Published On: Fri, May 29th, 2015

How well do Google Translate and Other Similar Products Match up to Human Translators?

Share This
Tags

The interface between human intelligence and artificial intelligence (AI) on the Internet is perhaps best exhibited in Google Translate and other similar products developed to perform simple foreign language translation. Since the introduction of Web-based translation technologies, the localization of websites and other content has become far easier than before. While artificial intelligence in machine translation deploys language interpretation by way of mathematically programmed algorithms developed to narrow the abductive reasoning of keyword searches, human intelligence continues to exceed AI software application capabilities.

 

Updates to Google’s translator first launched in 2001 have made it possible to translate both text content and audio recordings to some effect. Although the app is limited in capabilities where nuances and word choice are concerned, the potential to interpret narrative has been vastly improved over preliminary programs designed to perform word-for-word translation. The visual translation feature to Google Translate encourages user engagement. Nevertheless, there are still flaws in the proficiency and precision of Google’s popular translation app, and other similar programs, in that machine logic cannot capture the entire scope of subtleties within each language.

 

Dimensional elements to speech and written text deviate in terms of formal and colloquial composition. Other factors such as grammar, syntax, and word choice are further complicated by the semiotics of meaning. Unlike professional translation services, these apps are also entirely outside the level of proficiency demanded of professional lexicons required for translation of business and legal documents. Human translators are capable to grasp the context of a written document, and interpret the intention for which it is meant.

 

Apps do not offer such apprehension. In fact, context is a basic dimension of speech and written communication, yet to this point no machine translator on the market offers users an adequate interface from which to designated context. Tense changes in Romance languages such as Spanish, French, Italian, Portuguese and Italian are also not understood in their entirety by machine translators. Emotions, speech inflections, and conceptual frameworks have the potential to escape the capabilities of an app translator.


The benefits of the Google Translate app and other like machine translation technologies are evident, however. Business and vacation travelers are the top users of these easy-to-use interpreter programs. One only has to think about the last time they were in an ethnic restaurant, to acknowledge that translation apps are a great product for clarifying confusion in another language. Whether reviewing the menu, or reading a newspaper from abroad, translator apps are a user-friendly option for managing the flow of information in a polyglot world.

But what about business or commercial documents?


 

VERBALINK (Professional Transcription and Translation Services)

Earlier this year, Google released a significant upgrade of its Google Translate application (Read article published on The Yucatan Times). No longer content to offer translations of written text, the upgraded app includes a visual translator and an audio translation function.

We don’t offer visual translation just yet, but as one of the only language services companies that offers both document translation and audio translation, we were curious: How well do our human translators match up to the machines? Could the machines imbue their translations with the nuances and thoughtful choices that we expect from human translators?

We enlisted two of our Spanish translators, Adriana and Gaby, to participate in a two-part test. For the first test, Google Translate and Adriana were required to translate a promotional document about Argentina’s National Beekeeping Council from Spanish into English. For the second test, both contenders had to translate approximately one minute of spoken Spanish audio into English.

Since Gaby is one of our quality control specialists, we had her serve as our judge. Gaby reviewed the finished English document, the initial Spanish transcription of the Spanish audio recording, and the English translation of the audio recording. Translations were judged on several dimensions, including word choice, grammar and syntax, and overall accuracy and comprehension.

So how did our contestants do? While Google Translate has come a long way since its humble beginnings in 2001, it’s still not going to take the place of human translators anytime soon.

For both document and audio translation, it was able to adequately capture the overall gist of the content and would be useful for quick, general understanding. However, Google’s translations were too raw to be used for any professional purpose, and readability was often compromised by poor word choices and questionable grammar.

Our human translator Adriana fared better, with higher overall accuracy and a greater ability to capture the tone and style of the original documents. By being able to really consider the subject matter, context, and target audience, Adriana was able to create translations that didn’t read like translations.

Of course, a human translator won’t fit in your pocket, and they may not always be able to accompany you to that very authentic Thai restaurant or overseas vacation, so here’s what we recommend: When you need immediate results, and they don’t have to be perfect, it’s OK to use Google Translate. For any sort of business or commercial documents, or anything that has to be precise, it’s best to stick with human translators.

verbal-ink-translation-infographic-final1

 

Source: http://verbalink.com/

Mexico Travel Care

footer-john-2


Comments

comments

Leave a comment

XHTML: You can use these html tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>