Steps Towards Semantically Annotated Language Resources

Proceedings of the Fourth International Conference on Language Resources and Evaluation (LREC 2004)

Abstract

The use of textual resources such as text corpora, tree banks, large-scale lexica etc., has become a widely accepted commitment in the field of computational linguistics. However the scope of the annotations proposed has been unbalanced towards the 'surface' level. Only recently corpora with a deeper level of annotations have started to emerge. In this paper we describe a machine learning approach aimed at learning transformational rules that would allow the (partial) generation of semantic annotations starting from syntactic annotations generated by a parser (or created manually).