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

Multi-hop reasoning question and answer method based on cross-language knowledge graph

A knowledge graph, cross-language technology, applied in the field of multi-hop reasoning question answering, can solve problems such as error transmission, and achieve the effect of enhancing performance

Pending Publication Date: 2022-07-22
SOUTHEAST UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0009] Technical problem: The technical problem to be solved by the present invention is to provide a multi-hop reasoning question-answering method based on cross-language knowledge graphs for traditional question-answering methods that rely on pre-fused graphs when using cross-language knowledge graphs.

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
  • Multi-hop reasoning question and answer method based on cross-language knowledge graph
  • Multi-hop reasoning question and answer method based on cross-language knowledge graph
  • Multi-hop reasoning question and answer method based on cross-language knowledge graph

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0053] Embodiment 1: A multi-hop reasoning question answering method based on cross-language knowledge graph, the method comprises the following steps:

[0054] Step 1) Initialization of multi-hop reasoning question answering task;

[0055] Step 2) build a cross-graph reasoning model;

[0056] Step 3) Multi-hop reasoning-entity alignment question answering framework;

[0057] Step 4) iterative pseudo-entity alignment annotation mining and alignment model enhancement training process;

[0058] Step 5) Path ranking based on sequence similarity.

[0059] Among them, step 1) multi-hop reasoning question answering task initialization is as follows: first determine the target question type (multi-relational complex question), collect related questions for the target question and a sample set O composed of query annotations on the corresponding knowledge graph, The sample set is split, and the original sample is split into multiple one-hop inference sub-samples according to the nu...

Embodiment 2

[0094] Example 2: Reference figure 1 , figure 2 , image 3 and Figure 4 , a multi-hop reasoning question answering method based on cross-language knowledge graph, the method includes the following steps:

[0095] Step 1) Initialization of multi-hop reasoning question answering task;

[0096] First determine the target problem type (multi-relational complex problem), collect related questions and a sample set O composed of query annotations on the corresponding knowledge graph for the target problem, split the sample set, and divide the sample set according to the number of relations contained in the annotated samples. The original sample is split into multiple one-hop inference sub-samples, and L is obtained by replacing the original sample in O.

[0097] Step 2) build a cross-graph reasoning model;

[0098] The cross-graph reasoning model takes natural language questions, historical reasoning paths and main entities on the cross-language graph as input, and outputs pre...

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 multi-hop reasoning question-answering method based on a cross-language knowledge graph, which is mainly used for improving an existing intelligent question-answering system, so that the existing intelligent question-answering system can effectively fuse and utilize information in knowledge graphs of different languages when questions proposed by users are answered. The invention provides a cross-map reasoning method and further establishes an iterative framework for combining multi-hop reasoning and entity alignment by paying attention to the problem of error transmission caused by dependence on pre-fusion of a map when a traditional question and answer method uses a cross-language knowledge map. According to the method, multilingual-BERT is used for carrying out general representation on a text, so that reasoning information can be transmitted among maps of different languages, and a bidirectional LSTM coding problem is used for reasoning a path; and updating problem representation through an attention mechanism for relation prediction. And the entity alignment method combines the former with the entity alignment method, multiple groups of candidate paths are generated from the atlas by taking the input question as a query, and pseudo-alignment labels are iteratively extracted from the candidate paths to enhance the entity alignment method, so that the question and answer performance is improved.

Description

technical field [0001] The invention relates to a multi-hop reasoning question answering method based on a cross-language knowledge graph, and belongs to the technical field of natural language intelligent question answering. Background technique [0002] A knowledge graph is a structured information collection that contains rich entities and relationships. In order to cover more knowledge resources in the world, resources in different languages ​​are brought together to form a multilingual knowledge graph, which is based on DBpedia, YAGO and Wikidata. as a benchmark. [0003] In recent years, with the continuous development and enrichment of non-English knowledge graphs, the complementary relationship between English resources and non-English resources has become more and more obvious. Taking Chinese as an example, Baidu Baike, CN-DBpedia and Zhishi.me represent Chinese Knowledge graphs cover tens of millions of entity information. Therefore, how to integrate and utilize c...

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 Applications(China)
IPC IPC(8): G06F16/36G06N3/04G06N3/08G06N5/04
CPCG06F16/367G06N5/04G06N3/08G06N3/044
Inventor 谭亦鸣张欣宇漆桂林
Owner SOUTHEAST 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