The role of technology in the evolution of machine translation
New technologies have brought about significant changes in the way we live, and the field of translation is no exception. When we understand technology's role in translation correctly, it offers many advantages for both industry professionals and clients.
At Acolad, we acknowledge that technology has been part of the translation world for quite some time and is here to stay. That's why we want to stress the importance of not ignoring it and making the most of it in our daily work.
The evolution of technology in the world of translation
As we all know, technology has always marked the passage from one era to another in the history of mankind. This is also what happened in the translation sector.
In terms of milestones, we could go back to the discovery of the printing press, which multiplied the number of readers and translations. But we will keep it brief. In essence, the evolution of technology in translation comes down to the appearance of the first proposals for machine translation, which occurred in 1933. Its low acceptance gave rise to the famous CAT tools, or “computer-assisted translation”. The Internet's debut in 1969 was a game-changer, not only in translation but in many other areas as well.
The cloud has become the go-to platform for sharing information and translation resources. After discussing CAT tools, we'll explore machine translation and the application of neural networks in translation, which are closely tied to the emergence of artificial intelligence.
CAT tools: computer-assisted translation
The early struggles of machine translation prompted the emergence of Computer-Assisted Translation, commonly referred to as CAT.
CAT software relies on translation memories, which are essentially databases storing source language sentences and their corresponding translations to reuse within the same project or for future tasks.
This software can be very sophisticated, even in terms of managing terminology databases. In addition to guaranteeing the consistency and quality of translations, they allow both the end customer and the translator to save time and cut costs.
It's important to highlight that at Acolad, we have a strong belief in the role of technology in translation. That's why we use cutting-edge translation software to provide you with the outcomes you desire.
From machine translation to the latest advances in translation technology
As mentioned earlier, machine translation faced limited success during its initial stages. In fact, a 1966 ALPAC report went so far as to recommend discontinuing investment in machine translation development. This was due to its lower quality compared to human translation, as well as higher costs and longer turnaround times.
However, machine translation technologies didn't stop at that point. They've steadily improved, paving the way for a new horizon for industry professionals: post-editing.
Today, their role extends beyond mere translation. They also engage in post-editing and revising translations generated by automated systems, striving to create results that closely resemble human translations. This shift marks a significant change in the landscape.
There are three main types of machine translation (MT) today:
- one that uses linguistic information (commonly known as rule-based TA, or RBMT (Rule Based Machine Translation),
- the statistical one (or SMT, Statistical Machine Translation),
- and one based on neural networks (or NMT for Neural Machine Translation).
Let's focus on this last particular type of translation, which has made significant strides in recent years, thanks to the advancements in artificial intelligence.
Using neural networks allows translation by association, based on the input of millions of sentences. With each use, the volume of data increases and, with it, the quality of the translation. This technology now allows the translation of millions of words, in real time, with an accuracy and reliability close to that of a human being.
At Acolad, we have developed our own engine based on artificial neural networks, and we feed it daily with data from all the companies in the group in order to make it more efficient. Neural motors work almost like the human brain: they train and learn.
Artificial intelligence has developed to the point of being able to interpret the meaning of the message, even if it's poorly worded. Thanks to machine learning, the algorithm even manages to learn from mistakes. Our online machine translation platform, U-Translate, is the perfect example of the effectiveness of this technology.
Speech Recognition
However, artificial intelligence does not only affect the translation process itself. It also has a lot to contribute to new automatic speech recognition technologies, which can reliably transcribe speech to speed up the subsequent translation process.
Acolad uses a form of Automatic Speech Recognition, Large Vocabulary Continuous Speech Recognition (LVCSR). Based on the automatic identification of very short audio sequences, this technology makes it possible to produce an excellent quality transcription, provided the audio is recorded correctly.
The Acolad group at the forefront of innovation
As we've noticed, technology in the translation industry keeps evolving, and it's vital to make the most of it. Acolad Group is committed to staying at the cutting edge of innovation by working daily with the latest advances in translation technologies, in order to take full advantage of them without compromising the optimal quality that we are committed to providing.
