MARCELL: Multilingual Resources for CEF.AT in the Legal Domain

The goal of the MARCELL project is to assist in improving CEF.AT (automated translation) in the area of national legislation (laws, decrees, regulations, etc.) for the seven official languages of the European Union (Bulgarian, Croatian, Hungarian, Polish, Romanian, Slovak, Slovenian). The texts of national legislation are not automatically available to CEF.AT. The efficiency of existing machine learning systems and the quality of translations can be increased if these systems have access to national legislative texts. This is what the MARCELL project enables.

Within the framework of the MARCELL project, the available materials (collections of national legislation in the appropriate languages, the multilingual ontology-based thesaurus EUROVOC, and the IATE term collection) were used to create seven large monolingual national legal corpora, a comparable corpus in seven languages assigned to the highest-level domains by EUROVOC descriptors, and seven document processing software codes.