Computer Laboratory

Diarmuid Ó Séaghdha

Semantic classification with WordNet kernels

PDF

Diarmuid Ó Séaghdha

This paper presents methods for performing graph-based semantic classification using kernel functions defined on the WordNet lexical hierarchy. These functions are evaluated on the SemEval Task 4 relation classification dataset and their performance is shown to be competitive with that of more complex systems. A number of possible future developments are suggested to illustrate the flexibility of the approach.

@InProceedings{OSeaghdha:09,
  author = 	 {Diarmuid {\'O S\'eaghdha}},
  title = 	 {Semantic classification with {W}ord{N}et kernels},
  booktitle = 	 {Proceedings of the NAACL-HLT-09 Short Paper Session},
  year =	 2009,
  address =	 {Boulder, CO}
}