Task scheduling method and device based on AI algorithm

Abstract

The invention provides a task scheduling method and device based on an AI algorithm, and relates to the technical field of image processing, and the method comprises the steps: obtaining a to-be-processed file, matching a file type of the to-be-processed file with an algorithm group in user algorithm configuration data, and obtaining a first algorithm group matched with the file type of the to-be-processed file; and when the file ID of the to-be-processed file and the algorithm result corresponding to the algorithm ID of the first algorithm group are stored in the database, directly outputting the algorithm result. And when the file ID of the to-be-processed file and the algorithm result corresponding to the algorithm ID of the first algorithm group are not stored in the database, executing algorithm group arrangement on the to-be-processed file and the first algorithm group, outputting the algorithm result and storing the algorithm result in the database. The efficiency and the stability of AI algorithm task scheduling are improved, the operation cost is reduced, and the adaptability and the expandability of AI algorithm task scheduling are enhanced.

Description

Technical Field

[0001] The present application relates to the field of AI technology, and in particular to a task scheduling method and device based on AI algorithm. Background Art

[0002] In the context of the rapid development of mobile Internet and artificial intelligence technology, AI algorithm processing technology plays a vital role in many fields such as Internet social networking, video live broadcast, and monitoring. However, existing AI algorithms are usually limited to specific product platforms or different protocol standards, which greatly limits their flexibility and scalability. Although integrating AI algorithms into Internet applications can significantly improve data processing accuracy and user experience, thereby enhancing the company's revenue capacity and user retention rate, the rapid iteration of AI algorithms has brought huge challenges to system integrators. First, when integrating new AI algorithm capabilities, Internet applications often need to frequently reconstruct business process codes, which not only increases the instability of the system, but also greatly increases the manpower and IT integration costs. Secondly, due to the emergence of various algorithms such as text large models, vector large models, and multimodal large models, there are significant differences in the algorithm capabilities and protocol standards of different suppliers, which seriously leads to system instability and increases the computational cost of AI models.

[0003] Therefore, how to effectively solve the compatibility, flexibility and cost issues of AI algorithm task scheduling has become a technical problem that needs to be solved urgently. Summary of the invention

[0004] The present invention provides a task scheduling method and device based on an AI algorithm, which are used to improve the accuracy of iris quality evaluation.

[0005] In a first aspect, a task scheduling method based on an AI algorithm is provided. The method comprises: Acquire a file to be processed, match the file to be processed with an algorithm group in the user algorithm configuration data, and acquire a first algorithm group that matches the file type of the file to be processed; When the database has stored the file ID of the to-be-processed file and the algorithm result corresponding to the algorithm ID of the first algorithm group, directly output the algorithm result; When the algorithm result corresponding to the file ID of the to-be-processed file and the algorithm ID of the first algorithm group is not stored in the database, algorithm group arrangement is performed on the to-be-processed file and the first algorithm group, and the algorithm result is output and stored in the database.

[0006] Furthermore, the sources of the files to be processed include files actively uploaded by users and user file resource tables; The user file resource table is constructed based on the user's historical file upload events; The user file resource table includes the file ID and the file download address.

[0007] Furthermore, the user algorithm configuration data includes a user ID, a file type, an algorithm group code, and a list of algorithms included in the algorithm group; User algorithm configuration data is customized and stored based on the user's historical selection configuration in the global algorithm configuration data; The global algorithm configuration data includes the file type, the algorithm group code, and the list of algorithms contained in the algorithm group.

[0008] Furthermore, the file types of the files to be processed include text files, image files and other files.

[0009] Further, matching the file to be processed with the algorithm group in the user algorithm configuration data to obtain a first algorithm group that matches the file type of the file to be processed includes: Loop through the algorithm groups in the user algorithm configuration data according to the file type of the file to be processed, and select a first algorithm group that matches the file type of the file to be processed from the algorithm groups in the user algorithm configuration data; The types of algorithm groups in the user algorithm configuration data include text vectorization algorithm groups, image metadata extraction algorithm groups, and other algorithm groups.

