When Google launched Google Translate 10 years ago, the key algorithm behind the service was Phrase-Based Machine Translation. The translations provided by the service have since vastly improved due to the developments in machine intelligence, but the recent addition of neural networks has provided Google Translate with the biggest boost that it has ever received.
Language is naturally phrase-based, which is why translating between languages is not as simple as plugging in the translation of words in sentences. While computers have been developed to handle phrase-based translation, there are still nuances in languages that the machines are not able to understand.
Google has now deployed the Google Neural Machine Translation system, which utilizes machine learning and neural networks to provide a massive boost in translation accuracy. The GNMT does this by taking in a sentence to be translated as a whole while also considering the phrases within, as opposed to treating each phrase or word as stand-alone components.
Neutral networks, which are modelled after the neurons that are found in the human brain, come from the same artificial intelligence technology that powers the systems that are able to identify the spoken commands of users and identify the people or places shown in an image.
For Google Translate, the deep neural network used by Google is called an LSTM, which means long short-term memory. Such a neural network is able to retain information in both the short term and the long term, similar to human memory. As the system analyzes a sentence, it is able to remember all its parts, allowing it to consider the sentence as a whole instead of having to treat it as made up of different phrases and words.
According to Google, in certain languages, the usage of the GNMT system will reduce errors by a massive 60 percent. However, for now, the system is only being used for translations from Chinese to English, which is perhaps one of the best translation pairs to showcase the system's effectiveness.
Google, however, admitted that the GNMT still needs further improvement, as the system could still make significant mistakes that human translators will not commit, including mistranslating proper nouns and translating sentences without considering the provided context of the page or paragraph.
However, as the neural networks behind Google Translate continue to learn, the service could soon be the most effective way of breaking down language barriers. In addition to the recently added functionality that makes it easier for users to make translations and even while offline, Google Translate is being geared as the primary tool for adventures in foreign lands.