WO2023103390A1

WO2023103390A1 - Task processing method, task processing apparatus, electronic device and storage medium

Info

Publication number: WO2023103390A1
Application number: PCT/CN2022/106978
Authority: WO
Inventors: 褚振方; 钱正宇; 施恩; 胡鸣人; 袁正雄; 李金麒; 黄悦; 罗阳; 王国彬
Original assignee: 北京百度网讯科技有限公司
Priority date: 2021-12-07
Filing date: 2022-07-21
Publication date: 2023-06-15
Also published as: CN114201278A; CN114201278B

Abstract

The present disclosure relates to the technical field of artificial intelligence, and in particular relate to cloud computing, computer vision and deep learning technologies. Provided are a task processing method, a task processing apparatus, an electronic device and a storage medium. A specific implementation solution comprises: determining device scheduling information, wherein the device scheduling information comprises available resource information of an end-side device and respective resource consumption information of at least one artificial intelligence task; according to the device scheduling information, determining from a server and the end-side device a respective target device for executing the at least one artificial intelligence task; and controlling the target device respectively corresponding to the at least one artificial intelligence task to execute the at least one artificial intelligence task on the basis of an artificial intelligence model and task data which respectively correspond to the at least one artificial intelligence task.

Description

Task processing method, task processing device, electronic device, and storage medium

This application claims the priority of Chinese Patent Application No. 202111487774.8 submitted on December 7, 2021, the contents of which are hereby incorporated by reference.

technical field

The present disclosure relates to the technical field of artificial intelligence, in particular to cloud computing, computer vision and deep learning technologies. Specifically, it relates to a task processing method, a task processing device, electronic equipment, and a storage medium.

Background technique

With the continuous development of artificial intelligence technology, it has been applied in more and more fields. Complex scenes pose challenges to artificial intelligence technology. A complex scenario may refer to a scenario that requires the cooperation of at least one artificial intelligence task to achieve its business requirements.

Contents of the invention

The disclosure provides a task processing method, a task processing device, electronic equipment, and a storage medium.

According to an aspect of the present disclosure, a task processing method is provided, including: determining device scheduling information, wherein the device scheduling information includes available resource information of end-side devices and resource consumption information of at least one artificial intelligence task; according to the above Device scheduling information, determining from the server and the above-mentioned end-side devices the respective target devices for performing the above-mentioned at least one artificial intelligence task; and, controlling the respective target devices corresponding to the above-mentioned at least one artificial intelligence task, based on The artificial intelligence models and task data corresponding to the respective tasks execute at least one of the aforementioned artificial intelligence tasks.

According to another aspect of the present disclosure, a task processing apparatus is provided, including: a first determining module, configured to determine equipment scheduling information, wherein the equipment scheduling information includes available resource information of end-side equipment and at least one artificial intelligence task Respective resource consumption information; a second determining module, configured to determine from the server and the above-mentioned end-side device according to the above-mentioned device scheduling information, the respective target devices for executing the above-mentioned at least one artificial intelligence task; and, an execution module, for controlling The target devices corresponding to the at least one artificial intelligence task respectively execute the at least one artificial intelligence task based on the artificial intelligence model and task data corresponding to the at least one artificial intelligence task.

According to another aspect of the present disclosure, an electronic device is provided, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor , the above-mentioned instructions are executed by the above-mentioned at least one processor, so that the above-mentioned at least one processor can execute the above-mentioned method.

According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions, wherein the above-mentioned computer instructions are used to cause the above-mentioned computer to execute the above-mentioned method.

According to another aspect of the present disclosure, there is provided a computer program product, including a computer program, which implements the above method when executed by a processor.

It should be understood that what is described in this section is not intended to identify key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will be readily understood through the following description.

Description of drawings

The accompanying drawings are used to better understand the present solution, and do not constitute a limitation to the present disclosure. in:

FIG. 1 schematically shows an exemplary system architecture to which a task processing method and device can be applied according to an embodiment of the present disclosure;

FIG. 2 schematically shows a flowchart of a task processing method according to an embodiment of the present disclosure;

Fig. 3 schematically shows an example schematic diagram of a task processing system according to an embodiment of the present disclosure;

Fig. 4 schematically shows a block diagram of a task processing device according to an embodiment of the present disclosure; and

Fig. 5 schematically shows a block diagram of an electronic device suitable for implementing a task processing method according to an embodiment of the present disclosure.

Detailed ways

Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and they should be regarded as exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

Using the cooperation of at least one artificial intelligence task to realize the business requirements of complex scenarios requires the combination of hardware devices. It can be realized by means of end-cloud collaboration. That is, part of the artificial intelligence task in at least one artificial intelligence task is performed by the device, and another part of the artificial intelligence task is performed by the server. The resource information of different end-side devices is quite different. For example, for end-side devices with low computing power, it will be difficult to complete pre-assigned artificial intelligence tasks. For end-side devices with strong computing power, resources will be wasted.

To this end, an embodiment of the present disclosure proposes a task processing solution. That is, determine the available resource information of the end-side device and the respective resource consumption information of at least one artificial intelligence task, and determine the device scheduling information according to the available resource information and the respective resource consumption information of the at least one artificial intelligence task. According to the device scheduling information, the respective target devices for executing at least one artificial intelligence task are determined from the server and the end-side device. Control the target devices corresponding to at least one artificial intelligence task, and execute at least one artificial intelligence task based on the artificial intelligence model and task data corresponding to the at least one artificial intelligence task.

Based on the resource information of the end-side equipment and the resource consumption information of each artificial intelligence task, the target equipment for executing each artificial intelligence task is reasonably configured to realize the full use of end-side equipment and AI tasks on the basis of ensuring the completion of artificial intelligence tasks. The resources of the server improve the resource utilization of the equipment. In addition, the robustness and adaptability of task processing operations are also improved.

Fig. 1 schematically shows an exemplary system architecture to which a task processing method and apparatus can be applied according to an embodiment of the present disclosure.

It should be noted that, what is shown in FIG. 1 is only an example of the system architecture to which the embodiments of the present disclosure can be applied, so as to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other device, system, environment or scenario.

