Communication method and apparatus, device, and storage medium

Abstract

The present application relates to the technical field of communications, and discloses a communication method and apparatus, a device, and a storage medium, for use in solving the problem in the related art that a UE cannot accurately determine a calculation duration. The method comprises: acquiring at least one of first information used for indicating a processing capability of a terminal device and second information used for indicating model complexity of a first model; and determining a calculation duration on the basis of at least one of the first information and the second information, and target information, wherein the calculation duration refers to a duration for determining the target information, and the first model is used for determining target information.

Description

Communication method, device, equipment and storage medium

[0001] This application claims priority to the Chinese patent application filed with the State Intellectual Property Office on February 26, 2024, with application number 202410211128.6 and application name “Communication Method, Apparatus, Equipment and Storage Medium”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the field of communication technology, and in particular to a communication method, apparatus, device and storage medium. Background Art

[0003] With the continuous development of artificial intelligence, the application of neural networks in the field of communication technology continues to expand. For example, neural networks can be used to determine feedback information. Taking the feedback information as compressed information of downlink channel information as an example, after the user equipment (UE) obtains the downlink channel information based on the reference signal, it can use the downlink channel information as the input of the UE-side neural network to obtain compressed downlink channel information. Afterwards, the UE can send the compressed downlink channel information to the base station, and the base station will input the compressed downlink channel information into the neural network on the base station side to recover the downlink channel information.

[0004] As can be seen from the above, the computation time required for the UE to report feedback information can include the computation time of the neural network. If the UE cannot complete the computation by the reporting time required by the base station, it will not be able to send feedback information to the base station on schedule. Therefore, the UE needs to estimate the computation time required to obtain the feedback information so that the UE can determine whether the feedback information can be obtained before the reporting time required by the base station based on the computation time, thereby determining whether the feedback information can be sent to the base station on schedule.

[0005] Therefore, how to accurately determine the calculation time of feedback information is an urgent problem to be solved. Summary of the Invention

[0006] The embodiments of the present application provide a communication method, apparatus, device, and storage medium for solving the problem in related technologies that calculation duration cannot be accurately determined.

[0007] To achieve the above objectives, the present invention provides the following technical solutions:

[0008] In a first aspect, a communication method is provided. This method can be applied to any communication device, such as a terminal device or a network device. Specifically, the method is executed by the communication device, or by a module within the communication device, such as a chip, a chip system, or a circuit; or, alternatively, by a logic module or software that implements all or part of the functions of the communication device, without limitation. For ease of description, the following description uses execution by a communication device as an example.

[0009] The method includes: obtaining at least one of first information indicating the processing capability of the terminal device and second information indicating the model complexity of the first model, and determining the calculation duration based on at least one of the first information and the second information, as well as the target information.

[0010] The calculation time refers to the estimated time required to determine the target information.

[0011] The first model is used to determine target information. It is understood that the first model can be used to indicate a single model or multiple models. Specifically, if a single model is required to determine target information, the first model can be used to indicate the single model; if multiple models are required to determine target information, the first model can be used to indicate multiple models.

[0012] That is, when one model is needed to determine the target information, the number of models included in the first model is 1; when multiple models are needed to determine the target information, the number of models included in the first model is greater than 1.

[0013] The first information may directly indicate the processing capability of the terminal device, for example, the first information is information including the processing capability of the terminal device; the first information may also indirectly indicate the processing capability of the terminal device, for example, the first information is information including the processing capability level of the terminal device. This embodiment of the present application is not limited to this.

[0014] The second information is used to indicate the model complexity of the first model. The first model is used to determine the target information. Specifically, the second information can directly indicate the model complexity of the first model. For example, the second information is information including the model complexity of the first model. The second information can also indirectly indicate the model complexity of the first model, for example, the second information is information including the model complexity level of the first model. This embodiment of the present application is not limited to this.

[0015] It can be seen from the above method that when determining the calculation time, the calculation time is not determined based on information of one dimension, but is determined by integrating information of multiple dimensions, such as the processing capability of the communication equipment, the model complexity of the first model, and the target information. Therefore, the accuracy of the determined calculation time can be effectively improved.

[0016] In an optional implementation, the calculation duration is determined based on at least the first duration and the second duration.

[0017] The first duration is determined according to the type of target information.

[0018] The second duration is determined based on the first information indicating the processing capability of the terminal device and / or the second information indicating the model complexity of the first model. It will be understood that when the first model includes multiple models, that is, when multiple models are required to determine the target information, the second duration can be determined based on the first information indicating the processing capability of the terminal device and / or the second information indicating the model complexity of the multiple models.

[0019] By limiting the calculation time through the above content, the feasibility of this solution can be effectively improved.

[0020] In an optional implementation, the first duration is an estimated duration for determining the original information, or the first duration is an estimated duration for determining the characteristic information.

[0021] Both original information and feature information are used to determine target information.

[0022] By limiting the first duration through the above content, the feasibility of this solution can be effectively improved.

[0023] In an optional embodiment, when the target information is determined based on multiple original information, the calculation duration is determined based on at least a first duration and a third duration. The first duration is determined based on the type of the target information. The third duration is determined based on the number of the multiple original information and at least one of the first information and the second information.

[0024] Through the above content, in the scenario where the target information is determined based on multiple original information, when determining the calculation time, in addition to the various factors described above, such as the type of target information and the processing capability of the terminal device, it is also necessary to combine the number of multiple original information to determine the calculation time, so as to further improve the accuracy of the calculation time.

[0025] In an optional embodiment, when the target information is performance information of the second model, the calculation duration is determined based on at least a first duration and a fourth duration. The first duration is determined based on the type of the target information. The fourth duration is determined based on a method for determining the performance information of the second model and the first information.

[0026] Through the above content, in the scenario where the target information is the performance information of the second model, when determining the calculation time, in addition to the various factors described above, such as the type of target information and the processing capability of the terminal device, it is also necessary to combine the method of determining the performance information of the second model to determine the calculation time, so as to further improve the accuracy of the calculation time.

[0027] In an optional implementation, when there are multiple pieces of target information, the calculation duration is determined based on at least the first duration, the second duration, and the number of target information.

[0028] Through the above content, in the scenario where the number of target information is multiple, when determining the calculation time, in addition to the various factors described above, such as the type of target information, the processing capability of the terminal device, and the model complexity of the first model, the calculation time also needs to be determined in combination with the number of target information to further improve the accuracy of the calculation time.

[0029] In an optional implementation, calculating the duration further includes generating a duration of uplink control information (UCI) according to the target information.

[0030] According to the above content, when determining the calculation duration, the duration of generating the UCI may also be included, thereby further improving the accuracy of the calculation duration.

[0031] In an optional implementation, the communication device is a terminal device or a network device.

[0032] Based on this possible implementation method, the application scenario of this application is given to improve the application scope of this application.

[0033] In an optional implementation, when the communication device is a terminal device, the terminal device may receive instruction information from the network device, which is used to instruct the target information to be sent to the network device at the first moment.

[0034] In this way, the terminal device can determine the calculation duration after receiving the indication information, which enriches the process of this solution and makes this solution more complete.

[0035] In an optional implementation, the terminal device may also determine whether to send the target information to the network device based on the first moment and the calculation duration.

[0036] In this way, the terminal device can abandon sending the target information to the network device when the first moment and the calculation duration meet certain conditions, so as to save the transmission resources of the terminal device.

[0037] In an optional implementation, the terminal device may discard the target information if the time difference between the second moment and the first moment is less than the calculated duration, wherein the second moment is determined based on an instruction from the network device.

[0038] The above content limits the conditions for determining the sending of target information to the network device, which can effectively improve the feasibility of this solution.

[0039] In an optional embodiment, when the second moment is determined based on a reference signal indicated by the network device, the terminal device may determine the moment of receiving the reference signal as the second moment, or the terminal device may determine the moment of sending the reference signal by the network device as the second moment.

[0040] The second moment is limited through the above content, which effectively improves the feasibility of this solution based on the application scenario of this application.

[0041] In an optional implementation, when the communication device is a terminal device, the terminal device may send the first information to the network device.

[0042] In an optional implementation, when the communication device is a terminal device, the terminal device may send third information indicating the model complexity of one or more models to the network device.

[0043] In an optional implementation, when the communication device is a terminal device, the terminal device may receive third information indicating the model complexity of one or more models from the first device.

[0044] In this way, the terminal device does not need to determine the model complexity of each model by itself, which can save computing resources of the terminal device.

[0045] In an optional implementation, when the communication device is a network device, the network device may send instruction information to the terminal device for instructing the terminal device to send target information to the network device at the first moment.

[0046] In an optional implementation, the network device may also determine the first moment based on the calculation duration.

[0047] Through the above content, the network device determines the first moment based on the calculation duration, which can reduce the probability that the terminal device cannot complete the target information calculation before the first moment.

