WO2023230969A1 - Procédé et appareil de détermination de modèle d'intelligence artificielle, et dispositif de communication et support de stockage - Google Patents

Procédé et appareil de détermination de modèle d'intelligence artificielle, et dispositif de communication et support de stockage Download PDF

Info

Publication number
WO2023230969A1
WO2023230969A1 PCT/CN2022/096704 CN2022096704W WO2023230969A1 WO 2023230969 A1 WO2023230969 A1 WO 2023230969A1 CN 2022096704 W CN2022096704 W CN 2022096704W WO 2023230969 A1 WO2023230969 A1 WO 2023230969A1
Authority
WO
WIPO (PCT)
Prior art keywords
terminal
power consumption
model
capability
base station
Prior art date
Application number
PCT/CN2022/096704
Other languages
English (en)
Chinese (zh)
Inventor
乔雪梅
Original Assignee
北京小米移动软件有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京小米移动软件有限公司 filed Critical 北京小米移动软件有限公司
Priority to PCT/CN2022/096704 priority Critical patent/WO2023230969A1/fr
Priority to CN202280002058.2A priority patent/CN117501777A/zh
Publication of WO2023230969A1 publication Critical patent/WO2023230969A1/fr

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W72/00—Local resource management
    • H04W72/04—Wireless resource allocation

