SIMPLE: From Monolingual to Multilingual resources
June 3rd, 2000
The objective of SIMPLE (project LE4-8346, IV Framework Programme of the European Union) was to add a layer of semantic information to the set of harmonised multifunctional lexica, built for 12 European languages by the PAROLE Consortium, which provide morphological and syntactic information for about 20,000 lexical entries. SIMPLE has added semantic information to approximately 3,500 lexical entries (about 10,000 senses) for each of the 12 languages.
The innovative linguistic model produced by SIMPLE integrates proposals of the major current state-of-the-art approaches to lexical semantics, and provides detailed technical specifications which should ensure coherence and harmonisation in the semantic encoding for the 12 languages.
The results of SIMPLE are an example of how harmonized lexica could be structured, and of which lexical information could be proposed for applications using Human Language Technologies.
Several National Projects currently aim at extending the linguistic coverage of the initial lexical nuclei produced by PAROLE/SIMPLE, and will create large computational lexica for the respective languages, harmonised according to the common PAROLE/SIMPLE model. They have decided to establish a Coordination Group (the ENABLER initiative).
Special emphasis is placed on the potentiality of the model: type of information encoded, format, reusability of the content. The model and the semantic information have been designed to be the bases for future cross-lingual links among the harmonized lexica of the 12 European languages.
SIMPLE, at the time of this workshop, will be near to completion. The goal of the workshop is to present the SIMPLE model and the lexical data to the R&D community, in order to stimulate and collect feedback on the results achieved, to discuss their use in natural language processing applications, and in particular to evaluate the possibility of developing lexical resources for multilingual information processing, starting from the semantic information provided by the SIMPLE model and data.
a) The SIMPLE Project
10:00 - 10:15 Introduction to the SIMPLE project
(Núria Bel – Grup d’Investigació en Lingüistica Computacional – Universitat de Barcelona, SIMPLE Technical Manager)
10:15 - 10:30 Presentation of the SIMPLE model: Linguistic Specifications
(Nicoletta Calzolari, Alessandro Lenci, Nilda Ruimy – Institute of Computational Linguistics, SIMPLE Linguistic Specifications Group)
10:30 - 10:45 Presentation of the SIMPLE resources: Content and formal issues
(Marta Villegas – Institut d’Estudis Catalans &– Grup d’Investigació en Lingüistica Computacional – Universitat de Barcelona)
10:45 - 11:00 Extensions of PAROLE & SIMPLE resources: National Projects
(Antonio Zampolli - Università di Pisa, SIMPLE co-ordinator)
11:00 - 11:20 Coffee Break
b) Semantic requirements for mono and multilingual processing tasks and applications
11:20 - 11:30 The EAGLES Guidelines for lexical semantic encoding as a basis for the development of multilingual lexical resources
(Antonio Sanfilippo, past EAGLES Computational Lexicon Working Group co-ordinator)
11:30 - 11:40 Semantic information for Multilingual Information Processing
(Martha Palmer, University of Pennsylvania)
11:40 - 12:00 Lexical Resources for MT products: new prototypes.
(Gregor Thurmair, SAIL-LABS )
12:00 - 12:20 Lexical Resources for Multilingual Information Management
12:20 - 13:00 General Discussion
(Chairman: Ralph Grishman, New York University)
SIMPLE lexical specifications can be found in:
Nuria Bel, SIMPLE Technical Manager
Nicoletta Calzolari, SIMPLE Linguistic Specifications Group
Antonio Sanfilippo, past EAGLES Computational Lexicon Working Group co-ordinator
Antonio Zampolli, SIMPLE coordinator