[0048] In a second aspect, a communication device is provided. The communication device may be a communication device or located in a communication device, such as a functional module or chip located in the communication device. The communication device may include: functional units for executing any one of the methods provided in the first aspect, wherein the actions performed by each functional unit are implemented by hardware or by hardware executing corresponding software implementations. The device includes a processing module.

[0049] The processing module is configured to determine a calculation duration based on at least one of the first information and the second information, as well as the target information. The calculation duration refers to an estimated duration for determining the target information; the first information indicates the processing capability of the terminal device; the second information indicates the model complexity of the first model; and the first model is used to determine the target information.

[0050] In a third aspect, a communication device is provided, which includes at least one processor coupled to at least one memory: at least one processor is used to execute a computer program or instruction stored in at least one memory, so that the communication device executes any communication method provided by the first aspect or any optional embodiment of the first aspect.

[0051] In a fourth aspect, a communication system is provided, comprising: a plurality of communication devices; each communication device is configured to execute any one of the communication methods provided in the first aspect or any optional implementation of the first aspect.

[0052] In a fifth aspect, a computer-readable storage medium is provided, comprising computer-executable instructions. When the computer-executable instructions are executed on a computer, the computer is enabled to execute any one of the communication methods provided in the first aspect or any optional implementation of the first aspect.

[0053] In the sixth aspect, a chip is provided, which includes: a processor and an interface circuit; the interface circuit is used to receive code instructions and transmit them to the processor; the processor is used to run the code instructions to execute any one of the communication methods provided in the first aspect or any optional implementation method of the first aspect.

[0054] In a seventh aspect, a computer program product is provided, comprising computer execution instructions, which, when the computer execution instructions are run on a computer, enable the computer to execute any one of the communication methods provided in the first aspect or any optional implementation of the first aspect.

[0055] It should be noted that the technical effects brought about by any implementation method in the second to seventh aspects can be referred to the technical effects brought about by the corresponding implementation method in the first aspect or any optional implementation method of the first aspect, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] FIG1 is a schematic diagram of the structure of a neural network;

[0057] FIG2 is a system architecture diagram of a communication system provided in an embodiment of the present application;

[0058] FIG3 is a system architecture diagram of another communication system provided in an embodiment of the present application;

[0059] FIG4 is a schematic diagram of the composition of a communication device provided in an embodiment of the present application;

[0060] FIG5 is a flow chart of a communication method provided in an embodiment of the present application;

[0061] FIG6 is a schematic diagram of an interaction flow of a communication method provided in an embodiment of the present application;

[0062] FIG7 is a schematic diagram of an interaction flow of another communication method provided in an embodiment of the present application;

[0063] FIG8 is a schematic structural diagram of a communication device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0064] In the description of this application, unless otherwise specified, " / " means "or", for example, A / B can mean A or B. "And / or" in this article is merely a description of the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, "at least one" means one or more, and "a plurality" means two or more. Words such as "first" and "second" do not limit the quantity and execution order, and words such as "first" and "second" do not limit them to be necessarily different.

[0065] It should be noted that, in this application, words such as "exemplarily" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described in this application as "exemplary" or "for example" should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplarily" or "for example" is intended to present the relevant concepts in a concrete manner.

[0066] "Used to indicate" can include direct and indirect indications, as well as explicit and implicit indications. When describing "a certain indication information is used to indicate A" or "indication information of A," this can include whether the indication information directly indicates A or indirectly indicates A, but does not necessarily mean that the indication information contains A. The information indicated by a certain information is referred to as the information to be indicated. During implementation, there are many ways to indicate the information to be indicated. For example, but not limited to, the information to be indicated can be directly indicated, such as the information to be indicated itself or an index of the information to be indicated. The information to be indicated can also be indirectly indicated by indicating other information, where the other information is associated with the information to be indicated. Alternatively, only a portion of the information to be indicated can be indicated, while the rest of the information to be indicated is known or agreed upon in advance. For example, a pre-agreed (e.g., protocol-specified) order of information can be used to indicate specific information, thereby reducing indication overhead to a certain extent. Furthermore, common portions of various information can be identified and indicated uniformly, reducing the indication overhead associated with separately indicating the same information. Furthermore, the specific indication method can also be any of the various existing indication methods, such as, but not limited to, the aforementioned indication methods and their various combinations. The specific details of various indication methods can be referred to the prior art and will not be elaborated herein. As can be seen from the above, for example, when multiple pieces of information of the same type need to be indicated, different indication methods may be used for different pieces of information. During implementation, the desired indication method can be selected based on specific needs. The embodiments of this application do not limit the selected indication method. Thus, the indication methods involved in the embodiments of this application should be understood to encompass various methods by which the party to be indicated can be informed of the information to be indicated. The information to be indicated can be sent as a whole or as multiple sub-information sent separately, and the sending periods and / or sending times of these sub-information can be the same or different. The specific sending method is not limited in this application. The sending periods and / or sending times of these sub-information can be predefined, for example, according to a protocol, or can be configured by the transmitting device sending configuration information to the receiving device. The configuration information can, for example, but is not limited to, one or a combination of at least two of radio resource control signaling, media access control (MAC) layer signaling, and physical layer signaling. The radio resource control signaling includes, for example, radio resource control (RRC) signaling; the MAC layer signaling includes, for example, a MAC control element (CE); and the physical layer signaling includes, for example, downlink control information (DCI).

[0067] Artificial Intelligence (AI) can enable machines to learn, accumulate experience, and solve problems that humans can solve through experience, such as natural language understanding, image recognition, and chess.

[0068] Machine learning is an implementation of artificial intelligence. It's a method that empowers machines to perform tasks that are impossible to program directly. In practical terms, machine learning involves training models using data and then using those models to make predictions.

[0069] Neural Network (NN) is a specific implementation form of machine learning. As shown in Figure 1, the multi-layer structure of a neural network can include an input layer, a hidden layer, and an output layer. The input layer can process the received values ​​through neurons and pass them to the middle hidden layer. The hidden layer is processed by neurons and then passed to the output layer. The output layer obtains the final output after processing by neurons. The hidden layer can affect the ability of the neural network to extract information and fit functions. Increasing the number of hidden layers or expanding the width of each hidden layer can improve the function fitting ability of the DNN. It should be understood that the neural network in this application can also be referred to as a neural network model.

[0070] Neural networks can include feedforward neural networks (FNN), convolutional neural networks (CNN), and recurrent neural networks (RNN).

[0071] Feedforward neural networks are specialized for processing data with grid-like structures, such as time series data and image data. Instead of using all input information at once, feedforward neural networks employ convolution operations, capturing a portion of the information using a fixed-size window. This significantly reduces the computational complexity of model parameters. Furthermore, depending on the type of information captured by the window, different convolution kernels can be used for each window, enabling feedforward neural networks to better extract features from the input data.

[0072] A convolutional neural network is a type of neural network that uses feedback time series information. Its input consists of a new input value at the current moment and its own output value at the previous moment. Convolutional neural networks are suitable for capturing temporally correlated sequence features and are applicable to fields such as speech recognition and channel coding.

[0073] The characteristic of a feedforward neural network is that neurons in adjacent layers are fully connected to each other. A feedforward neural network usually requires a large amount of storage space, resulting in a high computational complexity of the feedforward neural network.

[0074] The network structures of various neural networks, such as feedforward neural networks, convolutional neural networks, and recurrent neural networks, are all based on neurons. As mentioned above, each neuron performs a weighted summation operation on its input values ​​and calculates the corresponding output value through a nonlinear function. We can call the weights and nonlinear functions of each neuron in a neural network the parameters of the neural network (or the model parameters of the neural network).

[0075] With the continuous development of artificial intelligence, the application of neural networks in the field of communication technology has continued to expand, using neural networks to assist communication equipment in completing some computing tasks / processing tasks to improve communication performance. For example, neural networks can be used to determine feedback information. Taking the feedback information as compressed information of downlink channel information as an example, after the user equipment obtains the downlink channel information based on the reference signal, it can use the downlink channel information as input to the UE-side neural network to obtain compressed downlink channel information. Afterwards, the UE can send the compressed downlink channel information to the base station, and the base station can input the compressed downlink channel information into the neural network on the base station side to recover the downlink channel information.

[0076] As can be seen from the above, the computation time required for the UE to report feedback information can include the computation time of the neural network. If the UE cannot complete the computation by the reporting time required by the base station, it will not be able to send feedback information to the base station on schedule. Therefore, the UE needs to determine the computation time required to obtain feedback information so that the UE can determine whether it can obtain feedback information before the reporting time required by the base station based on this computation time, and thus determine whether it can send feedback information to the base station on schedule.

[0077] In one example, the UE may determine the calculation duration based on the content of the feedback information.

[0078] For example, when the feedback information is channel state information (CSI), the UE may determine the calculation duration for reporting the CSI this time according to the durations corresponding to different predefined CSI codebooks.

[0079] In another example: the UE may determine the calculation duration based on the model complexity of the UE-side model.

[0080] For example, the UE may determine the calculation duration based on factors such as the model structure and parallelism of the neural network model.