As shown in FIG. 1 , a system architecture 100 according to this embodiment may include a server 101 , an end-side device 102 and a network 103 . The network 103 is used to provide a communication link medium between the server 101 and the end-side device 102 . Network 104 may include various connection types, such as wired and/or wireless communication links, and the like.

The user can use the end-side device 102 to interact with the control device 101 and the server 101 through the network 103 to receive or send messages and the like. Various communication client applications can be installed on the end-side device 102, such as knowledge reading applications, web browser applications, search applications, instant messaging tools, email clients and/or social platform software, etc. (just for example).

The end-side device 102 may be various electronic devices that have a display screen and support web browsing, including but not limited to smartphones, tablet computers, laptop computers, and desktop computers.

For example, the end-side device 102 determines device scheduling information. The device scheduling information includes available resource information of the end-side device 102 and respective resource consumption information of at least one artificial intelligence task. According to the device scheduling information, the respective target devices for executing at least one artificial intelligence task are determined from the server 101 and the end-side device 102 . Each target device is controlled to execute at least one artificial intelligence task using at least one artificial intelligence model corresponding to the target device and task data respectively corresponding to at least one artificial intelligence task.

The server 103 may be various types of servers that provide various services. For example, the server 103 can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system to solve the problem of traditional physical hosts and VPS services (Virtual Private Server, VPS). The management is difficult and the business scalability is weak. The server 103 can also be a server of a distributed system, or a server combined with a block chain.

It should be understood that the numbers of servers, end-side devices, and networks in FIG. 1 are only illustrative. According to implementation requirements, there can be any number of servers, end-side devices and networks.

Fig. 2 schematically shows a flowchart of a task processing method according to an embodiment of the present disclosure.

As shown in FIG. 2, the method 200 includes operations S210-S230.

In operation S210, device scheduling information is determined. The device scheduling information includes available resource information of the end-side device and respective resource consumption information of at least one artificial intelligence task.

In operation S220, according to the device scheduling information, the respective target devices for executing at least one artificial intelligence task are determined from the server and the end-side device.

In operation S230, the target devices respectively corresponding to the at least one artificial intelligence task are controlled, and the at least one artificial intelligence task is executed based on the artificial intelligence model and the task data respectively corresponding to the at least one artificial intelligence task.

According to an embodiment of the present disclosure, the device scheduling information may be used as a basis for determining a target device for performing an artificial intelligence task. The device scheduling information may include available resource information and at least one resource consumption information. The available resource information may refer to resource information that the terminal-side device can currently provide. Resource consumption information may refer to resource information that needs to be consumed to perform an artificial intelligence task. Resource information may include computing resource information and storage resource information. For example, the resource consumption information may be determined according to the computing resource information and storage resource information required to be consumed by the artificial intelligence model corresponding to the artificial intelligence task and the task data. The computing resource information and storage resource information that the artificial intelligence model and task data need to consume can be determined according to the size of the model structure of the artificial intelligence model and the data volume of the task data.

According to an embodiment of the present disclosure, the available resource information may include respective available resource information of at least one resource item. The resource consumption information may include respective resource consumption information of at least one resource item. The resource item may include at least one of the following: a resource item related to a CPU (Central Processing Unit, central processing unit), a resource item related to a GPU (Graphics Processing Unit, graphics processing unit), and a resource item related to memory. The resource item related to the CPU may include at least one of the following: main frequency of the CPU, turbo frequency of the CPU, number of cores of the CPU, number of threads of the CPU, multi-level cache of the CPU, and thermal design power consumption of the CPU. The resource item related to the GPU may include at least one of the following: a core of the GPU, a frequency of the GPU, and a capacity of the GPU. The memory-related resource item may include at least one of the following: memory size and memory frequency.

According to an embodiment of the present disclosure, at least one artificial intelligence task may be multiple artificial intelligence tasks in complex scenarios. There may be an association between multiple artificial intelligence tasks. The artificial intelligence task may include at least one of the following: an image processing task, a text processing task, and an audio processing task. The image processing task may include at least one of the following: an image recognition task, an object detection task, an image classification task, an image segmentation task, an image retrieval task, and the like. The text processing task may include at least one of the following: named entity recognition task, entity relationship extraction task, and translation translation task. The audio processing task may include at least one of the following: an audio recognition task, an audio classification task, and the like. Complex scenarios can refer to scenarios that require the cooperation of multiple artificial intelligence tasks to achieve their business requirements. For example, complex scenes may include long shelf item layer recognition scenes or emergency rescue robot work scenes, etc.

For example, taking the long shelf item layer recognition scenario as an example, the business requirements of the long shelf item layer recognition scenario may include local image, global image, local image recognition, global image recognition, shelf level recognition and item level recognition. A partial image may refer to an image corresponding to a partial shelf. The global image may be obtained by splicing multiple partial images. Correspondingly, multiple artificial intelligence tasks corresponding to the scene of recognizing the number of layers of long-shelf items may include image recognition tasks and object detection tasks. The image recognition task and the target detection task are artificial intelligence tasks with a relationship, that is, the image recognition task is performed first, and then it is determined whether to perform the target detection task according to the recognition result of the image recognition task. For example, the image recognition task is to identify whether there is an item "toothbrush" on the shelf. If the image recognition result shows that there is an item "toothbrush", perform the target detection task to determine the position of the item "toothbrush".

According to an embodiment of the present disclosure, after the device scheduling information is obtained, target devices for executing all artificial intelligence tasks can be determined from the server and the end-side device according to the device scheduling information. For example, the total resource consumption information is obtained according to the respective resource consumption information of at least one artificial intelligence task. According to the available resource information and the total resource consumption information, respective target devices for performing at least one artificial intelligence task are determined.

According to an embodiment of the present disclosure, for each artificial intelligence task, after determining the target device for executing the artificial intelligence task, the target device can be controlled to execute the artificial intelligence using the artificial intelligence model and task data corresponding to the artificial intelligence task. Task.

According to the embodiments of the present disclosure, based on the resource information of the end-side device and the resource consumption information of each artificial intelligence task, the target device for executing each artificial intelligence task is rationally configured, so that on the basis of ensuring that the artificial intelligence task can be completed, Make full use of the resources of end-side devices and servers, and improve the resource utilization of devices. In addition, the robustness and adaptability of task processing operations are also improved.

According to an embodiment of the present disclosure, operation S220 may include the following operations.

