Following SemEval (see e.g. http://alt.qcri.org/semeval2017/) and its local variants (such as GermEval, https://sites.google.com/view/germeval2017-absa/) we invite you to take part in PolEval competition for NLP tools for Polish. It is intended to improve state-of-the-art by engaging both local and international participants. Thanks to hospitality of LTC 2017 (Language & Technology Conference: Human Language Technologies as a Challenge for Computer Science and Linguistics, http://ltc.amu.edu.pl) PolEval 2017 results will be presented in a special session.
You will find all details at http://poleval.pl (in English) and a short summary below. Timeframe is short so please help advertising PolEval 2017 among potential participants (colleagues, students, NLP enthusiasts) - it can make a nice machine learning or neural network task even if they have not heard about Polish yet.
Tasks and subtasks -----------
1. POS Tagging 1a. Morphosyntactic disambiguation and guessing 1b. Lemmatisation 1c. Complete system: POS tagging 2. Sentiment analysis
Users are evaluating their systems on their own using test data and scorer provided by PolEval 2017 organisers. At the same time official acceptance of the results requires sharing the source code to let task managers verify the score.
Important dates -----------
- June 14: training data and evaluation procedure available - until June 30: scorer and test data available - until July 16: systems officially registered and source code shared - until July 23: results announced - until August 31: LTC 2017 papers describing participating systems registered for review - until September 7: decisions and review results sent to participants - until September 15: camera-ready papers sent to organizers - November 17-19: accepted papers presented at LTC 2017
Best regards, Maciej Ogrodniczuk, Łukasz Kobyliński, Aleksander Wawer Institute of Computer Science, Polish Academy of Sciences