Finding sufficient online literature to train its machine learning models was a problem while supporting various languages in the past. Finding enough human community members who can assist Google in improving the models so they are on par with more frequently used languages is another challenge. But according to the firm, both fronts have made progress in recent years.
According to a business representative, "Google Translate learns from current translations found on the web and when languages don't have an abundance of web content, it's proven challenging for our algorithm to handle them successfully." However, "we've been able to add support for these languages thanks to recent advancements in our machine learning technology, and active cooperation from our Google Translate Community members."
The update will start rolling out to 1 percent of Google Translate users starting today, the company says, with a steady ramping up to the full Translate user base across Android and iOS in the coming days. Translate will support both text translation and website translation for the five new languages and virtual keyboard input for three of them - Kinyarwanda, Tatar, and Uyghur.
The fact that Google supports the Uyghur language, which is spoken by 12 million people, the majority of whom reside in the autonomous Xinjiang region in northwest China, is of great geopolitical significance. Under the guise of national security, the Chinese government has been waging a campaign against the Turkic Muslim minority group living there that involves widespread surveillance, travel and work restrictions, and detention in Communist Party "re-education camps," raising concerns about potential human rights abuses abroad.
Comments
Post a Comment