Non Standard Treebank Romania Republic of Moldova in the Universal Dependencies

Abstract

Morphological and syntactic annotated corpora are very important for language computerization. A Dependency Treebank was created in 2007 at the Al. I. Cuza University. 4,500 sentences of it were introduced into the Universal Dependencies (UD), consisting of over 80 treebanks for 50 languages annotated in unique conventions. Added to other 5,000 sentences annotated at the Artificial Intelligence Institute in Bucharest, they form a Romanian Contemporary Standard Treebank, affiliated with UD. But our Treebank has other non-standard language sentences. We intend to affiliate with UD a Non-Standard Romanian Dependency Treebank, having 15,000 sentences, part of them collected and annotated in the Republic of Moldova.