[0081] However, the above two methods ignore the influence of other factors on the calculation duration, resulting in reduced accuracy of the calculation duration determined by the UE, and further making it impossible for the UE to accurately determine whether it needs to send compressed feedback information to the base station based on the calculation duration.

[0082] Therefore, how to accurately determine the calculation time is an urgent problem to be solved.

[0083] In view of this, an embodiment of the present application provides a communication method that can be applied to any communication device such as a terminal device or a network device. The communication device can obtain at least one of first information indicating the processing capability of the terminal device and second information indicating the model complexity of a first model, and determine a calculation duration based on at least one of the first information and the second information, as well as target information. The calculation duration refers to the estimated duration for determining the target information. The first model is used to determine the target information.

[0084] It can be seen from this that in an embodiment of the present application, when determining the calculation time, the calculation time is determined based on at least one of the processing capability of the terminal device and the model complexity of the first model, as well as multi-dimensional information such as target information. Therefore, the accuracy of the determined calculation time can be effectively improved.

[0085] FIG2 is a system architecture diagram of a communication system provided in an embodiment of the present application. The system architecture diagram may include multiple communication devices, such as terminal device 201, terminal device 202, and network device 203 as shown in FIG2. Terminal device 201 and terminal device 202 can access network device 203 and communicate with network device 203.

[0086] The terminal equipment involved in the embodiments of the present application may be a UE, an access terminal, a terminal unit, a user station, a terminal station, a mobile station, a mobile station, a remote station, a remote terminal, a user terminal (Terminal Equipment, TE), a mobile device, a wireless communication device, a terminal agent, a tablet computer (Pad), a handheld device with wireless communication function, a computing device or other processing device connected to a wireless modem, a vehicle-mounted device, a vehicle-mounted transceiver unit, a wearable device, or a terminal device in a fifth generation mobile communication technology (5th Generation, 5G) network or a public land mobile network (Public Land Mobile Network, PLMN) evolved after 5G. The access terminal may be a cellular phone, a cordless phone, a Session Initiation Protocol (SIP) phone, a Wireless Local Loop (WLL) station, a Personal Digital Assistant (PDA), a handheld device with wireless communication function, a computing device or other processing device connected to a wireless modem, a vehicle-mounted device, a drone, a robot, a smart point of sale (POS) machine, a customer-premises terminal device (Customer-Premises The terminal device may be a wireless terminal in the form of a consumer equipment (CPE) or wearable device, 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 smart city, a wireless terminal in smart home, etc. Alternatively, the terminal device may be a terminal with communication function in the Internet of Things (IOT), such as a terminal in V2X (e.g., a vehicle-to-everything (V2X) device), a terminal in D2D communication, or a terminal in M2M communication. The terminal device may be mobile or fixed.

[0087] The embodiments of this application do not limit the form of the terminal device. The device used to implement the functions of the terminal device can be the terminal device; it can also be a device that supports the terminal device to implement the functions, such as a chip system. The device can be installed in the terminal device or used in conjunction with the terminal device. In the embodiments of this application, the chip system can be composed of chips or include chips and other discrete devices.

[0088] The embodiment of the present application does not limit the number of terminal devices and may include more or fewer terminal devices than those in Figure 2.

[0089] The network device involved in this application may be a device for communicating with a terminal device, for example, it may include an evolved base station (NodeB or eNB or Evolutional Node B, e-NodeB) in a Long Term Evolution (LTE) system or an enhanced LTE (LTE-Advanced, LTE-A) system, such as a traditional macro base station eNB and a micro base station eNB in ​​a heterogeneous network scenario. Alternatively, it may include a Next Generation Node B (gNB) in an NR system. Alternatively, it may include a Transmission Reception Point (TRP), a home base station (e.g., Home Evolved NodeB, or Home Node B, HNB), a Base Band Unit (BBU), a Base Band Pool (BBU Pool), or a Wireless Fidelity (WiFi) Access Point (AP), etc. Alternatively, it may include a base station in a non-terrestrial network (NTN), which can be deployed on an aircraft or satellite. In an NTN, the network device can function as a Layer 1 (L1) relay, a base station, or an Integrated Access and Backhaul (IAB) node. Alternatively, the network device can be a device that implements base station functions in the IoT, such as drone communications, V2X, D2D, or machine-to-machine (M2M) communications.

[0090] Optionally, the base station in the embodiments of the present application may be an integrated base station, or may be a base station including a centralized unit (CU) and a distributed unit (DU). A base station including both a CU and a DU may also be referred to as a base station with separate CU and DU components, such as a base station including a gNB-CU and a gNB-DU. The CU may also be separated into a CU control plane (CU-CP) and a CU user plane (CU-CP), such as a base station including a gNB-CU-CP, a gNB-CU-UP, and a gNB-DU.

[0091] Optionally, the base station in the embodiment of the present application may include various forms of base stations, such as: macro base stations, micro base stations (also called small stations), relay stations, access points, home base stations, TRPs, transmitting points (TP), mobile switching centers, etc., and the embodiments of the present application do not make specific limitations on this.

[0092] In the embodiments of the present application, the form of the network device is not limited. The device used to implement the function of the network device can be a network device; it can also be a device that can support the network device to implement the function, such as a chip system. The device can be installed in the network device or used in conjunction with the network device.

[0093] It should be noted that Figure 2 is an exemplary figure, and the number of terminal devices shown in Figure 2 and the naming of the interfaces between the devices in Figure 2 are not limited. In addition to the network elements shown in Figure 2, the communication system shown in Figure 2 may also include other devices, such as access network devices, without limitation.

[0094] The communication system provided in the embodiments of the present application may further include an AI entity, which may be an AI network element or an AI module. The embodiments of the present application are not limited to this. The AI ​​entity may be deployed with a neural network model and may perform AI-related operations such as building a training data set and training a model.

[0095] In some embodiments, the AI ​​entity can be integrated into the network device 203 shown in Figure 2, so that after the network device 203 receives the data related to the AI ​​model reported by the terminal device, it can process the data through the AI ​​entity in the network device 203, obtain the corresponding processing results, and send the processing results to the terminal device.

[0096] In other embodiments, as shown in FIG3 , the AI ​​entity may be independent of the network device. After receiving the data related to the AI ​​model reported by the terminal device, the network device 203 may send the data to the AI ​​entity. After the AI ​​entity processes the data and obtains a corresponding processing result, it may send the processing result to the network device, and the network device may then send the processing result to the terminal device.

[0097] In specific implementations, the communication devices in Figures 2 or 3 may adopt the structure shown in Figure 4, or include the components shown in Figure 4. Figure 4 is a schematic diagram of the composition of a communication device 400 provided in an embodiment of the present application. The communication device 400 may be a terminal device or a chip or system on chip in a terminal device. Alternatively, the communication device 400 may be a network device or a chip or system on chip in a network device. As shown in Figure 4, the communication device 400 includes a processor 401, a communication interface 402, and a communication line 403.

[0098] Furthermore, the communication device 400 may further include a memory 404 , wherein the processor 401 , the memory 404 and the communication interface 402 may be connected via a communication line 403 .

[0099] The processor 401 is a central processing unit (CPU), a general-purpose processor, a network processor (NP), a digital signal processor (DSP), a microprocessor, a microcontroller, a programmable logic device (PLD), or any combination thereof. The processor 401 may also be other devices with processing functions, such as circuits, devices, or software modules. For example, the processor may have the ability to process network models (such as DNN network models), without limitation.

[0100] Communication interface 402 is used to communicate with other devices or other communication networks. Such other communication networks may be Ethernet, radio access networks (RAN), wireless local area networks (WLAN), etc. Communication interface 402 may be a module, circuit, communication interface, or any other device capable of implementing communication.

[0101] The communication line 403 is used to transmit information between the components included in the communication device 400.

[0102] The memory 404 is used to store instructions, where the instructions may be computer programs.

[0103] The memory 404 may be a read-only memory (ROM) or other type of static storage device that can store static information and / or instructions, or a random access memory (RAM) or other type of dynamic storage device that can store information and / or instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compact disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), magnetic disk storage media or other magnetic storage devices, etc., without limitation.

[0104] It should be noted that memory 404 can exist independently of processor 401 or can be integrated with processor 401. Memory 404 can be used to store instructions, program code, samples, etc. Memory 404 can be located within communication device 400 or outside of communication device 400, without limitation. Processor 401 is configured to execute instructions stored in memory 404 to implement the sample processing method for a distributed training system provided in the following embodiments of this application.

[0105] In one example, the processor 401 may include one or more CPUs, such as CPU0 and CPU1 in FIG. 4 .

[0106] As an optional implementation, the communication device 400 includes multiple processors. For example, in addition to the processor 401 in FIG. 4 , it may also include a processor 407 .

[0107] As an optional implementation, the communication apparatus 400 further includes an output device 405 and an input device 406. For example, the input device 406 is a keyboard, a mouse, a microphone, a joystick, or the like, and the output device 405 is a display screen, a speaker, or the like.

