Case retrieval analysis method and system based on large model

Abstract

The invention discloses a case retrieval and analysis method based on a large model, which comprises the following steps: acquiring statement data of a case, and preprocessing the statement data into first processing data; analyzing the first processing data into second processing data; generating a first event according to the second data, and searching a second event corresponding to the first event in the large model; searching for a first case term corresponding to the first event, searching for a second case term corresponding to the second event, and searching for a third case term corresponding to the second data. And determining a fourth case term related to the case according to the first case term, the second case term and the third case term. And writing a document related to the statement data according to the fourth case term and the second processing data.

Description

Technical Field

[0001] The present invention relates to the field of network information and legal technology, and in particular to a method and system for case retrieval and analysis based on a large model. Background Art

[0002] With the rapid development of artificial intelligence (AI) technology, large models (such as deep learning models) have been widely used in various fields. The legal field is an information-intensive field, and the writing and proofreading of legal documents involve a large number of legal clauses and case analysis. At present, the efficiency of manually writing legal documents is low, and there is a risk of omitting or misusing legal clauses, which makes it difficult to ensure the accuracy and legality of the documents.

[0003] Traditional legal document writing relies on legal professionals to manually search for relevant clauses and cases, which is time-consuming, labor-intensive and error-prone. Especially when faced with complex legal disputes, document writing and proofreading are difficult to complete efficiently at the same time.

[0004] Therefore, there is an urgent need for a technical solution that can automatically generate legal documents and combine semantic analysis with matching of legal clauses to improve the efficiency of legal document generation and proofreading and reduce human errors. Summary of the invention

[0005] In view of the above-mentioned defects, the technical problem solved by the present invention is to provide a method for legal case retrieval and analysis based on a big model, generate legal documents through the big model, combine semantic analysis and matching functions of legal clauses, proofread the legal documents, ensure the accuracy and legality of the content, thereby improving the generation and proofreading time of legal documents, reducing manual omissions, and bringing better legal services and experience to users.

[0006] A first aspect of the present invention provides a method for legal case retrieval and analysis based on a large model, the method comprising:

[0007] S101: Acquire case statement data, and pre-process the statement data into first processed data.

[0008] S102: Parse the first processed data into second processed data.

[0009] S103: Generate a first event according to the second data, search for a second event corresponding to the first event in the large model; search for a first case clause corresponding to the first event, and search for a second case clause corresponding to the second event.

[0010] Find the third case clause corresponding to the second data.

[0011] S104: Determine the fourth case clause related to the present case based on the first case clause, the second case clause and the third case clause.

[0012] S105: Compile a document related to the statement data according to the fourth case clause and the second processing data.

[0013] According to an embodiment of the present invention, the statement data includes: text description, video description, audio description and picture description.

[0014] According to an embodiment of the present invention, the pre-processing of the statement data into first processed data converts the video description, audio description and picture description into text data.

[0015] According to one embodiment of the present invention, the parsing of the first processed data into second processed data in step S102 includes: parsing the text data through natural language processing technology (NLP) and a large language model (LLM), and extracting facts, opinions, controversial points, and related information corresponding to the opinions or controversial points from the text data.

[0016] According to an embodiment of the present invention, the first event is generated according to the second data in step S103, and the first event describing the second data is regenerated again according to the facts, opinions, controversial points, and related information corresponding to the opinions or controversial points of the text data.

[0017] According to an embodiment of the present invention, the step S103 searches for a second event corresponding to the first event in the large model; searches for a first case clause corresponding to the first event, searches for a second case clause corresponding to the second event, and searches for a third case clause corresponding to the second data; including:

[0018] Search for the second event corresponding to the first event stored in the large model event library, as well as the first case clause and the second case clause corresponding to the first event and the second event, respectively. Search for the third case clause corresponding to the facts, opinions, controversial points, and related information corresponding to the opinions or controversial points in the text data.

[0019] According to an embodiment of the present invention, in step S104, determining a fourth case clause related to the present case according to the first case clause, the first case clause and the third case clause includes:

[0020] Determine whether there are overlapping case clauses among the first case clause, the second case clause, and the third case clause, including: if there are overlapping case clauses among the first case clause, the second case clause, and the third case clause, determine that the overlapping case clause among the first case clause, the second case clause, and the third case clause is the fourth case clause.

[0021] If there are no overlapping case clauses among the first case clause, the second case clause, and the third case clause, return to step S103.

[0022] According to one embodiment of the present invention, the method further includes: S106: semantically analyzing the written document and the fourth case clause to determine whether the document matches the fourth case clause and whether there is a logical error.

