Configuration method based on auxiliary information and related device

Abstract

The invention discloses a configuration method based on auxiliary information and a related device, and the method comprises the steps: a first device transmits first auxiliary information to a second device; the first auxiliary information is used for indicating computing resources on at least one time unit; the second device allocates tasks to the first device based on the computing resources on the at least one time unit; the first device receives first indication information from the second device; the first indication information is used for indicating a task allocated to the first equipment by the second equipment based on the computing resources on the at least one time unit. In the embodiment of the invention, the second equipment can perform task allocation for the first equipment based on the computing resources on the at least one time unit, so that the task scheduling at the first equipment side is executed on the time unit with sufficient computing resources, and the task scheduling efficiency is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication and artificial intelligence, and particularly relates to a configuration method based on assistance information and related devices. BACKGROUND

[0002] In order to cope with the vision of future intelligent universal access, intelligentization will further evolve at the wireless network architecture level, and artificial intelligence (AI) will be further deeply integrated with wireless network communication. A common scenario of AI integration with wireless network at the present stage is that the terminal side and the network side cooperatively perform an AI task, and in this scenario, there is frequent interaction of AI data between the terminal and the network. In a New Radio (NR) network, although the protocol defines some contents and manners of assistance information (AI) reported by the terminal, when a specific AI task is involved, the assistance information reported by the terminal in the existing protocol cannot enable the network side to accurately and comprehensively configure the related parameters of the AI task, resulting in low processing efficiency and scheduling efficiency of the AI task. SUMMARY

[0003] Embodiments of the present application provide a configuration method based on assistance information and related devices, which enrich the assistance information reported by the user side, and the network side can efficiently schedule the user side based on the assistance information, which is conducive to improving the processing efficiency of the task.

[0004] In a first aspect, embodiments of the present application provide a configuration method based on assistance information, applied to a first device, and the method comprises:

[0005] sending first assistance information to a second device; the first assistance information is used to indicate computing resources in at least one time unit;

[0006] receiving first indication information from the second device; the first indication information is used to indicate a task allocated by the second device to the first device based on the computing resources in the at least one time unit.

[0007] As can be seen, in the embodiments of the present application, the first device can send the computing resources in the at least one time unit as assistance information to the second device, so that the second device can allocate tasks for the first device based on the computing resources in the at least one time unit, so as to schedule the tasks on the first device side to be executed in the time unit with sufficient computing resources, thereby facilitating the improvement of the scheduling efficiency of the tasks. At the same time, the first device can fully utilize the computing resources in the time unit to execute the allocated tasks, which is conducive to improving the processing efficiency of the tasks.

[0008] In a possible implementation manner, the first assistance information indicates one or more of the following:

[0009] an available load of the computing resource in at least one time unit;

[0010] whether the used computing resource in the at least one time unit is less than or equal to a first load threshold;

[0011] whether the computing resource in the at least one time unit is closed.

[0012] In this implementation, the first device can send the usage of the computing resource in the at least one time unit to the second device as auxiliary information, so that the second device can perform efficient task scheduling for the first device based on the auxiliary information.

[0013] In a possible implementation, the method further includes:

[0014] if the task is allocated in M time units, converging the task in the M time units to N time units; where M is greater than N, and N is greater than or equal to 1;

[0015] sending second auxiliary information to the second device; the second auxiliary information is used to indicate that the first device converges the task in the M time units to the N time units;

[0016] receiving second indication information from the second device; the second indication information is used to indicate the transmission resource configured by the second device for the task based on the N time units.

[0017] In this implementation, the first device can converge the task in fewer time units for execution, and close the computing resource in other time units, thereby facilitating reduction of power consumption of the first device. Based on the operation of converging the task in fewer time units by the first device, the second device can schedule the transmission resource after the task is executed to the task on the first device side, thereby avoiding resource waste caused by scheduling the transmission resource to the first device before the task is executed.

[0018] In a possible implementation, the task is executed by a first neural network model on the first device side; the method further includes:

[0019] if the load of the first device is greater than or equal to a second load threshold, sending third auxiliary information to the second device; the third auxiliary information includes an update delay indication of the first neural network model;

[0020] receiving third indication information from the second device; the third indication information is used to instruct the first device to reduce the monitoring frequency of the first neural network model; the third indication information is generated based on the update delay indication.

[0021] In this implementation, the first device can monitor the local load during inference of the first neural network model, and send the update delay indication as auxiliary information to the second device when the local load is high. Then, the second device can instruct the first device to reduce the monitoring frequency of the first neural network model, thereby reducing the power consumption of the first device.

[0022] In a possible implementation, the method further includes:

[0023] In a case where the communication environment in which the first device is located changes, fourth auxiliary information is sent to the second device; the fourth auxiliary information includes an identifier of a second neural network model that is recommended to be used;

[0024] Fourth indication information is received from the second device; the fourth indication information is used to instruct the first device to start a third neural network model to perform a task; the third neural network model is the second neural network model or a neural network model whose performance difference with the performance of the second neural network model is less than or equal to a performance threshold.

[0025] In this implementation, the first device can send the identifier of the second neural network model that is recommended to be used to the second device in a case where the communication environment changes, and the second device can instruct the first device to start the third neural network model to perform a task based on the auxiliary information, so as to adapt the model to the communication environment, thereby facilitating improvement of the execution efficiency of the task.

[0026] In a possible implementation, the method further includes:

[0027] Fifth auxiliary information is sent to the second device; the fifth auxiliary information indicates one or more of the following: a data collection capability of the first device; a reporting time of collected data recommended by the first device; an activity track of the first device in a future period of time; a data preprocessing capability of the first device;

[0028] Fifth indication information is received from the second device; the fifth indication information indicates one or more of the following: a data collection quantity configured by the second device for the first device based on the data collection capability of the first device; a reporting time of collected data configured by the second device for the first device based on the reporting time of collected data recommended by the first device; a data collection task on an activity track configured by the second device for the first device based on the activity track of the first device in the future period of time; a type of data to be collected configured by the second device for the first device based on the data preprocessing capability of the first device;

[0029] The data collected by the first device is used to perform the task.

[0030] In the implementation, the first device can send the auxiliary information related to data collection to the second device, and the second device can configure the data collection task and the reporting time of the collected data for the first device based on the auxiliary information, which is beneficial to improving the scheduling efficiency of the network side. On the basis of improving the data collection efficiency, the execution efficiency of the task can also be improved.

[0031] In a possible implementation, before sending the first auxiliary information to the second device, the method further includes:

[0032] In a case where the training delay of the first neural network is greater than or equal to the delay threshold and / or the load of the first device when training the first neural network is greater than or equal to the third load threshold, the sixth auxiliary information is sent to the second device; the sixth auxiliary information is used to indicate the suggested adjustment parameter of the first neural network; the first neural network model is constructed based on the first neural network;

[0033] The neural network training parameter configured by the second device for the first device based on the suggested adjustment parameter of the first neural network is received.

[0034] In the implementation, the first device can send the suggested adjustment parameter to the second device during model training, and the second device can configure the neural network training parameter for the first device based on the information sent by the first device, thereby being beneficial to improving the training efficiency of the model and reducing the power consumption of the first device.

[0035] In a possible implementation, the sixth auxiliary information indicates one or more of the following:

[0036] The suggested unit calculation amount when training the first neural network;

[0037] The parameter quantity and network structure of the first neural network;

[0038] The suggestion to use the second neural network for training;

[0039] The suggested parameter format and parameter precision;

[0040] The suggested learning rate;

[0041] In a case where the first neural network is collaboratively trained by the first device and the second device, the time interval in which the first neural network participates in adjacent two rounds of training is suggested;

[0042] The neural network training parameter configured by the second device indicates one or more of the following:

[0043] The unit calculation amount or the adjustment value of the unit calculation amount when training the first neural network;

[0044] The network structure and parameter quantity of the first neural network that need to be adjusted;

[0045] training using the third neural network; the third neural network is the second neural network or a neural network whose scale difference with the second neural network is less than or equal to the scale threshold;

[0046] a learning rate of the first neural network;

[0047] in a case where the first neural network is collaboratively trained by the first device and the second device, a time interval at which the first neural network participates in two adjacent rounds of training.

[0048] In this implementation, the parameters suggested by the first device for adjustment are wide in scope, comprehensive, and highly consistent with AI scenarios, so that the second device can configure fine-grained neural network training parameters for the first device based on these parameters.

[0049] In a second aspect, an embodiment of the present application provides a configuration method based on auxiliary information, applied to a second device, and the method comprises:

[0050] receiving first auxiliary information from a first device; the first auxiliary information is used to indicate computing resources in at least one time unit;

[0051] sending first indication information to the first device; the first indication information is used to indicate tasks allocated by the second device for the first device based on the computing resources in at least one time unit.

[0052] As can be seen, in the embodiment of the present application, the second device can receive the computing resource situation in at least one time unit sent by the first device, so as to allocate tasks for the first device based on the computing resources in at least one time unit, so as to execute the task scheduling of the first device side in the time unit with sufficient computing resources, thereby facilitating the improvement of the scheduling efficiency of the task. At the same time, the first device can fully utilize the computing resources in the time unit to execute the allocated tasks, which is conducive to improving the processing efficiency of the tasks.

[0053] In a possible implementation, the first auxiliary information indicates one or more of the following:

[0054] available load of the computing resources in at least one time unit;

[0055] whether the used computing resources in at least one time unit are less than or equal to a first load threshold;

[0056] whether the computing resources in at least one time unit have been closed.

[0057] In this implementation, the second device can receive the use of the computing resources in at least one time unit sent by the first device, so as to perform efficient task scheduling for the first device based on the auxiliary information.