According to at least one piece of resource consumption information, total resource consumption information is obtained. In a case where it is determined that the available resource information satisfies the total resource consumption information, the end-side device is determined as a respective target device for executing at least one artificial intelligence task. In a case where it is determined that the available resource information does not meet the total resource consumption information, according to the device scheduling information, the respective target devices for executing at least one artificial intelligence task are determined from the server and the end-side device.

According to an embodiment of the present disclosure, the sum of at least one resource consumption information can be determined to obtain the total resource consumption information. That is, for each resource item in at least one resource item, the resource consumption information corresponding to the resource item in all the resource consumption information is summed to obtain a summation result corresponding to the resource item. In this way, summation results corresponding to all resource items can be obtained. According to summation results corresponding to all resource items, total resource consumption information is obtained.

According to an embodiment of the present disclosure, after obtaining the total resource consumption information, for each resource item in at least one resource item, it may be determined whether the available resource information corresponding to the resource item satisfies the resource consumption information.

For example, if it is determined that the available resource information corresponding to all resource items can satisfy the resource consumption information, it can be determined that the available resource information can satisfy the total resource consumption information. If it is determined that the available resource information corresponding to some or all resource items cannot satisfy the resource consumption information, it may be determined that the available resource information cannot satisfy the total resource consumption information.

For example, if it is determined that the number of target resource items is greater than or equal to the number threshold, it may be determined that the available resource information can satisfy the total resource consumption information. If it is determined that the number of target resource items is less than the number threshold, it may be determined that the available resource information cannot satisfy the total resource consumption information. The target resource item may refer to a resource item whose available resource information can satisfy the total resource consumption information. The number threshold can be configured according to actual business requirements, and is not limited here.

For example, if it is determined that the number of target resource items is less than the number threshold and the non-target resource items are non-main resource items, it may be determined that the available resource information can satisfy the total resource consumption information. The non-target resource item may refer to any resource item in at least one resource item except the target resource item. The resource items can be divided into main resource items and non-main resource items according to the importance of the resource items. The importance of resource items can be determined according to the degree of influence of resource items on the execution of artificial intelligence tasks by end-side devices. For example, the main resource item may include at least one of the following: CPU main frequency, memory size, and memory frequency.

According to an embodiment of the present disclosure, if it is determined that the available resource information can satisfy the total resource consumption information, then the end-side device may be determined as a target device for performing all artificial intelligence tasks. If it is determined that the available resource information cannot satisfy the total resource consumption information, then according to the equipment scheduling information, the target device corresponding to some of the artificial intelligence tasks in at least one artificial intelligence task can be determined as the server, and the target device corresponding to other artificial intelligence tasks can be determined as the server. The device is determined as an end-side device, so that the end-side device and the server cooperate to complete all artificial intelligence tasks.

According to an embodiment of the present disclosure, determining the respective target devices for executing at least one artificial intelligence task from the server and the end-side device according to the device scheduling information may include the following operations.

Determine the target artificial intelligence task according to the equipment scheduling information. The target artificial intelligence task is an artificial intelligence task that the end-side device can perform. The end-side device is determined as the target device for performing the target artificial intelligence task. A server is determined as a target device for performing a first other artificial intelligence task. The first other artificial intelligence task is an artificial intelligence task other than the target artificial intelligence task in at least one artificial intelligence task.

According to an embodiment of the present disclosure, a target artificial intelligence task may be determined from at least one artificial intelligence task according to available resource information and respective resource consumption information of at least one artificial intelligence task. Determine the target device corresponding to the target artificial intelligence task as the end-side device. A target device corresponding to the first other artificial intelligence task is determined as a server. The first other artificial intelligence task may be any artificial intelligence task in at least one artificial intelligence task except the target artificial intelligence task.

According to an embodiment of the present disclosure, the device scheduling information may further include demand priority information of at least one artificial intelligence task.

According to an embodiment of the present disclosure, determining a target artificial intelligence task according to device scheduling information may include the following operations.

The first candidate artificial intelligence task is determined according to the demand priority information of at least one artificial intelligence task. The second candidate artificial intelligence task is determined according to the available resource information of the terminal device and the resource consumption information of the first candidate artificial intelligence task. In a case where it is determined that the available resource information of the device is consistent with the resource consumption information of the second candidate artificial intelligence task, the second candidate artificial intelligence task is determined as the target artificial intelligence task. In a case where it is determined that the available resource information of the end-side device is greater than the resource consumption information of the second candidate artificial intelligence task, a third candidate artificial intelligence task is determined. The third candidate artificial intelligence task is an artificial intelligence task other than the first candidate artificial intelligence task in at least one artificial intelligence task. A target artificial intelligence task is determined according to the second candidate artificial intelligence task and the third candidate artificial intelligence task.

According to an embodiment of the present disclosure, the requirement priority information may represent the priority of the business requirement. The requirement priority information may be determined according to the urgency of the business requirement. If the urgency of the artificial intelligence task is higher, the priority of the business requirement represented by the requirement priority information is higher. The urgency can be determined according to at least one of the number of times the AI task is utilized within a predetermined time period and the number of times the AI task is invoked by other AI tasks within the predetermined time period.

According to an embodiment of the present disclosure, the first candidate artificial intelligence task may refer to an artificial intelligence task whose requirement priority information satisfies a predetermined priority condition. For example, at least one artificial intelligence task may be sorted according to the priority of business requirements represented by the demand priority information to obtain a ranking result, and the first candidate artificial intelligence task may be determined from the at least one artificial intelligence task according to the ranking result. The sorting can be sorted according to the priority of the business needs from high to low or according to the priority of the business needs from low to high.

According to an embodiment of the present disclosure, after the first candidate artificial intelligence task is determined, the second candidate artificial intelligence task that the end-side device can execute can be determined according to the available resource information of the end-side device and the resource consumption information of the first candidate artificial intelligence task. Task. That is, to achieve the purpose of performing artificial intelligence tasks with higher priority in business requirements by end-side devices, and improve task execution efficiency.

According to an embodiment of the present disclosure, if it is determined that the available resource information of the device-side device is consistent with the resource consumption information of the second candidate artificial intelligence task, it may indicate that the device-side device cannot perform other artificial intelligence tasks. In this case, the second candidate artificial intelligence task may be determined as the target artificial intelligence task.

