System document review task allocation method and device

Abstract

The invention discloses a system document review task allocation method and device. The method comprises the steps that a system document is obtained, the system document is input into a named entity recognition model, department anaphora information recorded in the system document is recognized, and paragraph position information of the department anaphora information is obtained through positioning; performing similarity calculation on the identified department anaphora information and each piece of department anaphora information in the dynamic dictionary to obtain a corresponding similarity calculation result; the department anaphora information with the similarity calculation result meeting a preset similarity threshold requirement is screened out, and department entity information corresponding to the department anaphora information is determined; extracting a corresponding paragraph from the system document according to paragraph position information of the department anaphora information; and according to department entity information corresponding to the department reference information, the extracted paragraphs are allocated to corresponding department entities for review. According to the method, the system document review tasks can be quickly distributed, and the entity review efficiency of each department is improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method and device for allocating review tasks for system documents. Background Art

[0002] This section is intended to provide a background or context to the embodiments of the invention recited in the claims. No description herein is admitted to be prior art by inclusion in this section.

[0003] At present, the internal systems and regulations of an enterprise are often written by one department, and other relevant departments will review and provide comments after receiving them. However, since system documents are usually long, it is difficult for each department to quickly locate the content related to their department when reviewing, which not only increases the workload of the review, but also easily leads to some key issues being ignored or omitted. The existing technology cannot quickly disassemble the system documents and assign them to relevant departments for review, resulting in low departmental review efficiency. Therefore, a method for assigning review tasks for system documents is urgently needed to solve the above problems. Summary of the invention

[0004] The embodiment of the present invention provides a method for allocating review tasks of system documents, which is used to realize rapid allocation of review tasks of system documents and improve the efficiency of entity review of each department. The method includes:

[0005] Obtaining a system document, inputting the system document into a named entity recognition model, identifying the department reference information recorded in the system document, and locating the paragraph position information of the department reference information; the named entity recognition model is obtained by training a NER model based on historical system documents and historical department reference information recorded in the historical system documents;

[0006] Calculate the similarity between the identified department reference information and each department reference information in the dynamic dictionary to obtain the corresponding similarity calculation result; the dynamic dictionary is used to update and store the corresponding relationship between the department reference information and the department entity information in real time through a sliding time window algorithm;

[0007] The department reference information whose similarity calculation results meet the preset similarity threshold requirements is screened out, and the department entity information corresponding to the screened department reference information in the dynamic dictionary is used as the department entity information corresponding to the department reference information recorded in the system document;

[0008] According to the paragraph position information of the department reference information, the corresponding paragraphs are extracted from the system document; according to the department entity information corresponding to the department reference information, the extracted paragraphs are assigned to the corresponding department entity for review.

[0009] The embodiment of the present invention also provides a review task allocation device for system documents, which is used to realize the rapid allocation of system document review tasks and improve the efficiency of entity review of each department. The device includes:

[0010] The department reference information acquisition module is used to obtain system documents, input the system documents into the named entity recognition model, identify the department reference information recorded in the system documents, and locate the paragraph position information of the department reference information; the named entity recognition model is obtained by training the NER model based on historical system documents and historical department reference information recorded in the historical system documents;

[0011] A similarity calculation result acquisition module is used to perform similarity calculation on the identified department reference information and each department reference information in the dynamic dictionary to obtain the corresponding similarity calculation result; the dynamic dictionary is used to update and store the corresponding relationship between the department reference information and the department entity information in real time through a sliding time window algorithm;

[0012] The department entity information determination model is used to screen out the department reference information whose similarity calculation results meet the preset similarity threshold requirements, and use the department entity information corresponding to the screened department reference information in the dynamic dictionary as the department entity information corresponding to the department reference information recorded in the system document;

[0013] The review task assignment module is used to extract corresponding paragraphs from the system document according to the paragraph position information of the department reference information; according to the department entity information corresponding to the department reference information, the extracted paragraphs are assigned to the corresponding department entity for review.

[0014] An embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned method for assigning review tasks for institutional documents when executing the computer program.

[0015] An embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for assigning review tasks for the above-mentioned system documents is implemented.

[0016] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the method for assigning review tasks for the above-mentioned system documents is implemented.

