{"id":6356,"date":"2022-11-15T08:02:26","date_gmt":"2022-11-15T07:02:26","guid":{"rendered":"https:\/\/www.textmaster.com\/blog\/?p=6356"},"modified":"2022-11-14T16:14:44","modified_gmt":"2022-11-14T15:14:44","slug":"future-translation","status":"publish","type":"post","link":"https:\/\/www.textmaster.com\/blog\/future-translation\/","title":{"rendered":"The future of translation: on and offline"},"content":{"rendered":"

Type “future of translation<\/strong>\u201d into a search engine and you’ll be presented with pessimistic articles questioning whether translators <\/strong>and interpreters<\/strong> will all be unemployed in a few years and whether machines<\/strong> will completely replace human translators. But what is the situation actually like? Is machine translation (MT) really putting thousands of jobs at risk?<\/p>\n

Here is what we know about the future of translation on and offline<\/strong>!<\/p>\n

What does the future of translation look like?<\/strong><\/h2>\n

In our digital times, businesses don’t have a minute to waste! So, they use new technologies<\/strong> to translate their company websites<\/strong><\/a>, newsletters<\/strong> and communications materials<\/strong>.<\/p>\n

If they didn’t, their competitors would have the edge.<\/p>\n

But it’s important to distinguish between two types of translation tools.<\/p>\n

First, you have translation tools with an integrated dictionary and rules. These are known as computer-assisted translation (CAT) tools<\/strong> and, just like humans, they have a translation memory<\/strong>. This means they can save words and phrases to support and facilitate the translator’s work<\/strong>. SDL Trados Studio, MemoQ and Wordfast are the best-known CAT tools.<\/p>\n

Second, you have machine translation engines like DeepL and Google Translate. While these can be handy if you want to translate a non-technical text, they remain flawed and a long way from being powerful enough to replace human and professional translators altogether…<\/p>\n

The future of translation: a homogeneous mix of humans and new technologies<\/strong><\/h2>\n

Humans will always be better at detecting cultural subtleties in a text than machines<\/strong>. However, human translators do need to climb on the bandwagon and use translation tools to save clients time and money.<\/p>\n

Many translation agencies<\/strong> are already using technical software<\/strong> to automate the translation process and CAT tools are constantly improving. Does that make the human translator obsolete?<\/p>\n

Not yet! For the moment, tools support translators in their work by allowing them to work faster<\/strong>. This would be impossible without artificial intelligence.<\/p>\n

The importance of artificial intelligence<\/strong><\/h2>\n

Translation and artificial intelligence (AI) are developing alongside each other. As a reminder, AI means machines that imitate human intelligence to carry out a range of virtual tasks<\/strong>.<\/p>\n

Based on neural algorithms (NMT)<\/a>, machine translation is improving daily because it can learn. Machines can now translate over 7000 languages and dialects successfully.<\/p>\n

Thanks to this technology, translation agencies can supply reliable translations more quickly.<\/p>\n

Futuristic translation services<\/strong><\/h2>\n

Researchers and experts are currently working on futuristic translation services <\/strong>like speech translation <\/strong>and neural machine translation (NMT)<\/strong>.<\/p>\n

Let’s explore these two futuristic solutions in the translation sector.<\/p>\n

Speech translation<\/strong><\/h3>\n

Speech translation is still in development. Why? Because this innovation requires three things:<\/p>\n

    \n
  1. Speech recognition and discourse analysis;<\/li>\n
  2. Translation (using a tool with a translation memory);<\/li>\n
  3. Text-to-speech delivery.<\/li>\n<\/ol>\n

    These three things must be meticulously carried out on all words spoken within a few tenths of a second.<\/p>\n

    Neural machine translation<\/strong><\/h3>\n

    Neural machine translation<\/strong> (NMT) is a technology that uses artificial neural networks. It has come on hugely in recent years. NMT uses data corpora: the more data there is in a corpus, the more reliable it is. This data includes documentation resources, terminology databases, glossaries<\/a>, etc.<\/p>\n

    Neural machine translation<\/strong><\/a> is now more like a human translator than a CAT tool.<\/p>\n

    What do these changes mean for translators?<\/strong><\/h2>\n

    Machine translation<\/strong> is the future. When it has progressed far enough, the translation profession may disappear<\/strong> over time to leave machines doing all the translation work.<\/p>\n

    At that point, humans will exclusively be in the role of post-editor. As a key element involved in producing faithful texts in a careful style<\/strong>, post-editors will never completely disappear.<\/p>\n

    In future, humans’ task will be detecting grammatical errors, mistranslations and misinterpretations in texts translated by a machine. To do this work, a post-editor needs to have real linguistic skills in the source and target languages. Thanks to him or her, machine translation will generate the same quality as a human translator.<\/p>\n

    The future of translation: things to remember<\/strong><\/h2>\n

    Although machine technical skills are increasingly approaching those of humans, nothing can replace the knowledge of a native translator<\/a>. Texts can be highly technical and without having studied the topic at hand, they are impossible to translate.<\/p>\n

    Of course, NMT<\/strong> can contextualise a text<\/strong> so the most appropriate word can be chosen. However, terminological and grammatical checks<\/strong> are and will remain an essential part of the translation process in every language pair.<\/p>\n

    In short, CAT tools<\/strong> are currently support tools<\/strong> that are becoming more and more important in the translation world. That said, post-editors<\/a> should always be there to check the output.<\/p>\n

    \n

    \"New<\/a><\/span>