[go: up one dir, main page]

WO2025092160A1 - 一种通信方法及相关设备 - Google Patents

一种通信方法及相关设备 Download PDF

Info

Publication number
WO2025092160A1
WO2025092160A1 PCT/CN2024/114227 CN2024114227W WO2025092160A1 WO 2025092160 A1 WO2025092160 A1 WO 2025092160A1 CN 2024114227 W CN2024114227 W CN 2024114227W WO 2025092160 A1 WO2025092160 A1 WO 2025092160A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
index
communication device
neural network
processing
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
PCT/CN2024/114227
Other languages
English (en)
French (fr)
Inventor
徐晨
张公正
王坚
李榕
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Publication of WO2025092160A1 publication Critical patent/WO2025092160A1/zh
Pending legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W8/00—Network data management
    • H04W8/02—Processing of mobility data, e.g. registration information at HLR [Home Location Register] or VLR [Visitor Location Register]; Transfer of mobility data, e.g. between HLR, VLR or external networks
    • H04W8/08—Mobility data transfer
    • H04W8/14—Mobility data transfer between corresponding nodes
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W8/00—Network data management
    • H04W8/22—Processing or transfer of terminal data, e.g. status or physical capabilities
    • H04W8/24—Transfer of terminal data

Definitions

  • the present application relates to the field of communications, and in particular to a communication method and related equipment.
  • Wireless communication can be the transmission communication between two or more communication nodes without propagation through conductors or cables.
  • the communication nodes generally include network equipment and terminal equipment.
  • communication nodes generally have signal transceiving capabilities and computing capabilities.
  • the computing capabilities of network devices mainly provide computing power support for signal transceiving capabilities (for example: sending and receiving signals) to achieve communication between network devices and other communication nodes.
  • the computing power of communication nodes may have surplus computing power in addition to providing computing power support for the above communication tasks. Therefore, how to utilize this computing power is a technical problem that needs to be solved urgently.
  • the present application provides a communication method and related equipment, which are used to enable the computing power of communication nodes to be applied to artificial intelligence (AI) processing of neural networks while also improving the flexibility of neural network deployment.
  • AI artificial intelligence
  • the present application provides a communication method, which is performed by a first communication device, which may be a communication device (e.g., a terminal device or a network device), or the first communication device may be a partial component in a communication device (e.g., a processor, a chip, or a chip system, etc.), or the first communication device may also be a logic module or software that can implement all or part of the functions of the communication device.
  • the communication method is described as being performed by the first communication device.
  • the first communication device may be a terminal device or a network device.
  • the second data sent by the first communication device is obtained based on the first data
  • the subsequent receiver of the second data (for example, the second communication device) can process the second data to obtain the third data.
  • the index of the first data in the first data set is used to determine the fourth data in the second data set
  • the fourth data is the label data corresponding to the first data.
  • the receiver of the second data can determine the label data in the second data set and process the third data based on the label data.
  • the second data and/or the third data are obtained based on a neural network, that is, the neural network used for AI processing can include a neural network deployed in the first communication device and/or a neural network deployed in the second communication device.
  • the index of the first data in the first data set is used to determine the label data (i.e., the fourth data) corresponding to the first data in the second data set, and the index satisfies at least one of the above items.
  • the second communication device can determine the label data in the second data set based on the resource carrying the data or the number of times the data in the data set is processed. In this way, the air interface overhead can be reduced to improve communication efficiency.
  • the second data is obtained by processing the first data based on the first neural network
  • the third data is obtained by processing the second data based on the second neural network.
  • the first data set and the second data set may be included in A data set.
  • N input data and M label data may be included, where N and M are both positive integers; and the first data set may include the N input data, and the second data set may include the M label data.
  • the first data set may be called an input data set, a neural network input data set, etc.
  • the second data set may be called a label data set, a neural network label data set, etc.
  • the AI neural network may be processed based on the same data set to implement iteration, update, etc. of the AI neural network.
  • the first data set may include N pieces of input data in one data set, and for this purpose, the first data set may also be replaced by other descriptions, such as N pieces of input data, N pieces of input data in a data set, etc.
  • the second data set may include M pieces of label data in the one data set, and for this purpose, the second data set may also be replaced by other descriptions, such as M pieces of label data, M pieces of label data in a data set, etc.
  • the value of N is equal to the value of M.
  • the value of N may be greater than or equal to the value of M.
  • the value of N may be less than or equal to the value of M.
  • AI neural network
  • AI neural network machine learning
  • AI processing AI neural network processing
  • the data involved (such as first data, second data, third data, and fourth data, etc.) can be replaced by information, signals, etc.
  • the first neural network can be called a neural network deployed at the transmitting end, an encoding neural network, an AI encoding neural network, etc.
  • the second neural network can be called a neural network deployed at the receiving end, a decoding neural network, an AI decoding neural network, etc.
  • the index is determined based on the resource carrying the second data, including: the value of the index is determined by at least one of the time domain resource index of the resource, the frequency domain resource index of the resource, and the resource block size of the resource.
  • the index of the first data in the first data set is determined based on the resources carrying the second data
  • the index can be specifically determined by at least one of the above items to improve the flexibility of the solution implementation.
  • the method when the index is determined based on the number of times data in the first data set is processed, the method further includes: the first communication device sends first information, and the first information is used to indicate the index of the first data in the first data set.
  • the second communication device may determine the index based on the number of times the data in the second data set is processed. Accordingly, the first communication device may also send the first information so that the second communication device can determine the index of the first data in the first data set based on the first information, and subsequently determine the fourth data in the second data set based on the index.
  • the first communication device can perform multiple processing based on the data in the first data set to obtain and send the processing results (for example, one of the processing results is the second data obtained based on the first data).
  • the first communication device can send corresponding indexes for some or all of the processing results (for example, send first information for the second data processing result). Thereafter, sending the corresponding index based on the part or all of the processing results can achieve alignment of the understanding of the index by the data sender and receiver, so as to avoid data processing errors caused by the misalignment of the understanding of the index by the data sender and receiver, thereby improving the robustness of the system.
  • alignment may mean that when there are interactive messages/data/information between different communication devices, the two have a consistent understanding of the meaning, configuration method, index in the data set, etc. of the interactive messages/data/information.
  • the first communication device may send an index corresponding to a partial processing result, without sending an index corresponding to the entire processing result, thereby saving overhead.
  • the first information is one of multiple information transmitted based on a first period; the method also includes: the first communication device receives or sends configuration information, and the configuration information is used to configure the first period.
  • the first information may be one of the periodic information transmitted based on the first period, before which the first period
  • a communication device can receive or send configuration information for configuring the first cycle, so that the first communication device can serve as both the configurator and the configured party of the first cycle, which can align the understanding of the first cycle by both the data sender and receiver, and improve the flexibility of the solution implementation.
  • the method further includes: the first communication device receiving indication information indicating the first data set; and/or the first communication device sending indication information indicating the second data set.
  • the first communication device can be used as the configured party of the first data set, and/or the first communication device can be used as the configured party of the second data set, so that the data sender and receiver can obtain the data set before exchanging data.
  • the first data set may be preconfigured for the first communication device, and/or the second data set may be preconfigured for the second communication device. In this way, overhead can be reduced.
  • the method further includes: the first communication device receiving gradient information and/or a result of a loss function determined based on the third data and the fourth data.
  • the receiver of the second data can process the second data to obtain the third data, and the receiver can also determine and send the corresponding gradient information and/or the result of the loss function based on the third data and the fourth data, so that the first communication device can update or iterate the first neural network based on the gradient information and/or the result of the loss function after receiving the gradient information and/or the result of the loss function.
  • the method further includes: the first communication device receiving or sending indication information indicating that the index satisfies the at least one item.
  • the first communication device can also receive or send indication information indicating that the index satisfies at least one item, so that the data sender and receiver can align their understanding of the index of the data in the data set based on the indication information, so as to avoid data processing errors caused by the misaligned understanding of the index by the data sender and receiver, thereby improving the robustness of the system.
  • the second aspect of the present application provides a communication method, which is performed by a second communication device, which may be a communication device (such as a terminal device or a network device), or the second communication device may be a partial component in a communication device (such as a processor, a chip or a chip system, etc.), or the second communication device may also be a logic module or software that can implement all or part of the functions of the communication device.
  • a second communication device which may be a communication device (such as a terminal device or a network device), or the second communication device may be a partial component in a communication device (such as a processor, a chip or a chip system, etc.), or the second communication device may also be a logic module or software that can implement all or part of the functions of the communication device.
  • the communication method is described as being performed by the second communication device, wherein the second communication device may be a terminal device or a network device.
  • a second communication device receives second data, which is obtained by processing first data, and the first data is data in the first data set; wherein the index of the first data in the first data set is used to determine fourth data in the second data set, and the second data set includes label data corresponding to the data in the first data set, and the fourth data is label data corresponding to the first data; the second communication device determines third data based on the second data; wherein the second data is obtained by processing the first data based on a first neural network, and/or the third data is obtained by processing the second data based on a second neural network; the index satisfies at least one of the following: the index is determined based on the resources carrying the second data; the index is determined based on the number of times the data in the second data set is processed.
  • the second data received by the second communication device is obtained based on the first data, and the second communication device can subsequently process the second data to obtain the third data.
  • the index of the first data in the first data set is used to determine the fourth data in the second data set, and the fourth data is the label data corresponding to the first data.
  • the second communication device can determine the label data in the second data set, and process the third data based on the label data.
  • the second data and/or the third data are obtained based on a neural network, that is, the neural network used for AI processing may include a neural network deployed in the first communication device and/or a neural network deployed in the second communication device.
  • the index of the first data in the first data set is used to determine the label data (i.e., the fourth data) corresponding to the first data in the second data set, and the index satisfies at least one of the above items.
  • the second communication device can determine the label data in the second data set based on the resource carrying the data or the number of times the data in the data set is processed. In this way, the air interface overhead can be reduced to improve communication efficiency.
  • the index is determined based on the resource carrying the second data, including: the value of the index is determined by at least one of the time domain resource index of the resource, the frequency domain resource index of the resource, and the resource block size of the resource.
  • the index may be specifically determined by at least one of the above items to enhance the flexibility of implementing the solution.
  • the method when the index is determined based on the number of times data in the second data set is processed, the method further includes: the second communication device sends first information, and the first information is used to indicate the index of the first data in the first data set.
  • the second communication device may determine the index based on the number of times the data in the second data set is processed. Accordingly, the second communication device may also receive the first information, so that the second communication device can determine the index of the first data in the first data set based on the first information, and subsequently determine the fourth data in the second data set based on the index.
  • the first communication device can perform multiple processing based on the data in the first data set to obtain and send the processing results (for example, one of the processing results is the second data obtained based on the first data).
  • the first communication device can send corresponding indexes for some or all of the processing results (for example, send first information for the second data processing result). Thereafter, sending the corresponding index based on the part or all of the processing results can achieve alignment of the understanding of the index by the data sender and receiver, so as to avoid data processing errors caused by the misalignment of the understanding of the index by the data sender and receiver, thereby improving the robustness of the system.
  • the first communication device may send an index corresponding to a partial processing result, without sending an index corresponding to the entire processing result, thereby saving overhead.
  • the first information is one of multiple information transmitted based on a first period; the method also includes: the second communication device receives or sends configuration information, and the configuration information is used to configure the first period.
  • the first information can be one of the periodic information transmitted based on the first cycle.
  • the second communication device can receive or send configuration information for configuring the first cycle, so that the second communication device can serve as both the configurator of the first cycle and the configured party of the first cycle. This can align the understanding of the first cycle between the data sender and receiver, and improve the flexibility of the solution implementation.
  • the method further includes: the second communication device sending indication information indicating the first data set; and/or the second communication device receiving indication information indicating the second data set.
  • the second communication device can serve as the configurator of the first data set, and/or the second communication device can serve as the configured party of the second data set, so that the data sender and receiver can acquire the data set before exchanging data.
  • the first data set may be preconfigured for the first communication device, and/or the second data set may be preconfigured for the second communication device. In this way, overhead can be reduced.
  • the method further includes: the second communication device sends gradient information and/or a result of a loss function determined based on the third data and the fourth data.
  • the second communication device can process the second data to obtain the third data, and the second communication device can also determine and send the corresponding gradient information and/or the result of the loss function based on the third data and the fourth data, so that the first communication device can update or iterate the first neural network based on the gradient information and/or the result of the loss function after receiving the gradient information and/or the result of the loss function.
  • the method further includes: the second communication device receiving or sending indication information indicating that the index satisfies the at least one item.
  • the second communication device can also receive or send indication information indicating that the index satisfies at least one item, so that the data sender and receiver can align their understanding of the index of the data in the data set based on the indication information, so as to avoid data processing errors caused by the misaligned understanding of the index by the data sender and receiver, thereby improving the robustness of the system.
  • the third aspect of the present application provides a communication method, wherein the first communication device may be a communication device (such as a terminal device or a network device), or the first communication device may be a partial component in a communication device (such as a processor, a chip or a chip system, etc.), or the first communication device may also be a logic module or software that can implement all or part of the functions of the communication device.
  • the communication method is described as being executed by a first communication device, wherein the first communication device may be a terminal device or a network device.
  • the first communication device processes the first data to obtain the second data; the first communication device sends the second data and fourth data, wherein the second data is used to determine the third data, and the fourth data is the label data corresponding to the first data; wherein, The second data is obtained by processing the first data based on the first neural network, and/or the third data is obtained by processing the second data based on the second neural network.
  • the second data sent by the first communication device is obtained based on the first data, and the subsequent receiver of the second data (for example, the second communication device) can process the second data to obtain the third data.
  • the first communication device can also send fourth data, and the fourth data is the label data corresponding to the first data.
  • the receiver of the second data and the fourth data can process the third data based on the label data.
  • the second data and/or the third data are obtained based on a neural network, that is, the neural network used for AI processing may include a neural network deployed in the first communication device and/or a neural network deployed in the second communication device.
  • the second data sent by the first communication device is the processing result of the first data
  • the fourth data sent by the first communication device is the label data corresponding to the first data.
  • the method further includes: the first communication device receiving gradient information and/or a result of a loss function determined based on the third data and the fourth data.
  • the receiver of the second data can process the second data to obtain the third data, and the receiver can also determine and send the corresponding gradient information and/or the result of the loss function based on the third data and the fourth data, so that the first communication device can update or iterate the first neural network based on the gradient information and/or the result of the loss function after receiving the gradient information and/or the result of the loss function.
  • the fourth aspect of the present application provides a communication method, which is performed by a second communication device, and the second communication device may be a communication device (such as a terminal device or a network device), or the second communication device may be a partial component in the communication device (such as a processor, a chip or a chip system, etc.), or the second communication device may also be a logic module or software that can realize all or part of the functions of the communication device.
  • the communication method is described as being performed by a second communication device, wherein the second communication device may be a terminal device or a network device.
  • the second communication device receives second data and fourth data, wherein the second data is obtained based on the first data, and the fourth data is the label data corresponding to the first data; the second communication device determines the third data based on the second data; wherein the second data is obtained based on the first neural network processing the first data, and/or the third data is obtained based on the second neural network processing the second data.
  • the second data received by the second communication device is obtained based on the first data, and the second communication device can subsequently process the second data to obtain the third data.
  • the second communication device can also receive fourth data, and the fourth data is the label data corresponding to the first data.
  • the second communication device can process the third data based on the label data.
  • the second data and/or the third data are obtained based on a neural network, that is, the neural network used for AI processing may include a neural network deployed in the first communication device and/or a neural network deployed in the second communication device.
  • the second data received by the second communication device is the processing result of the first data
  • the fourth data sent by the second communication device is the label data corresponding to the first data.
  • the method further includes: the second communication device sending gradient information and/or a result of a loss function determined based on the third data and the fourth data.
  • the second communication device can process the second data to obtain the third data, and the second communication device can also determine and send the corresponding gradient information and/or the result of the loss function based on the third data and the fourth data, so that the first communication device can update or iterate the first neural network based on the gradient information and/or the result of the loss function after receiving the gradient information and/or the result of the loss function.
  • a communication method is provided, which is performed by a first communication device, which may be a communication device (such as a terminal device or a network device), or the first communication device may be a partial component in a communication device (such as a processor, a chip or a chip system, etc.), or the first communication device may also be a logic module or software that can realize all or part of the functions of the communication device.
  • the communication method is described as being performed by a first communication device, wherein the first communication device may be a terminal device or a network device.
  • the first communication device processes the first data to obtain the second data; wherein the first data is the data in the first data set; the first communication device sends the second data and a first index, and the second data is used to determine the third data; wherein the first index is used to determine the fourth data in the second data set, and the second data set includes the label data corresponding to the data in the first data set, and the fourth data is the label data corresponding to the first data; wherein the second data is obtained based on the first neural network processing the first data, and/or the fourth data is obtained based on the neural network processing the second data.
  • the second data sent by the first communication device is obtained based on the first data, and the subsequent receiver of the second data (for example, the second communication device) can process the second data to obtain the third data.
  • the first communication device can also send a first index.
  • the receiver of the second data and the fourth data can process the third data based on the label data.
  • the second data and/or the third data are obtained based on a neural network, that is, the neural network used for AI processing may include a neural network deployed in the first communication device and/or a neural network deployed in the second communication device.
  • the second data sent by the first communication device is the processing result of the first data
  • the fourth data sent by the first communication device is the label data corresponding to the first data.
  • the second data is obtained based on the first neural network processing the first data
  • the third data is obtained based on the second neural network processing the second data.
  • the first data set and the second data set can be included in one data set.
  • N input data and M label data can be included, and N and M are both positive integers; and the first data set can include the N input data, and the second data set can include the M label data.
  • the first data set can be called an input data set, a neural network input data set, etc.
  • the second data set can be called a label data set, a neural network label data set, etc.
  • the AI neural network can be processed based on the same data set to achieve iteration, update, etc. of the AI neural network.
  • the first index is determined by a second index of the first data in the first data set.
  • the value of N is equal to the value of M.
  • the first index may be the same as the second index of the first data in the first data set.
  • the label data of the i-th input data in the first data set is the j-th data in the second data set, i is the first index, j is the second index, and i is equal to j. That is, the label data of the first data of N input data is the first data in M label data, the label data of the second data of N input data is the second data in M label data, and so on, the label data of the N-th data of N input data is the M-th data in M label data (N is equal to M).
  • the first index may be partially or completely different from the second index of the first data in the first data set.
  • the label data of the i-th input data in the first data set is the j-th data in the second data set
  • i is the first index
  • j is the second index
  • i and j may be partially or completely unequal.
  • the mapping relationship between i and j may be preconfigured.
  • the mapping relationship between i and j can be preconfigured, i can be traversed from 1 to N, and j can be traversed from N (N equals M) to 1.
  • N N equals M
  • the label data of the first copy of N input data is the third copy of M label data
  • the label data of the second copy of N input data is the second copy of M label data
  • the label data of the third copy of N input data is the first copy of M label data.
  • the first index can be different from the second index part of the first data in the first data set.
  • mapping relationship between i and j can be preconfigured
  • the mapping relationship between i and j can be configured or preconfigured to align the data sender and receiver.
  • the label data of the first data of N input data is the third data of M label data
  • the label data of the second data of N input data is the M label data.
  • the first data in the data set is the first data
  • the label data of the third data of the N input data is the second data in the M label data.
  • the first index can be completely different from the second index of the first data in the first data set.
  • mapping relationship between the first index and the second index of the first data in the first data set is preconfigured.
  • the value of N may be greater than or equal to the value of M.
  • the first index may be less than or equal to the second index of the first data in the first data set.
  • the value of N may be less than or equal to the value of M.
  • the first index may be greater than or equal to the second index of the first data in the first data set.
  • the receiver of the second data can process the second data to obtain the third data, and the receiver can also determine and send the corresponding gradient information and/or the result of the loss function based on the third data and the fourth data, so that the first communication device can update or iterate the first neural network based on the gradient information and/or the result of the loss function after receiving the gradient information and/or the result of the loss function.
  • a communication method is provided, which is performed by a second communication device, which may be a communication device (such as a terminal device or a network device), or the second communication device may be a partial component in the communication device (such as a processor, a chip or a chip system, etc.), or the second communication device may also be a logic module or software that can realize all or part of the functions of the communication device.
  • the communication method is described as being performed by a second communication device, wherein the second communication device may be a terminal device or a network device.
  • the second communication device receives second data and a first index, the second data is obtained based on the first data, and the first data is data in the first data set; wherein the first index is used to determine fourth data in the second data set, the second data set includes label data corresponding to the data in the first data set, and the fourth data is label data corresponding to the first data; the second communication device determines third data based on the second data; wherein the second data is obtained based on the first neural network processing the first data, and/or the third data is obtained based on the second neural network processing the second data.
  • the second data received by the second communication device is obtained based on the first data, and the second communication device can subsequently process the second data to obtain the third data.
  • the second communication device can also receive a first index. Accordingly, after receiving the second data, the second communication device can process the third data based on the label data.
  • the second data and/or the third data are obtained based on a neural network, that is, the neural network used for AI processing may include a neural network deployed in the first communication device and/or a neural network deployed in the second communication device.
  • the second data received by the second communication device is the processing result of the first data
  • the fourth data received by the second communication device is the label data corresponding to the first data.
  • the first index is determined by a second index of the first data in the first data set.
  • the method further includes: the second communication device sends gradient information and/or a result of a loss function determined based on the third data and the fourth data.
  • the second communication device can process the second data to obtain the third data, and the second communication device can also determine and send the corresponding gradient information and/or the result of the loss function based on the third data and the fourth data, so that the first communication device can update or iterate the first neural network based on the gradient information and/or the result of the loss function after receiving the gradient information and/or the result of the loss function.
  • a seventh aspect of the present application provides a communication device, which is a first communication device, and includes a transceiver unit and a processing unit, wherein the processing unit is used to process first data to obtain second data; wherein the first data is data in the first data set; the receiving unit is used to process the first data to obtain second data; The sending unit is used to send the second data, and the second data is used to determine the third data; wherein the index of the first data in the first data set is used to determine the fourth data in the second data set, and the second data set includes label data corresponding to the data in the first data set, and the fourth data is the label data corresponding to the first data; wherein the second data is obtained based on the first neural network processing the first data, and/or the third data is obtained based on the second neural network processing the second data; the index satisfies at least one of the following: the index is determined based on the resources carrying the second data; the index is determined based on the number of times the data in the first data set is processed.
  • the constituent modules of the communication device can also be used to execute the steps performed in each possible implementation method of the first aspect and achieve corresponding technical effects.
  • the constituent modules of the communication device can also be used to execute the steps performed in each possible implementation method of the first aspect and achieve corresponding technical effects.
  • a communication device which is a second communication device, and includes a transceiver unit and a processing unit, wherein the transceiver unit is used to receive second data, and the second data is obtained by processing first data, and the first data is data in the first data set; wherein the index of the first data in the first data set is used to determine fourth data in the second data set, and the second data set includes label data corresponding to the data in the first data set, and the fourth data is label data corresponding to the first data; the processing unit is used to determine third data based on the second data; wherein the second data is obtained by processing the first data based on a first neural network, and/or the third data is obtained by processing the second data based on a second neural network; the index satisfies at least one of the following: the index is determined based on the resources carrying the second data; the index is determined based on the number of times the data in the second data set is processed.
  • the constituent modules of the communication device can also be used to execute the steps performed in each possible implementation method of the second aspect and achieve corresponding technical effects.
  • the constituent modules of the communication device can also be used to execute the steps performed in each possible implementation method of the second aspect and achieve corresponding technical effects.
  • a ninth aspect of the present application provides a communication device, which is a first communication device, and includes a transceiver unit and a processing unit, wherein the processing unit is used to process first data to obtain second data; the transceiver unit is used to send the second data and fourth data, wherein the second data is used to determine third data, and the fourth data is label data corresponding to the first data; wherein the second data is obtained by processing the first data based on a first neural network, and/or the third data is obtained by processing the second data based on a second neural network.
  • the constituent modules of the communication device can also be used to execute the steps performed in each possible implementation method of the third aspect and achieve corresponding technical effects.
  • the constituent modules of the communication device can also be used to execute the steps performed in each possible implementation method of the third aspect and achieve corresponding technical effects.
  • the tenth aspect of the present application provides a communication device, which is a second communication device, and includes a transceiver unit and a processing unit.
  • the transceiver unit is used to receive second data and fourth data, wherein the second data is obtained based on the first data, and the fourth data is label data corresponding to the first data;
  • the processing unit is used to determine third data based on the second data; wherein the second data is obtained based on the first neural network processing the first data, and/or the third data is obtained based on the second neural network processing the second data.
  • the constituent modules of the communication device can also be used to execute the steps performed in each possible implementation method of the fourth aspect and achieve corresponding technical effects.
  • the constituent modules of the communication device can also be used to execute the steps performed in each possible implementation method of the fourth aspect and achieve corresponding technical effects.
  • a communication device which is a first communication device, and includes a transceiver unit and a processing unit, wherein the processing unit is used to process first data to obtain second data; wherein the first data is data in a first data set; the transceiver unit is used to send the second data and a first index, and the second data is used to determine third data; wherein the first index is used to determine fourth data in the second data set, and the second data set includes label data corresponding to the data in the first data set, and the fourth data is label data corresponding to the first data; wherein the second data is obtained by processing the first data based on a first neural network, and/or the fourth data is obtained by processing the second data based on a neural network.
  • the constituent modules of the communication device can also be used to execute the steps performed in each possible implementation method of the fifth aspect and achieve corresponding technical effects.
  • the constituent modules of the communication device can also be used to execute the steps performed in each possible implementation method of the fifth aspect and achieve corresponding technical effects.
  • the twelfth aspect of the present application provides a communication device, which is a second communication device, and includes a transceiver unit and a processing unit.
  • the transceiver unit is used to receive second data and a first index, and the second data is obtained based on the first data, and the first data is the data in the first data set; wherein the first index is used to determine fourth data in the second data set, and the second data set includes label data corresponding to the data in the first data set, and the fourth data is the label data corresponding to the first data;
  • the processing unit is used to determine third data based on the second data; wherein the second data is obtained based on the first neural network processing the first data, and/or the third data is obtained based on the second neural network processing the second data.
  • the component modules of the communication device can also be used to execute the operations performed in each possible implementation of the sixth aspect.
  • the steps and the corresponding technical effects can be achieved.
  • a thirteenth aspect of the present application provides a communication device, comprising at least one processor, wherein the at least one processor is coupled to a memory; the memory is used to store programs or instructions; the at least one processor is used to execute the program or instructions so that the device implements the method described in any possible implementation method of any one of the first to sixth aspects.
  • the present application provides a communication device, comprising at least one logic circuit and an input/output interface; the logic circuit is used to execute the method described in any possible implementation method of any one of the first to sixth aspects.
  • a fifteenth aspect of the present application provides a communication system, which includes the above-mentioned first communication device and second communication device.
  • a computer-readable storage medium which is used to store one or more computer-executable instructions.
  • the processor executes the method described in any possible implementation of any aspect of the first to sixth aspects above.
  • the seventeenth aspect of the present application provides a computer program product (or computer program).
  • the processor executes the method described in any possible implementation of any one of the first to sixth aspects above.
  • the present application provides a chip system, which includes at least one processor for supporting a communication device to implement the method described in any possible implementation method of any one of the first to sixth aspects.
  • the chip system may also include a memory for storing program instructions and data necessary for the first communication device.
  • the chip system may be composed of a chip, or may include a chip and other discrete devices.
  • the chip system also includes an interface circuit, which provides program instructions and/or data for the at least one processor.
  • the technical effects brought about by any design method in the seventh to eighteenth aspects can refer to the technical effects brought about by different design methods in the above-mentioned first to sixth aspects, and will not be repeated here.
  • FIGS. 1a to 1c are schematic diagrams of a communication system provided by the present application.
  • FIG. 1d, FIG. 1e, and FIG. 2a to FIG. 2f are schematic diagrams of the AI processing process involved in the present application.
  • FIG3 is an interactive schematic diagram of the communication method provided by the present application.
  • FIGS 4 and 5 are schematic diagrams of the AI processing process provided by the present application.
  • 6 to 7 are interactive schematic diagrams of the communication method provided by the present application.
  • 8 to 12 are schematic diagrams of the communication device provided in the present application.
  • Terminal device It can be a wireless terminal device that can receive network device scheduling and instruction information.
  • the wireless terminal device can be a device that provides voice and/or data connectivity to users, or a handheld device with wireless connection function, or other processing devices connected to a wireless modem.