[0058] In a possible implementation, in the case that the task is allocated on M time units, the method further includes:

[0059] receiving second assistance information from the first device; the second assistance information is used to indicate that the first device aggregates the task on the M time units to N time units; wherein M is greater than N, and N is greater than or equal to 1;

[0060] configuring a transmission resource for the task based on the N time units;

[0061] sending second indication information to the first device; the second indication information is used to indicate the transmission resource.

[0062] In this implementation, the second device can determine, through the second assistance information of the first device, that the first device aggregates the task to be executed on fewer time units, and then the second device can schedule the transmission resource after the task is executed to the task on the first device side based on the operation of the first device aggregating the task on fewer time units, so as to avoid the resource waste caused by scheduling the transmission resource to the first device before the task is executed.

[0063] In a possible implementation, the task is executed by a first neural network model on the first device side; the method further includes:

[0064] receiving third assistance information from the first device; the third assistance information includes an update delay indication of the first neural network model;

[0065] sending third indication information to the first device; the third indication information is used to indicate that the first device reduces the monitoring frequency of the first neural network model; and the third indication information is generated based on the update delay indication.

[0066] In this implementation, the second device can receive the update delay indication sent by the first device when the local load of the first device is high, so as to instruct the first device to reduce the monitoring frequency of the first neural network model, and then reduce the power consumption of the first device.

[0067] In a possible implementation, the method further includes:

[0068] receiving fourth assistance information from the first device; the fourth assistance information includes an identifier of a second neural network model recommended to be used;

[0069] determining a third neural network model based on the identifier of the second neural network model;

[0070] sending fourth indication information to the first device; the fourth indication information is used to instruct the first device to start the third neural network model to execute the task; and the third neural network model is the second neural network model or a neural network model whose performance difference with the second neural network model is less than or equal to a performance threshold.

[0071] In this implementation, the second device can receive the identification of the second neural network model suggested by the first device to use in the case of a change in the communication environment, and the second device can instruct the first device to start the third neural network model to perform the task based on the assistance information, so as to adapt the model to the communication environment, thereby facilitating improvement of the execution efficiency of the task.

[0072] In a possible implementation, the method further includes:

[0073] receiving fifth assistance information from the first device; the fifth assistance information indicates one or more of the following: data collection capability of the first device; reporting time of collected data suggested by the first device; activity track of the first device in a future period of time; data preprocessing capability of the first device;

[0074] sending fifth indication information to the first device; the fifth indication information indicates one or more of the following: data collection quantity configured by the second device for the first device based on the data collection capability of the first device; reporting time of collected data configured by the second device for the first device based on the reporting time of collected data suggested by the first device; data collection task on the activity track configured by the second device for the first device based on the activity track of the first device in the future period of time; and data type to be collected configured by the second device for the first device based on the data preprocessing capability of the first device;

[0075] The data collected by the first device is used to perform the task.

[0076] In this implementation, the second device can receive the data collection related assistance information from the first device, and based on the assistance information, the second device can configure data collection tasks and reporting time of collected data for the first device, thereby facilitating improvement of the scheduling efficiency of the network side. On the basis of improvement of the data collection efficiency, the execution efficiency of the task can also be improved.

[0077] In a possible implementation, before receiving the first assistance information from the first device, the method further includes:

[0078] receiving sixth assistance information from the first device; the sixth assistance information is used to indicate parameters for training the first neural network to suggest adjustment; and the first neural network model is constructed based on the first neural network;

[0079] configuring neural network training parameters for the first device based on the parameters for training the first neural network to suggest adjustment;

[0080] sending the configured neural network training parameters to the first device.

[0081] In this implementation, the first device can send the suggested adjustment parameter to the second device during model training, and the second device can configure the neural network training parameter for the first device based on the information sent by the first device, thereby improving the training efficiency of the model and reducing the power consumption of the first device.

[0082] In a possible implementation, the sixth auxiliary information indicates one or more of the following:

[0083] The suggested unit computation amount during training of the first neural network;

[0084] The parameter quantity and network structure of the first neural network;

[0085] Suggestion to use the second neural network for training;

[0086] The suggested parameter format and parameter precision;

[0087] The suggested learning rate;

[0088] In the case where the first neural network is collaboratively trained by the first device and the second device, the time interval during which the first neural network participates in adjacent two rounds of training is suggested;

[0089] The first device is configured with the neural network training parameter based on the suggested adjustment parameter for training the first neural network, including:

[0090] The unit computation amount or the adjustment value of the unit computation amount during training of the first neural network is configured based on the suggested unit computation amount of the first device;

[0091] And / or the network structure and parameter quantity of the first neural network that need to be adjusted are configured based on the parameter quantity and network structure of the first neural network suggested by the first device;

[0092] And / or the third neural network for training is configured for the first device based on the second neural network suggested by the first device; the third neural network is the second neural network or a neural network whose size difference with the second neural network is less than or equal to a size threshold;

[0093] And / or the learning rate of the first neural network is configured based on the learning rate suggested by the first device;

[0094] And / or in the case where the first neural network is collaboratively trained by the first device and the second device, the time interval during which the first neural network participates in adjacent two rounds of training is configured based on the time interval suggested by the first device.

[0095] In this implementation, the suggested adjustment parameter of the first device is wide-ranging, comprehensive, and highly consistent with the AI scene, so that the second device can configure the first device with fine-grained neural network training parameters based on these parameters.

[0096] In a third aspect, the embodiments of the present application provide a communication device applied to a first device, the device comprising a first transceiver and a first processing unit; the first transceiver is configured to:

[0097] send first auxiliary information to the second device; the first auxiliary information is used to indicate the computing resources in at least one time unit;

[0098] receive first indication information from the second device; the first indication information is used to indicate the task allocated by the second device to the first device based on the computing resources in at least one time unit.

[0099] In a possible implementation, the first auxiliary information indicates one or more of the following:

[0100] the available load of the computing resources in at least one time unit;

[0101] whether the used computing resources in at least one time unit are less than or equal to a first load threshold;

[0102] whether the computing resources in at least one time unit are closed.

[0103] In a possible implementation, the first processing unit is configured to, if the task is allocated in M time units, aggregate the task in the M time units into N time units; wherein M is greater than N, and N is greater than or equal to 1.

[0104] The first transceiver is further configured to send second auxiliary information to the second device; the second auxiliary information is used to indicate that the first device aggregates the task in the M time units into the N time units; receive second indication information from the second device; the second indication information is used to indicate the transmission resources configured by the second device for the task in the N time units.

[0105] In a possible implementation, the task is executed by a first neural network model on the first device side; the first transceiver is further configured to:

[0106] in a case where the load of the first device is greater than or equal to a second load threshold, send third auxiliary information to the second device; the third auxiliary information comprises an update delay indication of the first neural network model; receive third indication information from the second device; the third indication information is used to indicate that the first device reduces the monitoring frequency of the first neural network model; the third indication information is generated based on the update delay indication.

[0107] In a possible implementation, the first transceiver is further configured to:

[0108] in a case where the communication environment of the first device changes, send fourth auxiliary information to the second device; the fourth auxiliary information comprises an identifier of a second neural network model recommended to be used.

[0109] receiving fourth indication information from the second device; the fourth indication information is used to instruct the first device to start a third neural network model to perform the task; the third neural network model is the second neural network model or a neural network model whose performance is less than or equal to a performance threshold value compared with the performance of the second neural network model.

[0110] In a possible implementation, the first transceiver is further configured to:

[0111] sending fifth auxiliary information to the second device; the fifth auxiliary information indicates one or more of the following: a data collection capability of the first device; a recommended reporting time of collected data of the first device; an activity track of the first device in a future period of time; a data preprocessing capability of the first device;

[0112] receiving fifth indication information from the second device; the fifth indication information indicates one or more of the following: a data collection quantity configured by the second device for the first device based on the data collection capability of the first device; a reporting time of collected data configured by the second device for the first device based on the recommended reporting time of collected data of the first device; a data collection task on an activity track configured by the second device for the first device based on the activity track of the first device in the future period of time; a to-be-collected data type configured by the second device for the first device based on the data preprocessing capability of the first device;

[0113] wherein the data collected by the first device is used to perform the task.

[0114] In a possible implementation, the first transceiver is further configured to:

[0115] in a case where a training delay of the first neural network is greater than or equal to a delay threshold value and / or a load of the first device when training the first neural network is greater than or equal to a third load threshold value, sending sixth auxiliary information to the second device; the sixth auxiliary information is used to instruct a parameter recommended for adjustment in training the first neural network; the first neural network model is constructed based on the first neural network;

[0116] receiving, by the first device, a neural network training parameter configured by the second device for the first device based on the parameter recommended for adjustment in training the first neural network.

[0117] In a possible implementation, the sixth auxiliary information indicates one or more of the following:

[0118] a recommended unit computing quantity when training the first neural network;

[0119] a parameter quantity and a network structure of the first neural network;

[0120] a recommendation to use the second neural network for training;

[0121] a parameter format and a parameter precision recommended to be used;

[0122] a learning rate recommended to be used;

[0123] in a case where the first neural network is collaboratively trained by the first device and the second device, a time interval at which the first neural network participates in adjacent two rounds of training is recommended;

[0124] the neural network training parameter configured by the second device indicates one or more of the following:

[0125] a unit computation amount or an adjustment value of the unit computation amount when training the first neural network;

[0126] a network structure and a number of parameters that need to be adjusted of the first neural network;

[0127] training using a third neural network; the third neural network is the second neural network or a neural network whose size difference with the second neural network is less than or equal to a size threshold;

[0128] a learning rate of the first neural network;

[0129] in a case where the first neural network is collaboratively trained by the first device and the second device, a time interval at which the first neural network participates in adjacent two rounds of training.

[0130] It should be understood that, since the method embodiment and the device embodiment are different presentation forms of the same technical concept, the content of the first aspect of the embodiment of the present application should be synchronously adapted to the third aspect of the embodiment of the present application, and the same or similar beneficial effects can be achieved, which will not be repeated here.

