Method and system for automatically generating supervision items based on AI supervision large model

Abstract

The invention provides a method and system for automatically generating supervision items based on an AI supervision large model, and belongs to the technical field of data processing. The method comprises the steps of collecting information, training an AI supervision large model, uploading related document files describing newly-added supervision items and generating basic information, generating a data view, generating description information, verifying and optimizing, displaying and inputting and the like. The system comprises an acquisition module, a modeling module, an extraction and generation module, an association module, a matching and generation module, a verification module, a display and input module and the like. By adopting the method and the system for automatically generating the supervision items based on the AI supervision large model, the sources of the supervision items can be efficiently arranged, the time for studying and reading files in the process of newly adding the supervision items is shortened, the content which needs to be input and adjusted by supervision personnel in the process of newly adding the items is simplified, and the speed of newly adding the supervision items is greatly increased.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and system for automatically generating inspection items based on an AI inspection big model. Background Art

[0002] Supervision is an important part of the work of various institutions such as governments and universities. In the daily work of governments and institutions, various documents, meetings, speeches by leaders, central spirit, etc. are often involved, and the matters to be supervised and handled often come from these, and the supervisors enter the supervision matters based on this information and documents.

[0003] However, the existing supervision process, especially in the process of adding new supervision items, has the following main problems:

[0004] (1) The sources of inspection items are messy: In actual work, inspection items often come from multiple sources, such as official documents, meetings, etc., or third-party systems such as the mayor's mailbox, 12345 hotline, etc. Different sources of inspection items may represent different types of inspection items, such as meeting resolutions, leadership instructions, and people's livelihood issues. Therefore, in actual work, adding new inspection items based on different sources scattered everywhere will result in high management costs and a large workload for inspectors.

[0005] (2) The documents related to the inspection items contain a lot of content: When inspectors need to add inspection items to certain documents or materials, they often need to study the entire document material to correctly fill in the content and spirit of the document in the inspection item information. This work often takes a lot of time and is relatively inefficient. If there are places where the inspection item information does not fully describe the content and spirit of the document, it will also increase the communication and collaboration costs between the inspectors and the lead unit.

[0006] (3) Complex description of inspection items: When adding new inspection items, you often need to fill in multiple fields, such as the overall description of the item, the classification and level of the item, the description of the subtasks under the item, the inspection rules for each subtask, the responsible leading unit and individual, etc. All the above inspection item information needs to be filled in and selected manually, and the input content is complicated, which increases the tediousness of the inspectors' work content.

[0007] (4) Slow operation speed for adding new items: When manually entering the inspection item information, because it involves filling in many fields and may also require reference to relevant documents and materials, it often takes a long time to complete the entry of a new inspection item. At the same time, there may be a situation where multiple inspection items need to be entered in batches within a short period of time, which will put greater work pressure on the inspectors. Summary of the invention

[0008] The present invention aims to provide a method and system for automatically generating inspection items based on an AI inspection big model, which can efficiently organize the sources of inspection items, reduce the time for studying documents in the process of adding new inspection items, simplify the content that inspectors need to enter and adjust in the process of adding new items, and greatly speed up the speed of adding new inspection items.

[0009] The technical solution adopted by the present invention is:

[0010] A method for automatically generating inspection items based on an AI inspection model includes the following steps:

[0011] Step S1, collecting historical inspection data from different sources related to inspection items, and cleaning and sorting the inspection data to form a data set; wherein the inspection data includes official documents, meeting minutes, policy documents and historical inspection business data;

[0012] Step S2, building a machine deep learning model, and importing the data set for training to obtain an AI inspection model;

[0013] Step S3, uploading relevant document files describing newly added inspection items, extracting local information features from the relevant document files, and importing the local information features into the AI ​​inspection big model to generate basic information of the newly added inspection items; wherein the local information features include key entities, concepts and various nouns; the basic information includes specific item descriptions and responsible units;

[0014] Step S4, associating the local information features and the basic information with the existing inspection knowledge base, analyzing the association relationship, and generating a data view;

[0015] Step S5, matching the data view with the fields in the supervision item information form, and applying the rule engine to generate description information related to the newly added supervision item according to the preset rules; wherein the description information includes the title, content summary, item classification and task description of the decomposed task, responsible unit, and feedback rules;

[0016] Step S6, using the AI ​​inspection big model to verify the description information, and optimizing the description information according to the verification result;

[0017] Step S7, display the optimized description information in the inspection matter information form, and enter it after confirmation.