  • the terminal device can communicate with one or more core networks or the Internet via a radio access network (RAN).
  • RAN radio access network
  • the terminal device can be a mobile terminal device, such as a mobile phone (or "cellular" phone, mobile phone), a computer and a data card, for example, a portable, pocket-sized, handheld, computer-built-in or vehicle-mounted mobile device, which exchanges voice and/or data with the radio access network.
  • PCS personal communication service
  • SIP session initiation protocol
  • WLL wireless local loop
  • PDAs personal digital assistants
  • Pads computers with wireless transceiver functions, and other devices.
  • the wireless terminal device can also be called a system, a subscriber unit, a subscriber station, a mobile station, a mobile station (MS), a remote station, an access point (AP), a remote terminal device (remote terminal), an access terminal device (access terminal), a user terminal device (user terminal), a user agent (user agent), a subscriber station (SS), a customer premises equipment (CPE), a terminal, a user equipment (UE), a mobile terminal (MT), etc.
  • a system a subscriber unit, a subscriber station, a mobile station, a mobile station (MS), a remote station, an access point (AP), a remote terminal device (remote terminal), an access terminal device (access terminal), a user terminal device (user terminal), a user agent (user agent), a subscriber station (SS), a customer premises equipment (CPE), a terminal, a user equipment (UE), a mobile terminal (MT), etc.
  • the terminal device may also be a wearable device.
  • Wearable devices may also be referred to as wearable smart devices or smart wearable devices, etc., which are a general term for the application of wearable technology to intelligently design and develop wearable devices for daily wear, such as glasses, gloves, watches, clothing and shoes.
  • a wearable device is a portable device that is worn directly on the body or integrated into the user's clothes or accessories. Wearable devices are not only hardware devices, but also powerful functions achieved through software support, data interaction, and cloud interaction.
  • wearable smart devices include full-featured, large-size, and independent of smartphones to achieve complete or partial functions, such as smart watches or smart glasses, etc., as well as those that only focus on a certain type of application function and need to be used in conjunction with other devices such as smartphones, such as various types of smart bracelets, smart helmets, and smart jewelry for vital sign monitoring.
  • the terminal can also be a drone, a robot, a terminal in device-to-device (D2D) communication, a terminal in vehicle to everything (V2X), a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless terminal in industrial control, a wireless terminal in self driving, a wireless terminal in remote medical, a wireless terminal in smart grid, a wireless terminal in transportation safety, a wireless terminal in a smart city, a wireless terminal in a smart home, etc.
  • D2D device-to-device
  • V2X vehicle to everything
  • VR virtual reality
  • AR augmented reality
  • the terminal device may also be a terminal device in a communication system that evolves after the fifth generation (5th generation, 5G) communication system (e.g., a sixth generation (6th generation, 6G) communication system, etc.) or a terminal device in a public land mobile network (PLMN) that evolves in the future, etc.
  • 5G fifth generation
  • 6G sixth generation
  • PLMN public land mobile network
  • the 6G network can further expand the form and function of the 5G communication terminal
  • the 6G terminal includes but is not limited to a car, a cellular network terminal (with integrated satellite terminal function), a drone, and an Internet of Things (IoT) device.
  • IoT Internet of Things
  • the terminal device may also obtain AI services provided by the network device.
  • the terminal device may also have AI processing capabilities.
  • the network equipment can be a RAN node (or device) that connects a terminal device to a wireless network, which can also be called a base station.
  • RAN equipment are: base station, evolved NodeB (eNodeB), gNB (gNodeB) in a 5G communication system, transmission reception point (TRP), evolved Node B (eNB), radio network controller (RNC), Node B (NB), home base station (e.g., home evolved Node B, or home Node B, HNB), baseband unit (BBU), or wireless fidelity (Wi-Fi) access point AP, etc.
  • the network equipment may include a centralized unit (CU) node, a distributed unit (DU) node, or a RAN device including a CU node and a DU node.
  • CU centralized unit
  • DU distributed unit
  • RAN device including a CU node and a DU node.
  • the RAN node can also be a macro base station, a micro base station or an indoor station, a relay node or a donor node, or a wireless controller in a cloud radio access network (CRAN) scenario.
  • the RAN node can also be a server, a wearable device, a vehicle or an onboard device, etc.
  • the access network device in the vehicle to everything (V2X) technology can be a road side unit (RSU).
  • the RAN node can be a central unit (CU), a distributed unit (DU), a CU-control plane (CP), a CU-user plane (UP), or a radio unit (RU).
  • the CU and DU can be set separately, or can also be included in the same network element, such as a baseband unit (BBU).
  • BBU baseband unit
  • the RU can be included in a radio frequency device or a radio frequency unit, such as a remote radio unit (RRU), an active antenna unit (AAU) or a remote radio head (RRH).
  • CU or CU-CP and CU-UP
  • DU or RU may also have different names, but those skilled in the art can understand their meanings.
  • O-CU open CU
  • DU may also be called O-DU
  • CU-CP may also be called O-CU-CP
  • CU-UP may also be called O-CU-UP
  • RU may also be called O-RU.
  • CU, CU-CP, CU-UP, DU and RU are used as examples for description in this application.
  • Any unit of CU (or CU-CP, CU-UP), DU and RU in this application may be implemented by a software module, a hardware module, or a combination of a software module and a hardware module.
  • the protocol layer may include a control plane protocol layer and a user plane protocol layer.
  • the control plane protocol layer may include at least one of the following: a radio resource control (RRC) layer, a packet data convergence protocol (PDCP) layer, a radio link control (RLC) layer, a packet data convergence protocol (PDCP) layer, a radio link control (RLCP) layer, a radio link control (RLC ...
  • RRC radio resource control
  • PDCP packet data convergence protocol
  • RLC radio link control
  • PDCP packet data convergence protocol
  • RLCP radio link control
  • RLC radio link control
  • the user plane protocol layer may include at least one of the following: a service data adaptation protocol (SDAP) layer, a PDCP layer, an RLC layer, a MAC layer, or a physical layer.
  • SDAP service data adaptation protocol
  • the network device may be any other device that provides wireless communication functions for the terminal device.
  • the embodiments of the present application do not limit the specific technology and specific device form used by the network device. For the convenience of description, the embodiments of the present application do not limit.
  • the network equipment may also include core network equipment, such as mobility management entity (MME), home subscriber server (HSS), serving gateway (S-GW), policy and charging rules function (PCRF), public data network gateway (PDN gateway, P-GW) in the fourth generation (4G) network; access and mobility management function (AMF), user plane function (UPF) or session management function (SMF) in the 5G network.
  • MME mobility management entity
  • HSS home subscriber server
  • S-GW serving gateway
  • PDN gateway public data network gateway
  • P-GW public data network gateway
  • AMF access and mobility management function
  • UPF user plane function
  • SMF session management function
  • SMF session management function
  • 5G network equipment may also include other core network equipment in the 5G network and the next generation network of the 5G network.
  • the above-mentioned network device may also have a network node with AI capabilities, which can provide AI services for terminals or other network devices.
  • a network node with AI capabilities can provide AI services for terminals or other network devices.
  • it may be an AI node on the network side (access network or core network), a computing node, a RAN node with AI capabilities, a core network element with AI capabilities, etc.
  • the device for realizing the function of the network device may be a network device, or may be a device capable of supporting the network device to realize the function, such as a chip system, which may be installed in the network device.
  • the technical solution provided in the embodiment of the present application is described by taking the device for realizing the function of the network device as an example that the network device is used as the device.
  • Configuration and pre-configuration are used at the same time.
  • Configuration refers to the network device/server sending some parameter configuration information or parameter values to the terminal through messages or signaling, so that the terminal can determine the communication parameters or resources during transmission based on these values or information.
  • Pre-configuration is similar to configuration, and can be parameter information or parameter values pre-negotiated between the network device/server and the terminal device, or parameter information or parameter values used by the base station/network device or terminal device specified by the standard protocol, or parameter information or parameter values pre-stored in the base station/server or terminal device. This application does not limit this.
  • system and “network” in the embodiments of the present application can be used interchangeably.
  • “Multiple” refers to two or more.
  • “And/or” describes the association relationship of associated objects, indicating that three relationships may exist.
  • a and/or B can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural.
  • the character “/” generally indicates that the objects associated with each other are in an "or” relationship.
  • At least one of the following” or similar expressions refers to any combination of these items, including any combination of single items or plural items.
  • “at least one of A, B and C” includes A, B, C, AB, AC, BC or ABC.
  • the ordinal numbers such as “first” and “second” mentioned in the embodiments of the present application are used to distinguish multiple objects, and are not used to limit the order, timing, priority or importance of multiple objects.
  • Send and “receive” in the embodiments of the present application indicate the direction of signal transmission.
  • send information to XX can be understood as the destination of the information is XX, which can include direct sending through the air interface, and also include indirect sending through the air interface by other units or modules.
  • Receiveive information from YY can be understood as the source of the information is YY, which can include direct receiving from YY through the air interface, and also include indirect receiving from YY through the air interface from 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.
  • sending and receiving can be performed between devices, for example, between a network device and a terminal device, or can be performed within a device, for example, sending or receiving between components, modules, chips, software modules, or hardware modules within the device through a bus, wiring, or interface.
  • information may be processed between the source and destination of information transmission, such as coding, modulation, etc., but the destination can understand the valid information from the source. Similar expressions in this application can be understood similarly and will not be repeated.
  • indication may include direct indication and indirect indication, and may also include explicit indication and implicit indication.
  • the information indicated by a certain information is called information to be indicated.
  • information to be indicated In the specific implementation process, there are many ways to indicate the information to be indicated, such as but not limited to, directly indicating the information to be indicated, such as the information to be indicated itself or the index of the information to be indicated.
  • the information to be indicated may also be indirectly indicated by indicating other information, wherein the other information is associated with the information to be indicated; or only a part of the information to be indicated may be indicated, while the other part of the information to be indicated is known or agreed in advance.
  • the indication of specific information may be realized by means of the arrangement order of each information agreed in advance (such as predefined by the protocol), thereby reducing the indication overhead to a certain extent.
  • the present application does not limit the specific method of indication. It is understandable that, for the sender of the indication information, the indication information may be used to indicate the information to be indicated, and for the receiver of the indication information, the indication information may be used to determine the information to be indicated.
  • the present application can be applied to a long term evolution (LTE) system, a new radio (NR) system, or a communication system evolved after 5G (such as 6G, etc.), wherein the communication system includes at least one network device and/or at least one terminal device.
  • LTE long term evolution
  • NR new radio
  • 5G 5th Generation
  • 6G 6th Generation
  • FIG. 1a is a schematic diagram of a communication system in the present application.
  • FIG. 1a shows a network device and six terminal devices, which are terminal device 1, terminal device 2, terminal device 3, terminal device 4, terminal device 5, and terminal device 6.
  • terminal device 1 is a smart tea cup
  • terminal device 2 is a smart air conditioner
  • terminal device 3 is a smart gas station
  • terminal device 4 is a means of transportation
  • terminal device 5 is a mobile phone
  • terminal device 6 is a printer.
  • the AI configuration information sending entity may be a network device.
  • the AI configuration information receiving entity may be a terminal device 1-terminal device 6.
  • the network device and the terminal device 1-terminal device 6 form a communication system.
  • the terminal device 1-terminal device 6 may send data to the network device, and the network device needs to receive the data sent by the terminal device 1-terminal device 6.
  • the network device may send configuration information to the terminal device 1-terminal device 6.
  • terminal device 4-terminal device 6 can also form a communication system.
  • terminal device 5 serves as a network device, that is, an AI configuration information sending entity
  • terminal device 4 and terminal device 6 serve as terminal devices, that is, AI configuration information receiving entities.
  • terminal device 5 sends AI configuration information to terminal device 4 and terminal device 6 respectively, and receives data sent by terminal device 4 and terminal device 6; correspondingly, terminal device 4 and terminal device 6 receive AI configuration information sent by terminal device 5, and send data to terminal device 5.
  • different devices may also execute AI-related services.
  • the base station can perform communication-related services and AI-related services with one or more terminal devices, and communication-related services and AI-related services can also be performed between different terminal devices.
  • communication-related services and AI-related services can also be performed between the TV and the mobile phone.
  • an AI network element can be introduced into the communication system provided in the present application to implement some or all AI-related operations.
  • the AI network element may also be referred to as an AI node, an AI device, an AI entity, an AI module, an AI model, or an AI unit, etc.
  • the AI network element may be a network element built into a communication system.
  • the AI network element may be an AI module built into: an access network device, a core network device, a cloud server, or a network management (operation, administration and maintenance, OAM) to implement AI-related functions.
  • the OAM may be a core network device network management and/or Or as a network manager of an access network device.
  • the AI network element may also be a network element independently set up in a communication system.
  • the terminal or a chip built into the terminal may also include an AI entity for implementing AI-related functions.
  • AI artificial intelligence
  • AI Artificial intelligence
  • machines human intelligence for example, it can allow machines to use computer hardware and software to simulate certain intelligent behaviors of humans.
  • machine learning methods can be used.
  • machines use training data to learn (or train) a model.
  • the model represents the mapping from input to output.
  • the learned model can be used for reasoning (or prediction), that is, the model can be used to predict the output corresponding to a given input. Among them, the output can also be called the reasoning result (or prediction result).
  • Machine learning can include supervised learning, unsupervised learning, and reinforcement learning. Among them, unsupervised learning can also be called unsupervised learning.
  • supervised learning can use machine learning algorithms to learn the mapping relationship from sample values to sample labels based on the collected sample values and sample labels, and use AI models to express the learned mapping relationship.
  • the process of training a machine learning model is the process of learning this mapping relationship.
  • the sample value is input into the model to obtain the model's predicted value, and the model parameters are optimized by calculating the error between the model's predicted value and the sample label (ideal value).
  • the learned mapping can be used to predict new sample labels.
  • the mapping relationship learned by supervised learning can include linear mapping or nonlinear mapping. According to the type of label, the learning task can be divided into classification task and regression task.
  • the goal of supervised learning can be to learn the mapping relationship between the input data and the output data (i.e., the labeled data) in a given training set (containing multiple pairs of input data and labeled data), and at the same time, hope that the mapping relationship can also be applied to data outside the training set.
  • the training set is a collection of correct input and output pairs.
  • Neural network is a specific model in machine learning technology. According to the universal approximation theorem, neural network can theoretically approximate any continuous function, so that neural network has the ability to learn any mapping.
  • Traditional communication systems require rich expert knowledge to design communication modules, while deep learning communication systems based on neural networks can automatically discover implicit pattern structures from a large number of data sets, establish mapping relationships between data, and obtain performance that is superior to traditional modeling methods.
  • each neuron performs a weighted sum operation on its input values and outputs the operation result through an activation function.
  • FIG. 1d it is a schematic diagram of a neuron structure.
  • w i is used as the weight of xi to weight xi .
  • the bias for weighted summation of input values according to the weights is, for example, b.
  • the activation function can take many forms.
  • the output of the neuron is:
  • the output of the neuron is:
  • b can be a decimal, an integer (eg, 0, a positive integer or a negative integer), or a complex number, etc.
  • the activation functions of different neurons in a neural network can be the same or different.
  • a neural network generally includes multiple layers, each of which may include one or more neurons.
  • the expressive power of the neural network can be improved, providing a more powerful information extraction and abstract modeling capability for complex systems.
  • the depth of a neural network may refer to the number of layers included in the neural network, and the number of neurons included in each layer may be referred to as the width of the layer.
  • the neural network includes an input layer and an output layer. The input layer of the neural network processes the received input information through neurons, passes the processing results to the output layer, and the output layer obtains the output result of the neural network.
  • the neural network includes an input layer, a hidden layer, and an output layer.
  • the input layer of the neural network processes the received input information through neurons, passes the processing results to the middle hidden layer, the hidden layer calculates the received processing results, obtains the calculation results, and the hidden layer passes the calculation results to the output layer or the next adjacent hidden layer, and finally the output layer obtains the output result of the neural network.
  • a neural network may include one hidden layer, or include multiple hidden layers connected in sequence, without limitation.
  • the neural network is, for example, a deep neural network (DNN).
  • DNN can include a feedforward neural network (FNN), a convolutional neural network (FNN), or a CNN) and recurrent neural network (RNN).
  • FNN feedforward neural network
  • FNN convolutional neural network
  • RNN recurrent neural network
  • Figure 1e is a schematic diagram of a FNN network.
  • the characteristic of the FNN network is that the neurons in adjacent layers are fully connected to each other. This characteristic makes FNN usually require a large amount of storage space and leads to high computational complexity.
  • CNN is a neural network that is specifically designed to process data with a grid-like structure. For example, time series data (discrete sampling on the time axis) and image data (discrete sampling on two dimensions) can be considered to be data with a grid-like structure.
  • CNN does not use all the input information for calculations at once, but uses a fixed-size window to intercept part of the information for convolution operations, which greatly reduces the amount of calculation of model parameters.
  • each window can use different convolution kernel operations, which enables CNN to better extract the features of the input data.
  • RNN is a type of DNN network that uses feedback time series information. Its input includes the new input value at the current moment and its own output value at the previous moment. RNN is suitable for obtaining sequence features that are correlated in time, and is particularly suitable for applications such as speech recognition and channel coding.
  • a loss function can be defined.
  • the loss function describes the gap or difference between the output value of the model and the ideal target value.
  • the loss function can be expressed in many forms, and there is no restriction on the specific form of the loss function.
  • the model training process can be regarded as the following process: by adjusting some or all parameters of the model, the value of the loss function is less than the threshold value or meets the target requirements.
  • Models can also be referred to as AI models, rules or other names.
  • AI models can be considered as specific methods for implementing AI functions.
  • AI models characterize the mapping relationship or function between the input and output of a model.
  • AI functions may include one or more of the following: data collection, model training (or model learning), model information publishing, model inference (or model reasoning, inference, or prediction, etc.), model monitoring or model verification, or reasoning result publishing, etc.
  • AI functions can also be referred to as AI (related) operations, or AI-related functions.
  • Fully connected neural network also called multilayer perceptron (MLP).
  • an MLP consists of an input layer (left), an output layer (right), and multiple hidden layers (middle).
  • Each layer of the MLP contains several nodes, called neurons. The neurons in two adjacent layers are connected to each other.
  • w is the weight matrix
  • b is the bias vector
  • f is the activation function
  • a neural network can be understood as a mapping relationship from an input data set to an output data set.
  • neural networks are randomly initialized, and the process of obtaining this mapping relationship from random w and b using existing data is called neural network training.
  • the specific method of training is to use a loss function to evaluate the output results of the neural network.
  • the error can be back-propagated, and the neural network parameters (including w and b) can be iteratively optimized by the gradient descent method until the loss function reaches a minimum value, that is, the "better point (e.g., optimal point)" in FIG2b.
  • the neural network parameters corresponding to the "better point (e.g., optimal point)" in FIG2b can be used as the neural network parameters in the trained AI model information.
  • the gradient descent process can be expressed as:
  • ⁇ is the parameter to be optimized (including w and b)
  • L is the loss function
  • ⁇ is the learning rate, which controls the step size of gradient descent.
  • ⁇ is the learning rate, which controls the step size of gradient descent.
  • the back-propagation process utilizes the chain rule for partial derivatives.
  • the gradient of the previous layer parameters can be recursively calculated from the gradient of the next layer parameters, which can be expressed as:
  • w ij is the weight of node j connecting node i
  • si is the weighted sum of inputs on node i.
  • the FL architecture is the most widely used training architecture in the current FL field.
  • the FedAvg algorithm is the basic algorithm of FL. Its algorithm flow is as follows:
  • the center initializes the model to be trained And broadcast it to all client devices.
  • the central node aggregates and collects local training results from all (or some) clients. Assume that the client set that uploads the local model in round t is The center will use the number of samples of the corresponding client as the weight to perform weighted averaging to obtain a new global model. The specific update rule is: The center then sends the latest version of the global model Broadcast to all client devices for a new round of training.
  • the central node In addition to reporting local models You can also use the local gradient of training After reporting, the central node averages the local gradients and updates the global model according to the direction of the average gradient.
  • the data set exists in the distributed nodes, that is, the distributed nodes collect local data sets, perform local training, and report the local results (models or gradients) obtained from the training to the central node.
  • the central node itself does not have a data set, and is only responsible for fusing the training results of the distributed nodes to obtain the global model and send it to the distributed nodes.
  • Decentralized learning Different from federated learning, there is another distributed learning architecture - decentralized learning.
  • the design goal f(x) of a decentralized learning system is generally the mean of the goals fi (x) of each node, that is, Where n is the number of distributed nodes, x is the parameter to be optimized. In machine learning, x is the parameter of the machine learning (such as neural network) model.
  • Each node uses local data and local target fi (x) to calculate the local gradient Then it is sent to the neighboring nodes that can be communicated with. After any node receives the gradient information sent by its neighbor, it can update the parameter x of the local model according to the following formula:
  • ⁇ k represents represents the tuning coefficient
  • Ni is the set of neighbor nodes of node i
  • represents the number of elements in the set of neighbor nodes of node i, that is, the number of neighbor nodes of node i.
  • a wireless communication system e.g., the system shown in FIG. 1a or FIG. 1b
  • a communication node generally has signal transceiving capability and computing capability.
  • the computing capability of the network device is mainly to provide computing power support for the signal transceiving capability (e.g., sending and receiving signals) to realize the communication task between the network device and other communication nodes.
  • the computing power of communication nodes may have surplus computing power in addition to providing computing power support for the above communication tasks. Therefore, how to utilize this computing power is a technical problem that needs to be solved urgently.
  • the communication node can be used as a participating node in the AI learning system, and the computing power of the communication node can be applied to a certain link of the AI learning system.
  • deep learning models with massive parameters such as bidirectional encoder representations from transformers (BERT) and generative pre-trained transformers (GPT)
  • BERT bidirectional encoder representations from transformers
  • GPST generative pre-trained transformers
  • the reasoning process of the model will be limited by the capacity of the device, so generally large models are stored on cloud central servers.
  • each device in the network generates a huge amount of raw data every day, which requires multiple calls to the large model for reasoning.
  • the device (such as a communication node) can send data to the central server, the central server uses the data for reasoning, and then the central server returns the reasoning result to the device.
  • This process will consume a lot of communication resources for transmitting data, and the privacy of device data will also be at risk.
  • both communication nodes are taken as an example to participate in the AI learning system.
  • both Node 1 and Node 2 can be communication nodes, such as terminal devices or network devices.
  • the neural network used by the AI learning system can include at least a sub-neural network deployed at Node 1 for AI encoding, and/or, a sub-neural network deployed at Node 2 for AI decoding.
  • node 2 after node 1 processes the encoding result based on the sub-neural network for AI encoding, the encoding result is quantized and processed at the physical layer to obtain a wireless signal; accordingly, after node 2 receives the wireless signal through the transmission of the wireless channel, node 2 processes the signal at the physical layer and dequantizes the signal as the input of AI decoding, and the decoding result can be obtained after AI decoding processing. In addition, node 2 can also determine the gradient data based on the decoding result and the label data.
  • node 1 After node 2 obtains gradient data based on the sub-neural network processing of AI decoding, the gradient data is quantized and processed at the physical layer to obtain a wireless signal; correspondingly, after node 1 receives the wireless signal through transmission through the wireless channel, node 1 obtains gradient data after physical layer processing and dequantization processing. Subsequently, node 1 can optimize the neural network (such as training/updating/iteration, etc.) of the sub-neural network for AI encoding deployed in node 1 based on the gradient data.
  • the neural network such as training/updating/iteration, etc.
  • node 2 can also optimize the sub-neural network for AI encoding deployed in node 2 based on the gradient data (e.g., training/updating/iteration, etc.).
  • the node 2 can also calculate the result of the loss function based on the decoding result and the label data, and the result of the loss function can also be used for optimizing the neural network.
  • the above implementation is only explained by taking the node 2 determining the gradient data as an example.
  • the optimization process of the neural network may need to execute the above-mentioned AI encoding and AI decoding processes multiple times.
  • node 1 can perform AI encoding processing on multiple input data and then send it
  • node 2 can process the AI decoding results through multiple label data to obtain gradient data.
  • the multiple input data and the multiple label data can be data in the same training data set (refer to the implementation process of supervised learning in the previous article).
  • how to align the indexes of the input data used for the AI encoding and the label data used after AI decoding in the same training data set is a technical problem that needs to be solved urgently.
  • the present application provides a communication method and related equipment, which are used to enable the computing power of communication nodes to be applied to artificial intelligence (AI) processing of neural networks while also improving the flexibility of neural network deployment.
  • AI artificial intelligence
  • FIG3 is a schematic diagram of an implementation of the communication method provided in the present application.
  • the method includes the following steps.
  • the first communication device and the second communication device are used as the execution subjects of the interaction diagram as an example to illustrate the Method, but the present application does not limit the execution subject of the interaction diagram.
  • the execution subject of the method can be replaced by a chip, a chip system, a processor, a logic module or software in a communication device.
  • the first communication device can be a terminal device and the second communication device can be a network device, or the first communication device can be a network device and the second communication device can be a terminal device, or the first communication device and the second communication device are both terminal devices (for example, the method can be applied to the communication process of different terminal devices in a side link communication scenario).
  • a first communication device performs a first process on first data to obtain second data, wherein the first data is data in the first data set.
  • the first communication device sends second data, and correspondingly, the second communication device receives the second data.
  • the second communication device determines third data based on the second data.
  • the second data is obtained based on the first neural network processing the first data
  • the third data is obtained based on the second neural network processing the second data.
  • the first data set and the second data set can be included in one data set.
  • N input data and M label data can be included, and N and M are both positive integers; and the first data set can include the N input data, and the second data set can include the M label data.
  • the first data set can be called an input data set, a neural network input data set, etc.
  • the second data set can be called a label data set, a neural network label data set, etc.
  • the AI neural network can be processed based on the same data set to achieve iteration, update, etc. of the AI neural network.
  • the value of N is equal to the value of M.
  • the value of N may be greater than or equal to the value of M.
  • the value of N may be less than or equal to the value of M.
  • the first neural network can be called a neural network deployed at the transmitting end, an encoding neural network, an AI encoding neural network, etc.
  • the second neural network can be called a neural network deployed at the receiving end, a decoding neural network, an AI decoding neural network, etc.
  • AI neural network
  • AI neural network machine learning
  • AI processing AI neural network processing
  • the data involved (such as first data, second data, third data, and fourth data, etc.) can be replaced by information, signals, etc.
  • the index of the first data in the first data set is used to determine the fourth data in the second data set
  • the second data set includes the label data corresponding to the data in the first data set
  • the fourth data is the label data corresponding to the first data.
  • the method further includes: the second communication device sends the gradient information and/or loss function result determined based on the third data and the fourth data to the first communication device.
  • the receiver of the second data e.g., the second communication device
  • the receiver can process the second data to obtain the third data, and the receiver can also determine and send the corresponding gradient information and/or loss function result based on the third data and the fourth data.
  • the first communication device can update or iterate the first neural network based on the gradient information and/or loss function result.
  • the index of the first data in the first data set satisfies at least one of the following manner A and manner B.
  • Mode A The index is determined based on the resource carrying the second data.
  • the value of the index is determined by at least one of the time domain resource index of the resource, the frequency domain resource index of the resource, and the resource block size of the resource.
  • the index can be specifically determined by at least one of the above items to improve the flexibility of the solution implementation.
  • Method A can be understood as using the synchronized information between the data sender and receiver to generate the same index value on both sides, thereby achieving the index
  • the batch size of the training is N batch and the number of data set samples is N dataset
  • a round of training can be considered to use N batches of sample data each time until the data set samples are used up.
  • the following will take the resource carrying the second data as a time domain resource index as an example to provide some implementation examples.
  • the index value is generated using the system frame number nf in the time domain resource index of the data sender and receiver. For example, the index value of the data in the data set satisfies:
  • % represents the modulo operation
  • nf represents the system frame number
  • no is the offset value (used to traverse the data set).
  • the index value of the data in the data set may also be related to the system frame number nf and the subframe number nsf in the time domain resource index.
  • the index value of the data in the data set satisfies:
  • % represents the modulo operation
  • nf represents the system frame number
  • Nsf_f represents the number of subframes contained in each frame
  • nsf represents the subframe number
  • no is the offset value (used to traverse the data set).
  • the index value of the data in the data set may also be related to the system frame number nf and the time slot number nslot in the time domain resource index.
  • the index value of the data in the data set satisfies:
  • % represents the modulo operation
  • nf represents the system frame number
  • Nslot_f represents the number of time slots contained in each frame
  • nsf represents the subframe number
  • no is the offset value (used to traverse the data set).
  • the above implementation diagram only takes the resource carrying the second data as the time domain resource index as an example.
  • the time domain resource index can be replaced by other information related to the synchronized information between the sender and the receiver of the data, such as the number of physical resources in the frequency domain carrying the second data, including but not limited to the number of resource blocks, the number of subcarriers, etc.
  • method A can be understood as a real-time data alignment method, where real-time can be understood as a relatively fixed time interval between the process of the first communication device performing the first processing in step S301 and the process of the second communication device performing the second processing in step S303; and/or, the time interval between the process of the first communication device sending the processing result of the first processing (i.e., the second data) and the process of the second communication device sending the gradient data corresponding to the second processing (and/or the result of the loss function) is relatively fixed.
  • real-time can be understood as a relatively fixed time interval between the process of the first communication device performing the first processing in step S301 and the process of the second communication device performing the second processing in step S303; and/or, the time interval between the process of the first communication device sending the processing result of the first processing (i.e., the second data) and the process of the second communication device sending the gradient data corresponding to the second processing (and/or the result of the loss function) is relatively fixed.
  • FIG4 is an implementation example of mode A (i.e., real-time data alignment mode).
  • the first frame in every six frames e.g., frames with frame numbers 1/7/13
  • the fourth frame in every six frames e.g., frames with frame numbers 4/10/16
  • the time interval between the time domain resource carrying the second data and the time domain resource carrying the gradient data (and/or the result of the loss function) can be preconfigured.
  • the first communication device and the second communication device can also exchange other data, such as other communication signals shown in Figure 4, such as system information, reference signals, channel information obtained by measuring based on reference signals, etc.
  • Mode B The index is determined based on the number of times the data in the first data set is processed.
  • the method further includes: the first communication device sends first information, where the first information is used to indicate the index of the first data in the first data set.
  • the second communication device may determine the index based on the number of times data in the second data set is processed. Accordingly, the first communication device may also send the first information so that the second communication device can determine the index of the first data in the first data set based on the first information, and subsequently determine the fourth data in the second data set based on the index.
  • the first communication device can perform multiple processing based on the data in the first data set to obtain and send the processing results (for example, one of the processing results is the second data obtained based on the first data).
  • the first communication device can send corresponding indexes for some or all of the processing results (for example, send first information for the second data processing result). Thereafter, sending the corresponding index based on the part or all of the processing results can achieve alignment of the understanding of the index by the data sender and receiver, so as to avoid data processing errors caused by the misalignment of the understanding of the index by the data sender and receiver, thereby improving the robustness of the system.
  • the first communication device may send an index corresponding to a partial processing result for a partial processing result, without sending an index corresponding to the entire processing result for all processing results, which can save overhead.
  • the first information is one of a plurality of information transmitted based on a first cycle; the method also includes: the first communication device receives or sends configuration information, and the configuration information is used to configure the first cycle.
  • the first information may be one of the periodic information transmitted based on the first cycle.
  • the first communication device may receive or send configuration information for configuring the first cycle, so that the first communication device can serve as both a configurator and a configured party of the first cycle, which can align the understanding of the first cycle between the data sender and receiver, and can also improve the flexibility of the solution implementation.
  • the first communication device may set a sample index counter, which is used to accumulate the number of times the first processing is performed on the data in the first data set, and determine the index of the first data in the first data set based on the accumulated value in step S301; accordingly, the second communication device may set a sample index counter, which is used to accumulate the number of times the second processing is performed on the data in the second data set, and determine the index of the first data used to generate the second data in the first data set based on the accumulated value after receiving the second data in step S302 (or determine the index of the fourth data used after step S303 in the second data set).