[0010] Further, when the database does not store the algorithm result corresponding to the file ID of the to-be-processed file and the algorithm ID of the first algorithm group, performing algorithm group arrangement on the to-be-processed file and the first algorithm group includes: Loop through and query the algorithm results corresponding to the file ID of the to-be-processed file and the algorithm ID of the first algorithm group; Acquire a second algorithm group having no algorithm result corresponding to the file ID of the to-be-processed file from the algorithm ID of the first algorithm group; The second algorithm group is executed until the file ID of the to-be-processed file and each algorithm ID of the second algorithm group have corresponding algorithm results.

[0011] Optionally, the second algorithm group is executed until the file ID of the to-be-processed file and each algorithm ID of the second algorithm group have corresponding algorithm results, including: When the second algorithm group includes more than one algorithm, the algorithms are executed sequentially according to the sequence numbers in the second algorithm group.

[0012] Furthermore, the algorithm results are stored in a database, including: Write the algorithm results into the corresponding sub-database according to the data type of the algorithm results; The sub-databases include vector databases, non-relational databases and other storage databases.

[0013] Optionally, the algorithm result is written into a corresponding sub-database according to the data type of the algorithm result, including: The vector database can store image vector data, text vector data and other vector data; Image metadata tables can be stored in non-relational databases; Metadata tables can be stored in other storage databases.

[0014] In a second aspect, a task scheduling device based on an AI algorithm is provided. The device comprises: An acquisition module, used for acquiring a file to be processed, matching the file to be processed with an algorithm group in the user algorithm configuration data, and acquiring a first algorithm group matching a file type of the file to be processed; A first output module, used for directly outputting the algorithm result when the file ID of the to-be-processed file and the algorithm result corresponding to the algorithm ID of the first algorithm group are already stored in the database; The second output module is used to perform algorithm group arrangement on the file to be processed and the first algorithm group, output the algorithm result and store the algorithm result in the database when the algorithm result corresponding to the file ID of the file to be processed and the algorithm ID of the first algorithm group is not stored in the database.

[0015] The present invention provides a task scheduling method and device based on AI algorithm, the method comprising: obtaining a file to be processed, matching the file type of the file to be processed with the algorithm group in the user algorithm configuration data, and obtaining a first algorithm group matching the file type of the file to be processed. When the algorithm result corresponding to the file ID of the file to be processed and the algorithm ID of the first algorithm group is already stored in the database, the algorithm result is directly output. When the algorithm result corresponding to the file ID of the file to be processed and the algorithm ID of the first algorithm group is not stored in the database, the algorithm group arrangement is performed on the file to be processed and the first algorithm group, and the algorithm result is output and stored in the database. That is, by matching the file to be processed with the user algorithm configuration data, the appropriate first algorithm group is quickly located, the waiting time of the file to be processed in the processing flow is reduced, and the adaptability to different business scenarios and demand changes is enhanced. Moreover, when the database already has the corresponding algorithm result of the file to be processed on the first algorithm group, the result can be directly output, avoiding unnecessary repeated calculations and saving computing resources. When there is no corresponding algorithm result in the database, the algorithm group arrangement is performed, which improves resource utilization efficiency. In addition, the storage and reuse of algorithm results simplifies the execution process of AI algorithms, making the task scheduling of AI algorithms more efficient and stable.

[0016] It should be understood that the contents described in the summary of the invention are not intended to limit the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The above and other features, advantages and aspects of the embodiments of the present invention will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements.

[0018] Figure 1 This is a flowchart of a task scheduling method based on an AI algorithm according to an embodiment of the present invention; Figure 2 This is a flowchart of a task scheduling method based on an AI algorithm in which a user actively uploads a file to be processed according to an embodiment of the present invention; Figure 3 It is a flowchart of a task scheduling method based on an AI algorithm for obtaining files to be processed from a user file resource table according to an embodiment of the present invention; Figure 4 A flowchart of an algorithm group according to an embodiment of the present invention; Figure 5 Schematic diagram of the structure of a task scheduling device based on AI algorithm according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below in conjunction with the drawings in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. Based on one or more embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this document.

