Up until recently, spaCy’s functionality was limited to the actual text in the sentences, defining words only in the context of other words. In this talk, I present my recent work on adding Entity Linking functionality to spaCy, which allows grounding information from text into “real world” unique identifiers defined by a knowledge base (KB). This normalization step allows to integrate textual information with structured data in downstream tasks. My talk walks you through the challenges and potential solutions of this disambiguation task, including the implementation of an in-memory KB which is fast and efficient, and a flexible neural network architecture that encodes both the entity information as well as the sentence context.

→  Venue: Belgian NLP meetup

→  Slides: Speakerdeck