Machine Translation (MT) has been around for decades, but enterprise funding for research in TMs and CAT tools began to garner renewed interest towards the end of 1980s, with increasing tendency of MT being bundled with Artificial Intelligence (Lawson). Many forward-thinking translators who tried out MT during this period recognised two important things about MT: First, this strange thing might work after all. Secondly, it is unlikely to work well enough to threaten them (Lawson, emphasis mine).
Though back in 1980s this attitude and perception was true, the use of MT has drastically changed in the past decade. With globalization, multiculturalism, multilingual governance, technological advancements, and the age of the Internet, there is an immense amount of content from companies as well as individual users. Even if a fraction of this content needs to be translated, the use of machines is inevitable. That is why MT has proliferated the translation and localization industry.
This is turn has led to a problematic relationship between the users, proponents, and providers of MT on one hand, and traditional translators on the other. The fact is that translators are still the linchpin of the localization industry. Even with immensely improved automated translation output, translators will always be vital to the language quality review process.
As MT stands today, it needs the translator¹s skill to improve the output quality. What technology can do however, is develop a process, which will help translators increase their productivity and reduce the tedious, repetitive errors. For example, the Language Quality Review (LQR) process that helps improve the MT output is generally rife with collaborative difficulties for Project Managers. However, an automated language review tool that helps streamline and formalise the workflow would not only help reduce frustrating, repetitive MT errors, but also substantially increase translation productivity.
I will talk about innovations in the field of translation quality management in detail during my presentation at the ATC conference in September. For now, it is sufficient to point out again that with the introduction of new technologies, the translator¹s role has metamorphosed into more of a contributively dynamic part.
MT today has come a long way from its inception and developers are tirelessly striving to bring the quality of automated translation closer to the standard of professional human translators. However, most translators will argue that MT output can never match the multifaceted and nuanced translation style of the human brain. While this is true, recent innovations and developments in the field of Deep Learning and Neural Machine Learning at Montreal Institute for Learning Algorithms and KantanLabs, our very own R&D department makes me feel confident that machine will soon be able to deal with translation more ³intelligently² than ever before.
MT is here to stay the basic technology may evolve and improve, but in a world where we are inundated with information overload, embracing automated translation technology is the only way forward for anyone in the language industry.