[0023] If the document matches the fourth case clause and is logically correct, the document is output.

[0024] If the document does not match the fourth case clause and / or there is a logical error, return to step S103.

[0025] A second aspect of the present invention provides a system for case retrieval and analysis based on a large model, the system comprising:

[0026] The acquisition module is used to acquire the statement data of the case and pre-process the statement data into first processed data.

[0027] A parsing module is used to parse the first processed data into second processed data.

[0028] A search module is used to generate a first event based on the second data, and search for a second event corresponding to the first event in the large model; search for a first case clause corresponding to the first event, search for a second case clause corresponding to the second event, and search for a third case clause corresponding to the second data.

[0029] A matching module is used to determine a fourth case clause related to the present case based on the first case clause, the second case clause and the third case clause.

[0030] A writing module is used to write a document related to the statement data according to the fourth case clause and the second processing data.

[0031] A proofreading module is used to perform semantic analysis on the written document and the fourth case clause.

[0032] Determine whether the document matches the terms of the fourth case and whether there are any logical errors.

[0033] The output module is used to output the document.

[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 case retrieval and analysis method based on large models.

[0035] The 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 case retrieval and analysis method based on a large model.

[0036] The beneficial effects provided by the present invention are as follows: first, legal documents are automatically generated through a large model, and combined with semantic analysis and legal clause matching functions, the efficiency of legal document writing and proofreading is greatly improved, and the event of manual search for legal clauses and cases is saved; second, through the legal clause matching function, documents are automatically proofread to ensure the legality and accuracy of the document content, and reduce errors caused by human negligence; at the same time, it is suitable for a variety of legal cases or cases that require text analysis, such as handling contract disputes, labor disputes or property disputes, and text analysis in other educational scenarios can all be used for this method; it is also of great benefit to novices learning in a certain professional field. 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 case retrieval and analysis based on a large model disclosed in an embodiment of the present invention;

[0039] Figure 2 A flowchart of another method for case retrieval and analysis based on a large model disclosed in an embodiment of the present invention;

[0040] Figure 3 A flowchart of another method for case retrieval and analysis based on a large model disclosed in an embodiment of the present invention;

[0041] Figure 4 A system block diagram of case retrieval and analysis based on a large model disclosed in an embodiment of the present invention.

[0042] 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

[0043] 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.

[0044] The first aspect of the present invention provides a method for legal case retrieval and analysis based on a large model, such as Figure 1 The method comprises:

[0045] S101: Obtain the statement data of the case, and pre-process the statement data into the first processing data. The statement data includes: text description, video description, audio description and picture description. For example, in a labor dispute, the plaintiff (employee) requires the defendant (company) to pay the overdue wages and economic compensation. Of course, some users cannot and do not know how to describe the event, or have no focus or clue. They can use pictures, or voice descriptions, or videos. In short, it is not limited to text, video, voice or pictures. It can reflect the events that occurred and represent the disputes or appeals to be resolved. The statement data is pre-processed into the first processing data, and the video description, audio description and picture description are converted into text data. These text data can be scattered or illogical, but they are just the original descriptions of their respective events. Pictures describe the current pictures in text, videos describe the current videos in text, and audio directly converts audio into text descriptions.

[0046] S102: Parse the first processed data into the second processed data. Logically sort out all texts, audio or video, and pictures, and extract key information points, such as facts, opinions, controversial points, and related information corresponding to the opinions or controversial points in the text data. The case or event is described as facts, opinions, controversial points, and information related to the opinions or controversial points. Parsing can be performed through natural language processing technology (NLP) and large language model (LLM).

[0047] S103: Generate a first event according to the second data, search for a second event corresponding to the first event in the large model; search for a first case clause corresponding to the first event, and search for a second case clause corresponding to the second event.

[0048] Find the third case clause corresponding to the second data. Find the second event corresponding to the first event stored in the large model event library, as well as the first case clause and the second case clause corresponding to the first event and the second event, respectively. Find the third case clause corresponding to the facts, opinions, controversial points, and related information corresponding to the opinions or controversial points in the text data.

[0049] Then, the facts, opinions, controversial points and information related to opinions or controversial points in the second data are reassembled into an event through semantic analysis to restore the real event again. This can more objectively and truly restore the cause, process and result of the event, making it easier to handle the subsequent cases without affecting the rationality of the entire case due to evasive details.