According to an embodiment of the present disclosure, if it is determined that the available resource information of the end-side device is greater than the resource consumption information of the second candidate artificial intelligence task, it can be explained that the end-side device can perform other artificial intelligence tasks in addition to the second candidate artificial intelligence task. Smart tasks. Other artificial intelligence tasks may be third candidate artificial intelligence tasks. The second candidate artificial intelligence task and the third candidate artificial intelligence task may be determined as target artificial intelligence tasks.

According to an embodiment of the present disclosure, operation S230 may include the following operations.

Determine the correlation between at least one artificial intelligence task. Determine the task execution sequence according to the association relationship. According to the task execution order, control the target device corresponding to at least one artificial intelligence task, and execute at least one artificial intelligence task based on the artificial intelligence model and task data corresponding to the at least one artificial intelligence task.

According to an embodiment of the present disclosure, the association relationship may include having an association relationship and not having an association relationship. A task execution sequence of at least one artificial intelligence task may be determined according to the relationship between at least one artificial intelligence task. Then, according to the task execution sequence, control the target device corresponding to at least one artificial intelligence task, and execute the artificial intelligence task corresponding to the target device.

According to an embodiment of the present disclosure, the order of task execution may include parallel execution and serial execution. If there is no correlation between the artificial intelligence task and other artificial intelligence tasks, the artificial intelligence task can be executed in parallel with other artificial intelligence tasks. If there is an association relationship between the artificial intelligence task and other artificial intelligence tasks, each artificial intelligence task can be executed sequentially according to the task execution sequence determined by the association relationship.

According to an embodiment of the present disclosure, the above task processing method may further include the following operations.

Respective execution priority information of at least one artificial intelligence task is determined.

According to an embodiment of the present disclosure, determining the task execution order according to the association relationship may include the following operations.

Determine the task execution order according to the association relationship and at least one piece of execution priority information.

According to an embodiment of the present disclosure, the execution priority information may represent the priority of executing the artificial intelligence task. The execution priority information of each artificial intelligence task in at least one artificial intelligence task may be determined according to business requirements. Then, the task execution sequence is determined according to the association relationship and at least one piece of execution priority information.

For example, for an artificial intelligence task, if the association relationship is satisfied, if it is determined that the artificial intelligence task has the highest priority according to the execution priority information of the artificial intelligence task, the artificial intelligence task can be executed first.

According to an embodiment of the present disclosure, controlling the target device corresponding to at least one artificial intelligence task, and executing at least one artificial intelligence task based on the artificial intelligence model and task data corresponding to the at least one artificial intelligence task may include the following operations.

When it is determined that each target device for executing at least one artificial intelligence task includes a cloud server, a task execution request is sent to the server, so that the server executes the artificial intelligence task corresponding to the server in response to receiving the task execution request. The task execution request includes task information, and the task information includes an artificial intelligence task identifier and task data corresponding to the artificial intelligence task identifier, and the artificial intelligence task identifier is used to determine an artificial intelligence model corresponding to the artificial intelligence task identifier.

According to an embodiment of the present disclosure, a task execution request may refer to a request for requesting a server to execute an artificial intelligence task. If it is determined that the target device for performing the artificial intelligence task includes a server, a task execution request may be generated. The task execution request is sent to the server, so that the server parses the task execution request in response to receiving the task execution request, and obtains task data including the artificial intelligence task identifier and corresponding to the artificial intelligence task identifier. The server determines the artificial intelligence model corresponding to the artificial intelligence task identifier according to the artificial intelligence task identifier. The server can execute the artificial intelligence task corresponding to the artificial intelligence task identifier based on the artificial intelligence model and task data corresponding to the artificial intelligence model identifier.

According to an embodiment of the present disclosure, the task information included in the task execution request is obtained through encryption.

According to an embodiment of the present disclosure, an encryption algorithm may be used to encrypt task information to obtain encrypted task information. The encryption algorithm may include a symmetric encryption algorithm.

According to the embodiments of the present disclosure, the security of data and computing logic is effectively guaranteed by encrypting the information exchanged between the terminal and the cloud.

In the case where it is determined that the respective target devices for performing at least one artificial intelligence task also include end-side devices, sending at least one first log information to the server, so that the server performs at least one first log information and at least one second log information Aggregate to get aggregated information. Each first log information includes an execution result obtained by the terminal-side device executing an artificial intelligence task corresponding to the terminal-side device, and each second log information includes an execution result obtained by the server executing the artificial intelligence task corresponding to the server.

According to an embodiment of the present disclosure, the log information may include information related to performing an artificial intelligence task. Log information may include execution results. Log information may also include intermediate processing parameters and error messages.

According to an embodiment of the present disclosure, for each artificial intelligence task in at least one artificial intelligence task, if it is determined that the target device for executing the artificial intelligence task is an end-side device, the end-side device may obtain After the execution result, first log information including the execution result is generated. Then, the first log information is sent to the server, so that the server can receive at least one piece of first log information from the end-side device. If it is determined that the target device for executing the artificial intelligence task is a server, the server generates second log information including the execution result after executing the artificial intelligence task and obtaining an execution result. Thus, the server can obtain at least one piece of first log information and at least one piece of second log information. At least one piece of first log information and at least one piece of second log information may be aggregated to obtain aggregated information.

According to an embodiment of the present disclosure, the server aggregates at least one first log information and at least one second log information to obtain the aggregated information, so that the complete process can be obtained based on the aggregated information later, which facilitates troubleshooting and facilitates developers for debugging and optimization.

According to an embodiment of the present disclosure, each first log information is obtained through encryption processing.

Processing the execution results respectively corresponding to at least one artificial intelligence task to obtain the processing results. The processing results are analyzed to obtain extended results related to the processing results.

According to an embodiment of the present disclosure, if the target device includes a server, the execution results respectively corresponding to at least one artificial intelligence task include execution results processed by the server. The server can send the execution result obtained by the server to execute the artificial intelligence task to the end-side device. The execution results described here may be obtained through encryption.

According to an embodiment of the present disclosure, the end-side device may process execution results corresponding to at least one artificial intelligence task to obtain a processing result. For example, execution results corresponding to at least one artificial intelligence task may be aggregated to obtain a processing result. Then the processing result is analyzed to obtain the extended result related to the processing result. The extended result may be obtained by inferring the processing result.