[0131] In a fourth aspect, the embodiment of the present application provides a communication device applied to a second device or a chip in the second device, the device comprising a second transceiver unit and a second processing unit; the second transceiver unit is configured to:

[0132] receive first auxiliary information from a first device; the first auxiliary information is used to indicate a computing resource in at least one time unit;

[0133] send first indication information to the first device; the first indication information is used to indicate a task allocated by the second device to the first device based on the computing resource in the at least one time unit.

[0134] In a possible implementation, the first auxiliary information indicates one or more of the following:

[0135] an available load of the computing resource in the at least one time unit;

[0136] whether the used computing resource in the at least one time unit is less than or equal to a first load threshold;

[0137] Whether the computing resource on at least one time unit has been closed.

[0138] In a possible implementation, in a case where the task is allocated on M time units, the second transceiver is further configured to receive second assistance information from the first device; the second assistance information is used to indicate that the first device aggregates the task on the M time units to N time units; where M is greater than N, and N is greater than or equal to 1.

[0139] The second processing unit is configured to configure a transmission resource for the task based on the N time units.

[0140] The second transceiver is further configured to send second indication information to the first device; the second indication information is used to indicate the transmission resource.

[0141] In a possible implementation, the task is executed by a first neural network model on the first device side; the second transceiver is further configured to:

[0142] receive third assistance information from the first device; the third assistance information includes an update delay indication of the first neural network model;

[0143] send third indication information to the first device; the third indication information is used to instruct the first device to reduce a monitoring frequency of the first neural network model; the third indication information is generated based on the update delay indication.

[0144] In a possible implementation, the second transceiver is further configured to receive fourth assistance information from the first device; the fourth assistance information includes an identifier of a second neural network model recommended to be used;

[0145] The second processing unit is further configured to determine a third neural network model based on the identifier of the second neural network model.

[0146] The second transceiver is further configured to send fourth indication information to the first device; the fourth indication information is used to instruct the first device to start the third neural network model to execute the task; the third neural network model is the second neural network model or a neural network model whose performance difference with the second neural network model is less than or equal to a performance threshold.

[0147] In a possible implementation, the second transceiver is further configured to:

[0148] receive fifth assistance information from the first device; the fifth assistance information indicates one or more of the following: a collection capability of the first device for various types of data; a reporting time recommended by the first device for collecting various types of data; an activity track of the first device in a future period of time; a preprocessing capability of the first device for various types of data;

[0149] Sending fifth indication information to the first device; the fifth indication information indicates one or more of the following: a data collection amount configured by the second device for the first device based on the data collection capability of the first device; a reporting time of collected data configured by the second device for the first device based on a reporting time of collected data recommended by the first device; a data collection task on an activity trajectory configured by the second device for the first device based on an activity trajectory of the first device within a period of time in the future; and a type of to-be-collected data configured by the second device for the first device based on the data preprocessing capability of the first device.

[0150] The data collected by the first device is used to perform the task.

[0151] In one possible implementation, the second transceiver unit is further configured to receive sixth auxiliary information from the first device; the sixth auxiliary information is used to indicate parameters recommended for adjustment in training the first neural network; and the first neural network model is constructed based on the first neural network.

[0152] The second processing unit is further configured to configure neural network training parameters for the first device based on the parameters recommended for adjustment during training of the first neural network;

[0153] The second transceiver unit is further configured to send the configured neural network training parameters to the first device.

[0154] In a possible implementation, the sixth auxiliary information indicates one or more of the following:

[0155] The recommended unit computational effort when training the first neural network;

[0156] The number of parameters and network structure of the first neural network;

[0157] It is recommended to use a second neural network for training;

[0158] Recommended parameter format and parameter precision;

[0159] Recommended learning rate;

[0160] When the first neural network is trained collaboratively by the first device and the second device, a time interval between two consecutive rounds of training of the first neural network is recommended;

[0161] The second processing unit is specifically configured to: configure neural network training parameters for the first device based on the parameters recommended for adjustment during training of the first neural network;

[0162] a unit computation amount or an adjusted value of the unit computation amount when training the first neural network based on the unit computation amount configuration suggested by the first device;

[0163] and / or configuring the network structure and number of parameters that need to be adjusted for the first neural network based on the number of parameters and the network structure of the first neural network;

[0164] and / or configuring, based on the learning rate suggested by the first device, the learning rate of the first neural network;

[0165] and / or configuring, based on the learning rate suggested by the first device, the learning rate of the first neural network;

[0166] and / or in the case that the first neural network is collaboratively trained by the first device and the second device, configuring, based on the time interval suggested by the first device, the time interval for the first neural network to participate in training of adjacent two rounds.

[0167] It should be understood that, since the method embodiments and the device embodiments are different presentation forms of the same technical concept, the content of the second aspect of the embodiments of the present application should be adapted to the fourth aspect of the embodiments of the present application synchronously, and the same or similar beneficial effects can be achieved, which will not be repeated here.

[0168] In a fifth aspect, the embodiments of the present application provide a communication device, comprising a processor, a memory, a communication interface, and one or more programs, the one or more programs are stored in the memory and configured to be executed by the processor to cooperate with the communication interface to implement the method in any one of the embodiments of the first aspect or the second aspect.

[0169] In a sixth aspect, the embodiments of the present application provide a chip, comprising: a processor configured to invoke and run a computer program from a memory, so that a device installed with the chip executes the method in any one of the embodiments of the first aspect or the second aspect.

[0170] In a seventh aspect, the embodiments of the present application provide a computer readable storage medium, which stores a computer program for execution by a device, the computer program is executed to implement the method in any one of the embodiments of the first aspect or the second aspect.

[0171] In an eighth aspect, the embodiments of the present application provide a computer program product, when the computer program product is run by a device, so that the device executes the method in any one of the embodiments of the first aspect or the second aspect. BRIEF DESCRIPTION OF DRAWINGS

[0172] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the background art, the drawings needed to be used in the embodiments of the present application or the background art will be described below.

[0173] Figure 1 is a schematic diagram of a UE reporting auxiliary information;

[0174] Figure 2 A schematic diagram of a communication system provided for an embodiment of the present application;

[0175] Figure 3 A flowchart of a configuration method based on assistance information provided for an embodiment of the present application;

[0176] Figure 4 A schematic diagram of a second device assigning AI tasks to a first device provided for an embodiment of the present application;

[0177] Figure 5 A flowchart of another configuration method based on assistance information provided for an embodiment of the present application;

[0178] Figure 6 A schematic diagram of aggregating multiple tasks provided for an embodiment of the present application;

[0179] Figure 7 A flowchart of another configuration method based on assistance information provided for an embodiment of the present application;

[0180] Figure 8 A flowchart of another configuration method based on assistance information provided for an embodiment of the present application;

[0181] Figure 9 A flowchart of another configuration method based on assistance information provided for an embodiment of the present application;

[0182] Figure 10 A flowchart of another configuration method based on assistance information provided for an embodiment of the present application;

[0183] Figure 11 A structural schematic diagram of a communication device provided for an embodiment of the present application;

[0184] Figure 12 A structural schematic diagram of another communication device provided for an embodiment of the present application;

[0185] Figure 13 A structural schematic diagram of a communication device provided for an embodiment of the present application;

[0186] Figure 14 A structural schematic diagram of another communication device provided for an embodiment of the present application. DETAILED DESCRIPTION

[0187] The terms "first", "second", "third", and "fourth" and the like in the description and in the claims of the present application and the accompanying drawings are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. The terms "comprises", "comprising", "includes", "including", "contains", "containing" and any variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, system, product, or apparatus that comprises, includes or contains an item or list of items who does not also include items not expressly listed or other items inherent in such process, method, system, product, or apparatus. The terms "a", "an" and "the" and the like in the context of an embodiment description are to be construed to cover both a single object or a plurality of objects.

[0188] Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase "in an embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily mutually exclusive of other embodiments. It is expressly understood that any of the embodiments described herein can be incorporated in to other embodiments.

[0189] The terms "component", "module", "system", and the like as used herein generally refer to computer-related entities, hardware, software, a combination of both hardware and software, or software in execution. For example, a component can be, but is not limited to being, a process running on a processor, a processor, an object, an executable, a thread of execution, a program, and / or a computer. By way of illustration, both an application running on a device and the device can be a component. One or more components can reside within a process and / or thread of execution and a component can be localized, partially localized, and / or distributed across two or more computers. Also, these components can execute from various computer readable media having various data structures stored thereon. The components can communicate by way of local and / or remote processes such as in accordance with a signal having one or more data packets (e.g., data from one component interacting with another component in a local system, distributed system, and / or across a network such as the Internet with other systems via the signal).

[0190] First, the related terms and the related technical background in the present application are briefly introduced, so as to facilitate the understanding of the skilled in the art.

[0191] (1) Floating Point, FP;

[0192] (2) Integer, INT;

[0193] (3) Long Term Evolution, LTE;

[0194] (4) Machine Learning, ML;

[0195] (5) User Equipment: UE;

[0196] (6) UE assistance information: UAI;

[0197] (7) New Radio: NR;

[0198] (8) Radio Resource Control: RRC;

[0199] (9) Operations, Administration, and Maintenance: OAM;

[0200] (10) Over The Top: OTT.

[0201] In the NR network, the protocol defines a means for the terminal to report assistance information to inform the network side, and after the network side obtains the UE reported assistance information, it can perform subsequent scheduling according to the UE recommended configuration. As shown in the following figure, the overall process is as follows: Figure 1

[0202] 1: The network sends RRC Reconfiguration signaling to the UE;

[0203] 2: The UE reports UE assistance information to the network.

