Method and system for generating proofreading of legal text based on large model

Abstract

The invention discloses a method and a system for generating and proofreading legal texts based on a large model. The method comprises the following steps: receiving related description of a legal event, judging whether a logic error exists in a first legal text or not, correcting the logic error of the first legal text, and generating a second legal text. And matching the corrected legal text with the corresponding legal clause, judging whether the legal text is completely in one-to-one correspondence with the corresponding legal clause until the logic is correct, and generating and outputting a third legal text. Compared with a traditional manual processing method, a large amount of time is saved, meanwhile, errors in the legal text can be accurately recognized and corrected through the semantic analysis and legal clause matching technology, human errors are reduced, it is ensured that the generated legal text conforms to related laws and regulations, and legality and compliance of the text are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of legal text generation and proofreading, and in particular to a method and system for legal text generation and proofreading based on a large model. Background Art

[0002] In the legal field, the generation and proofreading of legal texts is a complex task that requires extremely high accuracy. Traditionally, the generation and proofreading of legal texts mainly rely on manual processing, which is not only time-consuming but also prone to human errors. With the advancement of artificial intelligence technology, especially the development of large models, new possibilities are provided for the intelligent generation and proofreading of legal texts.

[0003] The big model has powerful capabilities in natural language processing and can understand and generate complex legal language. However, relying solely on the model's generation capabilities is not enough to ensure the accuracy and legality of legal texts. Therefore, combining the big model with legal clause matching and semantic analysis technology for the generation and proofreading of legal texts has significant technical advantages and practical significance. Summary of the invention

[0004] In view of the above-mentioned defects, the technical problem solved by the present invention is to provide a method for generating and proofreading legal texts based on a big model. The method generates legal texts by utilizing a big model and combines semantic analysis and functional matching of legal clauses to proofread the generated documents to ensure the accuracy and legality of their contents, thereby improving the efficiency of legal text generation and proofreading and reducing human errors.

[0005] The first aspect of the present invention provides a method for generating and proofreading legal texts based on a large model, comprising: S101: receiving a description of a legal event.

[0006] S102: Generate a first legal text according to the description related to the legal event.

[0007] S103: Determine whether there is a logical error in the first legal text, if not, proceed to step S105. If there is a logical error, proceed to step S104.

[0008] S104: Correct the logical errors in the first legal text and generate a second legal text.

[0009] S105: Matching the first legal text or the second legal text with the corresponding legal clauses, including: determining whether the first legal text corresponds to the corresponding legal clauses one-to-one; determining whether the second legal text corresponds to the corresponding legal clauses one-to-one.

[0010] S106: If the first legal text and the second legal text correspond to their corresponding legal clauses one by one, output the first legal text and the second legal text.

[0011] S107: If there is no one-to-one correspondence between the first legal text, the second legal text and the corresponding legal clauses, proofread the first legal text and the second legal text; generate and output a third legal text.

[0012] According to an embodiment of the present invention, the description related to the legal event in step S101 includes: text description, video description, audio description and picture description related to the event.

[0013] According to an embodiment of the present invention, generating a first legal text according to the description related to the legal event in step S102 includes:

[0014] S201: Preprocess the text, video, audio and picture descriptions of the event to generate a text description.

[0015] S202: Extract key information points of the text description, including: statement of facts, appeal, dispute points, and evidence information.

[0016] S203: Generate a first legal text corresponding to the stated facts or claims or disputed points or evidence information.

[0017] According to an embodiment of the present invention, in step S203, generating a first legal text corresponding to the stated facts or claims or disputed points or evidence information includes:

[0018] According to the stated facts, a first corresponding event is searched in an event library, wherein the event library corresponds to a legal text library, and the event corresponds to its corresponding legal text.

[0019] A second corresponding event in the event library is searched according to the dispute point, and a corresponding third event in the event library is searched according to the evidence information.

[0020] A fourth corresponding event is searched in the event library according to the stated facts and the appeal points, and a fifth corresponding event is searched in the event library according to the dispute points and the evidence information.

[0021] Search the sixth corresponding event in the event library according to the stated facts, dispute points and evidence information; calculate the matching degree between the corresponding event and the stated facts, and store the matching degree values ​​between the stated facts and the first corresponding event, the second corresponding event, the third corresponding event, the fourth corresponding event, the fifth corresponding event and the sixth corresponding event.

