5th Belgium NLP Meetup

We don't yet know what 2018 will bring, but we do know a new Belgium NLP Meetup is taking shape. The fifth meetup in the series will take us back to Antwerp, where on February 1st Digipolis will host a lineup of speakers that is sure to pique your interest. We'll start the evening around 19h and hope to finish around 22h. Food and drinks will be provided. This is the program for the evening:

• Building an ISCO job classifier, a pragmatic approach.
  Joeri Mesens & Leen Van Acker, theMatchBox
  theMatchBox builds intelligent search and match solutions in different domains like HR recruiting, CMS systems, … To build quality search engines, they put a lot of effort in preparing data for search (tokenizing, lemmatizing, decompounding, spelling correction) and in extracting extra information from the data (e.g. job title or location extraction from a vacancy). Joeri and Leen will talk about how they built one of their latest information extraction algorithms: a job classifier that maps jobs to ISCO occupations.

• Synthetic Literature. Writing Science Fiction in a Co-Creative Process.
  Enrique Manjavacas & Mike Kestemont, University of Antwerp
  Enrique and Mike will talk about a co-creative text generation system they developed for the Dutch Book Week. Their system uses a character-level language model to generate text based on a large corpus of Dutch novels. It is accessed through a graphical user interface that helps writers elicit, modify and incorporate suggestions by the text generation system. In a first experiment, the system was used by Ronald Giphart, an established Dutch novelist, and the generated text will be published in a volume of science fiction stories.

• Matching CVs and vacancies for the City of Antwerp.
  Christophe Marrecau & Stijn Geuens, Digipolis
  Personnel management of the City of Antwerp processes 17,000 CVs annually. The data science team at Digipolis are investigating the possibilities of a CV matching tool based on Natural Language Processing to partially automate this process, and to allow potential applicants to see for themselves what vacancies are relevant to their CV. Christophe and Stijn will talk about their development of this application and the accompanying evolution from classic text mining algorithms to Deep Learning ambitions.