  • the first communication device can set a timer of a first period. When the timer expires, the first communication device can also send the first information in step S302 when sending the second data, so that the first communication device and the second communication device can synchronize the indexes applicable to both through the first information to prevent the two from losing step due to the misalignment of sample index counters.
  • the configuration information may be carried in an RRC message.
  • the configuration information may implement the configuration of the first period through a data synchronization period (DataSyncPeriod) element in the RRC message.
  • DataSyncPeriod data synchronization period
  • mode B can be understood as a data synchronization mode in a non-real-time system.
  • the non-real-time here can be understood as the time interval between the process of the first communication device performing the first processing in step S301 and the process of the second communication device performing the second processing in step S303 is not relatively fixed, and/or, the time interval between the process of the first communication device sending the processing result (i.e., the second data) of the first processing and the process of the second communication device sending the gradient data (and/or the result of the loss function) corresponding to the second processing is not relatively fixed.
  • FIG5 is an implementation example of method B (i.e., non-real-time data alignment method).
  • multiple AI tasks can be executed between the first communication device and the second communication device, and the execution cycles of different AI tasks or the triggering of data transmission and reception of different AI tasks may be different.
  • the scale of input data of different AI tasks may be different.
  • the scale of gradient data (and/or loss function results) of different AI tasks may be different.
  • the data involved in one AI task may include the second data transmitted with frame number 1 and the gradient data (and/or the result of the loss function) transmitted with frame number 4, that is, the interval between the two is 2 frames (that is, frames with frame numbers 2 and 3); the data involved in another AI task may include the second data transmitted with frame number 5 and the gradient data (and/or the result of the loss function) transmitted with frame number 10, that is, the interval between the two is 4 frames (that is, frames with frame numbers 6, 7, 8 and 9); the data involved in another AI task may include the second data transmitted with frame number 17 and the gradient data (and/or the result of the loss function) transmitted with frame number 18, that is, the interval between the two is 0 frame (that is, the two are two adjacent frames).
  • the index is determined based on the resource carrying the second data and the number of times the data in the first data set is processed.
  • the index of the first data in the first data set satisfies method A and method B
  • the index can be determined based on the resource carrying the second data (for example, at least one of the time domain resource index, frequency domain resource index, and resource block size of the resource) and the number of times the data in the first data set is processed.
  • the value of the index may be a mathematical operation result of the value of the time domain resource index of the resource and the value of the number of processing times (e.g., the sum of the two values, the difference of the two values, the product of the two values, etc.).
  • the value of the index may be a mathematical operation result of the value of the resource block size of the resource and the value of the number of processing times (e.g., the sum of the two values, the difference of the two values, the product of the two values, etc.).
  • the method shown in FIG3 further includes: the first communication device receives or sends indication information indicating that the index satisfies the at least one item (that is, the indication information is used to indicate mode A and/or mode B, or the indication information is used to indicate mode A or mode B or mode C).
  • the first communication device may also receive or send indication information indicating that the index satisfies the at least one item, so that the data sender and receiver can align their understanding of the index of the data in the data set based on the indication information, so as to avoid data processing errors caused by the misalignment of the understanding of the index between the data sender and receiver, thereby improving the robustness of the system.
  • the method shown in FIG3 further includes: the first communication device receiving an indication signal indicating the first data set; Specifically, the first communication device can be used as a configured party of the first data set, and/or the first communication device can be used as a configured party of the second data set, so that the data sender and receiver can obtain the data set before exchanging data.
  • the first data set may be preconfigured for the first communication device, and/or the second data set may be preconfigured for the second communication device. In this way, overhead can be reduced.
  • the second data sent by the first communication device in step S301 is obtained based on the first data, and the second communication device can subsequently process the second data in step S302 to obtain the third data.
  • the index of the first data in the first data set is used to determine the fourth data in the second data set, and the fourth data is the label data corresponding to the first data.
  • the second communication device can determine the label data in the second data set and process the third data based on the label data.
  • the second data and/or the third data are obtained based on a neural network, that is, the neural network used for AI processing can include a neural network deployed in the first communication device and/or a neural network deployed in the second communication device.
  • the communication device in the communication system acts as an AI participating node
  • the computing power of the communication device can be applied to the AI processing of the neural network, while also improving the flexibility of the neural network deployment.
  • the index of the first data in the first data set is used to determine the label data (i.e., the fourth data) corresponding to the first data in the second data set, and the index satisfies at least one of the above items.
  • the second communication device can determine the label data in the second data set based on the resource carrying the data or the number of times the data in the data set is processed. In this way, the air interface overhead can be reduced to improve communication efficiency.
  • FIG6 is a schematic diagram of an implementation of the communication method provided in the present application. The method includes the following steps.
  • the first communication device performs a first process on first data to obtain second data, wherein the first data is data in the first data set.
  • the second communication device determines third data based on the second data.
  • the method shown in FIG6 further includes: the first communication device receives the gradient information and/or the result of the loss function determined based on the third data and the fourth data.
  • the receiver of the second data e.g., the second communication device
  • the receiver can process the second data to obtain the third data, and the receiver can also determine and send the corresponding gradient information and/or the result of the loss function based on the third data and the fourth data.
  • the first communication device can update or iterate the first neural network based on the gradient information and/or the result of the loss function.
  • the implementation process of the first communication device and the second communication device (eg, the first data to the fourth data, the first processing and the second data, etc.) can refer to the previous FIG. 3 and related implementation methods.
  • the second data sent by the first communication device in step S602 is obtained based on the first data, and the second communication device can subsequently process the second data in step S603 to obtain the third data.
  • the first communication device can also send fourth data in step S602, and the fourth data is the label data corresponding to the first data.
  • the second communication device receives the fourth data in step S602, it can process the third data based on the fourth data (i.e., the label data corresponding to the first data).
  • the second data and/or the third data are obtained based on a neural network, that is, the neural network used for AI processing may include a neural network deployed in the first communication device and/or a neural network deployed in the second communication device.
  • the communication device in the communication system acts as an AI participating node
  • the computing power of the communication device can be applied to the AI processing of the neural network, while also improving the flexibility of the neural network deployment.
  • the second data sent by the first communication device is the processing result of the first data
  • the fourth data sent by the first communication device is the label data corresponding to the first data.
  • the first communication device can obtain the first data set by configuration or pre-configuration
  • the second communication device can obtain the second data set by configuration or pre-configuration
  • both the first communication device and the second communication device need to align the indexes by the resource carrying the second data or the number of times the data in the data set is processed.
  • the difference is that the first data set and the second data set can be deployed in the first communication device, and the second communication device There is no need to deploy a data set, and the index of the data set is only maintained on the first communication device side.
  • the advantage is that there is no need to maintain synchronized index values on both sides, which can reduce the overhead and implementation complexity of the second communication device.
  • the technical solution shown in FIG. 6 since the fourth data transmitted in step S602 is label data, the technical solution shown in FIG. 6 may be applicable to a scenario where the data scale of the label data is relatively small.
  • FIG. 7 is a schematic diagram of an implementation of the communication method provided in the present application. The method includes the following steps.
  • the first communication device performs a first process on first data to obtain second data, wherein the first data is data in the first data set.
  • the first communication device sends the second data and the first index, and correspondingly, the second communication device receives the second data and the first index.
  • the second communication device determines third data based on the second data.
  • the second data is obtained based on the first neural network processing the first data
  • the third data is obtained based on the second neural network processing the second data.
  • the first data set and the second data set can be included in one data set.
  • N input data and M label data can be included, and N and M are both positive integers; and the first data set can include the N input data, and the second data set can include the M label data.
  • the first data set can be called an input data set, a neural network input data set, etc.
  • the second data set can be called a label data set, a neural network label data set, etc.
  • the AI neural network can be processed based on the same data set to achieve iteration, update, etc. of the AI neural network.
  • the first index is determined by a second index of the first data in the first data set.
  • the value of N is equal to the value of M.
  • the first index may be the same as the second index of the first data in the first data set.
  • the label data of the i-th input data in the first data set is the j-th data in the second data set, i is the first index, j is the second index, and i is equal to j. That is, the label data of the first data of N input data is the first data in M label data, the label data of the second data of N input data is the second data in M label data, and so on, the label data of the N-th data of N input data is the M-th data in M label data (N is equal to M).
  • the first index may be partially or completely different from the second index of the first data in the first data set.
  • the label data of the i-th input data in the first data set is the j-th data in the second data set
  • i is the first index
  • j is the second index
  • i and j may be partially or completely unequal.
  • the mapping relationship between i and j may be preconfigured.
  • the mapping relationship between i and j can be preconfigured, i can be traversed from 1 to N, and j can be traversed from N (N equals M) to 1.
  • N N equals M
  • the label data of the first copy of N input data is the third copy of M label data
  • the label data of the second copy of N input data is the second copy of M label data
  • the label data of the third copy of N input data is the first copy of M label data.
  • the first index can be different from the second index part of the first data in the first data set.
  • mapping relationship between i and j can be preconfigured
  • the mapping relationship between i and j can be configured or preconfigured to align the data sender and receiver.
  • the label data of the first copy of N input data is the third copy of the M label data
  • the label data of the second copy of N input data is the first copy of the M label data
  • the label data of the third copy of N input data is the second copy of the M label data.
  • the first index can be completely different from the second index of the first data in the first data set.
  • mapping relationship between the first index and the second index of the first data in the first data set is preconfigured.
  • the value of N may be greater than or equal to the value of M.
  • the first index may be less than or equal to the second index of the first data in the first data set.
  • the value of N may be less than or equal to the value of M.
  • the first index may be greater than or equal to the second index of the first data in the first data set.
  • the method shown in FIG7 further includes: the first communication device receiving a communication signal based on the third data and the fourth data
  • the first communication device can update or iterate the first neural network based on the gradient information and/or the result of the loss function after receiving the gradient information and/or the result of the loss function.
  • the receiver of the second data e.g., the second communication device
  • the receiver can process the second data to obtain the third data, and the receiver can also determine and send the corresponding gradient information and/or the result of the loss function based on the third data and the fourth data.
  • the second data sent by the first communication device in step S702 is obtained based on the first data, and the second communication device can subsequently process the second data to obtain the third data in step S703.
  • the first communication device can also send a first index.
  • the recipient of the second data and the fourth data can process the third data based on the label data.
  • the second data and/or the third data are obtained based on a neural network, that is, the neural network used for AI processing may include a neural network deployed in the first communication device and/or a neural network deployed in the second communication device.
  • the second data sent by the first communication device is the processing result of the first data
  • the fourth data sent by the first communication device is the label data corresponding to the first data.
  • the first communication device can obtain the first data set by configuration or pre-configuration
  • the second communication device can obtain the second data set by configuration or pre-configuration
  • both the first communication device and the second communication device need to achieve index alignment through the number of times the data in the resource carrying the second data or the data set is processed.
  • the difference is that the first communication device and the second communication device do not need to set a sample index counter for synchronization, which can reduce the implementation complexity.
  • the technical solution shown in FIG. 7 since the data transmitted in step S702 includes the second index, the technical solution shown in FIG. 7 may be applicable to a scenario where the data set is small (or the number of indexes in the data set is small).
  • the embodiment of the present application provides a communication device 800, which can implement the functions of the second communication device or the first communication device in the above method embodiment, and thus can also achieve the beneficial effects of the above method embodiment.
  • the communication device 800 can be the first communication device (or the second communication device), or it can be an integrated circuit or component inside the first communication device (or the second communication device), such as a chip.
  • the transceiver unit 802 may include a sending unit and a receiving unit, which are respectively used to perform sending and receiving.
  • the device 800 when the device 800 is used to execute the method executed by the first communication device in the aforementioned embodiment, the device 800 includes a processing unit 801 and a transceiver unit 802; the processing unit 801 is used to process the first data to obtain the second data; wherein the first data is the data in the first data set; the transceiver unit 802 is used to send the second data, and the second data is used to determine the third data; wherein the index of the first data in the first data set is used to determine the fourth data in the second data set, and the second data set includes label data corresponding to the data in the first data set, and the fourth data is the label data corresponding to the first data; wherein the second data is obtained based on the first neural network processing the first data, and/or the third data is obtained based on the second neural network processing the second data; the index satisfies at least one of the following: the index is determined based on the resources carrying the second data; the index is determined based on the number of times the data in the first data set
  • the device 800 when the device 800 is used to execute the method executed by the second communication device in the aforementioned embodiment, the device 800 includes a processing unit 801 and a transceiver unit 802; the transceiver unit 802 is used to receive second data, which is obtained by processing the first data, and the first data is the data in the first data set; wherein the index of the first data in the first data set is used to determine fourth data in the second data set, and the second data set includes label data corresponding to the data in the first data set, and the fourth data is the label data corresponding to the first data; the processing unit 801 is used to determine third data based on the second data; wherein the second data is obtained by processing the first data based on the first neural network, and/or the third data is obtained by processing the second data based on the second neural network; the index satisfies at least one of the following: the index is determined based on the resources carrying the second data; the index is determined based on the number of times the data in the second data set is processed.
  • the device 800 when the device 800 is used to execute the method executed by the first communication device in the above embodiment, the device 800 includes a processing unit 801 and a transceiver unit 802; the processing unit 801 is used to process the first data to obtain the second data; The transceiver unit 802 is used to send the second data and the fourth data, wherein the second data is used to determine the third data, and the fourth data is the label data corresponding to the first data; wherein the second data is obtained by processing the first data based on the first neural network, and/or the third data is obtained by processing the second data based on the second neural network.
  • the device 800 when the device 800 is used to execute the method executed by the second communication device in the aforementioned embodiment, the device 800 includes a processing unit 801 and a transceiver unit 802; the transceiver unit 802 is used to receive second data and fourth data, wherein the second data is obtained based on the first data, and the fourth data is label data corresponding to the first data; the processing unit 801 is used to determine third data based on the second data; wherein the second data is obtained based on the first neural network processing the first data, and/or the third data is obtained based on the second neural network processing the second data.
  • the device 800 when the device 800 is used to execute the method executed by the first communication device in the aforementioned embodiment, the device 800 includes a processing unit 801 and a transceiver unit 802; the processing unit 801 is used to process the first data to obtain the second data; wherein the first data is the data in the first data set; the transceiver unit 802 is used to send the second data and a first index, and the second data is used to determine the third data; wherein the first index is used to determine the fourth data in the second data set, the second data set includes label data corresponding to the data in the first data set, and the fourth data is the label data corresponding to the first data; wherein the second data is obtained based on the first neural network processing the first data, and/or the fourth data is obtained based on the neural network processing the second data.
  • the device 800 when the device 800 is used to execute the method executed by the second communication device in the aforementioned embodiment, the device 800 includes a processing unit 801 and a transceiver unit 802; the transceiver unit 802 is used to receive second data and a first index, the second data is obtained based on the first data, and the first data is the data in the first data set; wherein the first index is used to determine fourth data in the second data set, the second data set includes label data corresponding to the data in the first data set, and the fourth data is the label data corresponding to the first data; the processing unit 801 is used to determine third data based on the second data; wherein the second data is obtained based on the first neural network processing the first data, and/or the third data is obtained based on the second neural network processing the second data.
  • Fig. 9 is another schematic structural diagram of a communication device 900 provided in the present application.
  • the communication device 900 includes a logic circuit 901 and an input/output interface 902.
  • the communication device 900 may be a chip or an integrated circuit.
  • the transceiver unit 802 shown in Fig. 8 may be a communication interface, which may be the input/output interface 902 in Fig. 9, which may include an input interface and an output interface.
  • the communication interface may be a transceiver circuit, which may include an input interface circuit and an output interface circuit.
  • the logic circuit 901 is used to process the first data to obtain the second data; wherein the first data is the data in the first data set; the input-output interface 902 is used to send the second data, and the second data is used to determine the third data; wherein the index of the first data in the first data set is used to determine the fourth data in the second data set, and the second data set includes label data corresponding to the data in the first data set, and the fourth data is the label data corresponding to the first data; wherein the second data is obtained based on the first neural network processing the first data, and/or the third data is obtained based on the second neural network processing the second data; the index satisfies at least one of the following: the index is determined based on the resources carrying the second data; the index is determined based on the number of times the data in the first data set is processed.
  • the input-output interface 902 is used to receive second data, which is obtained by processing first data, and the first data is data in the first data set; wherein the index of the first data in the first data set is used to determine fourth data in the second data set, and the second data set includes label data corresponding to the data in the first data set, and the fourth data is label data corresponding to the first data; the logic circuit 901 is used to determine third data based on the second data; wherein the second data is obtained by processing the first data based on the first neural network, and/or the third data is obtained by processing the second data based on the second neural network; the index satisfies at least one of the following: the index is determined based on the resources carrying the second data; the index is determined based on the number of times the data in the second data set is processed.
  • the logic circuit 901 is used to process the first data to obtain the second data; the input-output interface 902 is used to send the second data and fourth data, wherein the second data is used to determine the third data, and the fourth data is label data corresponding to the first data; wherein the second data is obtained by processing the first data based on the first neural network, and/or the third data is obtained by processing the second data based on the second neural network.
  • the input-output interface 902 is used to receive second data and fourth data, wherein the second data is obtained based on the first data, and the fourth data is label data corresponding to the first data; the logic circuit 901 is used to determine third data based on the second data; wherein the second data is obtained based on the first neural network processing the first data, and/or the third data is obtained based on the second neural network processing the second data.
  • the logic circuit 901 is used to process the first data to obtain the second data; wherein the first data is the data in the first data set; the input-output interface 902 is used to send the second data and the first index, and the second data is used to determine the third data; wherein the first index is used to determine the fourth data in the second data set, and the second data set includes label data corresponding to the data in the first data set, and the fourth data is the label data corresponding to the first data; wherein the second data is obtained by processing the first data based on the first neural network, and/or the fourth data is obtained by processing the second data based on the neural network.
  • the input-output interface 902 is used to receive second data and a first index, the second data is obtained based on the first data, and the first data is the data in the first data set; wherein the first index is used to determine fourth data in the second data set, the second data set includes label data corresponding to the data in the first data set, and the fourth data is the label data corresponding to the first data; the logic circuit 901 is used to determine third data based on the second data; wherein the second data is obtained based on the first neural network processing the first data, and/or the third data is obtained based on the second neural network processing the second data.
  • the logic circuit 901 and the input/output interface 902 may also execute other steps executed by the first communication device or the second communication device in any embodiment and achieve corresponding beneficial effects, which will not be described in detail here.
  • the processing unit 801 shown in FIG. 8 may be the logic circuit 901 in FIG. 9 .
  • the logic circuit 901 may be a processing device, and the functions of the processing device may be partially or completely implemented by software.
  • the functions of the processing device may be partially or completely implemented by software.
  • the processing device may include a memory and a processor, wherein the memory is used to store a computer program, and the processor reads and executes the computer program stored in the memory to perform corresponding processing and/or steps in any one of the method embodiments.
  • the processing device may include only a processor.
  • a memory for storing a computer program is located outside the processing device, and the processor is connected to the memory via a circuit/wire to read and execute the computer program stored in the memory.
  • the memory and the processor may be integrated together, or may be physically independent of each other.
  • the processing device may be one or more chips, or one or more integrated circuits.
  • the processing device may be one or more field-programmable gate arrays (FPGA), application specific integrated circuits (ASIC), system on chip (SoC), central processor unit (CPU), network processor (NP), digital signal processor (DSP), microcontroller unit (MCU), programmable logic device (PLD) or other integrated chips, or any combination of the above chips or processors.
  • FPGA field-programmable gate arrays
  • ASIC application specific integrated circuits
  • SoC system on chip
  • CPU central processor unit
  • NP network processor
  • DSP digital signal processor
  • MCU microcontroller unit
  • PLD programmable logic device
  • FIG 10 shows a communication device 1000 involved in the above embodiments provided in an embodiment of the present application.
  • the communication device 1000 can specifically be a communication device as a terminal device in the above embodiments.
  • the example shown in Figure 10 is that the terminal device is implemented through the terminal device (or a component in the terminal device).
  • the communication device 1000 may include but is not limited to at least one processor 1001 and a communication port 1002 .
  • the transceiver unit 802 shown in Fig. 8 may be a communication interface, which may be the communication port 1002 in Fig. 10, which may include an input interface and an output interface.
  • the communication port 1002 may also be a transceiver circuit, which may include an input interface circuit and an output interface circuit.
  • the device may also include at least one of the memory 1003 and the bus 1004.
  • the at least one processor 1001 is used to control and process the actions of the communication device 1000.
  • the processor 1001 can be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a transistor logic device, a hardware component or any combination thereof. It can implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application.
  • the processor can also be a combination that implements a computing function, such as a combination of one or more microprocessors, a combination of a digital signal processor and a microprocessor, and the like.
  • the communication device 1000 shown in Figure 10 can be specifically used to implement the steps implemented by the terminal device in the aforementioned method embodiment, and to achieve the corresponding technical effects of the terminal device.
  • the specific implementation methods of the communication device shown in Figure 10 can refer to the description in the aforementioned method embodiment, and will not be repeated here one by one.
  • FIG 11 is a structural diagram of the communication device 1100 involved in the above-mentioned embodiments provided in an embodiment of the present application.
  • the communication device 1100 can specifically be a communication device as a network device in the above-mentioned embodiments.
  • the example shown in Figure 11 is that the network device is implemented through the network device (or a component in the network device), wherein the structure of the communication device can refer to the structure shown in Figure 11.
  • the communication device 1100 includes at least one processor 1111 and at least one network interface 1114. Further optionally, the communication device also includes at least one memory 1112, at least one transceiver 1113 and one or more antennas 1115.
  • the processor 1111, the memory 1112, the transceiver 1113 and the network interface 1114 are connected, for example, through a bus. In an embodiment of the present application, the connection may include various interfaces, transmission lines or buses, etc., which are not limited in this embodiment.
  • the antenna 1115 is connected to the transceiver 1113.
  • the network interface 1114 is used to enable the communication device to communicate with other communication devices through a communication link.
  • the network interface 1114 may include a network interface between the communication device and the core network device, such as an S1 interface, and the network interface may include a network interface between the communication device and other communication devices (such as other network devices or core network devices), such as an X2 or Xn interface.
  • the transceiver unit 802 shown in Fig. 8 may be a communication interface, which may be the network interface 1114 in Fig. 11, and the network interface 1114 may include an input interface and an output interface.
  • the network interface 1114 may also be a transceiver circuit, and the transceiver circuit may include an input interface circuit and an output interface circuit.
  • the processor 1111 is mainly used to process the communication protocol and communication data, and to control the entire communication device, execute the software program, and process the data of the software program, for example, to support the communication device to perform the actions described in the embodiment.
  • the communication device may include a baseband processor and a central processor, the baseband processor is mainly used to process the communication protocol and communication data, and the central processor is mainly used to control the entire terminal device, execute the software program, and process the data of the software program.
  • the processor 1111 in Figure 11 can integrate the functions of the baseband processor and the central processor. It can be understood by those skilled in the art that the baseband processor and the central processor can also be independent processors, interconnected by technologies such as buses.
  • the terminal device can include multiple baseband processors to adapt to different network formats, the terminal device can include multiple central processors to enhance its processing capabilities, and the various components of the terminal device can be connected through various buses.
  • the baseband processor can also be described as a baseband processing circuit or a baseband processing chip.
  • the central processor can also be described as a central processing circuit or a central processing chip.
  • the function of processing the communication protocol and communication data can be built into the processor, or it can be stored in the memory in the form of a software program, and the processor executes the software program to realize the baseband processing function.
  • the memory is mainly used to store software programs and data.
  • the memory 1112 can exist independently and be connected to the processor 1111.
  • the memory 1112 can be integrated with the processor 1111, for example, integrated into a chip.
  • the memory 1112 can store program codes for executing the technical solutions of the embodiments of the present application, and the execution is controlled by the processor 1111.
  • the various types of computer program codes executed can also be regarded as drivers of the processor 1111.
  • FIG11 shows only one memory and one processor.
  • the memory may also be referred to as a storage medium or a storage device, etc.
  • the memory may be a storage element on the same chip as the processor, i.e., an on-chip storage element, or an independent storage element, which is not limited in the embodiments of the present application.
  • the transceiver 1113 can be used to support the reception or transmission of radio frequency signals between the communication device and the terminal, and the transceiver 1113 can be connected to the antenna 1115.
  • the transceiver 1113 includes a transmitter Tx and a receiver Rx.
  • one or more antennas 1115 can receive radio frequency signals
  • the receiver Rx of the transceiver 1113 is used to receive the radio frequency signal from the antenna, and convert the radio frequency signal into a digital baseband signal or a digital intermediate frequency signal, and provide the digital baseband signal or the digital intermediate frequency signal to the processor 1111, so that the processor 1111 further processes the digital baseband signal or the digital intermediate frequency signal, such as demodulation and decoding.
  • the transmitter Tx in the transceiver 1113 is also used to receive a modulated digital baseband signal or a digital intermediate frequency signal from the processor 1111, and convert the modulated digital baseband signal or the digital intermediate frequency signal into a radio frequency signal, and send the radio frequency signal through one or more antennas 1115.
  • the receiver Rx can selectively perform one or more stages of down-mixing and analog-to-digital conversion processing on the RF signal to obtain a digital baseband signal or a digital intermediate frequency signal, and the order of the down-mixing and analog-to-digital conversion processing is adjustable.
  • the transmitter Tx can selectively perform one or more stages of up-mixing and digital-to-analog conversion processing on the modulated digital baseband signal or digital intermediate frequency signal to obtain a RF signal, and the order of the up-mixing and digital-to-analog conversion processing is adjustable.
  • the digital baseband signal and the digital intermediate frequency signal can be collectively referred to as a digital signal.
  • the transceiver 1113 may also be referred to as a transceiver unit, a transceiver, a transceiver device, etc.
  • the device used to implement the receiving function in the transceiver unit may be regarded as a receiving unit
  • the device used to implement the sending function in the transceiver unit may be regarded as a sending unit, that is, the transceiver unit includes a receiving unit, a receiving unit, and a transmitting unit.
  • the receiving unit and the sending unit, the receiving unit can also be called a receiver, an input port, a receiving circuit, etc.
  • the sending unit can be called a transmitter, a transmitter or a transmitting circuit, etc.
  • the communication device 1100 shown in Figure 11 can be specifically used to implement the steps implemented by the network device in the aforementioned method embodiment, and to achieve the corresponding technical effects of the network device.
  • the specific implementation methods of the communication device 1100 shown in Figure 11 can refer to the description in the aforementioned method embodiment, and will not be repeated here.
  • FIG. 12 is a schematic diagram of the structure of the communication device involved in the above-mentioned embodiment provided in an embodiment of the present application.
  • the communication device 120 includes, for example, modules, units, elements, circuits, or interfaces, etc., which are appropriately configured together to perform the technical solutions provided in this application.
  • the communication device 120 may be the terminal device or network device described above, or a component (such as a chip) in these devices, to implement the method described in the following method embodiment.
  • the communication device 120 includes one or more processors 121.
  • the processor 121 may be a general-purpose processor or a dedicated processor, etc.
  • it may be a baseband processor or a central processing unit.
  • the baseband processor may be used to process communication protocols and communication data
  • the central processing unit may be used to control the communication device (such as a RAN node, a terminal, or a chip, etc.), execute software programs, and process data of software programs.
  • the processor 121 may include a program 123 (sometimes also referred to as code or instruction), and the program 123 may be executed on the processor 121 so that the communication device 120 performs the method described in the following embodiments.
  • the communication device 120 includes a circuit (not shown in FIG. 12 ).
  • the communication device 120 may include one or more memories 122 on which a program 124 (sometimes also referred to as code or instructions) is stored.
  • the program 124 can be run on the processor 121 so that the communication device 120 executes the method described in the above method embodiment.
  • the processor 121 and/or the memory 122 may include an AI module 127, 128, and the AI module is used to implement AI-related functions.
  • the AI module may be implemented by software, hardware, or a combination of software and hardware.
  • the AI module may include a wireless intelligent control (radio intelligence control, RIC) module.
  • the AI module may be a near real-time RIC or a non-real-time RIC.
  • data may also be stored in the processor 121 and/or the memory 122.
  • the processor and the memory may be provided separately or integrated together.
  • the communication device 120 may further include a transceiver 125 and/or an antenna 126.
  • the processor 121 may also be sometimes referred to as a processing unit, which controls the communication device (e.g., a RAN node or a terminal).
  • the transceiver 125 may also be sometimes referred to as a transceiver unit, a transceiver, a transceiver circuit, or a transceiver, etc., which is used to implement the transceiver function of the communication device through the antenna 126.
  • the transceiver unit 802 shown in Fig. 8 may be a communication interface, which may be the transceiver 125 in Fig. 12, and the transceiver 125 may include an input interface and an output interface.
  • the transceiver 125 may also be a transceiver circuit, which may include an input interface circuit and an output interface circuit.
  • An embodiment of the present application further provides a computer-readable storage medium, which is used to store one or more computer-executable instructions.
  • the processor executes the method described in the possible implementation methods of the first communication device or the second communication device in the aforementioned embodiment.
  • An embodiment of the present application also provides a computer program product (or computer program).
  • the processor executes the method that may be implemented by the above-mentioned first communication device or second communication device.
  • An embodiment of the present application also provides a chip system, which includes at least one processor for supporting a communication device to implement the functions involved in the possible implementation methods of the above-mentioned communication device.
  • the chip system also includes an interface circuit, which provides program instructions and/or data for the at least one processor.
  • the chip system may also include a memory, which is used to store the necessary program instructions and data for the communication device.
  • the chip system can be composed of chips, and may also include chips and other discrete devices, wherein the communication device can specifically be the first communication device or the second communication device in the aforementioned method embodiment.
  • An embodiment of the present application also provides a communication system, and the network system architecture includes the first communication device and the second communication device in any of the above embodiments.
  • the disclosed systems, devices and methods can be implemented in other ways.
  • the device embodiments described above are only schematic.
  • the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.
  • Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
  • each functional unit in each embodiment of the present application can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
  • the above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit. If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.
  • the technical solution of the present application is essentially or the part that contributes or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application.
  • the aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code.