[0108] It should be noted that the communication device 400 can be a desktop computer, a portable computer, a network server, a mobile phone, a tablet computer, a wireless terminal, an embedded device, a chip system, or a device having a structure similar to that shown in FIG4 . In addition, the structure shown in FIG4 does not limit the access network device, the core network device, and the terminal device. In addition to the components shown in FIG4 , the access network device, the core network device, and the terminal device may include more or fewer components than shown, or combine certain components, or arrange the components differently.

[0109] In the embodiment of the present application, the chip system can be composed of chips, or can include chips and other discrete devices.

[0110] In addition, the actions, terms, etc. involved in the various embodiments of this application can refer to each other without limitation. The names of the information exchanged between the various devices in the embodiments of this application or the names of the parameters in the information are only examples. Other names can also be used in the specific implementation without limitation. The execution subject of the embodiments of this application can be an access network device, or a device in the access network device, such as a chip. It can also be a core network device, or a device in the core network device, such as a chip. It can also be a terminal device, or a device in a terminal device, such as a chip.

[0111] 2, the communication method provided by the embodiment of the present application is described below, taking the communication device as the terminal device in FIG2 as an example. FIG5 is a flow chart of a communication method provided by the embodiment of the present application, as shown in FIG5, the method includes:

[0112] S501: Obtain at least one of first information and second information.

[0113] The first information is used to indicate the processing capability of the terminal device. Specifically, the processing capability of the terminal device can be alternatively described as the terminal device's ability to process information or the terminal device's computing capability. The following description uses the processing capability of the terminal device as an example. Specifically, the processing capability of the terminal device can be determined based on the number of processors in the terminal device and / or the processing capability of each processor.

[0114] The processing capability of each processor may include but is not limited to the computing speed, storage capacity, and model processing capability of the processor. These capabilities may be collectively referred to as the processing capability of the processor.

[0115] The first information may directly indicate the processing capability of the terminal device. For example, the first information may include information about the processing capability of the terminal device, and the first information may include information such as the number of processors in the terminal device and / or the processing capability of the processors. The first information may also indirectly indicate the processing capability of the terminal device. For example, the first information may include information about the processing capability level of the terminal device, where the processing capability level is determined based on the processing capability of the terminal device. This embodiment of the present application is not limited to this.

[0116] The second information is used to indicate the model complexity of the first model. The first model is used to determine the target information. Specifically, the second information can directly indicate the model complexity of the first model. For example, the second information is information including the model complexity of the first model. The second information can also indirectly indicate the model complexity of the first model, for example, the second information is information including the model complexity level of the first model. This embodiment of the present application is not limited to this.

[0117] It is understood that the first model can be used to indicate one model or multiple models. Specifically, when one model is required to determine target information, the first model can be used to indicate one model; when multiple models are required to determine target information, the first model can be used to indicate multiple models.

[0118] S502: Determine a calculation duration based on at least one of the first information and the second information, and the target information.

[0119] The calculation duration refers to the estimated duration for determining the target information. The estimated duration can refer to the average duration for obtaining the target information or the shortest duration for obtaining the target information within a preset time range.

[0120] In an optional implementation, the calculation duration may be determined based on at least the first duration and the second duration.

[0121] The first duration may be determined according to the type of target information.

[0122] The types of target information may include, but are not limited to, original information and feature information, which are not limited in the embodiments of the present application. Both original information and feature information are used to determine target information.

[0123] The characteristic information may include but is not limited to a characteristic vector. For example, the characteristic information of a channel is a channel characteristic vector, which is not limited in this embodiment of the present application.

[0124] Specifically, when determining the first duration, the terminal device may determine the first duration according to the type of the target information and a preset type-duration table.

[0125] The preset type-duration table may include a correspondence between information types and durations. For example, the preset type-duration table may include, but is not limited to, two correspondences: the duration corresponding to the original information and the duration corresponding to the characteristic information. The duration corresponding to the original information may be used to indicate the duration of obtaining the original information, and the duration corresponding to the characteristic information may be used to indicate the duration of obtaining the characteristic information.

[0126] It should be noted that the correspondence in the preset type-duration table can also be the correspondence between the information identifier, the type of information and the duration. The information identifier is used to represent the attributes of the information, such as channel information or model information, etc. The embodiments of the present application do not limit this.

[0127] It should be noted that the duration in the preset type-duration table can indicate the average duration for the terminal device to obtain information of this type within the preset time range, or it can be the shortest duration for the terminal device to obtain information of this type within the preset time range. The embodiments of the present application do not limit this.

[0128] The second duration can be determined based on the first information and / or the second information. That is, the second duration can be determined based on either the processing capability of the terminal device or the processing capability level of the terminal device, and / or the model complexity of the first model or the model complexity level of the first model. The following description uses the example of determining the second duration based on the processing capability level of the terminal device and the model complexity level of the first model.

[0129] It can be understood that when the first model contains multiple models, that is, multiple models are needed to determine the target information, the second duration can be determined based on the first information used to indicate the processing capability of the terminal device, and / or the second information used to indicate the model complexity of multiple models.

[0130] Specifically, when determining the second duration, the terminal device can first determine the first model used to determine the target information, and then determine the second duration based on the model complexity level of the first model, the processing capability level of the terminal device and the preset level-duration table; it can also determine the second duration based on the model complexity level of the first model and the preset complexity level-duration table; it can also determine the second duration based on the processing capability level of the terminal device and the capability level-duration table.

[0131] The preset level-duration table may include the correspondence between the model complexity level of the model, the processing capability level of the terminal device and the duration. For example, when the model complexity level of the model includes two levels (such as level 1 and level 2), and the processing capability level of the terminal device includes two levels (such as level A and level B), the level-duration table may include four correspondences, namely: the duration corresponding to when the model complexity of the model is level 1 and the processing capability level of the terminal device is level A, the duration corresponding to when the model complexity of the model is level 1 and the processing capability level of the terminal device is level B, the duration corresponding to when the model complexity of the model is level 2 and the processing capability level of the terminal device is level A; and the duration corresponding to when the model complexity of the model is level 2 and the processing capability level of the terminal device is level B. The durations contained in each correspondence may be equal or unequal, and this is not limited in the embodiments of the present application.

[0132] The preset complexity level-duration table may include the correspondence between the model complexity level and duration of the model. For example, when the model complexity level of the model includes two levels (such as level 1 and level 2), the complexity level-duration table may include two correspondences, namely: the duration corresponding to when the model complexity of the model is level 1; the duration corresponding to when the model complexity of the model is level 2. The durations contained in each correspondence may be equal or unequal, and this is not limited in the embodiments of the present application.

[0133] When the first model includes multiple models, the model complexity level of the first model can be the average of the model complexity levels of the multiple models, or the sum of the model complexity levels of the multiple models, which is not limited in this embodiment of the present application. Alternatively, when the first model includes multiple models, the second duration is the sum of the corresponding durations of each model, or the average of the corresponding durations of each model, which is not limited in this embodiment of the present application. The method for determining the corresponding duration of each model can refer to the method above.

[0134] When determining the calculation duration based on the first duration and the second duration, the sum of the first duration and the second duration can be used as the calculation duration, or the result of the weighted sum of the first duration and the second duration can be used as the calculation duration, or a calculation duration-first duration-second duration correspondence table can be preset. This embodiment of the present application is not limited to this, and the following description takes the sum of the first duration and the second duration as the calculation duration as an example.

[0135] For example, it is assumed that the preset type-duration table contains two corresponding relationships, namely: the duration corresponding to the original information is 10 milliseconds, and the duration corresponding to the characteristic information is 15 milliseconds. The preset level-duration table contains four corresponding relationships, namely: when the model complexity level of the model is level 1 and the processing capability level of the terminal device is level A, the corresponding duration is 5 milliseconds; when the model complexity level of the model is level 1 and the processing capability level of the terminal device is level B, the corresponding duration is 10 milliseconds; when the model complexity level of the model is level 2 and the processing capability level of the terminal device is level A, the corresponding duration is 10 milliseconds; when the model complexity level of the model is level 2 and the processing capability level of the terminal device is level B, the corresponding duration is 15 milliseconds.

[0136] In some embodiments, assuming that the target information is compressed channel information of channel M, the first model is an AI encoder, the model complexity level of the AI ​​encoder is level 2, and the processing capability level of the terminal device is level A. On this basis, the terminal device determining the duration of the target information may at least include: (1) obtaining an estimated duration of the channel information of channel M, i.e., a first duration; and (2) compressing the channel information of channel M through the AI ​​encoder to obtain an estimated duration of the compressed channel information of channel M, i.e., a second duration.

[0137] The channel information of channel M is original information. Therefore, when determining the calculation duration, the terminal device can use the duration of 10 milliseconds corresponding to the original information in the preset type-duration table as the first duration, and according to the model complexity level 2 of the AI ​​encoder and the processing capability level A of the terminal device, determine the second duration as 10 milliseconds in the preset level-duration table, and then determine the calculation duration as 20 milliseconds based on the first duration of 10 milliseconds and the second duration of 10 milliseconds.