[0204] At present, the standard supports part of the assistance information as follows:

[0205] (1) Delay budget report for adjusting the length of the Discontinuous Reception (DRX) cycle in the connected state;

[0206] (2) Overheating assistance information;

[0207] (3) Suggested DRX parameters to save power;

[0208] (4) Suggested maximum total bandwidth;

[0209] (5) Suggested maximum number of auxiliary carriers.

[0210] ​In existing solutions, the auxiliary information reported by the terminal is usually static, that is, the auxiliary information is reported once and will not be changed subsequently. This type of auxiliary information can be carried in the UE capability message, such as: the maximum number of central processing units (CPUs) of the UE, the maximum number of threads, the maximum memory, the maximum available computing power, etc. These auxiliary information are not comprehensive and static auxiliary information is not applicable to dynamically changing task scenarios. The network side cannot schedule and optimize user-side tasks based on this information. In addition, in the scenario where the terminal and the base station jointly train the neural network model, the terminal can report some current training status information to the base station side through layer 1 signaling or layer 2 signaling, but this is limited to letting the base station know this information. The base station will not take corresponding actions based on this information to perform scheduling optimization.

[0211] In order to overcome the shortcomings of the prior art, the present invention provides a configuration method based on auxiliary information. Figure 2 The communication system shown is implemented as Figure 2 As shown, the communication system includes a terminal device, a network device, and an AI network element, wherein the AI ​​network element can also be an AI module. When an AI network element is introduced into the communication system, the AI ​​network element is an independent network element; when an AI module is introduced into the communication system, the AI ​​module can be a module in a network element, and the corresponding network element can be a terminal device or a network device, etc.

[0212] The communication system can be a wireless communication system of a 5th Generation Mobile Network (5G) satellite communication, short-range communication, etc. It should be noted that the wireless communication system mentioned in the present application includes but is not limited to: a Narrow Band-Internet of Things (NB-IoT) system, a Global System for Mobile Communications (GSM) system, an Enhanced Data rate for GSM Evolution (EDGE) system, a Wideband Code Division Multiple Access (WCDMA) system, a Code Division Multiple Access (CDMA2000) system, a Time Division-Synchronization Code Division Multiple Access (TD-SCDMA) system, an LTE system, and three application scenarios of the next generation 5G mobile communication system.

[0213] The network device in the communication system can be any form of base station (Base Station, BS), including various forms of macro base station, micro base station (also known as small station), relay station, access point, etc. In systems using different wireless access technologies, the names of devices with base station functions may be different, for example, in LTE systems, it is called evolved NodeB (eNB or eNodeB); in the third generation (3rd Generation, 3G) system, it is called Node B, etc. For the convenience of description, all the embodiments of the present application are collectively referred to as network equipment or BS. The terminal equipment in the communication system includes but is not limited to: UE, user unit, user station, mobile station, MS, remote station, remote terminal equipment, mobile terminal equipment, user terminal equipment, wireless communication equipment, user agent, user device, cellular phone, cordless phone, session initiation protocol (SIP) phone, wireless local loop (WLL) station, personal digital assistant (PDA), handheld device with wireless communication function, computing device, processing device connected to wireless modem, vehicle-mounted device, wearable device, terminal equipment in Internet of Things, household appliance, virtual reality device, terminal equipment in future 5G network or terminal equipment in future evolved public land mobile network (PLMN), etc.

[0214] The technical solutions provided by the present application will be described in detail below in combination with specific embodiments.

[0215] Please refer to Figure 3 , Figure 3 The flowchart of a configuration method based on auxiliary information provided by the embodiment of the present application is shown in Figure 3 , which includes steps 301-303:

[0216] 301: The first device sends the first auxiliary information to the second device.

[0217] The first auxiliary information is used to indicate the computing resource (which can also include storage resource in a certain sense, such as: integrated computing and storage resource) in at least one time unit, and is specifically used to indicate one or more of the following:

[0218] A. Available load of the computing resource in at least one time unit.

[0219] B. whether the used computing resources in at least one time unit are less than or equal to a first load threshold. That is, the first device can inform the second device of the time units in which the computing resources are consumed less. The first load threshold can be predefined or configured by the network.

[0220] C. whether the computing resources in at least one time unit are closed. That is, the first device can select which time units to open or close the computing resources according to its own load condition, and can inform the second device of the time units in which the computing resources are closed.

[0221] For example, the auxiliary information in A and B above is usually the auxiliary information in the time units in which the computing resources are opened.

[0222] In this implementation, the first device can send the usage of the computing resources in at least one time unit to the second device as auxiliary information, so that the second device can perform efficient task scheduling for the first device based on the auxiliary information.

[0223] The at least one time unit can be all time units that the first device can count, or can be time units in a certain counting period. The time unit refers to a time period, for example, 1 microsecond can be a time unit, and one or more time slots can also be a time unit, which is a division in the time dimension. The specific division manner is not limited in the embodiments of the application. For example, the time unit can also be referred to as a time unit, a time slice, etc. The first device can be a terminal device, such as a UE, a chip in a terminal device, or a module in a terminal device, such as an AI module, etc.

[0224] 302: The second device assigns a task to the first device based on the computing resources in at least one time unit.

[0225] In the embodiments of the application, the second device can assign a task (or perform task scheduling) to the first device based on the first auxiliary information sent by the first device, for example, assign the task to the time unit with higher available load, assign the task to the time unit in which the computing resources are less than or equal to the first load threshold, or assign the task to the time unit in which the computing resources are not closed.

[0226] The task can be one or more tasks, such as a conventional communication data processing service or one or more AI tasks initiated by the first device. The one or more AI tasks can be autonomously initiated by the first device, can be scheduled to be initiated by other terminal devices, can be scheduled to be initiated by the second device, or can be scheduled to be initiated by the core network / OAM / OTT, etc. After the AI task is initiated, the second device assigns and schedules the first device to execute, for example, Figure 4As shown, the second device can allocate three AI tasks (task 1, task 2, task 3) on two time units.

[0227] The second device includes but is not limited to: an access network device, a base station, a core network device, an independent AI network element, an OAM device, an OTT device, a chip in the above device, or an AI module in the above device.

[0228] 303: The second device sends first indication information to the first device.

[0229] The first indication information is used to indicate the task allocated by the second device for the first device based on the computing resources on at least one time unit, so that the first device can execute the allocated task on the corresponding time unit. If the task is an AI task, it can be a training task or an inference task.

[0230] It can be seen that in the embodiments of the present application, the first device can send the computing resources on at least one time unit to the second device as auxiliary information, so that the second device can allocate tasks for the first device based on the computing resources on at least one time unit, so as to schedule the tasks on the first device side to be executed on the time unit with sufficient computing resources, thereby improving the scheduling efficiency of the tasks. At the same time, the first device can fully utilize the computing resources on the time unit to execute the allocated tasks, which is beneficial to improve the processing efficiency of the tasks.

[0231] Please refer to Figure 5 , Figure 5 Another flowchart of the configuration method based on auxiliary information provided by the embodiments of the present application is shown in Figure 5 , which includes steps 501-507:

[0232] 501: The first device sends first auxiliary information to the second device;

[0233] 502: The second device allocates tasks for the first device based on the computing resources on at least one time unit;

[0234] 503: The second device sends first indication information to the first device;

[0235] Steps 501-503 can correspond to the description of steps 301-303 in Figure 3 .

[0236] If the task is allocated on M time units, then:

[0237] 504: The first device aggregates the tasks on M time units to N time units;

[0238] In the embodiments of the present application, the first device can converge the tasks in M time units to N time units for execution according to the execution of the related tasks in each time unit, as shown in Figure 6 For other time units without tasks, the first device can close the computing resources. For example, when the first device takes a long time to execute the tasks in M time units and cannot meet the reporting time of the processing result of the task required by the second device, the first device can perform the operation of converging the tasks in M time units to N time units. Wherein, M is greater than N, and N is greater than or equal to 1;

[0239] 505: The first device sends second auxiliary information to the second device;

[0240] The second auxiliary information is used to indicate that the first device converges the tasks in M time units to N time units. That is, the first device can send the N time units obtained after converging the tasks to the second device as auxiliary information.

[0241] 506: The second device configures transmission resources for the task based on the N time units;

[0242] In the embodiments of the present application, after the second device learns that the tasks are executed in N time units, it can determine the time when the tasks are completed, so that it can schedule the transmission resources after the completion of the tasks to the first device to upload the related information of the tasks. For example, in the end-edge collaborative AI task, the related information can be intermediate processing results of the end-side model output, end-side computing power, channel state and other parameters.

[0243] In this implementation manner, the first device can converge the tasks in fewer time units for execution and close the computing resources in other time units, thereby facilitating the reduction of the power consumption of the first device. Based on the operation of the first device converging the tasks in fewer time units, the second device can schedule the transmission resources after the completion of the tasks to the first device side, thereby avoiding the waste of resources caused by scheduling the transmission resources to the first device before the completion of the tasks.

[0244] 507: The second device sends second indication information to the first device.

[0245] The second indication information is used to indicate the transmission resources configured by the second device for the task based on the N time units. That is, it informs the first device of the resources used for task transmission.

[0246] It should be understood that in the AI task scenario, the tasks allocated to the first device side are executed by the first neural network model of the first device. Please refer to Figure 7 , Figure 7Another flowchart of the configuration method based on the auxiliary information provided by the embodiments of the present application is shown in FIG. 7, which comprises steps 701-706. Figure 7

[0247] 701: The first device sends the first auxiliary information to the second device.

[0248] 702: The second device allocates the task for the first device based on the computing resources in at least one time unit.

[0249] 703: The second device sends the first indication information to the first device.

[0250] Correspondingly, steps 701-703 can correspond to the description of steps 301-303 in FIG. 3. Figure 3

[0251] If the load of the first device is greater than or equal to the second load threshold, then:

[0252] 704: The first device sends the third auxiliary information to the second device.