[0022] According to the matching value, the final corresponding event to the stated fact is determined.

[0023] A legal text corresponding to the corresponding event is searched according to the corresponding event, which is the first legal text.

[0024] According to an embodiment of the present invention, the determining whether there is a logical error in the first legal text in step S103 includes determining whether there is a logical error in the first legal text based on natural language processing (NLP) technology.

[0025] According to an embodiment of the present invention, the step S104 further includes determining whether the first legal text and the second legal text are completely consistent in terms of legal clauses. If the first legal text and the second legal text are inconsistent in terms of legal clauses, the second legal text shall prevail.

[0026] According to an embodiment of the present invention, in step S107, the first legal text and the second legal text are proofread; a third legal text is generated and output. The legal clauses in the first legal text and the second legal text are extracted, and the corresponding legal clauses in the legal text are logically corrected using natural speech processing technology. The corrected legal text is the third legal text.

[0027] A second aspect of the present invention provides a system for generating and proofreading legal texts based on a large model, characterized in that the system comprises: a receiving module for receiving relevant descriptions of legal events.

[0028] The first generating module is used to generate a first legal text according to the description related to the legal event.

[0029] The first judgment module is used to judge whether there is a logical error in the first legal text of the first generation module.

[0030] The second generating module is used to correct the logical errors in the first legal text and generate a second legal text.

[0031] The second judgment module is used to match the first legal text or the second legal text with the corresponding legal clauses, including: judging whether the first legal text corresponds to the corresponding legal clauses one by one; judging whether the second legal text corresponds to the corresponding legal clauses one by one.

[0032] The first output module is used to output the first legal text and the second legal text if the first legal text and the second legal text correspond to their corresponding legal clauses one by one.

[0033] The second output module is used to output a third legal text. If there is no one-to-one correspondence between the first legal text, the second legal text and the corresponding legal clauses, the first legal text and the second legal text are proofread, and a third legal text is generated based on the proofread first legal text and second legal text.

[0034] The third aspect of the present invention provides an intelligent device, including a transmitter, a receiver, a memory and a processor; the memory is used to store computer instructions; the processor is used to run the computer instructions stored in the memory to implement the above-mentioned method for generating and proofreading legal texts based on a large model.

[0035] A fourth aspect of the present invention provides a storage medium, comprising: a readable storage medium and computer instructions, wherein the computer instructions are stored in the readable storage medium; the computer instructions are used to implement the above-mentioned method for generating and proofreading legal texts based on a large model.

[0036] The beneficial effects provided by the present invention are as follows: First, the automatic generation and proofreading technology of large models significantly improves the speed of legal text generation and proofreading, saving a lot of time compared with traditional manual processing methods. Second, through semantic analysis and legal clause matching technology, errors in legal texts can be accurately identified and corrected, reducing human errors. Third, finally, combined with legal clause function matching, it ensures that the generated legal text complies with relevant laws and regulations, improving the legality and compliance of the document. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0038] Figure 1 A flow chart of a method for generating and proofreading a legal text based on a large model disclosed in an embodiment of the present invention;

[0039] Figure 2 A flow chart of a method for generating a first legal text according to a description related to a legal event disclosed in an embodiment of the present invention;

[0040] Figure 3 A block diagram of a system for generating and proofreading legal texts based on a large model disclosed in an embodiment of the present invention.

[0041] The above drawings show clear embodiments of the present disclosure, which will be described in more detail below. These drawings and text descriptions are not intended to limit the scope of the present disclosure in any way, but to illustrate the concepts of the present disclosure to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0042] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0043] In real life, there are many applicable scenarios, such as the generation and proofreading of contract texts, the intelligent generation and proofreading of legal opinions, the automatic generation and proofreading of complaints, etc.

[0044] For example, in the scenario of contract text generation, a commercial contract needs to be generated. The user enters the basic information of the contract through the system, such as the parties to the contract, the main terms and the purpose of the contract. The text generation module of the system generates a preliminary contract text based on this input information. Subsequently, the generated contract text is sent to the semantic analysis module, which identifies potential logical errors in the text. If there are inconsistencies in the terms, the legal terms matching module verifies whether the contract terms comply with relevant laws and regulations, such as the requirements for essential terms in the contract law.

[0045] Finally, the proofreading and revision module revises the contract text according to the analysis results and provides revision suggestions. The user can view and confirm the revised contract text on the user interface module.

