Objetivo The ACQUILEX Action aimed at solving problems in the following areas: -the design of a general computational model of a dictionary entry -the degree to which machine-readable dictionaries contain lexical, syntactic and semantic information which can be made explicit and can be reused in natural language processing systems -the development of methods and techniques for the extraction of syntactic and semantic information from machine-readable dictionaries -the semi-automatic construction of taxonomies starting from machine-readable dictionaries -the design and organisation of a single lexical knowledge-base with concepts and relations for the different languages involved.Techniques and methodologies for using existing machine readable dictionaries in the construction of components for natural language processing systems were developed. The main focus was on the extraction of lexical information from multiple machine readable sources in a multilingual context, with the overall goal of constructing a single multilingual lexical knowledge base.The results achieved so far involve:a second release of the lexical database software;conversion of the machine readable dictionary sources onto the lexical database (LDB);supply of semantic taxonomies;an English analyser system;first releases of the lexical knowledge base software;definition of common set of types and features structures.Research is currently being undertaken on:continuing the construction of taxonomies for the different monolingual dictionaries;merging of taxonomies derived from different dictionaries;implementing the lexical knowledge base system;definition of a system of typed feature structures for the formalization of syntactic and semantic information;development of a prototype of the natural language processing test bed system.APPROACH AND METHODS The approach was based on theoretical linguistics, computational lexicography and lexicology, and computational linguistics. The main themes were: -A formal analysis of the structure and content of existing machine-readable dictionaries, in order to design an explicit and standardised representation language for the computational model of the dictionary entry. -The design of procedures for the extraction of super-ordinates from natural language definitions, for their disambiguation, and for the construction of taxonomies throughout the lexicon. -The design of procedures for the linguistic and computational analysis of natural language definitions, with the aim of extracting all the semantic information implicit in them. -The study of ways of representing the semantic and syntactic information which is extracted in a unification-based lexical representation language using typed feature structures and supporting default inheritance. -The design and implementation of basic software for the creation, accessing and processing of lexical databases and lexical knowledge-bases. -The study of how to link taxonomies and conceptual or relational information coming from different sources (either monolingual or multilingual). -The design of a natural language processing test-bed for the information extracted from machine-readable dictionaries. PROGRESS AND RESULTS - STATUS AS OF OCTOBER 1991 The results achieved in the last year are: -a second release of the lexical database software -conversion of the machine-readable dictionary sources onto the LDB -supply of semantic taxonomies -English analyser system -first releases of the Lexical Knowledge Base Software -definition of common set of types and feature structures The Action is presently working on: -continuing the construction of taxonomies for the different monolingual dictionaries; -merging of taxonomies derived from different dictionaries; -implementing the lexical knowledge base system; -definition of a system of typed feature structures for the formalisation of syntactic and semantic information; -development of a prototype of the natural language processing test-bed system. POTENTIAL The research themes tackled within ACQUILEX aimed to meet one of the major bottlenecks of natural language processing: the availability of large computational lexicons with particular emphasis on making semantic information explicit and accessible. Ámbito científico humanitieslanguages and literaturegeneral language studieshumanitieslanguages and literaturelinguisticsnatural sciencescomputer and information sciencessoftwarenatural sciencescomputer and information sciencesdatabasesnatural sciencescomputer and information sciencesdata sciencenatural language processing Programa(s) FP2-ESPRIT 2 - European strategic programme (EEC) for research and development in information technologies (ESPRIT), 1987-1992 Tema(s) Data not available Convocatoria de propuestas Data not available Régimen de financiación Data not available Coordinador Università degli Studi di Pisa Aportación de la UE Sin datos Dirección Via Risorgimento 9 56126 Pisa Italia Ver en el mapa Coste total Sin datos Participantes (6) Ordenar alfabéticamente Ordenar por aportación de la UE Ampliar todo Contraer todo Cambridge University Press Reino Unido Aportación de la UE Sin datos Dirección Shaftesbury Road, CB2 2RU Cambridge Ver en el mapa Coste total Sin datos Consiglio Nazionale delle Ricerche (CNR) Italia Aportación de la UE Sin datos Dirección Via della Faggiola 32 56100 Pisa Ver en el mapa Coste total Sin datos UNIVERSITAT POLITECNICA DE CATALUNYA España Aportación de la UE Sin datos Dirección C/PAU GARGALLO, APARTADO 30002, 5 08028 BARCELONA Ver en el mapa Coste total Sin datos UNIVERSITEIT VAN AMSTERDAM Países Bajos Aportación de la UE Sin datos Dirección SPUISTRAAT 210 1012 VT AMSTERDAM Ver en el mapa Coste total Sin datos UNIVERSITY COLLEGE DUBLIN Irlanda Aportación de la UE Sin datos Dirección BELFIELD DUBLIN 4 Ver en el mapa Coste total Sin datos University of Cambridge Reino Unido Aportación de la UE Sin datos Dirección New Museums Site Pembroke Street CB2 3QG Cambridge Ver en el mapa Coste total Sin datos