[0253] The third auxiliary information comprises the update delay indication of the first neural network model, i.e., the first device can monitor the local load during the inference of the first neural network model, and when the local load is high, the first device can inform the second device that the performance of the first neural network model is sufficient and the local load of the first device is high, and the first neural network model does not need to be frequently updated. The second load threshold can be predefined or configured by the network side.

[0254] 705: The second device generates the third indication information based on the update delay indication.

[0255] The third indication information is used to instruct the first device to reduce the monitoring frequency of the first neural network model. The third indication information can comprise the monitoring parameter of the first neural network model configured by the second device for the first device, such as the monitoring period or the monitoring frequency. Generally, the first device has a monitoring period or a monitoring frequency for the first neural network model before sending the third auxiliary information, and each monitoring will trigger the second device to update the first neural network model, and frequent updates will cause certain consumption of the first device. Therefore, the first device can suggest the second device to reduce or delay the update of the first neural network model, and the second device can configure a lower monitoring frequency or a longer monitoring period for the first device after receiving the third auxiliary information.

[0256] 706: The second device sends the third indication information to the first device.

[0257] Correspondingly, the first device receives the third indication information from the second device, thereby reducing the monitoring frequency of the first neural network model.​​

[0258] In this implementation, the first device can monitor the local load during inference of the first neural network model, and send the update delay indication as auxiliary information to the second device when the local load is high. Then the second device can instruct the first device to reduce the monitoring frequency of the first neural network model, thereby reducing the power consumption of the first device.

[0259] Please refer to Figure 8 , Figure 8 Another flowchart of the configuration method based on auxiliary information provided by the embodiments of the present application is shown in Figure 8 , which includes steps 801-806:

[0260] 801: The first device sends first auxiliary information to the second device;

[0261] 802: The second device allocates tasks for the first device based on the computing resources of at least one time unit;

[0262] 803: The second device sends first indication information to the first device;

[0263] Corresponding descriptions of steps 301-303 in Figure 3 may be referred to for steps 801-803.

[0264] In the case where the communication environment of the first device changes, then:

[0265] 804: The first device sends fourth auxiliary information to the second device;

[0266] Correspondingly, the second device receives the fourth auxiliary information. The fourth auxiliary information includes the identifier of the second neural network model recommended to be used. The communication environment may, for example, be the channel environment, the geographical location, the service type, etc. That is, in the case where the communication environment changes, the first device can recommend the second device to use the second neural network model to perform the task.

[0267] 805: The second device determines a third neural network model based on the identifier of the second neural network model;

[0268] The third neural network model is the second neural network model or a neural network model whose performance difference with the second neural network model is less than or equal to the performance threshold.

[0269] 806: The second device sends fourth indication information to the first device.

[0270] The fourth indication information is used to instruct the first device to start the third neural network model to perform the task. That is, the second device can instruct the first device to start the second neural network model to perform the task, or instruct the first device to start a neural network model with a performance close to that of the second neural network model to perform the task. Specifically, the first device can perform the task using a new model, or continue to perform the task after the first neural network model has completed the task. The second device can issue an identifier of the third neural network model to the first device, or issue the third neural network model, or issue parameters of the third neural network model, and the like.

[0271] In this implementation, when the communication environment changes, the first device can send an identifier of a second neural network model recommended to be used to the second device, and the second device can instruct the first device to start a third neural network model to perform the task based on the auxiliary information, so as to adapt the model to the communication environment, thereby improving the execution efficiency of the task.

[0272] Please refer to Figure 9 , Figure 9 Another flowchart of a configuration method based on auxiliary information provided by the embodiments of the present application is shown in Figure 9 The method comprises steps 901-905:

[0273] 901: The first device sends fifth auxiliary information to the second device;

[0274] The fifth auxiliary information indicates one or more of the following:

[0275] a. Data collection capability of the first device. The data collection capability of the first device includes the collection capability of the first device for various types of data. The division of data types can be based on associated tasks (for example, image recognition requires image data collection, voice recognition requires voice data collection), geographical location of collection, sensor type of collected data, and the like. The data collection capability can be the unit time collection amount or single collection amount of the first device for various types of data, and the like.

[0276] b. Suggested reporting time of collected data by the first device. The reporting time includes one-time reporting time or periodic reporting time. For example, the first device estimates the data collection completion time based on the data amount to be collected configured by the second device and its own data collection capability, and sends a suggested one-time reporting time to the second device based on the time, that is, the first device can suggest a scheduling time of the collected data to the second device. Generally, the scheduling time is later than the data collection completion time of the first device. The first device can also suggest the second device to periodically upload the collected data, for example, the first device can indicate a suggested upload period to the second device.

[0277] c. the activity trajectory of the first device in a future period of time. The activity trajectory can also refer to the activity range of the first device in a future period of time.

[0278] d. the data preprocessing capability of the first device. The data preprocessing capability includes data filtering, data enhancement, data conversion, and the like.

[0279] 902: The second device sends fifth indication information to the first device.

[0280] Correspondingly, the first device receives the fifth indication information. The fifth indication information indicates one or more of the following:

[0281] (1) The data collection amount configured by the second device for the first device based on the data collection capability of the first device. For example, it can be the data collection amount for different types of data.

[0282] (2) The reporting time of the collected data configured by the second device for the first device based on the reporting time of the collected data suggested by the first device. Based on the suggested reporting time sent by the first device, the second device can indicate the one-time reporting time or the periodic reporting time to the first device. For example, based on the one-time reporting time suggested by the first device, the second device configures the one-time reporting time of the collected data for the first device, or based on the periodic reporting time suggested by the first device, the second device configures the periodic reporting time of the collected data for the first device, or based on the one-time reporting time suggested by the first device, the second device configures the periodic reporting time of the collected data for the first device (for example, in a scenario with urgent data demand), or based on the periodic reporting time suggested by the first device, the second device estimates the time when the first device completes the collection and configures the one-time reporting time of the collected data for the first device (for example, without urgent data demand). Exemplarily, the periodic reporting time configured by the second device for the first device can be indicated by the DRX cycle or the Discontinuous Transmission (DTX) cycle configured by the second device.

[0283] (3) The data collection task on the activity trajectory configured by the second device for the first device based on the activity trajectory of the first device in a future period of time. For example, it indicates that the first device collects what kind of data and the data amount of the data in a specific geographic location or geographic range. For the type of collected data, the second device can also comprehensively determine in combination with the data preprocessing capability of the first device.

[0284] (4) The type of data to be collected configured by the second device for the first device based on the data preprocessing capability of the first device. For example, if the data preprocessing capability of the first device is high, it can indicate that it collects data with high complexity (such as meteorological data, etc.), and if the data preprocessing capability of the first device is low, it can indicate that it collects some simple data.

[0285] The data collected by the first device is used to perform a task on the first device side, such as performing model training or model inference.

[0286] 903: The first device sends first auxiliary information to the second device;

[0287] 904: The second device allocates a task for the first device based on the computing resources of at least one time unit;

[0288] 905: The second device sends first indication information to the first device.

[0289] It should be noted that steps 901 and 902 can be performed before steps 903-905, or after steps 903-905, or simultaneously with steps 903-905, and the embodiments of the present application are not limited. Figure 3 corresponding description of steps 301-303 in

[0290] It should be noted that steps 901 and 902 can be performed before steps 903-905, or after steps 903-905, or simultaneously with steps 903-905, and the embodiments of the present application are not limited.

[0291] In this implementation, the first device can send data collection-related auxiliary information to the second device, and the second device can configure data collection tasks and reporting times of collected data for the first device based on the auxiliary information, which is beneficial to improving the scheduling efficiency of the network side. On the basis of improving the data collection efficiency, the execution efficiency of the task can also be improved.

[0292] Please refer to Figure 10 , Figure 10 Another configuration method based on auxiliary information provided by the embodiments of the present application is shown in the flowchart as Figure 10 shown, the method comprises steps 1001-1007:

[0293] 1001: The first device determines that the training time delay of the first neural network is greater than or equal to a time delay threshold and / or the load of the first device when training the first neural network is greater than or equal to a third load threshold;

[0294] The time delay threshold and the third load threshold can be predefined or configured by the network side. The training time delay can be the time delay of one or more rounds of training.

[0295] The first neural network model is constructed based on the first neural network. Specifically, the first neural network can be an initial neural network used to construct the first neural network model. The first device side performs a training task of the first neural network, and in the training process, the first device can monitor the training time delay of the first neural network and / or the load of the first device. If the training time delay of the first neural network is greater than or equal to the time delay threshold and / or the load of the first device when training the first neural network is greater than or equal to the third load threshold, then

[0296] 1002: The first device sends sixth assistance information to the second device;

[0297] The sixth assistance information is used to indicate parameters recommended for adjusting the first neural network during training. For example, the sixth assistance information can indicate one or more of the following:

[0298] 1) The recommended unit amount of calculation during training of the first neural network. For example, it can be the amount of floating-point operations per second.

[0299] 2) The number of parameters and network structure of the first neural network. The network structure can include the number of network layers, the number of neurons, etc.

[0300] 3) It is recommended to use the second neural network for training.

[0301] 4) The recommended parameter format and parameter precision. For example, the parameter format can be floating-point numbers, fixed-point numbers, etc., and the parameter precision can be a specific number of bits or a high or low precision.

[0302] 5) The recommended learning rate. For example, if it is determined to continue training using the first neural network, the learning rate of the first neural network is recommended to be adjusted. In some scenarios, such as when the first device recommends using another neural network for training, the learning rate can also be the learning rate of the replaced neural network.

