The Department of Linguistics and the Department of Computer Science at The University of Arizona invite you to a colloquium presentation by
University of Geneva
|Topic:||A Multilingual Paradigm for Automatic Verb Classification|
|Date:||Friday, March 14|
|Time:||2:00 - 3:15 PM|
A major problem in the automatic processing of language lies in the proper treatment of new or unforeseen words. How can a system know all the necessary information to deal with them? Of particular concern are verbs, crucial elements for interpretation. Our approach is to automatically group verbs into classes, which will provide the needed linguistic information for new verbs.
We will first show that simple linguistically-grounded statistics over a text corpus can be used to successfully classify verbs into classes that share both semantic and syntactic properties. We then illustrate the benefits of a multilingual approach to improve the coverage and accuracy of automatic verb classification. Our research incorporates two interrelated threads. In one thread, we exploit the similarities in the cross-linguistic classification of verbs, to extend work on English verb classification to a new language (Italian). In the other thread, we exploit the differences across languages in the syntactic expression of semantic properties, to show that complementary information about English verbs extracted from their translations in a second language (Chinese) improves classification performance of the English verbs.
Our results demonstrate an effective combination of deeper linguistic knowledge with the robustness and scalability of statistical techniques. This combination makes this work relevant for studies of human acquisition, on one hand, and large scale document processing, on the other hand.
ABOUT THE SPEAKER: Paola Merlo received her Laurea in Foreign Languages from the University of Venice, Ca' Foscari, in 1986, and her PhD in Linguistics, specializing in Computational Linguistics, from the University of Maryland in 1992, under the supervision of Amy Weinberg. She has worked at the University of Geneva since 1992, first as a post-doctoral assistant, then as Maître Assistante, and since 1999 as Maître d'Enseignement et de Recherche in Computer Science and Computational Linguistics. She is currently principal or co-principal investigator on research grants from the Swiss National Science Foundation, has published numerous journal articles, book chapters, and conference papers, is the author of Parsing with Principles and Classes of Information (Kluwer, 1996), and is co-editor of The Lexical Basis of Sentence Processing: Formal, Computational and Experimental Issues (Benjamins, 2002). She describes her main current area of research as "the automatic acquisition of verb knowledge from corpora and the automatic organisation of such knowledge in a classification".