Method and system for generating question sentences

A technology for question generation and model generation, which is applied in the field of question generation methods and systems, can solve problems such as slow execution speed, insufficient performance of question generation, and lack of related questions, so as to improve execution speed and accuracy, and improve readability Sex and diversity, reducing the effect of manual labeling

Active Publication Date: 2021-08-06
AEROSPACE INFORMATION RES INST CAS
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

At the same time, we found that due to the lack of support for relevant questions in the current dialogue system and reading comprehension system, the content of the dialogue system and reading comprehension is too single, which is not suitable for current people's needs
Although there are some question generation methods at present, the method of generating questions using traditional rules requires a lot of manual annotation, so the process of generating questions has insufficient generation performance, poor scalability, slow execution speed, and low generation performance. , not enough to meet the needs of the current people

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
  • Method and system for generating question sentences
  • Method and system for generating question sentences
  • Method and system for generating question sentences

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0047] Example 1, such as figure 1 shown and figure 2 Shown, technical scheme of the present invention is as follows:

[0048] S1 Recognize the text to be read and comprehend based on the named entity recognition tool, and get the answer part;

[0049] S2 brings the text to be read and comprehend and the corresponding answer part into the pre-trained question generation model to generate multiple questions for the answer;

[0050] S3 correcting the plurality of questions to obtain the questions corresponding to the text to be read and understood;

[0051] Wherein, the question generation model, based on the existing dialogue system and reading comprehension text, introduces a copy mechanism and a placeholder mechanism into the algorithm model of the multi-layer and multi-scale transformer network to replace the named entities in the reading comprehension text , to obtain the question expressed by the dialogue system. The transformer mentioned in the present invention is a...

Embodiment 2

[0066] In this embodiment, the text to be read and understood is set to include a statement sentence, which can be understood as an answer, for example: Beijing is the capital of China,

[0067] First, the sentence is preprocessed, including sentence segmentation, word segmentation, word vector embedding, regularization, cleaning, etc. of the text to obtain: word segmentation:

[0068] Then, use the existing named entity recognition tool to process the above-mentioned processed data to obtain the entity characteristics of each word, and get: and are place names

[0069] Finally, using the training method in steps 3-5 of Example 1, the named entity information is encoded and incorporated into the word embedding; then the word embedding model integrated with the named entity information is sent to the transformer question generation model to obtain Question: Where is the capital of China?

Embodiment 3

[0071] In order to realize the above method, the present invention also provides a system for generating question sentences, including:

[0072] The data preparation module is used to identify the text to be read and comprehend based on the named entity recognition tool, and obtain the answer part;

[0073] A question generation module, used to bring the text to be read and comprehend and the corresponding answer part into a pre-trained question generation model to generate multiple questions for the answer;

[0074] A question sentence determination module, which is used to correct a plurality of questions to obtain a question corresponding to the text to be read and understood;

[0075] Among them, the question generation model, based on the existing dialogue system and reading comprehension text, introduces the copy mechanism and placeholder mechanism in the algorithm model of the multi-layer and multi-scale transformer network to replace the named entities in the reading c...

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 present invention provides a method and system for generating question sentences, including: identifying the text to be read and understood based on a named entity recognition tool to obtain the answer part; bringing the text to be read and understood and the corresponding answer part into the pre-trained question generation model Generate multiple questions for the answers; correct the multiple questions to obtain the corresponding questions for the text to be read and comprehend; among them, the question generation model is based on the existing dialogue system and the reading comprehension text in multi-layer and multi-scale The algorithm model of the transformer network introduces a copy mechanism and a placeholder mechanism to replace the named entities in the reading comprehension text, which improves the execution speed and accuracy of generating questions, improves scalability, and greatly reduces manual annotation , while using the existing dialogue system to improve the readability and diversity of question generation.

Description

technical field [0001] The invention belongs to the technical field of processing natural language data, and in particular relates to a question generation method and system. Background technique [0002] With the explosive growth of network information, all kinds of information flood the entire network environment. People are now used to go to the Internet to search for some solutions to problems. When users are not very familiar with some search techniques, they often need to spend a lot of time to filter the results returned by the search engine. The birth of the interactive dialogue system and the reading comprehension system effectively solved the problem of complicated information mentioned above. The interactive dialogue system and reading comprehension system use natural language processing to analyze the questions submitted by users, obtain relevant answers and return them to users. [0003] Automatic question generation will provide question-answer pairs for int...

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): G06F40/295G06F40/211
CPCG06F40/211G06F40/295
Inventor 许光銮于泓峰张文凯田雨李沛光姚方龙武斌刘那与
Owner AEROSPACE INFORMATION RES INST CAS
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