16.3008, Diss: Computational Linguistics: Wagner: 'Learning T...'

[LINGUIST List]

LINGUIST List: Vol-16-3008. Tue Oct 18 2005. ISSN: 1068 - 4875.

Subject: 16.3008, Diss: Computational Linguistics: Wagner: 'Learning T...'

Moderators: Anthony Aristar, Wayne State U <aristar at linguistlist.org>
            Helen Aristar-Dry, Eastern Michigan U <hdry at linguistlist.org>
 
Reviews (reviews at linguistlist.org) 
        Sheila Dooley, U of Arizona  
        Terry Langendoen, U of Arizona  

Homepage: http://linguistlist.org/

The LINGUIST List is funded by Eastern Michigan University, Wayne
State University, and donations from subscribers and publishers.

Editor for this issue: Meredith Valant <av8736 at wayne.edu>
================================================================  

To post to LINGUIST, use our convenient web form at
http://linguistlist.org/LL/posttolinguist.html.


===========================Directory==============================  

1)
Date: 18-Oct-2005
From: Andreas Wagner < andreas.wagner at uni-due.de >
Subject: Learning Thematic Role Relations for Lexical Semantic Nets 

	
-------------------------Message 1 ---------------------------------- 
Date: Tue, 18 Oct 2005 10:40:21
From: Andreas Wagner < andreas.wagner at uni-due.de >
Subject: Learning Thematic Role Relations for Lexical Semantic Nets 
 


Institution: University of Tübingen 
Program: Department of Romance Languages 
Dissertation Status: Completed 
Degree Date: 2004 

Author: Andreas Wagner

Dissertation Title: Learning Thematic Role Relations for Lexical Semantic Nets 

Dissertation URL:  http://w210.ub.uni-tuebingen.de/dbt/volltexte/2005/2005/

Linguistic Field(s): Computational Linguistics

Subject Language(s): English (eng)


Dissertation Director(s):
Erhard W. Hinrichs
Martin Volk

Dissertation Abstract:

This thesis presents a strategy for the acquisition of thematic role
relations (such as AGENT, PATIENT, or INSTRUMENT) by means of statistical
corpus analysis, for the purpose of semi-automatically extending
lexical-semantic nets. In particular, this work focuses on resources in the
style of WordNet (Fellbaum 1998) and EuroWordNet (Vossen 1999).
Lexical-semantic nets represent the meanings of words via semantic
relations between words and/or word concepts. Semantic (thematic) role
relations are conceptual relations which hold between verbs and their
nominal arguments (e.g. {eat}--AGENT--{human} or  {eat}--PATIENT--{food}).
Such relations capture selectional restrictions of verbs. Therefore, the
task of acquiring thematic role relations is intrinsically related to the
task of acquiring selectional restrictions.

Consequently, the core of a strategy for learning role relations consists
in a method for learning selectional restrictions (or, more precisely,
selectional preferences). For the latter task, a number of methods have
been proposed which utilise syntactically analysed corpora and WordNet. To
acquire the selectional preferences of a certain verb for a certain
argument, the respective complement nouns of that verb are extracted from
the corpus, and statistical methods are applied to generalise over these
nouns; these generalisations are expressed as a set of WordNet noun
concepts. One of these approaches, namely the method proposed by (Abe & Li
1996), constitutes the starting point of my research. However, this
approach is not immediately applicable for learning role relations, but
requires modifications and extensions for that task. In particular, two
aspects have to be taken into account. Firstly, it is crucial that the
WordNet concepts acquired to represent selectional preferences of a verb
are located at an appropriate level of generalisation (e.g. {food} as
PATIENT of {eat}, rather than {cake} or {physical_object}). I develop a
modification of the approach which substantially improves its performance
in this respect. Secondly, as the existing methods generalise over
syntactic complements, they acquire selectional preferences for syntactic
rather than semantic arguments. To learn selectional preferences for
semantic roles, the syntactic arguments provided by the parsed corpus have
to be linked to their underlying roles so that the statistical learning
method can be applied to generalise, for example, over all (semantic)
Agents of the examined verb rather than over all its (syntactic) subjects.
Therefore, I develop a method for linking syntactic to semantic arguments.
A further aspect of the overall strategy I present is an appropriate method
for mapping the verbs and nouns in the training data to the corresponding
WordNet concepts, which is a prerequisite for applying the preference
acquisition algorithm.

To evaluate the role acquisition approach developed in this thesis, I
extract a gold standard from the EuroWordNet database and propose detailed
evaluation criteria. Overall, the evaluation results (accuracy rates of up
to 84%) show that the approach works effectively. 




-----------------------------------------------------------
LINGUIST List: Vol-16-3008