Search

CN-122020530-A - Multi-mode case analysis intelligent generation system and method for lawsuits

CN122020530ACN 122020530 ACN122020530 ACN 122020530ACN-122020530-A

Abstract

The invention relates to the technical field of computer judicial practice and discloses a multi-mode case analysis intelligent generation system and method for lawsuits, wherein the system automatically collects and processes multi-source materials such as audio, images, scanning pieces, electronic documents and the like related to the cases, converts the multi-source materials into unified texts through recognition and analysis, further intelligently extracts factors such as principal, litigation requests and key facts from the unified texts, constructs a structured case model integrating principal-legal relation-request-fact-evidence, automatically performs request right basic analysis, legal clause matching and evidence association based on the model, automatically schedules and completes all associated analysis tasks after a user selects a document type according to a predefined intelligent mapping relation, and generates various litigation documents conforming to judicial practice formats and expression specifications.

Inventors

  • CHENG ZIYANG

Assignees

  • 上海盛熙法索人工智能科技有限公司

Dates

Publication Date
20260512
Application Date
20260123

Claims (10)

  1. 1. The utility model provides a lawsuit lawyer oriented multimode case analysis intelligence generation system which characterized in that includes: The multi-mode case data acquisition module is used for receiving various forms of original materials related to the target case; The multi-mode preprocessing and unified representation module is connected with the multi-mode case data acquisition module and is used for carrying out voice recognition, optical character recognition and document analysis on the received original materials and uniformly storing the processed texts and associated metadata into a preset case original corpus; The case element extraction module is connected with the case original corpus and is used for automatically extracting case elements from the case original corpus based on a pre-training language model and field rules and storing extraction results into a case structured database in a structured form, wherein the case elements at least comprise principal information, case types, litigation requests, key facts, time axes and dispute focuses; The request right foundation and law applicability analysis module is connected with the case structured database and is used for automatically identifying legal relations corresponding to litigation requests and matching legal terms according to elements in the case structured database and combining a preset legal knowledge base to generate mapping relations among the litigation requests, the legal relations, the legal terms and key facts; The evidence association and evidence list generation module is connected with the case structured database and is used for carrying out type recognition and gist extraction on evidence materials in the case, establishing an association model among facts, evidence and litigation requests, generating a standardized evidence list based on the association model, marking evidence gaps and giving supplementary evidence collection prompts; the document generation and multi-document type arrangement module is respectively connected with the case structured database, the request right basis and law applicability analysis module and the evidence association and evidence list generation module and is used for automatically generating corresponding litigation document drafts according to the selected document types based on the case structured database, the mapping relation and the association model; The interactive editing and traceable display module is connected with the document generation and multi-document type arrangement module and used for providing a visual interface to display and support modification of the generated document draft and synchronously updating the modification content to the case structured database.
  2. 2. The intelligent litigation-oriented multimodal case analysis generating system of claim 1, wherein the raw material received by the multimodal case data collection module includes at least one of an audio file, an electronic document, an image-like material, and structured data.
  3. 3. The intelligent litigation-oriented multi-modal case analysis generating system according to claim 1, wherein the process of generating the litigation document by the document generation and multi-document type arrangement module adopts a mode of combining template constraint and intelligent generation, and the intelligent litigation-oriented multi-modal case analysis generating system comprises loading a structured format template according to a selected document type, filling a framework of the template by utilizing contents in a case structured database, and calling a language model subjected to field fine tuning to carry out color rendering and logic optimization on the filled natural language part.
  4. 4. The litigation lawyer-oriented multimodal case analysis intelligent generation system of claim 1, further comprising: the quality evaluation and iteration optimization module is connected with the document generation and multi-document type arrangement module and is used for automatically checking the generated document draft according to preset rules and marking or triggering and correcting the parts which do not meet the requirements.
  5. 5. The lawsuit lawyer-oriented multi-modal case analysis intelligent generation system of claim 1, wherein the system predefines a mapping relationship between document types and analysis modules, and wherein when a user selects a target document type in the interface of the interactive editing and traceable presentation module, the system automatically triggers and invokes one or more pre-analysis modules required for generating the target document type to perform an analysis task according to the mapping relationship.
  6. 6. A lawsuit lawyer oriented multi-modal case analysis intelligent generation method, characterized by being based on the intelligent generation method of the system of any one of claims 1-5, comprising the following steps: s1, acquiring multi-modal input materials of a target case, converting the multi-modal input materials into a unified text and metadata form through multi-modal preprocessing and unified representation processing, and storing the unified text and metadata form into an original corpus of the case; S2, extracting case elements from the case original corpus to form structured case data and storing the structured case data into a case structured database; S3, based on the case structured database, carrying out request right foundation and law applicability analysis to generate a litigation request, legal relation, legal clause and key fact mapping relation; s4, processing evidence materials in cases, establishing a correlation model between facts, evidence and litigation requests, and generating a standardized evidence list; S5, responding to the selection of a user on a target document type, automatically generating a corresponding litigation document draft based on the case structured database, the mapping relation and the association model, and And S6, providing an interactive editing interface, responding to the modification operation of the user on the manuscript, and synchronously updating the modification content to the case structured database.
  7. 7. The intelligent litigation-oriented multi-modal case analysis generation method of claim 6, wherein the automatically generating the corresponding litigation document draft after responding to the user selection of the target document type comprises: determining one or more pre-analysis modules required to be called for generating the selected target document type according to the mapping relation between the predefined document type and the analysis modules; Automatically triggering and sequentially executing the determined prepositive analysis modules to generate corresponding intermediate analysis results, wherein the intermediate analysis results at least comprise one or more of a case summary, a case diary, a request right analysis report and an evidence list; And generating the document draft of the target document type based on the case structured database and the intermediate analysis result.
  8. 8. The lawsuit lawyer-oriented multi-modal case analysis intelligent generation method of claim 6, wherein the pre-analysis module comprises a plurality of case summary generation modules, a case big event combing module, a request right basis pre-analysis module, a law rationality examination module, a case route determination module, an anti-dialectic analysis module and an evidence list combing module.
  9. 9. The intelligent generating method for multi-modal case analysis of litigation-oriented lawyers of claim 6, wherein the specific process of generating the litigation document draft in step S5 comprises: Loading a structured format template corresponding to the selected target document type; filling the content in the case structured database into a corresponding frame of the template; and calling the language model subjected to field fine tuning, and generating or optimizing text of the natural language part in the template after filling to form a manuscript.
  10. 10. The litigation lawyer-oriented multimodal case analysis intelligent generation method of claim 6, further comprising, after step S6: and S7, automatically checking format normalization and logic consistency of the finally generated document, and prompting potential defects.

