User login

About us
Text Analysis and Knowledge Engineering Lab is a research group at the Faculty of Electrical Engineering and Computing, University of Zagreb, Croatia. Our research focuses on the intersection of text mining, information retrieval, and natural language processing.

Latest news
RANLP 2013 in Hissar Print
Tuesday, 17 September 2013 17:27



TakeLab member Goran Glavaš participated at the Student Research Workshop of the International Conference on Recent Advances in Natural Language Processing (RANLP 2013) held in Hissar, Bulgaria (Sep 9-13) . Goran presented the paper "Event-Centered Simplification of News Stories" (done in collaboration with Sanja Štajner from the University of Wolverhampton, UK)  awarded with the Best Paper Award! More photos from Goran's Bulgarian experience inside!


Participation at ACL & BSNLP 2013 Print
Tuesday, 17 September 2013 15:59



TakeLab members Jan Šnajder, Mladen Karan, and Goran Glavaš participated in the world's most important NLP/CL conference -- ACL 2013 and the collocated Workshop on Balto-Slavic natural language processing (BSNLP) held from Aug 5 to Aug 10 in Sofia. Combined, TakeLab members published 7 papers (4 at ACL and 3 at BSNLP), which made TakeLab one of the most successful research groups at this year's ACL. Dive in for more information on published papers and photos from Sofia!


WSC in Hong Kong Print
Wednesday, 11 September 2013 14:35

hk2013-IMG 2621-mala

Takelab was invited to hold an invited lecture at the 59th ISI World Statistics Congress held in Hong Kong (25-30 Aug 2013). Our member Artur Šilić presented the work "A system for temporally-sliced correspondence analysis visualization" (Artur Šilić, Bojana Dalbelo Bašić, Annie Morin, Jean-Hugues Chauchat). This work is a part od Artur's PhD reaearch and also result of our long-lasting cooperation with colleagues from the institute IRISA in Rennes and the University of Lyon, France.


TakeLab at BSNLP 2013 Print
Sunday, 30 June 2013 15:35


After working on papers during the winter we decided to use spring to ... well ... work on more papers!. As a result, three papers co-authored by TakeLab members have been accepted for this years BSNLP workshop. The workshop will be held in conjunction with the ACL 2013 conference. We report our results on opinion/sentiment analysis, keyphrase extraction, and FAQ retrieval for Croatian. More details after the jump ... 

TakeLab excels at ACL 2013 Print
Wednesday, 29 May 2013 21:20


We've worked hard during the gloomy winter but now have something to show for it. Four papers co-autored by TakeLab members have been accepted for this year's ACL 2013, the most significant NLP conference, to be held in Sofia in August. We make contributions in the fields of computational morphology, distributional semantics, and event extraction. Read the nifty details after the jump...

TakeLab offers new courses Print
Wednesday, 29 May 2013 16:15

We at TakeLab are aware of the importance of teaching NLP, IR, and machine learning. After all, these are the things that we do for research, the things we do for living, and... the things we believe in! We're now finally up for the challenge and offer two new courses for MA and PhD students at UNIZG FER. Read more after the jump...

CICLing 2013 in Greece Print
Saturday, 13 April 2013 09:55


Takelab member Goran Glavaš participated in the 14th International Conference on Intelligent Text Processing and Computational Linguistics on the Samos island in Greece presenting the work entitled "Exploring Coreference Uncertainty of Generically Extracted Event Mentions" . Besides being an excellent academic event, conference also provided the opportunity to explore this beautiful Mediterranean island, home of Pythagoras, Epicurus, and Aesop.


Language Technologies Day Print
Wednesday, 05 December 2012 15:33

On November 30th, 2012, TakeLab participated in the "Language Technologies Day", an international conference as a part of the CESAR project. TakeLab members presented their latest products and language technologies to colleagues in academia and IT industry.
CroNER developed Print
Wednesday, 28 November 2012 14:15

croner slika

A state-of-the-art named entity extraction (NER) system for Croatian language has been developed in TakeLab. A web demonstration is available. The system can successfully extract many named entity types (most important being Person, Organisation, and Location) from text. A detailed description can be found in the paper.

TakeLab on Croatian Radio Print
Thursday, 29 November 2012 11:16
radio slika
TakeLab member Jan Šnajder was a guest on Croatian Radio in the "Oko znanosti" show.  He presented activities of TakeLab as a combination of text mining, information retrieval and natural language processing. Also, he explained the complexities as well as many useful and beneficial applications in these areas. The full show can be found online.