[0017] The embodiment of the present invention extracts the department reference information recorded in the system document through the NER model, thereby automatically extracting the specified information from the document; at the same time, the position of the department reference information in the system document is located through the NER model; the department reference information is matched with the department reference information in the dynamic dictionary to obtain the department entity information corresponding to the department reference information; wherein the dynamic dictionary automatically identifies and integrates the correspondence between the new department reference information and the department entity information, and the correspondence between the department reference information and the department entity information is updated in real time through the dynamic dictionary. Then, according to the paragraph position of the department reference information in the system document, the corresponding paragraph is extracted from the system document; according to the department entity information corresponding to the department reference information, the extracted paragraph is assigned to the corresponding department entity for review, thereby quickly and clearly assigning review tasks to the system document, and improving the efficiency of the review of each department entity. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0019] Figure 1 A flowchart of a method for assigning review tasks for system documents in an embodiment of the present invention;

[0020] Figure 2 A flowchart of calculating the similarity between department reference information and department entity information in a dynamic dictionary in an embodiment of the present invention;

[0021] Figure 3 A flowchart of determining department entity information corresponding to department reference information in an embodiment of the present invention;

[0022] Figure 4 It is a schematic diagram of a review task allocation device for system documents in an embodiment of the present invention;

[0023] Figure 5 Schematic diagram of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] To make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0025] Figure 1The flowchart of the method for assigning review tasks of system documents in an embodiment of the present invention includes:

[0026] Step 101, obtaining a system document, inputting the system document into a named entity recognition model, identifying the department reference information recorded in the system document, and locating the paragraph position information of the department reference information; the named entity recognition model is obtained by training a NER model based on historical system documents and historical department reference information recorded in the historical system documents;

[0027] Step 102, calculating the similarity between the identified department reference information and each department reference information in the dynamic dictionary to obtain a corresponding similarity calculation result; the dynamic dictionary is used to update and store the corresponding relationship between the department reference information and the department entity information in real time through a sliding time window algorithm;

[0028] Step 103, filtering out the department reference information whose similarity calculation results meet the preset similarity threshold requirement, and using the department entity information corresponding to the filtered department reference information in the dynamic dictionary as the department entity information corresponding to the department reference information recorded in the system document;

[0029] Step 104: extract corresponding paragraphs from the system document according to the paragraph position information of the department reference information; and assign the extracted paragraphs to the corresponding department entity for review according to the department reference information corresponding to the department entity information.

[0030] Each step is described in detail below.

[0031] In step 101, obtain the system document, input the system document into the named entity recognition model, identify the department reference information recorded in the system document, and locate the paragraph position information of the department reference information; the named entity recognition model is obtained by training the NER model based on historical system documents and historical department references recorded in the historical system documents.

[0032] In a specific embodiment, when training the NER model, data preparation is first performed to collect a large amount of text data from the internal system documents and operation manuals of the enterprise. The text data includes detailed descriptions of the names of each department and the responsibilities of the department to ensure the diversity and coverage of the training data, and the department reference names (informal names of departments) and department entity names (formal names of departments) are respectively labeled. Secondly, the labeled department reference names and department entity names are used to train the NER model during the model training process.

[0033] In step 102, the identified department reference information is similarly calculated with each department reference information in the dynamic dictionary to obtain a corresponding similarity calculation result; the dynamic dictionary is used to update and store the correspondence between the department reference information and the department entity information in real time through a sliding time window algorithm.

[0034] Figure 2 FIG. 4 is a flowchart of calculating the similarity between department reference information and department entity information in a dynamic dictionary in an embodiment of the present invention. Figure 2 As shown, in one embodiment, the identified department reference information is similarly calculated with each department reference information in the dynamic dictionary to obtain a corresponding similarity calculation result, including:

[0035] Step 201, performing word segmentation preprocessing on the department reference information identified from the system document and the department reference information in the dynamic dictionary, and the result of the preprocessing is text features; the text features include semantic features and text structure features;

[0036] Step 202 , respectively calculate the semantic feature similarity of the preprocessing result of the department reference information identified from the system document and the preprocessing result of the department reference information in the dynamic dictionary, and the text structure feature similarity to obtain the corresponding similarity calculation result.

[0037] In a specific embodiment, a dynamic dictionary is used to automatically identify and integrate the correspondence between new department reference information and department entity information. Thus, the correspondence between department reference information and department entity information is updated in real time through the dynamic dictionary. Among them, a sliding time window algorithm is used to monitor the frequency of occurrence of new department reference information in recent institutional documents in real time, and a preset frequency threshold is set. When the frequency of occurrence of a new department reference information in a specific time window exceeds the preset frequency threshold, the new department reference information and the corresponding department entity information are stored in the dynamic dictionary.

[0038] In step 103, the department reference information whose similarity calculation results meet the preset similarity threshold requirements is screened out, and the department entity information corresponding to the screened department reference information in the dynamic dictionary is used as the department entity information corresponding to the department reference information recorded in the system document.

