Produção de textos paralelos em língua portuguesa e uma interlíngua de libras

Abstract

This work embraces the Human-Computer Interaction and the Natural Language Processing areas of knowledge, touching accessibility concepts, based on one textual simplification architecture and on one corpus structure. The result of this research is the generation of parallel texts – in the Portuguese Language version and in Brazilian Sign Language (LIBRAS) version – with the objective of supplying a data repository created for future works in the Natural Language Processing area that have relation with the language of the Brazilian deaf people. This work presents a theoretical revision where are discussed questions like studies about LIBRAS linguistic aspects, acquisition of a second language in the deaf community as well as concepts of textual simplification. Complementing this discussion, related works are presented as contributions concerning to the application of the referred concepts. Based on this revision, this work studies the architecture and the functionalities of the Editor de Anotações de Simplificação, in the context of PorSimples project, as a platform for the development of a module of textual generation in a LIBRAS interlingua which is based in a simplification process. At last, the functionalities of the module are explored by a LIBRAS expert and the generated texts are analyzed searching for future improvements in the process of text production.