Landscapes

  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

一种通信方法及相关设备，用于使得通信节点的算力能够应用于神经网络的人工智能(artificial intelligence，AI)处理的同时，也能够提升神经网络部署的灵活性。在该方法中，第一通信装置发送的第二数据是基于第一数据得到的，后续该第二数据的接收方(例如第二通信装置)可以对第二数据进行处理得到第三数据。其中，该第一数据在该第一数据集中的索引用于在第二数据集中确定第四数据，该第四数据是该第一数据对应的标签数据。并且，第二数据和/或该第三数据是基于神经网络得到的，即用于AI处理的神经网络可以包括部署于第一通信装置的神经网络和/或部署于第二通信装置的神经网络。

Description

一种通信方法及相关设备
本申请要求于2023年11月03日提交国家知识产权局、申请号为202311461212.5、申请名称为“一种通信方法及相关设备”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及通信领域，尤其涉及一种通信方法及相关设备。
背景技术
无线通信，可以是两个或两个以上的通信节点间不经由导体或缆线传播而进行的传输通讯，该通信节点一般包括网络设备和终端设备。
目前，在无线通信系统中，通信节点一般具备信号收发能力和计算能力。以具备计算能力的网络设备为例，网络设备的计算能力主要是为信号收发能力提供算力支持(例如：对信号进行发送处理和接收处理)，以实现网络设备与其它通信节点的通信。
然而，在通信网络中，通信节点的计算能力除了为上述通信任务提供算力支持之外，还可能具备富余的计算能力。为此，如何利用这些计算能力，是一个亟待解决的技术问题。
发明内容
本申请提供了一种通信方法及相关设备，用于使得通信节点的算力能够应用于神经网络的人工智能(artificial intelligence，AI)处理的同时，也能够提升神经网络部署的灵活性。
本申请第一方面提供了一种通信方法，该方法由第一通信装置执行，该第一通信装置可以是通信设备(如，终端设备或网络设备)，或者，该第一通信装置可以是通信设备中的部分组件(例如处理器、芯片或芯片系统等)，或者该第一通信装置还可以是能实现全部或部分通信设备功能的逻辑模块或软件。在第一方面及其可能的实现方式中，以该通信方法由第一通信装置执行为例进行描述。示例性的，该第一通信装置可以为终端设备或网络设备。在该方法中，第一通信装置对第一数据进行处理，得到第二数据；其中，该第一数据为第一数据集中的数据；该第一通信装置发送该第二数据，该第二数据用于确定第三数据；其中，该第一数据在该第一数据集中的索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的；该索引满足以下至少一项：该索引是基于承载该第二数据的资源确定的；该索引是基于该第一数据集中的数据的处理次数确定的。
基于上述技术方案，第一通信装置发送的第二数据是基于第一数据得到的，后续该第二数据的接收方(例如第二通信装置)可以对第二数据进行处理得到第三数据。其中，该第一数据在该第一数据集中的索引用于在第二数据集中确定第四数据，该第四数据是该第一数据对应的标签数据。换言之，第二数据的接收方在接收第二数据之后，能够在第二数据集中确定标签数据，并基于该标签数据对第三数据进行处理。并且，第二数据和/或该第三数据是基于神经网络得到的，即用于AI处理的神经网络可以包括部署于第一通信装置的神经网络和/或部署于第二通信装置的神经网络。从而，在通信系统中的通信装置作为AI参与节点的情况下，使得通信装置的算力能够应用于神经网络的AI处理的同时，也能够提升神经网络部署的灵活性。
此外，第一数据在该第一数据集中的索引用于在第二数据集中确定该第一数据对应的标签数据(即第四数据)，并且，该索引满足上述至少一项。换言之，第二通信装置在接收第二数据之后，能够基于承载数据的资源或数据集中的数据的处理次数在第二数据集中确定标签数据，通过这种方式，可以降低空口开销，以提升通信效率。
应理解，在上述技术方案中，第二数据是基于第一神经网络对第一数据进行处理得到的，和/或，第三数据是基于第二神经网络对第二数据进行处理得到的。其中，第一数据集和第二数据集可以包含于 一个数据集。在该一个数据集中，可以包括N份输入数据和M份标签数据，N和M均为正整数；并且，该第一数据集可以包括该N份输入数据，该第二数据集可以包括该M份标签数据。换言之，该第一数据集可以称为输入数据集，神经网络输入数据集等，该第二数据集可以称为标签数据集，神经网络标签数据集等。此外，AI神经网络可以基于该同一数据集进行处理，以实现该AI神经网络的迭代、更新等。
可选地，第一数据集可以包括一个数据集中N份输入数据，为此，该第一数据集也可以替换为其它描述，例如，N份输入数据，数据集中的N份输入数据等。相应的，第二数据集可以包括该一个数据集中的M份标签数据，为此，该第二数据集也可以替换为其它描述，例如，M份标签数据，数据集中的M份标签数据等。
可选地，在N份输入数据中的每一份输入数据对应于M份标签数据中的不同标签数据的情况下，N的取值与M的取值相等。在N份输入数据中的至少两份输入数据对应于M份标签数据中的其中一个相同标签数据的情况下，N的取值可以大于或等于M的取值。在N份输入数据中的其中一份输入数据对应于M份标签数据中的至少两份标签数据的情况下，N的取值可以小于或等于M的取值。
本申请中，AI、神经网络、AI神经网络、机器学习、AI处理，AI神经网络处理等术语可以相互替换。
本申请中，涉及的数据(例如第一数据、第二数据、第三数据，以及第四数据等)，可以替换为信息、信号等。
可选地，在第二数据是基于第一神经网络对第一数据集中的第一数据进行处理得到的情况下，由于第二数据为第一通信装置对第一数据进行处理得到的发送数据，为此，该第一神经网络可以称为部署于发送端的神经网络，编码神经网络，AI编码神经网络等。类似地，在第三数据是基于第二神经网络对第二数据进行处理得到的情况下，由于第三数据为第二通信装置基于第二神经网络对接收的第二数据进行处理得到的数据，为此，该第二神经网络可以称为部署于接收端的神经网络，解码神经网络，AI解码神经网络等。
在第一方面的一种可能的实现方式中，该索引是基于承载该第二数据的资源确定的，包括：该索引的取值是该资源的时域资源索引、该资源的频域资源索引，该资源的资源块大小中的至少一项确定的。
基于上述技术方案，在第一数据在第一数据集中的索引是基于承载该第二数据的资源确定的情况下，该索引具体可以是通过上述至少一项确定的，以提升方案实现的灵活性。
在第一方面的一种可能的实现方式中，在该索引是基于该第一数据集中的数据的处理次数确定的情况下，该方法还包括：该第一通信装置发送第一信息，该第一信息用于指示该第一数据在该第一数据集中的索引。
基于上述技术方案，第一数据在第一数据集中的索引是基于该第一数据集中的数据的处理次数确定的情况下，由于第二通信装置可能无法感知第一数据集中的数据处理次数，为此，第二通信装置可以基于第二数据集中的数据的处理次数确定该索引。相应的，第一通信装置还可以发送第一信息，使得第二通信装置能够基于该第一信息确定该第一数据在该第一数据集中的索引，后续基于该索引在第二数据集中确定第四数据。
应理解，第一通信装置可以基于第一数据集中的数据执行多次处理得到并发送处理结果(例如，其中一个处理结果为基于第一数据得到的第二数据)的过程，相应的，在该多个处理结果中，第一通信装置可以针对部分或全部处理结果发送对应的索引(例如，针对第二数据这个处理结果发送第一信息)。此后，基于该部分或全部处理结果发送对应的索引能够实现数据收发双方对索引的理解的对齐，以避免数据收发双方对索引的理解不对齐导致的数据处理出错，进而提升系统的鲁棒性。
本申请中，“对齐”可以是指不同通信装置之间存在交互消息/数据/信息时，两者对于交互的消息/数据/信息的含义、配置方式、在数据集中的索引等理解一致。
可选地，第一通信装置可以针对部分处理结果发送该部分处理结果对应的索引，而无需针对全部处理结果发送该全部处理结果对应的索引，能够节省开销。
在第一方面的一种可能的实现方式中，该第一信息为基于第一周期传输的多个信息中的其中一个信息；该方法还包括：该第一通信装置接收或发送配置信息，该配置信息用于配置该第一周期。
基于上述技术方案，第一信息可以是基于第一周期传输的周期性信息中的其中一个，在此之前，第 一通信装置可以接收或发送用于配置该第一周期的配置信息，使得该第一通信装置既可以作为该第一周期的配置方，也可以作为第一周期的被配置方，能够使得数据收发双方对第一周期的理解的对齐的同时，也能够提升方案实现的灵活性。
在第一方面的一种可能的实现方式中，该方法还包括：该第一通信装置接收指示该第一数据集的指示信息；和/或，该第一通信装置发送指示该第二数据集的指示信息。
基于上述技术方案，第一通信装置可以作为第一数据集的被配置方，和/或，第一通信装置可以作为第二数据集的配置方，使得数据收发双方能够在交互数据之前实现数据集的获取。
可选地，第一数据集可以是预配置于该第一通信装置的，和/或，第二数据集可以是预配置于该第二通信装置的，通过这种方式，能够降低开销。
在第一方面的一种可能的实现方式中，该方法还包括：该第一通信装置接收基于该第三数据和该第四数据确定的梯度信息和/或损失函数的结果。
基于上述技术方案，第二数据的接收方(例如第二通信装置)能够基于第二数据进行处理得到第三数据，并且，该接收方还能够基于该第三数据和第四数据确定并发送相应的梯度信息和/或损失函数的结果。使得第一通信装置能够在接收梯度信息和/或损失函数的结果之后，能够基于该梯度信息和/或损失函数的结果实现对第一神经网络的更新或迭代。
在第一方面的一种可能的实现方式中，该方法还包括：该第一通信装置接收或发送指示该索引满足该至少一项的指示信息。
基于上述技术方案，第一通信装置还可以接收或发送指示该索引满足该至少一项的指示信息，使得数据收发双方能够基于该指示信息实现数据集中的数据的索引的理解的对齐，以避免数据收发双方对索引的理解不对齐导致的数据处理出错，进而提升系统的鲁棒性。
本申请第二方面提供了一种通信方法，该方法由第二通信装置执行，该第二通信装置可以是通信设备(如，终端设备或网络设备)，或者，该第二通信装置可以是通信设备中的部分组件(例如处理器、芯片或芯片系统等)，或者该第二通信装置还可以是能实现全部或部分通信设备功能的逻辑模块或软件。在第二方面及其可能的实现方式中，以该通信方法由第二通信装置执行为例进行描述，其中，该第二通信装置可以为终端设备或网络设备。在该方法中，第二通信装置接收第二数据，该第二数据是基于第一数据进行处理得到的，该第一数据为第一数据集中的数据；其中，该第一数据在该第一数据集中的索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；该第二通信装置基于该第二数据确定第三数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的；该索引满足以下至少一项：该索引是基于承载该第二数据的资源确定的；该索引是基于该第二数据集中的数据的处理次数确定的。
基于上述技术方案，第二通信装置接收的第二数据是基于第一数据得到的，后续该第二通信装置可以对第二数据进行处理得到第三数据。其中，该第一数据在该第一数据集中的索引用于在第二数据集中确定第四数据，该第四数据是该第一数据对应的标签数据。换言之，该第二通信装置在接收第二数据之后，能够在第二数据集中确定标签数据，并基于该标签数据对第三数据进行处理。并且，第二数据和/或该第三数据是基于神经网络得到的，即用于AI处理的神经网络可以包括部署于第一通信装置的神经网络和/或部署于第二通信装置的神经网络。从而，在通信系统中的通信装置作为AI参与节点的情况下，使得通信装置的算力能够应用于神经网络的AI处理的同时，也能够提升神经网络部署的灵活性。
此外，第一数据在该第一数据集中的索引用于在第二数据集中确定该第一数据对应的标签数据(即第四数据)，并且，该索引满足上述至少一项。换言之，第二通信装置在接收第二数据之后，能够基于承载数据的资源或数据集中的数据的处理次数在第二数据集中确定标签数据，通过这种方式，可以降低空口开销，以提升通信效率。
在第二方面的一种可能的实现方式中，该索引是基于承载该第二数据的资源确定的，包括：该索引的取值是该资源的时域资源索引、该资源的频域资源索引，该资源的资源块大小中的至少一项确定的。
基于上述技术方案，在第一数据在第一数据集中的索引是基于承载该第二数据的资源确定的情况下， 该索引具体可以是通过上述至少一项确定的，以提升方案实现的灵活性。
在第二方面的一种可能的实现方式中，在该索引是基于该第二数据集中的数据的处理次数确定的情况下，该方法还包括：该第二通信装置发送第一信息，该第一信息用于指示该第一数据在该第一数据集中的索引。
基于上述技术方案，第一数据在第一数据集中的索引是基于该第一数据集中的数据的处理次数确定的情况下，由于第二通信装置可能无法感知第一数据集中的数据处理次数，为此，第二通信装置可以基于第二数据集中的数据的处理次数确定该索引。相应的，第二通信装置还可以接收第一信息，使得第二通信装置能够基于该第一信息确定该第一数据在该第一数据集中的索引，后续基于该索引在第二数据集中确定第四数据。
应理解，第一通信装置可以基于第一数据集中的数据执行多次处理得到并发送处理结果(例如，其中一个处理结果为基于第一数据得到的第二数据)的过程，相应的，在该多个处理结果中，第一通信装置可以针对部分或全部处理结果发送对应的索引(例如，针对第二数据这个处理结果发送第一信息)。此后，基于该部分或全部处理结果发送对应的索引能够实现数据收发双方对索引的理解的对齐，以避免数据收发双方对索引的理解不对齐导致的数据处理出错，进而提升系统的鲁棒性。
可选地，第一通信装置可以针对部分处理结果发送该部分处理结果对应的索引，而无需针对全部处理结果发送该全部处理结果对应的索引，能够节省开销。
在第二方面的一种可能的实现方式中，该第一信息为基于第一周期传输的多个信息中的其中一个信息；该方法还包括：该第二通信装置接收或发送配置信息，该配置信息用于配置该第一周期。
基于上述技术方案，第一信息可以是基于第一周期传输的周期性信息中的其中一个，在此之前，第二通信装置可以接收或发送用于配置该第一周期的配置信息，使得该第二通信装置既可以作为该第一周期的配置方，也可以作为第一周期的被配置方，能够使得数据收发双方对第一周期的理解的对齐的同时，也能够提升方案实现的灵活性。
在第二方面的一种可能的实现方式中，该方法还包括：该第二通信装置发送指示该第一数据集的指示信息；和/或，该第二通信装置接收指示该第二数据集的指示信息。
基于上述技术方案，第二通信装置可以作为第一数据集的配置方，和/或，第二通信装置可以作为第二数据集的被配置方，使得数据收发双方能够在交互数据之前实现数据集的获取。
可选地，第一数据集可以是预配置于该第一通信装置的，和/或，第二数据集可以是预配置于该第二通信装置的，通过这种方式，能够降低开销。
在第二方面的一种可能的实现方式中，该方法还包括：该第二通信装置发送基于该第三数据和该第四数据确定的梯度信息和/或损失函数的结果。
基于上述技术方案，第二通信装置能够基于第二数据进行处理得到第三数据，并且，该第二通信装置还能够基于该第三数据和第四数据确定并发送相应的梯度信息和/或损失函数的结果。使得第一通信装置能够在接收梯度信息和/或损失函数的结果之后，能够基于该梯度信息和/或损失函数的结果实现对第一神经网络的更新或迭代。
在第二方面的一种可能的实现方式中，该方法还包括：该第二通信装置接收或发送指示该索引满足所述至少一项的指示信息。
基于上述技术方案，第二通信装置还可以接收或发送指示该索引满足该至少一项的指示信息，使得数据收发双方能够基于该指示信息实现数据集中的数据的索引的理解的对齐，以避免数据收发双方对索引的理解不对齐导致的数据处理出错，进而提升系统的鲁棒性。
本申请第三方面提供了一种通信方法，该第一通信装置可以是通信设备(如，终端设备或网络设备)，或者，该第一通信装置可以是通信设备中的部分组件(例如处理器、芯片或芯片系统等)，或者该第一通信装置还可以是能实现全部或部分通信设备功能的逻辑模块或软件。在第三方面及其可能的实现方式中，以该通信方法由第一通信装置执行为例进行描述，其中，该第一通信装置可以为终端设备或网络设备。在该方法中，第一通信装置对第一数据进行处理，得到第二数据；该第一通信装置发送该第二数据和第四数据，其中，该第二数据用于确定第三数据，该第四数据为该第一数据对应的标签数据；其中， 该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的。
基于上述技术方案，第一通信装置发送的第二数据是基于第一数据得到的，后续该第二数据的接收方(例如第二通信装置)可以对第二数据进行处理得到第三数据。其中，第一通信装置还可以发送第四数据，该第四数据是该第一数据对应的标签数据。相应的，第二数据和第四数据的接收方在接收第二数据之后，能够基于该标签数据对第三数据进行处理。并且，第二数据和/或该第三数据是基于神经网络得到的，即用于AI处理的神经网络可以包括部署于第一通信装置的神经网络和/或部署于第二通信装置的神经网络。从而，在通信系统中的通信装置作为AI参与节点的情况下，使得通信装置的算力能够应用于神经网络的AI处理的同时，也能够提升神经网络部署的灵活性。
此外，第一通信装置发送的第二数据为第一数据的处理结果，该第一通信装置发送的第四数据为该第一数据对应的标签数据，从而，通过发送第一数据的处理结果和第一数据对应的标签数据的方式，使得接收方能够基于该标签数据进一步处理的同时，也能够使得方案能够适用于标签数据的数据量较小的场景，以尽可能地减少标签数据的传输开销。
在第三方面的一种可能的实现方式中，该方法还包括：该第一通信装置接收基于该第三数据和该第四数据确定的梯度信息和/或损失函数的结果。
基于上述技术方案，第二数据的接收方(例如第二通信装置)能够基于第二数据进行处理得到第三数据，并且，该接收方还能够基于该第三数据和第四数据确定并发送相应的梯度信息和/或损失函数的结果。使得第一通信装置能够在接收梯度信息和/或损失函数的结果之后，能够基于该梯度信息和/或损失函数的结果实现对第一神经网络的更新或迭代。
本申请第四方面提供了一种通信方法，该方法由第二通信装置执行，该第二通信装置可以是通信设备(如，终端设备或网络设备)，或者，该第二通信装置可以是通信设备中的部分组件(例如处理器、芯片或芯片系统等)，或者该第二通信装置还可以是能实现全部或部分通信设备功能的逻辑模块或软件。在第四方面及其可能的实现方式中，以该通信方法由第二通信装置执行为例进行描述，其中，该第二通信装置可以为终端设备或网络设备。在该方法中，第二通信装置接收第二数据和第四数据，其中，该第二数据是基于第一数据得到的，该第四数据为该第一数据对应的标签数据；该第二通信装置基于该第二数据确定第三数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的。
基于上述技术方案，第二通信装置接收的第二数据是基于第一数据得到的，后续该第二通信装置可以对第二数据进行处理得到第三数据。其中，第二通信装置还可以接收第四数据，该第四数据是该第一数据对应的标签数据。相应的，该第二通信装置在接收第二数据之后，能够基于该标签数据对第三数据进行处理。并且，第二数据和/或该第三数据是基于神经网络得到的，即用于AI处理的神经网络可以包括部署于第一通信装置的神经网络和/或部署于第二通信装置的神经网络。从而，在通信系统中的通信装置作为AI参与节点的情况下，使得通信装置的算力能够应用于神经网络的AI处理的同时，也能够提升神经网络部署的灵活性。
此外，第二通信装置接收的第二数据为第一数据的处理结果，该第二通信装置发送的第四数据为该第一数据对应的标签数据，从而，通过发送第一数据的处理结果和第一数据对应的标签数据的方式，使得该第二通信装置能够基于该标签数据进一步处理的同时，也能够使得方案能够适用于标签数据的数据量较小的场景，以尽可能地减少标签数据的传输开销。
在第四方面的一种可能的实现方式中，该方法还包括：该第二通信装置发送基于该第三数据和该第四数据确定的梯度信息和/或损失函数的结果。
基于上述技术方案，第二通信装置能够基于第二数据进行处理得到第三数据，并且，该第二通信装置还能够基于该第三数据和第四数据确定并发送相应的梯度信息和/或损失函数的结果。使得第一通信装置能够在接收梯度信息和/或损失函数的结果之后，能够基于该梯度信息和/或损失函数的结果实现对第一神经网络的更新或迭代。
本申请第五方面提供了一种通信方法，该方法由第一通信装置执行，该第一通信装置可以是通信设备(如，终端设备或网络设备)，或者，该第一通信装置可以是通信设备中的部分组件(例如处理器、芯片或芯片系统等)，或者该第一通信装置还可以是能实现全部或部分通信设备功能的逻辑模块或软件。在第五方面及其可能的实现方式中，以该通信方法由第一通信装置执行为例进行描述，其中，该第一通信装置可以为终端设备或网络设备。在该方法中，第一通信装置对第一数据进行处理，得到第二数据；其中，该第一数据为第一数据集中的数据；该第一通信装置发送该第二数据和第一索引，该第二数据用于确定第三数据；其中，该第一索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第四数据是基于神经网络对该第二数据进行处理得到的。
基于上述技术方案，第一通信装置发送的第二数据是基于第一数据得到的，后续该第二数据的接收方(例如第二通信装置)可以对第二数据进行处理得到第三数据。其中，第一通信装置还可以发送第一索引。相应的，第二数据和第四数据的接收方在接收第二数据之后，能够基于该标签数据对第三数据进行处理。并且，第二数据和/或该第三数据是基于神经网络得到的，即用于AI处理的神经网络可以包括部署于第一通信装置的神经网络和/或部署于第二通信装置的神经网络。从而，在通信系统中的通信装置作为AI参与节点的情况下，使得通信装置的算力能够应用于神经网络的AI处理的同时，也能够提升神经网络部署的灵活性。
此外，第一通信装置发送的第二数据为第一数据的处理结果，该第一通信装置发送的第四数据为该第一数据对应的标签数据，从而，通过发送第一数据的处理结果和第一数据对应的标签数据的方式，使得接收方能够基于该标签数据进一步处理的同时，也能够使得方案能够适用于标签数据的数据量较小的场景，以尽可能地减少标签数据的传输开销。
应理解，在上述技术方案中，第二数据是基于第一神经网络对第一数据进行处理得到的，和/或，第三数据是基于第二神经网络对第二数据进行处理得到的。其中，第一数据集和第二数据集可以包含于一个数据集。在该一个数据集中，可以包括N份输入数据和M份标签数据，N和M均为正整数；并且，该第一数据集可以包括该N份输入数据，该第二数据集可以包括该M份标签数据。换言之，该第一数据集可以称为输入数据集，神经网络输入数据集等，该第二数据集可以称为标签数据集，神经网络标签数据集等。此外，AI神经网络可以基于该同一数据集进行处理，以实现该AI神经网络的迭代、更新等。
在一种可能的实现方式中，该第一索引是该第一数据在该第一数据集中的第二索引确定的。
一种实现示例中，在N份输入数据中的每一份输入数据对应于M份标签数据中的不同标签数据的情况下，N的取值与M的取值相等。
例如，第一索引可以与第一数据在该第一数据集中的第二索引相同。换言之，第一数据集中的第i份输入数据的标签数据为第二数据集中的第j份数据，i为第一索引，j为第二索引，并且，i与j相等。即N份输入数据的第一份数据的标签数据为M份标签数据中的第一份数据，N份输入数据的第二份数据的标签数据为M份标签数据中的第二份数据，以此类推，N份输入数据的第N份数据的标签数据为M份标签数据中的第M份数据(N与M相等)。
又如，第一索引可以与第一数据在该第一数据集中的第二索引部分不相同或完全不相同。换言之，第一数据集中的第i份输入数据的标签数据为第二数据集中的第j份数据，i为第一索引，j为第二索引，并且，i与j可以是部分或全部不相等。其中，i与j之间的映射关系可以是预配置的。
作为i与j之间的映射关系可以是预配置的一种示例，i可以是从1遍历至N，j可以是从N(N等于M)遍历至1。例如，以N和M的取值均为3为例，N份输入数据的第一份数据的标签数据为M份标签数据中的第三份数据，N份输入数据的第二份数据的标签数据为M份标签数据中的第二份数据，N份输入数据的第三份数据的标签数据为M份标签数据中的第一份数据。在该示例中，第一索引可以与第一数据在该第一数据集中的第二索引部分不相同。
作为i与j之间的映射关系可以是预配置的另一种示例，i和j之间的映射关系可以通过配置的方式或预配置的方式使得数据收发双方对齐。例如，仍以N和M的取值均为3为例，N份输入数据的第一份数据的标签数据为M份标签数据中的第三份数据，N份输入数据的第二份数据的标签数据为M份标签 数据中的第一份数据，N份输入数据的第三份数据的标签数据为M份标签数据中的第二份数据。在该示例中，第一索引可以与第一数据在该第一数据集中的第二索引完全不相同。
另一种实现示例中，该第一索引与第一数据在该第一数据集中的第二索引之间的映射关系是预配置的。
例如，在N份输入数据中的至少两份输入数据对应于M份标签数据中的其中一个相同标签数据的情况下，N的取值可以大于或等于M的取值。相应的，在这种情况下，第一索引可以小于或等于第一数据在该第一数据集中的第二索引。
又如，在N份输入数据中的其中一份输入数据对应于M份标签数据中的至少两份标签数据的情况下，N的取值可以小于或等于M的取值。相应的，在这种情况下，第一索引可以大于或等于第一数据在该第一数据集中的第二索引。
在第五方面的一种可能的实现方式中，该方法还包括：该第一通信装置接收基于该第三数据和该第四数据确定的梯度信息和/或损失函数的结果。
基于上述技术方案，第二数据的接收方(例如第二通信装置)能够基于第二数据进行处理得到第三数据，并且，该接收方还能够基于该第三数据和第四数据确定并发送相应的梯度信息和/或损失函数的结果。使得第一通信装置能够在接收梯度信息和/或损失函数的结果之后，能够基于该梯度信息和/或损失函数的结果实现对第一神经网络的更新或迭代。
本申请第六方面提供了一种通信方法，该方法由第二通信装置执行，该第二通信装置可以是通信设备(如，终端设备或网络设备)，或者，该第二通信装置可以是通信设备中的部分组件(例如处理器、芯片或芯片系统等)，或者该第二通信装置还可以是能实现全部或部分通信设备功能的逻辑模块或软件。在第六方面及其可能的实现方式中，以该通信方法由第二通信装置执行为例进行描述，其中，该第二通信装置可以为终端设备或网络设备。在该方法中，第二通信装置接收第二数据和第一索引，该第二数据是基于第一数据得到的，该第一数据为第一数据集中的数据；其中，该第一索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；该第二通信装置基于该第二数据确定第三数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的。
