information (or the model processing capability of the first node indicated by the second information). That is, the first communication device can determine the second AI model that the first node supports for deployment based on the second information (or the model processing capability of the first node indicated by the second information). Similarly, the first communication device can determine a third AI model adapted to the model processing capability of the second node based on the second information (or the model processing capability of the second node indicated by the second information). That is, the first communication device can determine the third AI model that the second node supports for deployment based on the second information (or the model processing capability of the second node indicated by the second information).

[0199] In this way, the first communication device can determine the second AI model corresponding to the first node and the third AI model corresponding to the second node based on the second information. Furthermore, the first communication device can determine whether these two AI models have a mapping relationship with the first AI model based on one or more mapping relationships (such as one or more mapping relationships in Example 1). If the first AI model has a mapping relationship with the second AI model and the third AI model, the first communication device can determine that these three AI models can perform data processing in the above manner; that is, the first communication device generates first information to provide an indication of the first multiple access method.

[0200] Optionally, in Example 2, after the first communication device determines the second AI model corresponding to the first node and the third AI model corresponding to the second node based on the second information, the first communication device can also deploy the AI ​​model to the first node and / or the second node. For example, the first communication device can also send the model parameters of the AI ​​model or a model file containing the model parameters (e.g., the model file is a PyTorch model file (or pt file), a binary file (or bin file), an open neural network exchange format file (or onnx file), or a PyTorch script file (pyTorch scriptfile, pth file), etc.) based on the node's model processing capability. Taking model parameters as an example, the first communication device can send the model parameters of the second AI model to the first node based on the first node's model processing capability, so that the first node obtains the second AI model based on the model parameters, thereby realizing the deployment of the second AI model. For example, the first communication device can send model parameters of a third AI model to the second node based on the second node's model processing capabilities, enabling the second node to obtain the third AI model based on these parameters and thus deploy it. In this way, the above solution can be applied to scenarios where no corresponding AI model is currently deployed on any node, and provides AI models adapted to the model processing capabilities of each node, thereby improving the processing performance of these AI models.

[0201] For example, the model parameters mentioned above may include one or more of the following: structural parameters (e.g., at least one of the following: number of neural network layers, neural network width, inter-layer connections, neuron weights, neuron activation function, or bias in activation function), input parameters (e.g., type of input parameters and / or dimension of input parameters), or output parameters (e.g., type of output parameters and / or dimension of output parameters).

[0202] Optionally, in Implementation Example 2, the first communication device can determine the AI ​​models supported by one or more nodes based on their model processing capabilities, and instruct some nodes to communicate using a first multiple access method based on the AI ​​models with existing mapping relationships. These nodes may include the aforementioned first node and second node. The AI ​​models supported by the one or more nodes may include AI models without the aforementioned mapping relationships, and nodes that deploy AI models without the aforementioned mapping relationships can be referred to as other nodes. The first communication device may not instruct these other nodes to communicate using the first multiple access method. Alternatively, the first communication device can test the multiple access process of these other nodes through the testing and verification process described in Implementation Example 4, as detailed below.

[0203] In Example 3, the first information is obtained based on the second information, which is used to indicate the second AI model deployed by the first node and the third AI model deployed by the second node.

[0204] In Implementation Example 3, the first communication device can determine the second AI model and the third AI model deployed by the first node based on the second information. Furthermore, the first communication device can determine whether these two AI models have a mapping relationship with the first AI model based on one or more mapping relationships (such as one or more mapping relationships in Implementation Example 1). If the first AI model has a mapping relationship with the second AI model and the third AI model, the first communication device can determine that these three AI models can perform data processing in the aforementioned manner; that is, the first communication device generates first information to provide an indication of the first multiple access method.

[0205] In Example 4, the first information is obtained based on the second information, which is used to indicate the processing performance of one or more nodes on the test data. This process can be implemented based on the test data, and therefore, it can also be called a test verification process.