Definitions

  • the present disclosure relates to the field of wireless communication technology but is not limited to the field of wireless communication technology, and in particular, to a method and device for determining an artificial intelligence (Artificial Intelligence, AI) model, communication equipment and storage media.
  • AI Artificial Intelligence
  • the first aspect of the embodiment of the present disclosure provides a method for determining an AI model, which is executed by a terminal.
  • the method includes:
  • the second aspect of the embodiment of the present disclosure provides a method for determining an AI model, which is executed by a base station.
  • the method includes:
  • a sixth aspect of the embodiments of the present disclosure provides a computer storage medium that stores an executable program; after the executable program is executed by a processor, it can realize the AI provided by the first aspect or the second aspect. How to determine the model.
  • the technical solution provided by the embodiments of the present disclosure determines an AI model that matches the power consumption capability of the terminal, where the power consumption capability is used to characterize the computing capability of the terminal under unit power consumption, or the power consumption capability is expressed in terms of To characterize the power consumption of the terminal under a unit calculation amount; the AI model is at least used for the terminal to perform AI inference operations.
  • the matching AI model is determined based on the power consumption capability of the terminal, and the AI model that is more in line with the terminal performance can be used to process services, reducing the situation where the power consumption and complexity of the AI model cannot match the terminal performance, thereby reducing the impact of the AI model on the terminal. Eliminate unnecessary consumption, improve the adaptability of terminals and AI models, and improve the smoothness of terminal business processing based on AI models.
  • Figure 5 is a schematic flowchart of a method for determining an AI model according to an exemplary embodiment
  • Figure 11 is a schematic structural diagram of a device for determining an AI model according to an exemplary embodiment
  • Figure 13 is a schematic structural diagram of a communication device according to an exemplary embodiment.
  • first, second, third, etc. may be used to describe various information in the embodiments of the present disclosure, the information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other.
  • first information may also be called second information, and similarly, the second information may also be called first information.
  • word “if” as used herein may be interpreted as "when” or "when” or "in response to determining.”
  • FIG. 1 shows a schematic structural diagram of a wireless communication system provided by an embodiment of the present disclosure.
  • the wireless communication system is a communication system based on cellular mobile communication technology.
  • the wireless communication system may include: several terminals 11 and several access devices 12.
  • the terminal 11 may be a device of an unmanned aerial vehicle.
  • the terminal 11 may also be a vehicle-mounted device, for example, it may be an on-board computer with a wireless communication function, or a wireless communication device connected to an external on-board computer.
  • the terminal 11 may also be a roadside device, for example, it may be a streetlight, a signal light or other roadside device with wireless communication function.
  • the AI model that determines that the power consumption matches the terminal's power consumption capability may include an AI model that determines that the power consumption matches a power consumption threshold corresponding to the terminal's power consumption capability.
  • it may include an AI model that determines that the power consumption is lower than the power consumption threshold corresponding to the terminal's power consumption capability, or determines that the power consumption is lower than the power consumption threshold corresponding to the terminal's power consumption capability and the difference from the power consumption threshold is greater than a preset value
  • selecting an AI model that matches the terminal's power consumption capability may include: selecting an AI model that matches the terminal's power consumption capability and has the lowest power consumption parameters.
  • sending the model information of the AI model to the base station may include: sending the model parameters and/or model identification of the AI model to the base station.
  • the model parameters can be used by the base station to generate the corresponding AI model
  • the model identifier can be used by the base station to obtain the corresponding AI model.
  • the capability information indicating the terminal's power consumption capability sent to the base station may also include a power consumption threshold indicating the AI model power consumption level supported by the terminal.
  • the power consumption threshold can be used by the base station to determine whether the power consumption level of the AI model matches that of the terminal.
  • Uplink Control Information (UCI);
  • MAC-CE Media Access Control-Control Element
  • S810 Determine an AI model that matches the terminal's power consumption capability, where the power consumption capability is used to characterize the terminal's computing capability under unit power consumption, or the power consumption capability is used to characterize the terminal's power consumption under unit computing volume; AI The model is at least used by the terminal to perform AI inference operations.
  • the types of AI models may include analytical AI models, visual AI models, textual AI models, interactive AI models, etc.
  • the AI inference operation may include AI inference operations in an AI collaboration scenario between the terminal and the base station.
  • the AI collaboration between the terminal and the base station may include CSI compression transmission service processing.
  • the terminal can perform CSI compression services through an AI model, and the compressed CSI is used to report to the base station.
  • the base station can perform CSI decompression through an AI model.
  • the AI model used by the terminal and the AI model used by the base station can be the same model or different models.
  • the AI model determined by the terminal that matches the power consumption capability of the terminal can be the same as the AI model determined by the base station that matches the terminal's power consumption capability, or have the same input-output relationship as the AI model determined by the base station that matches the terminal's power consumption capability. .
  • AI collaboration is performed between the terminal and the base station.
  • the training set used in the training process of the AI model used by the terminal may be the same as the training set used in the training process of the AI model used by the base station.
  • having the same input-output relationship means that when the inputs of two AI models are the same, the outputs are also the same.
  • two AI models have the same hyperparameters and/or have the same number of network layers or network nodes, etc.
  • the input data or output data of the two AI models may also have the same dimensions and length.
  • the AI inference operation may also include AI inference operations in a scenario where there is no AI collaboration between the terminal and the base station. For example, if there is no AI collaboration between the terminal and the base station, it may include DMRS-based channel estimation. Among them, the terminal performs channel estimation based on the AI model and detects data received based on the channel.
  • the base station can also provide the terminal with an AI model determined by the terminal itself. For example, the terminal determines by itself the model identifier of the AI model that matches the power consumption capability, and the base station based on this The model identifier provides the corresponding AI model to the terminal.
  • the model parameters may include the network architecture parameters of the AI model, such as the number of network layers included in the model, the number of nodes in each layer of the network, and hyperparameters.
  • indicating the determined AI model that matches the terminal's power consumption capability to the terminal may include sending model parameters of the trained AI model that matches the terminal's power consumption capability to the terminal.
  • the base station sends the model parameters of the generated and trained AI model that matches the terminal's power consumption capability to the terminal; the model parameters are used by the terminal to obtain the trained AI model.
  • the AI model that matches the terminal's power consumption capability may be that the power consumption level of the AI model matches the terminal's power consumption capability. For example, if the terminal's computing capability under unit power consumption is determined based on the terminal's power consumption capability, then the power consumption generated by applying the AI model to the terminal can be determined based on the terminal's computing capability under unit power consumption and the computing capability of the AI model, and then the power consumption generated by applying the AI model to the terminal can be determined. AI model that matches power consumption to terminal power consumption capabilities.
  • the AI model that determines that the power consumption matches the terminal's power consumption capability may include an AI model that determines that the power consumption matches a power consumption threshold corresponding to the terminal's power consumption capability.
  • it may include an AI model that determines that the power consumption is lower than the power consumption threshold corresponding to the terminal's power consumption capability, or determines that the power consumption is lower than the power consumption threshold corresponding to the terminal's power consumption capability and the difference from the power consumption threshold is greater than a preset value
  • the power consumption threshold can indicate the AI model power consumption level that the terminal can support.
  • the power consumption threshold can be the maximum power consumption or the average power consumption corresponding to the terminal's power consumption capability, etc.
  • determining the AI model whose power consumption is lower than the power consumption threshold corresponding to the terminal's power consumption capability may include determining the AI model with the lowest or highest power consumption among the AI models whose power consumption is lower than the power consumption threshold corresponding to the terminal's power consumption capability. AI model.
  • determining that the proportion of power consumption relative to the power consumption threshold corresponding to the terminal's power consumption capability is lower than the preset proportion of the AI model may include determining the proportion of power consumption relative to the power consumption threshold corresponding to the terminal's power consumption capability.
  • the power consumption threshold can be obtained from the terminal, or can also be determined according to a preset protocol between the terminal and the base station.
  • the AI model obtained by matching the terminal power consumption capability for the terminal to perform business processing can improve the adaptability of the AI model's power consumption level and the terminal's power consumption capability, and reduce the terminal's power consumption capability that cannot support more complex AI models.
  • the terminal can improve the fluency of using the AI model to perform business processing, thereby reducing business processing failures caused by the terminal interrupting the business process due to excessive power consumption of the AI model.
  • the received model information of the AI model determined by the terminal may include model parameters and/or model identification of the AI model, etc., wherein the model parameters may be used for the base station to generate the corresponding AI model, and the model identification may be used for For the base station to obtain the corresponding AI model.
  • the method may further include: training an AI model corresponding to the model information received from the terminal.
  • the training data set used by the base station to train the AI model may be the same as the training data set of the AI model determined by the terminal training terminal.
  • step S810 may include:
  • the capability information indicating the terminal's power consumption capability may also include a power consumption threshold indicating the AI model power consumption level supported by the terminal.
  • the power consumption threshold can be used by the base station to determine whether the power consumption level of the AI model matches that of the terminal.
  • the base station may receive capability information indicating the power consumption capability of the terminal, and indicate the model information or type information of the determined AI model that matches the capability information to the terminal. .
  • determining the AI model that matches the terminal's power consumption capability based on the power consumption parameters may include: determining the AI model that matches the terminal's power consumption capability based on the power consumption parameters of the alternative AI model and parameters such as inference delay and accuracy.
  • AI model represents the time taken by the AI model to complete an AI inference operation, such as indicating the average delay in completing the AI inference operation.
  • the terminal and the base station each need to use an AI model to perform business collaboration processing.
  • the base station may send model information or type information of an AI model that matches the terminal's power consumption capability to the terminal based on the received capability information indicating the power consumption capability.
  • the base station obtains the power consumption capability information provided by the terminal, which helps the base station obtain the power consumption capability of the terminal efficiently and accurately, thereby allowing the terminal and/or base station to more flexibly and intelligently determine the power consumption capability of the terminal in various scenarios such as AI collaboration. Matching AI model.
  • the method further includes:
  • receiving capability information indicating the terminal's power consumption capability may include:
  • the type information may indicate model identifications of multiple AI models belonging to one or more model types.
  • determining the type information of the AI model that matches the terminal's power consumption capability based on the capability information may include: determining the power consumption parameters of the AI model based on the capability information; selecting an AI model whose power consumption parameters match the terminal's power consumption capability. .
  • the power consumption parameters of the AI model can be determined based on the terminal power consumption capability and the AI model computing capability. For example, the power consumption parameters of the AI model are calculated according to the formula FLOPs of AI model/(FLOPS/mW).
  • determining the type information of the AI model that matches the terminal's power consumption capability based on the capability information may include: determining the power consumption parameters of the AI model based on the capability information; selecting an AI whose power consumption parameters are lower than the terminal's power consumption threshold. model, or select an AI model whose power consumption parameters are lower than the power consumption threshold and the difference from the power consumption threshold is greater than the preset value, or select an AI model whose power consumption parameters relative to the power consumption threshold are lower than the preset proportion.
  • the terminal can further select a matching AI model from the corresponding AI model based on the type determined by the base station, thereby further improving the matching between the AI model and the terminal's power consumption capability. nature, and flexibility in AI model selection.
  • determining an AI model that matches the terminal's power consumption capability based on the capability information includes:
  • the AI model selected from multiple candidate AI models based on the capability information that matches the terminal's power consumption capability may be an AI model whose power consumption parameters match the terminal's power consumption capability.
  • the power consumption parameters of the AI model can be determined based on the terminal power consumption capability and the AI model computing capability. For example, the power consumption parameters of the AI model are calculated according to the formula FLOPs of AI model/(FLOPS/mW).
  • selecting an AI model that matches the terminal's power consumption capability among multiple candidate AI models may include: selecting an AI model that matches the terminal's power consumption capability and has the lowest power consumption parameter among multiple candidate AI models. Model.
