Task scheduling method and device based on large model, equipment, medium and product

Abstract

The invention provides a task scheduling method and device based on a large model, equipment, a medium and a product, and relates to the technical field of intelligent command scheduling, and the method comprises the steps: obtaining business data and scheduling scene data of a to-be-scheduled task; respectively inputting the business data and the scheduling scene data into a pre-trained data stream framework, and determining an entity and an intention of the to-be-scheduled task; determining a target task corresponding to the entity and the intention of the to-be-scheduled task according to tasks bound for various intentions in advance; and issuing and executing the target task, and selecting a corresponding execution action according to the execution condition of the target task and the feedback information to complete task scheduling. According to the invention, intention arrangement and identification are carried out through the data stream framework, the task scheduling efficiency is enhanced, the scheduling cost can be effectively saved for the user, the scheduling accuracy and speed are enhanced, the professional level of automation and intelligence is embodied, and task execution and distribution in different scheduling scenes are rapidly adapted.

Description

Technical Field

[0001] The present invention relates to the field of intelligent command and dispatch technology, and in particular to a task dispatching method, device, equipment, medium and product based on a large model. Background Art

[0002] In various dispatching scenarios, there are problems such as complex and diverse scenarios, sudden situations, and low dispatching efficiency. How to issue tasks more quickly and accurately, accurately and quickly dispatch personnel, materials and other resources, and reduce the adverse effects of various emergencies has become an urgent problem to be solved. Both the user side and the front line have great demands. However, in traditional dispatching scenarios, task issuance is mainly handled by manual judgment, which cannot meet user needs and scenario applications in terms of labor costs and dispatching efficiency. Summary of the invention

[0003] The present invention provides a task scheduling method, device, equipment, medium and product based on a large model, which solves the defect that the existing scheduling method cannot meet the needs of users and scenario applications.

[0004] The present invention provides a task scheduling method based on a large model, comprising: Obtain the business data and scheduling scenario data of the tasks to be scheduled; Inputting the business data and the scheduling scenario data into a pre-trained data flow framework respectively to determine the entity and intent of the task to be scheduled; Determine the target task corresponding to the entity and the intent of the task to be scheduled according to the tasks pre-bound to various intents; The target task is issued and executed, and the corresponding execution action is selected according to the execution status and feedback information of the target task to complete the task scheduling.

[0005] As an embodiment, the data flow framework includes a process automation model, a big model and an artificial intelligence model. The process automation model is used to realize business flow visualization, and the big model is used to determine the entity and intention of the task to be scheduled based on the business data and the scheduling scenario data; the artificial intelligence model is used to adjust the parameters of the big model.

[0006] As an embodiment, the artificial intelligence model is also used to obtain data samples, entity samples and intent samples based on a preset data collection and preprocessing template, and use the data samples, entity samples and intent samples as data sets to adjust the parameters of the big model based on the big model adjustment algorithm.

[0007] As an embodiment, the artificial intelligence model is also used to learn the data set based on deep learning to identify keyword machine language corresponding to different intent samples.

[0008] As an embodiment, after the business data and the scheduling scenario data are respectively input into a pre-trained data flow framework to determine the entity and the intention of the task to be scheduled, the method further includes: Convert the entity of the task to be scheduled into an API input parameter based on a word slot algorithm; The API interface is called according to the API input parameters to realize the interaction between the data flow framework and the existing network data.

[0009] As an embodiment, after selecting a corresponding execution action according to the execution status of the target task and the feedback information, the method further includes: Iterate over the entities and intentions of the tasks to be scheduled.

[0010] The present invention also provides a task scheduling device based on a large model, comprising: The data acquisition module is used to obtain the business data and scheduling scenario data of the tasks to be scheduled; A first determination module is used to input the business data and the scheduling scenario data into a pre-trained data flow framework, respectively, to determine the entity and intention of the task to be scheduled, wherein the data flow framework is determined based on a large model; A second determination module is used to determine the target task corresponding to the entity and intent of the task to be scheduled according to the tasks pre-bound to various intents; The task scheduling module is used to issue and execute the target task, iterate the entity and intention of the task to be scheduled during the execution of the target task, and complete the task scheduling.

