Broadening the Scope of the EAGLES/ISLE Lexical Standardization Initiative

ISLE is a continuation of the long standing EAGLES initiative and it is supported by EC and NSF under the Human Language Technology (HLT) programme. Its objective is to develop widely agreed and urgently demanded standards and guidelines for infrastructural language resources, tools, and HLT products. EAGLES itself is a well-known trademark and point of reference for HLT projects and products and its previous results have already become de facto widely adopted standards. Multilingual computational lexicons, natural interaction and multimodality, and evaluation are the three areas targeted by ISLE. In the first section of the paper we describe the overall goals and methodology of EAGLES/ISLE, in the second section we focus on the work of the Computational Lexicon Working Group, introducing its work strategy and the preliminary guidelines of a standard framework for multilingual computational lexicons, based on a general schema for the “Multilingual ISLE Lexical Entry” (MILE). 1 Introducing EAGLES/ISLE ISLE (International Standards for Language Engineering) is a continuation of the long standing European EAGLES (Expert Advisory Group for Language Engineering Standards) initiative (Calzolari et al., 1996), carried out through a number of subsequent projects funded by the European Commission (EC) since 1993. ISLE is an initiative under the Human Language Technology (HLT) programme within the EU-US International Research Co-operation with the aim to develop and promote widely agreed and urgently demanded HLT standards, common guidelines and best practice recommendations for infrastructural language resources (Zampolli, 1998), (Calzolari, 1998), tools that exploit them, and language engineering products. Object of EAGLES/ISLE are large-scale language resources (such as text corpora, computational lexicons, speech corpora, multimodal resources), means of manipulating such knowledge via computational linguistic formalisms, mark-up languages and various software tools and means of assessing and evaluating resources, tools and products (EAGLES EWG final report, 1996). EAGLES was set up to determine which aspects of our field are open to short-term de facto standardisation and to encourage the development of such standards for the benefit of consumers and producers of language technology, through bringing together representatives of major collaborative European R&D projects, and of HLT industry, in relevant areas. In this respect, more than 150 leading industrial and academic players in the HLT field have actively participated in the definition of this initiative and have lent invaluable support to its execution. Successful standards are those which respond to commonly perceived needs or aid in overcoming common problems. In terms of offering workable, compromise solutions, they must be based on some solid platform of accepted facts and acceptable practices. The current ISLE project1 targets the three areas of : -multilingual computational lexicons2, -natural interaction and multimodality (NIMM)3, -evaluation of HLT systems4. For multilingual computational lexicons, ISLE goals are: i) extending EAGLES work on lexical semantics, necessary to establish interlanguage links; ii) designing and proposing standards for multilingual lexicons; iii) developing a prototype tool to implement lexicon guidelines and standards; iv) creating exemplary EAGLES-conformant sample lexicons and tagging exemplary corpora for validation purposes; v) developing standardised evaluation procedures for lexicons. For NIMM, ISLE work is targeted to develop guidelines for: i) the creation of NIMM data resources; ii) interpretative annotation of NIMM data, including spoken dialogue in NIMM contexts; iii) metadata descriptions for large NIMM resources; iv) annotation of discourse phenomena. For evaluation, ISLE is working on: i) quality models for machine translation systems; ii) maintenance of previous guidelines in an ISO based framework (ISO 9126, ISO 14598). Three Working Groups, and their sub-groups, carry out the work, according to the EAGLES 1 Coordinated by A. Zampolli for EU and M. Palmer for US, see http://www.ilc.pi.cnr.it/EAGLES96/isle/ISLE_Home_Pag