[0039] Figure 3 FIG. 1 is a flow chart of determining department entity information corresponding to department reference information in an embodiment of the present invention. Figure 3 As shown, in one embodiment, the identified department reference information is subjected to similarity calculation with each department reference information in the dynamic dictionary to obtain a corresponding similarity calculation result, and further comprising: the dynamic dictionary includes preset keywords for each department entity information;

[0040] Step 301, establishing corresponding fuzzy sets for preset keywords in each department entity information;

[0041] Step 302, when the similarity calculation result does not meet the preset similarity threshold requirement, construct a membership function for each fuzzy set;

[0042] Step 303, extracting texts related to department reference information from the system document, and calculating the membership of the text to each fuzzy set according to the membership function of each fuzzy set;

[0043] Step 304: when the degree of membership satisfies a preset degree of membership threshold of any fuzzy set, the department entity information corresponding to the fuzzy set is determined as the department entity information corresponding to the department reference information.

[0044] In one embodiment, calculating the membership of the text to each fuzzy set according to the membership function of each fuzzy set includes:

[0045] According to the fuzzy rules and membership function of each fuzzy set, the membership of the text to each fuzzy set is calculated; the fuzzy rules are used to define the relationship between the preset keywords and the department entity information.

[0046] In a specific embodiment, department reference information often contains imprecise or ambiguous expressions, such as "a department similar to sales", "a team responsible for approximately half of customer service", or "a department related to human resources". These department reference information are difficult to obtain corresponding department entity information using traditional matching methods. However, fuzzy rules can be used to effectively determine the department entity information corresponding to the department reference information based on these department reference information.

[0047] The specific implementation steps are as follows:

[0048] 1. Determine the preset keywords for each department entity information in the dynamic dictionary;

[0049] 2. Define fuzzy sets for preset keywords in the entity information of each department; for example, the fuzzy set of "sales relevance", the fuzzy set of "customer service responsibilities", the fuzzy set of "human resources relevance", etc.;

[0050] 3. Define fuzzy rules: Create a series of IF-THEN rules, for example:

[0051] IF sales relevance is high AND customer service responsibility is low THEN it may be the sales department;

[0052] IF HR relevance is medium OR Administration relevance is high THEN it may be the Human Resources department.

[0053] The above process of determining the department entity information corresponding to the department reference information through fuzzy rules can better handle the ambiguity of natural language and improve the accuracy of judgment.

[0054] In one embodiment, calculating the degree of membership of the text to each fuzzy set further includes:

[0055] When the membership degree does not meet the preset membership degree threshold of all fuzzy sets, the big model is used to query the department reference information to obtain the query result; the query result is the department entity information corresponding to the department reference information; the big model is trained by using the historical department reference information and the historical department entity information corresponding to the historical department reference information as the training set.

[0056] In one embodiment, the large model is used to retrieve text related to department reference information to obtain query results, including:

[0057] Construct a prompt message indicating the correspondence between the department reference information and the department entity information, and use the large model to query the correspondence between the department reference information and the department entity information to obtain the query result.

[0058] In a specific embodiment, there is usually some specific department reference information in the system document, which is different from the preset common department reference information. At this time, a prompt is constructed through the big model to inquire. When the big model can give the query results, the corresponding department entity information is determined according to the query results.

[0059] In step 104, according to the paragraph position information of the department reference information, the corresponding paragraph is extracted from the system document; according to the department entity information corresponding to the department reference information, the extracted paragraph is assigned to the corresponding department entity for review.

[0060] In a specific embodiment, according to the paragraph position information of the department reference information, an entire paragraph containing the department reference information is extracted from the system document to ensure the integrity of the paragraph and the coherence of the context. According to the department entity information corresponding to the department reference information, the corresponding paragraph corresponding to the same department is assigned to the auditor of the relevant department. An independent emphasized document is generated for each department, including the paragraphs that need to be reviewed and the corresponding summary of responsibilities. Among them, the responsibility summary process is: using the big model to analyze the paragraphs and summarize the management responsibilities and managed responsibilities of each department. For example: the paragraph assigned to C is: "Each functional department needs to provide xxx.\n Department C needs to collect and review xxxx"; replaced with: "Department C needs to provide xxx.\n Department C needs to collect and review xxxx"; the prompt input to the big model can be "Please summarize the management responsibilities and managed responsibilities of {Department C} based on the following system paragraph. The paragraph text is: Department C needs to provide xxx.\n Department C needs to collect and review xxxx".

[0061] Following the above steps to assign review tasks, auditors in each department only need to review the paragraphs related to themselves, greatly improving review efficiency.

