11th Belgium NLP Meetup

This is a past event

120 people went

Location image of event venue

Details

Summer seems to have suddenly turned into autumn, and this change of weather has reminded me it's time to plan the next Belgium NLP Meetup. On Tuesday October 1st, Radix.AI is inviting us to its office in Brussels for an evening filled with language and technology. As usual, doors open at 7pm, talks start around 7.30pm, and from 9pm there's time for networking and drinks.

Here's the program for the evening:

Word embeddings: a state of the art
Clément Ntwari Nshuti (Radix.AI)
Word embeddings are one of the main drivers behind the recent popularity of Natural Language Processing. It's no surprise, then, that they come in many different flavours. Clément will talk about the experiments Radix.AI has recently done with this technology. He will discuss the current state of the art in word embeddings and present the pros and cons of the various existing approaches.

Entity Linking for spaCy
Sofie Van Landeghem (Oxykodit)
spaCy is an open-source NLP library designed specifically for production use. 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, Sofie will present her 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. Sofie's talk will walk 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.

Extracting deep insights from 'boring' documents: a real-life story
Niek Bartholomeus (openthebox.be)
At openthebox.be we are building a knowledge graph of Belgian companies and their directors. In this talk I will explain how we applied deep learning on the publications of the Belgian Official Journal (Belgisch Staatsblad/Moniteur Belge) to extract interesting information and got some extremely high accuracies. Unfortunately, it doesn't stop there. A lot more hard work, sweat and tears is needed to bring some real value into the hands of your customer...