Trexy solution

Trexy is a cloud-based API that enables you to translate your content. It is already integrated into MOOC platforms and video repositories to create new educational experiences. It also offers tools for developers and researchers. Trexy API is the first free and open translation service to use Neural Machine Translation Models (NMT) for educational content, which immensely improves translation results. NMT has recently emerged as a disruptive technology, and has become the dominant paradigm in machine translation.

Discover how you can use, integrate, and customize this service in your own application or MOOC solution. Trexy API can be used on its own or its results can be customized pre-publishing with our partners.

Trexy API provides translation from English into nine European and two BRIC, languages, namely German (DE), Italian (IT), Portuguese (PT), Dutch (DU), Bulgarian (BG), Greek (EL), Polish (PL), Czech (CS), Croatian (CR), Russian (RU) and Chinese (ZH).

The problem Trexy solves

Business solution: Education providers, MOOC providers and Video repositories need to offer high-quality, integrated multilingual educational services to their customers. Trexy provides this service for all types of textual data that is language independent in nature and addresses all types of text genre.

Market solution: The European Digital Single Market is still fragmented by language barriers that hamper a wide penetration of cross-border commerce, social communication and exchange of cultural content. Current machine translation solutions typically perform well only for a limited number of target languages, and for a given text type. Therefore, Trexy provides a unique solution to tackle scalable, sustainable translation for the education market.

Back to Trexy here

The project behind Trexy

In 2015 eleven European Institutions joined forces to solve the problem of mass translation in education. They designed a European project called TraMOOC which provides a system and solution that is intended to overcome the language barriers in MOOCs, and is expected to deal with huge volumes, high variety of languages and text styles, and deliver results in reasonable time. Special attention was on selecting 9 EU and 2 BRIC languages that have “fragmentary” or “weak/no” machine translation support according to the META-net language white papers.

Launched on 1 February 2015 by a consortium made up of six European universities Humboldt Universität zu Berlin (UBER), Dublin City University (DCU), University of Edinburgh (UEDIN), Ionian University (IURC), Stichting Katholieke Universiteit (Radboud University), Stichting Katholieke Universiteit Brabant (Tilburg University), two industrial partners Deluxe Media Europe (DME), EASN-Technology Innovation Services (BVBA), and a non-for-profit Knowledge 4 All Foundation Ltd (K4A), and it will run until 31 January 2018.

The science behind Trexy

Machine translation: The high quality of translations is achieved through a hybrid translation schema that combined automatic processes (i.e. text processing tools and linguistic resources that are adapted to cope with multi-genre MOOC educational text) as well as limited, focused human intervention, i.e. crowdsourcing in a novel time- and cost-efficient setup for evaluating the generated translation.

Implicit and Explicit evaluation: We established a new, standardized, multi-level evaluation schema for determining the value of the produced machine translation. Taking into account the idiosyncrasies of the particular text type and its multi-genre character, the evaluation schema consisted of a human and an automatic aspect, as well as a traditional explicit (direct) and an innovative implicit mode, that will be facilitated via a separate text mining and topic identification.

Crowdsourcing: We employed crowdsourcing activities for infrastructure enhancement and translation evaluation where human intervention was required. The choice of crowdsourcing for handling these tasks is supported due to its direct link with Open Innovation, with crowd wisdom and co-creation.

Start with Trexy

Below are some of the key areas where developers can engage with the platform. For additional information get in contact.

Trexy’s tools for developers and researchers

This toolkit is a repository of tools and solutions created by the TraMOOC project with appropriate license models (development license, research license and an open license for testing purposes), and software packages for the development of state-of-the-art systems for language translation of educational text.

The algorithms, methodologies, processes and tools are the result of the TraMOOC project. The translation model training processes, the resource bootstrapping processes, the text mining methodologies form a library of algorithms that will prove invaluable to the scientific community.

Benefit from Trexy’s tools

Developers, Businesses, Educators can use Trexy API to play with language pairs and use business analysis to understand where is the value of having their content translated into other languages.

Machine Translation developers can use Trexy as a unified workbench for integrating MT solutions and a platform for promoting, testing and comparing their solutions.

Language Technology Engineers and community can get access to accurate and wide-coverage linguistic infrastructure for less widely spoken languages. This infrastructure includes exploitation, enrichment, and adaptation policy on bootstrapping, cross-language, and interdisciplinary approaches to create new or enrich/adapt available tools and resources.

Serving learners identity in MOOC courses

MOOC platform use case / openHPI

openHPI, the educational Internet platform of the German Hasso Plattner Institute, Potsdam runs MOOCs – Massive Open Online Courses – which are one of the most innovative forms of learning. Available free via the Internet, learners can access didactic course materials and learn in close contact with other students by means of social media.

Quick facts:

  • Platform with learning videos, interactive self-tests, tutorials, practical exercises, etc.
  • Part of a network of platforms
  • Runs openSAP: MOOC platform of SAP with1.500.000 enrollments
  • Runs openWHO: MOOC platform of the World Health Organization
  • Successfully registered 133,000 registered
  • So far 340,000 enrollments in 30 courses

Challenge:

In order to enable the translation of video content, an automated translation tool is required. Textual content consists of reading material, quizzes and assignments. User contributed content mainly consists of forum posts, allowing multi-lingual discussions would be a nice-to-have option. All translations currently done manually with some tool support for subtitling. The platform’s video player supports the display of transcripts and subtitles in any number of languages. The courses are in English and German and would benefit the extension into others.

Providing audiences with multilingual learning experiences

Video platform use case / VideoLectures.Net

The state-of-the-art in business, technology, industry and scientific breakthroughs are disseminated through a rich network of numerous yearly global conferences. However, only a tiny fraction of the interested scientific community has access to this knowledge, which exacerbates both an intellectual and development gap among those talented individuals with and without access. VideoLectres.Net fills this gap with its service.

Quick facts:

  • Portfolio of 23,000 peer-reviewed lectures
  • Collected from over 1000 events
  • Presented by over 14,000 authors
  • Attention of over 5000 unique visitors daily from academia and industry
  • Recipient of United Nations World Summit Award
  • Currently provides only EN to Slovene, Spanish, Catalan, French, Italian
  • Provides 16870 transcriptions in EN waiting to be translated

Challenge:

VideoLectures.Net is making up-to-date knowledge available to everyone, regardless of their ability to attend expensive scientific events from all disciplines. This includes video recording the top scientific events world-wide, translating the content into major world languages (through automatic subtitling), and making the content available freely to anyone with an internet connection. The challenge is how to serve the global education and learning community by using open and state-of-the-art translation technologies.