[0020] It should be noted that the description of the embodiments of the present invention is only for more clearly illustrating the technical solutions of the embodiments of the present invention, and does not constitute a limitation on the technical solutions provided by the embodiments of the present invention.

[0021] The embodiment of the present invention provides a task scheduling method based on AI algorithm. Figure 1 Flow chart of the task scheduling method based on AI algorithm in an embodiment of the present invention. Figure 1 , the method comprising: S101. Obtain a file to be processed, match the file to be processed with an algorithm group in user algorithm configuration data, and obtain a first algorithm group that matches the file type of the file to be processed.

[0022] In this embodiment, the first algorithm group matching the file type of the to-be-processed file is obtained by matching, without manual intervention, thereby improving the automation level of the file processing flow and ensuring that the to-be-processed file can be properly processed.

[0023] Optionally, the sources of the files to be processed include files actively uploaded by users and a user file resource table; The user file resource table is constructed based on the user's historical file upload events; The user file resource table includes the file ID and the file download address.

[0024] Figure 2 Flow chart of the task scheduling method based on AI algorithm in which a user actively uploads a file to be processed according to an embodiment of the present invention. Figure 2 As shown below. Figure 2 , and explains in detail step by step how to execute the task scheduling method based on AI algorithm when the user actively uploads the files to be processed.

[0025] S201. Monitor the user's file upload event, obtain the file to be processed uploaded by the user, and ensure the immediacy and accuracy of the response; S202. Upload the file to be processed to the user file resource table to facilitate subsequent file management and algorithm matching; S203. Matching the file to be processed with the algorithm group in the user algorithm configuration data, aiming to select the most appropriate algorithm for processing the file; S204. If there is no algorithm group that matches the file to be processed, the execution is terminated directly to avoid invalid operations and improve system operation efficiency; S205. Determine whether the database has stored the algorithm result corresponding to the file ID of the to-be-processed file and the algorithm ID of the first algorithm group. If yes, go to S206; if not, go to S207; S206. Directly output the algorithm result; S207. Arrange the algorithm for the files to be processed and the first algorithm group to ensure that the execution order and logic of the algorithm are correct, and continue to execute S208; S208. Store the algorithm results in the database for subsequent query and call, and continue to execute S206.

[0026] Figure 3 FIG. 1 is a flowchart of a task scheduling method based on an AI algorithm for obtaining files to be processed from a user file resource table according to an embodiment of the present invention. Figure 3 As shown below. Figure 3 , and explains in detail step by step how to execute the task scheduling method based on the AI ​​algorithm when obtaining the files to be processed from the user file resource table.

[0027] S301. Obtain the user file resource table, select the file to be processed from the user file resource table, and ensure that the files to be processed subsequently are screened and confirmed; S302. Match the file to be processed with the algorithm group in the user algorithm configuration data to optimize processing efficiency and result quality; S303. If there is no algorithm group that matches the file to be processed, the execution is terminated directly to avoid invalid operations and improve system operation efficiency; S304. Determine whether the database has stored the algorithm result corresponding to the file ID of the file to be processed and the algorithm ID of the first algorithm group. If yes, go to S305; if not, go to S306; S305. Directly output the algorithm result; S306. Arrange the algorithm for the files to be processed and the first algorithm group to ensure that the execution order and logic of the algorithm are correct, and continue to execute S307; S307. Store the algorithm result in the database for subsequent query and call, and continue to execute S305.

[0028] In this embodiment, the sources of the files to be processed include two ways: user active upload and user file resource table, which provides users with more personalized services. In addition, the user file resource table is managed by file ID and file download address, which helps to quickly locate and obtain the subsequent files to be processed.