[0050] Then, by searching the large model event library for a second event that corresponds to or is similar to the first event, the event similarity of similar cases can be judged to determine whether it is the second event corresponding to the first event. The second event here is the event template in the large model, and is also a time template that has been continuously learned by the large model. After determining the second event, find the second case clause corresponding to the second time, the first case clause corresponding to the first event, and the third case clause corresponding to the second data.

[0051] S104: According to the first case clause, the second case clause and the third case clause, determine the fourth case clause related to this case. Whether in a dispute case or other controversial cases or events, the basis is the legal clause. Here, the corresponding legal case is found through the event to find a breakthrough in solving the case.

[0052] Specifically, based on the first case clause, the first case clause and the third case clause, determine the fourth case clause related to this case, such as Figure 2 As shown, including:

[0053] S104: Determine whether the first case clause, the second case clause, and the third case clause have overlapping case clauses.

[0054] S104a: If there are overlapping case clauses among the first case clause, the second case clause, and the third case clause, determine the overlapping case clauses among the first case clause, the second case clause, and the third case clause as the fourth case clause. The overlapping clauses here include: if there are overlapping case clauses among the first case clause, the second case clause, and the third case clause, or if there are overlapping case clauses between the first case clause and the second case clause, or if there are overlapping case clauses between the first case clause and the third case clause, or if there are overlapping case clauses between the second case clause and the third case clause, the overlapping case clauses can be determined as the final fourth case clause.

[0055] S104b: If there are no overlapping clauses in the first case clause, the second case clause, and the third case clause, the process returns to step S103. If there are no overlapping clauses in the first case clause, the second case clause, and the third case clause, and there are no overlapping clauses in the three cases, the process returns to step 103.

[0056] To the maximum extent possible, the case can be handled and has a basis, and the event in the event library closest to it is found to the maximum extent possible. Until a similar event is found in step S103, the corresponding case is solved in sequence.

[0057] The step S104 further includes: judging whether the statement event is recorded in the large model event library, including:

[0058] A counting threshold is set. If the first case clause, the second case clause, and the third case clause do not have overlapping case clauses, and the number of times of returning to step 103 is greater than the counting threshold, the statement event is recorded in the event library;

[0059] If there are no overlapping case clauses among the first case clause, the second case clause and the third case clause, and the number of times of returning to step 103 is not greater than the threshold, they will not be recorded in the event library.

[0060] If there are no overlapping case clauses among the first case clause, the second case clause and the third case clause, and the number of times returning to step 103 is greater than (counting threshold + 1), then the output is "The current case cannot be accepted temporarily", indicating that the case cannot be accepted temporarily.

[0061] The number of cases in the event library is large enough, and the simulated events or derived times are sufficient to cover all time. What's more, the large model is also constantly learning. After the case is handled, it decides whether to store the event in the event library. If it is a very general and ordinary case with little reference value, it will not be included in the case library. Cases that take a long time to search and have few similar cases can be recorded in the case library. In the long run, if there are enough such cases in the case library and similar cases to new cases can be accurately judged, previous cases can be released to ensure that the case library is always in the learning stage and has sufficient case library sources.

[0062] The method also includes, Figure 3 As shown, in addition to the above method, the present invention also includes the steps

[0063] S106: Perform semantic analysis on the written document and the clauses of the fourth case: determine whether the document matches the clauses of the fourth case and whether there are any logical errors.

[0064] S106a: If the document matches the clause of the fourth case and is logically correct, output the document.

[0065] S106b: If the document does not match the fourth case clause and / or there is a logical error, return to step S103.

[0066] The second aspect of the present invention provides a system 400 for case retrieval and analysis based on a large model, such as Figure 4 As shown, the system 400 includes:

[0067] The acquisition module 401 is used to acquire the statement data of the case and pre-process the statement data into first processed data.

[0068] The parsing module 402 is configured to parse the first processed data into second processed data.

[0069] The search module 403 is used to generate a first event based on the second data, search for a second event corresponding to the first event in the large model; search for a first case clause corresponding to the first event, search for a second case clause corresponding to the second event, and search for a third case clause corresponding to the second data.

[0070] The matching module 404 is used to determine a fourth case clause related to the present case based on the first case clause, the second case clause and the third case clause.

[0071] The writing module 405 is used to write a document related to the statement data according to the fourth case clause and the second processing data.

[0072] The proofreading module 406 is used to perform semantic analysis on the written document and the clauses of the fourth case.

[0073] Determine whether the document matches the terms of the fourth case and whether there are any logical errors.

[0074] The output module 407 is used to output the document.

[0075] 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 case retrieval and analysis method based on a large model.

[0076] The 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 case retrieval and analysis method based on a large model.

