LINGUIST List 17.1429: Computational Ling: Poria: 'An Engineering App...'

LINGUIST List 17.1429

Tue May 09 2006

Diss: Computational Ling: Poria: 'An Engineering App...'

Editor for this issue: Meredith Valant <meredithlinguistlist.org>

Directory 1. Sanjay Poria, An Engineering Approach to Knowledge Acquisition by the Interactive Analysis of Dictionary Definitions

Message 1: An Engineering Approach to Knowledge Acquisition by the Interactive Analysis of Dictionary Definitions
Date: 09-May-2006
From: Sanjay Poria <sanjayporiahotmail.com>
Subject: An Engineering Approach to Knowledge Acquisition by the Interactive Analysis of Dictionary Definitions

Institution: University of Durham Program: Department of Computer Science Dissertation Status: Completed Degree Date: 1998

Author: Sanjay Poria

Dissertation Title: An Engineering Approach to Knowledge Acquisition by the Interactive Analysis of Dictionary Definitions

Linguistic Field(s): Computational Linguistics
Subject Language(s): None ()
Dissertation Director:
Roberto Garigliano
Dissertation Abstract:

It has long been recognised that everyday dictionaries are a potentialsource of lexical and world knowledge of the type required by many NaturalLanguage Processing (NLP) systems. This research presents a semi-automatedapproach to the extraction of rich semantic relationships from dictionarydefinitions. The definitions are taken from the recently published'Cambridge International Dictionary of English' (CIDE). The thesisillustrates how many of the innovative features of CIDE can be exploitedduring the knowledge acquisition process.

The approach introduced in this thesis uses the LOLITA NLP system toextract and represent semantic relationships, along with a human operatorto resolve the different forms of ambiguity which exist within dictionarydefinitions. Such a strategy combines the strengths of both participants inthe acquisition process: automated procedures provide consistency in theconstruction of complex and inter-related semantic relationships, while thehuman participant can use his or her knowledge to determine the correctinterpretation of a definition.

This semi-automated strategy eliminates the weakness of many existingapproaches because it guarantees feasibility and correctness: feasibilityis ensured by exploiting LOLITA's existing NLP capabilities so that humanswith minimal linguistic training can resolve the ambiguities withindictionary definitions; and correctness is ensured because incorrectlyinterpreted definitions can be manually eliminated.