A Guide to Machine Translation


BEST PRACTICES SERIES

Article ImageFor much of history, the primary obstacle preventing people from connecting with one another was distance. But in the mobile, social, always-on world in which we live today, distance is no longer the major challenge. Instead, our biggest challenge is the lack of a common language. It's the inability to understand one another that prevents us from making meaningful connections.

Isn't there an app for that? Yes, there is. Machine translation (MT) hails from the discipline of computational linguistics and aims to help humans who speak different languages communicate with one another. And while MT is not actually an "app," many app makers are attempting to harness the power of automated translation by building MT into their products and services. But there are challenges.

RULES-BASED VS. STATISTICAL MACHINE TRANSLATION

There are two distinctively different MT paradigms. One is called rules-based MT, in which words are translated from one language to another following specific linguistic rules. The other method is statistical MT. Statistical MT, as the name implies, involves using software designed to assess the statistical probability that text in one language is the equivalent of text in another. (Google Translate uses statistical MT.) More recently, market demands have created the need for hybrid approaches that leverage both rules and statistics to produce more accurate translations.

Critics often cite examples that illustrate the shortcomings of MT. They say it isn't ready for widespread usage because it cannot provide the same level of quality that human translators can. Others say that while MT might not be perfect, it is good enough for some situations.

Creating perfect translations with MT isn't easy. Language is complex. Understanding the intent of the original or "source" content is necessary in order to produce translated equivalent content in the desired or "target" language. Sentence length, the presence of idioms, jargon, and word order differences from one language to another also complicate the process.

MT software can be fine-tuned to accommodate linguistic differences by domain or profession. Without fine-tuning or leveraging the power of statistics, MT can produce some undesirable results.

THE FUTURE OF MACHINE TRANSLATION

Jaap van der Meer, of TAUS (Translation Automation Users Society), sees a day when MT will be ubiquitous-a utility such as electricity, water, and the internet. He says he believes there will soon come a time when MT will be embedded in everything from consumer electronics to automobiles, bank machines, digital kiosks, heavy machinery, and every other type of product imaginable. The way van der Meer sees it, MT will make the Internet of Things accessible to all, regardless of language.

If you are wondering whether you should leverage MT or not, you may be asking the wrong question. If your product or service is of significant interest to people who can't understand the language you use to produce content, chances are they are already using free MT engines, such as Google Translate, to translate your content. And they do it without your knowledge or permission. With that in mind, the question you should ask is, "What can we do to our source content to make it possible for MT engines to help people who speak other languages understand our content?" Of course, if you have a lot of legacy content, you must also decide which content is most important for prospective customers who speak other languages to understand, whether or not MT should be used, and, if so, how.

Some organizations use MT to create a first draft of sorts, passing the machine-translated content to human translators for post-MT cleanup. Human translators are important in situations in which it is necessary to ensure that machine-translated content meets quality standards. Others use MT to translate content that they provide to end users without post-translation editing. New approaches are in use in some industry sectors, including the use of the crowd to clean up content previously translated by machine.

The selection of the right MT system for your organization should be part of your overall global content strategy. The decision should be based on a thorough analysis of your content, customer, and business needs. Some of your content might require precision translation (the type you would expect expert human translators to provide), while other content might not.

One thing is certain: If you want your content to work for you and to attract as many prospects as possible, you must find ways of removing the obstacles language introduces. MT might just be the key to success.


Related Articles

Any organization with an interest in disseminating its content to a global audience needs to familiarize itself with the basics of machine translation. At a time when the globalization of commerce is yesterday's news, the number of companies with the capacity for international ecommerce has grown considerably. Machine translation is not necessarily reserved for the titans of industry, as many translation service providers are beginning to offer more scalable services to assist even small companies with translation projects.