  • selecting an AI model from multiple candidate AI models that matches the terminal's power consumption capability may include: selecting from multiple candidate AI models that matches the terminal's power consumption capability and the current operating conditions of the terminal. AI model. For example, select an AI model that matches the terminal's power consumption capability and the terminal's current operating parameters, or select an AI model that matches the business type of the terminal's current AI inference operation to be performed, etc.
  • the corresponding relationship between the numerical range and the power consumption capability identifier may include that different numerical ranges correspond to different power consumption capability identifiers, and there is no intersection between different numerical ranges.
  • the value range of the capability parameter FLOPs/mW is [A, B), and the corresponding power consumption capability identifier can be Capability 1, which represents the first type of power consumption capability.
  • the value range of the capability parameter FLOPs/mW is [B, C), and the corresponding power consumption capability identifier can be Capability 2, which represents the second type of power consumption capability.
  • the value range of the capability parameter FLOPs/mW is [C, D), and the corresponding power consumption capability identifier can be Capability 3, which represents the third type of power consumption capability.
  • the value range of the capability parameter FLOPs/mW is [D, F), and the corresponding power consumption capability identifier can be Capability 4, which represents the fourth type of power consumption capability.
  • the base station if the power consumption capability identifier received by the base station indicates that the terminal does not support AI capabilities, it is determined that the terminal capability parameters do not fall within the preset value range. At this time, the base station is not sure about the AI model that matches the terminal's power consumption capability.
  • the terminal capability parameter FLOPS/mW can be determined to be A, B or (A+B)/2, which is used for calculation according to the formula FLOPs of AI model/(FLOPS/mW) Obtain the power consumption parameters of the AI model, and then determine the AI model whose power consumption parameters match the terminal's power consumption capability.
  • the information received by the base station from the terminal can be simplified, and the numerical range of the terminal capability parameter can be characterized through power consumption capability identification or signaling, thereby facilitating the faster determination of the terminal's power consumption capability, reducing the amount of data interaction, and thus improving the determination of matching terminal power consumption. Capability of AI model efficiency.
  • Embodiments of the present disclosure provide an AI model selection solution based on terminal power consumption capabilities, which may be specifically as follows:
  • Scenario 1 For scenarios where the base station delivers and/or deploys an AI model, the solution may include:
  • the terminal reports specific power consumption hardware capabilities, such as reporting the number of floating point operations FLOPs/W per 1W of power consumption or the number of floating point operations FLOPs/mW per 1mW of power consumption.
  • the reporting can be carried out through signaling methods such as terminal capability information, terminal assistance information, MAC-CE, UCI or RRC.
  • Step 2 The base station side calculates the power consumption required to execute an AI model inference, and makes decisions based on this power consumption whether to deliver the model and which AI model to deliver.
  • the base station can calculate the power consumption of different AI models through the following formula: FLOPs of AI model/(FLOPS/mW), where FLOPs of AI model is the floating point operation number of the AI model. Determine which AI model to use based on the power consumption of the AI model, the inference delay of the AI model, and the accuracy of the AI model.
  • FLOPs of AI model is the floating point operation number of the AI model.
  • the power consumption threshold mainly affects the energy consumption of the terminal
  • the power consumption threshold can also be specified by the protocol to constrain the base station side behavior.
  • Scenario 2 For AI collaboration scenarios between the base station side and the terminal side (such as CSI compression), the following model deployment method may exist: both the terminal side and the base station side use the same data set for model training and deployment.
  • solutions may include the following methods:
  • Method 1 Both the base station side and the terminal side determine the AI model or CSI reporting method based on the terminal's hardware power consumption capability.
  • Step 1 Terminal capability reporting: The terminal reports specific power consumption hardware capabilities, such as FLOPs/W or FLOPs/mW, etc.
  • the reporting can be carried out through signaling such as terminal capability information, terminal assistance information, MAC-CE, UCI or RRC.
  • Step 2 Both the base station side and the terminal side calculate the power consumption required to perform an AI model inference, and make decisions based on this power consumption whether to enable the AI model and which AI model to enable.
  • the base station and terminal sides can calculate the power consumption of different AI models through the following formula: FLOPs of AI model/(FLOPS/mW). Compare the power consumption of different AI models to determine which AI model to use. In addition, if the power consumption of all AI models is large, for example, greater than a certain power consumption threshold, the traditional method will be used.
  • Method 2 The terminal side determines which AI model to use based on its own hardware power consumption capability and AI complexity, and indicates the AI model information to the base station.
  • Step 1 Based on its own hardware power consumption capabilities and the complexity of different AI models, the terminal side calculates the power consumption required to perform an AI model inference, and based on this power consumption, decides whether to enable the AI model and which AI model to enable. Waiting for decision.
  • the terminal side can calculate the power consumption of different AI models through the following formula: FLOPs of AI model/(FLOPS/mW). Compare the power consumption of different AI models to determine which AI model to use. In addition, if the power consumption of all AI models is large, for example, greater than a certain power consumption threshold, the traditional method will be used.
  • the reporting can be carried out through signaling methods such as terminal capability information, terminal assistance information, MAC-CE, UCI or RRC.
  • the AI models of all possible modes are numbered in a manner mutually agreed upon by both the base station and the terminal to simplify reporting signaling; or both parties use a simplified model representation method to report the model.
  • Method 3 The base station determines the AI model based on the hardware power consumption capability of the terminal, and issues the determined model type to the terminal again.
  • Step 1 Terminal capability reporting: The terminal reports specific power consumption hardware capabilities, such as FLOPs/W or FLOPs/mW, etc. For example, it can be reported through signaling such as terminal capability information, terminal assistance information, MAC-CE, UCI or RRC.
  • specific power consumption hardware capabilities such as FLOPs/W or FLOPs/mW, etc.
  • signaling such as terminal capability information, terminal assistance information, MAC-CE, UCI or RRC.
  • Step 2 The base station side calculates the power consumption required to perform an AI model inference, and makes decisions based on this power consumption whether to enable the AI model and which AI model to enable.
  • the base station and terminal side can calculate the power consumption of different AI models through the following formula: FLOPs of AI model/(FLOPS/mW). Compare the power consumption of different AI models to determine which AI model to use. In addition, if the power consumption of all AI models is large, for example, greater than a certain power consumption threshold, the traditional method will be used.
  • the power consumption capability of the hardware can be quantified through the power consumption capability identifier.
  • One possible quantification method is shown in the following table :
  • Capability 1 represents the first type of power consumption capability, and the value range of the corresponding power consumption capability FLOPs/mW is [A, B);
  • Capability 2 represents the second type of power consumption capability, and the value range of the corresponding power consumption capability FLOPs/mW is [B, C);
  • Capability 3 represents the third type of power consumption capability, and the corresponding value range of power consumption capability FLOPs/mW is [C, D);