基于上述技术方案，第二通信装置接收的第二数据是基于第一数据得到的，后续该第二通信装置可以对第二数据进行处理得到第三数据。其中，第二通信装置还可以接收第一索引。相应的，该第二通信装置在接收第二数据之后，能够基于该标签数据对第三数据进行处理。并且，第二数据和/或该第三数据是基于神经网络得到的，即用于AI处理的神经网络可以包括部署于第一通信装置的神经网络和/或部署于第二通信装置的神经网络。从而，在通信系统中的通信装置作为AI参与节点的情况下，使得通信装置的算力能够应用于神经网络的AI处理的同时，也能够提升神经网络部署的灵活性。
此外，第二通信装置接收的第二数据为第一数据的处理结果，该第二通信装置接收的第四数据为该第一数据对应的标签数据，从而，通过发送第一数据的处理结果和第一数据对应的标签数据的方式，使得该第二通信装置能够基于该标签数据进一步处理的同时，也能够使得方案能够适用于标签数据的数据量较小的场景，以尽可能地减少标签数据的传输开销。
可选地，该第一索引是该第一数据在该第一数据集中的第二索引确定的。
在第六方面的一种可能的实现方式中，该方法还包括：该第二通信装置发送基于该第三数据和该第四数据确定的梯度信息和/或损失函数的结果。
基于上述技术方案，第二通信装置能够基于第二数据进行处理得到第三数据，并且，该第二通信装置还能够基于该第三数据和第四数据确定并发送相应的梯度信息和/或损失函数的结果。使得第一通信装置能够在接收梯度信息和/或损失函数的结果之后，能够基于该梯度信息和/或损失函数的结果实现对第一神经网络的更新或迭代。
本申请第七方面提供了一种通信装置，该装置为第一通信装置，该装置包括收发单元和处理单元，该处理单元用于对第一数据进行处理，得到第二数据；其中，该第一数据为第一数据集中的数据；该收 发单元用于发送该第二数据，该第二数据用于确定第三数据；其中，该第一数据在该第一数据集中的索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的；该索引满足以下至少一项：该索引是基于承载该第二数据的资源确定的；该索引是基于该第一数据集中的数据的处理次数确定的。
本申请第七方面中，通信装置的组成模块还可以用于执行第一方面的各个可能实现方式中所执行的步骤，并实现相应的技术效果，具体均可以参阅第一方面，此处不再赘述。
本申请第八方面提供了一种通信装置，该装置为第二通信装置，该装置包括收发单元和处理单元，该收发单元用于接收第二数据，该第二数据是基于第一数据进行处理得到的，该第一数据为第一数据集中的数据；其中，该第一数据在该第一数据集中的索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；该处理单元用于基于该第二数据确定第三数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的；该索引满足以下至少一项：该索引是基于承载该第二数据的资源确定的；该索引是基于该第二数据集中的数据的处理次数确定的。
本申请第八方面中，通信装置的组成模块还可以用于执行第二方面的各个可能实现方式中所执行的步骤，并实现相应的技术效果，具体均可以参阅第二方面，此处不再赘述。
本申请第九方面提供了一种通信装置，该装置为第一通信装置，该装置包括收发单元和处理单元，该处理单元用于对第一数据进行处理，得到第二数据；该收发单元用于发送该第二数据和第四数据，其中，该第二数据用于确定第三数据，该第四数据为该第一数据对应的标签数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的。
本申请第九方面中，通信装置的组成模块还可以用于执行第三方面的各个可能实现方式中所执行的步骤，并实现相应的技术效果，具体均可以参阅第三方面，此处不再赘述。
本申请第十方面提供了一种通信装置，该装置为第二通信装置，该装置包括收发单元和处理单元，该收发单元用于接收第二数据和第四数据，其中，该第二数据是基于第一数据得到的，该第四数据为该第一数据对应的标签数据；该处理单元用于基于该第二数据确定第三数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的。
本申请第十方面中，通信装置的组成模块还可以用于执行第四方面的各个可能实现方式中所执行的步骤，并实现相应的技术效果，具体均可以参阅第四方面，此处不再赘述。
本申请第十一方面提供了一种通信装置，该装置为第一通信装置，该装置包括收发单元和处理单元，该处理单元用于对第一数据进行处理，得到第二数据；其中，该第一数据为第一数据集中的数据；该收发单元用于发送该第二数据和第一索引，该第二数据用于确定第三数据；其中，该第一索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第四数据是基于神经网络对该第二数据进行处理得到的。
本申请第十一方面中，通信装置的组成模块还可以用于执行第五方面的各个可能实现方式中所执行的步骤，并实现相应的技术效果，具体均可以参阅第五方面，此处不再赘述。
本申请第十二方面提供了一种通信装置，该装置为第二通信装置，该装置包括收发单元和处理单元，该收发单元用于接收第二数据和第一索引，该第二数据是基于第一数据得到的，该第一数据为第一数据集中的数据；其中，该第一索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；该处理单元用于基于该第二数据确定第三数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的。
本申请第十二方面中，通信装置的组成模块还可以用于执行第六方面的各个可能实现方式中所执行 的步骤，并实现相应的技术效果，具体均可以参阅第六方面，此处不再赘述。
本申请第十三方面提供了一种通信装置，包括至少一个处理器，所述至少一个处理器与存储器耦合；该存储器用于存储程序或指令；该至少一个处理器用于执行该程序或指令，以使该装置实现前述第一方面至第六方面任一方面中的任意一种可能的实现方式所述的方法。
本申请第十四方面提供了一种通信装置，包括至少一个逻辑电路和输入输出接口；该逻辑电路用于执行如前述第一方面至第六方面任一方面中的任意一种可能的实现方式所述的方法。
本申请第十五方面提供了一种通信系统，该通信系统包括上述第一通信装置以及第二通信装置。
本申请第十六方面提供一种计算机可读存储介质，该存储介质用于存储一个或多个计算机执行指令，当计算机执行指令被处理器执行时，该处理器执行如上述第一方面至第六方面中任一方面的任意一种可能的实现方式所述的方法。
本申请第十七方面提供一种计算机程序产品(或称计算机程序)，当计算机程序产品中的计算机程序被该处理器执行时，该处理器执行上述第一方面至第六方面中任一方面的任意一种可能的实现方式所述的方法。
本申请第十八方面提供了一种芯片系统，该芯片系统包括至少一个处理器，用于支持通信装置实现上述第一方面至第六方面中任一方面的任意一种可能的实现方式所述的方法。
在一种可能的设计中，该芯片系统还可以包括存储器，存储器，用于保存该第一通信装置必要的程序指令和数据。该芯片系统，可以由芯片构成，也可以包含芯片和其他分立器件。可选的，所述芯片系统还包括接口电路，所述接口电路为所述至少一个处理器提供程序指令和/或数据。
其中，第七方面至第十八方面中任一种设计方式所带来的技术效果可参见上述第一方面至第六方面中不同设计方式所带来的技术效果，在此不再赘述。
附图说明
图1a至图1c为本申请提供的通信系统的示意图；
图1d、图1e以及图2a至图2f为本申请涉及的AI处理过程的示意图；
图3为本申请提供的通信方法的交互示意图；
图4至图5为本申请提供的AI处理过程的示意图；
图6至图7为本申请提供的通信方法的交互示意图；
图8至图12为本申请提供的通信装置的示意图。
具体实施方式
首先，对本申请实施例中的部分用语进行解释说明，以便于本领域技术人员理解。
(1)终端设备：可以是能够接收网络设备调度和指示信息的无线终端设备，无线终端设备可以是指向用户提供语音和/或数据连通性的设备，或具有无线连接功能的手持式设备，或连接到无线调制解调器的其他处理设备。
终端设备可以经无线接入网(radio access network，RAN)与一个或多个核心网或者互联网进行通信，终端设备可以是移动终端设备，如移动电话(或称为“蜂窝”电话，手机(mobile phone))、计算机和数据卡，例如，可以是便携式、袖珍式、手持式、计算机内置的或者车载的移动装置，它们与无线接入网交换语音和/或数据。例如，个人通信业务(personal communication service，PCS)电话、无绳电话、会话发起协议(SIP)话机、无线本地环路(wireless local loop，WLL)站、个人数字助理(personal digital assistant，PDA)、平板电脑(Pad)、带无线收发功能的电脑等设备。无线终端设备也可以称为系统、订户单元(subscriber unit)、订户站(subscriber station)，移动站(mobile station)、移动台(mobile station，MS)、远程站(remote station)、接入点(access point，AP)、远程终端设备(remote terminal)、接入终端设备(access terminal)、用户终端设备(user terminal)、用户代理(user agent)、用户站(subscriber station，SS)、用户端设备(customer premises equipment，CPE)、终端(terminal)、用户设备(user equipment，UE)、移动终端(mobile terminal，MT)等。
作为示例而非限定，在本申请实施例中，该终端设备还可以是可穿戴设备。可穿戴设备也可以称为穿戴式智能设备或智能穿戴式设备等，是应用穿戴式技术对日常穿戴进行智能化设计、开发出可以穿戴的设备的总称，如眼镜、手套、手表、服饰及鞋等。可穿戴设备即直接穿在身上，或是整合到用户的衣服或配件的一种便携式设备。可穿戴设备不仅仅是一种硬件设备，更是通过软件支持以及数据交互、云端交互来实现强大的功能。广义穿戴式智能设备包括功能全、尺寸大、可不依赖智能手机实现完整或者部分的功能，例如：智能手表或智能眼镜等，以及只专注于某一类应用功能，需要和其它设备如智能手机配合使用，如各类进行体征监测的智能手环、智能头盔、智能首饰等。
终端还可以是无人机、机器人、设备到设备通信(device-to-device，D2D)中的终端、车到一切(vehicle to everything，V2X)中的终端、虚拟现实(virtual reality，VR)终端设备、增强现实(augmented reality，AR)终端设备、工业控制(industrial control)中的无线终端、无人驾驶(self driving)中的无线终端、远程医疗(remote medical)中的无线终端、智能电网(smart grid)中的无线终端、运输安全(transportation safety)中的无线终端、智慧城市(smart city)中的无线终端、智慧家庭(smart home)中的无线终端等。
此外，终端设备也可以是第五代(5th generation，5G)通信系统之后演进的通信系统(例如第六代(6th generation，6G)通信系统等)中的终端设备或者未来演进的公共陆地移动网络(public land mobile network，PLMN)中的终端设备等。示例性的，6G网络可以进一步扩展5G通信终端的形态和功能，6G终端包括但不限于车、蜂窝网络终端(融合卫星终端功能)、无人机、物联网(internet of things，IoT)设备。
在本申请实施例中，上述终端设备还可以获得网络设备提供的AI服务。可选地，终端设备还可以具有AI处理能力。
(2)网络设备：可以是无线网络中的设备，例如网络设备可以为将终端设备接入到无线网络的RAN节点(或设备)，又可以称为基站。目前，一些RAN设备的举例为：基站(base station)、演进型基站(evolved NodeB，eNodeB)、5G通信系统中的基站gNB(gNodeB)、传输接收点(transmission reception point，TRP)、演进型节点B(evolved Node B，eNB)、无线网络控制器(radio network controller，RNC)、节点B(Node B，NB)、家庭基站(例如，home evolved Node B，或home Node B，HNB)、基带单元(base band unit，BBU)，或无线保真(wireless fidelity，Wi-Fi)接入点AP等。另外，在一种网络结构中，网络设备可以包括集中单元(centralized unit，CU)节点、或分布单元(distributed unit，DU)节点、或包括CU节点和DU节点的RAN设备。
可选的，RAN节点还可以是宏基站、微基站或室内站、中继节点或施主节点、或者是云无线接入网络(cloud radio access network，CRAN)场景下的无线控制器。RAN节点还可以是服务器，可穿戴设备，车辆或车载设备等。例如，车辆外联(vehicle to everything，V2X)技术中的接入网设备可以为路侧单元(road side unit，RSU)。
在另一种可能的场景中，由多个RAN节点协作协助终端实现无线接入，不同RAN节点分别实现基站的部分功能。例如，RAN节点可以是集中式单元(central unit，CU)，分布式单元(distributed unit，DU)，CU-控制面(control plane，CP)，CU-用户面(user plane，UP)，或者无线单元(radio unit，RU)等。CU和DU可以是单独设置，或者也可以包括在同一个网元中，例如基带单元(baseband unit，BBU)中。RU可以包括在射频设备或者射频单元中，例如包括在射频拉远单元(remote radio unit，RRU)、有源天线处理单元(active antenna unit，AAU)或远程射频头(remote radio head，RRH)中。
在不同系统中，CU(或CU-CP和CU-UP)、DU或RU也可以有不同的名称，但是本领域的技术人员可以理解其含义。例如，在开放式接入网(open RAN，O-RAN或ORAN)系统中，CU也可以称为O-CU(开放式CU)，DU也可以称为O-DU，CU-CP也可以称为O-CU-CP，CU-UP也可以称为O-CU-UP，RU也可以称为O-RU。为描述方便，本申请中以CU，CU-CP，CU-UP、DU和RU为例进行描述。本申请中的CU(或CU-CP、CU-UP)、DU和RU中的任一单元，可以是通过软件模块、硬件模块、或者软件模块与硬件模块结合来实现。
接入网设备和终端设备之间的通信遵循一定的协议层结构。该协议层可以包括控制面协议层和用户面协议层。控制面协议层可以包括以下至少一项：无线资源控制(radio resource control，RRC)层、分组数据汇聚层协议(packet data convergence protocol，PDCP)层、无线链路控制(radio link control， RLC)层、媒体接入控制(media access control，MAC)层、或物理(physical，PHY)层等。用户面协议层可以包括以下至少一项：业务数据适配协议(service data adaptation protocol，SDAP)层、PDCP层、RLC层、MAC层、或物理层等。
对于ORAN系统中的网元及其可实现的协议层功能对应关系，可参照下表1。
表1
网络设备可以是其它为终端设备提供无线通信功能的装置。本申请的实施例对网络设备所采用的具体技术和具体设备形态不做限定。为方便描述，本申请实施例并不限定。
网络设备还可以包括核心网设备，核心网设备例如包括第四代(4th generation，4G)网络中的移动性管理实体(mobility management entity，MME)，归属用户服务器(home subscriber server，HSS)，服务网关(serving gateway，S-GW)，策略和计费规则功能(policy and charging rules function，PCRF)，公共数据网网关(public data network gateway，PDN gateway，P-GW)；5G网络中的访问和移动管理功能(access and mobility management function，AMF)、用户面功能(user plane function，UPF)或会话管理功能(session management function，SMF)等网元。此外，该核心网设备还可以包括5G网络以及5G网络的下一代网络中的其他核心网设备。
本申请实施例中，上述网络设备还可以具有AI能力的网络节点，可以为终端或其他网络设备提供AI服务，例如，可以为网络侧(接入网或核心网)的AI节点、算力节点、具有AI能力的RAN节点、具有AI能力的核心网网元等。
本申请实施例中，用于实现网络设备的功能的装置可以是网络设备，也可以是能够支持网络设备实现该功能的装置，例如芯片系统，该装置可以被安装在网络设备中。在本申请实施例提供的技术方案中，以用于实现网络设备的功能的装置是网络设备为例，描述本申请实施例提供的技术方案。
(3)配置与预配置：在本申请中，会同时用到配置与预配置。其中，配置是指网络设备/服务器通过消息或信令将一些参数的配置信息或参数的取值发送给终端，以便终端根据这些取值或信息来确定通信的参数或传输时的资源。预配置与配置类似，可以是网络设备/服务器预先与终端设备协商好的参数信息或参数值，也可以是标准协议规定的基站/网络设备或终端设备采用的参数信息或参数值，还可以是预先存储在基站/服务器或终端设备的参数信息或参数值。本申请对此不做限定。
进一步地，这些取值和参数，是可以变化或更新的。
(4)本申请实施例中的术语“系统”和“网络”可被互换使用。“多个”是指两个或两个以上。“和/或”，描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A、同时存在A和B、单独存在B的情况，其中A，B可以是单数或者复数。字符“/”一般表示前后关联对象是一种“或”的关系。“以下至少一项(个)”或其类似表达，是指的这些项中的任意组合，包括单项(个)或复数项(个)的任意组合。例如“A，B和C中的至少一项”包括A，B，C，AB，AC，BC或ABC。以及，除非有特别说明，本申请实施例提及“第一”、“第二”等序数词是用于对多个对象进行区分，不用于限定多个对象的顺序、时序、优先级或者重要程度。
(5)本申请实施例中的“发送”和“接收”，表示信号传递的走向。例如，“向XX发送信息”可以理解为该信息的目的端是XX，可以包括通过空口直接发送，也包括其他单元或模块通过空口间接发送。“接收来自YY的信息”可以理解为该信息的源端是YY，可以包括通过空口直接从YY接收，也可以包括通过空口从其他单元或模块间接地从YY接收。“发送”也可以理解为芯片接口的“输出”，“接收”也可以理解为芯片接口的“输入”。
换言之，发送和接收可以是在设备之间进行的，例如，网络设备和终端设备之间进行的，也可以是在设备内进行的，例如，通过总线、走线或接口在设备内的部件之间、模组之间、芯片之间、软件模块或者硬件模块之间发送或接收。
可以理解的是，信息在信息发送的源端和目的端之间可能会被进行必要的处理，比如编码、调制等，但目的端可以理解来自源端的有效信息。本申请中类似的表述可以做相似的理解，不再赘述。
(6)在本申请实施例中，“指示”可以包括直接指示和间接指示，也可以包括显式指示和隐式指示。将某一信息(如下文所述的指示信息)所指示的信息称为待指示信息，则具体实现过程中，对待指示信息进行指示的方式有很多种，例如但不限于，可以直接指示待指示信息，如待指示信息本身或者该待指示信息的索引等。也可以通过指示其他信息来间接指示待指示信息，其中该其他信息与待指示信息之间存在关联关系；还可以仅仅指示待指示信息的一部分，而待指示信息的其他部分则是已知的或者提前约定的，例如可以借助预先约定(例如协议预定义)的各个信息的排列顺序来实现对特定信息的指示，从而在一定程度上降低指示开销。本申请对于指示的具体方式不作限定。可以理解的是，对于该指示信息的发送方来说，该指示信息可用于指示待指示信息，对于指示信息的接收方来说，该指示信息可用于确定待指示信息。
本申请中，除特殊说明外，各个实施例之间相同或相似的部分可以互相参考。在本申请中各个实施例、以及各实施例中的各个方法/设计/实现方式中，如果没有特殊说明以及逻辑冲突，不同的实施例之间、以及各实施例中的各个方法/设计/实现方式之间的术语和/或描述具有一致性、且可以相互引用，不同的实施例、以及各实施例中的各个方法/设计/实现方式中的技术特征根据其内在的逻辑关系可以组合形成新的实施例、方法、或实现方式。以下所述的本申请实施方式并不构成对本申请保护范围的限定。
本申请可以应用于长期演进(long term evolution，LTE)系统、新无线(new radio，NR)系统，或者是5G之后演进的通信系统(例如6G等)。其中，该通信系统中包括至少一个网络设备和/或至少一个终端设备。
请参阅图1a，为本申请中通信系统的一种示意图。图1a中，示例性的示出了一个网络设备和6个终端设备，6个终端设备分别为终端设备1、终端设备2、终端设备3、终端设备4、终端设备5以及终端设备6等。在图1a所示的示例中，是以终端设备1为智能茶杯，终端设备2为智能空调，终端设备3为智能加油机，终端设备4为交通工具，终端设备5为手机，终端设备6为打印机进行举例说明的。
如图1a所示，AI配置信息发送实体可以为网络设备。AI配置信息接收实体可以为终端设备1-终端设备6，此时，网络设备和终端设备1-终端设备6组成一个通信系统，在该通信系统中，终端设备1-终端设备6可以发送数据给网络设备，网络设备需要接收终端设备1-终端设备6发送的数据。同时，网络设备可以向终端设备1-终端设备6发送配置信息。
示例性的，在图1a中，终端设备4-终端设备6也可以组成一个通信系统。其中，终端设备5作为网络设备，即AI配置信息发送实体；终端设备4和终端设备6作为终端设备，即AI配置信息接收实体。例如车联网系统中，终端设备5分别向终端设备4和终端设备6发送AI配置信息，并且接收终端设备4和终端设备6发送的数据；相应的，终端设备4和终端设备6接收终端设备5发送的AI配置信息，并向终端设备5发送数据。
以图1a所示通信系统为例，不同的设备之间(包括网络设备与网络设备之间，网络设备与终端设备之间，和/或，终端设备和终端设备之间)除了执行通信相关业务之外，还有可能执行AI相关业务。
如图1b所示，以网络设备为基站为例，基站可以与一个或多个终端设备之间可以执行通信相关业务和AI相关业务，不同终端设备之间也可以执行通信相关业务和AI相关业务。
如图1c所示，以终端设备包括电视和手机为例，电视和手机之间也可以执行通信相关业务和AI相关业务。
本申请提供的技术方案可以应用于无线通信系统(例如图1a、图1b或图1c所示系统)，例如本申请提供的通信系统中可以引入AI网元来实现部分或全部AI相关的操作。AI网元也可以称为AI节点、AI设备、AI实体、AI模块、AI模型、或AI单元等。所述AI网元可以是内置在通信系统的网元中。例如，AI网元可以是内置在：接入网设备、核心网设备、云服务器、或网管(operation,administration and maintenance，OAM)中的AI模块，用以实现AI相关的功能。所述OAM可以是作为核心网设备网管和/ 或作为接入网设备的网管。或者，所述AI网元也可以是通信系统中独立设置的网元。可选的，终端或终端内置的芯片中也可以包括AI实体，用于实现AI相关的功能。
下面将本申请中可能涉及到的人工智能(artificial intelligence，AI)进行简要介绍。
人工智能(artificial intelligence，AI)，可以让机器具有人类的智能，例如可以让机器应用计算机的软硬件来模拟人类某些智能行为。为了实现人工智能，可以采用机器学习方法。机器学习方法中，机器利用训练数据学习(或训练)得到模型。该模型表征了从输入到输出之间的映射。学习得到的模型可以用于进行推理(或预测)，即可以利用该模型预测出给定输入所对应的输出。其中，该输出还可以称为推理结果(或预测结果)。
机器学习可以包括监督学习、无监督学习、和强化学习。其中，无监督学习还可以称为非监督学习。
以监督学习为例，监督学习可以依据已采集到的样本值和样本标签，利用机器学习算法学习样本值到样本标签的映射关系，并用AI模型来表达学到的映射关系。训练机器学习模型的过程就是学习这种映射关系的过程。在训练过程中，将样本值输入模型得到模型的预测值，通过计算模型的预测值与样本标签(理想值)之间的误差来优化模型参数。映射关系学习完成后，就可以利用学到的映射来预测新的样本标签。监督学习学到的映射关系可以包括线性映射或非线性映射。根据标签的类型可将学习的任务分为分类任务和回归任务。
换言之，监督学习的目标可以是给定一个训练集(包含多对输入数据及标签数据)，学习训练集中的输入数据和输出数据(即标签数据)的映射关系，同时，希望其映射关系还能应用于训练集之外的数据。训练集为正确的输入输出对的集合。
神经网络(neural network,NN)是机器学习技术中的一种具体的模型。根据通用近似定理，神经网络在理论上可以逼近任意连续函数，从而使得神经网络具备学习任意映射的能力。传统的通信系统需要借助丰富的专家知识来设计通信模块，而基于神经网络的深度学习通信系统可以从大量的数据集中自动发现隐含的模式结构，建立数据之间的映射关系，获得优于传统建模方法的性能。
神经网络的思想来源于大脑组织的神经元结构。例如，每个神经元都对其输入值进行加权求和运算，通过一个激活函数输出运算结果。
如图1d所示，为神经元结构的一种示意图。假设神经元的输入为x＝[x0,x1,…,xn]，与各个输入对应的权值分别为w＝[w,w1,…,wn]，其中，n为正整数，wi和xi可以是小数、整数(例如0、正整数或负整数等)、或复数等各种可能的类型。wi作为xi的权值，用于对xi进行加权。根据权值对输入值进行加权求和的偏置例如为b。激活函数的形式可以有多种，假设一个神经元的激活函数为：y＝f(z)＝max(0,z)，则该神经元的输出为：再例如，一个神经元的激活函数为：y＝f(z)＝z，则该神经元的输出为：其中，b可以是小数、整数(例如0、正整数或负整数)、或复数等各种可能的类型。神经网络中不同神经元的激活函数可以相同或不同。
