Meedan

Meedan is a non-profit social technology company which aims to increase cross-language interaction on the web, with particular emphasis on translation and aggregation services in Arabic and English. Through its use of Machine Translation (MT), Machine Augmented Translation (MAT), and distributed human translation, Meedan’s goal is to increase dialogue and exchange between Arabic and English speakers primarily by launching a cross-language forum for conversation and media sharing. This service will be designed to stand as “a digital gathering place for a linguistically, culturally, and geographically diverse community of Arabic and English speaking Internet users”.[1]

Background

Meedan was founded by Ed Bice in 2005 and incorporated as a nonprofit charitable organization in 2006. Bice, who was formerly Executive Director of The People's Opinion Project, remains CEO with a board comprising Jon Corshen, President of GoingOn Technologies, a social media company for onDemand software, Adnan Hassan, Chairman and CEO of Mecasa Advisors and a former senior adviser and member of the Board of Directors at the World Bank, and Hanan Heakal, Senior Human Resources Manager at Procter & Gamble in Cairo, Egypt.

Vision and Aims

Meedan – whose name means ‘gathering place’ or ‘town square’ in Arabic – sees “dialogue and collaboration” as the key to creating “understanding and tolerance” between the Arab and Western worlds. A fundamental premise of the organization is that social technology on the web can play a part in enabling citizen-to-citizen diplomacy between the peoples of Arab and Western countries, thereby helping to improve cross-cultural understanding. Meedan’s vision is thus to “innovate technologies that enable dialogues across linguistic, cultural, and ideological boundaries”.[2] The company describes itself as “multi-cultural, non-partisan, non-ideological, and supportive of diverse viewpoints”.[3]

Features

Dialogue platform

Meedan’s core product is a forum for cross-language conversation and media sharing in Arabic and English. Users can browse aggregated sources around world events – blogs and mainstream sources; opinion and reporting; Arabic and English writing – and help expand the news narrative by posting articles and comments themselves. All sources and comments are mirrored across Arabic and English using a combination of machine and human translation. Users can also make friends with other users and upload their profile to their My Meedan page.

Translation

Meedan’s translation technology has been developed in collaboration with The Watson Research Group at IBM with which it entered into a research partnership in 2006. Development has focused on Machine and Machine Augmented Translation tools that enable users to improve translations collaboratively, including IBM’s Transbrowser – a browser tool for creating a translation layer on the web. Perhaps most significantly, Meedan is open licensing the linguistic data generated from its efforts. Improved translation automation and augmentation has profound implications in global media, education, and business and information services; Meedan hopes to play a small part by leading an effort to steer the global computational linguistics community to embrace open licensing of linguistic corpora. Meedan aims to contribute to the Arabic English linguistic corpus, and therefore improve MT quality, by providing translation services to individuals and partner organizations. Meedan is especially interested in advancing progress in the under served direction of source English to target Arabic Machine Translation.

Challenges

Translation

Machine Translation is often considered to be unreliable in rendering correct “context-dependent meanings” as well as coherent and grammatically correct sentences (particularly when dealing with languages, such as Arabic and English, with very different sentence structures).[4] Meedan’s software, however, appears to be more than mere Machine Translation. With a large linguistic corpus and overseen by a staff of over 20 professional translators, Meedan’s translations are more likely to be accurate in both their grammar and meaning. Moreover, Meedan’s users can edit and improve translations themselves. These improved translations are then stored on the Machine Translation software to be used in the same context in the future. Thus, as more users participate in conversations and help broaden the machine translation’s ‘vocabulary’, so the potential errors of Machine Translation should be reduced.

The dependence on human translators to enhance the quality of the translation is not free of problems. Besides syntactic and semantic factors there are factors of style and aesthetics of the target language. While many so-called professional translators could produce comprehensible text of higher quality than machine translation, the quality of the text in the target language is difficult to assess. Awkward or unfluent texts, when admitted in the system, may establish norms and frame of reference that don't serve the purpose of enhancing the overall quality of translation, and an example for translators, or aspiring translators to imitate. This, however, is arguably outside of the scope of the project.

Steering Committee

Meedan's Steering Committee was established in October 2008 to serve as a high level advisory panel for the project. Members include: the Director of the Bibliotheca Alexandrina, Dr. Ismail Serageldin; Dr Sabri Saidam, minister of telecoms and IT for the Palestinian Authority serving in the ninth Palestinian cabinet; and prominent media personality Muna Abu Sulayman, Executive Director of the Alwaleed Bin Talal Foundation, the philanthropic arm of HRH Prince Alwaleed Bin Talal's Kingdom Holding Company.

Funding and Research Partnerships

Meedan’s funding partners include

Meedan has received more than $3.2 million in research and development support from IBM to further the development of its Arabic-English Automated Translation technology and to support Meedan’s ongoing work on “a social media sharing platform bridging the Arabic and English speaking communities”. Other technology partners include, dotSUB, GeoCommons, MakerLab, Carrot Search, and Rylander Design.[6]

In addition, Meedan has a number of social network partners:

References

  1. ↑ "Ideas", http://www.meedan.net
  2. ↑ "Vision", http://www.meedan.net
  3. ↑ "Mission", http://www.meedan.net
  4. ↑ Giles, Jim. "Learning to talk". http://www.nytimes.com/2008/12/14/magazine/14wwln-idealab-t.html?_r=4&ref=magazine. Retrieved on 30 January 2009.
  5. ↑ "Partners", http://www.meedan.net
  6. ↑ "Partners", http://www.meedan.net
This article is issued from Wikipedia. The text is licensed under Creative Commons - Attribution - Sharealike. Additional terms may apply for the media files.