Fine-grained and precise alignment method for multilingual texts based on deep semantics

A fine-grained, multi-language technology, applied in semantic analysis, text database query, unstructured text data retrieval, etc., can solve problems affecting fine-grained alignment accuracy, fine-grained alignment accuracy, and low efficiency of student comprehension, etc. problems, to achieve the effect of speeding up understanding, improving accuracy, and improving accuracy

Active Publication Date: 2022-05-17
LUDONG UNIVERSITY
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] When performing fine-grained alignment between two or more texts in multiple languages, due to the differences in grammar and expression habits between languages, there are certain deviations in the calculation of semantics after mutual translation, and then alignment, resulting in The accuracy of fine-grained alignment is affected. When performing fine-grained alignment between two or more texts in the same language, due to differences in polysemous words or emotional words with strong subjectivity, when word embedding features are used for semantic calculation, Some of the results of fine-grained alignment are difficult to align, which also affects the accuracy of fine-grained alignment. In addition, in the prior art, the results of fine-grained alignment are usually only displayed through data. In special occasions, especially in colleges and universities, When teaching this aspect, the results are relatively abstract, resulting in low efficiency for students to understand

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
  • Fine-grained and precise alignment method for multilingual texts based on deep semantics
  • Fine-grained and precise alignment method for multilingual texts based on deep semantics
  • Fine-grained and precise alignment method for multilingual texts based on deep semantics

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0051] see figure 1 , a method for fine-grained precise alignment of multilingual texts based on deep semantics, characterized in that it includes the following steps:

[0052] S1. First extract the word embedding features in two or more texts of the target, construct a corresponding fine-grained light array according to the word embedding features, and input them into the neural network to calculate the semantic similarity of the word embedding features. For the same or high similarity The semantics of the word embedding features are aligned, and the light balls in the fine-grained light array corresponding to the aligned word embedding features are lit;

[0053] S2. Extract the sentence where the unaligned word embedding feature is located, obtain the sentence embedding feature, and input it into the neural network to calculate the semantic similarity of the sentence embedding feature, align the semantics of the same or high similarity sentence embedding feature, and embed t...

Embodiment 2

[0057] A fine-grained and precise alignment method for multilingual texts based on deep semantics, characterized in that it includes the following steps:

[0058] S1. According to the title, extract the article embedding features in two or more texts of the target, and input them into the neural network to calculate the semantic similarity of the article embedding features, align the semantics of the same or high similarity article embedding features, and align Unaligned chapters are marked in fine-grained light arrays;

[0059] S2. Extract paragraph embedding features for unaligned chapters, and input them into the neural network to calculate the semantic similarity of paragraph embedding features, align the semantics of paragraph embedding features with the same or high similarity, and fine-grained light for unaligned paragraphs mark in the array;

[0060] S3. Segment the unaligned paragraphs according to the punctuation marks, extract the sentence embedding features, and i...

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 discloses a fine-grained and precise alignment method for multilingual texts based on deep semantics, which belongs to the field of text alignment. The fine-grained and precise alignment method for multilingual texts based on deep semantics gradually calculates words, sentences, paragraphs and chapters by extracting word embedding features Semantic similarity, so as to gradually improve the accuracy of text fine-grained alignment. At the same time, in the alignment process, through the construction of fine-grained light arrays, after each step, the air expansion tube is inflated in the radial direction and vertical direction respectively by inflation. and elongation, so that the fluorescent liquid enters the mask, lights up the lamp ball corresponding to the aligned word embedding feature, and when it is lit for the second time or the last few times, because the inflation time gradually becomes longer, the lit lamp The downward extension of the ball is getting longer and longer, and the brightness is getting brighter, so that after each step, the improvement of the accuracy of fine-grained alignment is more visually reflected, further assisting students to speed up their understanding of this content speed.

Description

technical field [0001] The invention relates to the field of text alignment, and more specifically, to a method for fine-grained and precise alignment of multilingual texts based on deep semantics. Background technique [0002] Entity linking is the process of mapping entity references in natural language to correct candidate entities in a knowledge base. In layman's terms, the fine-grained model is to subdivide the objects in the business model to obtain a more scientific and reasonable object model. Intuitively speaking, it is to divide many objects. [0003] The memory space in modern computers is divided by byte. In theory, it seems that access to any type of variable can start from any address, but the actual situation is that when accessing a specific variable, it is often accessed at a specific memory address, which is All types of data need to be arranged in space according to certain rules, rather than sequentially arranged one by one, which is alignment. [0004]...

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
Patent Type & Authority Patents(China)
IPC IPC(8): G06F16/33G06F40/30G09F19/10
CPCG06F16/3344G06F16/3335G06F40/30G09F19/10
Inventor 刘伍颖
Owner LUDONG UNIVERSITY
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