[0018] Furthermore, the specific process of cleaning and sorting in step S1 includes: deleting duplicate values, processing abnormal values, unifying formats and types, and normalizing data.

[0019] Furthermore, before extracting the local information features from the relevant document files in step S3, the relevant document files are first checked for content and format.

[0020] Furthermore, before checking the content and format of the relevant document file in step S3, the relevant document file is parsed by using a document parser.

[0021] Furthermore, the local information features in step S3 are extracted by using natural language processing technology.

[0022] Furthermore, the local information features and the basic information in step S5 are associated with the existing inspection knowledge base using graph database technology and machine learning algorithms.

[0023] Furthermore, when the optimized description information is displayed in the inspection matter information form in step S7, the inspector may modify the description information in the inspection matter information form according to actual conditions and form a modification record.

[0024] Furthermore, after the modification record in step S7 is formed, it is fed back to the AI ​​inspection big model to optimize the AI ​​inspection big model.

[0025] Based on the same inventive concept, the present invention also provides a system for automatically generating inspection items based on an AI inspection big model, so as to implement the aforementioned method for automatically generating inspection items based on an AI inspection big model, including:

[0026] A collection module, which is used to collect historical inspection data from different sources related to inspection matters, and clean and organize the inspection data to form a data set; wherein the inspection data includes official documents, meeting minutes, policy documents and historical inspection business data;

[0027] A modeling module, which is used to build a machine deep learning model and import the data set for training to obtain an AI inspection model;

[0028] An extraction and generation module, which is used to upload relevant document files describing newly added inspection items, extract local information features from the relevant document files, and import the local information features into the AI ​​inspection big model to generate basic information of the newly added inspection items; wherein the local information features include key entities, concepts and various nouns; the basic information includes specific item descriptions and responsible units;

[0029] An association module, the association module is used to associate the local information features and the basic information with the existing inspection knowledge base, analyze the association relationship, and generate a data view;

[0030] A matching and generating module, which is used to match the data view with the fields in the supervision item information form, and apply a rule engine to generate description information related to the newly added supervision item according to preset rules; wherein the description information includes a title, a summary of the content, a task description of the item classification and decomposition task, a responsible unit, and feedback rules;

[0031] A verification module, the verification module is used to verify the description information using the AI ​​inspection big model, and optimize the description information according to the verification result;

[0032] The display and entry module is used to display the optimized description information in the inspection matter information form and enter it after confirmation.

[0033] The beneficial effects of the present invention are:

[0034] The present invention provides a method and system for automatically generating inspection items based on an AI inspection big model. The method includes: collecting information, training an AI inspection big model, uploading relevant document files describing newly added inspection items and generating basic information, generating data views, generating description information, verification and optimization, display and entry, and other steps. The system includes an acquisition module, a modeling module, an extraction and generation module, an association module, a matching and generation module, a verification module, and a display and entry module. The method and system for automatically generating inspection items based on an AI inspection big model in the present invention can efficiently organize the sources of inspection items, reduce the time for studying documents in the process of adding new inspection items, simplify the content that inspectors need to enter and adjust in the process of adding new items, and greatly speed up the speed of adding new inspection items. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Figure 1 This is a flow chart of the method for automatically generating inspection items based on the AI ​​inspection big model in Example 1. DETAILED DESCRIPTION

[0037] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0038] The disclosure below provides many different embodiments or examples to realize different structures of the present invention. In order to simplify the disclosure of the present invention, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present invention.

