Hi there,
some comments:

>
> apertium-tagger-training-tools: Machine learning of better taggers using
> target-language information
You can read about this in Felipe Sánchez-Martínez, Juan Antonio 
Pérez-Ortiz, Mikel L. Forcada. Using target-language information to 
train part-of-speech taggers for machine translation. In Machine 
Translation, volume 22, issue 1-2, p. 29-66. 
(http://www.springerlink.com/content/m452802q3536044v/fulltext.pdf)

If you are behind a paywall, ask me for a copy of our paper.
> apertium-transfer-tools: Machine learning of transfer rules using
> parallel corpora

You can read about this in Felipe Sánchez-Martínez, Mikel L. Forcada. 
Inferring shallow-transfer machine translation rules from small parallel 
corpora. In Journal of Artificial Intelligence Research. volume 34, p. 
605-635 (http://www.dlsi.ua.es/~fsanchez/pub/pdf/sanchez-martinez09b.pdf).

Felipe Sánchez-Martínez and Juan Antonio Pérez-Ortiz are currently 
advising on Víctor Sánchez-Cartagena's thesis which builds upon the 
results in that paper. Nothing has been published yet as far as I know.
> apertium-lex-tools: Machine learning of lexical selection rules using
> both monolingual and parallel corpora.
Fran is preparing a full paper on this (this was his PhD thesis), but he 
published some preliminary results: Francis M. Tyers, Felipe 
Sánchez-Martínez, Mikel L. Forcada. Flexible finite-state lexical 
selection for rule-based machine translation. In Proceedings of the 16th 
Annual Conference of the European Association for Machine Translation, 
p. 213-220, May 28-30, 2012, Trento, Italy 
(http://www.dlsi.ua.es/~fsanchez/pub/pdf/tyers12a.pdf)
>
>> >Is machine learning now synonymous with statistical machine
>> >translation?
> I don't think so. There are plenty of ways to apply machine learning
> techniques to rule-based machine translation.
Completely agreed.
>
>> >Do you think this task described in the ideas page would be a good
>> >candidate for a machine learning algorithm:
>> >
>> >
>> >Apertium assimilation evaluation toolkit
> No, I don't think that this would be a good candidate for a machine
> learning algorithm.
The ideas behind this are described in a paper too: J. O'Regan, M.L. 
Forcada, " Peeking through the language barrier: the development of a 
free/open-source gisting system for Basque to English based on 
apertium.org ", Procesamiento del Lenguaje Natural, (XXIX Congreso de la 
Sociedad Española de Procesamiento del Lenguaje Natural, Madrid, Spain, 
16-18.09.2013) 51, 15-22 (http://www.dlsi.ua.es/~mlf/docum/forcada13p.pdf).

I think there is a bit of room for machine learning if one obtains 
success rates for various hole-poking frequencies and then tries to 
combine them in some way that correlates with another evaluation 
measure, for instance. But that would not be the toolkit per se, but 
some research on top of results obtained with it.
> Some of the ideas that would require machine
> learning would be:
>
> Corpus-based lexicalised feature transfer
> Improvements in lexical-selection module
> Accent and diacritic restoration
>
> There may be more tasks in the future.
>
>> >Lastly, I can't remember or access information on alternative ways to
>> >view Apertium other than through Apertium Viewer (I am using Mac).
> You can use the command line:)
Yeah! There is also apertium-caffeine 
(http://wiki.apertium.org/wiki/Apertium-Caffeine) but I would also 
recommend installing apertium from the repository and compiling it 
yourself, so you can tie it down and do perverse things to it to see how 
it reacts ;-)

Some of us hang on the #apertium IRC channel (at freenode.net), and our 
wiki is a great source of information:
http://wiki.apertium.org/wiki/Main_Page


Have fun

Mikel

-- 
Mikel L. Forcada (http://www.dlsi.ua.es/~mlf/)
Departament de Llenguatges i Sistemes Informàtics
Universitat d'Alacant
E-03071 Alacant, Spain
Phone: +34 96 590 9776
Fax: +34 96 590 9326


------------------------------------------------------------------------------
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments & Everything In Between.
Get a Quote or Start a Free Trial Today. 
http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
_______________________________________________
Apertium-stuff mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/apertium-stuff

Reply via email to