[0011] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements any one of the large model-based task scheduling methods or the large model-based task scheduling method described above.

[0012] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the computer program implements any of the large model-based task scheduling methods or the large model-based task scheduling methods described above.

[0013] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the large model-based task scheduling methods or the large model-based task scheduling methods described above.

[0014] The large-model-based task scheduling method, device, equipment, medium and product provided by the present invention obtain business data and scheduling scenario data of the task to be scheduled; respectively input the business data and the scheduling scenario data into a pre-trained data flow framework to determine the entity and intent of the task to be scheduled; determine the target task corresponding to the entity and intent of the task to be scheduled based on the tasks pre-bound to various intents; issue and execute the target task, select the corresponding execution action based on the execution status and feedback information of the target task, and complete task scheduling. The present invention arranges and identifies intents through a data flow framework, enhances task scheduling efficiency, and can also effectively save scheduling costs for users, enhance scheduling accuracy and speed, embody the professional level of automation and intelligence, and quickly adapt to task execution and allocation in different scheduling scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 It is a flowchart of the task scheduling method based on the large model provided by the present invention.

[0017] Figure 2 It is a structural schematic diagram of a task scheduling device based on a large model provided by the present invention.

[0018] Figure 3 It is a structural diagram of a task scheduling system based on a large model provided by the present invention.

[0019] Figure 4 It is a logical schematic diagram of the fine-tuning module of the command and dispatch scenario provided by the present invention.

[0020] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] The following is a detailed description of a large model-based task scheduling method, device, equipment, medium and product provided by the present invention in conjunction with the accompanying drawings.

[0023] Figure 1 This is one of the flow charts of the task scheduling method based on the large model provided by the present invention, such as Figure 1 As shown, the present invention provides a task scheduling method based on a large model, which at least includes steps S100 to S400.

[0024] Step S100, obtain the business data and scheduling scenario data of the task to be scheduled. The task to be scheduled is the command and dispatch task in different emergency scenarios. The business data includes the trigger rules, faults, processes, indicators, plans, emergency measures, etc. in the process of task issuance, as well as the feedback results, reply content, and returned pictures in the process of task execution. The scheduling scenario data refers to the emergency scenario and also includes some public information, such as weather, natural disasters, emergencies, etc.

[0025] Step S200, respectively input the business data and the scheduling scenario data into the pre-trained data flow framework to determine the entity and intent of the task to be scheduled. The data flow framework uses a large model and AI algorithm to optimize the task scheduling in the command and dispatch application scenario, uses the deep learning intent recognition algorithm to perform intent recognition, automatically learns the feature representation in the text corpus, and classifies intent according to these features. It can effectively optimize the poor accuracy, high cost, low efficiency and other problems in task scheduling, and can continuously optimize the task scheduling process with the help of AI training, thereby realizing intelligent scheduling.

[0026] The entities of the tasks to be scheduled include keywords such as task, time, location, and participants.

[0027] Task: Specific work to be performed, such as "data processing", "report generation", etc.

[0028] Time: The specific time or time range for task execution, such as "tomorrow morning at 9 o'clock", "next Friday", etc.

[0029] Location: The geographical location associated with the task, such as "Office", "Remote", etc.

[0030] Priority: The urgency of the task, such as "high priority", "low priority", etc.

[0031] Participant: The person or team that performs the task, such as "John", "Development Team", etc.

[0032] Resources: Resources required to complete the task, such as "server", "software", etc.

[0033] The intention of the task to be scheduled is related to the task, including creating a task, modifying a task, deleting a task, querying a task, assigning a task, etc.

[0034] Create Task: Create a new scheduled task.

[0035] Modify Task: Change the details of an existing task, such as time, participants, or priority.

[0036] Delete Task: Remove a scheduled task.

[0037] Query tasks: Get the list or status of the current tasks to be scheduled.

