Program/One World, One Wiki!
- Format
Poster
- Keywords
- LanguageConverter
- machine translation
- native script editing
- non-latin scripts
- multilingual UI
- Cyrillic
- Chinese
- When and where
- Friday 20 July, 20:00 - 21:30
- Plenary hall
-->
- Presenter(s)
- Topic area
- Relationship to theme
Language & literacy
Access & accessibility
Back to Program
Edit this Session
Abstract
Instead of today's many siloed wikis, separated by language and project, we aspire to re-establish a unified community of collaborators in the spirit of ubuntu. We will still respect language and cultural differences — there will still be English, German, Hebrew, Arabic, etc. Wikipedias; they will disagree at times — but instead of separate domains, we propose a single user experience with integrated navigation between projects and languages and the possibility of split screen views aligning related content. On a single page we can work on articles in different languages, or simultaneously edit textbook content and encyclopedia articles. Via machine translation we can facilitate conversations and collaborations spanning languages and projects, without forcing a single culture or perspective. The improvements made to the project by each community can be shared with all others to improve the sum of all knowledge.
This talk will describe a number of concrete projects in pursuit of this goal. We will describe current work on LanguageConverter to allow communities to edit together without being separated by different ways of writing their language, using “selective serialization” technology to allow editing in a user’s native script without corrupting other authors’ texts.
A prototype translation suggestion tool suggests an edit in one language whenever an edit is made to a parallel text in another language. In addition to manually-created correspondences, our translation engine can automatically search for potential new correspondences or prune old entries when the translation drifts.
We will discuss the use of "zero-shot translation" machine translation models to allow our training data from "big" wikis to improve the translation of "small" wikis. Every contributed correspondence or translation further improves the ability of our tools to make additional articles from other languages available.
By bridging languages, our individual contributions can better fill knowledge gaps everywhere.
Outcomes
Increase awareness and use of LanguageConverter.
Describe a vision for a multilingual UI which can inspire related projects by the attendees.
Encourage discussion of the potential of machine translation for our projects, and encourage thinking of our translation correspondences as a valuable output of our project.
Signup
- <add your username here if you are interested in attending>