Language technology, a friend or foe?

Have you ever tried out Google Translate or a dictionary application when travelling? Have you tested the text-to-speech features on your smart phone? If you answered yes, then you have been using language technology.

Language technology means using a computer or other form of technology to process natural language or create text in spoken or written form.

This may include machine translators like Google Translate, chatbots that respond to questions, or text analytics. Language technology may also include your digital assistant tool at home, such as Siri or Alexa that you can ask to turn on your music. Language technology also includes the subtitles in your favourite film, and even the fun language learning apps you may sometimes use to practice a foreign language. Language technology is found in our everyday life, and it can be found in tools to help translators as well.

At Lingo, we use speech-to-text software, and very carefully selected AI technology offered by our computer-assisted translation software. This includes translation memories, some machine translation options, and text recognition tools. We would love to tell you more about our native translator’s work with machine-translated text editing, also known as post-editing.

What are CAT tools?

CAT (Computer-Assisted Translation) tools are computer software programs where translators translate text. They help to preserve the format of the text, and save a translation project. Translation agencies translate many varieties of texts, so being able to save key words and phrases is useful.

When a translation project is saved in the system and text segments are stored in a large database called the translation memory, we can use the saved phrases and words to help us translate future texts. This can be helpful especially when we want to keep vocabulary, style or tone of voice consistent across the same customer’s texts.

What is machine translation?

Machine translation refers to the process of using an automated translator that translates text from one language to another. Machine translators are commonly built upon statistics and complex processing. Even though machine translators are improving, it is important to be carefully consider what texts are translated with a machine translator..

If a quality machine translator is used, then it may be a productive choice, but oftentimes it’s hard to decipher which machine translators are worth using and the result is a poor translation.

Machine translators have a hard time processing large amounts of text data at once with complete accuracy, and may not be able to correctly translate words that are specific to certain topics or domains.

Machine translators may not produce the preferred language variant. Sometimes machine translations/translators do not formulate phrases in the correct order, and the text becomes difficult to read. This also makes it difficult to correct afterwards.

Machine translators are good/useful if you need to quickly understand a general idea about a text or if you need to find out the meaning of a word or phrase, but for accurate translations, it is best to seek a human translator.

Post-editing 

If a text is translated with a machine translator, then it must be post-edited. Post-editing is essential to ensure translation quality. Post-editing means that a human translator or post-editor checks the work of the machine translator. They read the original text, and the text that the machine translator creates. They make the necessary changes, but also check that the meaning of the text is correct, the sentences are written correctly, and that it is easy to read.

The translator or post-editor must be an expert, with linguistic knowledge of both the source and target languages. The post-editor should have excellent cultural knowledge, a broad vocabulary in the domain in question, excellent grammatical skills in the target language, and good post-editing skills.

The post-editing process requires a different process than proofreading, and may require more time and resources. The effort put into post-editing is hard to show, and sometimes effort is not reflected in the time spent or in the corrections made. The effort that post-editors put in is in the process.

At Lingo, we have excellent, professional translators that can cooperate with the machine translation output and we offer our customers quality post-edited translations. While we use digital/AI-powered tools, we value and rely on the expertise of human translators and linguists. Leaving it to the experts is the best way to ensure a high-quality translation!

Read our article on how Lingo’s translation process works here: https://lingo.fi/mitenkaannostehdaan/ (in Finnish).

Predicting the future

In the future, Artificial Intelligence (A.I.) could change the way we translate texts, because A.I. is rapidly improving especially in the world of translation and linguistics. Currently many machine translators such as Google Translate use complex processing called neural networks to produce translated text. This was an important step, as we have seen improvements from previous machine translators that have been mostly built on statistical information.

In the future, A.I. machine translators can translate into many more languages and store much more information. With the possibility to process large amounts of language data, we can access more languages, and communicate with more of the world.

The future may include improvements in voice recognition, and real-time translated audio output, as well as improvements in grammar, vocabulary, cultural transfer, and more.

It will be interesting to see if A.I. will be able to adapt to language change or blooming language communities. At the pace A.I. is evolving, we can hold onto our seats and wait and see what the next leaps will offer.

Patricia Mikkola

Lingo Languages Oy
agile reliable multilingual