[0039] The embodiments of the invention are described in detail below with reference to the accompanying drawings.

[0040] Example 1

[0041] Attached Figure 1 Schematic diagram of the process of automatically generating inspection items based on the AI ​​inspection model in Example 1. Figure 1 As shown in , the method for automatically generating inspection items based on the AI ​​inspection big model includes the following steps:

[0042] (1) Data collection and processing: Collect historical inspection data from different sources related to inspection matters, such as various official documents, meeting minutes, historical inspection business data, policy documents, etc., and clean and organize the inspection data in a unified manner, such as deleting duplicate values, processing outliers, unifying formats and types, and normalizing data, to form a data set that can be used for model training. At the same time, divide the data in the data set into training set, test set, and validation set according to the ratio (for example, 8:1:1).

[0043] (2) Model construction and training: Build a machine deep learning model used to generate inspection information, and train the machine deep learning model based on the prepared data set. During the training phase, the model can learn the mapping relationship between features and targets, so that the model can learn enough rules and patterns from the training set data; during the verification phase, the verification set data can be used to verify the training effect of the model and to adjust the various parameters of the model to ensure that the performance of the model on the location data can be effectively evaluated and adjusted; during the testing phase, the test set data can be used to evaluate the final performance of the model to ensure that the model has sufficient generalization ability and the ability to avoid overfitting. After completing the training, the obtained AI inspection model has the ability to intelligently understand and analyze the input data and generate corresponding inspection information.

[0044] (3) File upload and preprocessing: When inspectors need to add new inspection items, they first upload relevant document files describing the new inspection items, such as various official documents, meeting minutes, etc., and then conduct a preliminary content and format check on the relevant document files. Furthermore, before executing the content and format check, a document parser can be used to parse the uploaded relevant document files to ensure that the uploaded relevant document files meet the processing requirements, avoid conversion and generation problems caused by format errors, etc., and improve the accuracy of subsequent recognition and generation of content.

[0045] (4) Content extraction and understanding: Natural language processing (NLP) technology is used to deeply learn and understand the semantics of relevant document files with the help of AI inspection big model. In view of the local information features such as key entities, concepts and various nouns that may be mentioned in the relevant document files, the entity recognition technology in the AI ​​inspection big model can be introduced to automatically extract relevant information and form the basic information of the inspection items, such as specific item descriptions, responsible units, etc. In this embodiment, the entity recognition technology in the AI ​​inspection big model is applied to greatly improve the accuracy of content extraction and understanding, and ensure that the inspection item information is correctly identified and understood.

[0046] (5) Information association and integration: Associate local information features and basic information with the existing inspection knowledge base, and analyze the association relationship. Apply graph database technology and various machine learning algorithms to efficiently match and predict document information and knowledge base content, discover potential information associations, and form a comprehensive data view to prepare for the generation of subsequent matter information.

[0047] (6) Item information generation and logic construction: Match the data view with the fields in the inspection item information form, and apply the rule engine to preliminarily generate descriptive information of the inspection items according to the preset rules and ensure that it meets the logical and actual needs. The generated item information includes title, content summary, item classification and task description of the decomposed task, responsible unit, feedback rules, etc.

[0048] (7) Item verification and optimization: Use the AI ​​inspection model to conduct preliminary verification of the confidence in the description to ensure its rationality and feasibility, and optimize the description information based on the verification results.

[0049] (8) Fine-tuning of item information: The optimized description information is displayed in the inspection item information form. The inspector can modify the description information in the inspection item information form according to the actual situation and record the modification. At the same time, the modification record is fed back to the AI ​​inspection big model after it is formed, and the behavior details output by the AI ​​inspection big model are adjusted according to the modification behavior of the inspector, so as to achieve adaptive improvement and optimization of the generated content.