[0038] Assign tasks: Assign tasks to specific participants or teams.

[0039] Complete Task: Marks the task as completed.

[0040] Step S300, determining a target task corresponding to the entity and intent of the task to be scheduled according to the tasks pre-bound to various intents.

[0041] Tasks that are pre-bound for various intents refer to the use of intent orchestration technology to customize the process of the intents that need to be identified and the interface binding operations, thereby supporting the flow according to the preset intent when the task is executed.

[0042] In the scheduling scenario, the intent is triggered according to the scenario type and event information, the corresponding task is started by the intent, and the issuance and scheduling of the task are controlled according to the preset intent.

[0043] Through intent recognition, task initiation rules can be triggered. The big model outputs different intentions based on different business scenario information, and the intentions initiate different task scheduling processes based on actual conditions.

[0044] Step S400, issue and execute the target task, select the corresponding execution action according to the execution status and feedback information of the target task, and complete the task scheduling. During the execution of the task, the execution action will be automatically selected according to the execution status and feedback information, including branch selection, waiting, and continuing, and the closed-loop operation will be automatically completed according to the result.

[0045] Optionally, when implementing the task scheduling method of the present invention, that is, before executing step S100, it is necessary to collect and store a large amount of corpus data, including but not limited to indicators, faults, complaints, work orders, resources, etc., and train and fine-tune the data flow framework based on the large model base through the fine-tuning algorithm of the command and dispatch scenario. By analyzing the data points that the user is concerned about, the intentions in the command and dispatch scenario can be identified, such as emergencies, large-scale decommissioning, failures, etc. The data interaction between the large model and the API is realized according to the word slot algorithm to support data query or task execution after intent recognition. The trained intent is bound to the task scheduling process to automatically drive the task flow during the task execution phase, collect feedback information and make judgments.

[0046] It is understandable that the present invention arranges and identifies intents through a data stream framework, which enhances the efficiency of task scheduling, can effectively save scheduling costs for users, enhance the accuracy and speed of scheduling, embody the professional level of automation and intelligence, and quickly adapt to task execution and allocation in different scheduling scenarios. It can achieve rapid matching, accurate scheduling, and plan recommendation of command and dispatch task processes in different emergency scenarios, thereby further streamlining the scheduling process, reducing operation and maintenance costs, and improving response timeliness.

[0047] Based on the above embodiment, as an optional embodiment, the data flow framework includes a process automation (RPA) model, a big model and an artificial intelligence model. The process automation model is used to realize business flow visualization, and the big model is used to determine the entity and intention of the task to be scheduled according to the business data and the scheduling scenario data; the artificial intelligence model is used to adjust the parameters of the big model.

[0048] It can be understood that the present invention integrates process automation technology, big model technology, and artificial intelligence technology for command and dispatch scenarios, so that business flows are visible, AI algorithm parameters can be configured, and fine-tuning models can be selected, so as to better realize task identification and scheduling and adapt to various command and dispatch business scenarios.

[0049] Based on the above embodiments, as an optional embodiment, the artificial intelligence model is also used to obtain data samples, entity samples and intent samples based on a preset data collection and preprocessing template, and use the data samples, entity samples and intent samples as data sets to adjust the parameters of the big model based on the big model adjustment algorithm.

[0050] Large model adjustment algorithms include the LoRA algorithm, the P-tuning-2 algorithm, etc. The LoRA algorithm is a method for fine-tuning pre-trained models through low-rank matrix decomposition. It mainly introduces trainable low-rank adapters in the weight update process of the model, rather than directly adjusting large-scale model parameters. The P-tuning-2 algorithm is a method for enhancing the performance of pre-trained models by introducing learnable prompts. This method assumes that in a large number of downstream tasks, the model can guide input and output through prompts. Unlike traditional prompt methods, P-tuning-2 learns a prompt sequence suitable for a specific task and embeds it into the input so that the language model can better understand the task intent, thereby improving the accuracy and effectiveness of the model.