For example, at least one artificial intelligence task includes an image recognition task and an object detection task. The target device performs the image recognition task and the target detection task and the processing result is "there is smoke outside classroom A, and there is someone in classroom A". After analyzing the processing results, the extended result obtained is "There may be a fire in classroom A, and it is necessary to call the fire alarm and emergency numbers for help."

In the case that the execution result of the target artificial intelligence task is determined to be obtained, the execution result of the target artificial intelligence task is displayed. A target AI task is an AI task that demonstrates that the level satisfies a predetermined condition.

According to an embodiment of the present disclosure, the display level may refer to a level of display execution results. The higher the display level, the sooner it can be displayed after the execution result is obtained. If the execution results of the target artificial intelligence task are obtained, the execution results of the target artificial intelligence task may be displayed when all or none of the execution results of other artificial intelligence tasks in at least one artificial intelligence task have been obtained.

Call the data interface. Using the data interface to acquire task data corresponding to at least one artificial intelligence task.

According to an embodiment of the present disclosure, task data corresponding to an artificial intelligence task can be obtained through a data interface. End-side devices can provide data interfaces to users through SDK (Software Development Kit, SDK, software tool development kit) or HTTP (HyperText Transfer Protocol, hypertext transfer protocol). The server can provide the interface of relevant functions to the end-side device through HTTP or RPC (Remote Procedure Call Protocol, remote procedure call protocol), so that the end-side device can encapsulate the interface.

Referring to FIG. 3 , the method according to the embodiments of the present disclosure will be further described in conjunction with specific embodiments.

Fig. 3 schematically shows an example diagram of a task processing system according to an embodiment of the present disclosure.

As shown in FIG. 3 , in 300 , the end-side device 300 - 1 includes a data interface unit 301 , an end-side device performance evaluation unit 302 , a device scheduling unit 303 , a data security unit 306 and an application log return unit 312 . The server 300 - 2 includes a data security unit 306 and an application log aggregation unit 311 . The at least one artificial intelligence task includes artificial intelligence task 304 , artificial intelligence task 305 , artificial intelligence task 307 , artificial intelligence task 308 , artificial intelligence task 309 , and artificial intelligence task 310 .

The end-side device 300-1 obtains the tasks corresponding to the artificial intelligence task 304, the artificial intelligence task 305, the artificial intelligence task 307, the artificial intelligence task 308, the artificial intelligence task 309 and the artificial intelligence task 310 through the data interface provided by the data interface unit 301. data.

The device-side device performance evaluation unit 302 determines available resource information of the device-side device 300-1.

The device scheduling unit 303 determines the total resource consumption information according to the respective resource consumption information of the AI task 304 , the AI task 305 , the AI task 307 , the AI task 308 , the AI task 309 and the AI task 310 .

The device scheduling unit 303 determines that the available resource information of the device-side device 300-1 does not satisfy the total resource consumption information. According to the resource consumption information and demand priority information of the artificial intelligence task 304, the artificial intelligence task 305, the artificial intelligence task 307, the artificial intelligence task 308, the artificial intelligence task 309 and the artificial intelligence task 310, it is determined to be used to execute the artificial intelligence task 304 and The target device of the artificial intelligence task 305 is the end-side device 300-1. It is determined that the target device for executing the artificial intelligence task 307, the artificial intelligence task 308, the artificial intelligence task 309, and the artificial intelligence task 310 is the server 300-2. It is determined that the artificial intelligence task 304 and the artificial intelligence task 305 are executed serially, that is, the artificial intelligence task 304 is executed first, and then the artificial intelligence task 305 is executed. The artificial intelligence task 307 , the artificial intelligence task 308 , the artificial intelligence task 309 and the artificial intelligence task 310 are executed in parallel and are all executed in parallel with the artificial intelligence task 304 and the artificial intelligence task 305 .

The end-side device 300-1 generates a task execution request, and the task execution request includes task information, and the task information includes the AI task identification and business information corresponding to the AI task 307, AI task 308, AI task 309, and AI task 310 respectively. data. The data security unit 306 of the end-side device 300-1 encrypts the task information. The end-side device 300-1 sends the task execution request to the server 300-2.

The data security unit 306 of the server 300-2 decrypts the task information to obtain the AI task identification and business data corresponding to the AI task 307, AI task 308, AI task 309 and AI task 310 respectively. The server 300-2 determines the AI model according to the ID of the AI task. Based on the artificial intelligence models and business data corresponding to the artificial intelligence task 307, artificial intelligence task 308, artificial intelligence task 309 and artificial intelligence task 310, respectively execute the artificial intelligence task 307, the artificial intelligence task 308, the artificial intelligence task 309 and the artificial intelligence task The task 310 is to obtain execution results corresponding to the artificial intelligence task 307 , the artificial intelligence task 308 , the artificial intelligence task 309 and the artificial intelligence task 310 .

The end-side device 300 - 1 first executes the artificial intelligence task 304 based on the artificial intelligence model and business data corresponding to the artificial intelligence task 304 , and obtains an execution result corresponding to the artificial intelligence task 304 . Then, according to the execution result corresponding to the artificial intelligence task 304, it is determined that the artificial intelligence task 305 needs to be executed. Then, based on the artificial intelligence model and business data corresponding to the artificial intelligence task 305, the artificial intelligence task 305 is executed, and an execution result corresponding to the artificial intelligence task 305 is obtained. The server 300-2 sends the execution results corresponding to the artificial intelligence task 307, the artificial intelligence task 308, the artificial intelligence task 309 and the artificial intelligence task 310 to the end-side device 300-1 through the data security unit 306.

The end-side device 300-1 sends the first log information respectively corresponding to the artificial intelligence task 304 and the artificial intelligence task 305 to the server 300-2 through the application log return unit 312 and the data security unit 306. The application log aggregation unit 311 of the server 300-2 is based on the second log information corresponding to the AI task 307, the AI task 308, the AI task 309, and the AI task 310 and the AI task 304 and the AI task 305 respectively. The aggregated information is generated corresponding to the first log information.

