Exploitable results
The project addresses the need for high-quality, large-scale grammatical resources as a basis for natural language (NL) technology. Commercial resources so far often lack properties such as stability, extensibility, modularity, and quality. Unification-based grammars could fill this gap, but so far they have been limited to academia and not of much practical relevance, as they are too inefficient to be practically applicable. LS-GRAM narrows down the gap between the scientific concept of unification grammar and real industrial applications. To overcome the inefficiency of unification grammars the project uses a 'lean' formalism included in the ALEP system. The basic idea is that the grammar developer sacrifices some formal constructs and in order to gain the efficiency he needs for developing realistic grammars. Essentially, the formalism provides constructs which are directly compilable into Prolog, making use of efficient term unification. Results achieved so far consist of very carefully documented grammars for 3 languages (Spanish, English and German), grammar design for 6 languages (Greek, Danish, Dutch, French, Italian and Portuguese), 9 corpus investigations, specifications for semantic output structures, and an ALEP rule coding manual. Also nearly a dozen publications have resulted from the project.
Buscando datos de OpenAIRE...
Se ha producido un error en la búsqueda de datos de OpenAIRE
No hay resultados disponibles