  • Capability 4 represents the fourth type of power consumption capability, corresponding to the power consumption capability FLOPs/mW The numerical range is [D, F).
  • the power consumption capability of the terminal that does not meet any of the above ranges can be implicitly reported to the base station through an indication of "AI capability not supported".
  • an embodiment of the present disclosure provides an AI model determination device, which is applied to a terminal and may include:
  • the terminal determines by itself the AI model that matches the terminal’s power consumption capability
  • the device further includes:
  • the first sending unit is configured to send capability information indicating the power consumption capability of the terminal to the base station; where the model information is sent by the base station based on the capability information.
  • the second sending unit is configured for the base station to perform AI cooperation with the terminal, and to send capability information indicating the terminal's power consumption capability to the base station;
  • a receiving unit configured to receive the type information of the AI model that matches the terminal power consumption capability returned by the base station according to the capability information
  • the first determining unit 110 is specifically configured as:
  • the terminal's power consumption capability select an AI model that matches the terminal's power consumption capability from multiple alternative AI models;
  • the third sending unit is configured to send model information of the AI model to the base station; the model information is used by the base station to determine the AI model selected by the terminal.
  • the third sending unit is specifically configured as:
  • the base station cooperates with the terminal in AI and sends the model information of the AI model to the base station.
  • the first determining unit 110 is specifically configured as:
  • an embodiment of the present disclosure provides an AI model determination device, which is applied to a base station and may include:
  • the second determination unit 210 is configured to determine an AI model that matches the terminal's power consumption capability, where the power consumption capability is used to characterize the terminal's computing capability under unit power consumption, or the power consumption capability is used to characterize the terminal's computing capability under unit power consumption. Power consumption under certain conditions; the AI model is at least used by the terminal to perform AI inference operations.
  • the second determining unit 210 is specifically configured as:
  • the second determining unit 210 is specifically configured as:
  • the device further includes:
  • the fourth sending unit is configured to send model information of the AI model to the terminal.
  • the second determining unit 210 is specifically configured as:
  • the base station cooperates with the terminal in AI and receives capability information indicating the terminal's power consumption capability;
  • the fifth sending unit is configured to determine, based on the capability information, the type information of the AI model that matches the power consumption capability of the terminal; and send the type information to the terminal.
  • the second determining unit 210 is specifically configured as:
  • the second determining unit 210 is specifically configured as:
  • Memory used to store instructions executable by the processor
  • the processor is configured to execute the AI model determination method provided by any of the foregoing technical solutions.
  • the processor may include various types of storage media, which are non-transitory computer storage media that can continue to store information stored thereon after the communication device is powered off.
  • the communication device includes: a terminal or a network element.
  • Figure 12 is a block diagram of a terminal 800 according to an exemplary embodiment.
  • the terminal 800 may be a mobile phone, a computer, a digital broadcast user device, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, or the like.
  • the terminal 800 may include one or more of the following components: a processing component 802, a memory 804, a power supply component 806, a multimedia component 808, an audio component 810, an input/output (I/O) interface 812, a sensor component 814, and communications component 816.
  • Power supply component 806 provides power to various components of terminal 800.
  • Power component 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to terminal 800.
  • the front camera and/or the rear camera may receive external multimedia data.
  • Each front-facing camera and rear-facing camera can be a fixed optical lens system or have a focal length and optical zoom capabilities.
  • Audio component 810 is configured to output and/or input audio signals.
  • the audio component 810 includes a microphone (MIC) configured to receive external audio signals when the terminal 800 is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode.
  • the received audio signal may be further stored in memory 804 or sent via communication component 816 .
  • audio component 810 also includes a speaker for outputting audio signals.
  • Sensor component 814 includes one or more sensors that provide various aspects of status assessment for terminal 800 .
  • the sensor component 814 can detect the open/closed state of the device 800, the relative positioning of components, such as the display and keypad of the terminal 800, and the sensor component 814 can also detect the position change of the terminal 800 or a component of the terminal 800. , the presence or absence of user contact with the terminal 800 , the orientation or acceleration/deceleration of the terminal 800 and the temperature change of the terminal 800 .
  • Sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact.
  • Sensor assembly 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications.
  • the sensor component 814 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.
  • the communication component 816 is configured to facilitate wired or wireless communication between the terminal 800 and other devices.
  • the terminal 800 can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof.
  • the communication component 816 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel.
  • the communications component 816 also includes a near field communications (NFC) module to facilitate short-range communications.
  • NFC near field communications
  • the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.
  • RFID radio frequency identification
  • IrDA infrared data association
  • UWB ultra-wideband
  • Bluetooth Bluetooth
  • the terminal 800 may be configured by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable Gate array (FPGA), controller, microcontroller, microprocessor or other electronic components are implemented for executing the above method.
  • ASICs application specific integrated circuits
  • DSPs digital signal processors
  • DSPDs digital signal processing devices
  • PLDs programmable logic devices
  • FPGA field programmable Gate array
  • controller microcontroller, microprocessor or other electronic components are implemented for executing the above method.
  • a non-transitory computer-readable storage medium including instructions such as a memory 804 including instructions, executable by the processor 820 of the terminal 800 to generate the above method is also provided.
  • the non-transitory computer-readable storage medium may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.
  • an embodiment of the present disclosure shows the structure of a communication device 900.
  • the communication device 900 may be provided as a network side device.
  • the communication device 900 may be the aforementioned base station.
  • communications device 900 includes a processing component 922, which further includes one or more processors, and memory resources represented by memory 932 for storing instructions, such as application programs, executable by processing component 922.
  • the application program stored in memory 932 may include one or more modules, each corresponding to a set of instructions.
  • the processing component 922 is configured to execute instructions to perform any of the above-mentioned methods performed at the base station, for example, at least one of the methods shown in FIGS. 2 to 9 .
  • Communication device 900 may also include a power supply component 926 configured to perform power management of communication device 900, a wired or wireless network interface 950 configured to connect communication device 900 to a network, and an input-output (I/O) interface 958 .
  • the communication device 900 may operate based on an operating system stored in the memory 932, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM or the like.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

