Microsoft Translator has just launched Levantine Arabic as a new speech translation language, which is an Arabic dialect spoken in various countries including Jordan, Lebanon and Syria. It is aimed at assisting educators, businesses, and other non-profit organizations in communication with the Levantine speakers beyond the language barrier. It is expected to make Skype calls, presentations and general meetings much clearer and convenient.
According to Microsoft, Levantine is their 11th speech language and is the spoken dialect of more than 32 million Arabic native speakers. The complication in building this speech translation program was that since it is a spoken language that is rarely written, it did not have the vast amounts of parallel data available for training a usable machine translation system. Without the required amount of data available for training the neutral machine translation system, the system would not be able to procure translations which are good enough to be used in real-life. However, Microsoft researchers have been successful in developing a novel approach that utilizes monolingual data for training a system for any spoken language. The team was thus able to create a workable Levantine to English translation system even though there was not enough available parallel data. Hany-Hassan Awadalla, Microsoft’s principal research scientist commented on this, “We adapted a system trained on standard Arabic-to-English translation to be used on a spoken Arabic dialect (Levantine) using only monolingual data of the spoken dialect. We developed an approach to generate synthetic parallel data from monolingual data.”
Levantine Arabic is now available throughout the Translation apps, Skype translation feature in Windows 10, Presentation Translator for Powerpoint and in the unified Speech translation. According to this service, developers will also be able to customize speech transcriptions, text-to-speech, translations, before integrating them into their workflows, apps and websites.
Next on Microsoft’s agenda is partnering with No Lost Generation Tech Task Force, led by Netforce and Norwegian Refugee Council for co-creating an Al-powered solution which links youth affected by Iraqi and Syrian conflicts through educational resources. The goal of this initiative is to enable the conflict-affected youth to discover and access learning resources anytime and anywhere possible.