Law case deep reasoning method

Abstract

The invention relates to a law case deep reasoning method which comprises the following steps: S1, receiving a dispute type, a lawsuit request and a spoken language case description selected by a user; s2, constructing a knowledge graph according to the dispute type and the litigation request; s3, on the basis of the knowledge graph, processing the colloquial case description to obtain an analysisresult; and S4, performing support rate reasoning of the litigation request based on the knowledge graph according to the analysis result, and outputting a reasoning result. According to the method,the support rate of the litigation request is predicted in a deep reasoning mode, a corresponding support reason is given, the support reason can be displayed to a user in a visual reasoning process mode, and compared with a current result report, the method is more intelligent and more visual.

Description

technical field [0001] The invention belongs to the technical field of legal services, and in particular relates to a deep reasoning method for legal cases. Background technique [0002] Nowadays, with the popularity of deep learning and the improvement of algorithm computing power, artificial intelligence is gradually penetrating into all walks of life, and even surpasses human level in some fields, and the legal industry is no exception. In 2013, the disclosure of a large amount of legal data on China's judicial documents online made it possible to intelligentize the law. [0003] From the current point of view, AI has two main applications in the legal field: one is lawyer tools, through the information extraction of judgment documents and the application of various templates, it automates part of the case-handling process and improves lawyers’ case-handling efficiency, such as searching for similar cases, Automated contract analysis, litigation document generation, etc....

AI Technical Summary

Problems solved by technology

[0009] 1. Users can only provide information by answering questions, and do not support colloquial question descriptions, so the degree of freedom is not high

Users need to match their own situation with the answers to the questions. Because some questions involve professional legal knowledge, users cannot understand and choose well.

[0010] 2. The questions and answers are pre-set and cannot be adjusted according to the user's answers, and the number of questions is too large, resulting in a low proportion of users completing the consultation

[0011] 3. The question bank is sorted out manually, which requires the participation of a large number of legal professionals. It does not use the learning ability of machine learning, nor does it reflect artificial intelligence

[0012] 4. The final report is generated based on the template integration based on the answer results, and the report results are relatively fixed, lacking the components of intelligent reasoning

Images