Populating a linked data entity name system

Abstract

Resource Description Framework (RDF) is a graph-based data model used to publish data as a Web of Linked Data (Bizer et al . 2009). RDF is an emergent foundation for large-scale data integration , the problem of providing a unified view over multiple data sources. The structure in RDF data can be conveniently visualized using directed labeled graphs , as illustrated in the real-world graph fragments in Figure 1. Nodes in the graph represent entities (e.g. the node with label dbpedia:Allen_, Paul represents the entity Paul Allen in the DBpedia knowledge graph) and edges represent either attributes of an entity (e.g. '01/21/1953' is the birthdate of Paul Allen) or relationships between two entities (e.g. Paul Allen is the co-founder of the company entity, Microsoft). Facts in the knowledge base are formally represented as a set of triples , with a triple comprising a labeled edge (denoted as a property ) in the RDF graph along with its incoming and outgoing nodes.

Knowledge Graphs: A Practical Review of the Research Landscape

“…On the other hand, it may lead to the proverbial reinventing of the wheel, including a profusion of independently developed terminology for the same fundamental phenomenon. This is already evident in at least one research sub-area (namely, entity resolution or ER) that existed before KGs in multiple research areas, including data mining [29], Semantic Web [30,31], and databases [32]. (Each of the citations is an example survey of ER in that specific area.…”

“…It must be noted, however, that many communities are involved in the KG identification stage, as covered earlier. Important tasks (which tend to draw on similar techniques) include entity resolution [29,31,145], link prediction [23,132,146], triples classification [147,148] and representation learning [95,96]. To the researchers working on these problems, the provenance of the initial KG is not important; rather, their interest is in assuming it as input and yielding outputs that potentially improve it (e.g., clusters of resolved entities).…”

Cost-Efficient Prompt Engineering for Unsupervised Entity Resolution

“…Nevertheless, we are far from solving the problem at human performance levels, and errors can be costly. While initially (and in some contexts still), rule-based and manually engineered solutions were prevalent [47,67], machine learning methods became increasingly popular as ER solutions, as with many other problems amenable to learning from data, starting from the early-mid 2000s [63,32]. Over the last decade, deep learning solutions have also been applied to ER [16,15,43,31,35], and even more recently, transformer-based models such as BERT [38].…”