[0029] Optionally, the user algorithm configuration data includes a user ID, a file type, an algorithm group code, and a list of algorithms included in the algorithm group; User algorithm configuration data is customized and stored based on the user's historical selection configuration in the global algorithm configuration data; The global algorithm configuration data includes the file type, the algorithm group code, and the list of algorithms contained in the algorithm group.

[0030] For example, the user configuration algorithm data can be shown in the following table: User ID File Type Algorithm Group Code Algorithm List 123 image A1B2C3 Image segmentation algorithm, target detection algorithm, face recognition algorithm 69 text D4E5F6 Named entity recognition, sentiment analysis algorithm, text classification algorithm, machine translation algorithm 119 other G7H8I9 Speech Recognition Algorithms ... ... ... ... In this embodiment, users can select and configure from the global algorithm configuration data as needed, providing more suitable and professional algorithm models for different types of files to be processed, thereby improving the user experience and facilitating flexible replacement or adjustment according to actual application scenarios.

[0031] Optionally, the file types of the files to be processed include text files, image files and other files.

[0032] Among them, text files include but are not limited to news reports, scientific research papers and feedback emails, image files include but are not limited to product pictures, medical images and landscape photos, and other files include but are not limited to audio files, data reports and 3D modeling.

[0033] In this embodiment, by classifying the files to be processed, a suitable algorithm group can be quickly matched.

[0034] Optionally, matching the file to be processed with the algorithm group in the user algorithm configuration data to obtain a first algorithm group that matches the file type of the file to be processed includes: Loop through the algorithm groups in the user algorithm configuration data according to the file type of the file to be processed, and select a first algorithm group that matches the file type of the file to be processed from the algorithm groups in the user algorithm configuration data; The types of algorithm groups in the user algorithm configuration data include text vectorization algorithm groups, image metadata extraction algorithm groups, and other algorithm groups.

[0035] For example, when a user uploads a text file of the scientific literature category, the system receives the text file uploaded by the user and identifies the file type as "text", then loops through the algorithm groups in the user algorithm configuration data to find an algorithm group that matches the "text" file type. Among them, in the user algorithm configuration data, each algorithm group can have the following remarks: Text vectorization algorithm group: used to process text data and convert it into a numerical vector; Image metadata extraction algorithm group: used to extract metadata information from image files; Other algorithm groups: used for audio to text algorithms, etc. Since the "text vectorization algorithm group" is specifically used to process text files, it can be matched with the text files of the scientific literature category uploaded by the user as the first algorithm group.

[0036] In this embodiment, the intelligent and personalized task scheduling of the AI ​​algorithm is achieved by identifying the file type of the file to be processed and automatically matching the algorithm group in the user algorithm configuration data, which is of great significance for building efficient and intelligent task scheduling of the AI ​​algorithm.

[0037] S102. When the database has stored the algorithm result corresponding to the file ID of the to-be-processed file and the algorithm ID of the first algorithm group, directly output the algorithm result.

[0038] For example, the file ID of the to-be-processed file stored in the database is 315, and there is a first algorithm group with an algorithm ID of D1328K. The to-be-processed file with file ID 315 and the first algorithm group with algorithm ID D1328K have an algorithm result D1 stored in the database, and the algorithm result is directly output to the user.

[0039] In this embodiment, when the algorithm result of the file to be processed already exists in the database, the result is directly output, avoiding repeated calculation process, thereby reducing the burden of computing resources, saving costs, and helping enterprises improve service quality.

[0040] S103. When the database does not store the algorithm result corresponding to the file ID of the to-be-processed file and the algorithm ID of the first algorithm group, perform algorithm group arrangement on the to-be-processed file and the first algorithm group, output the algorithm result and store the algorithm result in the database.

[0041] In this embodiment, by executing algorithm groups to arrange calculations and store algorithm results on demand, the use efficiency and overall utilization of computing resources are optimized, the adaptability and flexibility of AI algorithm task scheduling are improved, and the later maintenance and expansion of the enterprise are facilitated.

