The SENSEVAL2 Panel on Domains, Topics and Senses

Paul Buitelaar
DFKI GmbH
Stuhlsatzenhausweg 3
D66123 Saarbruecken, Germany
paulb@dfki.de

1 Introduction: Why Domains
Matter in Sense Disambiguation
An important aspect of sense disambiguation is
the wider semantic space (domain, topic) in
which the ambiguous word occurs. This may be
most clearly illustrated by some crosslingual
examples, as they would appear in (machine)
translation. Consider for instance the English
word housing. In a more general ``sense'', this
translates in German into Wohnung. In an
engineering setting however it translates into
Gehuse. Also verbs may be translated
differently (i.e. have a different sense) according
to the semantic space in which they occur. For
instance, English warming up translates into
erhitzen in a more general sense, but into
aufwrmen in the sports domain.
Because of the apparent relevance then of
domains or topics on sense disambiguation, a
panel was organized at SENSEVAL2 to discuss
some current and previous work in this area. The
paper presents a more extended overview based
on the relevant literature, besides giving a
summary of the discussion that developed after
the panel presentations.

References
Agirre E., Ansa O., Hovy E., Martinez D.
Enriching very large ontologies using the WWW. In:
Proceedings of the Ontology Learning Workshop
ECAI 2000.
Agirre E., Ansa O., Martinez D., Hovy E.
Enriching WordNet concepts with topic signatures.
In: Proceedings NAACL WordNet Workshop, 2001.
Basili R., Della Rocca M., Pazienza M.T.
Contextual Word Sense Tuning and Disambiguation.
Applied Artificial Intelligence, vol. 11, 1997.
Buitelaar P., Sacaleanu B. Ranking and Selecting
Synsets by Domain Relevance. In: Proceedings
NAACL WordNet Workshop, 2001.
Cucchiarelli A., Velardi P. Finding a Domain
Appropriate Sense Inventory for Semantically
Tagging a Corpus. In: Journal of Natural Language
Engineering, 1998
Escudero G., Mrquez L., Rigau G. An Empirical
Study of the Domain Dependence of Supervised Word
Sense Disambiguation Systems. In: EMNLP 2000.
Gale W., Church K., Yarowsky D. One Sense per
Discourse. In: Proceedings of the 4 th DARPA
Speech and Natural Language Workshop, 1992.
Hearst M., Schtze H. Customizing a Lexicon to
Better Suit a Computational Task. In: Proceedings
ACL SIGLEX Workshop 1993.
Guthrie J. A., Guthrie I., Wilks Y., Aidinejad H.
Subject Dependent CoOccurrence and Word Sense
Disambiguation. In: Proceedings of ACL 1991.
Kilgarriff A. Bridging the gap between lexicon
and corpus: convergence of formalisms. In:
Proceedings of LREC Workshop on Adapting
Lexical Resources, 1998.
Krovetz R. More than one sense per discourse.
NEC Research Memorandum, 1998.
Magnini B., Cavagli G. Integrating Subject
Field Codes into WordNet. In: Proceedings LREC
2000.
Martinez D., Agirre E. One Sense per Collocation
and Genre/Topic Variations. In: Proceedings
EMNLP 2000.
Stevenson M., Wilks Y. Combining Weak
Knowledge Sources for Sense Disambiguation. In:
Proceedings IJCAI 1999.
Turcato D., Popowich F., Toole J., Fass D.,
Nicholson D., Tisher G. Adapting a synonym
database to specific domains. In: Proceedings of the
ACL workshop on recent advances in NLP and IR.
Hong Kong, 2000.
Vossen P. Extending, Trimming and Fusing
WordNet for Technical Documents. In: Proceedings
NAACL WordNet Workshop, 2001.
Walker D., Amsler R. The use of
machinereadable dictionaries in sublanguage
analysis In: Analyzing Language in Restricted
Domains, 1986.

