Google Neural Machine Translation (GNMT) System
Google is considered as the leading contributor of artificial intelligence-assisted language translation. They launched a neural machine translation (GNMT) system called Google Neural Machine Translation on November 2016. This new system of translation can translate full sentences instead of just words.
It improves the quality of translation. Its network encoding the semantics of the sentence, rather than by memorizing phrase-to-phrase translations. Google published a research paper about it. GNMT contains a deep LSTM network with 8 encoder and 8 decoder layers using both attention as well as residual connections. This new system of translation did not create its own universal interlingua.
GNMT directly translates one language into another. GNMT uses an artificial neural network to increase both accuracy as well as fluency in Google Translate. This system is available via the Google Cloud Translation API. It supports new languages such as English, Arabic, Russian, Hebrew and Polish. This system can reduce errors in translation.
This most advanced machine translation tool is not as good as a human translator. It utilizes state-of-the-art training techniques. GNMT is used by Google Translate mobile and web apps for 100% of machine translations from Chinese to English. But this system drop words, translate sentences in isolation rather than considering the context of the paragraph or page and mistranslate proper names or rare terms.