Hi Venkat,
here are some comments after a quick read. I am sure Fran Tyers can give
you better feedback (I just advised on the thesis, he worked it out).
A quick summary of what it is in Nikita's work that needs improvement
and why would be welcome for us who are not too familiar.
I would consider KenLM in addition to IRSTLM (I think it works faster
and it is all free/open-source, at least the part we would need here).
I am not sure if you need all the power of GIZA++. Other faster
aligners may be used and installed, and perhaps they are OK for the
task (e.g. https://github.com/clab/fast_align). In fact, some of the
task could simply be done using the bilingual dictionary to find hits,
without need for detailed alignments (poor man's in Nikita's work).
Hope this helps
Mikel
El 01/04/17 a les 10:47, Venkat Parthasarathy ha escrit:
Have uploaded my proposal to the wiki
<http://wiki.apertium.org/wiki/User:Venkat1997/proposal>. Please
provide feedback!
------------------------------------------------------------------------------
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
[email protected]
https://lists.sourceforge.net/lists/listinfo/apertium-stuff
--
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
[email protected]
https://lists.sourceforge.net/lists/listinfo/apertium-stuff
