[developers] Coref over ERSs?

Nikhil Krishnaswamy nkrishna at brandeis.edu
Wed Mar 13 18:46:46 CET 2019


Hi Woodley,

Thanks for getting in touch.  Insofar as I envision using MRS as a
resource, it would be plain text in single sentences or well-formed
sentence fragments.  The pipeline we're developing is still malleable
though, so it would be fairly simple to change formats or insert a
preprocessing step depending on the tools or resources already available
that we might want to make use of.

Thanks,
Nikhil

Nikhil Krishnaswamy, Ph.D.
Postdoctoral Researcher, Department of Computer Science




On Wed, Mar 13, 2019 at 1:43 PM Woodley Packard <sweaglesw at sweaglesw.org>
wrote:

> Hi Nikhil,
>
> In the past I worked on coreference resolution in MRS, although never
> quite to the point of a publication or software release.  Are you
> interested primarily in coreference within a single sentence or across
> multiple sentences?  Also, what format are you considering consuming MRS in
> (simple text based, DMRS, EDs, DM, ...)?  It’s possible some of my (oldish)
> tools could be of use to you.
>
> Regards,
> Woodley
>
> On Mar 7, 2019, at 8:54 AM, Emily M. Bender <ebender at uw.edu> wrote:
>
> Dear all,
>
> Does anyone know of systems that do coreference resolution over ERSs? See
> the question below from Nikhil Krishnaswamy who is interested in using
> these representations but needs coref:
>
> >>I had a question about ERS representations that I hope you'd be able to
> help with: in those cases where the ERG produces a representation without
> coref resolution, do you know of any resources that can take the output
> representation and use those syntactic or grammatical features to perform
> coref resolution?  So, in a simple case like "Grab the block and lift it
> up," where "the block" and "it" are not bound in the ERS but are both the
> endpoints of ARG2 arcs from the head verb, is there anything that can take
> that and actually perform the binding?
>
> Thanks,
> Emily
>
> --
> Emily M. Bender
> Professor, Department of Linguistics
> University of Washington
> Twitter: @emilymbender
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.delph-in.net/archives/developers/attachments/20190313/90d43dc1/attachment.html>


More information about the developers mailing list