CN-121981094-A - Document auxiliary writing method and device, electronic equipment and storage medium

CN121981094ACN 121981094 ACN121981094 ACN 121981094ACN-121981094-A

Abstract

The application relates to a document auxiliary writing method, a device, electronic equipment and a storage medium, wherein the method comprises the steps of displaying a first document, defect information of the first document and a first guidance problem of the first document to a user; the method comprises the steps of obtaining correction comments of a user aiming at a first guidance problem, adjusting a first document based on the correction comments of the first guidance problem to obtain an adjusted first document, responding to first confirmation indication information of the user aiming at the adjusted first document, taking the adjusted first document as a second document, generating and displaying a second guidance problem of the second document, obtaining feedback information of the user aiming at the second guidance problem, and optimizing the second document based on the feedback information of the second guidance problem to obtain a target document corresponding to task description information. The method realizes that the task description information which is input from the preliminary is guided to be input by the user step by step to obtain the target document, and further improves the quality of document auxiliary writing.

Inventors

ZHANG WENDA
CAI JIAQI
SHEN XINWEN
Yang Qiaofu
ZHANG XIN

Assignees

中国国际海运集装箱(集团)股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260109

Claims (10)

1. A method of document assisted writing, the method comprising: Extracting at least one first document element and key information corresponding to the first document element based on task description information provided by a user; summarizing and sorting based on the at least one first document element and key information corresponding to the first document element to obtain a first document corresponding to the task description information, and determining defect information existing in the first document; Presenting the first document, defect information of the first document and a first guiding problem of the first document to the user, wherein the first guiding problem is used for guiding the user to provide correction comments for the first document and/or the defect information; acquiring the correction opinion of the user aiming at the first guiding problem, so as to adjust the first document based on the correction opinion of the first guiding problem, and obtain an adjusted first document; Responding to first confirmation indication information of a user for the adjusted first document, taking the adjusted first document as a second document, and generating and displaying a second guiding problem of the second document, wherein the second guiding problem comprises an optimization problem related to at least one document content in the second document; And acquiring feedback information of the user aiming at the second guidance problem, so as to optimize the second document based on the feedback information of the second guidance problem, and obtain a target document corresponding to the task description information.
2. The method of claim 1, wherein the first introductory questions comprise a first question for directing a user to provide a correction opinion on the first document and/or defect information of the first document and a second question for directing a user to indicate whether to continue correcting or terminate writing on the first document.
3. The method of claim 2, wherein the method further comprises, prior to the taking the adjusted first document as the second document: And carrying out at least one round of dialogue through a first model based on the first document to obtain the adjusted first document, wherein the at least one round of dialogue comprises: Processing the first document of the current round of dialogue and the history dialogue content, and determining defect information of the first document of the current round of dialogue; generating a first guiding problem of the first document of the current round of dialogue based on the defect information of the first document of the current round of dialogue, displaying the first guiding problem to ask the user, and acquiring feedback information of the user for the first guiding problem; When the user feeds back the first document of the local dialogue and/or the correction opinion of the defect information of the first document of the local dialogue aiming at the first problem, the first document of the local dialogue is adjusted based on the correction opinion of the first problem, so that an adjusted first document of the local dialogue is obtained; and taking the adjusted first document as the first document of the next round of dialogue, and circularly executing the steps until the user feeds back first confirmation instruction information for stopping correcting the first document of the current round of dialogue aiming at the second problem, wherein the adjusted first document of the current round of dialogue is taken as the second document.
4. The method of claim 1, wherein the second introductory questions include a third question for asking an optimization question associated with at least one document content in the second document and a fourth question for guiding a user to indicate whether to continue optimizing or terminate writing the second document.
5. The method according to claim 4, wherein before obtaining the target document corresponding to the task description information, the method further comprises: And carrying out at least one round of dialogue through a second model based on the second document to obtain a target document corresponding to the task description information, wherein the at least one round of dialogue comprises: processing a second document of the current round of dialogue and historical dialogue content, generating a second guiding problem of the second document of the current round of dialogue, displaying the second guiding problem to ask the user, and acquiring feedback information of the user for the second guiding problem; When the user feeds back feedback information aiming at the optimization problem, optimizing the second document of the current round of dialogue based on the feedback information of the optimization problem to obtain a new second document of the optimized current round of dialogue; And circularly executing the steps until the user feeds back second confirmation instruction information for stopping optimizing the second document of the current round of dialogue aiming at the fourth problem, and taking the new second document of the current round of dialogue as a target document corresponding to the task description information.
6. The method of claim 5, wherein the optimization problem comprises at least one problem sequence number and a problem corresponding to the problem sequence number, the second guided problem further comprises a first optimization mode prompt message and a second optimization mode prompt message, the first optimization mode prompt message is used for prompting the user to feed back feedback information in a preset format to reply to the optimization problem, the second optimization mode prompt message is used for indicating the user to feed back a target problem sequence number to trigger a third model to process the problem corresponding to the target problem sequence number, The optimizing the second document based on the feedback information of the second guiding problem to obtain the target document corresponding to the task description information includes: Under the condition that the feedback information comprises a target problem sequence number is detected, processing a problem corresponding to the target problem sequence number through the third model based on the target problem sequence number in the feedback information to obtain and display target content; Acquiring indication information of the user on the target content, optimizing the second document based on the indication information to obtain a target document corresponding to the task description information, wherein the third model comprises at least one large language model; The indication information includes at least one of first indication information, second indication information, and third indication information; the first indication information is used for representing that all the target content or part of the target content selected by a user is incorporated into the second document; The second indication information is used for representing that the target content is modified based on modification information of a user, and the modified target content is incorporated into the second document; the third indication information is used for indicating that the operation of incorporating the target content into the second document is not performed.
7. The method of any one of claims 1 to 6, wherein the defect information includes at least one of first defect information, second defect information, and third defect information, Wherein the first defect information is defect information associated with a document structure of the first document, The second defect information is defect information associated with a linguistic representation of the first document, The third defect information is defect information associated with document content of the first document; The determining defect information existing in the first document includes: And detecting the first document based on defect detection logic corresponding to the defect information type, and determining defect information of the first document.
8. A document auxiliary writing apparatus, the apparatus comprising: The acquisition module is used for extracting at least one first document element and key information corresponding to the first document element based on task description information provided by a user; The first questioning module is used for summarizing and sorting based on the at least one first document element and key information corresponding to the first document element to obtain a first document corresponding to the task description information, and determining defect information in the first document; Presenting the first document, defect information of the first document and a first guiding problem of the first document to the user, wherein the first guiding problem is used for guiding the user to provide correction comments for the first document and/or the defect information; The first determining module is used for acquiring the correction opinion of the user aiming at the first guiding problem so as to adjust the first document based on the correction opinion of the first guiding problem to obtain an adjusted first document; The second questioning module is used for responding to first confirmation indication information of a user for the adjusted first document, taking the adjusted first document as a second document, generating and displaying a second guiding problem of the second document, wherein the second guiding problem comprises an optimization problem related to at least one document content in the second document; And the second determining module is used for acquiring feedback information of the user aiming at the second guidance problem, so as to optimize the second document based on the feedback information of the second guidance problem and obtain a target document corresponding to the task description information.
9. An electronic device, comprising: a memory having a computer program stored thereon; A processor for executing the computer program in the memory to implement the steps of the method of any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that a computer program is stored thereon, which, when being executed by a processor, realizes the steps of the method according to any of claims 1 to 7.