The device scheduling unit 303 of the end-side device 300-1 processes the execution results corresponding to the artificial intelligence task 304, the artificial intelligence task 305, the artificial intelligence task 307, the artificial intelligence task 308, the artificial intelligence task 309 and the artificial intelligence task 310, Get the processing result. The processing results are analyzed to obtain extended results related to the processing results.

The above are only exemplary embodiments, but are not limited thereto, and may also include other task processing methods known in the art, as long as the resources of end-side devices and servers can be fully utilized on the basis of ensuring that artificial intelligence tasks can be completed, and the equipment resource utilization. In addition, the robustness and adaptability of task processing operations can also be improved.

Fig. 4 schematically shows a block diagram of a task processing device according to an embodiment of the present disclosure.

As shown in FIG. 4 , the task processing apparatus 400 may include a first determination module 410 , a second determination module 420 and an execution module 430 .

The first determining module 410 is configured to determine device scheduling information. The device scheduling information includes available resource information of end-side devices and resource consumption information of at least one artificial intelligence task;

The second determination module 420 is configured to determine the respective target devices for executing at least one artificial intelligence task from the server and the end-side device according to the device scheduling information.

The execution module is used to control the target device corresponding to at least one artificial intelligence task, and execute at least one artificial intelligence task based on the artificial intelligence model and task data corresponding to the at least one artificial intelligence task.

According to an embodiment of the present disclosure, the second determining module may include a first obtaining submodule, a first determining submodule, and a second determining submodule.

The first obtaining submodule is configured to obtain total resource consumption information according to at least one piece of resource consumption information.

The first determination sub-module is configured to determine the end-side device as a respective target device for executing at least one artificial intelligence task when it is determined that the available resource information satisfies the total resource consumption information.

The second determining submodule is used to determine the respective target devices for executing at least one artificial intelligence task from the server and the end-side device according to the device scheduling information when it is determined that the available resource information does not satisfy the total resource consumption information.

According to an embodiment of the present disclosure, the second determination module may include a third determination submodule, a fourth determination submodule, and a fifth determination submodule.

The third determination sub-module is configured to determine the target artificial intelligence task according to the device scheduling information, wherein the target artificial intelligence task is an artificial intelligence task that can be performed by the terminal device.

The fourth determining submodule is used to determine the end-side device as the target device for performing the target artificial intelligence task.

The fifth determination sub-module is used to determine the server as the target device for performing the first other artificial intelligence task, wherein the first other artificial intelligence task is an artificial intelligence task other than the target artificial intelligence task in at least one artificial intelligence task .

According to an embodiment of the present disclosure, the device scheduling information further includes demand priority information of at least one artificial intelligence task.

According to an embodiment of the present disclosure, the third determination submodule may include a first determination unit, a second determination unit, a third determination unit, a fourth determination unit, and a fifth determination unit.

The first determining unit is configured to determine a first candidate artificial intelligence task according to the demand priority information of at least one artificial intelligence task.

The second determining unit is configured to determine a second candidate artificial intelligence task according to available resource information of the device on the device and resource consumption information of the first candidate artificial intelligence task.

The third determining unit is configured to determine the second candidate artificial intelligence task as the target artificial intelligence task when it is determined that the available resource information of the device is consistent with the resource consumption information of the second candidate artificial intelligence task.

The fourth determining unit is configured to determine a third candidate artificial intelligence task when it is determined that the available resource information of the end-side device is greater than the resource consumption information of the second candidate artificial intelligence task, wherein the third candidate artificial intelligence task is at least one Artificial intelligence tasks other than the first candidate artificial intelligence task in the artificial intelligence task.

The fifth determination unit is configured to determine the target artificial intelligence task according to the second candidate artificial intelligence task and the third candidate artificial intelligence task.

According to an embodiment of the present disclosure, the executing module may include a sixth determining submodule, a seventh determining submodule, and an executing submodule.

The sixth determining submodule is used to determine the relationship between at least one artificial intelligence task.

The seventh determination sub-module is used to determine the task execution sequence according to the association relationship.

The execution sub-module is used to control the target device corresponding to at least one artificial intelligence task according to the order of task execution, and execute at least one artificial intelligence task based on the artificial intelligence model and task data corresponding to the at least one artificial intelligence task.

According to an embodiment of the present disclosure, the task processing apparatus 400 may further include a third determining module.

The third determination module is used to determine the respective execution priority information of at least one artificial intelligence task.

According to an embodiment of the present disclosure, the seventh determining submodule may include a sixth determining unit.

The sixth determination unit is configured to determine the task execution order according to the association relationship and at least one piece of execution priority information.

According to an embodiment of the present disclosure, the executing module may include an eighth determining submodule.

The eighth determination submodule is used to send a task execution request to the server when it is determined that the respective target devices for executing at least one artificial intelligence task include a cloud server, so that the server executes the task corresponding to the server in response to receiving the task execution request. artificial intelligence tasks. The task execution request includes task information, and the task information includes an artificial intelligence task identifier and task data corresponding to the artificial intelligence task identifier, and the artificial intelligence task identifier is used to determine an artificial intelligence model corresponding to the artificial intelligence task identifier.

According to an embodiment of the present disclosure, the task processing apparatus 400 may further include a fourth determining module.

The fourth determination module is configured to send at least one first log information to the server when it is determined that the respective target devices for performing at least one artificial intelligence task also include end-side devices, so that the server can compare the at least one first log information and At least one piece of second log information is aggregated to obtain aggregated information. Each first log information includes an execution result obtained by the terminal-side device executing an artificial intelligence task corresponding to the terminal-side device, and each second log information includes an execution result obtained by the server executing the artificial intelligence task corresponding to the server.

According to an embodiment of the present disclosure, the task processing apparatus 400 may further include a processing module and an analysis module.

The processing module is configured to process the execution results corresponding to at least one artificial intelligence task to obtain the processing results.

The analysis module is used to analyze the processing results and obtain extended results related to the processing results.

According to an embodiment of the present disclosure, the task processing apparatus 400 may further include a presentation module.

The display module is configured to display the execution result of the target artificial intelligence task when it is determined that the execution result of the target artificial intelligence task is obtained, wherein the target artificial intelligence task is an artificial intelligence task whose display level satisfies a predetermined condition.

According to an embodiment of the present disclosure, the task processing apparatus 400 may further include a calling module and an obtaining module.