[0303] 6) In the case where the first neural network is trained by the first device and the second device in cooperation, the time interval at which the first neural network is recommended to participate in adjacent two rounds of training. For example, in the scenario of joint training of a neural network by multiple terminal devices and a base station, each terminal device can recommend the time interval for the base station to participate in training based on its own capability. Specifically, a terminal device with weaker capability can have a lower frequency of participation in training and a relatively longer time interval, while a terminal device with stronger capability can frequently participate in model training and have a relatively shorter time interval.

[0304] For example, the sixth assistance information can also indicate the training method of the first neural network recommended by the first device, such as unilateral training, joint training, etc.

[0305] 1003: The second device configures the neural network training parameters for the first device based on the parameters recommended for adjusting the first neural network during training.

[0306] In the embodiments of the present application, the second device can configure the unit calculation amount or the adjustment value of the unit calculation amount when training the first neural network based on the unit calculation amount suggested by the first device; and / or the second device can configure the network structure and the parameter quantity of the first neural network to be adjusted (such as reducing the number of network layers and the number of parameters) based on the parameter quantity and the network structure of the first neural network; and / or the second device can configure the third neural network for the first device to train based on the second neural network suggested by the first device, the third neural network being the second neural network or a neural network whose size is less than or equal to the size threshold from the size of the second neural network (such as a neural network with a size similar to that of the second neural network); and / or the second device can configure the learning rate of the first neural network based on the learning rate suggested by the first device; and / or in the case that the first neural network is trained by the first device and the second device in cooperation, the second device can configure the time interval at which the first neural network participates in adjacent two rounds of training based on the time interval suggested by the first device (such as: increasing the time interval at which the first device participates in training, or reducing the frequency at which the first device participates in training). Based on the time interval suggested by each first device, the second device can configure a minimum time interval and a maximum time interval.

[0307] 1004: The second device sends the configured neural network training parameters to the first device;

[0308] In the embodiments of the present application, the first device trains the first neural network or the third neural network based on the neural network training parameters configured by the second device, and obtains a first neural network model. After the first neural network model is constructed, the first device can execute related tasks through the first neural network model.

[0309] 1005: The first device sends first auxiliary information to the second device;

[0310] 1006: The second device allocates tasks for the first device based on the computing resources in at least one time unit;

[0311] 1007: The second device sends first indication information to the first device.

[0312] Steps 1005-1007 can correspond to the respective descriptions of steps 301-303 in Figure 3

[0313] In this implementation manner, the first device can send the suggested adjustment parameters to the second device when training the model, and the second device can configure the neural network training parameters for the first device based on the information sent by the first device, thereby facilitating the improvement of the training efficiency of the model and the reduction of the power consumption of the first device. In addition, the parameters suggested by the first device for adjustment are wide-ranging and comprehensive, and have high compatibility with AI scenarios, so that the second device can configure fine-grained neural network training parameters for the first device based on these parameters.​

[0314] It should be noted that Figure 5 、 Figures 7-10 The illustrated embodiments may also be combined in different forms to obtain one or more embodiments, which are all within the protection scope of this application.

[0315] The method of the embodiment of the present application is described above, and the device of the embodiment of the present application is provided below.

[0316] See Figure 11 , Figure 11 This is a schematic diagram of the structure of a communication device provided in an embodiment of the present application. Figure 11 As shown, the device includes a first transceiver unit 1101 and a first processing unit 1102. The first transceiver unit 1101 is used to:

[0317] Sending first auxiliary information to the second device; the first auxiliary information is used to indicate computing resources over at least one time unit;

[0318] Receive first indication information from the second device; the first indication information is used to indicate the task assigned by the second device to the first device based on computing resources over at least one time unit.

[0319] It can be seen that in Figure 11 In the illustrated apparatus, computing resources for at least one time unit can be sent as auxiliary information to a second device, enabling the second device to allocate tasks to the first device based on the computing resources for at least one time unit. This allows the first device to schedule tasks for execution in time units with sufficient computing resources, thereby improving task scheduling efficiency. Furthermore, the first device can fully utilize the computing resources in the time unit to execute the assigned tasks, thereby improving task processing efficiency.

[0320] In a possible implementation, the first auxiliary information indicates one or more of the following:

[0321] the available load of the computing resources over at least one time unit;

[0322] whether the computing resources used in at least one time unit are less than or equal to a first load threshold;

[0323] Whether the compute resource has been shut down for at least one time unit.

[0324] In one possible implementation, the first processing unit 1102 is configured to, if the tasks are allocated to M time units, aggregate the tasks on the M time units into N time units; wherein M is greater than N, and N is greater than or equal to 1;

[0325] The first transceiver unit 1101 is also used to send second auxiliary information to the second device; the second auxiliary information is used to instruct the first device to aggregate tasks on M time units into N time units; and receive second indication information from the second device; the second indication information is used to instruct the second device to configure transmission resources for the task based on N time units.

[0326] In one possible implementation, the task is performed by a first neural network model on the first device side; the first transceiver unit 1101 is further configured to:

[0327] When the load of the first device is greater than or equal to the second load threshold, third auxiliary information is sent to the second device; the third auxiliary information includes an update delay indication of the first neural network model; third indication information is received from the second device; the third indication information is used to instruct the first device to reduce the monitoring frequency of the first neural network model; the third indication information is generated based on the update delay indication.

[0328] In a possible implementation, the first transceiver unit 1101 is further configured to:

[0329] When the communication environment of the first device changes, sending fourth auxiliary information to the second device; the fourth auxiliary information includes an identifier of the second neural network model recommended for use;

[0330] Receive fourth indication information from the second device; the fourth indication information is used to instruct the first device to start a third neural network model to perform a task; the third neural network model is the second neural network model or a neural network model whose performance difference with the performance of the second neural network model is less than or equal to a performance threshold.

[0331] In a possible implementation, the first transceiver unit 1101 is further configured to:

[0332] Sending fifth auxiliary information to the second device; the fifth auxiliary information indicates one or more of the following: data collection capabilities of the first device; a time recommended by the first device for reporting collected data; an activity trajectory of the first device within a future period of time; and a data preprocessing capability of the first device;

[0333] Receive fifth indication information from the second device; the fifth indication information indicates one or more of the following: a data collection amount configured by the second device for the first device based on the data collection capability of the first device; a reporting time of collected data configured by the second device for the first device based on a reporting time of collected data recommended by the first device; a data collection task on an activity trajectory configured by the second device for the first device based on an activity trajectory of the first device within a period of time in the future; and a type of to-be-collected data configured by the second device for the first device based on the data preprocessing capability of the first device.

[0334] The data collected by the first device is used to perform a task.

[0335] In a possible implementation, the first transceiver 1101 is further configured to:

[0336] In a case where the training delay of the first neural network is greater than or equal to the delay threshold and / or the load of the first device when training the first neural network is greater than or equal to the third load threshold, the sixth assistance information is sent to the second device; the sixth assistance information is used to indicate parameters recommended to be adjusted when training the first neural network; the first neural network model is constructed based on the first neural network;

[0337] The neural network training parameters configured by the second device for the first device based on the parameters recommended to be adjusted when training the first neural network are received.

[0338] In a possible implementation, the sixth assistance information indicates one or more of the following:

[0339] A unit of computation recommended when training the first neural network;

[0340] A number of parameters and a network structure of the first neural network;

[0341] A recommendation to use the second neural network for training;

[0342] A parameter format and a parameter precision recommended to use;

[0343] A learning rate recommended;

[0344] In a case where the first neural network is collaboratively trained by the first device and the second device, a time interval at which the first neural network participates in adjacent two rounds of training is recommended;

[0345] The neural network training parameters configured by the second device indicate one or more of the following:

[0346] A unit of computation or an adjustment value of the unit of computation when training the first neural network;

[0347] A network structure and a number of parameters of the first neural network that need to be adjusted;

[0348] Training using the third neural network; the third neural network is the second neural network or a neural network whose size difference with the second neural network is less than or equal to a size threshold;

[0349] A learning rate of the first neural network;

[0350] In a case where the first neural network is collaboratively trained by the first device and the second device, a time interval at which the first neural network participates in adjacent two rounds of training is recommended.

[0351] It should be noted that, Figure 11The implementation of each unit described can also refer to Figures 3 to 10 The corresponding description of the embodiment shown. And, Figure 11 The beneficial effects brought about by the described communication device can be referred to Figures 3 to 10 The corresponding description of the illustrated embodiment will not be repeated here.

[0352] See Figure 12 , Figure 12 This is a structural diagram of another communication device provided in an embodiment of the present application. Figure 12 As shown, the device includes a second transceiver unit 1201 and a second processing unit 1202. The second transceiver unit 1201 is used to:

[0353] Receiving first auxiliary information from a first device; the first auxiliary information is used to indicate computing resources over at least one time unit;

[0354] Sending first indication information to the first device; the first indication information is used to instruct the second device to allocate a task to the first device based on computing resources over at least one time unit.

[0355] It can be seen that in Figure 12 The illustrated apparatus can receive computing resource information for at least one time unit from a first device, thereby allocating tasks to the first device based on the computing resources in the at least one time unit. This allows tasks on the first device to be scheduled for execution in time units with sufficient computing resources, thereby improving task scheduling efficiency. Furthermore, the first device can fully utilize the computing resources in the time unit to execute the assigned tasks, thereby improving task processing efficiency.

[0356] In one possible implementation, the first auxiliary information indicates one or more of the following:

[0357] the available load of the computing resources over at least one time unit;

[0358] whether the computing resources used in at least one time unit are less than or equal to a first load threshold;

[0359] Whether the compute resource has been shut down for at least one time unit.

[0360] In one possible implementation, when the task is allocated to M time units, the second transceiver unit 1201 is further configured to receive second auxiliary information from the first device; the second auxiliary information is used to instruct the first device to aggregate the tasks on the M time units into N time units; where M is greater than N, and N is greater than or equal to 1;

[0361] The second processing unit 1202 is configured to configure the transmission resource for the task based on N time units.