[0050] (9) Completed items: After confirming the descriptive information in the inspection item information form, enter it into the inspection system, and the inspector will sign for it and perform other subsequent processes.

[0051] Compared with the prior art, the method of automatically generating inspection items based on the AI ​​inspection big model in this embodiment has the following beneficial effects:

[0052] (1) Efficiently organize the sources of inspection items: With the help of the rapidly generated information on various inspection items, inspectors can manage various sources including official documents, meetings, mayor's mailbox, 12345 hotline, administrative approval, etc. more quickly and efficiently. The AI ​​inspection big model can quickly read various materials and convert them into inspection items. Inspectors no longer need to manage inspection items from multiple different sources separately, but can use the AI ​​inspection big model to perform a series of reading, transcription, management and other operations, thereby improving the efficiency of inspectors in managing inspection items.

[0053] (2) Reduce the time spent on reading documents when adding new inspection items: When adding new inspection items, inspectors no longer need to read all the documents related to the item in full. The AI ​​inspection model will automatically and quickly read the specific content of the relevant document files and extract key information to fill in the inspection item information. Inspectors only need to check the descriptive information related to the new inspection items generated by the AI ​​inspection model based on the key content in the relevant document files.

[0054] (3) Simplify the content that inspectors need to enter and adjust in the process of adding new items: For the entry of new inspection items, inspectors no longer need to manually enter the content based on various documents. Instead, they can use the AI ​​inspection model to automatically generate descriptive information related to the new inspection items, and then fine-tune the information according to the actual situation. This greatly reduces the workload required by inspectors, especially the amount of information entry, and improves the work efficiency of inspectors.

[0055] (4) Significantly speed up the process of adding new items: With the application of the AI ​​inspection model in the process of adding new inspection items, it can significantly speed up the process of inspectors adding new items, especially in batches, thus significantly reducing the repetitive workload of inspectors, improving their work efficiency, and promoting the inspection work at a high speed.

[0056] Example 2

[0057] In this embodiment, a system for automatically generating inspection items based on an AI inspection big model is provided to implement the method for automatically generating inspection items based on an AI inspection big model in the aforementioned embodiment 1. The system for automatically generating inspection items based on an AI inspection big model includes: a collection module, a modeling module, an extraction and generation module, an association module, a matching and generation module, a verification module, and a display and entry module.

[0058] Specifically, the acquisition module is used to collect historical inspection data from different sources related to inspection items, and clean and organize the inspection data to form a data set; wherein the inspection data includes official documents, meeting minutes, policy documents and historical inspection business data. The modeling module is used to build a machine deep learning model, and import the data set for training to obtain an AI inspection model. The extraction and generation module is used to upload relevant document files describing new inspection items, extract local information features from the relevant document files, and import the local information features into the AI ​​inspection model to generate the basic information of the new inspection items; wherein the local information features include key entities, concepts and various nouns; the basic information includes specific item descriptions and responsible units. The association module is used to associate the local information features and the basic information with the existing inspection knowledge base, analyze the association relationship, and generate a data view. The matching and generation module is used to match the data view with the fields in the inspection item information form, and apply the rule engine to generate descriptive information related to the newly added inspection item according to preset rules; wherein the descriptive information includes the title, content summary, task description of the item classification and decomposition task, responsible unit, and feedback rules. The verification module is used to verify the descriptive information using the AI ​​inspection big model, and optimize the descriptive information according to the verification results. The display and entry module is used to display the optimized descriptive information in the inspection item information form, and enter it after confirmation.

[0059] This embodiment greatly reduces the average time required to generate each inspection item, improves the management efficiency of inspection items, replaces the repetitive work of inspection staff to a certain extent, and improves the work efficiency of inspection staff.

