AKMiner: Domain-Specific Knowledge Graph Mining from Academic Literatures

The lack of entity label values is one of the problems faced by the application of Knowledge Graph. The method of automatically assigning entity label values still has shortcomings, such as costing more resources during training, leading to inaccurate label value assignment because of lacking entity semantics. In this paper, oriented to domain-specific Knowledge Graph, based on the situation that the initial entity label values of all triples are completely unknown, an Entity Label Value Assignment Method (ELVAM) based on external resources and entropy is proposed. ELVAM first constructs a Relationship Triples Cluster according to the relationship type, and randomly extracts the triples data from each cluster to form a Relationship Triples Subset; then collects the extended semantic text of the entities in the subset from the external resources to obtain nouns. Information Entropy and Conditional Entropy of the nouns are calculated through Ontology Category Hierarchy Graph, so as to obtain the entity label value with moderate granularity. Finally, the Label Triples Pattern of each Relationship Triples Cluster is summarized, and the corresponding entity is assigned the label value according to the pattern. The experimental results verify the effectiveness of ELVAM in assigning entity label values in Knowledge Graph.

Download Full-text

Constructing biomedical domain-specific knowledge graph with minimum supervision

Knowledge and Information Systems ◽

10.1007/s10115-019-01351-4 ◽

2019 ◽

Vol 62 (1) ◽

pp. 317-336

Author(s):

Jianbo Yuan ◽

Zhiwei Jin ◽

Han Guo ◽

Hongxia Jin ◽

Xianchao Zhang ◽

...

Keyword(s):

Knowledge Graph ◽

Biomedical Domain ◽

Specific Knowledge ◽

Domain Specific ◽

Domain Specific Knowledge

Download Full-text

Building Domain-Specific Knowledge Graph for Unmanned Combat Vehicle Decision Making under Uncertainty

2019 Chinese Automation Congress (CAC) ◽

10.1109/cac48633.2019.8996418 ◽

2019 ◽

Author(s):

Shuai Wang ◽

Yu Zhang ◽

Zhiyong Liao

Keyword(s):

Decision Making ◽

Knowledge Graph ◽

Decision Making Under Uncertainty ◽

Specific Knowledge ◽

Domain Specific ◽

Combat Vehicle ◽

Domain Specific Knowledge

Download Full-text

Building a Domain-Specific Knowledge Graph for Business Networking Analysis

Intelligent Information and Database Systems - Lecture Notes in Computer Science ◽

10.1007/978-3-030-73280-6_29 ◽

2021 ◽

pp. 362-372

Author(s):

Didier Gohourou ◽

Kazuhiro Kuwabara

Keyword(s):

Knowledge Graph ◽

Specific Knowledge ◽

Domain Specific ◽

Domain Specific Knowledge ◽

Business Networking

Download Full-text

Deep Learning-Based Knowledge Graph Generation for COVID-19

Sustainability ◽

10.3390/su13042276 ◽

2021 ◽

Vol 13 (4) ◽

pp. 2276

Author(s):

Taejin Kim ◽

Yeoil Yun ◽

Namgyu Kim

Keyword(s):

Knowledge Base ◽

Language Model ◽

Fine Tuning ◽

Knowledge Graph ◽

Specific Information ◽

Specific Knowledge ◽

Text Documents ◽

Domain Specific ◽

Open Information Extraction ◽

Domain Specific Knowledge

Many attempts have been made to construct new domain-specific knowledge graphs using the existing knowledge base of various domains. However, traditional “dictionary-based” or “supervised” knowledge graph building methods rely on predefined human-annotated resources of entities and their relationships. The cost of creating human-annotated resources is high in terms of both time and effort. This means that relying on human-annotated resources will not allow rapid adaptability in describing new knowledge when domain-specific information is added or updated very frequently, such as with the recent coronavirus disease-19 (COVID-19) pandemic situation. Therefore, in this study, we propose an Open Information Extraction (OpenIE) system based on unsupervised learning without a pre-built dataset. The proposed method obtains knowledge from a vast amount of text documents about COVID-19 rather than a general knowledge base and add this to the existing knowledge graph. First, we constructed a COVID-19 entity dictionary, and then we scraped a large text dataset related to COVID-19. Next, we constructed a COVID-19 perspective language model by fine-tuning the bidirectional encoder representations from transformer (BERT) pre-trained language model. Finally, we defined a new COVID-19-specific knowledge base by extracting connecting words between COVID-19 entities using the BERT self-attention weight from COVID-19 sentences. Experimental results demonstrated that the proposed Co-BERT model outperforms the original BERT in terms of mask prediction accuracy and metric for evaluation of translation with explicit ordering (METEOR) score.

Download Full-text