Description

Multi-mode case analysis intelligent generation system and method for lawsuits Technical Field The invention relates to the technical field of computer judicial practice, in particular to a multi-mode case analysis intelligent generation system and method for lawsuits. Background With the increasing complexity of social and economic activities, the number of civil business litigation cases is continuously increased, the fact chain of the cases is prolonged, evidence materials are numerous, the information of the population is disordered, and huge information integration and litigation document writing pressure is brought to first-line litigation lawyers. Under the prior art conditions, lawyers rely mainly on the following ways to handle cases, but all have limitations: The traditional lawyers need to acquire dictation information through telephone and face-to-face interviews, and combine scattered materials from multiple sources such as WeChat, mail, contract, financial certificate and the like to manually comb to form a big diary, a case summary and an evidence list. This approach is highly dependent on lawyers' personal experience and effort, inefficient and time consuming, and is highly prone to human factor-induced missing of critical information or deviation in understanding. And the current part of legal information service providers provide legal regulations, database retrieval functions of judge documents and general document templates with fixed formats, lawyers can manually rewrite and splice on the basis of retrieval results, however, the tools still essentially belong to the primary stage of 'retrieval-template filling', intelligent extraction of case core elements, deep analysis of request right basis and automatic association of evidence and facts cannot be automatically completed on the basis of specific unstructured case materials, and continuous and integrated generation of multiple types of litigation documents cannot be realized. Moreover, with the development of artificial intelligence technology, a general Large Language Model (LLM) is tried to be used for document drafting assistance, lawyers can input cases in natural language description to obtain preliminary draft, but the general model cannot directly process multi-mode original materials such as audio, scanning pieces and pictures, the input threshold is high, specific case elements, evidence chain logics and fact relations in the legal field are lack of accurate grasp, fact accuracy and logic tightness of generated contents are insufficient, deep customization aiming at characteristics of litigation procedures, court document format specifications and legal term habits is lacking, a gap often exists between a generated result and actual requirement, and unified structured case knowledge representation is lacking as an intermediate layer, so that each link of 'case analysis-evidence organization-document generation' is mutually split, and traceable and editable integrated workflow cannot be formed. In summary, in the prior art, when complex litigation cases are processed, lawyers are forced to input a great amount of time to perform repeated basic arrangement and document polishing work, the overall case handling efficiency is low, and the service quality is difficult to standardize and scale, so that a systematic technical scheme capable of automatically extracting elements from multi-mode materials, constructing a unified structured case model and intelligently generating series documents and analysis reports according to the Chinese litigation practical process is needed to overcome the defects, and a multi-mode case analysis intelligent generation system and method for the litigation lawyers is provided. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a multi-mode case analysis intelligent generation system and method for lawsuits to solve the background technical problem. In order to achieve the above purpose, the present invention provides the following technical solutions: in a first aspect, a litigation lawyer oriented multimodal case analysis intelligent generation system includes: The multi-mode case data acquisition module is used for receiving various forms of original materials related to the target case; The multi-mode preprocessing and unified representation module is connected with the multi-mode case data acquisition module and is used for carrying out voice recognition, optical character recognition and document analysis on the received original materials and uniformly storing the processed texts and associated metadata into a preset case original corpus; The case element extraction module is connected with the case original corpus and is used for automatically extracting case elements from the case original corpus based on a pre-training language model and field rules and storing extraction results into a case structured database in a structured form, wherein the case eleme