Despite the remarkable technological strides being taken in the translation industry, the translators themselves aren’t too concerned for their jobs. The industry is booming, with the number of people employed doubling in the last seven years, and that figure set to grow by 29 percent in the coming years. But that doesn’t mean the role of the industry isn’t changing. More and more people are being employed as post-editors to correct flawed translations made by machines, and software is increasingly being incorporated into the role of a translator to economize the process.
In V1’s interview with professional translator, Lyse Leroy, she suggested that if it were even possible for machines to mimic the linguistic nuances required for complex, sensitive texts, it’s a long way off in the future.
That doesn’t mean that companies aren’t trying, though. One of the key industry terms that’s come up this year has been neural machine translation (NMT). The technology – which was launched in 2016, and is being incorporated in Google, Yandex and Microsoft translation services – is a huge improvement on previous machine translation (MT) models, with studies showing that it produces 60 percent fewer errors and can even match up to human translators in some tests.
NMT is a leap forward from the formerly favored phrase-based machine translation (PBMT). Rather than breaking a sentence down into individual words and phrases, NMT considers the entire sentence input for translation. According to Google, “The advantage of this approach is that it requires fewer engineering design choices than previous Phrase-Based translation systems.”
However, this latest advance might not be the breakthrough that tech-enthusiasts have been waiting for. According to Dr. John Tinsley – CEO of Iconic translation firm and expert in MT – NMT is nothing too special and is “not going to replace human translators.” For now, it seems that the role of the translator still lies in the realm of the human, and is set to stay that way for the foreseeable future.
