Enhance matrix-based legal case candidate paragraph selection method and device

A paragraph and legal technology, applied in the field of selection of candidate paragraphs in legal cases, can solve problems such as inability to correctly select candidate paragraphs, and learning the reasoning relationship of paragraphs

Active Publication Date: 2021-09-07
CHONGQING UNIV OF POSTS & TELECOMM
View PDF10 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The present invention aims at the disadvantages that the existing candidate paragraph acquisition technology cannot correctly select the candidate paragraphs related to the question and learn the reasoning relationship between the paragraphs when processing the legal field-related documents with multi-hop characteristics, and proposes a method based on the enhance matrix The selection method and device of the legal case candidate paragraphs, the method includes the following steps:

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Enhance matrix-based legal case candidate paragraph selection method and device
  • Enhance matrix-based legal case candidate paragraph selection method and device
  • Enhance matrix-based legal case candidate paragraph selection method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0050] figure 1 It is a structure diagram for obtaining a candidate paragraph of a legal case in an embodiment of the present invention, such as figure 1 As shown, this example first obtains a legal reading comprehension data set with reasoning properties and labels the data set, obtains a data set of candidate paragraphs of legal cases, enters the legal text into the Bert model for representation, and adds attention to the vector to calculate the distance between paragraphs The similarity matrix R constructs the enhance matrix matrix EM according to the word relationship between paragraphs and the word relationship between paragraphs and questions, further uses the enhance matrix to enhance the paragraph similarity matrix R, and obtains the final paragraph vector by updating the feature representation, Put the paragraph vector feature into the sigmoid classifier to get the final candidate paragraph selection result.

[0051] In this embodiment, the inferential legal reading ...

Embodiment 2

[0072] Figure 6 Shown is a selection device of a legal case candidate paragraph based on an enhance matrix according to an embodiment of the present invention, including:

[0073] The candidate paragraph data processing module is used to perform processing on the acquired reading comprehension data set with reasoning properties, and select candidate paragraphs and label them according to the characteristics that the data set requires that the answer must be obtained through the reasoning of at least 2 candidate paragraphs Labeling, to obtain a data set that uses whether it is a candidate paragraph as a label;

[0074] The Bert characterization module is used to execute the characterization algorithm for the paragraphs and questions in each sample in the candidate paragraph data set, merge the question text and the paragraph text, intercept or complete the input with a length of 512, and input it to the pre-trained Bert Model, obtain the representation of the question and par...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to the fields of natural language processing, artificial intelligence and the like, in particular to an enhance matrix-based legal case candidate paragraph selection method and device, and the method comprises the steps: processing a legal reading understanding data set with an inference property, and obtaining a legal data set with candidate paragraph labels; connecting the question and the paragraph in each sample, outputting two paragraph vector matrixes with different model parameters through an attention operation and a Bert model, and calculating a similarity matrix R according to the two matrixes; constructing an EM matrix for each sample, and performing increment processing on the similarity matrix R by using the EM matrix; inputting the processed features into a dichotomy task classifier for training, and obtaining n paragraphs with the highest probability as candidate paragraphs through a trained model; the paragraph selection accuracy is improved, and noise information transmitted to downstream tasks is reduced to the greatest extent.

Description

technical field [0001] The invention relates to the fields of natural language processing, artificial intelligence, etc., and in particular to a method and device for selecting candidate paragraphs of legal cases based on an enhance matrix. Background technique [0002] With the development of society and the advent of the era of big data, the courts have to deal with a large number of litigation cases every year. The people's courts need to summarize the focus of disputes based on the parties' claims, defense opinions, and evidence exchange. This process requires judges to follow legal regulations. This process is very labor-intensive. With the development of artificial intelligence, many technologies in the field of language processing have been applied to the summary of the focus of disputes in legal cases, and good results have been achieved. Achievements; after obtaining the focus of dispute, it is also necessary to consider providing evidence for "why this focus of dis...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06F40/216G06F40/284G06K9/62G06N3/04G06N3/08G06Q50/18
CPCG06F40/216G06F40/284G06Q50/18G06N3/08G06N3/047G06N3/048G06F18/22
Inventor 胡峰董磊邓维斌
Owner CHONGQING UNIV OF POSTS & TELECOMM
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap