Currently, we have large amounts of information from any channel or part of the world. The confluence of cultures and constant communication between countries has turned society into a global group in which everything is interrelated.
Every day we receive a multitude of news from all over the world, access information and communicate with people who are beyond the reach of our language and culture.
This simply would not be possible without translation.
Translation allows us to communicate anywhere in the world. If we look at it in market terms, we know that the current market is a global one. This means that, thanks to translation, language barriers can be overcome, allowing access to international markets.
The important role that translation plays in our lives, combined with technological advances in recent decades, has created a union—almost a fusion—of these two disciplines, resulting in the concept of translation technology. In other words, this is the use of software to improve the way written text is converted from one language into another.
This arises from the need to obtain better and faster translation services, and has led to all kinds of translation technologies, the one with the greatest projection being automatic translation or Machine Translation (MT) .
However, is this the perfect option to synchronize language efficiency and quality? Where is the future of translation technology headed?
History and evolution of translation technology
First of all, let's clarify the concept of machine translation: it is understood as the automated process of translation, from a language A to a language B , without the help or human intervention.
Its origins date back to the Second World War, when interest in translating and decrypting messages from the other side was top priority. In 1949, The Rockefeller Foundation, with Warren Weaver (cryptographer) leading the way, began to develop the first hypotheses on the use of computers in the translation process.
In the years following, different universities began conducting studies on this new field. It wasn’t until 1954 that the first IBM prototype was unveiled at Georgetown University. It was a system that was capable of translating 49 sentences and had 6 grammar rules implemented .
This prototype generated great interest, which continues to this day.
Types of machine translation
Translation technology is not only limited to TM, it also encompasses other tools such as: translation memories, virtual interpreting technologies, speech-to-text technologies, and terminology management tools.
However, machine translation has accumulated great success as a translation tool, due to its lesser dependence on human action.
Since the creation of machine translation, there have been three distinct types:
Rule-Based Machine Translation (RBMT)
This type of machine translation is based on an engine that analyzes the source language syntax and uses linguistic rules and dictionaries to carry the meaning over to the target language.
Statistical Machine Translation (SMT)
The translations are generated according to statistical models, from a corpus of bilingual texts. This means that it works by using other translated texts, inputted beforehand, as a reference.
Neural machine translation
The translation technology sector took a major leap forward with the emergence of Neural Machine Translation (NMT). This type of machine translation uses artificial intelligence (AI) to generate translations.
This technological leap represented a change for machine translation. The mechanism attempts to emulate how a human translator would think to provide the most accurate result.
The engine learns from previous translations and is able to make predictions. It doesn’t simply translate word by word, but can understand full sentences and the context within paragraphs.
This breakthrough generated considerable interest among those seeking to translate large volumes of words, as well as highly standardized commercial documents.
However, it’s not a perfect process… yet. The use of this new translation technology tool still produces certain defects, such as: little readability, some slip with negative particles, terms not suitable for the context, etc.
It should not be forgotten that it is a technological system in the process of being perfected, not a human professional.
Uses of MT
Machine translator, as previously mentioned, has certain drawbacks when dealing with particular types of text. However, for texts with less linguistic complexity, it is a highly effective and profitable solution .
Some advantages of machine translation are:
- High level of terminological consistency and precision.
- An efficient and effective process.
- Lower translation cost.
Either way, it’s important to keep the purpose of the translation in mind, as machine translation may not be the most suitable method.
MT has a long road ahead of it, but it is speeding along. It’s true that improvements in this translation technology have resulted in lower costs for certain projects, but the creative limitations of this tool must be taken into account.
Ideally, once the machine translation tool has been used, a professional translator proofreads the texts.
At the end of the day, there are certain translation needs that can only be met with a human touch.
It’s likely that, with time, we will see the balance tip and, instead of talking about computer-assisted human translation, we’ll be talking about human-assisted computer translation.
As you can see, there are multiple translation technologies and these are already essential in daily translation work. The question is to properly choose the one that best suits your needs.
We hope this article has been helpful to you, you can find more on our blog .