Claims

1. A method for automatically generating inspection items based on an AI inspection big model, characterized in that: The following steps are involved: Step S1, collecting historical inspection data from different sources related to inspection items, and cleaning and sorting the inspection data to form a data set; wherein the inspection data includes official documents, meeting minutes, policy documents and historical inspection business data; Step S2, building a machine deep learning model, and importing the data set for training to obtain an AI inspection model; Step S3, uploading relevant document files describing newly added inspection items, extracting local information features from the relevant document files, and importing the local information features into the AI ​​inspection big model to generate basic information of the newly added inspection items; wherein the local information features include key entities, concepts and various nouns; the basic information includes specific item descriptions and responsible units; Step S4, associating the local information features and the basic information with the existing inspection knowledge base, analyzing the association relationship, and generating a data view; Step S5, matching the data view with the fields in the supervision item information form, and applying the rule engine to generate description information related to the newly added supervision item according to the preset rules; wherein the description information includes the title, content summary, item classification and task description of the decomposed task, responsible unit, and feedback rules; Step S6, using the AI ​​inspection big model to verify the description information, and optimizing the description information according to the verification result; Step S7, display the optimized description information in the inspection matter information form, and enter it after confirmation.

2. The method for automatically generating inspection items based on the AI ​​inspection big model according to claim 1 is characterized in that: The specific process of cleaning and sorting in step S1 includes: deleting duplicate values, processing abnormal values, unifying formats and types, and normalizing data.

3. The method for automatically generating inspection items based on the AI ​​inspection big model according to claim 1 is characterized in that: Before extracting the local information features from the relevant document files in step S3, the relevant document files are firstly checked for content and format.

4. The method for automatically generating inspection items based on the AI ​​inspection big model according to claim 3 is characterized in that: Before checking the content and format of the relevant document file in step S3, the relevant document file is first parsed by using a document parser.

5. The method for automatically generating inspection items based on the AI ​​inspection big model according to any one of claims 1 to 4, characterized in that: The local information features in step S3 are extracted by using natural language processing technology.

6. The method for automatically generating inspection items based on the AI ​​inspection big model according to claim 1 is characterized in that: The local information features and the basic information in step S5 are associated with the existing inspection knowledge base using graph database technology and machine learning algorithms.

7. The method for automatically generating inspection items based on the AI ​​inspection big model according to any one of claims 1 to 4 and 6, characterized in that: When the optimized description information is displayed in the inspection item information form in step S7, the inspector may modify the description information in the inspection item information form according to actual conditions and form a modification record.

8. The method for automatically generating inspection items based on the AI ​​inspection big model according to claim 7 is characterized in that: After the modification record in step S7 is formed, it is fed back to the AI ​​inspection big model to optimize the AI ​​inspection big model.

9. A system for automatically generating inspection items based on an AI inspection big model, to implement the method for automatically generating inspection items based on an AI inspection big model as described in any one of claims 1 to 8, characterized in that: include: A collection module, which is used to collect historical inspection data from different sources related to inspection matters, and clean and organize the inspection data to form a data set; wherein the inspection data includes official documents, meeting minutes, policy documents and historical inspection business data; A modeling module, which is used to build a machine deep learning model and import the data set for training to obtain an AI inspection model; An extraction and generation module, which is used to upload relevant document files describing newly added inspection items, extract local information features from the relevant document files, and import the local information features into the AI ​​inspection big model to generate basic information of the newly added inspection items; wherein the local information features include key entities, concepts and various nouns; the basic information includes specific item descriptions and responsible units; An association module, the association module is used to associate the local information features and the basic information with the existing inspection knowledge base, analyze the association relationship, and generate a data view; A matching and generating module, which is used to match the data view with the fields in the supervision item information form, and apply a rule engine to generate description information related to the newly added supervision item according to preset rules; wherein the description information includes a title, a summary of the content, a task description of the item classification and decomposition task, a responsible unit, and feedback rules; A verification module, the verification module is used to verify the description information using the AI ​​inspection big model, and optimize the description information according to the verification result; The display and entry module is used to display the optimized description information in the inspection matter information form and enter it after confirmation.

Images