[0206] For example, in Figure 4b In the illustrated scenario, the second information can at least indicate the processing performance of the first node on the test data and the processing performance of the second node on the test data. Taking the test data as the second data as an example, before the first communication device sends the first information (such as before steps S401, S401a, S401b, or S401c), the above method further includes: the first communication device sending the second data based on the first multiple access method; sending the second data based on the first multiple access method includes: sending the second data to the first node and the second node through some or all of the same resources; wherein, the second data is obtained by processing the first test data and the second test data through the first AI model, the first test data corresponds to the first node, and the second test data corresponds to the second node. Accordingly, the first information mentioned above can be generated based on the second information, which is used to determine the processing performance of the first node on the second data and the processing performance of the second node on the second data.

[0207] Optionally, the second information may further indicate the model processing capabilities of the first node and the second node, or the second information may further indicate the AI ​​models deployed by the first node and the AI ​​models deployed by the second node. In this way, the first communication device can determine the second AI model corresponding to the first node and the third AI model corresponding to the second node based on Implementation Example 2 or Implementation Example 3. Furthermore, the first communication device can determine whether these two AI models have a mapping relationship with the first AI model based on one or more mapping relationships. Subsequently, in Implementation Example 3, if there is no mapping relationship between the first AI model and at least one of the two AI models (the second AI model and the third AI model), the first communication device can send second data based on the first multiple access method to implement the aforementioned test verification process through the test process corresponding to the second data.

[0208] In Example 4, the first communication device can instruct one or more nodes to communicate using a first multiple access method based on their processing performance of the test data. For example, the first communication device can instruct nodes whose processing performance of the test data meets the pre-configured data quality requirements to communicate using the first multiple access method, so that the multiple access method used by these nodes during communication can be adapted to their processing performance of the test data, thereby improving the model processing efficiency.

[0209] Optionally, the aforementioned processing performance can be characterized by one or more parameters, such as processing latency, the similarity between the processing result and the labeled data, or the performance of downstream tasks (e.g., the performance of classification tasks is measured by accuracy). The similarity can include one or more of the following: mean square error (MSE), normalized mean square error (NMSE), Kullback-Leibler divergence (KL), Jensen-Shannon divergence (JS), cosine similarity, Euclidean norm (also known as L2 norm), and Frobenius norm (also known as F norm).

[0210] For example, a node whose performance in processing test data meets the pre-configured data quality requirements can be a node whose processing latency for the test data is within a certain range (e.g., less than 1 minute).

[0211] For example, taking the similarity as MSE as mentioned above, a node whose processing performance of the test data meets the pre-configured data quality requirements can be a node whose processing result obtained from processing the test data is within a certain range (e.g., greater than 80%) compared to the MSE of the label data.

[0212] Optionally, in Implementation Example 4, the second information can indicate processing performance in various ways. For example, the second information can indicate the processing performance of the first node on the second data and the processing performance of the second node on the second data. Alternatively, the second information can indicate the processing result of the first node on the second data and the processing result of the second node on the second data, enabling the recipient of the second information to determine the corresponding performance based on these processing results.

[0213] Optionally, in Implementation Example 4, the first communication device can determine, based on the processing performance of one or more nodes on the test data, that some nodes communicate using a first multiple access method. These nodes may include a first node and a second node whose processing performance on the test data meets pre-configured data quality requirements. The one or more nodes may include other nodes whose processing performance on the test data does not meet the pre-configured data quality requirements, and the first communication device may not instruct these other nodes to communicate using the first multiple access method.

[0214] In one possible implementation, the first communication device can obtain the aforementioned second information in a variety of ways.

[0215] In one method, the first communication device can receive information from the first node and information from the second node. That is, the second information can include information from the first node and information from the second node. The second information is obtained by communicating with these two nodes.

[0216] For example, in the above implementation example two, information from the first node can indicate the model processing capability of the first node, and information from the second node can indicate the model processing capability of the second node.