Description

Document auxiliary writing method and device, electronic equipment and storage medium Technical Field The present application relates to the field of artificial intelligence technologies, and in particular, to a document assisted writing method and apparatus, an electronic device, and a storage medium. Background As enterprise research and development activities become more frequent, the demand for writing high quality technical documents (e.g., technical specifications, project legislation reports, product specifications, etc.) continues to grow. However, the document writing work is usually undertaken by technicians, and although the technical staff has professional field knowledge, the structured writing experience and document specification awareness are generally lacking, so that the submitted original content often has the problems of information fragmentation, key detail deletion, logic incoherence, fuzzy expression and the like, and is difficult to directly use, so that a large amount of manual intervention is still required for the original content, and the quality and the efficiency of document writing are low. Disclosure of Invention The embodiment of the application provides a document auxiliary writing method, a device, electronic equipment and a storage medium, which are used for improving the quality and efficiency of document auxiliary writing so as to solve the problems. To achieve the above object, according to a first aspect of the present application, there is provided a document auxiliary writing method, the method comprising: Extracting at least one first document element and key information corresponding to the first document element based on task description information provided by a user; summarizing and sorting based on the at least one first document element and key information corresponding to the first document element to obtain a first document corresponding to the task description information, and determining defect information existing in the first document; Presenting the first document, defect information of the first document and a first guiding problem of the first document to the user, wherein the first guiding problem is used for guiding the user to provide correction comments for the first document and/or the defect information; acquiring the correction opinion of the user aiming at the first guiding problem, so as to adjust the first document based on the correction opinion of the first guiding problem, and obtain an adjusted first document; Responding to first confirmation indication information of a user for the adjusted first document, taking the adjusted first document as a second document, and generating and displaying a second guiding problem of the second document, wherein the second guiding problem comprises an optimization problem related to at least one document content in the second document; And acquiring feedback information of the user aiming at the second guidance problem, so as to optimize the second document based on the feedback information of the second guidance problem, and obtain a target document corresponding to the task description information. Optionally, the first guiding problem includes a first problem for guiding a user to provide a correction opinion on the first document and/or defect information of the first document, and a second problem for guiding a user to indicate whether to continue correction on the first document or terminate writing. Optionally, before the step of using the adjusted first document as the second document, the method further includes: And carrying out at least one round of dialogue through a first model based on the first document to obtain the adjusted first document, wherein the at least one round of dialogue comprises: Processing the first document of the current round of dialogue and the history dialogue content, and determining defect information of the first document of the current round of dialogue; generating a first guiding problem of the first document of the current round of dialogue based on the defect information of the first document of the current round of dialogue, displaying the first guiding problem to ask the user, and acquiring feedback information of the user for the first guiding problem; When the user feeds back the first document of the local dialogue and/or the correction opinion of the defect information of the first document of the local dialogue aiming at the first problem, the first document of the local dialogue is adjusted based on the correction opinion of the first problem, so that an adjusted first document of the local dialogue is obtained; and taking the adjusted first document as the first document of the next round of dialogue, and circularly executing the steps until the user feeds back first confirmation instruction information for stopping correcting the first document of the current round of dialogue aiming at the second problem, wherein the adjusted first document of the