[0138] In some other embodiments, assuming that the target information is a compressed channel feature vector of channel M, the first model is an AI encoder, the model complexity level of the AI ​​encoder is level 2, and the processing capability level of the terminal device is level A. On this basis, the terminal device determines the duration of the target information, which may at least include: (1) decomposing the channel M using a preset decomposition method to obtain an estimated duration of the channel feature vector, i.e., a first duration; and (2) compressing the channel feature vector of channel M using the AI ​​encoder to obtain an estimated duration of the compressed channel feature vector of channel M, i.e., a second duration.

[0139] The preset decomposition method may include but is not limited to singular value decomposition (SVD), which is not limited in the embodiment of the present application.

[0140] The channel feature vector of channel M belongs to feature information. Therefore, when determining the calculation duration, the terminal device can use the duration of 15 milliseconds corresponding to the feature information in the preset type-duration table as the first duration, and according to the model complexity level 2 of the AI ​​encoder and the processing capability level A of the terminal device, determine the second duration as 10 milliseconds in the preset level-duration table, and then determine the calculation duration as 25 milliseconds based on the first duration of 15 milliseconds and the second duration of 10 milliseconds.

[0141] It can be seen from the above method that when determining the calculation time, the calculation time is not determined based on information of one dimension, but is determined by integrating information of multiple dimensions, such as the processing capability of the terminal device, the model complexity of the first model, and the target information. Therefore, the accuracy of the determined calculation time can be effectively improved.

[0142] In an optional embodiment, when the target information is determined based on multiple original information, the calculation duration can be determined based on at least the first duration and the third duration. For example, the calculation duration can be determined based on the first duration and the third duration, or based on the first duration, the second duration, and the third duration, which is not limited in this embodiment of the present application.

[0143] The third duration can be determined based on the number of multiple original information and at least one of the first information and the second information. Specifically, the third duration can be determined based on a specific application scenario. For example, in some embodiments, when the application scenario is to predict the channel information at a future time based on the channel information of multiple known channels, the third duration can be determined based on the channel information of multiple known channels, the algorithm complexity of the prediction algorithm used to predict the channel information at a future time, and the processing capability of the terminal device using the following formula 1. This embodiment of the present application does not limit this.

[0144] The channel information of the known channel is an optional original information. T3=X1*X2*X3+C1 (Formula 1)

[0145] Among them, T3 represents the third time length, X1 indicates the algorithm complexity of the prediction algorithm, X2 indicates the number of multiple original information, X3 indicates the processing capability level of the terminal device, and C1 is the bias.

[0146] Assuming that the application scenario is to predict the channel information at a future time based on the channel information of channel M and channel B, the third duration is determined based on the channel information of multiple known channels, the algorithm complexity of the prediction algorithm used to predict the channel information at a future time, and the processing capability level of the terminal device, and the first model is an AI encoder.

[0147] On this basis, in some embodiments, the terminal device determines the duration of the target information, which may include: (1) obtaining the expected duration of the channel information of channel M and the channel information of channel B, i.e., the first duration; (2) based on the channel information of channel M and the channel information of channel B, using a prediction algorithm to predict the expected duration of the channel information at a future moment, i.e., the third duration.

[0148] In other embodiments, the terminal device may determine the duration of the target information by: (1) obtaining the estimated duration of the channel information of channel M and the channel information of channel B, i.e., the first duration; (2) using a prediction algorithm to predict the estimated duration of the channel information at a future moment based on the channel information of channel M and the channel information of channel B, i.e., the third duration; (3) compressing the predicted channel information at a future moment through an AI encoder to obtain the estimated duration of the compressed channel information at the future moment, i.e., the second duration.

[0149] Among them, the method for determining the first duration and the second duration can refer to the above content and will not be repeated here.

[0150] In some embodiments, when determining the third duration, the terminal device can first determine the duration corresponding to the prediction algorithm and the processing capability level of the terminal device based on the algorithm complexity of the prediction algorithm, the processing capability level of the terminal device and a preset algorithm complexity-capability level-duration table, and finally determine the third duration based on the duration and the amount of original information.

[0151] The preset algorithm complexity-capability level-duration table may include the correspondence between the algorithm complexity, the processing capability level of the terminal device and the duration. For example, when the algorithm complexity is 10 and the processing capability level of the terminal device is A, the corresponding duration; when the algorithm complexity is 10 and the processing capability level of the terminal device is B, the corresponding duration.

[0152] In other embodiments, when determining the third duration, the terminal device can first determine the third duration based on the algorithm complexity of the prediction algorithm, the processing capability level of the terminal device, the amount of original information and the preset algorithm complexity-capability level-quantity-duration table.

[0153] The preset algorithm complexity-capability level-quantity-duration table may include the correspondence between the algorithm complexity, the processing capability level of the terminal device, the quantity of original information and the duration. For example, when the algorithm complexity is 10, the processing capability level of the terminal device is level A, and the number of original information is 2, the corresponding duration; when the algorithm complexity is 10, the processing capability level of the terminal device is level B, and the number of original information is 2, the corresponding duration.

[0154] In other embodiments, when determining the third duration, the terminal device may first determine the third duration based on the algorithm complexity of the prediction algorithm, the amount of original information, and a preset algorithm complexity-quantity-duration table.

[0155] In other embodiments, when determining the third duration, the terminal device may first determine the third duration based on the processing capability level of the terminal device, the amount of original information, and a preset capability level-quantity-duration table.

[0156] Assuming that the application scenario is to predict the channel feature vector at a future time based on the channel information of channel M and channel B, the third duration is determined based on the channel information of multiple known channels, the algorithm complexity of the prediction algorithm used to predict the channel information at a future time, and the processing capability level of the terminal device, and the first model is an AI encoder.

[0157] On this basis, in some embodiments, the terminal device determines the duration of the target information, which may include at least: (1) decomposing channel M and channel B respectively using a preset decomposition method to obtain the estimated duration of the channel characteristic vector of channel M and the channel characteristic vector of channel B, i.e., the first duration; (2) based on the channel characteristic vector of channel M and the channel characteristic vector of channel B, using a prediction algorithm to predict the estimated duration of the channel characteristic vector at a future moment, i.e., the third duration.

[0158] In other embodiments, the terminal device determines the duration of the target information, which may include at least: (1) decomposing channel M and channel B respectively using a preset decomposition method to obtain the estimated duration of the channel characteristic vector of channel M and the channel characteristic vector of channel B, i.e., the first duration; (2) based on the channel characteristic vector of channel M and the channel characteristic vector of channel B, using a prediction algorithm to predict the estimated duration of the channel characteristic vector at a future moment, i.e., the third duration; (3) compressing the predicted channel characteristic vector at a future moment through an AI encoder to obtain the estimated duration of the compressed channel characteristic vector at the future moment, i.e., the second duration.

[0159] Among them, the method for determining the first duration, the second duration and the third duration can refer to the above content and will not be repeated here.

[0160] Through the above content, in the scenario where the target information is determined based on multiple original information, when determining the calculation time, in addition to the various factors described above, such as the type of target information, the processing capability of the terminal device, and the model complexity of the first model, it is also necessary to combine the number of multiple original information to determine the calculation time, thereby further improving the accuracy of the calculation time.

[0161] In an optional embodiment, when the target information is performance information of the second model, the calculation duration can be determined based on at least the first duration and the fourth duration. For example, the calculation duration can be determined based on the first duration and the fourth duration, or based on the first duration, the second duration, and the fourth duration, which is not limited in this embodiment of the present application.

[0162] The fourth duration may be determined based on the method for determining the performance information of the second model and the first information.

[0163] The second model may be the same as the first model or different from the first model, and this embodiment of the present application does not limit this.

[0164] The performance information of the second model can be determined based on the type of the second model. For example, when the second model is an AI decoder, the performance information of the second model can be the information recovery accuracy of the AI ​​decoder, that is, the accuracy of the decoded information obtained by the AI ​​decoder when decoding compressed information. The embodiments of the present application do not limit the performance information of the second model.

[0165] The method for determining the performance information of the second model can be indicated by the network device, or it can be randomly selected by the terminal device from multiple preset methods for determining model performance information. The embodiment of the present application does not limit the method for determining the performance information of the second model.

[0166] Assuming that the target information is the information recovery accuracy of the AI ​​decoder deployed on the network device side, the method for determining the performance information of the second model is: determining the performance information of the second model based on the original channel information.

[0167] On this basis, in some embodiments, the terminal device determines the duration of the target information, which may include: (1) obtaining the expected duration of the original channel information, i.e., the first duration; and (2) determining the expected duration of the information recovery accuracy of the AI ​​decoder based on the method for determining the performance information of the second model and the processing capability level of the terminal device, i.e., the fourth duration.

[0168] Among them, the method for determining the first duration can refer to the above content and will not be repeated here.

