INTEGRATING MACHINE TRANSLATION INTO MOOCS
Dublin City University (IRELAND)
About this paper:
Conference name: 9th International Conference on Education and New Learning Technologies
Dates: 3-5 July, 2017
Location: Barcelona, Spain
Abstract:Massive Open Online Courses (MOOCs) offer valuable learning opportunities in several disciplines to many students, regardless of their background, location, and personal circumstances. MOOCs are typically available in the local language that is shared between tutors and students, which has the added bonus of encouraging interactions on social platforms and fora accompanying formal instruction. However, language barriers impede broad use of high-quality MOOC materials across national and language boundaries: English is often chosen as the common language of MOOCs with international users, even though this prevents large groups of users from fully engaging in a fulfilling MOOC experience, thus wasting precious learning opportunities for innumerable potential students around the world.
This presentation introduces TraMOOC, a European research project involving a consortium of academic partners and companies, that attempts to tackle the issue of empowering international learners in a digital multilingual world by providing reliable machine translation (MT) specifically tailored to MOOCs from English into 11 languages, i.e. Bulgarian, Chinese, Croatian, Czech, Dutch, German, Greek, Italian, Polish, Portuguese, and Russian. In an increasingly globalised and mobile society, in which academic institutions are under growing pressure to become internationalised, there is a strong need for high-quality digital teaching and learning resources to be distributed across linguistic and cultural boundaries. Accordingly, there are obvious benefits in disseminating MOOC content in multiple languages, so the research conducted in the TraMOOC project can be of great interest to a wide range of professionals in the educational sector, both in Europe and beyond. We describe how the project is addressing the challenges involved in developing an innovative, high-quality MT service for producing accurate translations of heterogeneous multi-genre MOOC materials originally available only in English, encompassing subtitles of video lectures, assignments, tutorials, and social web text posted on student blogs and fora. We explain why we preferred the state-of-the-art approach to MT system development based on neural networks over the competing well-established statistical framework, and provide evidence for its superior performance on MOOC data, focusing on translations from English into German, Greek, Portuguese and Russian for materials drawn from business, medical, physics, and social science MOOCs. We report the main results of a large-scale evaluation that was conducted as part of the TraMOOC project, involving both leading automatic metrics to assess MT output quality as well as experiments with professional translators, who were asked to compare the quality of different systems, rate the adequacy and fluency of MT output, manually tag errors according to a taxonomy, and judge how much effort was required to correct the mistakes found in the MT output.
The conclusion discusses the key lessons learned in the ongoing TraMOOC project that are relevant to the wider community of international professionals with an interest in the multilingual aspects of innovative education and new learning technologies, assessing the pros and cons of integrating state-of-the-art MT into MOOC platforms. In particular, we share reflections on how best to integrate MT into MOOCs, to build multilingual communities in which students can truly learn and interact across language barriers.
Keywords: MOOC, machine translation, translation, languages, e-learning.