The organizations will develop scalable language technology to address ongoing humanitarian crises
DANBURY, Conn., USA – 7 August, 2019. Translators without Borders (TWB), with the support of Microsoft Philanthropies, is working to develop innovative language-focused solutions that can be used to help the over 177 million people who live in areas affected by humanitarian crises. These solutions can improve both the effectiveness and accountability of humanitarian action.
Microsoft’s foundational funding, along with support from the Cisco Foundation, will allow TWB to grow Gamayun, the language equality initiative. This funding will be used to develop a replicable and scalable machine translation model for low-resource languages. Because these languages are not currently well-served by language technology or services, the crisis-affected people who speak them struggle to access critical information or voice their needs, in their own language.
Machine translation has the power to bridge this gap, revolutionizing humanitarian communication. Because a massive amount of data is needed to build a successful machine translation engine, TWB will first use the new funding to develop voice and text datasets for two languages: Kanuri and Rohingya.
“People living through unimaginable crises often cannot access information or communicate their needs, simply because little language data and technology exist in their language,” explains Jane Meseck, Microsoft Philanthropies’ Senior Director, Global Programs & Partnerships. “We’re excited to help change this by supporting the development of datasets and machine translation tools through Gamayun.”
In northeast Nigeria, humanitarians are working with 1.9 million people displaced by violence and poverty. Although Kanuri is the most widely spoken language in the response, only about half of speakers report receiving information from responders in the language. In Bangladesh, Rohingya is the only language that all Rohingya people speak and understand. Because it is a predominantly oral language without a standardized script, humanitarians have struggled to communicate in a way that is consistent and widely understood.
With both Rohingya and Kanuri, establishing consistent, culturally-relevant, and humanitarian-specific voice- and text-based data and technology will help responders and crisis-affected people communicate.
TWB will then establish a replicable and scalable machine translation engine for at least one of those languages over the next year. The organization is also working with humanitarian partners to prioritize new languages and use cases.
Additionally, TWB will grow language community teams, establish an advisory panel, and advocate for the need for and use of equitable language technology globally.
This work follows a successful pilot project organized through the Hague’s Humanitarian Action Challenge. In late 2018, TWB worked with the World Food Programme (WFP) and Mercy Corps to develop datasets and build a machine translation prototype for Levantine Arabic, a language spoken in Syria and neighboring countries.
Gamayun is developing machine translation for marginalized languages, so that everyone can access the information they need, in a language they understand. These technologies also give people a voice in their own futures — allowing them to share their needs, concerns, and ideas in their own language. Not only does this increase humanitarian accountability and effectiveness, it can also improve global two-way communication in all sectors.
About Translators without Borders
Translators without Borders (TWB) envisions a world where knowledge knows no language barriers. The US-based non-profit provides people access to vital knowledge in their language by connecting non-profit organizations with a community of language professionals, building local language solutions and raising awareness of the power of language. Originally founded in 1993 in France (as Traducteurs sans Frontières), TWB translates millions of words of lifesaving and life-changing information a year.
For media inquiries, contact [email protected]