Disambiguating automatic semantic annotation based on a thesaurus structure
1549Authors
nan
Abstract
AbstractThe use/use for relationship a thesaurus is usually more complex than the (para-) synonymy recommended in the ISO-2788 standard describing the content of these controlled vocabularies. The fact that a non preferred term can refer to multiple preferred terms (only the latter are relevant in controlled indexing) makes this relationship difficult to use in automatic annotation applications : it generates ambiguity cases. In this paper, we present the CARROT algorithm, meant to rank the output of our Information Extraction pipeline, and how this algorithm can be used to select the relevant preferred term out of different possibilities. This selection is meant to provide suggestions of keywords to human annotators, in order to ease and speed up their daily process and is based on the structure of their thesaurus. We achieve a 95 % success, and discuss these results along with perspectives for this experiment.
Keywords
Full Content
TALN 2007, Toulouse, 5–8 juin 2007 Disambiguating automatic semantic annotation based on a thesaurus structure Véronique MALAISÉ1, Luit GAZENDAM2, Hennie BRUGMAN3 1 Vrije Universiteit, Amsterdam 2 Telematica Institute, Enschedé, Netherlands 3 Max Planck Institute for Psycholinguistics, Nijmegen, Netherlands vmalaise@few.vu.nl, Luit.Gazendam@telin.nl, Hennie.Brugman@mpi.nl Résumé. La relation voir/employé pour d’un thesaurus est souvent plus complexe que la (para-)synonymie recommandée par l’ISO-2788, standard décrivant le contenu de ces vocabu- laires contrôlés. Le fait qu’un non descripteur puisse renvoyer à plusieurs descripteurs (seuls les descripteurs sont pertinents dans le cadre de l’indexation contrôlée) fait que cette relation est complexe à utiliser dans un contexte d’annotation automatique : elle génère des cas d’am- biguité. Dans ce papier, nous présentons CARROT, un algorithme que nous avons mis au point pour classer les résultats de notre chaîne de traitements pour l’Extraction d’Information, et son utilisation dans le cadre de la sélection du descripteur pertinent lorsque plusieurs choix sont possibles. Cette sélection s’adresse à des documentalistes, dans le but de simplifier et d’accé- lérer leur travail, et se base sur la structure de leur thesaurus. Nous arrivons à un succès de 95 % dans nos suggestions ; nous discutons ces résultats et présentons des perspectives à cette expérimentation. Abstract. The use/use for relationship a thesaurus is usually more complex than the (para- ) synonymy recommended in the ISO-2788 standard describing the content of these controlled vocabularies. The fact that a non preferred term can refer to multiple preferred terms (only the latter are relevant in controlled indexing) makes this relationship difficult to use in automatic annotation applications : it generates ambiguity cases. In this paper, we present the CARROT algorithm, meant to rank the output of our Information Extraction pipeline, and how this al- gorithm can be used to select the relevant preferred term out of different possibilities. This selection is meant to provide suggestions of keywords to human annotators, in order to ease and speed up their daily process and is based on the structure of their thesaurus. We achieve a 95 % success, and discuss these results along with perspectives for this experiment. Mots-clés : désambiguisation sémantique, algorithme de classement, annotation auto- matique. Keywords: word sense disambiguation, ranking algorithm, automatic annotation. 197 Véronique MALAISÉ, Luit GAZENDAM, Hennie BRUGMAN 1 Introduction Thesauri are controlled vocabularies, often used for indexing and retrieving documents from collections. The standard thesauri contain two types of elements, preferred and non preferred terms, related with a link called use/use for. This link is considered as (para-)synonymy in the ISO-2788 standard (ISO, 1986) and can thus be useful for (semi-) automatic indexing appli- cations : it enables a program to index a document with a preferred term (which is the type of thesaurus based controlled annotation we are interested in) either if the document contains an occurrence of the preferred term or if it contains occurrences of the corresponding non pre- ferred term. In reality, this use/use for relationship is often more complex, and can generate ambiguity problems when used “as is” in an automatic …
Content truncated to 500 words
Labeled Text
AbstractThe use/use for relationship a thesaurus is usually more complex than the (para-) synonymy recommended in the ISO-2788 standard describing the content of these controlled vocabularies. The fact that a non preferred term can refer to multiple preferred terms (only the latter are relevant in controlled indexing) makes this relationship difficult to use in automatic annotation applications : it generates ambiguity cases. In this paper, we present the CARROT algorithm, meant to rank the output of our Information Extraction pipeline, and how this algorithm can be used to select the relevant preferred term out of different possibilities. This selection is meant to provide suggestions of keywords to human annotators, in order to ease and speed up their daily process and is based on the structure of their thesaurus. We achieve a 95 % success, and discuss these results along with perspectives for this experiment.
Publication Details
- Paper ID:
- 1549
- Year:
- 2007
- Publication:
- TIPSTER
- Publication Type:
- conf
- DOI:
- nan