Jiří Materna


Parameter Estimation for LDA-Frames
Jiří Materna
Proceedings of the 2013 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies


Using Ontologies for Semi-automatic Linking VerbaLex with FrameNet
Jiří Materna | Karel Pala
Proceedings of the Seventh International Conference on Language Resources and Evaluation (LREC'10)

This work presents a method of linking verbs and their valency frames in VerbaLex database developed at the Centre for NLP at the Faculty of Informatics Masaryk University to the frames in Berkeley FrameNet. While completely manual work may take a long time, the proposed semi-automatic approach requires a smaller amount of human effort to reach sufficient results. The method of linking VerbaLex frames to FrameNet frames consists of two phases. The goal of the first one is to find an appropriate FrameNet frame for each frame in VerbaLex. The second phase includes assigning FrameNet frame elements to the deep semantic roles in VerbaLex. In this work main emphasis is put on the exploitation of ontologies behind VerbaLex and FrameNet. Especially, the method of linking FrameNet frame elements with VerbaLex semantic roles is built using the information provided by the ontology of semantic types in FrameNet. Based on the proposed technique, a semi-automatic linking tool has been developed. By linking FrameNet to VerbaLex, we are able to find a non-trivial subset of the interlingual FrameNet frames (including their frame-to-frame relations), which could be used as a core for building FrameNet in Czech.