The calling module is used to call the data interface.

An acquisition module, configured to acquire task data corresponding to at least one artificial intelligence task by using a data interface.

In the technical solution of this disclosure, the collection, storage, use, processing, transmission, provision, disclosure, and application of the user's personal information involved are all in compliance with relevant laws and regulations, necessary confidentiality measures have been taken, and they do not violate the Public order and good customs.

In the technical solution of the present disclosure, before acquiring or collecting the user's personal information, the user's authorization or consent is obtained.

According to the embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

According to an embodiment of the present disclosure, an electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by at least one processor, and the instructions are processed by at least one The processor is executed, so that at least one processor can perform the method as described above.

According to an embodiment of the present disclosure, there is a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the method as described above.

According to an embodiment of the present disclosure, a computer program product includes a computer program, and the computer program implements the method as described above when executed by a processor.

Fig. 5 schematically shows a block diagram of an electronic device suitable for implementing a task processing method according to an embodiment of the present disclosure. Electronic device is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. Electronic devices may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are by way of example only, and are not intended to limit implementations of the disclosure described and/or claimed herein.

As shown in FIG. 5 , an electronic device 500 includes a computing unit 501, which can perform calculations according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. Various appropriate actions and processes are performed. In the RAM 503, various programs and data necessary for the operation of the electronic device 500 can also be stored. The computing unit 501, ROM 502, and RAM 503 are connected to each other through a bus 504. An input/output (I/O) interface 505 is also connected to the bus 504 .

Multiple components in the electronic device 500 are connected to the I/O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a magnetic disk, an optical disk etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, and the like. The communication unit 509 allows the electronic device 500 to exchange information/data with other devices through a computer network such as the Internet and/or various telecommunication networks.

The computing unit 501 may be various general-purpose and/or special-purpose processing components having processing and computing capabilities. Some examples of computing units 501 include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), various dedicated artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, digital signal processing processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 501 executes various methods and processes described above, such as task processing methods. For example, in some embodiments, the task processing method may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 508 . In some embodiments, part or all of the computer program can be loaded and/or installed on the electronic device 500 via the ROM 502 and/or the communication unit 509. When the computer program is loaded into the RAM 503 and executed by the computing unit 501, one or more steps of the task processing method described above can be performed. Alternatively, in other embodiments, the computing unit 501 may be configured to execute the task processing method in any other suitable manner (for example, by means of firmware).

Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips Implemented in a system of systems (SOC), complex programmable logic device (CPLD), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include being implemented in one or more computer programs executable and/or interpreted on a programmable system including at least one programmable processor, the programmable processor Can be special-purpose or general-purpose programmable processor, can receive data and instruction from storage system, at least one input device, and at least one output device, and transmit data and instruction to this storage system, this at least one input device, and this at least one output device an output device.

Program codes for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special purpose computer, or other programmable data processing devices, so that the program codes, when executed by the processor or controller, make the functions/functions specified in the flow diagrams and/or block diagrams Action is implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media would include one or more wire-based electrical connections, portable computer discs, hard drives, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, compact disk read only memory (CD-ROM), optical storage, magnetic storage, or any suitable combination of the foregoing.

To provide for interaction with the user, the systems and techniques described herein can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user. ); and a keyboard and pointing device (eg, a mouse or a trackball) through which a user can provide input to the computer. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and can be in any form (including Acoustic input, speech input or, tactile input) to receive input from the user.

The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., as a a user computer having a graphical user interface or web browser through which a user can interact with embodiments of the systems and techniques described herein), or including such backend components, middleware components, Or any combination of front-end components in a computing system. The components of the system can be interconnected by any form or medium of digital data communication, eg, a communication network. Examples of communication networks include: Local Area Network (LAN), Wide Area Network (WAN) and the Internet.

A computer system may include clients and servers. Clients and servers are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, a server of a distributed system, or a server combined with a blockchain.

It should be understood that steps may be reordered, added or deleted using the various forms of flow shown above. For example, each step described in the present disclosure may be executed in parallel, sequentially, or in a different order, as long as the desired result of the technical solution disclosed in the present disclosure can be achieved, no limitation is imposed herein.

The specific implementation manners described above do not limit the protection scope of the present disclosure. It should be apparent to those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made depending on design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

A task processing method, comprising:

determining device scheduling information, wherein the device scheduling information includes available resource information of end-side devices and resource consumption information of at least one artificial intelligence task;

According to the device scheduling information, determine the respective target devices for executing the at least one artificial intelligence task from the server and the end-side device; and

Controlling target devices corresponding to the at least one artificial intelligence task, and executing the at least one artificial intelligence task based on the artificial intelligence model and task data corresponding to the at least one artificial intelligence task.
The method according to claim 1, wherein, according to the device scheduling information, determining from the server and the end-side device the respective target devices for executing the at least one artificial intelligence task comprises:

Obtaining total resource consumption information according to at least one piece of resource consumption information;

When it is determined that the available resource information satisfies the total resource consumption information, determine the end-side device as a respective target device for executing the at least one artificial intelligence task; and

When it is determined that the available resource information does not satisfy the total resource consumption information, according to the device scheduling information, determine from the server and the end-side device the respective resources for executing the at least one artificial intelligence task target device.
The method according to claim 2, wherein, according to the device scheduling information, determining from the server and the end-side device the respective target devices for executing the at least one artificial intelligence task comprises:

Determine a target artificial intelligence task according to the device scheduling information, where the target artificial intelligence task is an artificial intelligence task that can be performed by the device on the device side;

determining the end-side device as a target device for performing the target artificial intelligence task; and

determining the server as a target device for executing a first other artificial intelligence task, wherein the first other artificial intelligence task is an artificial intelligence task in the at least one artificial intelligence task other than the target artificial intelligence task .
The method according to claim 3, wherein the equipment scheduling information further includes demand priority information of the at least one artificial intelligence task;

Wherein, the determining the target artificial intelligence task according to the device scheduling information includes:

determining a first candidate artificial intelligence task according to respective demand priority information of the at least one artificial intelligence task;

determining the second candidate artificial intelligence task according to the available resource information of the terminal-side device and the resource consumption information of the first candidate artificial intelligence task;