[0217] For example, in the above implementation example three, the information from the first node can indicate the second AI model that the first node has deployed (such as the model identifier, model index, etc. of the second AI model), and the information from the second node can indicate the third AI model that the second node has deployed (such as the model identifier, model index, etc. of the third AI model).

[0218] For example, in the above implementation example four, information from the first node can indicate the processing result of the first node on the second data, and information from the second node can indicate the processing result of the second node on the second data.

[0219] Method 2: The first communication device can obtain the aforementioned second information through a pre-configured method or by other device instructions.

[0220] For example, in the above implementation example three, the first communication device can determine the model processing capabilities of one or more nodes through information exchange processes (such as capability reporting processes, data transmission processes, etc.) with the first and second nodes based on historical communication. Alternatively, network-side devices (such as network management devices, access network devices, core network devices, or application servers, etc.) can store or maintain the model processing capabilities of one or more nodes, and the first communication device can determine the model processing capabilities of the one or more nodes through the instructions of the network-side devices. The one or more nodes may include the first node and / or the second node.

[0221] For example, in the above implementation example four, the network-side device (such as network management device, access network device, core network device or application server, etc.) can perform tests based on the second data to obtain the processing performance of the first node on the second data and the processing performance of the second node on the second data; thereafter, the first communication device can determine the processing performance of the first node on the second data and the processing performance of the second node on the second data based on the instructions of the network-side device.

[0222] In one possible implementation of Example 4, the second information is further used to determine a mapping relationship between the first AI model and the second AI model deployed by the first node, and the third AI model deployed by the second node. Therefore, when the second information is used to determine the processing performance of the first node and the second node on the second data, the first communication device can also determine the mapping relationship between different models based on the second information. For example, the first communication device can determine that the first and second nodes can communicate via a first multiple access method based on their processing performance on the test data. Correspondingly, a mapping relationship exists between the first AI model used to obtain the second data, the second AI model used by the first node to process the second data, and the third AI model used by the second node to process the second data. This allows the first communication device to maintain or save this mapping relationship, and subsequently, it can instruct the first multiple access method based on the maintained or saved mapping relationship, thereby improving the efficiency of the multiple access method instruction and reducing processing complexity.

[0223] Optionally, one or more mapping relationships in Implementation Example 1 above can be combined with Implementation Example 2, Implementation Example 3 or Implementation Example 4 above.

[0224] As an example, the first communication device can determine whether there is a personalized model that has a mapping relationship with the public model based on one or more mapping relationships in Implementation Example 1; if there is, the first communication device generates first information based on the mapping relationship to implement the indication of the first multiple access mode; if there is no such information, the first communication device determines second information based on the aforementioned Implementation Example 4, and generates first information based on the second information.

[0225] As another example, after the first communication device generates the first information based on the second information in the aforementioned implementation example four, the first communication device can determine the personalized model that has a mapping relationship with the public model based on the second information; thereafter, the first communication device can update this mapping relationship to one or more mapping relationships in implementation example one.

[0226] Please see Figure 6 The diagram below shows a possible exemplary block diagram of the communication device involved in the embodiments of this application. Figure 6 As shown, the communication device 600 may include modules or units for implementing the methods described in the embodiments above. In one possible design, the communication device 600 includes a communication unit 603. Optionally, the communication device 600 may further include a storage unit 601 and / or a processing unit 602. The storage unit 601 stores device program code and / or data, and the processing unit 602 processes data and / or signals. For example, the processing unit 602 may perform processing based on the program code and / or data stored in the storage unit 601.

[0227] In one possible design, the communication device 600 can be the first communication device in the above embodiments. For example, the first communication device can be a terminal or a communication module in a terminal, or a circuit or chip in a terminal responsible for communication functions, or a circuit or chip in a network device responsible for communication functions. The communication unit 603 is used to send first information, which indicates a first multiple access method. The communication unit 603 is also used to send first data based on the first multiple access method. Sending the first data based on the first multiple access method includes: the communication unit 603 sending the first data to a first node and a second node using some or all of the same resources; wherein the first data is obtained by processing first initial data and second initial data through a first artificial intelligence (AI) model, the first initial data corresponding to the first node, and the second initial data corresponding to the second node.