[0051] It can be understood that the present invention designs corpus collection and preprocessing templates, collects sentence patterns, entities, and intent information, uses LoRA and P-tuning-2 technologies, trains and outputs intent recognition models and entity recognition models, and obtains a more suitable scheduling algorithm, which can ultimately identify the business intent and entity information in the scenario, and improve the accuracy and stability of entity recognition and intent recognition.

[0052] Based on the above embodiment, as an optional embodiment, the artificial intelligence model is also used to learn the data set based on deep learning to identify the keyword machine language corresponding to different intent samples.

[0053] Specifically, the present invention provides intent orchestration capabilities for intent orchestration and task driving, which can decompose the analyzed business process tasks and orchestrate them into corresponding intents, and bind the intents to the scheduling process. When keywords appear in the scheduling process or a certain node is reached, the relevant intents will be triggered, thereby starting the relevant tasks.

[0054] It can be understood that the present invention learns the training data based on the deep learning model, identifies the keyword machine language patterns behind different intentions, and optimizes the model parameters based on the training data, which can effectively improve the recognition accuracy of the model.

[0055] Based on the above embodiment, as an optional embodiment, after the business data and the scheduling scenario data are respectively input into a pre-trained data flow framework to determine the entity and intention of the task to be scheduled, the following steps are also included.

[0056] Based on the word slot algorithm, the entity of the task to be scheduled is converted into an API input parameter; the API interface is called according to the API input parameter to realize the interaction between the data flow framework and the existing network data.

[0057] Among them, the slot filling algorithm is a technology suitable for natural language processing (NLP) and dialogue systems. It is mainly used to extract and identify key information in user input. Slots usually refer to variables or parameters that need to be filled in specific dialogue scenarios. These variables can be entities extracted from natural language text.

[0058] It is understandable that when the data flow framework interacts with the existing network API through the big model, the identified entities are converted into API input parameters through the word slot algorithm to realize the interaction between the big model and the existing network data. After production testing, it is more stable and accurate than the general solution.

[0059] Based on the above embodiment, as an optional embodiment, after selecting the corresponding execution action according to the execution status and feedback information of the target task, it also includes: iterating the entity and intention of the task to be scheduled. In task scheduling, during the execution of the task, the execution action will be automatically selected according to the execution status and feedback information, including branch selection, waiting, and continuing, and then the closed-loop operation of the entire scheduling process will be automatically completed through iterative intention recognition.

[0060] It can be understood that the present invention combines task scheduling algorithms and strategies according to actual business scheduling scenarios and training scheduling intentions to complete automated and intelligent task distribution in command and scheduling scenarios, thereby reducing manpower input and increasing scheduling efficiency.

[0061] The task scheduling device based on a large model provided by the present invention is described below. The task scheduling device based on a large model described below and the task scheduling method based on a large model described above can refer to each other.

[0062] Figure 2 is a schematic diagram of the structure of the task scheduling device based on the large model provided by the present invention, such as Figure 2 As shown, the present invention also provides a task scheduling device based on a large model, including the following modules.

[0063] The data acquisition module 210 is used to acquire the business data and scheduling scenario data of the task to be scheduled; A first determination module 220, for inputting the business data and the scheduling scenario data into a pre-trained data flow framework, respectively, to determine the entity and intent of the task to be scheduled, wherein the data flow framework is determined based on a large model; The second determination module 230 is used to determine the target task corresponding to the entity and intent of the task to be scheduled according to the tasks pre-bound to various intents; The task scheduling module 240 is used to issue and execute the target task, iterate the entity and intention of the task to be scheduled during the execution of the target task, and complete task scheduling.

[0064] As an embodiment, the data flow framework includes a process automation model, a big model and an artificial intelligence model. The process automation model is used to realize business flow visualization, and the big model is used to determine the entity and intention of the task to be scheduled based on the business data and the scheduling scenario data; the artificial intelligence model is used to adjust the parameters of the big model.