[0169] In other embodiments, when determining the fourth duration, the terminal device can determine the expected duration for determining the performance information of the second model, i.e., the fourth duration, based on the method identifier for determining the performance information of the second model, the processing capability level of the terminal device, and the preset method identifier-capability level-duration table.

[0170] The preset method identifier-capability level-duration table may include a method for determining the performance information of the second model, a correspondence between the processing capability level of the terminal device and the duration.

[0171] Assuming that the target information is the information recovery accuracy of the AI ​​decoder deployed on the network device side, the method for determining the performance information of the second model is: determining the performance information of the second model based on the reference model, and the input of the reference model is the output of the encoder, and the output of the reference model is the original channel information after decoding.

[0172] The reference model is deployed on the terminal device side and is the same as or similar to the second model.

[0173] The reference model may be indicated by the network device or randomly selected by the terminal device, which is not limited in the present embodiment.

[0174] On this basis, in some embodiments, the terminal device determines the duration of the target information, which may include: (1) obtaining the expected duration of the original channel information, i.e., the first duration; (2) compressing the original channel information through the AI ​​encoder to obtain the expected duration of the compressed original channel information, i.e., the second duration; (3) inputting the compressed original channel information into the reference model to obtain the expected duration of the decoded original channel information, and determining the expected duration of the information recovery accuracy of the AI ​​decoder based on the target accuracy evaluation algorithm and the processing capability level of the terminal device, i.e., the fourth duration.

[0175] Among them, the method for determining the first duration and the second duration can refer to the above content and will not be repeated here.

[0176] The target accuracy assessment algorithm can be sent by the network device to the terminal device, or it can be randomly selected by the terminal device from multiple preset accuracy assessment algorithms, or it can be preset, and the embodiments of the present application do not limit this.

[0177] When determining the fourth duration, the terminal device may first determine the estimated duration for inputting the compressed original channel information into the reference model to obtain the decoded original channel information based on the identifier of the AI ​​decoder or reference model and the preset decoder-duration table, or may determine it with reference to the above-mentioned method for determining the second duration. Afterwards, the terminal device may determine the estimated duration for calculating the performance information of the second model based on the algorithm complexity of the target accuracy evaluation algorithm, the processing capability level of the terminal device, and the preset algorithm complexity-capability level-duration table, and finally determine the fourth duration based on the estimated duration for inputting the compressed original channel information into the reference model to obtain the decoded original channel information and the estimated duration for calculating the performance information of the second model.

[0178] The decoder-duration table may include a correspondence between a decoder or reference model identifier and a duration, for example, the duration corresponding to decoder A and the duration corresponding to decoder B.

[0179] The algorithm complexity-capability level-duration table may include the correspondence between the algorithm complexity, the capability level of the terminal device and the duration. For example, the algorithm complexity is 20 and the corresponding duration when the capability level of the terminal device is A; the algorithm complexity is 25 and the corresponding duration when the capability level of the terminal device is B.

[0180] Assuming that the target information is the information recovery accuracy of the AI ​​decoder deployed on the network device side, the method for determining the performance information of the second model is: determining the performance information of the second model based on the channel feature vector of the original channel information.

[0181] On this basis, in some embodiments, the duration of the target information determined by the terminal device may include: (1) obtaining the expected duration of the channel feature vector of the original channel information, i.e., the first duration; and (2) determining the expected duration of the information recovery accuracy of the AI ​​decoder based on the method for determining the performance information of the second model and the processing capability level of the terminal device, i.e., the fourth duration.

[0182] The method for determining the first duration and the fourth duration can refer to the above content and will not be repeated here.

[0183] Assuming that the target information is the information recovery accuracy of the AI ​​decoder deployed on the network device side, the method for determining the performance information of the second model is: determining the performance information of the second model based on the reference model, and the input of the reference model is the output of the encoder (that is, the channel feature vector of the compressed original channel information), and the output of the reference model is the decoded original channel information (that is, the channel feature vector of the decoded original channel information).

[0184] On this basis, the terminal device determines the duration of the target information, which may at least include: (1) obtaining the expected duration of the channel characteristic vector of the original channel information, i.e., the first duration; (2) compressing the channel characteristic vector of the original channel information through the AI ​​encoder to obtain the expected duration of the channel characteristic vector of the compressed original channel information, i.e., the second duration; (3) inputting the channel characteristic vector of the compressed original channel information into the reference model to obtain the expected duration of the channel characteristic vector of the decoded original channel information, and determining the expected duration of the information recovery accuracy of the AI ​​decoder based on the target accuracy evaluation algorithm and the processing capability level of the terminal device, i.e., the fourth duration.

[0185] Among them, the method for determining the first duration, the second duration and the fourth duration can refer to the above content and will not be repeated here.

[0186] Through the above content, in the scenario where the target information is based on the performance information of the second model, when determining the calculation time, in addition to the various factors described above, such as the type of target information, the processing capability of the terminal device, and the model complexity of the first model, it is also necessary to combine the method for determining the performance information of the second model to determine the calculation time, thereby further improving the accuracy of the calculation time.

[0187] In addition, when the network device intends to determine the performance information of the second model, in addition to directly indicating the target information as the performance information of the second model and instructing the terminal device to report the target information, the network device can also indicate the target information as channel information and compressed channel information, or, channel characteristic vector and compressed channel characteristic vector and other information, so that the network device can determine the performance information of the second model through the second model deployed by itself and the target information.

[0188] The following describes a method for a terminal device to determine the calculation duration of target information by taking the target information as channel information and compressed channel information as an example.

[0189] In some embodiments, assuming that the target information includes channel information of channel M and compressed channel information of channel M, the first model is an AI encoder, the model complexity level of the AI ​​encoder is level 2, and the processing capability level of the terminal device is level A. On this basis, the terminal device determines the duration of the target information by at least: (1) obtaining the estimated duration of the channel information of channel M, i.e., the first duration; and (2) compressing the channel information of channel M through the AI ​​encoder to obtain the estimated duration of the compressed channel information of channel M, i.e., the second duration.

[0190] The channel information of channel M is original information. Therefore, when determining the calculation duration, the terminal device can use the duration of 10 milliseconds corresponding to the original information in the preset type-duration table as the first duration, and according to the model complexity level 2 of the AI ​​encoder and the processing capability level A of the terminal device, determine the second duration as 10 milliseconds in the preset level-duration table, and then determine the calculation duration as 20 milliseconds based on the first duration of 10 milliseconds and the second duration of 10 milliseconds.

[0191] In an optional implementation, when there are multiple pieces of target information, the calculation duration can be determined based on at least the first duration, the second duration, and the number of target information.

[0192] Multiple target information can be indicated by the network device or randomly selected by the terminal device, and this embodiment of the present application does not limit this.

[0193] Specifically, after determining the first duration and the second duration, the terminal device can determine the calculation duration according to the following formula 2: T = (T1 + T2) * n + C2 (Formula 2)

[0194] Wherein, T represents the calculation duration, T1 represents the first duration, T2 represents the second duration, n represents the amount of target information, and C2 represents the bias.

[0195] For example, it is assumed that the preset type-duration table contains two corresponding relationships, namely: the duration corresponding to the original information is 10 milliseconds, and the duration corresponding to the characteristic information is 15 milliseconds. The preset level-duration table contains four corresponding relationships, namely: when the model complexity level of the model is level 1 and the processing capability level of the terminal device is level A, the corresponding duration is 5 milliseconds; when the model complexity level of the model is level 1 and the processing capability level of the terminal device is level B, the corresponding duration is 10 milliseconds; when the model complexity level of the model is level 2 and the processing capability level of the terminal device is level A, the corresponding duration is 10 milliseconds; when the model complexity level of the model is level 2 and the processing capability level of the terminal device is level B, the corresponding duration is 15 milliseconds.

[0196] In some embodiments, it is assumed that the number of target information is 2, where one target information is the channel information of channel M and the compressed channel information of channel M, and the other target information is the channel information of channel B and the compressed channel information of channel B. The first model is an AI encoder, and the model complexity level of the AI ​​encoder is level 2, and the processing capability level of the terminal device is level A.

[0197] On this basis, the terminal device determines the duration of the target information, which may at least include: (1) obtaining the duration of the channel information of channel M, or obtaining the estimated duration of the channel information of channel B, i.e., the first duration; (2) compressing the channel information of channel M through the AI ​​encoder to obtain the estimated duration of the compressed channel information of channel M, or compressing the channel information of channel B through the AI ​​encoder to obtain the estimated duration of the compressed channel information of channel B, i.e., the second duration.

[0198] The channel information of channel M and the channel information of channel B are both original information, and the AI ​​encoder used to compress the channel information of channel M and the channel information of channel B is also the same, that is, the model complexity level of the AI ​​encoder is equal. Therefore, when determining the calculation duration, the terminal device can use the duration of 10 milliseconds corresponding to the original information in the preset type-duration table as the first duration, and based on the AI ​​encoder's model complexity level 2 and the terminal device's processing capability level A, determine the second duration of 10 milliseconds in the preset level-duration table. Then, based on the first duration of 10 milliseconds, the second duration of 10 milliseconds, and the number of original information 2, the above formula 2 is used to determine the calculation duration as 40 milliseconds.

