ABSTRACT VIEW
MITIGATING LANGUAGE BARRIERS IN EDUCATION: DEVELOPING MULTILINGUAL DIGITAL LEARNING MATERIALS WITH MACHINE TRANSLATION
L. Poláková, M. Popel, V. Kloudová, M. Novák, M. Anisimova, J. Balhar
Charles University (CZECH REPUBLIC)
In this paper, we introduce an applied research project that brings together the fields of digital education, linguistics, translation studies, and machine translation. It focuses on the development and dissemination of multilingual digital learning materials for primary and secondary school students. The main goal of the project is to use machine translation methods to help mitigate language (and social) barriers faced by non-Czech-speaking, primarily Ukrainian, children in the Czech Republic’s education system. Due to recent political events, the Czech Republic is currently hosting a generation of Ukrainian children whose educational success largely depends on their integration into both society and the school system. They need access to high-quality education in their native language, knowledge of the Czech language, and access to subject-based education in Czech.

Within the project, which started thanks to the cooperation of a leading Czech academic institution and the largest Czech publisher of textbooks and educational materials, we develop multilingual interactive content for the company’s educational web portal. The web portal contains more than 8,000 interactive and, to a large extent, multimodal exercises, originally in Czech (drag-and-drop exercises, yes/no and multiple-choice quizzes, fill-in-the-blank tasks, matching pairs, crosswords, mind maps, dictation exercises, etc.). In the initial phase of the project, we focused on developing a machine translation system between Czech and Ukrainian and adapting it to the educational domain. In subsequent phases, we expanded support to English and German, so that the web portal's content is available in four languages. The resulting applications are freely available to educators and researchers.

In the full version of the paper, we will discuss: a) the results of an initial survey conducted among Czech teachers to assess the specific needs of non-Czech-speaking children in Czech schools; b) the development and evaluation of a direct (not via pivot) machine translation system between Czech and Ukrainian (with special attention to the ability to translate formatted input such as XML and PDF, as well as addressing the challenges of translating technical and scientific terms); and c) the implementation of the translation on the web portal.

Keywords: Multilingual education, digital learning materials, machine translation, Czech, Ukrainian.

Event: EDULEARN25
Session: Emerging Technologies in Education
Session time: Tuesday, 1st of July from 08:30 to 13:45
Session type: POSTER