[0062] The present invention also provides a review task allocation device for system documents, as described in the following embodiments. Since the principle of the device to solve the problem is similar to the review task allocation method for system documents, the implementation of the device can refer to the implementation of the review task allocation method for system documents, and the repeated parts will not be repeated.

[0063] Figure 4 This is a schematic diagram of a review task allocation device 400 for a system document in an embodiment of the present invention, the device comprising:

[0064] The department reference information acquisition module 401 is used to acquire the system document, input the system document into the named entity recognition model, identify the department reference information recorded in the system document, and locate the paragraph position information of the department reference information; the named entity recognition model is obtained by training the NER model based on the historical system document and the historical department reference information recorded in the historical system document;

[0065] The similarity calculation result acquisition module 402 is used to perform similarity calculation on the identified department reference information and each department reference information in the dynamic dictionary to obtain the corresponding similarity calculation result; the dynamic dictionary is used to update and store the corresponding relationship between the department reference information and the department entity information in real time through a sliding time window algorithm;

[0066] The department entity information determination model 403 is used to screen out the department reference information whose similarity calculation results meet the preset similarity threshold requirements, and use the department entity information corresponding to the screened department reference information in the dynamic dictionary as the department entity information corresponding to the department reference information recorded in the system document;

[0067] The review task assignment module 404 is used to extract corresponding paragraphs from the system document according to the paragraph position information of the department reference information; and assign the extracted paragraphs to the corresponding department entity for review according to the department entity information corresponding to the department reference information.

[0068] In one embodiment, the similarity calculation result acquisition module 402 is specifically used to:

[0069] The department reference information identified from the system documents and the department reference information in the dynamic dictionary are preprocessed by word segmentation, and the result of the preprocessing is text features; the text features include semantic features and text structure features;

[0070] The semantic feature similarity of the preprocessing result of the department reference information identified from the system document and the preprocessing result of the department reference information in the dynamic dictionary, as well as the text structure feature similarity are calculated respectively to obtain the corresponding similarity calculation results.

[0071] In one embodiment, it further includes a membership degree calculation module, which is specifically used for: the dynamic dictionary includes preset keywords of each department entity information;

[0072] Establish corresponding fuzzy sets for the preset keywords in the entity information of each department;

[0073] When the similarity calculation result does not meet the preset similarity threshold requirement, a membership function is constructed for each fuzzy set;

[0074] Extract the text involving department reference information in the system document, and calculate the membership of the text to each fuzzy set according to the membership function of each fuzzy set;

[0075] When the membership degree satisfies a preset membership degree threshold of any fuzzy set, the department entity information corresponding to the fuzzy set is determined as the department entity information corresponding to the department reference information.

[0076] In one embodiment, calculating the membership of the text to each fuzzy set according to the membership function of each fuzzy set includes:

[0077] According to the fuzzy rules and membership function of each fuzzy set, the membership of the text to each fuzzy set is calculated; the fuzzy rules are used to define the relationship between the preset keywords and the department entity information.

[0078] In one embodiment, a large model query module is also included, which is specifically used to:

[0079] When the membership degree does not meet the preset membership degree threshold of all fuzzy sets, the big model is used to query the department reference information to obtain the query result; the query result is the department entity information corresponding to the department reference information; the big model is trained by using the historical department reference information and the historical department entity information corresponding to the historical department reference information as the training set.

[0080] In one embodiment, the large model query module is specifically used to:

[0081] Construct a prompt message indicating the correspondence between the department reference information and the department entity information, and use the large model to query the correspondence between the department reference information and the department entity information to obtain the query result.

[0082] An embodiment of the present invention further provides a computer device, Figure 5It is a schematic diagram of a computer device in an embodiment of the present invention. The computer device 500 includes a memory 510, a processor 520, and a computer program 530 stored in the memory 510 and executable on the processor 520. When the processor 520 executes the computer program 530, the review task allocation method for the above-mentioned system document is implemented.

[0083] An embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for assigning review tasks for the above-mentioned system documents is implemented.

[0084] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the method for assigning review tasks for the above-mentioned system documents is implemented.

[0085] The embodiment of the present invention extracts the department reference information recorded in the system document through the NER model, thereby automatically extracting the specified information from the document; at the same time, the position of the department reference information in the system document is located through the NER model; the department reference information is matched with the department reference information in the dynamic dictionary to obtain the department entity information corresponding to the department reference information; wherein the dynamic dictionary automatically identifies and integrates the correspondence between the new department reference information and the department entity information, and the correspondence between the department reference information and the department entity information is updated in real time through the dynamic dictionary. Then, according to the paragraph position of the department reference information in the system document, the corresponding paragraph is extracted from the system document; according to the department entity information corresponding to the department reference information, the extracted paragraph is assigned to the corresponding department entity for review, thereby quickly and clearly assigning review tasks to the system document, and improving the efficiency of the review of each department entity.