[0042] Optionally, when the database does not store the algorithm result corresponding to the file ID of the to-be-processed file and the algorithm ID of the first algorithm group, performing algorithm group arrangement on the to-be-processed file and the first algorithm group includes: Loop through and query the algorithm results corresponding to the file ID of the to-be-processed file and the algorithm ID of the first algorithm group; Acquire a second algorithm group having no algorithm result corresponding to the file ID of the to-be-processed file from the algorithm ID of the first algorithm group; The second algorithm group is executed until the file ID of the to-be-processed file and each algorithm ID of the second algorithm group have corresponding algorithm results.

[0043] In this embodiment, by looping through and executing the second algorithm group that has no algorithm result corresponding to the file ID of the file to be processed, it is ensured that all related algorithms are executed, all necessary processing is effectively completed, and repeated work of multiple queries and processing is avoided.

[0044] Optionally, the second algorithm group is executed until the file ID of the to-be-processed file and each algorithm ID of the second algorithm group have corresponding algorithm results, including: When the second algorithm group includes more than one algorithm, the algorithms are executed sequentially according to the sequence numbers in the second algorithm group.

[0045] Among them, the execution type of the second algorithm group can be divided into sequential execution and execution once. When the second algorithm group contains more than one algorithm, it is executed in sequence according to the serial number in the second algorithm group. When the second algorithm group contains only one algorithm, it only needs to be executed once.

[0046] In this embodiment, when the second algorithm group includes multiple algorithms, they are executed sequentially according to the sequence numbers, ensuring that the algorithms are executed according to the predetermined priority and logical order, which helps to avoid conflicts or errors that may be caused by improper algorithm execution order.

[0047] Optionally, the algorithm results are stored in a database, including: Write the algorithm results into the corresponding sub-database according to the data type of the algorithm results; The sub-databases include vector databases, non-relational databases and other storage databases.

[0048] In this embodiment, the algorithm result is written into the corresponding sub-database according to the data type of the algorithm result, which can improve the efficiency of storing and querying the algorithm result data and facilitate subsequent management and maintenance.

[0049] Optionally, the algorithm result is written into a corresponding sub-database according to the data type of the algorithm result, including: The vector database can store image vector data, text vector data and other vector data; Image metadata tables can be stored in non-relational databases; Metadata tables can be stored in other storage databases.

[0050] Among them, the non-relational database, also known as the HBase database, is mainly used for the storage and processing of image metadata. It is characterized by high scalability, support for large-scale data storage and high-performance data retrieval.

[0051] In this embodiment, by setting up a vector database, a non-relational database, and other storage databases, a suitable database can be selected for storage according to the data type of the algorithm result, thereby improving the security of the data and reducing the risk of data leakage. In addition, other sub-databases can be expanded and set up as needed to make the overall architecture more refined and standardized.

[0052] Figure 4 A flowchart of the execution algorithm group of an embodiment of the present invention is shown in FIG. Figure 4 As shown below. Figure 4 , which explains in detail, step by step, how to perform algorithm group orchestration.

[0053] S401. Loop through the algorithm results corresponding to the file ID of the file to be processed and the algorithm ID of the first algorithm group to ensure that the algorithm results corresponding to the file to be processed and the first algorithm group can be checked; S402. Check whether the algorithm ID of the first algorithm group and the file ID of the file to be processed have corresponding algorithm results. If yes, go to S403; if not, go to S404; S403. Output the algorithm result corresponding to the algorithm ID and the file ID; S404. Obtain a second algorithm group of algorithm results that have no corresponding algorithm results to the file ID of the file to be processed; S405. Execute the second algorithm group, wherein the execution type is divided into sequential execution and execution once, and the algorithms in the sequentially executed algorithm group will be executed in sequence according to the established order; and the algorithm group executed once contains only one algorithm; S406. Obtain the algorithm result, store the algorithm result in the database, write the corresponding sub-dataset according to the algorithm result type, continue to execute S403, and loop to execute S402 until the file ID of the to-be-processed file and the algorithm ID of each second algorithm group have corresponding algorithm results. The database includes multiple sub-databases, namely, a vector database, a non-relational database, and other storage databases.