[0228] Optionally, the communication unit 603 is also used to receive second information; the device also includes a processing unit 602, which is used to generate the first information based on the second information.

[0229] In another possible design, the communication device 600 can be a second communication device in the above embodiments. For example, the second communication device can be a terminal or a communication module in a terminal, or a circuit or chip in a terminal responsible for communication functions, or a circuit or chip in a network device responsible for communication functions. The communication unit 603 is used to receive first information, which indicates a first multiple access mode; the communication unit 603 is also used to receive first data based on the first multiple access mode.

[0230] In one possible design, when the communication device 600 is a terminal or a communication module within a terminal, the function of the processing unit 602 can be implemented by one or more processors. Specifically, the processor may include a modem chip, or a system-on-a-chip (SoC) chip or a SIP chip containing a modem core. The function of the communication unit 603 can be implemented by transceiver circuitry.

[0231] In one possible design, when the communication device 600 is a circuit or chip in a terminal responsible for communication functions, such as a modem chip or a system-on-a-chip (SoC) or SIP chip containing a modem core, the function of the processing unit 602 can be implemented by a circuit system in the aforementioned chip that includes one or more processors or processor cores. The function of the communication unit 603 can be implemented by an interface circuit or data transceiver circuit on the aforementioned chip.

[0232] It is understood that the division of units in the above-described device is merely a logical functional division. One function can correspond to one functional unit, or two or more functions can be integrated into one functional unit. In actual implementation, all or some units can be integrated onto a single physical entity, or distributed across different physical entities. Furthermore, the aforementioned functional units can be implemented in hardware, software, or a combination of both. Whether a function is executed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for specific applications, but such implementations should not be considered beyond the scope of this application.

[0233] In one example, the functional unit in any of the above devices may be one or more integrated circuits configured to implement the above methods, such as: one or more ASICs, or one or more CPUs, one or more microcontroller units (MCUs), one or more digital signal processors (DSPs), or one or more field programmable gate arrays (FPGAs), or a combination of at least two of these integrated circuit forms.

[0234] In one example, storage unit 601 may include random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory and / or registers, etc.

[0235] Please see Figure 7 This is a schematic diagram of the structure of a terminal 700 provided in an embodiment of this application. The terminal 700 can correspond to... Figure 4a or Figure 4b The first or second communication device shown is used to implement the operation of the first or second communication device in the above embodiments. For example... Figure 7 As shown, the terminal includes: one or more antennas 710, a radio frequency processing system 720, and a processor system 730.

[0236] It is understood that the first or second communication device can be a terminal 700 or a processor system 730.

[0237] In the downlink or sidelink direction, the RF processing system 720 receives RF signals through the antenna 710 and sends the RF-processed signals to the processor system 730 for further processing. In the uplink or sidelink direction, the processor system 730 processes the terminal-side information and sends it to the RF processing system 720, which then processes the signal and transmits it through the antenna 710.

[0238] In one example, the radio frequency (RF) processing system 720 serves as the communication interface for external communication of the terminal and may include an RF front end (RFFE) 721 and an RF transceiver 722. The RFFE 721 is primarily used for one or more processing operations, such as shaping, passband selection, or gain adjustment, on the RF signals received by the antenna or those to be transmitted through the antenna. It may include one or more components such as RF switches, duplexers, filters, power amplifiers, antenna tuners, and low-noise amplifiers. The RFFE 721 can be a circuit system composed of multiple discrete components or integrated into one or more chips. The RF transceiver 722 processes the RF signals received by the RFFE into baseband / IF signals for further processing by the processor system 730, and processes the baseband / IF signals provided by the processor system 730 into RF signals for transmission to the RFFE 721. The baseband / IF signals transmitted between the RF transceiver 722 and the processor system 730 can be digital or analog signals. The RF transceiver 722 can be implemented by one or more chips, which are commonly referred to as RF ICs.