[0086] The acquisition, transmission, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0087] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0088] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0089] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0090] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0091] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0092] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for assigning review tasks for system documents, characterized in that: include: Obtaining a system document, inputting the system document into a named entity recognition model, identifying the department reference information recorded in the system document, and locating the paragraph position information of the department reference information; the named entity recognition model is obtained by training a named entity recognition NER model based on historical system documents and historical department reference information recorded in the historical system documents; Calculate the similarity between the identified department reference information and each department reference information in the dynamic dictionary to obtain the corresponding similarity calculation result; the dynamic dictionary is used to update and store the corresponding relationship between the department reference information and the department entity information in real time through a sliding time window algorithm; The department reference information whose similarity calculation results meet the preset similarity threshold requirements is screened out, and the department entity information corresponding to the screened department reference information in the dynamic dictionary is used as the department entity information corresponding to the department reference information recorded in the system document; According to the paragraph position information of the department reference information, the corresponding paragraphs are extracted from the system document; according to the department entity information corresponding to the department reference information, the extracted paragraphs are assigned to the corresponding department entity for review.

2. The method according to claim 1, characterized in that The identified department reference information is similarly calculated with each department reference information in the dynamic dictionary to obtain the corresponding similarity calculation results, including: The department reference information identified from the system documents and the department reference information in the dynamic dictionary are preprocessed by word segmentation, and the result of the preprocessing is text features; the text features include semantic features and text structure features; The semantic feature similarity of the preprocessing result of the department reference information identified from the system document and the preprocessing result of the department reference information in the dynamic dictionary, as well as the text structure feature similarity are calculated respectively to obtain the corresponding similarity calculation results.

3. The method according to claim 1, characterized in that Calculating the similarity between the identified department reference information and each department reference information in the dynamic dictionary to obtain a corresponding similarity calculation result, further comprising: the dynamic dictionary includes preset keywords for each department entity information; Establish corresponding fuzzy sets for the preset keywords in the entity information of each department; When the similarity calculation result does not meet the preset similarity threshold requirement, a membership function is constructed for each fuzzy set; Extract the text involving department reference information in the system document, and calculate the membership of the text to each fuzzy set according to the membership function of each fuzzy set; When the membership degree satisfies a preset membership degree threshold of any fuzzy set, the department entity information corresponding to the fuzzy set is determined as the department entity information corresponding to the department reference information.

4. The method according to claim 3, characterized in that According to the membership function of each fuzzy set, the membership of the text to each fuzzy set is calculated, including: According to the fuzzy rules and membership function of each fuzzy set, the membership of the text to each fuzzy set is calculated; the fuzzy rules are used to define the relationship between the preset keywords and the department entity information.

5. The method according to claim 3, characterized in that Calculating the degree of membership of the text to each fuzzy set also includes: When the membership degree does not meet the preset membership degree threshold of all fuzzy sets, the big model is used to query the department reference information to obtain the query result; the query result is the department entity information corresponding to the department reference information; the big model is trained by using the historical department reference information and the historical department entity information corresponding to the historical department reference information as the training set.

6. The method according to claim 5, characterized in that The large model is used to retrieve the text involving department reference information to obtain query results, including: Construct prompt information of the correspondence between department reference information and department entity information, use the big model to query the correspondence between department reference information and department entity information, and obtain the query result.

7. A review task allocation device for system documents, characterized in that: include: The department reference information acquisition module is used to obtain system documents, input the system documents into the named entity recognition model, identify the department reference information recorded in the system documents, and locate the paragraph position information of the department reference information; the named entity recognition model is obtained by training the NER model based on historical system documents and historical department reference information recorded in the historical system documents; A similarity calculation result acquisition module is used to perform similarity calculation on the identified department reference information and each department reference information in the dynamic dictionary to obtain the corresponding similarity calculation result; the dynamic dictionary is used to update and store the corresponding relationship between the department reference information and the department entity information in real time through a sliding time window algorithm; The department entity information determination model is used to screen out the department reference information whose similarity calculation results meet the preset similarity threshold requirements, and use the department entity information corresponding to the screened department reference information in the dynamic dictionary as the department entity information corresponding to the department reference information recorded in the system document; The review task assignment module is used to extract corresponding paragraphs from the system document according to the paragraph position information of the department reference information; according to the department entity information corresponding to the department reference information, the extracted paragraphs are assigned to the corresponding department entity for review.

8. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images