E.g., 11/18/2019
E.g., 11/18/2019

The Value of People Remains Despite Growth of Machine Translation Market

By: Jeremy Coombs, Vice President, International Operations - MultiLing Corporation

28 April 2015

New reports forecast growth of the machine translation market at nearly 25 percent in the next four years, with technological improvements likely along the way. With such expectations, it’s tempting to believe computers will soon translate well enough that there will be no need for human translators.

It’s not time to give up on humans in the near future, however. While technology such as machine translation plays an important role in making valuable content in one language accessible to all, humans specialized in their native languages, cultures, and fields of expertise still need to be involved in the majority of translations. This is especially true for intellectual property, particularly patent applications, where even one mistranslated word can result in invalidated patents and millions of lost revenue.

Consider the following points:

1. Machine translation uses the past as its guide

Every year, Oxford Dictionary announces a list of the world’s new words. For example, in 2014, the new word used most often (twice as much as it was in 2013) was “vape,” as in electronic cigarettes. Selfie, the 2013 word, increased 17,000 percent in use over two years, before it reached the top.

So even if you compile a massive database of all language used everywhere today, and update it in real time, it can only tell you how people have used language, not how they are using it today, and definitely not how it is being used tomorrow. In the patent world, we are always working with new inventions, many of which may need to use words in new ways, and perhaps even include new words – language you won’t find in yesterday’s documents.

2. Language is bound to culture and context.

Since language is bound so tightly to culture, “literal” translations are often incredibly misleading. In one language, the culture may dictate using a verbose greeting. If it is translated literally, it will sound ridiculous in another language or culture. It is possible for an advanced statistical machine translation to look for an equivalent, rather than literal translation. And then there is the potential problem that an “equivalent” word will most likely be different for different contexts, so the translation decision is linked to context and culture more so than to language. The complexity of understanding culture and context limits machine translation.

Touting the ongoing need for people – especially with native language, culture and technical skills – isn’t meant to diminish the value that technology-assisted translations offer them, however. In fact, machine-assisted translations provide valuable translation assets to customers in addition to the final translated document. Translation memory (allowing translation teams to reuse and leverage past approved translations) makes the work of human translators faster and more accurate. It can also save money for customers. We’ve found that it is not an uncommon scenario within large global corporations for different departments to request translation for the exact same document. Document management on our end, and using translation memory as a first pass, easily identifies these duplicate requests and has literally saved clients $100,000+ per year.

Additionally, terminology management software continuously updates glossaries with translations of frequently used terms that can be specific to a client’s internal vernacular and trade sector. This database can grow to hold tens of thousands of terms that can be leveraged across an enterprise for consistent translations.

Even the European Patent Office (EPO) concedes that its online machine translation tool – Patent Translate – “cannot provide legally binding translations.” But this technology definitely makes certain processes – such as the discovery of patent applications –quicker and simpler, which is what it is designed to do. Patent Translate can provide the gist of any patent or patent-related document and help determine if it is relevant. Then you can determine better where you need to invest in human translation.

But that brings us back to the need for human translators, especially for any document that can make or break your business. In this new era, it’s not so much a question of machine or man, but of how to manage the right mix for specific translation needs. Considering the purpose of the translation – marketing, literary, legal, etc. – will dictate the mix of resources to produce the best, most efficient translations needed by the customer.  

Jeremy Coombs

Jeremy Coombs is the vice president of international operations at MultiLing, the innovative leader in intellectual property (IP) translations and related support services for foreign patent filings. Jeremy Coombs joined MultiLing in 1999 and has become one of the company’s principle operations and technology minds.  In his current role as Vice President of International Operations, he manages MultiLing’s international offices across Asia, Europe and South America.  Combining his love of project management and computer science, he also helps guide the development of MultiLing’s translation memory system, terminology management system and project tracking/workflow system. In addition to his work at MultiLing, he volunteers time to Translators Without Borders and has influenced international standards through his participation in ASTM and ISO committees. Coombs received a bachelor’s degree in linguistics with an emphasis on computer science and Scandinavian studies from Brigham Young University in 2000.  He is a PMP accredited member of the Project Management Institute. Jeremy is a native English speaker and is also fluent in Finnish and Swedish.

randomness