ParGram January 2013

pargram@mailman.uni-konstanz.de

3 participants
3 discussions

Paper on Parallel Treebanking - Call for Contribution
by Sebastian Sulger 22 Jan '13

22 Jan '13

Dear ParGram Community, we are preparing an ACL paper on parallel treebanking at the moment, incorporating insights from the ParGram project. The long paper deadline for ACL2013 is February 20th. The main focus of the treebank would be on its typological diversity and depth of linguistic representation. As of now, we have received c- and f-structures produced by ParGram grammars for 8 languages. The essential argumentation for the paper would be as follows: - The treebank is yet quite small, but includes many languages. - Its representations are produced in an automatic fashion by grammars. - One can rely on parallelism in analyses and features/values, since ParGram has (th|f)ought about these a lot for >10 years. - The representations are deep. - The multitude of languages/depth of representation makes the treebank attractive for typologists and syntacticians/morphologists. Following is a short excerpt from the introduction: "An obvious application for parallel treebanking is machine translation, where treebank size is a deciding factor for or against choosing a particular treebank. When conducting in-depth linguistic studies of typological features, other factors (such as number of included languages, number of covered phenomena, depth of linguistic analysis) become more important. The treebanking effort reported on in this paper supports work of the latter mentality. We have created a parallel treebank which is yet quite small, but includes a multitude of eight typologically quite diverse languages as well as different constructions." Of course, one would then try to pick out interesting linguistic issues in the representation, or try to look at how different phenomena are handled across the grammars. If anyone of you is interested in contributing to the paper (be it by writing the paper itself, or by sending more structures for their language), please contact us. Also, if anyone has an opinion about the above argumentation, please feel free to share it with us. So far, best wishes to all of you, Jani (from the Urdu ParGram team in Konstanz) -- Sebastian Sulger FB Sprachwissenschaft Universitaet Konstanz http://ling.uni-konstanz.de/pages/home/sulger

1 0

publishing an XLE grammar?
by Adam Przepiorkowski 11 Jan '13

11 Jan '13

Dear ParGrammarians, We are planning to publish the current version of the Polish XLE grammar on an open licence, and we're looking for other XLE grammars published this way, to learn about best practices. Is it a problem that the grammar contains the PARGRAM feature declaration? Please advise. All best, Adam P. -- Adam Przepiórkowski ˈadam ˌpʃɛpjurˈkɔfskʲi http://clip.ipipan.waw.pl/ ____ Computational Linguistics in Poland http://jlm.ipipan.waw.pl/ ___________ Journal of Language Modelling http://zil.ipipan.waw.pl/ ____________ Linguistic Engineering Group http://nkjp.pl/ _________________________ National Corpus of Polish

1 0

Stochastic disambiguation with XLE?
by Katarzyna Krasnowska 09 Jan '13

09 Jan '13

Dear users of XLE, I would like to use the probabilistic parsing functionality of the XLE parser, but have trouble getting it to run. I would be very grateful if someone who has trained a property weights file for XLE and then used it for statistical disambiguation of parses could explain how to do it properly. Regards, Katarzyna Krasnowska PhD student Polish Academy of Sciences

1 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

ParGram January 2013