The countless mistranslations encountered as a result of machine translation have left many struggling to understand the benefits of such systems, even as their popularity grows among the general population.
However, recent advancements in machine translation technology suggest such systems could soon be delivering a higher standard of results that may prompt people to take the technology a little more seriously.
At Bubbles, we like to keep on top of the latest translation technology and take a balanced and pragmatic view of the role of machine translation.
(Spoiler alert – We’re impressed with the technology, but an algorithm understanding a sentence doesn’t mean it can understand the nuance of context within content like professional translators!)
An Introduction to Parsey McParseface
Riding high on the back of the Boaty McBoatface ‘scandal’, Google’s new Parsey McParseface system promises to be the next key step in the evolution of machine translation. Until now, machine translation has been limited to working out the different types of words within a sentence, and how their structural interrelationships alter the sentence. This approach works in theory, but it often fails in practice, leading to some clunky mistranslations.
Google believes Parsey McParseface could change all this and has already made the bold statement that it understands 94 per cent of standard sentences. The system is built on algorithms that are capable of learning to analyse the linguistic structure of language, and explaining the functional role of each word in a sentence. If further advancements increase the percentage of sentences understood by the machine, this new level of translation accuracy could help to reduce errors in machine translation.
The technology behind Google’s latest creation is already finding its way into new translation machines, one of the most significant being the Pilot earbuds. The world’s first language-translating earpiece, Pilot works in pairs to provide live language translation via a phone app.
Simply put, it can translate your words into your listener’s language, and their words back into your language. Although not technically the first system to use this type of technology, with the Skype Translator another similar option, the development of Pilot certainly marks a new, more interactive era for machine translation.
But Are We Really There Yet?
Although the new technology associated with Parsey McParseface could prove to be a big leap forward for machine translation, other aspects of language and communication remain a mystery to machines and most professional linguists and translators still think we’re a long way off truly accurate automated translation.
Local complexities associated with specific regions can be easily missed by machine translators, while cultural issues affecting the significance of certain words or phrases are also likely to be ignored, with potentially significant consequences.
When it comes to translation, Bubbles believes you can’t beat the real thing. Our translators combine an in depth knowledge of their chosen language with an expertly honed gift for understanding language itself. Their natural comprehension ensures the most accurate and appropriate translations possible.
Our human translators are also capable of altering messages that fail to translate directly, creating new words and phrases that convey the true values and ideas behind the original content.