Verbs represent a way In which ontological relationships between concepts and instances are expressed in natural language utterances. Moreover, an organized network of semantically related verbs can play a crucial role in applications. For example, if a Question-Answering system could exploit the direction of the entailment relation win → play, it may expand the question "Who played against Liverpool?" with "X won against Liverpool" and it may avoid the expansion of "Who won against Liverpool?" in "X played against Liverpool" that would be wrong. In this paper, we present a survey of the methods proposed to extract verb relations in corpora. These methods can be divided in two classes: those using the Harris distributional hypothesis and those based on point-wise assertions. These methods are analysed and compared. © Springer-Verlag Berlin Heidelberg 2006.
Pazienza, M.t., Pennacchiotti, M., Zanzotto, F.m. (2006). Discovering verb relations in corpora: Distributional versus non-distributional approaches. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (pp.1042-1052). Springer Verlag [10.1007/11779568_111].
Discovering verb relations in corpora: Distributional versus non-distributional approaches
PAZIENZA, MARIA TERESA;ZANZOTTO, FABIO MASSIMO
2006-01-01
Abstract
Verbs represent a way In which ontological relationships between concepts and instances are expressed in natural language utterances. Moreover, an organized network of semantically related verbs can play a crucial role in applications. For example, if a Question-Answering system could exploit the direction of the entailment relation win → play, it may expand the question "Who played against Liverpool?" with "X won against Liverpool" and it may avoid the expansion of "Who won against Liverpool?" in "X played against Liverpool" that would be wrong. In this paper, we present a survey of the methods proposed to extract verb relations in corpora. These methods can be divided in two classes: those using the Harris distributional hypothesis and those based on point-wise assertions. These methods are analysed and compared. © Springer-Verlag Berlin Heidelberg 2006.File | Dimensione | Formato | |
---|---|---|---|
2006_IEA_PazienzaPennacchiottiZanzotto.pdf
solo utenti autorizzati
Licenza:
Copyright dell'editore
Dimensione
175.86 kB
Formato
Adobe PDF
|
175.86 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.