Wikidata knowledge base completion using multilingual Wikipedia fact extraction (wikidatacon2019)

In this session we’ll talk about the SLING project at Google. The aim of the project is to learn to read and understand Wikipedia articles in many languages in terms existing knowledge, i.e., specific entities and properties in Wikidata. A key part of the project is that we use the same representation for both knowledge and document annotation, namely frame semantics. The Sling parser can be trained to produce frame semantic representations of text directly without any explicit intervening linguistic representation. The project is a work in progress and we have built a number of the components needed, like the SLING frame store (for building and manipulating frame semantic graph structures) and the Wiki flow pipeline which can take a raw dump of Wikidata and convert this into one big frame graph loadable into memory for fast graph traversal. The SLING Python API provides easy access to all this information. about this event: https://www.wikidata.org/wiki/Wikidata:WikidataCon_2019/Program/Sessions/Wikidata_knowledge_base_completion_using_multilingual_Wikipedia_fact_extraction