[0199] In other embodiments, it is assumed that the number of target information is 2, wherein one target information is the channel feature vector of channel M and the compressed channel feature vector of channel M, and the other target information is the channel feature vector of channel B and the compressed channel feature vector of channel B, the first model is an AI encoder, and the model complexity level of the AI ​​encoder is level 2, and the processing capability level of the terminal device is level A.

[0200] On this basis, the terminal device determines the duration of the target information, which may at least include: (1) obtaining the duration of the channel characteristic vector of channel M, or obtaining the estimated duration of the channel characteristic vector of channel B, i.e., the first duration; (2) compressing the channel characteristic vector of channel M through an AI encoder to obtain the estimated duration of the compressed channel characteristic vector of channel M, or compressing the channel characteristic vector of channel B through an AI encoder to obtain the estimated duration of the compressed channel characteristic vector of channel B, i.e., the second duration.

[0201] The channel characteristic vector of channel M and the channel characteristic vector of channel B are both characteristic information, and the AI ​​encoder used to compress the channel characteristic vector of channel M and the channel characteristic vector of channel B is also the same, that is, the model complexity level of the AI ​​encoder is equal. Therefore, when determining the calculation duration, the terminal device can use the duration of 15 milliseconds corresponding to the original information in the preset type-duration table as the first duration, and based on the model complexity level 2 of the AI ​​encoder and the processing capability level A of the terminal device, determine the second duration of 10 milliseconds in the preset level-duration table. Then, based on the first duration of 15 milliseconds, the second duration of 10 milliseconds and the number of original information 2, the above formula 1 is used to determine the calculation duration as 50 milliseconds.

[0202] Through the above content, in the scenario where the number of target information is multiple, when determining the calculation time, in addition to the various factors described above, such as the type of target information, the processing capability of the terminal device, and the model complexity of the first model, the calculation time also needs to be determined in combination with the number of target information to further improve the accuracy of the calculation time.

[0203] In an optional embodiment, when determining the first duration, in addition to determining it according to a preset type-duration table, it can also be determined according to the complexity of the algorithm for determining the original information and / or the processing capability of the terminal device, or, it can be determined according to the complexity of the algorithm for determining the characteristic information and / or the processing capability of the terminal device. This embodiment of the present application does not limit this.

[0204] In an optional implementation, the calculation duration may further include generating a duration of the UCI according to the target information.

[0205] Specifically, the terminal device may use the preset duration as the duration for generating the UCI carrying the target information according to the target information.

[0206] According to the above content, when determining the calculation duration, the estimated duration for generating the UCI may also be included, thereby further improving the accuracy of the calculation duration.

[0207] The communication method provided in an embodiment of the present application will be introduced in an interactive manner below. FIG6 is a schematic diagram of an interactive flow of a communication method provided in an embodiment of the present application. The method is implemented by a terminal device and a network device. As shown in FIG6 , the method includes:

[0208] S601: The terminal device sends first information to the network device.

[0209] S601 is described below by taking as examples the case where the first information is information including the processing capability of the terminal device and the case where the first information is information including the capability level corresponding to the processing capability of the terminal device.

[0210] In some embodiments, the terminal device can first obtain its own capabilities, that is, the number of processors it contains. For each processor, the terminal device can obtain the computing speed, storage capacity, and model identification that the processor can process, etc., and then generate information containing its own capabilities, that is, first information, and send the first information to the network device.

[0211] In other embodiments, the terminal device can first obtain its own capabilities, that is, the number of processors it contains. For each processor, the terminal device can obtain the computing speed, storage capacity, and model identification that the processor can process, etc., and then determine the capability level corresponding to its own capabilities based on its own capabilities and a preset capability level table, and generate information containing the capability level corresponding to its own capabilities, that is, first information, and finally send the first information to the network device.

[0212] The preset capability level table may include a correspondence between capabilities and capability levels.

[0213] S602: The terminal device sends third information to the network device.

[0214] The third information is used to indicate the model complexity of one or more models. Specifically, the third information can directly indicate the model complexity of each model, for example, the third information is information containing the model complexity of each model; the third information can also indirectly indicate the model complexity of each model, for example, the third information is information containing the model complexity level of each model. This embodiment of the present application is not limited to this.

[0215] The model complexity of each model can be represented by a preset complexity index or by the model complexity of a preset class of models to which the model belongs. A class of models to which the model belongs may include multiple models for implementing the same type of functionality. For example, assuming that model A belongs to class B, the model complexity of model A can be represented by a preset complexity index or by the model complexity of a preset class of models of class B.

[0216] The preset complexity index may include but is not limited to at least one of a floating point operations per second (FLOPS) index and a multiply-accumulate operations (MACs) index, and the embodiment of the present application does not limit this.

[0217] Multiple models refer to various models that can be used by the terminal device.

[0218] The following will illustrate S502 by taking the model complexity of each model represented by the FLOPS indicator, the third information being information containing the model complexity of each model, and the third information being information containing the model complexity level of each model as examples.

[0219] In some embodiments, the terminal device can first determine the various models that it can use, and then determine the model complexity of each model in turn based on the FLOPS indicator, and generate information containing the model complexity of each model, that is, the third information, and finally send the third information to the network device.

[0220] For the method of determining the model complexity of each model based on the FLOPS indicator, please refer to the relevant technology and will not be repeated here.

[0221] In other embodiments, the terminal device may first determine the various models it can use, and then sequentially determine the model complexity of each model based on the FLOPS metric. For each model, the terminal device may determine the model complexity level of the model based on the model complexity of the model and a preset model complexity-level table, and generate information containing the model complexity level of each model, i.e., third information, and finally send the third information to the network device.

[0222] The preset model complexity-level table may include a correspondence between model complexity and complexity levels.

[0223] S603: The network device sends instruction information to the terminal device.

[0224] The indication information is used to instruct the terminal device to send target information to the network device at the first moment.

[0225] In an optional embodiment, before sending the indication information to the terminal device, the network device can also determine the calculation duration by the method described before Figure 5, and determine the first moment based on the calculation duration to ensure that the terminal device can send the target information to the network device at the first moment.

[0226] S604: The terminal device determines the calculation duration based on at least one of the first information and the second information, and the target information.

[0227] Please refer to S502 for details, which will not be repeated here.

[0228] S605, the terminal device determines whether to discard the target information based on the first moment and the calculated duration; if not, execute S606, and if so, execute S608.

[0229] In an optional implementation, when the time difference between the second moment and the first moment is less than the calculated duration, the terminal device discards the target information, that is, decides not to send the target information to the network device.

[0230] The second moment is determined based on an instruction from the network device.

[0231] In some embodiments, the second time may be determined based on a reference signal indicated by the network device. Accordingly, the terminal device may determine the time of receiving the reference signal as the second time. The terminal device may discard the target information if the time difference between the time of receiving the reference signal and the first time is less than the calculated duration.

[0232] In some other embodiments, the second time may be determined based on the indication information in S603, and accordingly, the terminal device may determine the time of receiving the indication information as the second time. The terminal device may discard the target information if the time difference between the time of receiving the indication information and the first time is less than the calculated duration.

[0233] In other embodiments, the second time may be determined based on indication information other than the indication information in S503.

[0234] S606: The terminal device determines the target information.

[0235] S607: The terminal device sends target information to the network device.

[0236] S608: Abandon the target information.

[0237] After the terminal device determines not to send the target information to the network device based on the first moment and the calculated duration, the terminal device can directly discard the target information.

[0238] In FIG6 above, the third information is sent from the terminal device to the network device. In addition, the third information may also be sent from the network device or a device other than the network device to the terminal device. The following takes the case where the third information is sent from the network device to the terminal device as an example, and provides a schematic diagram of an interaction flow of another communication method provided in an embodiment of the present application, as shown in FIG7. The method includes:

[0239] S701: The terminal device sends first information to the network device.

[0240] S702: The network device sends third information to the terminal device.

[0241] The third information may be sent from the network device to the terminal device, or may be sent from a device other than the terminal device and the network device to the terminal device, and this embodiment of the present application does not limit this.

[0242] S703: The network device sends instruction information to the terminal device.

[0243] The instruction information is used to instruct to send target information to the network device at the first moment.

[0244] S704: The terminal device determines the calculation duration based on at least one of the first information and the second information, and the target information.

[0245] Please refer to S502 for details, which will not be repeated here.

[0246] S705: The terminal device determines whether to discard the target information based on the first moment and the calculated duration. If not, execute S706; if so, execute S708.

[0247] S706: The terminal device determines the target information.

[0248] S707: The terminal device sends target information to the network device.

[0249] S708: Discard the target information.

[0250] The above mainly introduces the solutions provided by the embodiments of the present application from the perspective of the interaction between various network elements. Accordingly, the embodiments of the present application also provide a communication device, which is used to implement the various methods described above. The communication device can be a component of the communication equipment in the above method embodiments. It can be understood that in order to implement the above functions, the communication device includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should easily appreciate that, in combination with the units and algorithm steps of the various examples described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in hardware or in a computer software-driven hardware manner depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0251] In the embodiment of the present application, the communication device can be divided into functional modules according to the above method embodiment. For example, each functional module can be divided according to each function, or two or more functions can be integrated into one processing module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules. It should be understood that the division of modules in the embodiment of the present application is schematic and is only a logical functional division. In actual implementation, there may be other division methods.

[0252] For example, taking the communication device as the communication equipment in the above method embodiment as an example, the communication equipment at least includes the acquisition module 801 and the processing model 802 shown in FIG8 .

[0253] The acquisition module 801 is configured to acquire at least one of the first information and the second information.

[0254] Processing module 802 is configured to determine a calculation duration based on at least one of the first information and the second information, as well as the target information. The calculation duration is the estimated duration for determining the target information. The first information indicates the processing capability of the terminal device. The second information indicates the model complexity of the first model. The first model is used to determine the target information.

[0255] In the embodiments of the present application, the communication device is presented in the form of various functional modules divided in an integrated manner. The "module" here can refer to a specific ASIC, circuit, processor and memory that executes one or more software or firmware programs, integrated logic circuit, and / or other devices that can provide the above functions. In a simple embodiment, those skilled in the art can imagine that the sub-node can take the form of the communication device 400 shown in Figure 4.

[0256] For example, the processor 401 in the communication device 400 shown in FIG4 may call computer-executable instructions stored in the memory 404 to enable the communication device 400 to execute the communication method in the above method embodiment.

[0257] Specifically, the functions / implementation process of the processing module can be implemented by the processor 401 in the communication device 400 shown in FIG4 calling the computer-executable instructions stored in the memory 404. Alternatively, the functions / implementation process of the processing module can be implemented by the processor 401 in the communication device 400 shown in FIG4 calling the computer-executable instructions stored in the memory 404.

[0258] Since the communication device provided in the embodiment of the present application can execute the above-mentioned communication method, the technical effects that can be obtained can be referred to the above-mentioned method embodiment and will not be repeated here.

[0259] It should be understood that one or more of the above modules or units can be implemented by software, hardware, or a combination of the two. When any of the above modules or units is implemented in software, the software exists in the form of computer program instructions and is stored in a memory, and a processor can be used to execute the program instructions and implement the above method flow. The processor can be built into an SoC (system on chip) or an ASIC, or it can be an independent semiconductor chip. In addition to the core used to execute software instructions to perform calculations or processing within the processor, it can further include necessary hardware accelerators, such as field programmable gate arrays (FPGAs), PLDs (programmable logic devices), or logic circuits that implement dedicated logic operations.

[0260] When the above modules or units are implemented in hardware, the hardware can be any one or any combination of a CPU, a microprocessor, a digital signal processing (DSP) chip, a microcontroller unit (MCU), an artificial intelligence processor, an ASIC, a SoC, an FPGA, a PLD, a dedicated digital circuit, a hardware accelerator or a non-integrated discrete device, which can run the necessary software or not rely on the software to execute the above method flow.

[0261] Optionally, an embodiment of the present application further provides a communication device (for example, the communication device may be a chip or a chip system), which includes a processor for implementing the method in any of the above method embodiments. In one possible design, the communication device also includes a memory. The memory is used to store necessary program instructions and data, and the processor can call the program code stored in the memory to instruct the communication device to execute the method in any of the above method embodiments. Of course, the memory may not be in the communication device. When the communication device is a chip system, it may be composed of a chip, or it may include a chip and other discrete devices, which is not specifically limited in the embodiment of the present application.

[0262] In one possible implementation, an embodiment of the present application also provides a computer-readable storage medium, which stores a computer program or instruction. When the computer program or instruction is run on a communication device, the communication device can execute the method described in any of the above method embodiments or any of its implementations.

[0263] In a possible implementation, an embodiment of the present application further provides a distributed training system, which includes the access network device described in the above method embodiment, the core network device described in the above method embodiment, and the terminal device described in the above method embodiment.

[0264] In a possible implementation, an embodiment of the present application further provides a communication method, which includes the method described in any of the above method embodiments or any of its implementations.

[0265] In the above embodiments, all or part of the embodiments may be implemented using software, hardware, firmware, or any combination thereof. When implemented using a software program, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions according to the embodiments of the present application are generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0266] Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. A computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media. Available media can be magnetic media (e.g., floppy disks, hard disks, tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid state drives (SSDs)).

[0267] Although the present application is described herein in conjunction with various embodiments, in the process of implementing the claimed application, those skilled in the art may understand and implement other variations of the disclosed embodiments by reviewing the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude multiple situations. A single processor or other unit may implement several functions listed in the claims. Certain measures are recorded in different dependent claims, but this does not mean that these measures cannot be combined to produce good results.

[0268] Although the present application has been described with reference to specific features and embodiments thereof, it is apparent that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and the drawings are merely illustrative of the present application as defined by the appended claims and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art may make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, the present application is intended to include such modifications and variations as fall within the scope of the claims of the present application and their equivalents.

Claims

1. A communication method, characterized in that: Applied to a communication device, the method includes: Acquire at least one of the first information and the second information; Based on at least one of the first information and the second information, and the target information, the calculation duration is determined; the calculation duration refers to the estimated duration for determining the target information; the first information is used to indicate the processing capability of the terminal device; the second information is used to indicate the model complexity of the first model; the first model is used to determine the target information.

2. The method according to claim 1, characterized in that The calculation duration is determined based on at least a first duration and a second duration; the first duration is determined based on the type of the target information; and the second duration is determined based on the first information and / or the second information.

3. The method according to claim 2, characterized in that The first duration is an estimated duration for determining the original information, or the first duration is an estimated duration for determining the characteristic information; both the original information and the characteristic information are used to determine the target information.

4. The method according to claim 1, wherein In the case where the target information is determined based on multiple original information, the calculation duration is determined based on at least the first duration and the third duration; the first duration is determined based on the type of the target information; the third duration is determined based on the number of the multiple original information, and at least one of the first information and the second information.

5. The method according to claim 1, wherein In a case where the target information is performance information of the second model, the calculation duration is determined based on at least the first duration and the fourth duration; The first duration is determined according to the type of the target information; The fourth duration is determined according to a method for determining the performance information of the second model and the first information.

6. The method according to claim 2, characterized in that In the case that there are multiple pieces of target information, the calculation duration is determined based on at least the first duration, the second duration, and the number of the target information.

7. The method according to any one of claims 1 to 6, characterized in that The calculation duration also includes a duration for generating uplink control information UCI according to the target information.

8. The method according to any one of claims 1 to 7, characterized in that The communication device is a terminal device or a network device.

9. The method according to claim 8, characterized in that The communication device is a terminal device; The method further comprises: Receive instruction information from a network device; the instruction information is used to instruct to send the target information to the network device at a first moment.

10. The method according to claim 9, characterized in that The method further comprises: Based on the first moment and the calculated duration, determine whether to send the target information to the network device.

11. The method according to claim 10, characterized in that The determining, based on the first moment and the calculated duration, whether to send the target information to the network device includes: In a case where a time difference between a second moment and the first moment is less than the calculated duration, the target information is discarded; the second moment is determined based on an instruction of the network device.

12. The method according to claim 11, characterized in that The second moment is determined based on a reference signal indicated by the network device; The method further comprises: The time at which the reference signal is received is determined as the second time.

13. The method according to claim 8, characterized in that The communication device is a terminal device; The method further comprises: The first information is sent to the network device.

14. The method according to claim 8, characterized in that The communication device is a terminal device; The method further comprises: Sending third information to the network device; the third information is used to indicate the model complexity of one or more models.

15. The method according to claim 8, characterized in that The communication device is a terminal device; The method further comprises: Receive third information from the first device; the third information is used to indicate model complexity of the multiple models.

16. The method according to claim 8, characterized in that The communication device is a network device; The method further comprises: Sending indication information to the terminal device; the indication information is used to instruct the terminal device to send the target information to the network device at the first moment.

17. The method according to claim 16, characterized in that The method further comprises: Based on the calculated duration, the first moment is determined.

18. A communication device, characterized in that: The device is located in a communication device, and includes: A processing module, used to determine a calculation duration based on at least one of the first information and the second information, and target information; the calculation duration refers to the estimated duration for determining the target information; the first information is used to indicate the processing capability of the terminal device; the second information is used to indicate the model complexity of the first model; the first model is used to determine the target information.

19. A communication device, characterized in that: The communication device includes at least one processor coupled to at least one memory: The at least one processor is configured to execute a computer program or instruction stored in the at least one memory, so that the access network device executes the communication method according to any one of claims 1 to 17.

20. A computer-readable storage medium, characterized in that The method comprises a program code, which, when executed on a computer or a processor, enables the computer or the processor to execute the communication method according to any one of claims 1 to 17.

Images