About STIL

STIL, the Foundation for Inductive Learning Applications, was founded in 1997 as a link to the outside world for language technology software and expertise developed at Tilburg University's Induction of Linguistic Knowledge (ILK) research group and the University of Antwerp's Computational Linguistics and Psycholinguistics Research Center (CLiPS).

        

ILK and CLiPS are known for their joint development of TiMBL, a generic machine learning algorithm emulating memory-based learning. Based on TiMBL the groups have developed

  • MBT for sequence tagging;
  • MBSP for shallow parsing of English text;
  • Frog for integrated morpho-syntactic tagging and parsing of Dutch text.
The groups are also developing software for web mining, context-sensitive spelling correction for Dutch and English, language modeling, machine translation, co-reference resolution, word sense disambiguation, author verification and attribution, text analytics in various domains, sentence compression, and paraphrasing.

STIL can offer solutions to businesses and non-profit organizations exploring what language technology could offer them, or with a need to integrate language technology components in their information systems.

STIL offers consultancy and expertise on

  • Language technology middleware: lemmatization, decompounding, part-of-speech tagging, constituent chunking, shallow parsing, full parsing.
  • Text analytics middleware: named entity recognition, text classification, time and event detection, negation and modality detection.
  • Text preprocessing: spelling correction (including OCR correction), spelling variation detection, tokenization
  • Customization of shallow parsing to domain-specific text analytics (bio-informatics, medicine, newswire, spontaneous speech).
  • Customization of text analytics tools to domain-specific needs (intelligence, business document management, cultural heritage, humanities research).
For further information on consultancy and software licenses, contact stil@uvt.nl.

News

STIL Scriptieprijs / Thesis prize

The winner of the STIL Thesis Award 2014/2015 is Nikos Voskarides for his thesis Explaining relationships between entities. The prize was awarded on December 18, 2015, at CLIN-26 in Amsterdam.

Since 2010, STIL sponsors a prize of 1000 Euros for a Master's thesis that describes high quality research in computational linguistics or its applications. The prize is awarded at the yearly CLIN meeting.

MA theses submitted to and evaluated at a university or organisation of higher education in The Netherlands or Flanders are eligible for the prize. The thesis should address a topic in computational linguistics or its applications, and may be written in Dutch or English.

The candidates will be evaluated by a jury presided by the organizer of CLIN that year, and will consist further of the organizers of the two preceding CLIN meetings. The jury will decide in consensus to which candidate the prize will be awarded. The jury can also decide not to award a prize or to award the prize to a maximum of two candidates. In the latter case, the money prize will be shared.

All STIL Thesis Award winners:

MA theses submitted to and evaluated at a university or organisation of higher education in The Netherlands or Flanders are eligible for the prize. The thesis should address a topic in computational linguistics or its applications, and may be written in Dutch or English.   > READ MORE

STIL provides Dutch data for Ubisoft game My Word Coach

Ubisoft releases the brain training game My Word Coach (Dutch), a vocabulary booster for Nintendo DS and Wii. Background work on the Dutch localization was performed by Walter Daelemans (University of Antwerp) and Antal van den Bosch (Tilburg University) for STIL. News covered (in Dutch) by ComputerIdee and De Pers; an item at Radio 1 show Radio Online; Tilburg University press release; articles in Kennislink, Brabants Dagblad, Volkskrant. Game review at Gamespot. April 8, 2008

Demo of the day: MBSP, memory-based shallow parser for generic or bio-medical text. Enter an English sentence, and press "parse".


(Maximum characters: 200)

optimized for English bio-medical text (BioMinT MBSP: eng-biomint)
optimized for generic English text (WSJ-trained MBSP: eng)

© 2005-2014 STIL, stil@uvt.nl