January 8, 2013 – New developments support improvements in community-based content translation
The ACCEPT (Automated Community Content Editing PorTal) Project, a research project aimed at investigating and developing new language technologies and user capabilities for users of forum and community-based content, has announced its first results from 12 months of testing. The project, which spans 36 months, is designed specifically to improve the quality of user-generated content within community forums to make it easier to translate, and to improve machine translation (MT) engines.
The project is supported by a 1.8 million euro grant from the European Union.
The first results cover four main development areas:
- An integrated pre-editing plug-in for forum users, designed to encourage participants to address critical quality issues as they post content in the community.
- Pre-editing strategies for statistical machine translation (SMT) engines.
- New developments concerning the SMT engine MOSES, to determine which data combinations lead to the best results for user-generated content.
- The Evaluation API, that provides a framework to collect user feedback from online content repositories, such as forums.
The ACCEPT Consortium, which administers the project, has also begun planning for a special interest group of potential users from outside the project in order to get feedback. The first members of the group will have access to the project in the first quarter of 2013.
The ACCEPT Project “works on enabling machine translation for the emerging community content paradigm, allowing citizens across the EU better access to communities in both commercial and non-profit environments.”
For this project, ACCEPT joins forces with the machine translation engine experts at Lexcelera/LexWorks and Symantec as well as the Moses inventor Philipp Koehn and his team from the University of Edinburgh. The other consortium partners are Acrolinx, a major content optimization software provider, the university of Geneva and Translators Without Borders. “Machine translation is here to stay,” commented Lori Thicke, Founder of Lexcelera/LexWorks. “This project is important because it’s helping us build the technology to make MT even more useful to more global communities.”
As reported by TechWhirl.