RELATOR
RELATOR |
|
The RELATOR project is a CEC-funded initiative which addresses the vital area of linguistic resources for spoken and written language processing. The RELATOR Consortium is currently establishing a repository of existing linguist resources and welcomes industrial participation at this crucial pilot stage. This document describes the RELATOR initiative and its importance to the European linguistic community.
RELATOR is a European-wide consortium of researchers who, with the support of the European Commission, are striving to establish a European repository of linguistic resources. Linguistic resources comprise a variety of spoken and written language materials, including lexicons, grammars, corpora, and spoken language databases. RELATOR will be focussing on -- but will be by no means limited to -- the nine working languages of the European Union. Initially, the RELATOR Consortium will address existing linguistic resources. Over time, it is envisaged that the RELATOR initiative will be expanded to encompass the development of new materials. RELATOR will be positioned as the European counterpart to the US Linguistic Data Consortium, an ARPA-funded initiative which has been addressing the needs of American researchers. RELATOR will ensure that the requirements of the European language processing community receive commensurate attention.
Who is behind RELATOR?
The Coordinating Partner of the RELATOR project is the Institute
for Computational Linguistics at the University of Pisa. Additional
Partners include CNRS-LIMSI, The Centre for Cognitive Science
(Edinburgh), and the German Institute for AI (DFKI) at the
University of Saarland. Associate partners are the Danish Center
for Language Technology, INESC Lisbon, and the Institut de la
Communication Parlee at Grenoble.
The RELATOR Consortium comprises representatives of major
coordinating bodies and associations, most notably ELSNET, ESCA,
and EACL, and will build on the previous like-minded efforts of
these groups. The Consortium is assisted by an Industrial Steering
Committee, which is made up of representatives of leading IT
companies, publishers, and other providers of electronic
information services.
RELATOR is supported by the European Commission within the
framework of the Linguistic Research and Engineering program. This
four-year program funds a select number of European research and
technical development projects in the field of language processing.
RELATOR is seen as a vital supplementary activity to the program's
varied array of basic research, technical development, and
evaluation-oriented projects.
Why was RELATOR established?
Language processing is an enabling technology which has the
potential of having profound impact across a vast spectrum of
modern day life, including information technology, education, and
telecommunications. However, a major bottleneck in the development
of applications in these sectors has been the lack of linguistic
resources which are essential for building portable and robust
systems with broad coverage. Being an immature science, language
processing also requires a substantial amount of further basic
research, much of which will largely depend on the availability of
large-scale linguistic resources. These matters are slowly being
addressed for the English language, but for other languages the
problem remains acute.
In the past, industry has been tackling some of these needs, but
efforts have been piecemeal at best -- and proprietary. Meanwhile,
fundamental economic changes have been taking place in the IT
industry as a whole, resulting in less funding available for the
medium- and long-term investments required to build commercial
language processing applications, a problem seen even in the
world's largest concerns. Moreover, the problem will only get
worse, as the globalization of the economy makes ever increasing
demands on national language support.
Given this troubling context, any duplication of effort in the
development of resources must be avoided. RELATOR was therefore
established to provide a mechanism for distributing this burden
among the providers and beneficiaries.
How will RELATOR achieve its goals?
The RELATOR Consortium's first actions were to appoint an
industrial steering committee, an editor/coordinator, and three
high-level consultants. The next step will be to attract interest
from a representative cross- section of industry. In the start-up
phase, the emphasis will be on identifying existing resources,
defining collective requirements, and establishing a prototypical
repository. Subsequently, the RELATOR consortium will experiment
with disseminating these materials by means of a distributed
electronic network. It will also investigate the possibility of
establishing CD-ROM pressing facilities in-house.
One of the most complex and urgent issues RELATOR will be grappling
with is the intellectual property considerations which surround the
collection and distribution of linguistic resources. At the behest
of the RELATOR Consortium, the Institut de Recherches Comparatives
sur les Institutions et le Droit (Paris) will be identifying the
relevant legal problems and will compile a report containing
suggestions for conditions governing use, production, and brokerage
of linguistic resources. It will also supply model licensing
agreements and suggest possible roles for national and
international funding agencies.
For RELATOR to succeed, it must ultimately satisfy the needs of a
broad range of European industry. For this reason, the Consortium
will be dependent on extensive input and feedback from industry
throughout the entire trajectory of the pilot project and beyond.
Timing
The European Commission has funded RELATOR for an initial timeframe
of eighteen months, starting January, 1994. Spring 1994 will see
the first meeting of the Industrial Steering Committee. A campaign
will also be launched at this time to attract a small number of
industrial partners. Within the first six months, a workshop will
be held to mobilize the European linguistic community regarding
RELATOR. Fall 1994 should see the establishment of an experimental
distributed database. A second workshop, possibly held in
conjunction with the International Forum and Conference on Language
Engineering and Industry (when?) will be held to gain feedback on the
interim report. A final workshop will gather recommendations to be
made in the final report.
While the initiative could receive further EU funding in subsequent
years, RELATOR will ultimately take the form of an independent,
self-sustaining entity.
Why you/your company should participate in RELATOR
Compiling linguistic resources for spoken and written language
processing applications is an expensive and time-consuming
undertaking. Because linguistic resources are essential, they are
largely pre- competitive. RELATOR aims to make such materials
available in a useful form at low cost to its members, thereby
freeing their resources for strategic implementation issues, such
as domain-specific requirements, performance, functionality, user
and program interfaces, and hardware platforms. Moreover, having
access to standardized linguistic resources which your company has
helped define could radically increase your time-to-market -- and
your commercial chances -- in new language areas. An investment of
time and resources in RELATOR today could mean manifold returns for
your company in the future.
Christoph Weyers (weyers@dfki.de)