Judgment document dispute focus extraction method based on multi-task small sample learning

Abstract

The invention discloses a judgment document dispute focus extraction method based on multi-task small sample learning. According to the method, the problems that disputed problems are distributed to multiple classes, the classes are unbalanced due to the huge difference of the number of disputed focus problems contained in each class, consequently, the performance of text clustering is reduced, most of current algorithm models have high requirements for data, and a large amount of data annotation work needs to be achieved are solved. A certain amount of labeled data is obtained after clustering, cluster labels are automatically obtained through LDA, model training, model clipping and dispute focus classification matching are carried out after data enhancement processing to extract judicialdispute focuses, and help can be better provided for lawyer judge to retrieve class cases.

Description

technical field [0001] The invention relates to text classification and matching in the field of focus of disputes in the judicial field, in particular to a method for extracting focus of disputes in adjudication documents based on multi-task small-sample learning. Background technique [0002] With the continuous progress of our country's social development, the judicial reform will be further promoted. The reform starts from the demands of the people for justice, and focuses on strengthening the supervision and restriction of power. With the rapid development of information technology, the release of online judgments has played a vital role in promoting judicial transparency. As the new litigation system gradually takes shape, Chinese courts organize debates around controversial issues. [0003] Factual questioning helps to focus fact-finding in court trials, while legal questioning helps courtroom debate organization and law application. The judgment embodies the proce...

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is that controversial issues are assigned to many classes, and the huge difference in the number of controversial focus issues contained in each class leads to an imbalance of classes, which leads to a decrease in the performance of text clustering ; Most of the current algorithm models have high requirements for data, and a large amount of data labeling work is required

[0008] The present invention provides a multi-task-based small-sample learning method for extracting the focus of disputes in referee documents, which solves the imbalance problem caused by the difference in the number of the above-mentioned classes and the problem that the algorithm model requires a large number of labels for the data.

Images