article img

Best use case for machine translation: When to use Machine translation with post edition?

What is Machine Translation ?

Machine translation (MT) is a form of language translation where a computer program is used to process text and translate it from one language to another. The process involves the use of algorithms, software and statistical models to create translations. The translations created by machine translation are often of lower quality than those created by human translators, but advances in technology are making machine translation more accurate and efficient over time.


When to use MT?

Machine translation is best used when you need quick translations of large amounts of text, such as websites or documents, where accuracy is not paramount. It is also a good choice if you need to get a basic understanding of the content in a foreign language.

A human translator is nevertheless advised for the translation of client-facing materials, marketing content, and other open publications. A human expert can add needed cultural subtleties and deeper meanings to the text. For example, a literary or journalistic text's multiple levels of meaning can only be decoded by a person. Although it does not completely replace human translators, machine translation can assist them in translating texts that would not otherwise be translated at all.


Free Machine Translation

Typically, free online translators like Google Translate spring to mind when people think about machine translation. Make sure to understand the terms and conditions, free online Machine Translation programs do not preserve the confidentiality of your documents. Whatever you upload to the software is now available to the rest of the internet. Savings on costs are significantly outweighed by security concerns and a lack of control over your content. 


Uses of full machine translation

Machine translation is currently a viable option for non-public, corporate-internal communication and information sharing, especially when the text's content needs to be conveyed swiftly, but the audience is small. It's crucial that everyone is informed about the usage of machine translation.  

The prototype of websites can also benefit from machine translation. We are all aware that translating into some languages yields more text than translating into others. Machine translation makes it possible to swiftly and affordably test a new web design in various languages without wasting time and money on quality translation. 

Product review texts for online businesses are a nice example of texts with a brief life cycle.
Using machine translation to translate evaluations for an online company that operates in several nations might be a great method to satisfy customers. Here, automated translation supports customers' purchasing choices. Especially when knowing that users really love reviews in their own languages: see our article about it.


Machine translation with post edition (MTPE)

In the post-editing translation process, the source text is first machine translated into the target language. A human editor then evaluates and edits the content to ensure that it is completely accurate and that the original sentiment is retained in the translation. Additionally, the outcome of post-editing can be utilized to instruct the machine translation engine to enhance its translations in the future. Working with language professionals enables you to create memory of past translation to improve the MT model for improved later use. 

Post-edited texts typically have sentence lengths that are much more closely aligned with the source text, less lexical variety, and less lexical density. Texts produced using MT and then post edited are typically less rich and varied. They are usually greatly influenced by the parent language and have a tendency to be more homogeneous than full human translations. This is why post editors cannot really improve a bad quality MT result, it almost requires them re-translate everything from scratch.  

Post editing and translating are different jobs

Post-editing and translating may look similar, but are different jobs. When post-editing, there is already something there, making it easier to be tempted to quickly accept the machine-translated output without giving it a thorough review. This often leads to errors, so paying attention to detail is key in order to avoid making mistakes.

This is made more difficult by the fact that recent advances in machine translation, particularly neural machine translation, have greatly improved the output's flow and grammatical accuracy. The result might seem so natural that post-editors mistakenly believe there is less to edit.  

On the other hand, when you’re translating from scratch, all the text is yours to create, and you’re not limited by the MT output. This allows you to take more time on each sentence, considering the nuances of each word, phrase, and structure. You can also take a more creative approach, which typically leads to more natural-sounding translations that are better suited to the target audience.

MT or MTPE is the right solution when time and cost savings are the ultimate goals. It is a less expensive and faster solution than traditional human translation because it involves the use of automated technology to do a substantial portion of the work. However, it is important to note that MTPE does not guarantee that the quality of the translated text will be as high as a full human translation, sometimes post editing can’t do much about it. Thus, it is essential to consider the intended use of the translated text when deciding if MTPE is the appropriate solution.

When translating safety data sheets or lists of ingredients on a cereal box, machine translation can be faster and more efficient in these scenarios, as it is better at recognizing simple repetitive patterns. However, for text that requires more creativity, such as a website, it might be a better idea to go for transcreation and localization instead of machine translation to get a better outcome.

NEWS : Powerling has recently acquired WCS group.

For more information, click here!