[Corpora] [Corpora-List] Named entities and abstract codes

[Evans, Richard J]

Dear Leon,

I seem to remember than in the ACE tasks, the terms "entities" (involved in the key facts of interest in the information extraction task) and "mentions" (references to those entities) were being used. Using ACE terminology, can we consider "named entities" to be those entities whose mentions are typically proper names? Then in clinical IE tasks, for example, entities include symptoms, clinical findings, and anatomical locations, etc., which are not referred to using proper names and so are not named entities. By contrast, in MUC they are locations, organisations, etc., which are named entities.

Best regards,

Richard



Richard Evans,
Research Fellow,
Computational Linguistics Research Group,
Research Institute of Information and Language Processing,
University of Wolverhampton,
United Kingdom.
http://clg.wlv.ac.uk/~richard<http://clg.wlv.ac.uk/~richard/>
________________________________
From: corpora-bounces at uib.no [corpora-bounces at uib.no] on behalf of Satoshi Sekine [sekine at cs.nyu.edu]
Sent: 03 November 2014 15:35
To: leon at dcs.shef.ac.uk
Cc: corpora at lists.uib.no
Subject: Re: [Corpora-List] Named entities and abstract codes

Dear Leon,


I think I'm not directly answering your question (asking related literature), but I just want to give a comment. If you are building an application which needs to recognize abstract codes, don't be afraid to make such category. We don't have to start designing it from the definition of "named entity". Rather, we should start design it from what is important for your application. We may be able to call them "targets of interest" or "important designator".

Actually, when we designed 200 extended named entity, which includes color name, animal, ID number (like yours), job title etc, we got comments that these are not named entities. We spent sometime trying to coin a new term to describe it, but it was not successful. (if any of you come up with a good name, please let me know).

http://nlp.cs.nyu.edu/ene/version7_1_0Beng.html
http://cs.nyu.edu/~sekine/papers/lrec04-65.pdf

Thanks,
Satoshi Sekine



On 2014/11/03, at 7:24, Leon Derczynski wrote:
Dear list,

Are "abstract codes" named entities? For example, "The *15:07 train to Sheffield*", "*Flight MH17*", "The new *Canon 50D*", "pass me *document 123*"?

One generic definition of a named entity is, a phrase that is a rigid designator. When we talk about "Lars von Trier", it is fairly uncontroversial to claim that this fits the rigid designator definition well. But using a less descriptive, more abstract phrase, like a document identifier, does the definition fit just as well? Is there some related literature?

All the best,


Leon

--
Leon R A Derczynski
Research Associate, NLP Group

Department of Computer Science
University of Sheffield, UK

Voted number one for student experience
Times Higher Education Student Experience Survey 2014-2015<tel:2014-2015>

http://www.dcs.shef.ac.uk/~leon/
_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no<mailto:Corpora at uib.no>
http://mailman.uib.no/listinfo/corpora

------------------------------
Satoshi Sekine
sekine at cs.nyu.edu<mailto:sekine at cs.nyu.edu>




-- 
Scanned by iCritical.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20141103/b0e04dad/attachment.htm>
-------------- next part --------------
_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora