covidia

C🦠vidia: COVID-19 Interdisciplinary Academic Knowledge Graph

• The SPARQL Endpoint is https://covidia.acemap.info/sparql
• The permanent data access point is https://dataset.acemap.info/covidia/covidia-dump-v1-05-22

Overview

Covidia is an Interdisciplinary Academic Knowledge Graph for COVID-19 by fusing papers' illustrations, text, inner knowledge and bibliometric data. To our knowledge, Covidia is currently the first interdisciplinary academic knowledge graph during the pandemic. Figure 1 shows the schema of the Covidia. And if you want to explore the Covidia data, try https://snorql.acemap.cn/Covidia.

Figure 1. Schema graph of the COVID-19 Interdisciplinary Academic Knowledge Graph (Covidia)

And schemagraph file stored in N-Triple form and the ontology file including concept's data properties and object properties stored in pair can be downloaded from:

schemagraph	ontologies
schema	ontology

Also, the developers can use the PREFIX set as follows, the ace/covid/geo is the prefix for concept URI, acer/covidr is the prefix for object properties and covidp/acep belongs to data properties.

• PREFIX rdf: http://www.w3.org/1999/02/22-rdf-syntax-ns#\
• PREFIX rdfs: http://www.w3.org/2000/01/rdf-schema#\
• PREFIX foaf: http://xmlns.com/foaf/0.1/\
• PREFIX skos: http://www.w3.org/2004/02/skos/core#\
• PREFIX acep: https://www.acekg.cn/property/\
• PREFIX acer: https://www.acekg.cn/relation/\
• PREFIX ace: https://www.acekg.cn/concept/\
• PREFIX geo: https://www.acekg.cn/concept/geo/\
• PREFIX covidiap: https://covidia.acemap.info/property/\
• PREFIX covidiar: https://covidia.acemap.info/relation/\
• PREFIX covidia: https://covidia.acemap.info/concept/\
• PREFIX owl: http://www.w3.org/2002/07/owl#\

Resources Statement

Based on Covidia, we put forward serveral resources, including a COVID-19 interdiciplainary knowledge graph dataset, 4 benchmark collections for network science, natural language processing and data minig, as well as 3 COVID-19-related applications. The detail accessment and statement are listed as follows.

Resource	URL	Resouce Type
Covidia Repository	https://github.com/davendw49/covidia	Dataset
Covidia Social Network	https://covidia.acemap.info/benchmarks/sn.zip	Benchmark
Paper Classification	https://covidia.acemap.info/benchmarks/ipc.zip	Benchmark
Knowledge Entity Linking	https://covidia.acemap.info/benchmarks/kel.zip	Benchmark
Relation Classification	https://covidia.acemap.info/benchmarks/rc.zip	Benchmark
Covidia SPARQL Endpoint	http://covidia.acemap.info/sparql	SPARQL Endpoint
Covidia Search Engine	https://covidia.acemap.info/#/	Application
Covidia Semantic Search	https://snorql.acemap.cn/Covidia/	Application
QA System	https://covidia.acemap.info/#/qa	Application

In addition, we have a sustainability plan specified for long-term maintenance of Covidia under the MIT License. The URI of each instance from Covidia is reachable, to meet the needs of Linked Open Data.

Related Work

	Platforms	Subject	Sources	Timeline	Institute	Paper Category	Paper Location	Paper Count	DataDump	Updating
CORD-19	Dataset	Bio&Med	Semantic Scholar	All time	AI2	N/A	N/A	970,835	✅	❌
LitCOVID	Search Engine	Bio&Med	PubMed	All time	NCBI	8 Medical Class	Geolocations	349,546	❌	✅
COVIDScholar	Search Engine	Bio&Med	CORD-19 and LitCOVID	All time	Lawrence Berkeley National Lab	5 COVID-19 related Classes	N/A	970,835	❌	❌
COVID-19 Data Poral	Data Sharing Platform	Bio&Med	EuropePMC	2020.04~	the European Union	4 Medical Classes	N/A	959,926	❌	✅
Dimensions	Search Engine	Interdisciplinary	Subset of Digital Science	2020.01~	Digital Science	22 Fields	Country	1,790,530	✅	❌
COVID-SEE	Search Engine	Bio&Med	CORD-19	All time	University of Melbourne	N/A	N/A	970,835	❌	❌
Docsearch	Medical Platform	Bio&Med	PubMed	All time	DR.EVIDENCE	Medical Classes	Country	40,400,000	❌	✅
COVID-19 Intelligent Insight	Search Engine	Bio&Med	Preprint	2019.12~	Sinequa	N/A	N/A	128,000	❌	❌
Covidia	Search engine and KG	Interdisciplinary	Acemap and Preprint	2019.12~	SJTU	22 Discipline	GPE	2,102,872	✅	✅

Link to Wiki and DBpedia

See alignment:

• entityLinking_db.txt collects the entities that we use for the knowledge points extraction
• sameAs_wiki.txt collects the entities that we use for axios <http://www.w3.org/2002/07/owl#sameAs>

Usage and Hands-on

• We have several routes for different applications for Covidia Search Engine.
• Academic Paper Search Engine: The users can try some keywords related to COVID-19 topics or disciplines (e.g. Vaccine in China). After starting the query, we can see the filter and the result list on the page, and the user can click on the title to redirect to the original URL and click on the similar paper to find the related paper by vector searching the vector is generated by the encoder trained with multiple-disciplines classification model.
• POI Paper Search: The users can try some keywords related to COVID-19 topics or disciplines (e.g. Vaccine in China). After starting the query, we can see the paper in the area of the bounding box the user selects. The positions of the papers are extracted by the BERT-based method and SpaCy.
• KG Augmented QA System: The users can try some question related to COVID-19. On the right of the panel we can see the tools provided, which the users can choose for the engine for augments. If the user chooses covidia_search, the KG provided by Covidia is the augmentation, while the bing_search uses Bing customer search API for the augmentation.
• For Covidia SPARQL Endpoint and Covidia Semantic Search: The user can use sparql to query the structure data from the Covidia.

For more detail, we now introduce the Covidia platform:

Covidia Dumps and Related Code

All the supplements can be downloaded from Google Drive

Acknowledgements

We thanks to the experts that provide precious annotations. They are from the Institute of Geographical Science, Natural Resources Research, Chinese Academy of Sciences, led by Chenghu Zhou. And the whole project is supported by Deep-time Digital Earth Big Science Project.

Citation

We now have a paper under review on ISWC Resource Track:

@inproceedings{covidia,
    title = "Covidia: COVID-19 Interdisciplinary Academic Knowledge Graph Construction and Application",
    author = "Cheng Deng, Jiaxin Ding, Luoyi Fu, Weinan Zhang, Xinbing Wang and Chenghu Zhou",
}