Summary of the paper

Title Constructing a Textual Semantic Relation Corpus Using a Discourse Treebank
Authors Rui Wang and Caroline Sporleder
Abstract In this paper, we present our work on constructing a textual semantic relationcorpus by making use of an existing treebank annotated with discourserelations. We extract adjacent text span pairs and group them into sixcategories according to the different discourse relations between them. Afterthat, we present the details of our annotation scheme, which includes sixtextual semantic relations, 'backward entailment', 'forward entailment','equality', 'contradiction', 'overlapping', and 'independent'. We also discusssome ambiguous examples to show the difficulty of such annotation task, whichcannot be easily done by an automatic mapping between discourse relations andsemantic relations. We have two annotators and each of them performs the tasktwice. The basic statistics on the constructed corpus looks promising: weachieve 81.17% of agreement on the six semantic relation annotation with a .718kappa score, and it increases to 91.21% if we collapse the last two labels witha .775 kappa score.
Language Corpus (creation, annotation, etc.)
Topics Textual Entailment and Paraphrasing, Discourse annotation, representation and processing, Corpus (creation, annotation, etc.)
Full paper Constructing a Textual Semantic Relation Corpus Using a Discourse Treebank
Bibtex @InProceedings{WANG10.820,
  author = {Rui Wang and Caroline Sporleder},
  title = {Constructing a Textual Semantic Relation Corpus Using a Discourse Treebank},
  booktitle = {Proceedings of the Seventh conference on International Language Resources and Evaluation (LREC'10)},
  year = {2010},
  month = {may},
  date = {19-21},
  address = {Valletta, Malta},
  editor = {Nicoletta Calzolari (Conference Chair), Khalid Choukri, Bente Maegaard, Joseph Mariani, Jan Odjik, Stelios Piperidis, Mike Rosner, Daniel Tapias},
  publisher = {European Language Resources Association (ELRA)},
  isbn = {2-9517408-6-7},
  language = {english}
 }
Powered by ELDA © 2010 ELDA/ELRA