|
The purpose of the project is to develop a system that will maintain consistent content in multi-language Wikis, while allowing the user to edit the page in the language they are most comfortable with. To achieve this, the system must identify novel information, translate it and insert it in the right place in the other language versions of the page. The first to use this technology will be the project partners Deutsche Welle, The Netherlands Institute for Sound and Vision, and WikiMedia.
HITS' part in the project consists of structural analysis of Wiki pages. This will consist of identifying the concepts of interest within pages, and linking them to a multi-language repository of concepts. These "anchors" will be used to find candidates for similar and novel text fragments from different language versions of a page, and for segmenting the content into coherent fragments. In a later stage, we will extract attributes of interest from pages -- similar to Wikipedia's infoboxes -- using the users' edits as an indication of what they find interesting and useful.
The project partners are:
|
|
