Google, via an official announcement today, is debuting its highly scrupulous ‘Neural Machine Translation’ tech to Indian languages. What that effectively means, is that with the help of some extremely intricate machine learning algorithms, Google Translate will now provide at least 55%-85% better translation results, for 9 Indian languages.
In layman terms, neural translation offers translates full sentences at a time, instead of pieces of a sentence. This change improves the quality of translation in a single jump than seen in the last ten years combined.
This will give you a better idea of what has changed :
GNMT (Google Neural Machine Translation) was announced by the company last year, via a research paper. With GNMT, translation errors were reduced by more than 55%-85% on several major language pairs measured on sampled sentences from Wikipedia and news websites with the help of bilingual human raters. Using human-rated side-by-side comparison as a metric, the GNMT system produces translations that are vastly improved compared to the previous phrase-based production system.
While GNMT had rolled out to major languages like Chinese, Spanish and others already, Indian languages on Google Translate were still translated to English using the old, less accurate Phrase-based translations. That will now change. Starting today, Google Translate will use GNMT to translate between English and nine widely used Indian languages — Hindi, Bengali, Marathi, Tamil, Telugu, Gujarati, Punjabi, Malayalam and Kannada.
So why is this rollout crucial for Google ? Well, India today has 234 million Indian Language users who’re online, compared to 175 Million english web users. Now, a joint report by Google and KPMG expects another 300 million Indian language users to come online in the next four years. With today’s launches (GNMT and Gboard for 22 Indian tongues), Google is making sure it isn’t left behind in the linguistic domain, an area where a lot of local competitors have come up and have offered decent solutions to cater to this humongous non English speaking populace.