Les modes de réalisation de la présente divulgation concernent un procédé et un appareil de détermination de modèle d'IA, un dispositif de communication et un support de stockage. Le procédé de détermination de modèle d'IA est exécuté au moyen d'un terminal, et consiste à : déterminer un modèle d'IA, qui correspond à la capacité de consommation d'énergie d'un terminal, la capacité de consommation d'énergie étant utilisée pour représenter la capacité de fonctionnement du terminal par unité de consommation d'énergie, ou, la capacité de consommation d'énergie étant utilisée pour représenter la consommation d'énergie du terminal par unité de capacité de fonctionnement ; et le modèle d'IA étant au moins utilisé par le terminal pour exécuter une opération d'inférence d'IA (S210).
PCT/CN2022/096704 2022-06-01 2022-06-01 Procédé et appareil de détermination de modèle d'intelligence artificielle, et dispositif de communication et support de stockage WO2023230969A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2022/096704 WO2023230969A1 (fr) 2022-06-01 2022-06-01 Procédé et appareil de détermination de modèle d'intelligence artificielle, et dispositif de communication et support de stockage
CN202280002058.2A CN117501777A (zh) 2022-06-01 2022-06-01 人工智能模型的确定方法及装置、通信设备及存储介质

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2022/096704 WO2023230969A1 (fr) 2022-06-01 2022-06-01 Procédé et appareil de détermination de modèle d'intelligence artificielle, et dispositif de communication et support de stockage