When it is determined that the available resource information of the device is consistent with the resource consumption information of the second candidate artificial intelligence task, determining the second candidate artificial intelligence task as the target artificial intelligence task;

When it is determined that the available resource information of the end-side device is greater than the resource consumption information of the second candidate artificial intelligence task, determine a third candidate artificial intelligence task, wherein the third candidate artificial intelligence task is the at least AI tasks in an AI task other than the first candidate AI task; and

The target artificial intelligence task is determined according to the second candidate artificial intelligence task and the third candidate artificial intelligence task.
The method according to any one of claims 1 to 4, wherein the controlling the target device corresponding to the at least one artificial intelligence task is based on the artificial intelligence model and the artificial intelligence model corresponding to the at least one artificial intelligence task Task data for performing said at least one artificial intelligence task, comprising:

determining the relationship between the at least one artificial intelligence task;

Determining a task execution sequence according to the association relationship; and

According to the task execution order, control the target devices corresponding to the at least one artificial intelligence task, and execute the at least one artificial intelligence task based on the artificial intelligence model and task data corresponding to the at least one artificial intelligence task.
The method according to claim 5, further comprising:

determining respective execution priority information of the at least one artificial intelligence task;

Wherein, the determining the order of task execution according to the association relationship includes:

The task execution order is determined according to the association relationship and at least one piece of the execution priority information.
The method according to claim 2, wherein said controlling the target device respectively corresponding to said at least one artificial intelligence task executes said At least one artificial intelligence task, including:

In the case where it is determined that the respective target devices for executing the at least one artificial intelligence task include the cloud server, sending a task execution request to the server, so that the server executes and executes in response to receiving the task execution request. The artificial intelligence task corresponding to the server,

Wherein, the task execution request includes task information, and the task information includes an artificial intelligence task identifier and task data corresponding to the artificial intelligence task identifier, and the artificial intelligence task identifier is used to determine artificial intelligence model.
The method according to claim 7, wherein the task information included in the task execution request is obtained through encryption.
The method according to claim 7 or 8, further comprising:

When it is determined that the respective target devices for executing the at least one artificial intelligence task also include the end-side device, send at least one first log information to the server, so that the server can perform the at least one first log information Aggregating log information and at least one second log information to obtain aggregated information,

Wherein, each of the first log information includes the execution result obtained by the end-side device executing the artificial intelligence task corresponding to the end-side device, and each of the second log information includes the server execution and the server The execution result obtained by the corresponding artificial intelligence task.
The method according to any one of claims 1 to 9, further comprising:

Processing the execution results corresponding to the at least one artificial intelligence task to obtain a processing result; and

The processing result is analyzed to obtain an extended result related to the processing result.
The method according to any one of claims 1-10, further comprising:

If it is determined that the execution result of the target artificial intelligence task is obtained, the execution result of the target artificial intelligence task is displayed, wherein the target artificial intelligence task is an artificial intelligence task whose presentation level satisfies a predetermined condition.
The method according to any one of claims 1-11, further comprising:

call the data interface; and

Using the data interface to acquire task data corresponding to the at least one artificial intelligence task.
A task processing device, comprising:

A first determining module, configured to determine device scheduling information, wherein the device scheduling information includes available resource information of end-side devices and resource consumption information of at least one artificial intelligence task;

A second determining module, configured to determine from the server and the end-side device the respective target devices for executing the at least one artificial intelligence task according to the device scheduling information; and

An execution module, configured to control target devices corresponding to the at least one artificial intelligence task, and execute the at least one artificial intelligence task based on the artificial intelligence model and task data corresponding to the at least one artificial intelligence task.
The device according to claim 13, wherein the second determining module comprises:

A first obtaining submodule, configured to obtain total resource consumption information according to at least one piece of resource consumption information;

A first determining submodule, configured to determine the end-side device as a respective target device for executing the at least one artificial intelligence task when it is determined that the available resource information satisfies the total resource consumption information; and

The second determining submodule is configured to determine, from the server and the end-side device, the resource used to execute the Respective target devices for at least one artificial intelligence task.
The device according to claim 14, wherein the second determining module comprises:

The third determination sub-module is configured to determine a target artificial intelligence task according to the device scheduling information, wherein the target artificial intelligence task is an artificial intelligence task that can be executed by the terminal-side device;

A fourth determining submodule, configured to determine the end-side device as a target device for performing the target artificial intelligence task; and

A fifth determining submodule, configured to determine the server as a target device for executing a first other artificial intelligence task, wherein the first other artificial intelligence task is the at least one artificial intelligence task other than the target device AI tasks other than AI tasks.
The apparatus according to claim 15, wherein the equipment scheduling information further includes demand priority information of each of the at least one artificial intelligence task;

Wherein, the third determining submodule includes:

A first determination unit, configured to determine a first candidate artificial intelligence task according to the respective demand priority information of the at least one artificial intelligence task;

The second determining unit is configured to determine the second candidate artificial intelligence task according to the available resource information of the device on the terminal side and the resource consumption information of the first candidate artificial intelligence task;

A third determining unit, configured to determine the second candidate artificial intelligence task as the target when it is determined that the available resource information of the device is consistent with the resource consumption information of the second candidate artificial intelligence task artificial intelligence tasks;

A fourth determining unit, configured to determine a third candidate artificial intelligence task when it is determined that the available resource information of the device is greater than the resource consumption information of the second candidate artificial intelligence task, wherein the third candidate The artificial intelligence task is an artificial intelligence task of the at least one artificial intelligence task other than the first candidate artificial intelligence task; and

A fifth determining unit, configured to determine the target artificial intelligence task according to the second candidate artificial intelligence task and the third candidate artificial intelligence task.
The device according to any one of claims 13-16, wherein the execution module includes:

The sixth determination submodule is used to determine the relationship between the at least one artificial intelligence task;

The seventh determining submodule is used to determine the order of task execution according to the association relationship; and

The execution submodule is configured to control the target devices corresponding to the at least one artificial intelligence task according to the task execution order, and execute the said at least one artificial intelligence task based on the artificial intelligence model and task data corresponding to the at least one artificial intelligence task respectively. At least one AI task.
An electronic device comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,

The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor, so that the at least one processor can execute any one of claims 1-12. Methods.
A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the method according to any one of claims 1-12.
A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1-12.