Content Translation: Computer-assisted translation tool for Wikipedia articles

Content Translation: Computer-assisted translation tool for Wikipedia articles Niklas Laxstr¨om University of Helsinki Dept. of Modern Languages and Wikimedia Foundation nlaxstrom@wikimedia.org Pau Giner Wikimedia Foundation pginer@wikimedia.org Santhosh Thottingal Wikimedia Foundation sthottingal@wikimedia.org Abstract The quality and quantity of articles in each Wikipedia language varies greatly. Trans- lating from another Wikipedia is a natural way to add more content, but the trans- lation process is not properly supported in the software used by Wikipedia. Past computer-assisted translation tools built for Wikipedia are not commonly used. We created a tool that adapts to the specific needs of an open community and to the kind of content in Wikipedia. Qualitative and quantitative data indicates that the new tool helps users translate articles easier and faster. 1 Introduction Wikipedia is the most multilingual encyclopedic knowledge archive, with over 280 languages with varying amount of content. Knowledge available for a user is limited by the languages used to ac- cess it. Translation has been a common way to expand knowledge across languages in Wikipedia. The editing activity of the top 46 language editions of Wikipedia shows that 25% of edits by multilin- gual users are for the same article in different lan- guages (Hale, 2013). It is not necessary to use any tool to trans- late Wikipedia articles. However, it is a com- plicated process and mainly done by experienced Wikipedia editors. There were many attempts to build tools to support translation of articles. None has seen widespread use: in our research only few users reported using those tools when translating Wikipedia articles. c⃝ 2015 The authors. This article is licensed under a Creative Commons 3.0 licence, attribution, CC BY. In this paper we present a new approach to sup- port translation which has been designed taking into account the unique needs of Wikipedia content and their community. Content Translation (CX) is a new tool that automates many steps of the trans- lation process and validates the approach in prac- tice. It was first enabled on 8 Wikipedias as an opt- in feature to create new articles in January 2015. Selected language pairs have machine translation (MT) support. 2 Previous work MediaWiki, the software powering Wikipedia, is translated to hundreds of languages using the Translate extension. No such solution was avail- able for translating Wikipedia articles, leaving a gap in the translation support. There were at least ten instances of translation tools built for Wikipedia1. Those tools can be di- vided into two groups based on whether the tool creators already possessed MT software. The first group is composed of companies such as Google and Microsoft, but also smaller companies and re- searchers. The other group of tools has been cre- ated just for Wikipedia article translation, mostly by volunteers. Among the earliest tools were GTT by Google and WikiBhasha by Microsoft, using their own MT services (Garc´ıa and Stevenson, 2009; Kumarana et al., 2011). Later, Casmacat for professionals and researchers (Alabau et al., 2013) and CoSyne for multilingual MediaWikis (Bronner et al., 2012), unlike Wikipedias …

Content truncated to 500 words