Gate Lookup annotations + Coreferencing

[Abdullah Bawakid <a....@gmail.com>]

Hello,

I am new to Apache UIMA and have been using Gate for the past two months for
few tasks. Now, that I started exploring the Apache UIMA, I am wondering
whether there is already an Analysis Engine (or annotator if using the GATE
terminology) or a tool that does the coreferencing between Named Entities
(i.e. same person appearing in several sentences).

Also, there is an annotation type called "lookups" that comes with Gate's
ANNIE. It gives various details about words in the form of majorType and
minorType. My interest is in particular for linguistic quantifiers (words
such as "more", "least",etc.). When using ANNIE, the lookup annotator
annotates these words and assign the majortype feature to "quantifier" and
minortype to "increase" and "decrease". Is there an available annotator that
does this ?

If not, what does the Apache UIMA community suggest for handling the
abovementioned issues ? or should I simply stick to Gate ?

Thanks
Abdul

Re: Gate Lookup annotations + Coreferencing

[Julien Nioche <li...@gmail.com>]

Hi Abdul,

I am new to Apache UIMA and have been using Gate for the past two months for
> few tasks. Now, that I started exploring the Apache UIMA, I am wondering
> whether there is already an Analysis Engine (or annotator if using the GATE
> terminology) or a tool that does the coreferencing between Named Entities
> (i.e. same person appearing in several sentences).


Not that I am aware of. What you could do would be to use another resource
for NE in UIMA (like OpenCalais) and port the code of the Coreference to
UIMA. Needs a bit of Java programming but that should be doable.

Also, there is an annotation type called "lookups" that comes with Gate's
> ANNIE. It gives various details about words in the form of majorType and
> minorType. My interest is in particular for linguistic quantifiers (words
> such as "more", "least",etc.). When using ANNIE, the lookup annotator
> annotates these words and assign the majortype feature to "quantifier" and
> minortype to "increase" and "decrease". Is there an available annotator
> that
> does this ?


The GATE gazetteer lists could be adapted to be used by the Dictionary
Annotator of UIMA (see
http://incubator.apache.org/uima/sandbox.html#dict.annotator). That should
be just a matter of converting the lists to the right format. I haven't used
the  Dictionary Annotator though so I can't tell you more about that.


>
> If not, what does the Apache UIMA community suggest for handling the
> abovementioned issues ? or should I simply stick to Gate ?


Another option would be to use the GATE plugin in UIMA. That would allow you
to use the original GATE components and their resources via the UIMA API.
This is certainly useful if your long term plans are to use UIMA only
otherwise it is probably easier to simply stick to GATE

HTH

Julien

-- 
DigitalPebble Ltd
http://www.digitalpebble.com