[0065] As an embodiment, the artificial intelligence model is also used to obtain data samples, entity samples and intent samples based on a preset data collection and preprocessing template, and use the data samples, entity samples and intent samples as data sets to adjust the parameters of the big model based on the big model adjustment algorithm.

[0066] As an embodiment, the artificial intelligence model is also used to learn the data set based on deep learning to identify keyword machine language corresponding to different intent samples.

[0067] As an embodiment, the first determining module 220 is further configured to: Convert the entity of the task to be scheduled into an API input parameter based on a word slot algorithm; The API interface is called according to the API input parameters to realize the interaction between the data flow framework and the existing network data.

[0068] As an embodiment, the task scheduling module 240 is further used for: Iterate over the entities and intentions of the tasks to be scheduled.

[0069] It should be noted that the task scheduling device based on the big model provided by the present invention can execute the task scheduling method based on the big model described in any of the above embodiments during specific operation, and has the technical effect corresponding to the method, which will not be elaborated in this embodiment.

[0070] The task scheduling system based on a large model provided by the present invention is described below. The task scheduling system based on a large model described below and the task scheduling method based on a large model described above can be referenced to each other.

[0071] Figure 3 is a schematic diagram of the structure of the task scheduling system based on the large model provided by the present invention, such as Figure 3As shown, the present invention provides a task scheduling system based on a big model, including a data flow framework, a fine-tuning module for a command and dispatch scenario, and a task scheduling module. The data flow framework integrates RPA capabilities, big model capabilities, and AI capabilities, so that business flows are visible, AI algorithm parameters are configurable, and fine-tuning models are selectable; the fine-tuning module for the command and dispatch scenario adjusts parameters of the data flow framework based on LoRA and P-Turning-2 algorithms, trains and outputs intent recognition models and entity recognition models to support the flow of task scheduling processes, and the task scheduling module is used to issue and execute the target task according to the recognized intent, select the corresponding execution action according to the execution status and feedback information of the target task, and complete the task scheduling.

[0072] Figure 4 is a logical schematic diagram of the fine-tuning module of the command and dispatch scenario provided by the present invention, such as Figure 4 As shown in the figure, the logic of the fine-tuning module of the command and dispatch scenario includes five steps: data collection, data analysis, model selection, model adjustment, and model evaluation. First, the data before and during task scheduling, such as communication data and monitoring data, are collected, and data analysis tools and algorithms are used to analyze the data. According to the analysis results, the model is selected as the basis for fine-tuning, and the model is fine-tuned using technologies such as P Turning-2 and LoRA to optimize the performance on specific tasks. Finally, the fine-tuned model is evaluated to meet the needs of task scheduling in the command and dispatch scenario.

[0073] In summary, the present invention conducts AI training on a large amount of corpus data, outputs an intent recognition model and an entity recognition model, and on this basis, realizes intent recognition and entity recognition in the scheduling process. Through intent recognition, the task start rule can be triggered. The large model outputs different intents according to different business scenario information, and the intent starts different task scheduling processes according to the actual situation; through entity recognition, the data interaction between the large model and the existing network API is realized. During the task execution process, relevant data can be queried through entity recognition, and different interfaces can be called to provide data support for the flow of tasks. In short, the present invention not only embodies the powerful capabilities of the large model, but also relies on the data flow framework and fine-tuning technology provided to effectively support the arrangement and recognition of intents, and realize automation and intelligent capabilities. At the same time, relying on word slot technology and knowledge base learning technology, it provides strong data support for task scheduling, which greatly improves the flow efficiency and accuracy of tasks in actual command and scheduling scenarios.

[0074] Figure 5 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 5As shown, the electronic device may include: a processor 510, a communication interface 520, a memory 530 and a communication bus 540, wherein the processor 510, the communication interface 520 and the memory 530 communicate with each other through the communication bus 540. The processor 510 may call the logic instructions in the memory 530 to execute the task scheduling method based on the large model, including: obtaining the business data and scheduling scenario data of the task to be scheduled; inputting the business data and the scheduling scenario data into the pre-trained data flow framework respectively to determine the entity and intent of the task to be scheduled; determining the target task corresponding to the entity and intent of the task to be scheduled according to the tasks pre-bound for various intents; issuing and executing the target task, selecting the corresponding execution action according to the execution status and feedback information of the target task, and completing the task scheduling.