[0054] In the present invention, the automation and intelligence of algorithm group arrangement are realized, which greatly improves the speed and accuracy of data processing, provides users with more efficient and convenient services, and enhances user experience. In addition, the database includes a vector database, a non-relational database, and multiple sub-databases of other storage databases. This structural design helps to efficiently optimize data storage and query efficiency, ensuring rapid retrieval and efficient management of algorithm results.

[0055] Based on the task scheduling method based on the AI ​​algorithm provided in this embodiment, by obtaining the file to be processed, the file type of the file to be processed is matched with the algorithm group in the user algorithm configuration data, and the first algorithm group matching the file type of the file to be processed is obtained. When the algorithm result corresponding to the file ID of the file to be processed and the algorithm ID of the first algorithm group is already stored in the database, the algorithm result is directly output. When the algorithm result corresponding to the file ID of the file to be processed and the algorithm ID of the first algorithm group is not stored in the database, the algorithm group arrangement is performed on the file to be processed and the first algorithm group, and the algorithm result is output and stored in the database. That is to say, by matching the file to be processed with the user algorithm configuration data, the appropriate first algorithm group is quickly located, the waiting time of the file to be processed in the processing flow is reduced, and the adaptability to different business scenarios and demand changes is enhanced. In addition, when the database already has the corresponding algorithm result of the file to be processed on the first algorithm group, the result can be directly output, avoiding unnecessary repeated calculations and saving computing resources. When there is no corresponding algorithm result in the database, the algorithm group arrangement will be performed, which improves resource utilization efficiency. In addition, the storage and reuse of the algorithm results simplify the process of AI algorithm execution, making the task scheduling of the AI ​​algorithm more efficient and stable.

[0056] Combination of the above Figure 1-Figure 4The task scheduling method based on AI algorithm provided in the embodiment of the present application is described in detail. Figure 5 A task scheduling device based on an AI algorithm for executing a task scheduling method based on an AI algorithm provided in an embodiment of the present application is described in detail.

[0057] Figure 5 FIG. 1 is a schematic diagram of the structure of a task scheduling device based on an AI algorithm according to an embodiment of the present invention. Figure 5 , the device comprises: The acquisition module 501 is used to acquire a file to be processed, match the file to be processed with an algorithm group in the user algorithm configuration data, and acquire a first algorithm group that matches the file type of the file to be processed; The first output module 502 is used to directly output the algorithm result when the file ID of the to-be-processed file and the algorithm result corresponding to the algorithm ID of the first algorithm group are already stored in the database; The second output module 503 is used to perform algorithm group arrangement on the file to be processed and the first algorithm group, output the algorithm result and store the algorithm result in the database when the algorithm result corresponding to the file ID of the file to be processed and the algorithm ID of the first algorithm group is not stored in the database.

[0058] The technical effects of the task scheduling device based on the AI ​​algorithm refer to the technical effects of the task scheduling method based on the AI ​​algorithm mentioned above, which will not be repeated here.

[0059] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a different order from the order marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0060] The units or modules involved in the embodiments of the present invention may be implemented by software or hardware. The units or modules described may also be set in a processor, for example, it may be described as follows: a processor includes an acquisition module 501, a first output module 502, and a second output module 503, wherein the names of these modules do not constitute a limitation on the modules themselves in some cases.

[0061] The above description is only a preferred embodiment of the present invention and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present invention is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present invention (but not limited to) to form a technical solution.