Publications (1)

Publication Number Publication Date
WO2023230969A1 true WO2023230969A1 (fr) 2023-12-07

Family

ID=89026787

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/096704 WO2023230969A1 (fr) 2022-06-01 2022-06-01 Procédé et appareil de détermination de modèle d'intelligence artificielle, et dispositif de communication et support de stockage

Country Status (2)

Country Link
CN (1) CN117501777A (fr)
WO (1) WO2023230969A1 (fr)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190318245A1 (en) * 2016-12-26 2019-10-17 Huawei Technologies Co., Ltd. Method, terminal-side device, and cloud-side device for data processing and terminal-cloud collaboration system
CN111047045A (zh) * 2018-10-12 2020-04-21 中科寒武纪科技股份有限公司 机器学习运算的分配系统及方法
CN111837425A (zh) * 2020-06-10 2020-10-27 北京小米移动软件有限公司 一种接入方法、接入装置及存储介质
CN112204532A (zh) * 2018-08-30 2021-01-08 华为技术有限公司 一种终端对ai任务支持能力的评测方法及终端
CN112329909A (zh) * 2019-08-05 2021-02-05 杭州海康威视数字技术股份有限公司 生成神经网络模型的方法、装置及存储介质
WO2021203437A1 (fr) * 2020-04-10 2021-10-14 Oppo广东移动通信有限公司 Procédé d'attribution de ressources, dispositif, appareil et support d'enregistrement
CN113747462A (zh) * 2020-05-30 2021-12-03 华为技术有限公司 一种信息处理方法及相关设备
CN114064280A (zh) * 2021-11-20 2022-02-18 东南大学 一种多约束下的端边协同推断方法

