ABSTRACT

The increasing availability of structured data on the Web stimulated a renewed interest in its graph nature. Applications like the Google Knowledge Graph (KG) [227] and the Facebook Graph (FG) [192] build large graphs of entities (e.g., people, places) and their semantic relations (e.g., born in, located in). The KG, by matching keywords in a search request against entities in the graph, enhances Google’s results with structured data in the same spirit of Wikipedia info boxes. The FG by looking at semantic relations between Facebook entities enables searching within this huge social graph. However, both approaches adopt proprietary architectures with idiosyncratic data models and limited support in terms of APIs to access their data and querying capabilities.