Is Machine Translation Taking over Human Translation?

At this point in history, humans and machines are working symbiotically in language translation. Is one, however, about to take over the reins at the expense of the other?

Computational linguistics has seen an emerging subfield in the name of Machine Translation. This particular area focuses on the translation of text from a language to another. Various algorithms in deep learning have been developed to progress the process. Google Translate, amongst the most popular online translation services, is leading the way. Machine Translation entails deep learning, large sets of sentences translated sentences from a source to a target language. The sentences form the data basis from which the computers learn the translations of words between two or more different languages.

The Case for Machines in Translation

How is the machine able to translate, in most cases correctly, what is ideally a human invention? A question that has puzzled many was the tremendous advancements of Neural Machine Translation. In the most basic form, machine translation is typically a word-for-word translation. The user inputs the words they intend to be translated: The computer outputs a replacement of each of the words one by one. However, this gets rather complicated when it comes to sentences and proses. Most translation tools are still miles away from following grammatical rules and understanding context. In short, machine learning is similar to looking up words in a bilingual dictionary.

In this disruptive age, machines are rapidly taking over most of what previously entailed human labor. Translation has been no exception either. Communication is one of the human fundamentals. Some would argue that communication, and by extension language, is what ultimately set us apart from the rest of the primates. However, humans have evolved and developed roughly 7,000 different languages. With the world shrinking into a global village, the need for translation services to bridge the gap has never been more necessary. Thus inevitably, technological development has come in handy to connect folks of all walks of life.

The applications of machine translation are not only varied but also immense. Almost all sectors of human interaction nowadays have been benefitting from the applicability and accessibility of machine translation:

  • Education has been boosted through the ease of sharing ideas, translation of academic papers, and collaborations.
  • Media has seen the ease of access to information, including sharing across various platforms.
  • Business and commerce have seen growth in international trade and travel, finance, a foreign transaction of goods and services, and multinational investments.
  • Government diplomacy has also benefitted from better foreign relations.

Besides the ease of access, the cost of most tools also tilts the scale in favor of machine learning. Currently, you can easily find a free online translator that will generally do a decent job. While the accuracy of the end product is debatable, the overall cost precedes hiring the services of an expert for translation.

The Case for Humans in Translation

Most online tools can only deliver about 60% accuracy on any given text. Despite the milestones, there is still a long way to go. This simple fact has led to the notion that machines are unlikely to replace human translation. Some are firm believers that computers will never achieve the accuracy levels of humans when it comes to translation.

The main reason why it is argued that machines are nowhere near human learning is because of the yet rudimentary nature of existing algorithms. The mere fact that machines are at best reliable on a word-for-word basis tells it all. Human language is built on more than words; word order, structure, and sentences that follow certain and specific grammatical and syntactical rules. Until the age where machines can process and replicate such complexities, humans will always come out on top.

Furthermore, machines are as of yet (and perhaps indefinitely) incapable of comprehending culture. Different languages, from different cultures, have unique and specific lexical rules and inclinations. In this case, it would be rather challenging for machines to understand language quirks, such as personal names, metaphors and idioms, nuances, and local slang. A machine herein would output a faulty word-for-word translation, therefore changing the intended meaning. In this case, only a human translator with an excellent comprehension of the language can appropriately translate to other languages.

More so, a language may have several dialects where similar words might have different meanings or connotations. Humans are also able to pick up utterances in different accents. Since they entirely rely on pronunciation, machines are thus far unable to differentiate words that a skilled human quickly would. This thus contributes to the significant discrepancy between human and machine translation.

Machines are also incapable of distinguishing words contextually.  This is particularly a considerable problem when it comes to words with double meanings. A human translator, on the other hand, can easily pick out the correct meaning of a word contextually. Thus, finding an appropriate equivalent in the second language is met with ease.

Machine translation, rather, unfortunately, lack a human touch. Any form of human writing contains a mood or tone that only a fellow human can decipher. It is easy to see how a poem translated by a machine can lose an evocative feeling, or a joke losing its humor. A human translator, contrastingly, can match and reproduce a translation without compromising on the humanness.

Conclusion

It takes another human to comprehend the complexities of language fully. For that matter, it is unimaginable that machine translation will ultimately force human translation out of the field. Nevertheless, the progress made on the machine translation end is not one to be overlooked. Despite the shortcomings, the core goal to unite people is highly promising. There is no limit to human endeavor once we have broken language barriers. Therefore, the future is auspicious for collaboration and interdependence of both human and machine translation. But for now, human translation still sits at the apex.

Image credit: Machine Translation via metamorworks/Shutterstock

Leave A Reply

Your email address will not be published.

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. AcceptRead More