Neural machine translation: definition and disadvantages
Neural machine translation represents the latest evolution of “traditional” machine translation systems. The first examples of machine translation date to the 1950s; people started using neural translation systems in the 2000s.
What is neural machine translation?
Neural machine translation is a software-based translation system that uses a network of synthetic neurons and artificial intelligence.
Artificial intelligence can perform deep learning, which allows it to continuously "learn" new concepts by analysing previously translated texts.
One of the key strengths of neural translation is its efficiency. NMT (Neural Machine Translation) can translate very long texts into several languages in a short space of time.
Therefore, the tool could be helpful when translating long or specialised texts into several languages or when you have limited time to complete a very long translation job.
In addition, neural machine translation can be more precise than less advanced translation software. It can identify authors’ intentions and reflect them in the target language.
It follows that, compared to other other machine translations, neural machine translations flow better and are more consistent with the source text’s intent; they also boast better internal terminological consistency.
What are the limitations of neural machine translation?
Like all types of machine translation, neural machine translation has certain limitations when translating texts that use specific turns of phrase or lots of symbolism.
The reason is that machine translation (whether neural or not) is only effective on texts you need to translate literally.
On the other hand, poetic and narrative texts are unsuitable for neural machine translation as they often use turns of phrase and wordplay. The reason is that, however sophisticated it may be, artificial intelligence can not yet understand and correctly translate implicit nuances related to language and meaning,
Why is it still necessary to use a professional translator?
The development of increasingly sophisticated machine translation techniques might make you think the job of professional translators has become superfluous.
In reality, human translators still need to proofread texts carefully during the post-editing phase. Human translators can correct translation errors and smooth out any inaccuracies or inconsistencies in the translated text. They also ensure that translations can achieve the original text’s purpose more effectively.
To sum up, you still need to use professional translators to produce quality, specialised translations.
Eurotrad's professional translators also provide a timely and high-quality post-editing service for texts on any topic: get in touch for a quote.
SHARE THIS POST
WRITTEN BY
A bridge between you and the world, this is our purpose: to help you on your internationalization journey
You will always be updated on the latest news and promotions
