In our ongoing effort to empower language communities to preserve their languages and cultures, we are excited to introduce Yucatec Maya and Querétaro Otomi to Microsoft Translator’s ever-growing list of supported languages. These language systems were developed in collaboration with community partners in Mexico, who created the automatic translations systems to permanently bridge the translation gap between these endangered languages and the rest of the world. The systems themselves were built using the Microsoft Translator Hub, a Translator product which is available for free to allow any group to create its own unique translation systems.
Using the Hub, our community partners took important steps to preserve their language and culture. The Yucatec Maya translation system was built by the Universidad Intercultural Maya de Quintana Roo (UIMQROO), a university in the southwestern Mexican state of Quintana Roo that was created to provide higher education to the Maya population of the region. Native to the Yucatan region of Mexico and Belize, Yucatec Maya is spoken by fewer than 800,000 people, with less than 59,000 monolingual speakers. The language is descended from the language of the ancient Mayan empire, which is well-known for its art, architecture, as well as its expertise in astronomy. The Querétaro Otomi language system was created by the Instituto Queretano de la Cultura y las Artes (IQCA), an institute in western central Mexico whose mission is to encourage artistic and cultural development and to promote equity and equality of opportunity within the State of Querétaro. Querétaro Otomi is an endangered language from the region that is only spoken by 33,000 people and has fewer than 2,000 monolingual speakers. The release of Maya and Otomi helps to celebrate the UNESCO’s International Mother Language Day, an annual international event which aims “to promote the preservation and protection of all languages used by peoples of the world.” According to UNESCO, “if nothing is done, half of 6,000-plus languages spoken today will disappear by the end of this century.” Maya and Otomi are indigenous languages from Mexico which are both currently threatened. Although they are still in use, the number of speakers is decreasing and younger people are not speaking them as actively as their elders. The new automatic translation systems will help the Maya and Otomi people safeguard their language and culture for generations to come. Over the years, Translator has worked closely with a variety of language community partners to encourage language preservation and, through it, intercultural communication. In the past, these community partners have used the Hub to create translation systems for languages such as Hmong Daw, Welsh, and Urdu. The Hub allows organizations such as UIMQROO and IQCA to leverage the computing power of Microsoft Translator’s machine-learning back end as well as its existing translation models to create unique and customized translation systems. The Translator Hub is a powerful tool for organizations that have specific translation needs, such as language preservation. It also allows organizations to create domain-specific systems, including industry-specific translation systems (for instance, for the medical or financial sectors) and business-specific systems that are customized to the company’s internal style and terminology. In addition to the Hub, Translator also supports a wide variety of products to connect individuals across language barriers, including the Translator API, which can be used to translate web pages and apps in real time into 45+ languages, as well as powering the translation features in the Microsoft Office suite of products. Most recently, Microsoft Translator and Skype introduced Skype Translator, a next–generation speech-to-speech translation platform which allows users to converse in different languages in near-real time. To learn more about International Mother Language Day, and what Microsoft is doing to support technology on this front, please visit the Official Microsoft Blog. Learn More about the Translator Hub and Language Preservation:
Over the next few weeks, we'll be showcasing how various teams around Microsoft have been able to use Translator to improve their internal and external operations in areas such as readiness, communications, customer support, forums and user groups, and web localization. Translator has proved to be a valuable tool for many teams across Microsoft, and we're happy to be able to share their stories.
How do you make libraries of billions of pages of scientific research, published in multiple languages, accessible to people around the globe? This was the problem faced by the WorldWideScience Alliance, a multinational partnership whose mission is to eliminate barriers in finding and sharing research across national boundaries. To solve this problem, WorldWideScience reached out to Deep Web Technologies which specializes in multilingual, federated search solutions across multiple industries. The result of this partnership was the WorldWideScience.org web portal that can search 100 different databases across 70 different countries, and then rank and translate the results into the user's preferred language. Millions of scientific articles are published around the world each year in wide variety of languages, but only some of them are available on the worldwide web using conventional search engines. Scientific publications are typically found in what is called the "Deep Web", which consists of documents, images, and records located in an often unconnected series databases throughout the globe. Using Microsoft Translator, Deep Web Technologies was able to create WorldWideScience.org — a consolidated portal, accessible through the web, which is able to search these worldwide databases and translate the results into 10 different languages. Users can choose to view scientific papers, multimedia, or research data on their desktop computers or mobile devices. The Microsoft Translator API provided Deep Web Technologies with the high level of scalability and reliability required for the project, and the translation API was easily integrated with the wide variety of data sources WorldWideScience.org pulls from. The completed portal improves global access to scientific research, encourages international collaboration, and provides new opportunities to share data. WorldWideScience.org now increases access to scientific and technological research worldwide, facilitates international collaboration, and provides new opportunities for research in multiple industries. The site handles approximately 70,000 queries and 1 million page views each month, and all traffic, including that from automated crawlers and search engines, amounts to approximately 70 million transactions per year. The multilingual, federated search solution implemented by WorldWideScience Alliance and Deep Web Technologies is applicable across a wide range of industries, and could be used for solutions ranging from customer support to organizational readiness. To learn more about WorldWideScience.org and multilingual, federated search technology, read the full case study.
Like any project management workflow, managing your organization's translation and localization is a constant balancing act between speed, quality and price. In a recent webinar, "Translation Trends 2015" hosted by MemSource, Microsoft Translator's Group Program Manager Chris Wendt showed how improvements in collaboration technology for translation could help raise the bar for all three of these elements. The primary choice faced by businesses when deciding to translate their content is whether to use human or machine translation to accomplish the task. To date, human translation has been able to provide high quality translation, but at a slower speed and higher cost than machine translation. In contrast, machine translation is instantaneous and inexpensive, but can be less accurate than human translation. Many organizations have had great success using machine translation with human translation integrated into their post-editing workflows — it has been shown to lead to productivity increases of up to 25%. Integrating human translation into post-publishing workflows using the latest collaborative translation memory software can have an even greater impact. Post-publishing translation allows website owners to leverage their community to refine the output of machine and human translation. This community includes subject matter experts, enthusiasts, employees, and other professional translators. In a recent research study at the University of Illinois at Urbana-Champaign, it was shown that the quality of machine translation, when interpreted by a subject matter expert, is of higher quality than human translation when that translator is not an expert in the field. Using a post-publish, post-edit workflow, organizations can raise the bar in speed, quality, and price— translation is done faster than by human translation, is of higher quality than machine translation alone, and decreases the cost of dedicated human translation services. For your post-publish, post-edit workflow to be successful, your organization needs to have several elements in place. The first is a machine translation API such as Microsoft Translator. This provides the initial translation used for your content. The second is a collaborative translation framework or translation memory system. This will allow you to coordinate your body of contributors to the translation project. Lastly, you will need to provide training for using these assets— making sure to include subject matter experts as well as translators. To learn more about post-publish, post-edit translation, and to see presentations from Microsoft Translator's Chris Wendt, MemSource CEO David Canek, Torben Dahl Jensen from TextMinded, and Moravia's Jan Hofmeister click on the link below. View the full webinar