[0077] The beneficial effects provided by the present invention are as follows: first, by automatically generating legal documents through a large model, and combining semantic analysis and legal clause matching functions, the efficiency of legal document writing and proofreading is greatly improved, and the event of manually searching for legal clauses and cases is saved; secondly, through the legal clause matching function, the document is automatically proofread to ensure the legality and accuracy of the document content, and reduce errors caused by manual negligence; at the same time, it is suitable for a variety of laws or cases requiring text analysis, such as handling contract disputes, labor disputes or property disputes, and text analysis in other educational scenarios can all be used for this method; it is also very helpful for novices to learn in a certain professional field. Obviously, the above-mentioned specific implementation cases are only examples for illustrating the application of this method, rather than limiting the implementation methods. For general technicians in this field, other different forms of changes and modifications can be made on the basis of the above description to study other related issues. Therefore, the scope of protection of the present invention should be the scope of protection of the claims.

[0078] 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.

[0079] 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.

[0080] 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.

[0081] 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.

[0082] 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.

[0083] 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 case retrieval and analysis based on a large model, characterized in that: The method comprises: S101: Acquire case statement data, and pre-process the statement data into first processed data; S102: parsing the first processed data into second processed data; S103: Generate a first event according to the second data, search for a second event corresponding to the first event in the large model; search for a first case clause corresponding to the first event, search for a second case clause corresponding to the second event, and search for a third case clause corresponding to the second data; S104: Determine a fourth case clause related to the present case based on the first case clause, the second case clause and the third case clause; S105: Compile a document related to the statement data according to the fourth case clause and the second processing data.

2. The method according to claim 1, characterized in that: The statement data includes: text description, video description, audio description and picture description.

3. The method according to claim 2, characterized in that The preprocessing of the statement data into first processed data converts the video description, audio description and picture description into text data.

4. The method according to claim 1, characterized in that: The parsing of the first processed data into second processed data in step S102 includes: parsing the text data through natural language processing technology (NLP) and large language model (LLM), and extracting facts, opinions, controversial points, and related information corresponding to the opinions or controversial points from the text data.

5. The method according to claim 4, characterized in that In step S103, the first event is generated according to the second data, and the first event describing the second data is regenerated again according to the facts, viewpoints, controversial points, and related information corresponding to the viewpoints or controversial points of the text data.

6. The method according to claim 5, characterized in that The step S103 includes searching for a second event corresponding to the first event in the large model; searching for a first case clause corresponding to the first event, searching for a second case clause corresponding to the second event, and searching for a third case clause corresponding to the second data; including: Searching for a second event corresponding to the first event, and a first case clause and a second case clause corresponding to the first event and the second event, respectively, stored in the large model event library; Find the third case clauses corresponding to the facts, opinions, controversial points, and related information corresponding to the opinions or controversial points in the text data.

7. The method according to claim 6, characterized in that In step S104, the fourth case clause related to the present case is determined according to the first case clause, the first case clause and the third case clause, including: Determining whether the first case clause, the second case clause, and the third case clause have overlapping case clauses includes: If there is an overlapping case clause among the first case clause, the second case clause, and the third case clause, the overlapping case clause among the first case clause, the second case clause, and the third case clause is determined to be the fourth case clause; If there are no overlapping case clauses among the first case clause, the second case clause, and the third case clause, return to step S103.

8. The method according to claim 1, characterized in that The method further comprises: S106: Conduct semantic analysis on the written document and the clauses of the fourth case: Determine whether the document matches the clauses of the fourth case and whether there are any logical errors; If the document matches the clause of the fourth case and is logically correct, output the document; If the document does not match the fourth case clause and / or there is a logical error, return to step S103.

9. A system for case retrieval and analysis based on a large model, characterized in that: The system comprises: An acquisition module, used for acquiring statement data of a case, and preprocessing the statement data into first processed data; A parsing module, used for parsing the first processed data into second processed data; A search module, used to generate a first event according to the second data, search for a second event corresponding to the first event in the large model; search for a first case clause corresponding to the first event, search for a second case clause corresponding to the second event, and search for a third case clause corresponding to the second data; A matching module, used for determining a fourth case clause related to the present case according to the first case clause, the second case clause and the third case clause; a writing module, for writing a document related to the statement data according to the fourth case clause and the second processing data; A proofreading module is used to perform semantic analysis on the written document and the fourth case clause: Determine whether the document matches the clauses of the fourth case and whether there are any logical errors; The output module is used to output the document.

10. 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 case retrieval and analysis based on a large model as described in any one of claims 1 to 8.

Images