Machine translation or MT (Machine Translation) is the translation of a text to be translated by a computer software developed for the translation process. In machine translation, the text to be translated is automatically transferred from the source language to the target language by the computer software that will perform the translation.
MT is dynamic and self-learning, constantly renewing and updating itself for more accurate translations, especially in today's fast-paced technological world. We observe that machine translation has developed at an extraordinary pace in recent years and has become more preferable to human translation. It should be noted that MT should not be confused with Computer Aided Translation tools. CATs (Computer Aided Translation) are memory and terminology software designed to facilitate human translation processes. However, MT compiles statistical data and produces the target text by scanning millions of previously translated pairs, i.e. source and target texts, while taking into account the grammar and spelling rules of the language. We will publish a more detailed blog post on CATs soon.
Advantages vs. Disadvantages of Machine Translation
Speed Advantage
MT is extremely fast compared to human translation. Thanks to this speed advantage, many brands prefer machine translation. Since human translation processes are usually completed in a certain period of time, brands have to wait for this work and incur a certain time cost to complete the work.
With machine translation, you load a text with tens of thousands of words that need to be translated into computer software, click on the translate button and within seconds you have a translation of your text. No matter how you look at it, this is a time-saver.
Cost Advantage
Machine translation is also cheaper in terms of cost than human translation. As clearly seen in the example of Google Translate, developed by Google in recent years, it is possible to perform machine translation without incurring any costs.
Diversity in Language Meaning
Machine translation can include any language loaded into computer software. This means that MT can translate from many source languages to many target languages at the same time.
Disadvantages of Machine Translation
Meaning Errors, Mistranslations
While machine translation offers advantages such as speed, cost and language diversity, it also brings some disadvantages.
The most prominent of these disadvantages is mistranslation. Since MT is computer software, it may not be able to handle the different meanings of some words as accurately as human translation. In particular, hundreds of words that are spelled the same but have completely different meanings when the context is different remain a major problem even for the best MT software. We can assume that MT still needs time to correctly translate very common words such as face, husband, eye, travel, al, hearth, etc., which have at least two and sometimes four or five different common meanings. This may lead to the frequent occurrence of significant semantic errors in a machine translation. At least for the next five to ten years, after machine translation, the target text should be checked and corrected by a translator specialized in this field.
Post-editing after Machine Translation (MT):
- a) Full Post-editing: This is the process of corrections made by the human translator and training of the MT process in order to produce an acceptable target text after machine translation. The stylistic features of the text are also reflected in the target text.
- b) Fast Post-editing: This is the process of post-editing by a human translator to produce a intelligible target text after machine translation.
Machine translation is still evolving, is still being debugged and is rapidly becoming preferable to human translation.
The issues that customers or users should pay attention to regarding MT are the following:
1- Is security/privacy ensured? It is necessary to prevent the information uploaded to the software from being leaked by the software company and used by third parties intentionally or unintentionally.
2- Is the software designed in a specific domain or are language pairs from all types of domains included in the statistics? For a highly efficient target text, software that is built in each specific domain separately should be preferred.
3- Is post-MT proofreading done by specialized translators? Translators need to be familiar with MT dynamics, feedback methods, and have specific domain knowledge.
Comments are closed