Research On The Method Of Generating Judicial Judgment Document Summaries Combined With Legal Provision

The court has accumulated a large number of judgment documents in the judicial practice for a long time,and the judgment documents are important legal documents that record the trial process and judgment results of the people’s court.With the continuous improvement of the national judicial system and the continuous improvement of people’s legal awareness,judicial cases are also exploding.It is increasingly difficult for relevant personnel to quickly search and understand the required guiding cases from the massive cases;An automatic text summarization method of judicial judgment documents is studied to help judges understand the trial process and context of cases quickly and effectively,and to improve the work efficiency of judges.As well as assisting relevant participants to quickly search and understand the required guiding cases from a large number of cases,it has very important research significance and application value.The task of generating summary of judicial judgment documents is to automatically extract a condensed version reflecting the theme from judicial judgment documents.As an important part of the judicial documents,the law plays an important role in the content and semantic structure of the documents.The traditional method of summary generation does not consider the characteristics of judicial documents and ignores the important role of law,which leads to the loss of important information and the degradation of summary generation performance.Starting from the important part of the judicial logic,this paper comprehensively considers the important content of the judicial documents and the co-occurrence relationship between the basis of the judgment and the legal provisions.From the perspective of traditional automatic text summary,this paper studies the method of generating the summary of judicial documents integrating the legal provisions.The main work of this paper is as follows.1.Summary method of generative judgment documents based on the fusion of laws: by studying the text structure and logic of judgment documents in the judicial field,the relationship between the laws and the important content of the case in the logic context of the judge’s trial is explored;Identify written judgment involved in legal terms,using the terms of the specific content,build encoder based on semantics of external law knowledge,the generation process,the law can be incorporated into the written judgment in this paper,the generation process of the auxiliary decision making written judgment,enhance the controllability of the model of a generation,further keep the important information in the generation process,improve the performance generated models.2.Segmented summary of judgment documents by fusion of laws: based on the research of algorithm of summary of judgment documents by fusion of laws,considering the semantic loss of long text judgment documents in the generative model,a segmented summary generation model of "extraction + generation" is constructed;In order to alleviate the influence of the redundant information in the extraction stage on the performance of the model in the extraction stage,reinforcement learning is used to train the extraction and generation models,so that the extraction model adaptively adjusts according to the final generated abstract,and does not affect the generation stage of the model.The performance of the extraction stage model and the whole model is improved effectively.