[0362] The second transceiver unit 1201 is further configured to send second indication information to the first device; the second indication information is used to indicate the transmission resource.

[0363] In a possible implementation, the task is executed by a first neural network model on the first device side; the second transceiver unit 1201 is further configured to:

[0364] receive third auxiliary information from the first device; the third auxiliary information comprises an update delay indication of the first neural network model;

[0365] send third indication information to the first device; the third indication information is used to instruct the first device to reduce the monitoring frequency of the first neural network model; the third indication information is generated based on the update delay indication.

[0366] In a possible implementation, the second transceiver unit 1201 is further configured to receive fourth auxiliary information from the first device; the fourth auxiliary information comprises an identifier of a second neural network model recommended to be used;

[0367] The second processing unit 1202 is further configured to determine a third neural network model based on the identifier of the second neural network model;

[0368] The second transceiver unit 1201 is further configured to send fourth indication information to the first device; the fourth indication information is used to instruct the first device to start the third neural network model to execute the task; the third neural network model is the second neural network model or a neural network model whose performance difference with the second neural network model is less than or equal to a performance threshold.

[0369] In a possible implementation, the second transceiver unit 1201 is further configured to:

[0370] receive fifth auxiliary information from the first device; the fifth auxiliary information indicates one or more of the following: a collection capability of the first device for various types of data; a reporting time recommended by the first device for collecting various types of data; an activity track of the first device in a future period of time; a preprocessing capability of the first device for various types of data;

[0371] The fifth indication information is sent to the first device; the fifth indication information indicates one or more of the following: the amount of data collection configured by the second device for the first device based on the data collection capability of the first device; the reporting time of the collected data configured by the second device for the first device based on the reporting time of the collected data recommended by the first device; the data collection task on the activity track configured by the second device for the first device based on the activity track of the first device in the future period of time; and the type of data to be collected configured by the second device for the first device based on the data preprocessing capability of the first device.

[0372] The data collected by the first device is used to perform the task.

[0373] In a possible implementation, the second transceiver 1201 is further configured to receive sixth auxiliary information from the first device; the sixth auxiliary information is used to indicate parameters recommended for training the first neural network; and the first neural network model is constructed based on the first neural network.

[0374] The second processing unit 1202 is further configured to configure the neural network training parameters for the first device based on the parameters recommended for training the first neural network.

[0375] The second transceiver 1201 is further configured to send the configured neural network training parameters to the first device.

[0376] In a possible implementation, the sixth auxiliary information indicates one or more of the following:

[0377] The unit computing amount recommended for training the first neural network;

[0378] The number of parameters and the network structure of the first neural network;

[0379] The second neural network is recommended to be used for training;

[0380] The parameter format and the parameter precision recommended to be used;

[0381] The learning rate recommended;

[0382] In the case where the first neural network is trained by the first device and the second device in cooperation, the time interval recommended for the first neural network to participate in adjacent two rounds of training;

[0383] In the case where the neural network training parameters are configured for the first device based on the parameters recommended for training the first neural network, the second processing unit 1202 is specifically configured to:

[0384] Configure the unit computing amount or the adjustment value of the unit computing amount when training the first neural network based on the unit computing amount recommended by the first device;

[0385] And / or configuring the network structure and the number of parameters of the first neural network based on the number of parameters and the network structure of the first neural network;

[0386] And / or configuring the third neural network for the first device to train based on the second neural network recommended by the first device; the third neural network is the second neural network or a neural network whose scale difference with the second neural network is less than or equal to the scale threshold;

[0387] And / or configuring the learning rate of the first neural network based on the learning rate recommended by the first device;

[0388] And / or in the case that the first neural network is collaboratively trained by the first device and the second device, configuring the time interval of the first neural network participating in the training of adjacent two rounds based on the time interval recommended by the first device.

[0389] It should be noted that, Figure 12 The implementation of each unit described can also correspond to the respective description of the embodiments shown in Figures 3 to 10 And, Figure 12 The beneficial effects brought by the communication device described can refer to the respective description of the embodiments shown in Figures 3 to 10 , which will not be described here.

[0390] Based on the description of the above method embodiments and device embodiments, the embodiments of the present application further provide a communication device. Please refer to Figure 13 , Figure 13 The structure schematic diagram of a communication device provided by the embodiments of the present application, which at least includes a processor 1301, a memory 1302 and a communication interface 1303, the processor 1301, the memory 1302 and the communication interface 1303 are connected with each other through a bus 1304. The communication device can be used to execute the related steps of the configuration method based on the auxiliary information. The communication device can be a terminal device in a wireless communication system, such as the first device in the above method embodiments, etc. The processor 1301 in the communication device is used to read the computer program code stored in the above memory 1302, and execute the method of any one of the embodiments shown in Figures 3 to 10 .

[0391] The memory 1302 includes but is not limited to a random access memory (RAM), a read-only memory (ROM), an erasable programmable read only memory (EPROM), or a compact disc read-only memory (CD-ROM), which is used to store related computer programs and data.

[0392] The processor 1301 can be one or more CPUs, and in the case of one CPU, the CPU can be a single core CPU or a multi-core CPU.

[0393] For example, the processor 1301 in the communication device can be configured to read one or more programs stored in the memory 1302, and perform the following operations:

[0394] sending first assistance information to the second device; the first assistance information is used to indicate the computing resources in at least one time unit;

[0395] receiving first indication information from the second device; the first indication information is used to indicate the tasks allocated by the second device to the first device based on the computing resources in at least one time unit.

[0396] It should be noted that the implementation of each operation can also correspond to the description of the method of any one of the embodiments shown in Figures 3 to 10 .

[0397] It should be noted that although the communication device shown in Figure 13 only includes the processor 1301, the memory 1302, the communication interface 1303 and the bus 1304, in the specific implementation process, those skilled in the art should understand that the communication device also includes other devices necessary for normal operation. At the same time, according to specific needs, those skilled in the art should understand that the communication device can also include hardware devices for realizing other additional functions. In addition, those skilled in the art should understand that the communication device can also only include devices necessary for the implementation of the embodiments of the present application, and does not necessarily include all the devices shown in Figure 13 .

[0398] Please refer to Figure 14 , Figure 14 for another structural schematic diagram of a communication device provided by the embodiments of the present application, which at least includes a processor 1401, a memory 1402 and a communication interface 1403, and the processor 1401, the memory 1402 and the communication interface 1403 are connected with each other through a bus 1404. The communication device can be used to execute the related steps of the configuration method based on assistance information. The communication device can be a network device in a wireless communication system, such as the second device in the above method embodiments. The processor 1401 in the communication device is configured to read the computer program code stored in the memory 1402, and execute the method of any one of the embodiments shown in Figures 3 to 10 .

[0399] The memory 1402 includes, but is not limited to, RAM, ROM, EPROM, or CD-ROM, and is used to store relevant computer programs and data.

[0400] The processor 1401 can be one or more CPUs, and in the case of one CPU, the CPU can be a single core CPU or a multi-core CPU.

[0401] For example, the processor 1401 in the communication device can be used to read one or more programs stored in the memory 1402, and perform the following operations:

[0402] receiving first assistance information from the first device; the first assistance information is used to indicate the computing resources in at least one time unit;

[0403] sending first indication information to the first device; the first indication information is used to indicate the task allocated by the second device to the first device based on the computing resources in at least one time unit.

[0404] It should be noted that the implementation of each operation can also correspond to the description of the method of any one of the embodiments shown in Figures 3 to 10 .

[0405] It should be noted that although the communication device shown in Figure 14 only shows the processor 1401, the memory 1402, the communication interface 1403 and the bus 1404, in the specific implementation process, those skilled in the art should understand that the communication device also includes other devices necessary for normal operation. At the same time, according to specific needs, those skilled in the art should understand that the communication device can also include hardware devices for realizing other additional functions. In addition, those skilled in the art should understand that the communication device can also only include devices necessary for the implementation of the embodiments of the present application, and does not necessarily include all the devices shown in Figure 14 .

[0406] The embodiments of the present application also provide a chip, comprising: a processor, configured to call and run a computer program from a memory, so that the device installed with the chip performs the method described in any one of the above Figures 3 to 10 embodiments. The chip can be a chip in a communication device.

[0407] The embodiments of the present application also provide a computer readable storage medium (Memory), which stores a computer program, and when the computer program is run, the method described in the above Figures 3 to 10The method described in any one of the embodiments. It can be understood that the computer readable storage medium herein can include the built-in storage medium in the device, and of course can also include the extended storage medium supported by the device. The computer readable storage medium provides a storage space which stores an operating system of the device. And, one or more computer programs suitable for being loaded and executed by the processor of the device are also stored in the storage space. It should be noted that the computer readable storage medium herein can be a high-speed RAM, or a non-volatile memory such as at least one disk memory; optionally, it can also be at least one computer readable storage medium located away from the aforementioned processor.

[0408] The embodiment of the present application further provides a computer program product, which comprises computer program codes, Figures 3 to 10 The method flow described in any one of the embodiments is realized.

[0409] In the above embodiments, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0410] It should be understood that the processor mentioned in the embodiments of the present application can be a CPU, and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0411] It should also be understood that the memory referred to in the embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be ROM, Programmable ROM (PROM), EPROM, Electrically EPROM (EEPROM), or flash memory. The volatile memory can be RAM used as an external cache. By way of example and not limitation, many forms of RAM can be used, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous Dynamic RAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DRRAM).

[0412] It should be noted that when the processor is a general processor, a DSP, an ASIC, a FPGA or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, the memory (storage module) is integrated in the processor.

[0413] It should be noted that the memory described herein is intended to include, but not limited to, these and any other suitable types of memory.

