Forgive the catchy title, but this is a good article that I am using as a baseline for developing the UIMA annotator. In my opinion as a novice in OpenNLP and UIMA, I see the two main challenges in creating a Coreference Resolution annotator are (1) creating a Parse object from existing Chunk annotations or avoiding the creation of the parse object all together given that the sentence has already been parsed upstream and (2) determining the appropriate annotation that will exemplify the DiscourseEnitty and corresponding Mentions. If/when I finish, I will follow up on the thread.
http://blog.dpdearing.com/2012/11/making-coreference-resolution-with-opennlp-1-5-0-your-bitch/#comment-1373 On Jan 25, 2013, at 1:17 PM, Thomas Plagwitz <[email protected]> wrot > > > > > > I am an outsider > (background ESL with regard to this, among other SLA), but for what it is > worth: > > > > I > would definitely find it helpful to have an OpenNLP Coreference > Resolution in UIMA, especially in the hope that this would make it easy to > integrate that component into dkPRo (which has the StanfordCoreferenceResolver > in its GPL branch, but apparently only a stub for Coref in its ASL branch). > > > Thanks, Thomas > ---------------------------------------------------------------------------Dr. > Thomas Plagwitz |Director, Language Resource Center UNC Charlotte | Dept. of > Languages and Culture Studies9201 University City Blvd. | Charlotte, NC > 28223Phone: 704-687-8762 | Fax: 704-687-3496 http://plagwitz.org | > http://lrc.uncc.edu > >> From: [email protected] >> To: [email protected] >> Subject: Progress on UIMA - OpenNLP annotator for Coreference Resolution >> Date: Wed, 23 Jan 2013 22:28:11 +0000 >> >> Has there been any progress on the development of a UIMA - OpenNLP annotator >> for Coreference Resolution? And, if not would it be helpful to others if >> that annotator were developed? >> >> >> I did read the following post, but was unclear as to whether or not >> development of the annotator would proceed. >> http://mail-archives.apache.org/mod_mbox/opennlp-users/201112.mbox/%[email protected]%3E > >