此外，神经网络一般包括多个层，每层可包括一个或多个神经元。通过增加神经网络的深度和/或宽度，能够提高该神经网络的表达能力，为复杂系统提供更强大的信息提取和抽象建模能力。其中，神经网络的深度可以是指神经网络包括的层数，每层包括的神经元个数可以称为该层的宽度。在一种实现方式中，神经网络包括输入层和输出层。神经网络的输入层将接收到的输入信息经过神经元处理，将处理结果传递给输出层，由输出层得到神经网络的输出结果。在另一种实现方式中，神经网络包括输入层、隐藏层和输出层。神经网络的输入层将接收到的输入信息经过神经元处理，将处理结果传递给中间的隐藏层，隐藏层对接收的处理结果进行计算，得到计算结果，隐藏层将计算结果传递给输出层或者下一个相邻的隐藏层，最终由输出层得到神经网络的输出结果。其中，一个神经网络可以包括一个隐藏层，或者包括多个依次连接的隐藏层，不予限制。
神经网络例如为深度神经网络(deep neural network,DNN)。根据网络的构建方式，DNN可以包括前馈神经网络(feedforward neural network,FNN)、卷积神经网络(convolutional neural networks, CNN)和递归神经网络(recurrent neural network,RNN)。
图1e为一种FNN网络示意图。FNN网络的特点为相邻层的神经元之间两两完全相连。该特点使得FNN通常需要大量的存储空间、导致较高的计算复杂度。
CNN是一种专门来处理具有类似网格结构的数据的神经网络。例如，时间序列数据(时间轴离散采样)和图像数据(二维离散采样)都可以认为是类似网格结构的数据。CNN并不一次性利用全部的输入信息做运算，而是采用一个固定大小的窗截取部分信息做卷积运算，这就大大降低了模型参数的计算量。另外根据窗截取的信息类型的不同(如同一副图中的人和物为不同类型信息)，每个窗可以采用不同的卷积核运算，这使得CNN能更好的提取输入数据的特征。
RNN是一类利用反馈时间序列信息的DNN网络。它的输入包括当前时刻的新的输入值和自身在前一时刻的输出值。RNN适合获取在时间上具有相关性的序列特征，特别适用于语音识别、信道编译码等应用。
在上述机器学习的模型训练过程中，可以定义损失函数。损失函数描述了模型的输出值和理想目标值之间的差距或差异。损失函数可以通过多种形式体现，对于损失函数的具体形式不予限制。模型训练过程可以看作以下过程：通过调整模型的部分或全部参数，使得损失函数的值小于门限值或者满足目标需求。
模型还可以被称为AI模型、规则或者其他名称等。AI模型可以认为是实现AI功能的具体方法。AI模型表征了模型的输入和输出之间的映射关系或者函数。AI功能可以包括以下一项或多项：数据收集、模型训练(或模型学习)、模型信息发布、模型推断(或称为模型推理、推理、或预测等)、模型监控或模型校验、或推理结果发布等。AI功能还可以称为AI(相关的)操作、或AI相关的功能。
下面将结合附图，对神经网络的实现过程进行示例性描述。
1.全连接神经网络,又叫多层感知机(multilayer perceptron，MLP)。
如图2a所示，一个MLP包含一个输入层(左侧)，一个输出层(右侧)，及多个隐藏层(中间)。其中，MLP的每层包含若干个节点，称为神经元。其中,相邻两层的神经元间两两相连。
可选的，考虑相邻两层的神经元，下一层的神经元的输出h为所有与之相连的上一层神经元x的加权和并经过激活函数，可以表示为：
h＝f(wx+b)。
其中，w为权重矩阵，b为偏置向量，f为激活函数。
进一步可选的，神经网络的输出可以递归表达为：
y＝fn(wnfn-1(…)+bn)。
其中,n是神经网络层的索引，1<＝n<＝N,其中N为神经网络的总层数。
换言之，可以将神经网络理解为一个从输入数据集合到输出数据集合的映射关系。而通常神经网络都是随机初始化的，用已有数据从随机的w和b得到这个映射关系的过程被称为神经网络的训练。
可选的，训练的具体方式为采用损失函数(loss function)对神经网络的输出结果进行评价。
如图2b所示，可以将误差反向传播，通过梯度下降的方法即能迭代优化神经网络参数(包括w和b)，直到损失函数达到最小值，即图2b中的“较优点(例如最优点)”。可以理解的是，图2b中的“较优点(例如最优点)”对应的神经网络参数可以作为训练好的AI模型信息中的神经网络参数。
进一步可选的，梯度下降的过程可以表示为：
其中，θ为待优化参数(包括w和b)，L为损失函数，η为学习率，控制梯度下降的步长，表示求导运算，表示对L求θ的导数。
进一步可选的，反向传播的过程利用到求偏导的链式法则。
如图2c所示，前一层参数的梯度可以由后一层参数的梯度递推计算得到，可以表达为：
其中，wij为节点j连接节点i的权重，si为节点i上的输入加权和。
2.联邦学习(Federated Learning，FL)。
联邦学习这一概念的提出有效地解决了当前人工智能发展所面临的困境，其在充分保障用户数据隐私和安全的前提下，通过促使各个边缘设备和中心端服务器协同合作来高效地完成模型的学习任务。
如图2d所示，FL架构是当前FL领域最为广泛的训练架构，FedAvg算法是FL的基础算法，其算法流程大致如下：
(1)中心端初始化待训练模型并将其广播发送给所有客户端设备。
(2)在第t∈[1,T]轮中，客户端k∈[1,K]基于局部数据集对接收到的全局模型进行E个epoch的训练以得到本地训练结果将其上报给中心节点。
(3)中心节点汇总收集来自全部(或部分)客户端的本地训练结果，假设第t轮上传局部模型的客户端集合为中心端将以对应客户端的样本数为权重进行加权求均得到新的全局模型，具体更新法则为其后中心端再将最新版本的全局模型广播发送给所有客户端设备进行新一轮的训练。
(4)重复步骤(2)和(3)直至模型最终收敛或训练轮数达到上限。
除了上报本地模型还可以将训练的本地梯度进行上报，中心节点将本地梯度求平均，并根据这个平均梯度的方向更新全局模型。
可以看到，在FL框架中，数据集存在于分布式节点处，即分布式节点收集本地的数据集，并进行本地训练，将训练得到的本地结果(模型或梯度)上报给中心节点。中心节点本身没有数据集，只负责将分布式节点的训练结果进行融合处理，得到全局模型，并下发给分布式节点。
3.去中心式学习。与联邦学习不同，另一种分布式学习架构——去中心式学习。
如图2e所示，考虑没有中心节点的完全分布式系统。去中心式学习系统的设计目标f(x)一般是各节点目标fi(x)的均值，即其中n是分布式节点数量，x是待优化参数，在机器学习中，x就是机器学习(如神经网络)模型的参数。各节点利用本地数据和本地目标fi(x)计算本地梯度然后将其发送给通信可达的邻居节点。任一节点收到其邻点发来的梯度信息后，可以按照下式更新本地模型的参数x：
其中，表示第i个节点中第k+1(k为自然数)次更新后的本地模型的参数，表示第i个节点中第k次更新后的本地模型的参数(若k为0，则表示为第i个节点的未参与更新的本地模型的参数)，αk表 示调优系数，Ni是节点i的邻居节点集合，|Ni|表示节点i的邻居节点集合中的元素数量，即节点i的邻居节点数量。通过节点间的信息交互，去中心式学习系统最终将学到一个统一的模型。
本申请提供的技术方案可以应用于无线通信系统(例如图1a或图1b所示系统)，在无线通信系统中，通信节点一般具备信号收发能力和计算能力。以具备计算能力的网络设备为例，网络设备的计算能力主要是为信号收发能力提供算力支持(例如：对信号进行发送处理和接收处理)，以实现网络设备与其它通信节点的通信任务。
在通信网络中，通信节点的计算能力除了为上述通信任务提供算力支持之外，还可能具备富余的计算能力。为此，如何利用这些计算能力，是一个亟待解决的技术问题。
在一种可能的实现方式中，通信节点可以作为AI学习系统的参与节点，将该通信节点的算力应用于AI学习系统的某一个环节。随着大模型时代的到来，拥有海量参数的深度学习模型如基于变换器的双向编码器表示(bidirectional encoder representations from transformers，BERT)，基于变换器的生成式预训练(generative pre-trained transformer，GPT)等能完成越来越复杂的任务，并能达到较好的性能。但是对于大模型来说，即使是模型的推理过程也会收到设备容量的限制，所以一般大模型存放于云端中心服务器上。同时，网络中的每台设备每天会产生巨量的原始数据，这些数据需要多次调用大模型推理。一般来说，可以是设备(例如通信节点)将数据发送给中心服务器，中心服务器使用数据进行推理，然后中心服务器返回推理结果给设备。这个过程将会消耗大量通信资源用于传输数据，同时设备数据的隐私也会承受风险。
为了更好地节省通信开销及保护用户数据的隐私，学者提出了深度神经网络的分布式推理技术。其做法是将模型分发给设备，利用设备本地的算力来推理模型，以此减少通信开销及获得数据的隐私保护。
示例性的，在图2f所示示例中，以节点1和节点2这两个通信节点参与AI学习系统为例。其中，该节点1和节点2均可以为通信节点，例如终端设备或网络设备。其中，AI学习系统所使用的神经网络可以至少包括部署于节点1的用于AI编码的子神经网络，和/或，部署于节点2的用于AI解码的子神经网络。
作为图2f的一种实现示例，节点1基于用于AI编码的子神经网络进行处理得到编码结果之后，该编码结果经过量化、物理层处理之后得到无线信号；相应的，通过无线信道的传输，节点2接收该无线信号之后，节点2经过物理层处理、解量化处理之后作为AI解码的输入，经过AI解码处理能够得到解码结果。并且，该节点2还可以基于该解码结果与标签数据确定梯度数据。
此后，节点2基于AI解码的子神经网络处理得到梯度数据之后，该梯度数据经过量化、物理层处理之后得到无线信号；相应的，通过无线信道的传输，节点1接收该无线信号之后，节点1经过物理层处理、解量化处理之后得到梯度数据，后续该节点1能够基于该梯度数据对部署于节点1中的用于AI编码的子神经网络进行神经网络的优化(例如训练/更新/迭代等)。
可选地，节点2得到梯度数据之后，该节点2也能基于该梯度数据对部署于节点2中的用于AI编码的子神经网络进行神经网络的优化(例如训练/更新/迭代等)。
需要说明的是，该节点2还可以基于该解码结果与标签数据计算损失函数的结果，并且，该损失函数的结果也可以用于神经网络的优化，上述实现仅以节点2确定梯度数据为例进行说明。
此外，神经网络的优化过程可能需要执行多次上述AI编码、AI解码的过程。而在多次执行过程中，节点1可以通过多个输入数据执行AI编码处理后发送，并且，节点2可以通过多个标签数据对AI解码结果进行处理得到梯度数据。在这种情况下，多个输入数据和多个标签数据可以为同一训练数据集中的数据(可参考前文监督学习的实现过程)。然而，对于不同的节点而言，在某次AI编码、AI解码的过程中，如何对齐该次AI编码所使用的输入数据、AI解码后所使用的标签数据在同一训练数据集中的索引，是一个亟待解决的技术问题。
为了解决上述问题，本申请提供了一种通信方法及相关设备，用于使得通信节点的算力能够应用于神经网络的人工智能(artificial intelligence，AI)处理的同时，也能够提升神经网络部署的灵活性。下面将结合附图进行详细介绍。
请参阅图3，为本申请提供的通信方法的一个实现示意图，该方法包括如下步骤。
需要说明的是，在图3中以第一通信装置和第二通信装置作为该交互示意的执行主体为例来示意该 方法，但本申请并不限制该交互示意的执行主体。例如，在图3和后文图6中，方法的执行主体可以替换为通信装置中的芯片、芯片系统、处理器、逻辑模块或软件等。其中，该第一通信装置可以为终端设备且第二通信装置可以为网络设备，或者，第一通信装置可以为网络设备且第二通信装置可以为终端设备，或者，该第一通信装置和第二通信装置均为终端设备(例如该方法可以应用于侧行链路通信场景下不同终端设备的通信过程)。
S301.第一通信装置对第一数据进行第一处理，得到第二数据。其中，该第一数据为第一数据集中的数据。
S302.第一通信装置发送第二数据，相应的，第二通信装置接收该第二数据。
S303.第二通信装置基于第二数据确定第三数据。
应理解，在上述技术方案中，第二数据是基于第一神经网络对第一数据进行处理得到的，和/或，第三数据是基于第二神经网络对第二数据进行处理得到的。可选的，第一数据集和第二数据集可以包含于一个数据集。在该一个数据集中，可以包括N份输入数据和M份标签数据，N和M均为正整数；并且，该第一数据集可以包括该N份输入数据，该第二数据集可以包括该M份标签数据。换言之，该第一数据集可以称为输入数据集，神经网络输入数据集等，该第二数据集可以称为标签数据集，神经网络标签数据集等。此外，AI神经网络可以基于该同一数据集进行处理，以实现该AI神经网络的迭代、更新等。
可选地，在N份输入数据中的每一份输入数据对应于M份标签数据中的不同标签数据的情况下，N的取值与M的取值相等。在N份输入数据中的至少两份输入数据对应于M份标签数据中的其中一个相同标签数据的情况下，N的取值可以大于或等于M的取值。在N份输入数据中的其中一份输入数据对应于M份标签数据中的至少两份标签数据的情况下，N的取值可以小于或等于M的取值。
可选地，在第二数据是基于第一神经网络对第一数据集中的第一数据进行处理得到的情况下，由于第二数据为第一通信装置对第一数据进行处理得到的发送数据，为此，该第一神经网络可以称为部署于发送端的神经网络，编码神经网络，AI编码神经网络等。类似地，在第三数据是基于第二神经网络对第二数据进行处理得到的情况下，由于第三数据为第二通信装置基于第二神经网络对接收的第二数据进行处理得到的数据，为此，该第二神经网络可以称为部署于接收端的神经网络，解码神经网络，AI解码神经网络等。
本申请中，AI、神经网络、AI神经网络、机器学习、AI处理，AI神经网络处理等术语可以相互替换。
本申请中，涉及的数据(例如第一数据、第二数据、第三数据，以及第四数据等)，可以替换为信息、信号等。
此外，在上述技术方案中，第一数据在该第一数据集中的索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据。换言之，第二通信装置在步骤S303中基于第二数据确定第三数据之后，该第二通信装置可以基于第一数据对应的标签数据(即第四数据)与该第三数据，确定相应的梯度信息和/或损失函数的结果。
可选地，在图3所示方法中，在步骤S303之后，该方法还包括：该第二通信装置向该第一通信装置发送基于该第三数据和该第四数据确定的梯度信息和/或损失函数的结果。具体地，第二数据的接收方(例如第二通信装置)能够基于第二数据进行处理得到第三数据，并且，该接收方还能够基于该第三数据和第四数据确定并发送相应的梯度信息和/或损失函数的结果。使得第一通信装置能够在接收梯度信息和/或损失函数的结果之后，能够基于该梯度信息和/或损失函数的结果实现对第一神经网络的更新或迭代。
此外，第一数据在该第一数据集中的索引满足以下方式A和方式B中的至少一项。
方式A.该索引是基于承载该第二数据的资源确定的。
具体地，在第一数据在该第一数据集中的索引满足方式A的情况下，该索引的取值是该资源的时域资源索引、该资源的频域资源索引，该资源的资源块大小中的至少一项确定的。从而，在第一数据在第一数据集中的索引是基于承载该第二数据的资源确定的情况下，该索引具体可以是通过上述至少一项确定的，以提升方案实现的灵活性。
其中，方式A可以理解为采用数据收发双方已同步的信息在两侧生成相同的索引值，从而实现对索 引的理解的对齐。举例如下，假设训练的批大小为Nbatch，数据集样本数量为Ndataset，一轮训练可认为每次采用Nbatch个样本数据，直至将数据集样本用完。下面将以承载该第二数据的资源为时域资源索引为例，提供一些实现示例。
一种实现示例中，利用数据收发双方的时域资源索引中的系统帧号nf来生成索引值。例如，数据集中的数据的索引值满足：
(nf+no)×Nbatch％Ndataset，((nf+no)×Nbatch+1)％Ndataset，…，((nf+no+1)×Nbatch-1)％Ndataset；
其中，％表示取余运算，nf表示系统帧号，no为偏移值(用于遍历数据集)。
另一种实现示例中，数据集中的数据的索引值也可以与时域资源索引中的系统帧号nf以及子帧号nsf相关。例如，数据集中的数据的索引值满足：
(nf×Nsf_f+nsf+no)×Nbatch％Ndataset，((nf×Nsf_f+nsf+no)×Nbatch+1)％Ndataset，…，((nf×Nsf_f+nsf+no+1)×Nbatch-1)％Ndataset；
其中，％表示取余运算，nf表示系统帧号，Nsf_f表示每帧包含的子帧个数，nsf表示子帧号，no为偏移值(用于遍历数据集)。
另一种实现示例中，数据集中的数据的索引值也可以与时域资源索引中的系统帧号nf以及时隙号nslot相关。例如，数据集中的数据的索引值满足：
(nf×Nslot_f+nslot+no)×Nbatch％Ndataset，((nf×Nslot_f+nslot+no)×Nbatch+1)％Ndataset，…，((nf×Nslot_f+nslot+no+1)×Nbatch-1)％Ndataset。
其中，％表示取余运算，nf表示系统帧号，Nslot_f表示每帧包含的时隙个数，nsf表示子帧号，no为偏移值(用于遍历数据集)。
应理解，上述实现示意仅以承载该第二数据的资源为时域资源索引为例，在上述实现示例中，时域资源索引可以替换为其他数据收发双方已同步的信息相关，如承载第二数据的频域的物理资源数量，包含但不限于资源块的数量，子载波的数量等。
可选地，方式A可以理解为一种实时的数据对齐方式，此处的实时可以理解为第一通信装置在步骤S301中执行第一处理的过程以及第二通信装置在步骤S303中执行第二处理的过程之间的时间间隔相对固定；和/或，第一通信装置发送第一处理的处理结果(即第二数据)的过程以及以及第二通信装置发送基于第二处理对应的梯度数据(和/或损失函数的结果)的过程之间的时间间隔相对固定。
图4为方式A(即实时的数据对齐方式)的一种实现示例。在该示例中，每六个帧中的第一个帧(例如帧号为1/7/13的帧)用于传输第一通信装置发送的第二数据，每六个帧中的第四个帧(例如帧号为4/10/16的帧)用于传输第二通信装置发送的梯度数据(和/或损失函数的结果)。换言之，承载第二数据的时域资源与承载梯度数据(和/或损失函数的结果)的时域资源之间的时间间隔可以是预配置的。
可以理解的是，在图4中，第一通信装置和第二通信装置之间除了交互第二数据和梯度数据(和/或损失函数的结果)之外，还可以交互其它的数据，例如图4所示的其他通信信号，例如系统信息、参考信号、基于参考信号进行测量得到的信道信息等。
方式B.该索引是基于该第一数据集中的数据的处理次数确定的。
在第一数据在该第一数据集中的索引满足方式B的情况下，该方法还包括：该第一通信装置发送第一信息，该第一信息用于指示该第一数据在该第一数据集中的索引。
具体地，在方式B中，由于第二通信装置可能无法感知第一数据集中的数据处理次数，为此，第二通信装置可以基于第二数据集中的数据的处理次数确定该索引。相应的，第一通信装置还可以发送第一信息，使得第二通信装置能够基于该第一信息确定该第一数据在该第一数据集中的索引，后续基于该索引在第二数据集中确定第四数据。
应理解，第一通信装置可以基于第一数据集中的数据执行多次处理得到并发送处理结果(例如，其中一个处理结果为基于第一数据得到的第二数据)的过程，相应的，在该多个处理结果中，第一通信装置可以针对部分或全部处理结果发送对应的索引(例如，针对第二数据这个处理结果发送第一信息)。此后，基于该部分或全部处理结果发送对应的索引能够实现数据收发双方对索引的理解的对齐，以避免数据收发双方对索引的理解不对齐导致的数据处理出错，进而提升系统的鲁棒性。
可选地，第一通信装置可以针对部分处理结果发送该部分处理结果对应的索引，而无需针对全部处理结果发送该全部处理结果对应的索引，能够节省开销。在方式B的一种可能的实现方式中，该第一信息为基于第一周期传输的多个信息中的其中一个信息；该方法还包括：该第一通信装置接收或发送配置信息，该配置信息用于配置该第一周期。具体地，第一信息可以是基于第一周期传输的周期性信息中的其中一个，在此之前，第一通信装置可以接收或发送用于配置该第一周期的配置信息，使得该第一通信装置既可以作为该第一周期的配置方，也可以作为第一周期的被配置方，能够使得数据收发双方对第一周期的理解的对齐的同时，也能够提升方案实现的灵活性。
例如，第一通信装置可以设置样本索引计数器，该样本索引计数器用于对基于第一数据集中的数据进行第一处理的次数而累加，并基于该累加的值在步骤S301中确定第一数据在第一数据集中的索引；相应的，第二通信装置可以设置样本索引计数器，该样本索引计数器用于对基于第二数据集中的数据进行第二处理的次数而累加，并基于该累加的值在步骤S302中接收第二数据之后，确定用于生成该第二数据的第一数据在第一数据集中的索引(或确定该步骤S303之后所使用的第四数据在第二数据集中的索引)。
此外，第一通信装置可以设置第一周期的定时器，在该定时器超期时，该第一通信装置在步骤S302中发送第二数据的时候，还可以在该步骤S302中发送第一信息，以便于第一通信装置和第二通信装置通过该第一信息对两者所适用的索引进行同步，以防止两者由于样本索引计数器不对齐导致的失步。
示例性的，该配置信息可以承载于RRC消息，例如，该配置信息可以通过RRC消息中的数据同步周期(DataSyncPeriod)信元，实现该第一周期的配置。
可选地，方式B可以理解为一种非实时系统中的数据同步方式。此处的非实时可以理解为第一通信装置在步骤S301中执行第一处理的过程以及第二通信装置在步骤S303中执行第二处理的过程之间的时间间隔并非是相对固定的，和/或，第一通信装置发送第一处理的处理结果(即第二数据)的过程以及以及第二通信装置发送基于第二处理对应的梯度数据(和/或损失函数的结果)的过程之间的时间间隔并非是相对固定的。
图5为方式B(即非实时的数据对齐方式)的一种实现示例。在该示例中，第一通信装置和第二通信装置之间可以执行多个AI任务，并且，不同的AI任务的执行周期或者不同AI任务的数据收发的触发可能是不同的。例如，不同的AI任务的输入数据的规模大小可能是不同的。又如，不同的AI任务的梯度数据(和/或损失函数的结果)的规模大小可能是不同的。
在图5所示示例中，一个AI任务涉及的数据可以包括帧号为1传输的第二数据以及帧号为4传输的梯度数据(和/或损失函数的结果)，即两者间隔为2个帧(即帧号为2和3的帧)；另一个AI任务涉及的数据可以包括帧号为5传输的第二数据以及帧号为10传输的梯度数据(和/或损失函数的结果)，即两者间隔为4个帧(即帧号为6、7、8和9的帧)；另一个AI任务涉及的数据可以包括帧号为17传输的第二数据以及帧号为18传输的梯度数据(和/或损失函数的结果)，即两者间隔为0个帧(即两者为相邻的两个帧)。
方式C，该索引是基于承载该第二数据的资源以及该第一数据集中的数据的处理次数确定的。
在第一数据在该第一数据集中的索引满足方式A和方式B的情况下，该索引可以是基于承载该第二数据的资源(例如该资源的时域资源索引、频域资源索引，资源块大小中的至少一项)以及该第一数据集中的数据的处理次数确定的。
例如，该索引的取值可以为该资源的时域资源索引的数值与该处理次数的数值的数学运算结果(例如两个数值的和、两个数值的差、两个数值的乘积等)。又如，该索引的取值可以为该资源的资源块大小的数值与该处理次数的数值的数学运算结果(例如两个数值的和、两个数值的差、两个数值的乘积等)。
在一种可能的实现方式中，图3所示方法还包括：该第一通信装置接收或发送指示该索引满足该至少一项的指示信息(即该指示信息用于指示方式A和/或方式B，或，该指示信息用于指示方式A或方式B或方式C)。具体地，第一通信装置还可以接收或发送指示该索引满足该至少一项的指示信息，使得数据收发双方能够基于该指示信息实现数据集中的数据的索引的理解的对齐，以避免数据收发双方对索引的理解不对齐导致的数据处理出错，进而提升系统的鲁棒性。
在一种可能的实现方式中，图3所示方法还包括：该第一通信装置接收指示该第一数据集的指示信 息；和/或，该第一通信装置发送指示该第二数据集的指示信息。具体地，第一通信装置可以作为第一数据集的被配置方，和/或，第一通信装置可以作为第二数据集的配置方，使得数据收发双方能够在交互数据之前实现数据集的获取。
可选地，第一数据集可以是预配置于该第一通信装置的，和/或，第二数据集可以是预配置于该第二通信装置的，通过这种方式，能够降低开销。
基于图3所示技术方案，第一通信装置在步骤S301中发送的第二数据是基于第一数据得到的，后续第二通信装置可以在步骤S302中对第二数据进行处理得到第三数据。其中，该第一数据在该第一数据集中的索引用于在第二数据集中确定第四数据，该第四数据是该第一数据对应的标签数据。换言之，第二数据的接收方在步骤S302中接收第二数据，并在步骤S303中确定第三数据之后，该第二通信装置能够在第二数据集中确定标签数据，并基于该标签数据对第三数据进行处理。并且，第二数据和/或该第三数据是基于神经网络得到的，即用于AI处理的神经网络可以包括部署于第一通信装置的神经网络和/或部署于第二通信装置的神经网络。从而，在通信系统中的通信装置作为AI参与节点的情况下，使得通信装置的算力能够应用于神经网络的AI处理的同时，也能够提升神经网络部署的灵活性。
此外，第一数据在该第一数据集中的索引用于在第二数据集中确定该第一数据对应的标签数据(即第四数据)，并且，该索引满足上述至少一项。换言之，第二通信装置在接收第二数据之后，能够基于承载数据的资源或数据集中的数据的处理次数在第二数据集中确定标签数据，通过这种方式，可以降低空口开销，以提升通信效率。
请参阅图6，为本申请提供的通信方法的一个实现示意图，该方法包括如下步骤。
S601.第一通信装置对第一数据进行第一处理，得到第二数据。其中，该第一数据为第一数据集中的数据。
S602.第一通信装置发送第二数据和第四数据，相应的，第二通信装置接收该第二数据和第四数据。
S603.第二通信装置基于第二数据确定第三数据。
在一种可能的实现方式中，图6所示方法还包括：该第一通信装置接收基于该第三数据和该第四数据确定的梯度信息和/或损失函数的结果。具体地，第二数据的接收方(例如第二通信装置)能够基于第二数据进行处理得到第三数据，并且，该接收方还能够基于该第三数据和第四数据确定并发送相应的梯度信息和/或损失函数的结果。使得第一通信装置能够在接收梯度信息和/或损失函数的结果之后，能够基于该梯度信息和/或损失函数的结果实现对第一神经网络的更新或迭代。
需要说明的是，图6所示技术方案中，第一通信装置和第二通信装置的实现过程(例如第一数据至第四数据、第一处理和第二数据等)可以参考前文图3及相关实现方式。
基于图6所示技术方案，第一通信装置在步骤S602中发送的第二数据是基于第一数据得到的，后续第二通信装置可以在步骤S603中对第二数据进行处理得到第三数据。其中，第一通信装置还可以在步骤S602中发送第四数据，该第四数据是该第一数据对应的标签数据。相应的，第二通信装置在步骤S602中接收第四数据之后，能够基于该第四数据(即该第一数据对应的标签数据)对第三数据进行处理。并且，第二数据和/或该第三数据是基于神经网络得到的，即用于AI处理的神经网络可以包括部署于第一通信装置的神经网络和/或部署于第二通信装置的神经网络。从而，在通信系统中的通信装置作为AI参与节点的情况下，使得通信装置的算力能够应用于神经网络的AI处理的同时，也能够提升神经网络部署的灵活性。
此外，第一通信装置发送的第二数据为第一数据的处理结果，该第一通信装置发送的第四数据为该第一数据对应的标签数据，从而，通过发送第一数据的处理结果和第一数据对应的标签数据的方式，使得接收方能够基于该标签数据进一步处理的同时，也能够使得方案能够适用于标签数据的数据量较小的场景，以尽可能地减少标签数据的传输开销。
应理解，在前述图3所示实现过程中，第一通信装置可以通过配置的方式或者预配置的方式获取第一数据集，第二通信装置可以通过配置的方式或者预配置的方式获取第二数据集，并且，第一通信装置和第二通信装置都需要通过承载第二数据的资源或数据集中的数据的处理次数实现索引的对齐。而在图6所示技术方案中，不同的地方是，第一数据集和第二数据集可以部署于第一通信装置，而第二通信装 置可以无需部署数据集，并且，数据集的索引也仅在第一通信装置侧维护，优点是无需在两侧维护同步的索引值，能够降低第二通信装置的开销以及实现复杂度。
可选地，在图6所示技术方案中，由于在步骤S602中传输的第四数据为标签数据，为此，图6所示技术方案可以适用于标签数据的数据规模较小的场景。
请参阅图7，为本申请提供的通信方法的一个实现示意图，该方法包括如下步骤。
S701.第一通信装置对第一数据进行第一处理，得到第二数据。其中，该第一数据为第一数据集中的数据。
S702.第一通信装置发送第二数据和第一索引，相应的，第二通信装置接收该第二数据和第一索引。
S703.第二通信装置基于第二数据确定第三数据。
应理解，在上述技术方案中，第二数据是基于第一神经网络对第一数据进行处理得到的，和/或，第三数据是基于第二神经网络对第二数据进行处理得到的。其中，第一数据集和第二数据集可以包含于一个数据集。在该一个数据集中，可以包括N份输入数据和M份标签数据，N和M均为正整数；并且，该第一数据集可以包括该N份输入数据，该第二数据集可以包括该M份标签数据。换言之，该第一数据集可以称为输入数据集，神经网络输入数据集等，该第二数据集可以称为标签数据集，神经网络标签数据集等。此外，AI神经网络可以基于该同一数据集进行处理，以实现该AI神经网络的迭代、更新等。
在一种可能的实现方式中，该第一索引是该第一数据在该第一数据集中的第二索引确定的。
一种实现示例中，在N份输入数据中的每一份输入数据对应于M份标签数据中的不同标签数据的情况下，N的取值与M的取值相等。
例如，第一索引可以与第一数据在该第一数据集中的第二索引相同。换言之，第一数据集中的第i份输入数据的标签数据为第二数据集中的第j份数据，i为第一索引，j为第二索引，并且，i与j相等。即N份输入数据的第一份数据的标签数据为M份标签数据中的第一份数据，N份输入数据的第二份数据的标签数据为M份标签数据中的第二份数据，以此类推，N份输入数据的第N份数据的标签数据为M份标签数据中的第M份数据(N与M相等)。
又如，第一索引可以与第一数据在该第一数据集中的第二索引部分不相同或完全不相同。换言之，第一数据集中的第i份输入数据的标签数据为第二数据集中的第j份数据，i为第一索引，j为第二索引，并且，i与j可以是部分或全部不相等。其中，i与j之间的映射关系可以是预配置的。
作为i与j之间的映射关系可以是预配置的一种示例，i可以是从1遍历至N，j可以是从N(N等于M)遍历至1。例如，以N和M的取值均为3为例，N份输入数据的第一份数据的标签数据为M份标签数据中的第三份数据，N份输入数据的第二份数据的标签数据为M份标签数据中的第二份数据，N份输入数据的第三份数据的标签数据为M份标签数据中的第一份数据。在该示例中，第一索引可以与第一数据在该第一数据集中的第二索引部分不相同。
作为i与j之间的映射关系可以是预配置的另一种示例，i和j之间的映射关系可以通过配置的方式或预配置的方式使得数据收发双方对齐。例如，仍以N和M的取值均为3为例，N份输入数据的第一份数据的标签数据为M份标签数据中的第三份数据，N份输入数据的第二份数据的标签数据为M份标签数据中的第一份数据，N份输入数据的第三份数据的标签数据为M份标签数据中的第二份数据。在该示例中，第一索引可以与第一数据在该第一数据集中的第二索引完全不相同。
另一种实现示例中，该第一索引与第一数据在该第一数据集中的第二索引之间的映射关系是预配置的。
例如，在N份输入数据中的至少两份输入数据对应于M份标签数据中的其中一个相同标签数据的情况下，N的取值可以大于或等于M的取值。相应的，在这种情况下，第一索引可以小于或等于第一数据在该第一数据集中的第二索引。
又如，在N份输入数据中的其中一份输入数据对应于M份标签数据中的至少两份标签数据的情况下，N的取值可以小于或等于M的取值。相应的，在这种情况下，第一索引可以大于或等于第一数据在该第一数据集中的第二索引。
在一种可能的实现方式中，图7所示方法还包括：该第一通信装置接收基于该第三数据和该第四数 据确定的梯度信息和/或损失函数的结果。具体地，第二数据的接收方(例如第二通信装置)能够基于第二数据进行处理得到第三数据，并且，该接收方还能够基于该第三数据和第四数据确定并发送相应的梯度信息和/或损失函数的结果。使得第一通信装置能够在接收梯度信息和/或损失函数的结果之后，能够基于该梯度信息和/或损失函数的结果实现对第一神经网络的更新或迭代。
基于图7所示技术方案，第一通信装置在步骤S702中发送的第二数据是基于第一数据得到的，后续该第二通信装置在步骤S703可以对第二数据进行处理得到第三数据。其中，第一通信装置还可以发送第一索引。相应的，第二数据和第四数据的接收方在接收第二数据之后，能够基于该标签数据对第三数据进行处理。并且，第二数据和/或该第三数据是基于神经网络得到的，即用于AI处理的神经网络可以包括部署于第一通信装置的神经网络和/或部署于第二通信装置的神经网络。从而，在通信系统中的通信装置作为AI参与节点的情况下，使得通信装置的算力能够应用于神经网络的AI处理的同时，也能够提升神经网络部署的灵活性。
此外，第一通信装置发送的第二数据为第一数据的处理结果，该第一通信装置发送的第四数据为该第一数据对应的标签数据，从而，通过发送第一数据的处理结果和第一数据对应的标签数据的方式，使得接收方能够基于该标签数据进一步处理的同时，也能够使得方案能够适用于标签数据的数据量较小的场景，以尽可能地减少标签数据的传输开销。
应理解，在前述图3所示实现过程中，第一通信装置可以通过配置的方式或者预配置的方式获取第一数据集，第二通信装置可以通过配置的方式或者预配置的方式获取第二数据集，并且，第一通信装置和第二通信装置都需要通过承载第二数据的资源或数据集中的数据的处理次数实现索引的对齐。而在图7所示技术方案中，不同的地方是，第一通信装置和第二通信装置可以无需设置用于同步的样本索引计数器，能够降低实现复杂度。
可选地，在图7所示技术方案中，由于在步骤S702中传输的数据包括第二索引，为此，图7所示技术方案可以适用于数据集规模较小(或数据集中的索引的数量较少)的场景。
请参阅图8，本申请实施例提供了一种通信装置800，该通信装置800可以实现上述方法实施例中第二通信装置或第一通信装置的功能，因此也能实现上述方法实施例所具备的有益效果。在本申请实施例中，该通信装置800可以是第一通信装置(或第二通信装置)，也可以是第一通信装置(或第二通信装置)内部的集成电路或者元件等，例如芯片。
需要说明的是，收发单元802可以包括发送单元和接收单元，分别用于执行发送和接收。
一种可能的实现方式中，当该装置800为用于执行前述实施例中第一通信装置所执行的方法时，该装置800包括处理单元801和收发单元802；该处理单元801用于对第一数据进行处理，得到第二数据；其中，该第一数据为第一数据集中的数据；该收发单元802用于发送该第二数据，该第二数据用于确定第三数据；其中，该第一数据在该第一数据集中的索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的；该索引满足以下至少一项：该索引是基于承载该第二数据的资源确定的；该索引是基于该第一数据集中的数据的处理次数确定的。
一种可能的实现方式中，当该装置800为用于执行前述实施例中第二通信装置所执行的方法时，该装置800包括处理单元801和收发单元802；该收发单元802用于接收第二数据，该第二数据是基于第一数据进行处理得到的，该第一数据为第一数据集中的数据；其中，该第一数据在该第一数据集中的索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；该处理单元801用于基于该第二数据确定第三数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的；该索引满足以下至少一项：该索引是基于承载该第二数据的资源确定的；该索引是基于该第二数据集中的数据的处理次数确定的。
一种可能的实现方式中，当该装置800为用于执行前述实施例中第一通信装置所执行的方法时，该装置800包括处理单元801和收发单元802；该处理单元801用于对第一数据进行处理，得到第二数据；该 收发单元802用于发送该第二数据和第四数据，其中，该第二数据用于确定第三数据，该第四数据为该第一数据对应的标签数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的。
一种可能的实现方式中，当该装置800为用于执行前述实施例中第二通信装置所执行的方法时，该装置800包括处理单元801和收发单元802；该收发单元802用于接收第二数据和第四数据，其中，该第二数据是基于第一数据得到的，该第四数据为该第一数据对应的标签数据；该处理单元801用于基于该第二数据确定第三数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的。
一种可能的实现方式中，当该装置800为用于执行前述实施例中第一通信装置所执行的方法时，该装置800包括处理单元801和收发单元802；该处理单元801用于对第一数据进行处理，得到第二数据；其中，该第一数据为第一数据集中的数据；该收发单元802用于发送该第二数据和第一索引，该第二数据用于确定第三数据；其中，该第一索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第四数据是基于神经网络对该第二数据进行处理得到的。
一种可能的实现方式中，当该装置800为用于执行前述实施例中第二通信装置所执行的方法时，该装置800包括处理单元801和收发单元802；该收发单元802用于接收第二数据和第一索引，该第二数据是基于第一数据得到的，该第一数据为第一数据集中的数据；其中，该第一索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；该处理单元801用于基于该第二数据确定第三数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的。
需要说明的是，上述通信装置800的单元的信息执行过程等内容，具体可参见本申请前述所示的方法实施例中的叙述，此处不再赘述。
请参阅图9，为本申请提供的通信装置900的另一种示意性结构图，通信装置900包括逻辑电路901和输入输出接口902。其中，通信装置900可以为芯片或集成电路。
其中，图8所示收发单元802可以为通信接口，该通信接口可以是图9中的输入输出接口902，该输入输出接口902可以包括输入接口和输出接口。或者，该通信接口也可以是收发电路，该收发电路可以包括输入接口电路和输出接口电路。
可选地，该逻辑电路901用于对第一数据进行处理，得到第二数据；其中，该第一数据为第一数据集中的数据；该输入输出接口902用于发送该第二数据，该第二数据用于确定第三数据；其中，该第一数据在该第一数据集中的索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的；该索引满足以下至少一项：该索引是基于承载该第二数据的资源确定的；该索引是基于该第一数据集中的数据的处理次数确定的。
可选地，该输入输出接口902用于接收第二数据，该第二数据是基于第一数据进行处理得到的，该第一数据为第一数据集中的数据；其中，该第一数据在该第一数据集中的索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；该逻辑电路901用于基于该第二数据确定第三数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的；该索引满足以下至少一项：该索引是基于承载该第二数据的资源确定的；该索引是基于该第二数据集中的数据的处理次数确定的。
可选地，该逻辑电路901用于对第一数据进行处理，得到第二数据；该输入输出接口902用于发送该第二数据和第四数据，其中，该第二数据用于确定第三数据，该第四数据为该第一数据对应的标签数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的。
可选地，该输入输出接口902用于接收第二数据和第四数据，其中，该第二数据是基于第一数据得到的，该第四数据为该第一数据对应的标签数据；该逻辑电路901用于基于该第二数据确定第三数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的。
可选地，该逻辑电路901用于对第一数据进行处理，得到第二数据；其中，该第一数据为第一数据集中的数据；该输入输出接口902用于发送该第二数据和第一索引，该第二数据用于确定第三数据；其中，该第一索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第四数据是基于神经网络对该第二数据进行处理得到的。
可选地，该输入输出接口902用于接收第二数据和第一索引，该第二数据是基于第一数据得到的，该第一数据为第一数据集中的数据；其中，该第一索引用于在第二数据集中确定第四数据，该第二数据集包括该第一数据集中的数据对应的标签数据，该第四数据是该第一数据对应的标签数据；该逻辑电路901用于基于该第二数据确定第三数据；其中，该第二数据是基于第一神经网络对该第一数据进行处理得到的，和/或，该第三数据是基于第二神经网络对该第二数据进行处理得到的。
其中，逻辑电路901和输入输出接口902还可以执行任一实施例中第一通信装置或第二通信装置执行的其他步骤并实现对应的有益效果，此处不再赘述。
在一种可能的实现方式中，图8所示处理单元801可以为图9中的逻辑电路901。
可选的，逻辑电路901可以是一个处理装置，处理装置的功能可以部分或全部通过软件实现。其中，处理装置的功能可以部分或全部通过软件实现。
可选的，处理装置可以包括存储器和处理器，其中，存储器用于存储计算机程序，处理器读取并执行存储器中存储的计算机程序，以执行任意一个方法实施例中的相应处理和/或步骤。
可选地，处理装置可以仅包括处理器。用于存储计算机程序的存储器位于处理装置之外，处理器通过电路/电线与存储器连接，以读取并执行存储器中存储的计算机程序。其中，存储器和处理器可以集成在一起，或者也可以是物理上互相独立的。
可选地，该处理装置可以是一个或多个芯片，或一个或多个集成电路。例如，处理装置可以是一个或多个现场可编程门阵列(field-programmable gate array，FPGA)、专用集成芯片(application specific integrated circuit，ASIC)、系统芯片(system on chip，SoC)、中央处理器(central processor unit，CPU)、网络处理器(network processor，NP)、数字信号处理电路(digital signal processor，DSP)、微控制器(micro controller unit，MCU)，可编程控制器(programmable logic device，PLD)或其它集成芯片，或者上述芯片或者处理器的任意组合等。
请参阅图10，为本申请的实施例提供的上述实施例中所涉及的通信装置1000，该通信装置1000具体可以为上述实施例中的作为终端设备的通信装置，图10所示示例为终端设备通过终端设备(或者终端设备中的部件)实现。
其中，该通信装置1000的一种可能的逻辑结构示意图，该通信装置1000可以包括但不限于至少一个处理器1001以及通信端口1002。
其中，图8所示收发单元802可以为通信接口，该通信接口可以是图10中的通信端口1002，该通信端口1002可以包括输入接口和输出接口。或者，该通信端口1002也可以是收发电路，该收发电路可以包括输入接口电路和输出接口电路。
进一步可选的，该装置还可以包括存储器1003、总线1004中的至少一个，在本申请的实施例中，该至少一个处理器1001用于对通信装置1000的动作进行控制处理。
此外，处理器1001可以是中央处理器单元，通用处理器，数字信号处理器，专用集成电路，现场可编程门阵列或者其他可编程逻辑器件、晶体管逻辑器件、硬件部件或者其任意组合。其可以实现或执行结合本申请公开内容所描述的各种示例性的逻辑方框，模块和电路。该处理器也可以是实现计算功能的组合，例如包含一个或多个微处理器组合，数字信号处理器和微处理器的组合等等。所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统，装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
需要说明的是，图10所示通信装置1000具体可以用于实现前述方法实施例中终端设备所实现的步骤，并实现终端设备对应的技术效果，图10所示通信装置的具体实现方式，均可以参考前述方法实施例中的叙述，此处不再一一赘述。
请参阅图11，为本申请的实施例提供的上述实施例中所涉及的通信装置1100的结构示意图，该通信装置1100具体可以为上述实施例中的作为网络设备的通信装置，图11所示示例为网络设备通过网络设备(或者网络设备中的部件)实现，其中，该通信装置的结构可以参考图11所示的结构。
通信装置1100包括至少一个处理器1111以及至少一个网络接口1114。进一步可选的，该通信装置还包括至少一个存储器1112、至少一个收发器1113和一个或多个天线1115。处理器1111、存储器1112、收发器1113和网络接口1114相连，例如通过总线相连，在本申请实施例中，该连接可包括各类接口、传输线或总线等，本实施例对此不做限定。天线1115与收发器1113相连。网络接口1114用于使得通信装置通过通信链路，与其它通信设备通信。例如网络接口1114可以包括通信装置与核心网设备之间的网络接口，例如S1接口，网络接口可以包括通信装置和其他通信装置(例如其他网络设备或者核心网设备)之间的网络接口，例如X2或者Xn接口。
其中，图8所示收发单元802可以为通信接口，该通信接口可以是图11中的网络接口1114，该网络接口1114可以包括输入接口和输出接口。或者，该网络接口1114也可以是收发电路，该收发电路可以包括输入接口电路和输出接口电路。
处理器1111主要用于对通信协议以及通信数据进行处理，以及对整个通信装置进行控制，执行软件程序，处理软件程序的数据，例如用于支持通信装置执行实施例中所描述的动作。通信装置可以包括基带处理器和中央处理器，基带处理器主要用于对通信协议以及通信数据进行处理，中央处理器主要用于对整个终端设备进行控制，执行软件程序，处理软件程序的数据。图11中的处理器1111可以集成基带处理器和中央处理器的功能，本领域技术人员可以理解，基带处理器和中央处理器也可以是各自独立的处理器，通过总线等技术互联。本领域技术人员可以理解，终端设备可以包括多个基带处理器以适应不同的网络制式，终端设备可以包括多个中央处理器以增强其处理能力，终端设备的各个部件可以通过各种总线连接。该基带处理器也可以表述为基带处理电路或者基带处理芯片。该中央处理器也可以表述为中央处理电路或者中央处理芯片。对通信协议以及通信数据进行处理的功能可以内置在处理器中，也可以以软件程序的形式存储在存储器中，由处理器执行软件程序以实现基带处理功能。
存储器主要用于存储软件程序和数据。存储器1112可以是独立存在，与处理器1111相连。可选的，存储器1112可以和处理器1111集成在一起，例如集成在一个芯片之内。其中，存储器1112能够存储执行本申请实施例的技术方案的程序代码，并由处理器1111来控制执行，被执行的各类计算机程序代码也可被视为是处理器1111的驱动程序。
图11仅示出了一个存储器和一个处理器。在实际的终端设备中，可以存在多个处理器和多个存储器。存储器也可以称为存储介质或者存储设备等。存储器可以为与处理器处于同一芯片上的存储元件，即片内存储元件，或者为独立的存储元件，本申请实施例对此不做限定。
收发器1113可以用于支持通信装置与终端之间射频信号的接收或者发送，收发器1113可以与天线1115相连。收发器1113包括发射机Tx和接收机Rx。具体地，一个或多个天线1115可以接收射频信号，该收发器1113的接收机Rx用于从天线接收该射频信号，并将射频信号转换为数字基带信号或数字中频信号，并将该数字基带信号或数字中频信号提供给该处理器1111，以便处理器1111对该数字基带信号或数字中频信号做进一步的处理，例如解调处理和译码处理。此外，收发器1113中的发射机Tx还用于从处理器1111接收经过调制的数字基带信号或数字中频信号，并将该经过调制的数字基带信号或数字中频信号转换为射频信号，并通过一个或多个天线1115发送该射频信号。具体地，接收机Rx可以选择性地对射频信号进行一级或多级下混频处理和模数转换处理以得到数字基带信号或数字中频信号，该下混频处理和模数转换处理的先后顺序是可调整的。发射机Tx可以选择性地对经过调制的数字基带信号或数字中频信号时进行一级或多级上混频处理和数模转换处理以得到射频信号，该上混频处理和数模转换处理的先后顺序是可调整的。数字基带信号和数字中频信号可以统称为数字信号。
收发器1113也可以称为收发单元、收发机、收发装置等。可选的，可以将收发单元中用于实现接收功能的器件视为接收单元，将收发单元中用于实现发送功能的器件视为发送单元，即收发单元包括接收 单元和发送单元，接收单元也可以称为接收机、输入口、接收电路等，发送单元可以称为发射机、发射器或者发射电路等。
需要说明的是，图11所示通信装置1100具体可以用于实现前述方法实施例中网络设备所实现的步骤，并实现网络设备对应的技术效果，图11所示通信装置1100的具体实现方式，均可以参考前述方法实施例中的叙述，此处不再一一赘述。
请参阅图12，为本申请的实施例提供的上述实施例中所涉及的通信装置的结构示意图。
可以理解的是，通信装置120包括例如模块、单元、元件、电路、或接口等，以适当地配置在一起以执行本申请提供的技术方案。所述通信装置120可以是前文描述的终端设备或网络设备，也可以是这些设备中的部件(例如芯片)，用以实现下述方法实施例中描述的方法。通信装置120包括一个或多个处理器121。所述处理器121可以是通用处理器或者专用处理器等。例如可以是基带处理器、或中央处理器。基带处理器可以用于对通信协议以及通信数据进行处理，中央处理器可以用于对通信装置(如，RAN节点、终端、或芯片等)进行控制，执行软件程序，处理软件程序的数据。
可选的，在一种设计中，处理器121可以包括程序123(有时也可以称为代码或指令)，所述程序123可以在所述处理器121上被运行，使得所述通信装置120执行下述实施例中描述的方法。在又一种可能的设计中，通信装置120包括电路(图12未示出)。
可选的，所述通信装置120中可以包括一个或多个存储器122，其上存有程序124(有时也可以称为代码或指令)，所述程序124可在所述处理器121上被运行，使得所述通信装置120执行上述方法实施例中描述的方法。
可选的，所述处理器121和/或存储器122中可以包括AI模块127,128，所述AI模块用于实现AI相关的功能。所述AI模块可以是通过软件，硬件，或软硬结合的方式实现。例如，AI模块可以包括无线智能控制(radio intelligence control，RIC)模块。例如AI模块可以是近实时RIC或者非实时RIC。
可选的，所述处理器121和/或存储器122中还可以存储有数据。所述处理器和存储器可以单独设置，也可以集成在一起。
可选的，所述通信装置120还可以包括收发器125和/或天线126。所述处理器121有时也可以称为处理单元，对通信装置(例如RAN节点或终端)进行控制。所述收发器125有时也可以称为收发单元、收发机、收发电路、或者收发器等，用于通过天线126实现通信装置的收发功能。
其中，图8所示收发单元802可以为通信接口，该通信接口可以是图12中的收发器125，该收发器125可以包括输入接口和输出接口。或者，该收发器125也可以是收发电路，该收发电路可以包括输入接口电路和输出接口电路。
本申请实施例还提供一种计算机可读存储介质，该存储介质用于存储一个或多个计算机执行指令，当计算机执行指令被处理器执行时，该处理器执行如前述实施例中第一通信装置或第二通信装置可能的实现方式所述的方法。
本申请实施例还提供一种计算机程序产品(或称计算机程序)，当计算机程序产品被该处理器执行时，该处理器执行上述第一通信装置或第二通信装置可能实现方式的方法。
本申请实施例还提供了一种芯片系统，该芯片系统包括至少一个处理器，用于支持通信装置实现上述通信装置可能的实现方式中所涉及的功能。可选的，所述芯片系统还包括接口电路，所述接口电路为所述至少一个处理器提供程序指令和/或数据。在一种可能的设计中，该芯片系统还可以包括存储器，存储器，用于保存该通信装置必要的程序指令和数据。该芯片系统，可以由芯片构成，也可以包含芯片和其他分立器件，其中，该通信装置具体可以为前述方法实施例中第一通信装置或第二通信装置。
本申请实施例还提供了一种通信系统，该网络系统架构包括上述任一实施例中的第一通信装置和第二通信装置。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统，装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。所述集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说做出贡献的部分或者该技术方案的全部或部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。