[0239] In one example, the processor system 730 may include one or more processors for processing signals and executing one or more communication protocols. Optionally, the processor system 730 may also include a memory 736. In one example, the one or more processors include at least one baseband processor 731 (also known as a modem processor). The memory 736 is used to store data and / or computer program instructions. Optionally, the processor system 730 may also include one or more application processors 732 for implementing processing of the terminal operating system and application layer. Optionally, the processor system 730 may also include one or more of a voice subsystem 733, a multimedia subsystem 734, or an interface circuit 735. The voice subsystem 733 is used to process voice signals, the multimedia subsystem 734 is used to handle multimedia-related operations, such as video encoding / decoding, image processing, etc., and the interface circuit 735 is used to enable communication with other terminal components, such as a display 740, an input device 750, a memory 760, etc. The above-mentioned components in the processor system 730 can communicate with each other via a bus or communication interface circuit.

[0240] In one example, the processor system 730 can be packaged as a single processor chip, such as a SoC chip or a SIP chip. In another example, the processor system 730 can be a system composed of multiple chips; for example, the baseband processor 731 can be packaged as a single chip, or packaged with part or all of the circuitry of the radio frequency processing system into a single chip.

[0241] In one example, memory 736 can be on-chip memory, i.e., located on the processor system 730 chip. In another example, memory 760 can be off-chip memory, i.e. located outside the processor system 730 chip.

[0242] In one example, the baseband processor 731 may include one or more processor cores 7311 and interface circuitry 7314. The one or more processor cores 7311 are used to process signals and execute one or more communication protocols. Optionally, the baseband processor 731 may also include a memory 7312 for storing at least a portion of the corresponding computer program instructions and / or data. In one example, the one or more processor cores 7311 execute the computer program instructions stored in the memory 7312 to perform the relevant operations (such as sending or receiving first information) in the above method embodiments. In this disclosure, memory 7312 is used to store corresponding computer program instructions and / or data. This can mean that memory 7312 stores all corresponding computer program instructions and / or data for execution by processor core 7311; or it can mean that memory 7312 stores a portion of corresponding computer program instructions and / or data, including the computer program instructions and / or data currently required to be executed by processor core 7311. Memory 7312 can store different portions of computer program instructions and / or data multiple times for execution by processor core 7311 to implement the relevant operations in the above method embodiments. Interface circuit 7314 serves as a communication interface for communication with other components, such as transmitting signals with radio frequency processing system 720, communicating with other subsystems and related components of processor system 730 via bus, such as transmitting data control signals with application processor 732, and transmitting data or computer program instructions with memory 736 or memory 760. Optionally, in order to reduce the load on the processor core, a baseband signal processing circuit 7313 can be set to perform at least some of the baseband signal processing work, including one or more of signal demodulation, modulation, encoding or decoding.

[0243] In one example, the communication device provided in this application may be a terminal 700, a communication module including a processor system 730 and a radio frequency system 720, or a baseband processor 731.

[0244] The processor, processor system, application processor, baseband processor, processor circuit or processor core mentioned above can be collectively referred to as a processor. The processor may include one or more of the following: CPU, DSP, microprocessor unit (MPU), MCU, GPU, FPGA, artificial intelligence processor (AI processor) or NPU.

[0245] The aforementioned memory may include one or more of the following storage media: random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), phase-change memory (PCM), resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), cache, register, read-only memory (ROM), flash memory, erasable programmable read-only memory (EPROM), hard disk, etc. In one example, computer program instructions for executing the above embodiments may be stored in non-volatile memory, such as at least a portion of the aforementioned memory 760 (e.g., one or more of ROM, flash memory, EPROM, or hard disk). When the terminal is running, the corresponding computer program instructions may be partially or wholly loaded onto a memory with a faster transfer speed than the processor, such as at least a portion of memory 736 and / or memory 7312 (e.g., one or more of RAM, SRAM, DRAM, PCM, RERAM, MRAM, FRAM, cache, or register), for the processor to execute in order to implement the steps in the above method embodiments.

[0246] In one example, the RF transceiver 722 and the RF front-end 721 can also be packaged in a single chip. In another example, the RF transceiver 722, the RF front-end 721, and the baseband processor 731 can also be packaged in a single chip.

[0247] This application also provides a computer-readable storage medium for storing one or more computer-executable instructions. When the computer-executable instructions are executed by a processor, the processor performs the method described in the possible implementations of the first or second communication device in the foregoing embodiments.

[0248] This application also provides a computer program product (or computer program) that, when executed by a processor, executes the method described above for the possible implementation of the first or second communication device.

[0249] This application also provides a chip system including at least one processor for supporting a communication device in implementing the functions involved in the possible implementations of the communication device described above. Optionally, the chip system further includes an interface circuit that provides program instructions and / or data to the at least one processor. In one possible design, the chip system may further include a memory for storing the program instructions and data necessary for the communication device. The chip system may be composed of chips or may include chips and other discrete devices, wherein the communication device may specifically be the first communication device or the second communication device in the aforementioned method embodiments.

[0250] This application also provides a communication system, which includes a first communication device and a second communication device in any of the above embodiments.

[0251] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0252] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0253] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0254] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, etc.) containing computer-usable program code.

[0255] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0256] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0257] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

Claims

1. A communication method, characterized in that, include: Send a first message, which is used to indicate a first multiple access mode; First data is sent based on the first multiple access method; Sending the first data based on the first multiple access method includes: sending the first data to a first node and a second node through some or all of the same resources; wherein the first data is obtained by processing the first initial data and the second initial data through a first artificial intelligence (AI) model, the first initial data corresponds to the first node, and the second initial data corresponds to the second node.

2. The method according to claim 1, characterized in that, The method further includes: Receive the second message; The first information is generated based on the second information.

3. The method according to claim 2, characterized in that, The second information is used to indicate the model processing capabilities of the first node and the second node.

4. The method according to claim 3, characterized in that, The method further includes: Based on the second information, a second AI model corresponding to the first node and a third AI model corresponding to the second node are determined; wherein, the first AI model has a mapping relationship with the second AI model and the third AI model.

5. The method according to claim 2, characterized in that, The second information is used to indicate the second AI model that has been deployed on the first node and the third AI model that has been deployed on the second node; The first AI model has a mapping relationship with the second AI model and the third AI model.

6. The method according to claim 2, characterized in that, The method further includes: Sending second data based on the first multiple access method; the sending of second data based on the first multiple access method includes: sending the second data to the first node and the second node through some or all of the same resources; wherein, the second data is obtained by processing the first test data and the second test data through the first AI model, the first test data corresponds to the first node, and the second test data corresponds to the second node; The second information is used to determine the processing performance of the first node on the second data and the processing performance of the second node on the second data.

7. The method according to claim 6, characterized in that, The second information is also used to determine that there is a mapping relationship between the first AI model and the second AI model deployed on the first node and the third AI model deployed on the second node.

8. The method according to claim 1, characterized in that, The first AI model has a mapping relationship with the second AI model deployed on the first node and the third AI model deployed on the second node.

9. The method according to any one of claims 1 to 8, characterized in that, The first data includes one or more tokens.

10. A communication device, characterized in that, Includes a unit for performing the method as described in any one of claims 1 to 9.

11. A readable storage medium, characterized in that, The storage medium stores a computer program or instructions, which, when executed by a communication device, implement the method as described in any one of claims 1 to 9.

12. A computer program product, characterized in that, Includes a computer program or instructions that, when executed, implement the method as described in any one of claims 1 to 9.

Images