What is machine translation and why has it changed language teaching forever?Happily, some language teachers have learned anew the age-old maxim: if you can’t beat ‘em, join ‘em. Let’s leave aside language learning products like Babbel, DuoLingo or Rosetta Stone, which are alternatives to classroom education. Machine translation services can become a language teacher’s best allies in making classroom education, and even homework assignment, more interesting and enjoyable.
 |
| Image credit:https://pixabay.com/vectors/cloud-computing-cloud-device-data-1989339/ |
What is machine translation?
More correctly, it’s machine translation software, driven by Artificial Intelligence techniques. We’ll be looking at some of the ways in which language teachers can transform the wonders of online translation and language software to enrich the studying approaches and learning experiences of their students as well as to make their own teaching experiences more fulfilling and easier.
Forgetting the human vs. machine war and learning to love software translation
Let’s start with the bottom line: machine translation is still no match for human translation when it comes to creative and persuasive language. Machines can’t deal well with humor, idiomatic expression, or poetry. Translation engines lose when they go up against professional linguists in informal, fictional, poetic, or free-style texts.
Software simply can’t rival human translation services for high-end translation challenges. Certainly, software translators are weaker in less common languages than the most popular, because less effort has gone into refining them. However, since the introduction of neural machine translation techniques in 2015, the race has gotten closer. When dealing with highly structured texts like news reports, financial reports, and sports reports, as well as some legal documents, machine translation has proven superior in direct competitions – with human judges!
Neural machine translation represents a revolution in the way artificial intelligence software approached text for transition purposes. What is neural machine translation? Before 2015, A text was approached sentence by sentence, word for word, from start to finish. With neural network translation, however, the whole text is read, back and forth, “digested” by smart software with internal connections grasped as interrelated networks, then compared with a vast database of previously analyzed texts. Each text is decoded then encoded, alternative translations ranked and revised in lightning speed. Foreign language teachers don’t need to know the AI mechanics: it’s enough to know that software machine translators do a pretty good job these days. And students know it.
Using machine translation services inside and outside the classroomThere are dozens of machine translation software options out there, but with a couple of exceptions, you can get by choosing between Google Translate and Microsoft Translator (formerly Bing Translator). The exceptions are in certain language pairs. If you’re translating to or from Chinese, consider Baidu. If German, consider DeepL. Otherwise, best to stick with the software giants.
Both have similar features, but differing interfaces. It’s a matter of personal preference, although the Big G tends to win more often in international AI translation competitions. There’s straight text translation, which can be done either by copy-pasting into a text box (after choosing your language pair) or translating a full document. Then there’s two-way voice interpretation, camera translation, and (Microsoft leads here) one-to-many translation where a lecturer can talk in one language and each listener can hear a simultaneous translation in their preferred language.
Here are a few fun ways to work for teachers to work with translation software:
The astounding progress that AI has made in recent years compels us to wonder (and worry) about the fate of language education in the face of the machine translation challenge. Traditional language teachers are struggling to cope with the vast capabilities of neural networks as applied to education. While machine translation is still far from perfect, structured texts can be translated reliably by software. And AI is even making inroads on more complex and creative texts.
We’ve provided some examples where language teachers can use AI translation software to enrich their classroom experiences and ease their teaching burdens. It seems that teachers must learn to live with language-learning and translation software rather than fighting against it. In this way, machine translation can be transformed into a benign if powerful teaching tools in the service of humans rather than replacing them.
This is a sponsored post.
Forgetting the human vs. machine war and learning to love software translation
Let’s start with the bottom line: machine translation is still no match for human translation when it comes to creative and persuasive language. Machines can’t deal well with humor, idiomatic expression, or poetry. Translation engines lose when they go up against professional linguists in informal, fictional, poetic, or free-style texts.
Software simply can’t rival human translation services for high-end translation challenges. Certainly, software translators are weaker in less common languages than the most popular, because less effort has gone into refining them. However, since the introduction of neural machine translation techniques in 2015, the race has gotten closer. When dealing with highly structured texts like news reports, financial reports, and sports reports, as well as some legal documents, machine translation has proven superior in direct competitions – with human judges!
Neural machine translation represents a revolution in the way artificial intelligence software approached text for transition purposes. What is neural machine translation? Before 2015, A text was approached sentence by sentence, word for word, from start to finish. With neural network translation, however, the whole text is read, back and forth, “digested” by smart software with internal connections grasped as interrelated networks, then compared with a vast database of previously analyzed texts. Each text is decoded then encoded, alternative translations ranked and revised in lightning speed. Foreign language teachers don’t need to know the AI mechanics: it’s enough to know that software machine translators do a pretty good job these days. And students know it.
Using machine translation services inside and outside the classroomThere are dozens of machine translation software options out there, but with a couple of exceptions, you can get by choosing between Google Translate and Microsoft Translator (formerly Bing Translator). The exceptions are in certain language pairs. If you’re translating to or from Chinese, consider Baidu. If German, consider DeepL. Otherwise, best to stick with the software giants.
Both have similar features, but differing interfaces. It’s a matter of personal preference, although the Big G tends to win more often in international AI translation competitions. There’s straight text translation, which can be done either by copy-pasting into a text box (after choosing your language pair) or translating a full document. Then there’s two-way voice interpretation, camera translation, and (Microsoft leads here) one-to-many translation where a lecturer can talk in one language and each listener can hear a simultaneous translation in their preferred language.
Here are a few fun ways to work for teachers to work with translation software:
- Give a translation assignment and then compare results to see which translation software students used to “assist” them. Discuss “translation plagiarism.”
- Play “translation telephone” and let the student see how translating with machines from one language to another then another and then another transmogrifies meaning irretrievably: a warning not to rely too much on machine translation
- Teach them how to use voice translation to practice their accents and pronunciation
- Use camera translation to translate signs and menus, showing how idioms and wordplay can befuddle software translators.
The astounding progress that AI has made in recent years compels us to wonder (and worry) about the fate of language education in the face of the machine translation challenge. Traditional language teachers are struggling to cope with the vast capabilities of neural networks as applied to education. While machine translation is still far from perfect, structured texts can be translated reliably by software. And AI is even making inroads on more complex and creative texts.
We’ve provided some examples where language teachers can use AI translation software to enrich their classroom experiences and ease their teaching burdens. It seems that teachers must learn to live with language-learning and translation software rather than fighting against it. In this way, machine translation can be transformed into a benign if powerful teaching tools in the service of humans rather than replacing them.
This is a sponsored post.
