Findings of the Association for Computational Linguistics: EMNLP 2020 2020
DOI: 10.18653/v1/2020.findings-emnlp.133
|View full text |Cite
|
Sign up to set email alerts
|

Open Domain Question Answering based on Text Enhanced Knowledge Graph with Hyperedge Infusion

Abstract: The incompleteness of knowledge base (KB) is a vital factor limiting the performance of question answering (QA). This paper proposes a novel QA method by leveraging text information to enhance the incomplete KB. The model enriches the entity representation through semantic information contained in the text, and employs graph convolutional networks to update the entity status. Furthermore, to exploit the latent structural information of text, we treat the text as hyperedges connecting entities among it to compl… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1

Citation Types

1
24
0

Year Published

2021
2021
2023
2023

Publication Types

Select...
4
2
1
1

Relationship

0
8

Authors

Journals

citations
Cited by 44 publications
(25 citation statements)
references
References 18 publications
(24 reference statements)
1
24
0
Order By: Relevance
“…Meanwhile, finer 368 PLMs can improve the semantic feature representations of question subgraphs and 369 achieve better results in question answering. This is also consistent with the view and 370 experimental results in[33].382dimensional hyper-graph with hierarchical structure (Figure6). The metaknowledge 383 network expressed by that type of graph model includes two dimensions: hierarchical 384 dimension and semantic dimension.…”
supporting
confidence: 89%
See 1 more Smart Citation
“…Meanwhile, finer 368 PLMs can improve the semantic feature representations of question subgraphs and 369 achieve better results in question answering. This is also consistent with the view and 370 experimental results in[33].382dimensional hyper-graph with hierarchical structure (Figure6). The metaknowledge 383 network expressed by that type of graph model includes two dimensions: hierarchical 384 dimension and semantic dimension.…”
supporting
confidence: 89%
“…Ref [34]. proposes an embedding framework MINES for multi-388 dimensional networks with hierarchical structure, which uses hierarchical structure for 389 multi-dimensional network embedding; Ref [33]. proposes an open domain question 390 answering method based on hyper-edge fusion.…”
mentioning
confidence: 99%
“…Sun et al [2018] and proposed to complement the subgraph extracted from incomplete KBs with extra question-related text sentences to form a heterogeneous graph and conduct reasoning on it. Instead of directly complementing sentences to question-specific graph as nodes, Xiong et al [2019] and Han et al [2020a] proposed to fuse extra textual information into the entity representation to supplement knowledge. They first encoded sentences and entities conditioned on questions, and then supplemented the incomplete KB by aggregating representations of sentences to enhance corresponding entity representations.…”
Section: Information Retrieval-based Methodsmentioning
confidence: 99%
“…Supplement KB with extra corpus [Sun et al, 2018;, fuse extra textual information into entity representations [Xiong et al, 2019;Han et al, 2020a] or leverage KB embeddings [Saxena et al, 2020].…”
Section: Reasoning Under Incomplete Kbmentioning
confidence: 99%
“…In recent years, many researchers have combined continuous text knowledge and knowledge graphs as a data source for KBQA. For example, Han et al [10] proposed a question-and-answer model based on text-enhanced knowledge graphs, which encoded entities in the KB through text information and applied graph convolutional networks [11] to perform reasoning on KB. Das et al [12] proposed a general model that used a combination of text and KB, combining with a memory network to complete question and answer.…”
Section: Introductionmentioning
confidence: 99%