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190318245A1 (en) * 2016-12-26 2019-10-17 Huawei Technologies Co., Ltd. Method, terminal-side device, and cloud-side device for data processing and terminal-cloud collaboration system
CN112204532A (zh) * 2018-08-30 2021-01-08 华为技术有限公司 一种终端对ai任务支持能力的评测方法及终端
CN111047045A (zh) * 2018-10-12 2020-04-21 中科寒武纪科技股份有限公司 机器学习运算的分配系统及方法
CN112329909A (zh) * 2019-08-05 2021-02-05 杭州海康威视数字技术股份有限公司 生成神经网络模型的方法、装置及存储介质
WO2021203437A1 (fr) * 2020-04-10 2021-10-14 Oppo广东移动通信有限公司 Procédé d'attribution de ressources, dispositif, appareil et support d'enregistrement
CN113747462A (zh) * 2020-05-30 2021-12-03 华为技术有限公司 一种信息处理方法及相关设备
CN111837425A (zh) * 2020-06-10 2020-10-27 北京小米移动软件有限公司 一种接入方法、接入装置及存储介质
CN114064280A (zh) * 2021-11-20 2022-02-18 东南大学 一种多约束下的端边协同推断方法

Also Published As

Publication number Publication date
CN117501777A (zh) 2024-02-02

Similar Documents

Publication Publication Date Title
CN112005565B (zh) 用户设备辅助信息的上报方法及装置、用户设备、存储介质
WO2023173262A1 (fr) Procédé et appareil de traitement d'informations, dispositif de communication et support de stockage
WO2022077393A1 (fr) Procédé et appareil de détermination d'une cellule cible d'un ue, dispositif de communication et support de stockage
WO2022088181A1 (fr) Procédé de communication, appareil de communication et support de stockage
WO2023245576A1 (fr) Procédé et appareil de détermination de modèle d'ia, dispositif de communication et support de stockage
WO2021196216A1 (fr) Procédé et appareil de traitement de radiomessagerie, équipement d'utilisateur, station de base et support de stockage
WO2020258103A1 (fr) Procédé de surveillance, procédé et dispositif d'envoi d'instructions, appareil de communication et support de stockage
US20240057210A1 (en) Terminal control method, information processing method, communication device, and storage medium
WO2023230969A1 (fr) Procédé et appareil de détermination de modèle d'intelligence artificielle, et dispositif de communication et support de stockage
WO2022077232A1 (fr) Procédé et appareil de communication sans fil, dispositif de communication et support de stockage
WO2022147695A1 (fr) Procédé et appareil de sélection d'ue relais, procédé et appareil de traitement d'informations, dispositif et support
WO2022016450A1 (fr) Procédé et appareil de multiplexage de canal logique, dispositif de communication et support de stockage
WO2024055217A1 (fr) Procédé et appareil de détermination d'état relaxé de bfd, dispositif de communication et support de stockage
WO2023201667A1 (fr) Procédé et appareil de communication sans fil, et dispositif de communication et support d'enregistrement
WO2023102701A1 (fr) Procédé et appareil de traitement d'informations, dispositif de communication et support de stockage
WO2023283953A1 (fr) Procédé et appareil de traitement de communication, dispositif de communication, et support de stockage
WO2022261983A1 (fr) Procédé et appareil de traitement d'informations, dispositif de communication, et support de stockage
WO2023082104A1 (fr) Procédé et appareil de déclenchement de mesure d'ue, dispositif de communication et support de stockage
WO2023225822A1 (fr) Procédé et appareil de surveillance, dispositif de communication et support de stockage
WO2023230970A1 (fr) Procédé et appareil de mesure de signal, dispositif de communication et support de stockage
EP4207915A1 (fr) Procédé et appareil de programmation de transmission, dispositif de communication et support de stockage
WO2023155111A1 (fr) Procédé et appareil de traitement d'informations, dispositif de communication et support de stockage
WO2023178488A1 (fr) Procédé et appareil de mesure, dispositif de communication et support d'enregistrement
CN114616848B (zh) 数据流处理方法及装置、通信设备及存储介质
WO2024016349A1 (fr) Procédé et appareil pour fournir un service de détection, dispositif de communication et support d'enregistrement

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 202280002058.2

Country of ref document: CN

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

Ref document number: 22944300

Country of ref document: EP

Kind code of ref document: A1