Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Syntactic relationship enhanced machine reading understanding multi-hop reasoning model and method

A technology of reading comprehension and reasoning methods, which is applied in the field of machine reading comprehension multi-hop reasoning models, which can solve problems such as poor performance of answering methods and inaccurate answer basis, and achieve the effect of improving interpretability and improving answering methods

Active Publication Date: 2021-02-26
SHANXI UNIV
View PDF14 Cites 18 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] Articles on MRC tasks often contain a wealth of dependent syntactic relationships. These syntactic relationships can well guide the model to perform multi-hop reasoning and mine answer basis. However, most multi-hop reasoning models do not consider these syntactic relationships, and the obtained answer basis is still Not very accurate; at the same time, the existing answering methods for opinion-based questions are not very good at identifying answer clues

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
  • Syntactic relationship enhanced machine reading understanding multi-hop reasoning model and method
  • Syntactic relationship enhanced machine reading understanding multi-hop reasoning model and method
  • Syntactic relationship enhanced machine reading understanding multi-hop reasoning model and method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0073] figure 1 It is a model frame diagram of the present invention, which mainly includes a text coding module, an associated element relationship diagram building module, a question answering module, and an answer prediction module, wherein the question answering module includes an answer generation module and a multi-hop reasoning module, and the specific contents of each module are as follows:

[0074] A text encoding module encodes the questions and articles to obtain semantic representations of the questions and articles;

[0075] Associating element relationship diagram construction module, identifying the key elements in each sentence of the article and the dependent syntactic relationship between them, defining element association rules, and using the association rules to construct the association element relationship diagram;

[0076] The multi-hop reasoning module performs multi-hop reasoning based on the relational element relationship graph and the graph attentio...

Embodiment 2

[0080] figure 2 An example from the reading comprehension dataset for the 2020 China "Law Research Cup" Judicial Artificial Intelligence Challenge (CAIL2020). Such as figure 2 As mentioned above, the article is a real case in Chinese judgment documents, the question is "Is there a time limit in the contract?", the answer is "yes", and the answer is based on the sentence numbers "4, 6" in the article.

[0081] 1. First, use the text encoding module to encode the questions and articles, and obtain the semantic vectorized representations of the questions and articles. The present invention uses the RoBERTa model as an encoder to map the article and each word or phrase of the article to a high-dimensional vector space to obtain the semantic representation of each word or word. The calculation formula is as follows:

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 deep learning, natural language processing and the like, and particularly relates to a syntactic relationship enhanced machine reading understanding multi-hop reasoning model and method. The model comprises a text coding module, an association element relationship graph construction module, a multi-hop reasoning module, an answer generation module and an answer prediction module. According to the invention, syntactic relationships are fused into the graph construction process, an associated element relationship graph is constructed, multi-hop reasoning iscarried out by utilizing a graph attention network based on the relationship graph, and answer support sentences are mined; meanwhile, a multi-head self-attention mechanism is introduced to further mine text clues of viewpoint type questions in the article, and an automatic solution method of the viewpoint type questions is improved; and finally, a plurality of tasks are subjected to joint optimization learning, so that when the model answers the questions, fact description for supporting the answers can be given, the interpretability of the model is improved, and meanwhile, the existing answering method for viewpoint type questions is improved.

Description

technical field [0001] The invention relates to the fields of deep learning, natural language processing, etc., and in particular to a multi-hop reasoning model and method for machine reading comprehension with enhanced syntactic relations. Background technique [0002] Machine Reading Comprehension (MRC) is an important research task to understand the semantics of articles and answer related questions through computers. The research on machine reading comprehension plays an important role in improving the machine's natural language understanding ability. widespread concern in the industry. Early machine reading comprehension research mainly adopted the method based on artificial rule bases. The establishment and maintenance of rule bases usually required a lot of manpower, and it was difficult to answer questions other than rules. In recent years, with the rapid development of machine learning, especially deep learning, the automatic answering effect of machine reading com...

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): G06F16/33G06F16/332G06F16/35G06F40/211G06F40/216G06F40/30G06K9/62G06N3/04G06N3/08
CPCG06F16/3329G06F16/3344G06F16/3346G06F16/35G06F40/211G06F40/216G06F40/30G06N3/08G06N3/048G06N3/045G06F18/241G06F18/2415
Inventor 张虎王宇杰李茹梁吉业谭红叶
Owner SHANXI UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com