Claims (27)

  1. 一种通信方法，其特征在于，包括：
    对第一数据进行处理，得到第二数据；其中，所述第一数据为第一数据集中的数据；
    发送所述第二数据，所述第二数据用于确定第三数据；其中，所述第一数据在所述第一数据集中的索引用于在第二数据集中确定第四数据，所述第二数据集包括所述第一数据集中的数据对应的标签数据，所述第四数据是所述第一数据对应的标签数据；其中，所述第二数据是基于第一神经网络对所述第一数据进行处理得到的，和/或，所述第三数据是基于第二神经网络对所述第二数据进行处理得到的；
    所述索引满足以下至少一项：
    所述索引是基于承载所述第二数据的资源确定的；
    所述索引是基于所述第一数据集中的数据的处理次数确定的。
  2. 根据权利要求1所述的方法，其特征在于，所述索引是基于承载所述第二数据的资源确定的，包括：
    所述索引的取值是所述资源的时域资源索引、所述资源的频域资源索引，所述资源的资源块大小中的至少一项确定的。
  3. 根据权利要求1或2所述的方法，其特征在于，所述索引是基于所述第一数据集中的数据的处理次数确定的，所述方法还包括：
    发送第一信息，所述第一信息用于指示所述第一数据在所述第一数据集中的索引。
  4. 根据权利要求3所述的方法，其特征在于，所述第一信息为基于第一周期传输的多个信息中的其中一个信息；所述方法还包括：
    接收或发送配置信息，所述配置信息用于配置所述第一周期。
  5. 根据权利要求1至4任一项所述的方法，其特征在于，所述方法还包括：
    接收指示所述第一数据集的指示信息；
    和/或，
    发送指示所述第二数据集的指示信息。
  6. 根据权利要求1至5任一项所述的方法，其特征在于，所述方法还包括：
    接收基于所述第三数据和所述第四数据确定的梯度信息和/或损失函数的结果。
  7. 根据权利要求1至6任一项所述的方法，其特征在于，所述方法还包括：
    接收或发送指示所述索引满足所述至少一项的指示信息。
  8. 一种通信方法，其特征在于，包括：
    接收第二数据，所述第二数据是基于第一数据进行处理得到的，所述第一数据为第一数据集中的数据；其中，所述第一数据在所述第一数据集中的索引用于在第二数据集中确定第四数据，所述第二数据集包括所述第一数据集中的数据对应的标签数据，所述第四数据是所述第一数据对应的标签数据；
    基于所述第二数据确定第三数据；其中，所述第二数据是基于第一神经网络对所述第一数据进行处理得到的，和/或，所述第三数据是基于第二神经网络对所述第二数据进行处理得到的；
    所述索引满足以下至少一项：
    所述索引是基于承载所述第二数据的资源确定的；
    所述索引是基于所述第二数据集中的数据的处理次数确定的。
  9. 根据权利要求8所述的方法，其特征在于，所述索引是基于承载所述第二数据的资源确定的，包括：
    所述索引的取值是所述资源的时域资源索引、所述资源的频域资源索引，所述资源的资源块大小中的至少一项确定的。
  10. 根据权利要求8或9所述的方法，其特征在于，所述索引是基于所述第二数据集中的数据的处理次数确定的，所述方法还包括：
    发送第一信息，所述第一信息用于指示所述第一数据在所述第一数据集中的索引。
  11. 根据权利要求10所述的方法，其特征在于，所述第一信息为基于第一周期传输的多个信息中的其中一个信息；所述方法还包括：
    接收或发送配置信息，所述配置信息用于配置所述第一周期。
  12. 根据权利要求8至11任一项所述的方法，其特征在于，所述方法还包括：
    发送指示所述第一数据集的指示信息；
    和/或，
    接收指示所述第二数据集的指示信息。
  13. 根据权利要求8至12任一项所述的方法，其特征在于，所述方法还包括：
    发送基于所述第三数据和所述第四数据确定的梯度信息和/或损失函数的结果。
  14. 根据权利要求8至13任一项所述的方法，其特征在于，所述方法还包括：
    接收或发送指示所述索引满足所述至少一项的指示信息。
  15. 一种通信方法，其特征在于，包括：
    对第一数据进行处理，得到第二数据；
    发送所述第二数据和第四数据，其中，所述第二数据用于确定第三数据，所述第四数据为所述第一数据对应的标签数据；其中，所述第二数据是基于第一神经网络对所述第一数据进行处理得到的，和/或，所述第三数据是基于第二神经网络对所述第二数据进行处理得到的。
  16. 根据权利要求15所述的方法，其特征在于，所述方法还包括：
    接收基于所述第三数据和所述第四数据确定的梯度信息和/或损失函数的结果。
  17. 一种通信方法，其特征在于，包括：
    接收第二数据和第四数据，其中，所述第二数据是基于第一数据得到的，所述第四数据为所述第一数据对应的标签数据；
    基于所述第二数据确定第三数据；其中，所述第二数据是基于第一神经网络对所述第一数据进行处理得到的，和/或，所述第三数据是基于第二神经网络对所述第二数据进行处理得到的。
  18. 根据权利要求17所述的方法，其特征在于，所述方法还包括：
    发送基于所述第三数据和所述第四数据确定的梯度信息和/或损失函数的结果。
  19. 一种通信方法，其特征在于，包括：
    对第一数据进行处理，得到第二数据；其中，所述第一数据为第一数据集中的数据；
    发送所述第二数据和第一索引，所述第二数据用于确定第三数据；其中，所述第一索引用于在第二数据集中确定第四数据，所述第二数据集包括所述第一数据集中的数据对应的标签数据，所述第四数据 是所述第一数据对应的标签数据；
    其中，所述第二数据是基于第一神经网络对所述第一数据进行处理得到的，和/或，所述第四数据是基于神经网络对所述第二数据进行处理得到的。
  20. 根据权利要求19所述的方法，其特征在于，所述方法还包括：
    接收基于所述第三数据和所述第四数据确定的梯度信息和/或损失函数的结果。
  21. 一种通信方法，其特征在于，包括：
    接收第二数据和第一索引，所述第二数据是基于第一数据得到的，所述第一数据为第一数据集中的数据；其中，所述第一索引用于在第二数据集中确定第四数据，所述第二数据集包括所述第一数据集中的数据对应的标签数据，所述第四数据是所述第一数据对应的标签数据；
    基于所述第二数据确定第三数据；其中，所述第二数据是基于第一神经网络对所述第一数据进行处理得到的，和/或，所述第三数据是基于第二神经网络对所述第二数据进行处理得到的。
  22. 根据权利要求21所述的方法，其特征在于，所述方法还包括：
    发送基于所述第三数据和所述第四数据确定的梯度信息和/或损失函数的结果。
  23. 一种通信装置，其特征在于，包括用于执行如权利要求1至22任一项所述的方法的模块。
  24. 一种通信装置，其特征在于，包括至少一个处理器，所述至少一个处理器与存储器耦合；所述至少一个处理器用于执行如权利要求1至22中任一项所述的方法。
  25. 根据权利要求24所述的通信装置，其特征在于，所述通信装置为芯片或芯片系统。
  26. 一种可读存储介质，其特征在于，所述存储介质中存储有计算机程序或指令，当所述计算机程序或指令被通信装置执行时，实现如权利要求1至22中任一项所述的方法。
  27. 一种计算机程序产品，其特征在于，包括指令，当所述指令在计算机上运行时，使得所述计算机执行如权利要求1至22中任一项所述的方法。
PCT/CN2024/114227 2023-11-03 2024-08-23 一种通信方法及相关设备 Pending WO2025092160A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202311461212.5A CN119946611A (zh) 2023-11-03 2023-11-03 一种通信方法及相关设备
CN202311461212.5 2023-11-03

Publications (1)

Publication Number Publication Date
WO2025092160A1 true WO2025092160A1 (zh) 2025-05-08

Family

ID=95535084

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2024/114227 Pending WO2025092160A1 (zh) 2023-11-03 2024-08-23 一种通信方法及相关设备

Country Status (2)

Country Link
CN (1) CN119946611A (zh)
WO (1) WO2025092160A1 (zh)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114337911A (zh) * 2020-09-30 2022-04-12 华为技术有限公司 一种基于神经网络的通信方法以及相关装置
CN114422380A (zh) * 2020-10-09 2022-04-29 维沃移动通信有限公司 神经网络信息传输方法、装置、通信设备和存储介质
US20230188302A1 (en) * 2020-08-31 2023-06-15 Qualcomm Incorporated Configurable metrics for channel state compression and feedback

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20230188302A1 (en) * 2020-08-31 2023-06-15 Qualcomm Incorporated Configurable metrics for channel state compression and feedback
CN114337911A (zh) * 2020-09-30 2022-04-12 华为技术有限公司 一种基于神经网络的通信方法以及相关装置
CN114422380A (zh) * 2020-10-09 2022-04-29 维沃移动通信有限公司 神经网络信息传输方法、装置、通信设备和存储介质

Also Published As

Publication number Publication date
CN119946611A (zh) 2025-05-06

Similar Documents

Publication Publication Date Title
WO2025092160A1 (zh) 一种通信方法及相关设备
WO2025092159A1 (zh) 一种通信方法及相关设备
WO2025179919A1 (zh) 一种通信方法及相关装置
WO2025118980A1 (zh) 一种通信方法及相关设备
WO2025025193A1 (zh) 一种通信方法及相关设备
WO2025179920A1 (zh) 一种通信方法及相关装置
WO2025103115A1 (zh) 一种通信方法及相关设备
WO2025059907A1 (zh) 一种通信方法及相关设备
WO2025086262A1 (zh) 一种通信方法及相关设备
WO2025059908A1 (zh) 一种通信方法及相关设备
WO2025189861A1 (zh) 一种通信方法及相关装置
WO2025190244A1 (zh) 一种通信方法及相关装置
WO2025175756A1 (zh) 一种通信方法及相关设备
WO2025190246A1 (zh) 一种通信方法及相关装置
WO2025189860A1 (zh) 一种通信方法及相关装置
WO2025208880A1 (zh) 一种通信方法及相关装置
WO2025107835A1 (zh) 一种通信方法及相关设备
WO2025190252A1 (zh) 一种通信方法及相关装置
WO2025139534A1 (zh) 一种通信方法及相关设备
WO2025019990A1 (zh) 一种通信方法及相关设备
WO2025098020A1 (zh) 一种通信方法及相关设备
WO2025140282A1 (zh) 一种通信方法及相关设备
WO2025190248A1 (zh) 一种通信方法及相关装置
WO2025189831A1 (zh) 一种通信方法及相关装置
WO2025118759A1 (zh) 一种通信方法及相关设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 24884167

Country of ref document: EP

Kind code of ref document: A1