[0414] It should be understood that in various embodiments of the present application, the size of the sequence number of each process described above does not mean the order of execution, the execution order of each process should be determined by its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0415] In several embodiments provided by the present application, it should be understood that the disclosed system, device and method can be implemented by other ways. For example, the above-described device embodiments are only exemplary, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point, the coupling or direct coupling or communication connection between the displayed or discussed each other can be through some interface, device or unit indirect coupling or communication connection, which can be electrical, mechanical or other forms.

[0416] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, i.e., may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0417] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit, if implemented in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium.

[0418] In the present application, "at least one" means one or more, and "multiple" means two or more. The "and / or" describes the association relationship between the associated objects, which means that there can be three kinds of relationships, for example, A and / or B, which can represent the following three cases: A exists alone, A and B exist together, and B exists alone, where A and B can be singular or plural. In the textual description of the present application, the character " / ", generally indicates that the front and rear associated objects are in an "or" relationship.

[0419] The steps in the method of the embodiments of the present application can be adjusted, combined and deleted in sequence according to actual needs.

[0420] The modules in the device of the embodiments of the present application can be combined, divided and deleted according to actual needs.

[0421] The above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A configuration method based on assistance information, characterized in that, The method is applied to a first device; the method comprises: sending first auxiliary information to a second device; the first auxiliary information is used to indicate computing resources in at least one time unit; receiving first indication information from the second device; the first indication information is used to indicate a task allocated by the second device for the first device based on the computing resources in the at least one time unit.

2. The method of claim 1, wherein, The first auxiliary information indicates one or more of: available load of the computing resources in the at least one time unit; whether the used computing resources in the at least one time unit are less than or equal to a first load threshold; whether the computing resources in the at least one time unit are closed.

3. The method according to claim 1 or 2, characterized in that, The method further comprises: if the task is allocated in M time units, converging the task in the M time units to N time units; wherein M is greater than N, and N is greater than or equal to 1; sending second auxiliary information to the second device; the second auxiliary information is used to indicate that the first device converges the task in the M time units to N time units; receiving second indication information from the second device; the second indication information is used to indicate transmission resources configured by the second device for the task based on the N time units.

4. The method according to any one of claims 1 to 3, characterized in that, The task is executed by a first neural network model on the first device side; the method further comprises: if the load of the first device is greater than or equal to a second load threshold, sending third auxiliary information to the second device; the third auxiliary information comprises an update delay indication of the first neural network model; receiving third indication information from the second device; the third indication information is used to indicate that the first device reduces the monitoring frequency of the first neural network model; the third indication information is generated based on the update delay indication.

5. The method according to any one of claims 1 to 3, characterized in that, The method further comprises: if the communication environment where the first device is located changes, sending fourth auxiliary information to the second device; the fourth auxiliary information comprises an identifier of a second neural network model recommended to be used; receiving fourth indication information from the second device; the fourth indication information is used to indicate that the first device starts a third neural network model to execute the task; the third neural network model is the second neural network model or a neural network model whose performance difference with the performance of the second neural network model is less than or equal to a performance threshold.

6. The method according to any one of claims 1 to 5, characterized in that, The method further comprises: sending fifth auxiliary information to the second device; the fifth auxiliary information indicates one or more of: data collection capability of the first device; reporting time of collected data recommended by the first device; activity track of the first device in a future period of time; data preprocessing capability of the first device; receiving fifth indication information from the second device; the fifth indication information indicates one or more of the following: a data collection amount configured for the first device by the second device based on a data collection capability of the first device; a reporting time of collected data configured for the first device by the second device based on a reporting time of collected data suggested by the first device; a data collection task on an activity trajectory of the first device in a future period of time configured for the first device by the second device based on the activity trajectory; a type of data to be collected configured for the first device by the second device based on a data preprocessing capability of the first device; wherein the data collected by the first device is used to perform the task.

7. The method of claim 4, wherein, Before sending the first auxiliary information to the second device, the method further comprises: in a case where a training delay of the first neural network is greater than or equal to a delay threshold and / or a load of the first device when training the first neural network is greater than or equal to a third load threshold, sending sixth auxiliary information to the second device; the sixth auxiliary information is used to indicate parameters suggested for adjustment when training the first neural network; the first neural network model is constructed based on the first neural network; receiving neural network training parameters configured for the first device by the second device based on parameters suggested for adjustment when training the first neural network.

8. The method of claim 7, wherein, The sixth auxiliary information indicates one or more of the following: a unit computation amount suggested when training the first neural network; a number of parameters and a network structure of the first neural network; suggested use of a second neural network for training; a parameter format and a parameter precision suggested for use; a learning rate suggested; in a case where the first neural network is collaboratively trained by the first device and the second device, a time interval suggested for the first neural network to participate in adjacent two rounds of training; The neural network training parameters configured by the second device indicate one or more of the following: a unit computation amount or an adjustment value of a unit computation amount when training the first neural network; a network structure and a number of parameters of the first neural network that need to be adjusted; use of a third neural network for training; the third neural network is the second neural network or a neural network whose size difference with the second neural network is less than or equal to a size threshold; a learning rate of the first neural network; in a case where the first neural network is collaboratively trained by the first device and the second device, a time interval for the first neural network to participate in adjacent two rounds of training.

9. A configuration method based on assistance information, characterized in that, applied to a second device; the method comprises: receiving first auxiliary information from a first device; the first auxiliary information is used to indicate a computing resource in at least one time unit; sending first indication information to the first device; the first indication information is used to indicate a task allocated for the first device by the second device based on the computing resource in the at least one time unit.

10. The method of claim 9, wherein, The first auxiliary information indicates one or more of the following: an available load of the computing resource in the at least one time unit; whether the used computing resource in the at least one time unit is less than or equal to a first load threshold; whether the computing resource on the at least one time unit is closed.

11. The method according to claim 9 or 10, characterized in that, in a case where the task is allocated on M time units, the method further comprises: receiving second assistance information from the first device; the second assistance information is used to indicate that the first device converges the task on the M time units to N time units; wherein M is greater than N, and N is greater than or equal to 1; configuring transmission resource for the task based on the N time units; sending second indication information to the first device; the second indication information is used to indicate the transmission resource.

12. The method according to any one of claims 9-11, characterized in that, the task is executed by a first neural network model on the first device side; the method further comprises: receiving third assistance information from the first device; the third assistance information comprises an update delay indication of the first neural network model; sending third indication information to the first device; the third indication information is used to instruct the first device to reduce the monitoring frequency of the first neural network model; the third indication information is generated based on the update delay indication.

13. The method according to any one of claims 9-11, characterized in that, the method further comprises: receiving fourth assistance information from the first device; the fourth assistance information comprises an identifier of a second neural network model recommended to be used; determining a third neural network model based on the identifier of the second neural network model; sending fourth indication information to the first device; the fourth indication information is used to instruct the first device to start the third neural network model to execute the task; the third neural network model is the second neural network model or a neural network model whose performance difference with the second neural network model is less than or equal to a performance threshold.

14. The method according to any one of claims 9 to 13, characterized in that, the method further comprises: receiving fifth assistance information from the first device; the fifth assistance information indicates one or more of the following: data collection capability of the first device; reporting time of collected data recommended by the first device; activity track of the first device in a future period of time; data preprocessing capability of the first device; sending fifth indication information to the first device; the fifth indication information indicates one or more of the following: data collection quantity configured by the second device for the first device based on the data collection capability of the first device; reporting time of collected data configured by the second device for the first device based on the reporting time of collected data recommended by the first device; data collection task on the activity track configured by the second device for the first device based on the activity track of the first device in the future period of time; and data type to be collected configured by the second device for the first device based on the data preprocessing capability of the first device; wherein the data collected by the first device is used to execute the task.

15. The method of claim 12, wherein, before receiving the first assistance information from the first device, the method further comprises: receiving sixth assistance information from the first device; the sixth assistance information is used to indicate that the first neural network is recommended to be adjusted; the first neural network model is constructed based on the first neural network. configuring the first device with neural network training parameters based on the suggested adjusted parameters for training the first neural network; sending the configured neural network training parameters to the first device.

16. The method of claim 15, wherein, The sixth assistance information indicates one or more of: a suggested unit computation amount for training the first neural network; a number of parameters and a network structure of the first neural network; a suggestion to use a second neural network for training; a suggested parameter format and parameter precision; a suggested learning rate; in the case that the first neural network is collaboratively trained by the first device and the second device, a suggested time interval for the first neural network to participate in adjacent two rounds of training; The configuring the first device with neural network training parameters based on the suggested adjusted parameters for training the first neural network comprises: configuring a unit computation amount or an adjustment value of a unit computation amount for training the first neural network based on the suggested unit computation amount by the first device; and / or configuring a network structure and a number of parameters to be adjusted for the first neural network based on the number of parameters and the network structure of the first neural network; and / or configuring a third neural network for the first device to train based on the suggested second neural network by the first device; the third neural network being the second neural network or a neural network whose size is less than or equal to a size threshold from the size of the second neural network; and / or configuring a learning rate for the first neural network based on the suggested learning rate by the first device; and / or in the case that the first neural network is collaboratively trained by the first device and the second device, configuring a time interval for the first neural network to participate in adjacent two rounds of training based on the suggested time interval by the first device.

17. A communications device, characterized by comprising means for performing the method of any one of claims 1-8, or comprising means for performing the method of any one of claims 9-16.

18. A communication device, characterized by comprising a processor, a memory, a communication interface, and one or more programs stored in the memory and configured to be executed by the processor to cooperate with the communication interface to implement the method of any one of claims 1-8 or claims 9-16.

19. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program for a device to execute, the computer program being executed to implement the method of any one of claims 1-8 or claims 9-16.

20. A computer program product, characterised in that, The computer program product, when run on a device, causes the device to perform the method of any one of claims 1-8 or claims 9-16.

Images