Claims

1. A task scheduling method based on AI algorithm, characterized in that: The method comprises: Acquire a file to be processed, match the file to be processed with an algorithm group in the user algorithm configuration data, and acquire a first algorithm group that matches the file type of the file to be processed; When the file ID of the to-be-processed file and the algorithm result corresponding to the algorithm ID of the first algorithm group are already stored in the database, directly outputting the algorithm result; When the algorithm result corresponding to the file ID of the to-be-processed file and the algorithm ID of the first algorithm group is not stored in the database, algorithm group arrangement is performed on the to-be-processed file and the first algorithm group, the algorithm result is output and stored in the database.

2. The task scheduling method based on AI algorithm according to claim 1 is characterized in that: The sources of the files to be processed include the files actively uploaded by the user and the user file resource table; The user file resource table is constructed according to the historical file upload events of the user; The user file resource table includes a file ID and a file download address.

3. The task scheduling method based on AI algorithm according to claim 1 is characterized in that: The user algorithm configuration data includes a user ID, a file type, an algorithm group code, and a list of algorithms included in the algorithm group; The user algorithm configuration data is customized and stored according to the user's historical selection configuration in the global algorithm configuration data; The global algorithm configuration data includes the file type, the algorithm group code and the algorithm list included in the algorithm group.

4. The task scheduling method based on AI algorithm according to claim 1 is characterized in that: The file types of the files to be processed include text files, image files and other files.

5. The task scheduling method based on AI algorithm according to claim 1 is characterized in that: The step of matching the file to be processed with the algorithm group in the user algorithm configuration data to obtain a first algorithm group that matches the file type of the file to be processed includes: Looping through the algorithm groups in the user algorithm configuration data according to the file type of the to-be-processed file, and screening out a first algorithm group that matches the file type of the to-be-processed file from the algorithm groups in the user algorithm configuration data; The types of algorithm groups in the user algorithm configuration data include text vectorization algorithm groups, image metadata extraction algorithm groups, and other algorithm groups.

6. The task scheduling method based on AI algorithm according to claim 1, characterized in that: When the file ID of the to-be-processed file and the algorithm result corresponding to the algorithm ID of the first algorithm group are not stored in the database, performing algorithm group arrangement on the to-be-processed file and the first algorithm group includes: Circularly query the file ID of the to-be-processed file and the algorithm result corresponding to the algorithm ID of the first algorithm group; Acquire, from the algorithm ID of the first algorithm group, a second algorithm group having no algorithm result corresponding to the file ID of the to-be-processed file; The second algorithm group is executed until the file ID of the to-be-processed file and each algorithm ID of the second algorithm group have corresponding algorithm results.

7. The task scheduling method based on AI algorithm according to claim 6 is characterized in that: The executing the second algorithm group until the file ID of the to-be-processed file and each algorithm ID of the second algorithm group have corresponding algorithm results, comprises: When the second algorithm group includes more than one algorithm, the algorithms are executed sequentially according to the sequence numbers in the second algorithm group.

8. The task scheduling method based on AI algorithm according to claim 1, characterized in that: The step of storing the algorithm result in a database comprises: Writing the algorithm result into the corresponding sub-database according to the data type of the algorithm result; The sub-databases include vector databases, non-relational databases and other storage databases.

9. The task scheduling method based on AI algorithm according to claim 8, characterized in that: The step of writing the algorithm result into a corresponding sub-database according to the data type of the algorithm result comprises: The vector database may store picture vector data, text vector data and other vector data; The non-relational database may store a picture metadata table; The other storage database may store a metadata table.

10. A task scheduling device based on AI algorithm, characterized in that: The device comprises: An acquisition module, used for acquiring a file to be processed, matching the file to be processed with an algorithm group in the user algorithm configuration data, and acquiring a first algorithm group matching a file type of the file to be processed; A first output module, configured to directly output the algorithm result when the file ID of the to-be-processed file and the algorithm result corresponding to the algorithm ID of the first algorithm group are already stored in the database; The second output module is used to perform algorithm group arrangement on the file to be processed and the first algorithm group, output the algorithm result and store the algorithm result in the database when the algorithm result corresponding to the file ID of the file to be processed and the algorithm ID of the first algorithm group is not stored in the database.

Images