[0075] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0076] On the other hand, the present invention also provides a computer program product, which includes a computer program, and the computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the large model-based task scheduling method provided by the above methods, including: obtaining business data and scheduling scenario data of the task to be scheduled; inputting the business data and the scheduling scenario data into a pre-trained data flow framework respectively to determine the entity and intention of the task to be scheduled; determining the target task corresponding to the entity and intention of the task to be scheduled based on the tasks pre-bound to various intentions; issuing and executing the target task, selecting the corresponding execution action based on the execution status and feedback information of the target task, and completing the task scheduling.

[0077] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the large model-based task scheduling method provided by the above-mentioned methods, including: obtaining business data and scheduling scenario data of the task to be scheduled; inputting the business data and the scheduling scenario data into a pre-trained data flow framework respectively to determine the entity and intent of the task to be scheduled; determining a target task corresponding to the entity and intent of the task to be scheduled based on tasks pre-bound to various intents; issuing and executing the target task, selecting a corresponding execution action based on the execution status and feedback information of the target task, and completing task scheduling.

[0078] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0079] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A task scheduling method based on a large model, characterized in that: include: Obtain the business data and scheduling scenario data of the tasks to be scheduled; Inputting the business data and the scheduling scenario data into a pre-trained data flow framework respectively to determine the entity and intent of the task to be scheduled; Determine the target task corresponding to the entity and the intent of the task to be scheduled according to the tasks pre-bound to various intents; The target task is issued and executed, and the corresponding execution action is selected according to the execution status and feedback information of the target task to complete the task scheduling.

2. The task scheduling method based on a large model according to claim 1 is characterized in that: The data flow framework includes a process automation model, a large model and an artificial intelligence model. The process automation model is used to realize business flow visualization, and the large model is used to determine the entity and intention of the task to be scheduled according to the business data and the scheduling scenario data; The artificial intelligence model is used to adjust the parameters of the large model.

3. The task scheduling method based on a large model according to claim 2 is characterized in that: The artificial intelligence model is also used to obtain data samples, entity samples and intent samples based on a preset data collection and preprocessing template, and use the data samples, entity samples and intent samples as data sets to adjust the parameters of the large model based on the large model adjustment algorithm.

4. The task scheduling method based on a large model according to claim 3 is characterized in that: The artificial intelligence model is also used to learn the data set based on deep learning to identify the keyword machine language corresponding to different intent samples.

5. The task scheduling method based on a large model according to claim 1 is characterized in that: After the business data and the scheduling scenario data are respectively input into the pre-trained data flow framework to determine the entity and intention of the task to be scheduled, the method further includes: Convert the entity of the task to be scheduled into an API input parameter based on a word slot algorithm; The API interface is called according to the API input parameters to realize the interaction between the data flow framework and the existing network data.

6. The task scheduling method based on a large model according to claim 1 is characterized in that: After selecting a corresponding execution action according to the execution status of the target task and the feedback information, the method further includes: Iterate over the entities and intentions of the tasks to be scheduled.

7. A task scheduling device based on a large model, characterized in that: include: The data acquisition module is used to obtain the business data and scheduling scenario data of the tasks to be scheduled; A first determination module is used to input the business data and the scheduling scenario data into a pre-trained data flow framework, respectively, to determine the entity and intention of the task to be scheduled, wherein the data flow framework is determined based on a large model; A second determination module is used to determine the target task corresponding to the entity and intent of the task to be scheduled according to the tasks pre-bound to various intents; The task scheduling module is used to issue and execute the target task, iterate the entity and intention of the task to be scheduled during the execution of the target task, and complete the task scheduling.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the large model-based task scheduling method as described in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the large model-based task scheduling method as described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the large model-based task scheduling method as described in any one of claims 1 to 6 is implemented.

Images