[0046] Specifically, Figure 1 As shown, the method for generating and proofreading a legal text based on a large model includes:

[0047] S101: Receive the relevant description of the legal event. The relevant event description here is not limited to the text description, video description, audio description and picture description related to the event, or the touch screen command can replace the text, voice, video or picture description of the event. For example, in a labor dispute, the user needs a lawsuit regarding the labor contract dispute.

[0048] S102: Generate a first legal text based on the description of the legal event. Figure 2 As shown. Includes:

[0049] S201: Preprocess the text, video, audio and picture descriptions of the event to generate a text description.

[0050] For example, some elderly people or disabled people can provide their legal demands by touching the screen buttons, and the system can convert the demands into text.

[0051] S202: extract key information points of the text description, including: statement of facts, appeal, dispute points, and evidence information.

[0052] For example, in a labor dispute case, the statement of facts may be: the employer deducts 1,000 yuan of wages each month due to a certain clause in the examination contract, and the claim requires the employer to pay 4,000 yuan of wages each month in accordance with the labor contract. The disputed point: the employee requires the employer to pay 4,000 yuan of wages each month, and the employer deducts 1,000 yuan from the employee each month in accordance with the contract terms specified by it, Article 23. The evidence information is the monthly pay slip and the labor contract.

[0053] S203: Generate a first legal text corresponding to the stated facts or claims or disputed points or evidence information.

[0054] Obviously, the point of dispute between the worker and the employer is the additional deduction of 1,000 yuan per month. In response to this point of dispute, the model in the event library under the large model is searched. Regarding the employer deducting the worker's salary without reason or in accordance with the contract specified by the company, the specific percentage of the deducted amount to the salary can be used as a key information point to find the closest benchmark event.

[0055] Of course, this is a relatively simple case. When encountering a complex case, you cannot find a similar benchmark event from a single factor. You need to find a benchmark event from multiple dimensions, such as the statement of facts, the appeal point, the dispute point, or the evidence information, in a unified, dual or diversified way. Specifically,

[0056] According to the stated facts, a first corresponding event is searched in an event library, wherein the event library corresponds to a legal text library, and the event corresponds to its corresponding legal text.

[0057] A second corresponding event in the event library is searched according to the dispute point, and a corresponding third event in the event library is searched according to the evidence information.

[0058] A fourth corresponding event is searched in the event library according to the stated facts and the appeal points, and a fifth corresponding event is searched in the event library according to the dispute points and the evidence information.

[0059] Search the sixth corresponding event in the event database based on the stated facts, dispute points and evidence information.

[0060] The matching degree between the corresponding event and the stated fact is calculated, and the matching degree values ​​between the stated fact and the first corresponding event, the second corresponding event, the third corresponding event, the fourth corresponding event, the fifth corresponding event, and the sixth corresponding event are stored.

[0061] According to the matching value, the event with the highest matching degree is determined as the final corresponding event to the stated fact, which is also the closest benchmark event. Moreover, the event of this statement is included in the event library as the benchmark event for the next time or other new events. In short, through the continuous addition of new events and feedback, the system can continue to learn, and when new events are added, it can find the benchmark event in the fastest way. In this way, a lot of data processing time will be saved in case handling, and redundant data processing of one-to-one events and legal texts will be minimized.

[0062] A legal text corresponding to the corresponding event is searched according to the corresponding event, which is the first legal text.

[0063] S103: Determine whether there is a logical error in the first legal text. If there is no logical error, proceed to step S105; if there is a logical error, proceed to step S104.

[0064] Use natural language processing technology to perform basic logical error processing such as word segmentation, part-of-speech tagging, and dependency analysis on the text. Through these technologies, the system can understand the structure of the text and identify the basic components of the sentence, such as the subject, predicate, and object. Check the contextual consistency in the text. Contextual consistency refers to whether the content between sentences in a paragraph or chapter is consistent and whether it follows a reasonable cause-and-effect relationship.

[0065] Ensure that entities (such as names of people, places, and companies) that appear in the text are consistent in context. For example, if a legal text mentions "Zhang San and Li Si signed a contract" and then "Zhang San and Wang Wu's contract terms" appear, the system needs to check whether the entities of these contracts are consistent.

[0066] Analyze whether the semantic relationship in the text is consistent. For example, whether the "seller" mentioned in the contract refers to the same entity throughout the text to avoid inconsistent references to the "seller" in different locations.

[0067] S104: Correct the logical errors in the first legal text and generate a second legal text. Determine whether the first legal text and the second legal text are completely consistent in terms of legal clauses. If the first legal text and the second legal text are inconsistent in terms of legal clauses, the second legal text shall prevail.

[0068] Because in legal texts, logical consistency and causality are very important. For example, if the first part of the document describes that a contract was executed, but the later part mentions that the contract was not executed, this inconsistency will be detected.

[0069] Guaranteed, these low-level errors will not appear in the legal text.

[0070] S105: Matching the first legal text or the second legal text with the corresponding legal clauses, including: determining whether the first legal text corresponds to the corresponding legal clauses one-to-one; determining whether the second legal text corresponds to the corresponding legal clauses one-to-one.

[0071] S106: If the first legal text and the second legal text correspond to their corresponding legal clauses one by one, output the first legal text and the second legal text.

[0072] The one-to-one correspondence here can be understood as follows. For example, the legal text describes that according to Articles 46, 40 and 42 of the Labor Contract Law, the employer must pay the employee three months' salary totaling 18,000 yuan. The legal terms described in the legal text, Articles 46, 40 and 42 of the Labor Contract Law, must be exactly the same as the legal terms, including data such as payment for three months. What is being said here is not a replicative complete consistency. For example, the three months' salary here is based on those terms.

[0073] Natural language processing technology can also be used here to search for logical errors in the legal text and its corresponding legal clauses. If it is found that there are logical errors in the text logic of the first legal text and the second legal text and the legal clauses in which they appear, the logic is corrected, and the corrected legal text is the third legal text.

[0074] The second aspect of the present invention provides a system 800 for generating and proofreading legal texts based on a large model, such as Figure 3 The system comprises: a receiving module 801, which is used to receive the relevant description of the legal event.

[0075] The first generating module 802 is used to generate a first legal text according to the description related to the legal event.

[0076] The first judgment module 803 is used to judge whether there is a logical error in the first legal text of the first generation module.

[0077] The second generating module 804 is used to correct the logical errors in the first legal text and generate a second legal text.

[0078] The second judgment module 805 is used to match the first legal text or the second legal text with the corresponding legal clauses, including: judging whether the first legal text corresponds to the corresponding legal clauses one by one; judging whether the second legal text corresponds to the corresponding legal clauses one by one.

[0079] The first output module 806 is used to output the first legal text and the second legal text if the first legal text and the second legal text correspond to their corresponding legal clauses one by one.

[0080] The second output module 807 is used to output a third legal text. If there is no one-to-one correspondence between the first legal text, the second legal text and the corresponding legal clauses, the first legal text and the second legal text are proofread, and a third legal text is generated based on the proofread first legal text and second legal text.

[0081] The third aspect of the present invention provides an intelligent device, including a transmitter, a receiver, a memory and a processor; the memory is used to store computer instructions; the processor is used to run the computer instructions stored in the memory to implement the above-mentioned method for generating and proofreading legal texts based on a large model.

[0082] A fourth aspect of the present invention provides a storage medium, comprising: a readable storage medium and computer instructions, wherein the computer instructions are stored in the readable storage medium; the computer instructions are used to implement the above-mentioned method for generating and proofreading legal texts based on a large model.

[0083] The beneficial effects provided by the present invention are as follows: First, the automatic generation and proofreading technology of large models significantly improves the speed of legal text generation and proofreading, saving a lot of time compared with traditional manual processing methods. Second, through semantic analysis and legal clause matching technology, errors in legal texts can be accurately identified and corrected, reducing human errors. Third, finally, combined with legal clause function matching, it ensures that the generated legal text complies with relevant laws and regulations, improving the legality and compliance of the document.

[0084] Obviously, the above specific implementation cases are only examples for illustrating the application of the present method, rather than limiting the implementation methods. For those skilled in the art, other different forms of changes and modifications can be made on the basis of the above description to study other related issues. Therefore, the protection scope of the present invention should be the protection scope of the claims.

[0085] A person of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk, etc., various media that can store program codes.

[0086] The electronic device and other 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. Ordinary technicians in this field can understand and implement it without creative work.

[0087] 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, it 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.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, rather than to limit them. Although the embodiments of the present invention have 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 replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

[0089] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are to be regarded as exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.

