Whilst there is still no substitute for real translators, translation is getting more and more advanced when it comes to software. Google, who have been innovators in a wide array of arenas, have been working on a unique translation app that would give two people a chance to talk in two languages and understand each other completely.

News of a patent application has been discovered, which abstracts how two keyboards can operate at either end of a device, providing mutually exclusive translation input.

The text input would then use Google Translate to relay text in a respective user’ language. This would give them the chance to communicate in real time without the traditional language barriers.

How it would work

The system would include a communications module, user interface module and display units. These would probably be tablet-sized to accommodate the amount of text required. In theory, it would work as thus:

  1.  Comms module receives Language A communication and requests translated communication in Language B that corresponds to Language A in real time.
  2.  Comms module receives a Language B communication and requests second translated communications in Language A that corresponds to Language B.
  3.  The user interface creates first output corresponding to Language A communication and second output relating to Language B communications.
  4.  A display would show both Language A and Language B side-by-side

Whilst it sounds quite complicated, the system is just the first step to a more accurate automated translation service.

Google have already made great steps in translation with the aforementioned Google Translate algorithm. Translation software such as this works by detecting patterns in documents that have already been translated by humans. The software then makes an educated guess to what the translation should be, providing a high (but not 100%) level of accuracy.

This is known as ‘statistical machine translation’. Whilst this technology has made significant advances, it is still prone to mistakes in the translation process.

Google have already made great steps in translation with the aforementioned Google Translate algorithm. Translation software such as this works by detecting patterns in documents that have already been translated by humans. The software then makes an educated guess to what the translation should be, providing a high (but not 100%) level of accuracy.

This is known as ‘statistical machine translation’. Whilst this technology has made significant advances, it is still prone to mistakes in the translation process. Howeve,  many non-European languages are still translated inaccurately with this software.

Real-time translation is something that only skilled human translators are able to do currently, and it is believed that it would take a lot more work before Google is able to improve and better it’s translation output when it comes to real-time situations.