concept

Artificial Intelligence Knowledge Graph

Also known as: AI-KG

Facts (10)

Sources

Construction of Knowledge Graphs: State and Challenges - arXiv arxiv.org arXiv 9 facts

procedureThe Artificial Intelligence Knowledge Graph (AI-KG) determines possible entities and relations by operating an overlapping strategy between Stanford CoreNLP tools (Open Information Extraction and POS-Tagger) and topics detected by the CSO Classifier.

claimThe internal ontology of the Artificial Intelligence Knowledge Graph (AI-KG) is built upon SKOS, PROV-O, and OWL.

procedureThe construction pipeline for the Artificial Intelligence Knowledge Graph (AI-KG) consists of four components: (1) extractors, (2) entities handler, (3) relations handler, and (4) triple selector.

measurementThe Artificial Intelligence Knowledge Graph (AI-KG) contains over 820,000 entities derived from over 333,000 research publications, integrating data from the Microsoft Academic Graph, the Computer Science Ontology, and Wikidata.

claimThe Artificial Intelligence Knowledge Graph (AI-KG) supports fact provenance by maintaining identifiers to original research papers and information regarding the extraction tools applied.

procedureThe Triple Selector component of the Artificial Intelligence Knowledge Graph (AI-KG) categorizes facts into valid and non-valid triples, using an occurrence frequency threshold for POS-Tagger-derived relations and trusting triples extracted by DyGIE++ due to its high precision in previous benchmarks.

procedureThe Artificial Intelligence Knowledge Graph (AI-KG) clusters semantically similar entities using word embeddings and hierarchical clustering, followed by manual revision.

claimThe Artificial Intelligence Knowledge Graph (AI-KG) was eventually replaced by the CS-KG, and while periodic updates were planned, the update process was never described.

claimThe extractors in the Artificial Intelligence Knowledge Graph (AI-KG) pipeline utilize DyGIE++, the CSO Classifier, and Stanford CoreNLP to extract entities and relations from text.

Knowledge Graphs: Opportunities and Challenges - Springer Nature link.springer.com Springer Apr 3, 2023 1 fact

measurementThe Artificial Intelligence Knowledge Graph (AI-KG) describes 800,000 entities, including tasks, methods, materials, and metrics, extracted from the 330,000 most cited articles in the field of artificial intelligence.