[0090] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A method for generating and proofreading legal texts based on a large model, characterized in that: The method comprises: S101: receiving a description of a legal event; S102: Generate a first legal text according to the description related to the legal event; S103: Determine whether the first legal text contains any logical errors; if not, proceed to step S105; if present, proceed to step S104; S104: Correcting logical errors in the first legal text to generate a second legal text; S105: Determine whether the first legal text and the second legal text correspond to the corresponding legal clauses one by one; S106: If the first legal text and the second legal text correspond to the corresponding legal clauses one by one, output the first legal text and the second legal text; S107: If there is no one-to-one correspondence between the first legal text, the second legal text and the corresponding legal clauses, proofread the first legal text and the second legal text; generate and output a third legal text.

2. The method according to claim 1, characterized in that The description related to the legal event in step S101 includes: text description, video description, audio description and picture description related to the event.

3. The method according to claim 2, characterized in that The step S102 generates a first legal text according to the description of the legal event, including: S201: Preprocessing the text, video, audio and picture descriptions of the event to generate a text description; S202: extracting key information points of the text description, including: statement of facts, appeal, dispute points, and evidence information; S203: Generate a first legal text corresponding to the stated facts or claims or disputed points or evidence information.

4. The method according to claim 3, characterized in that In step S203, a first legal text corresponding to the stated facts or claims or disputed points or evidence information is generated, including: Searching for a first corresponding event in an event library according to the stated fact, wherein the event library corresponds to a legal text library, and the event corresponds to its corresponding legal text; Searching for a second corresponding event in the event library according to the dispute point, and searching for a corresponding third event in the event library according to the evidence information; Searching for a fourth corresponding event in the event database according to the stated facts and the appeal points, and searching for a fifth corresponding event in the event database according to the dispute points and the evidence information; Searching for a sixth corresponding event in the event library according to the stated facts, dispute points, and evidence information; calculating a matching degree between the corresponding event and the stated facts, and storing matching degree values ​​between the stated facts and the first corresponding event, the second corresponding event, the third corresponding event, the fourth corresponding event, the fifth corresponding event, and the sixth corresponding event; Determine the final corresponding event to the stated fact according to the matching value; A legal text corresponding to the corresponding event is searched according to the corresponding event, which is the first legal text.

5. The method according to claim 4, characterized in that The step S103 of determining whether there is a logical error in the first legal text includes determining whether there is a logical error in the first legal text based on natural language processing (NLP) technology.

6. The method according to claim 1, characterized in that The step S104 further includes determining whether the first legal text and the second legal text are completely consistent in terms of legal clauses; If the legal provisions based on the first legal text and the second legal text are inconsistent, the second legal text shall prevail.

7. The method according to claim 6, characterized in that In the step S107, the first legal text and the second legal text are proofread; a third legal text is generated and output; Extracting the legal clauses in the first legal text and the second legal text, and using natural speech processing technology to perform logical correction on the corresponding legal clauses in the legal texts; The revised legal text is the third legal text.

8. A system for generating and proofreading legal texts based on a large model, characterized in that: The system comprises: A receiving module, used to receive relevant descriptions of legal events; A first generating module, used to generate a first legal text according to the description related to the legal event; A first judgment module, used to judge whether there is a logical error in the first legal text of the first generation module; A second generating module, used for correcting the logical errors in the first legal text and generating a second legal text; The second judgment module is used to match the first legal text or the second legal text with the corresponding legal clause, including: judging whether the first legal text corresponds to the corresponding legal clause one by one; judging whether the second legal text corresponds to the corresponding legal clause one by one; A first output module, used to output the first legal text and the second legal text, if the first legal text and the second legal text correspond to their corresponding legal clauses one by one; The second output module is used to output a third legal text. If there is no one-to-one correspondence between the first legal text, the second legal text and the corresponding legal clauses, the first legal text and the second legal text are proofread, and a third legal text is generated based on the proofread first legal text and second legal text.

9. A smart device, characterized in that: include: transmitter, receiver, memory and processor; The memory is used to store computer instructions; the processor is used to execute the computer instructions stored in the memory to implement the method for generating and proofreading legal texts based on a large model as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: include: A readable storage medium and computer instructions, wherein the computer instructions are stored in the readable storage medium; The computer instructions are used to implement the method for generating and proofreading legal texts based on a large model as described in any one of claims 1 to 7.

Images