Dear Apertiumers,
As I said in my earlier message, I travelled to Olot to attend a
Wikimedia prehackathon organized by Amical Wikimedia a Catalan
pro-Wikimedia organization. As you know, Apertium is one of the MT
systems used in their Content Translation facility, which is used to
translate Wikipedia articles.
Discussions there gave rise to two possible projects:
1. Using the database of post-edited machine translations in Content
Translation. I have revived an old GSoC idea and I have added it to
http://wiki.apertium.org/wiki/Ideas_for_Google_Summer_of_Code
("Improving language pairs mining Mediawiki Content Translation
postedits"). I have also dug out some code I had and created a GitHub
repository. I would love having feedback and additional mentors for this
idea.
2. Mediawiki is creating a lexical database called Lexeme:
https://www.mediawiki.org/wiki/Extension:WikibaseLexeme . The data there
is expected to have a CC0 license and will be wiki-edited. This means
that we would be able to use their data to improve Apertium language
pairs (which are GPL) (this means that Apertium data could not be
uploaded to Lexeme unless all authors agreed in releasing it as CC0
too). People there may be open to patterning Lexeme entries so that they
contain all the information needed in Apertium. In the horizon, the idea
of using a wiki way to generate Apertium dictionaries is attractive. Not
a mature GSoC idea yet, but I would like to see if there would be
interest, and I would then turn it into a more detailed idea (after
talking to people in Wikimedia Foundation such as Amir Aharoni and Laura
Pintscher). There is already a ticket in their web:
https://phabricator.wikimedia.org/T186420
[1]
https://www.mediawiki.org/wiki/Content_translation/Product_Definition/LinearDoc#Machine_translation
Any ideas/comments welcome!
Cheers
Mikel
--
Mikel L. Forcada http://www.dlsi.ua.es/~mlf/
Departament de Llenguatges i Sistemes Informàtics
Universitat d'Alacant
E-03690 Sant Vicent del Raspeig
Spain
Office: +34 96 590 9776
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Apertium-stuff mailing list
Apertium-stuff@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/apertium-stuff
