Posts

Translation in Your Own Voice

At a presentation in China, Microsoft recently demonstrated an improved machine translation technology that allows for real-time translation in your own voice. Using the system, Microsoft’s Chief Research Officer Rick Rashid was able to give a presentation in Chinese in his own voice, even though he doesn’t speak the language.

How does it work? Prompted by the attention his presentation generated, Rashid wrote a blog post to explain the technology behind the system:

“In my presentation, I showed how we take the text that represents my speech and run it through translation- in this case, turning my English into Chinese in two steps. The first takes my words and finds the Chinese equivalents, and while non-trivial, this is the easy part. The second reorders the words to be appropriate for Chinese, an important step for correct translation between languages.”

To more accurately perform the first step of the process, Microsoft is using a technique called Deep Neural Networks, which it says mimics the patterns of the human brain to make speech recognition more accurate. One caveat: the Deep Neural Networks technology may be better at recognizing words, but it’s still no substitute for the brain of a skilled interpreter. As Rashid wrote:

“We have been able to reduce the word error rate for speech by over 30% compared to previous methods. This means that rather than having one word in 4 or 5 incorrect, now the error rate is one word in 7 or 8…Of course, there are still likely to be errors in both the English text and the translation into Chinese, and the results can sometimes be humorous. Still, the technology has developed to be quite useful.”

While this technology is amazing and will definitely have its uses in the future, I doubt it will replace the knowledge and understanding that a trained translator brings to the job any time soon. What do you think?

Talking Business: How to Avoid a Translation Fail

Some phrases just don’t translate-especially when you are relying on a computer to do the heavy lifting. The International Trade website has published a list of English business phrases that don’t translate well, and it illustrates this point beautifully.

Take, for example, the common English expression “give me a ballpark figure.” Translated into Russian literally, as a computer would do it, you get “Give to me the diagram of the baseball stadium.” Unless you’re in the baseball stadium construction business, that simply won’t do. In Spanish, “We’ll hit the ground running” turns into a phrase that brings to mind an action movie: “We will strike the earth operation.” The best of the bunch is probably the literal Chinese translation of the phrase “We need to get our ducks in a row.” Once translated, it becomes “We need to obtain our duck continuously.” What?!?!

So, how do you avoid sounding like an idiot when you deal with foreign clients? The best course of action is to avoid machine translation if at all possible-it simply isn’t reliable enough yet. If you do need to use machine translation for a business project, write in simple language, avoiding metaphors, figurative language, jargon and colloquial expressions.

Richard Brooks, General Manager of UK based translation firm K International, has the following advice for UK businesses:

“Idioms are common place in workplaces across Britain and its fine (within reason) to use them in your local marketing activities. The tricky part comes in when you need to translate that message for use in another region.

Computers (at the moment) simply cannot understand the real meaning behind these idioms. For copy, that when translated is intended to convert potentially interested parties into sales revenue then a real human being must be used in the translation process.

For the best results recreating your message for use in another country a service such as transcreation should be used which includes incountry testing and cultural focus groups.

Get it right and you’ll have a winning marketing campaign that will spread like wildfire (excuse the idiom) in the blogs and social media networks, get it wrong and people will think you’re an idiot”

Assuming you have a competent interpreter, human-powered translation is always superior because human interpreters recognize expressions like these and know how to translate them appropriately to convey the correct meaning.

Translation Value

Translation: Price is what you pay, Value is what you get

Picture the scene, it’s the weekend, the sun is shining, not a cloud in the sky, a perfect day for a leisurely drive. About an hour into your jaunt around the local country roads, you notice a strange clunking sound coming from under the bonnet. It looks like a trip to the garage is in order. Once you get home you call the guy (or girl) you always call when your car needs attention. You drop it off at the garage and wait for the workshop to call, what are the first three things you want to learn from that call?… Most people would likely answer along the lines of “can they fix it, what is it going to cost and how long is it going to take”, probably in that order.

Now you are probably wondering what going for a drive and suffering an impending breakdown has to do with anything, well I’ll get to that. About a year ago I was talking to a chap in a pub, the best stories always start with that line right? His name is Dave, you wouldn’t say he was anything out of the ordinary, casually dressed, glasses, drives a van, all very run of the mill, he wouldn’t mind me saying that he’d probably agree. Anyway, I sat at the bar waiting for my friends to finally show up and just happened to strike up a conversation with him. He told me about how he works in a garage and has done probably longer than I’ve been alive, another classic line from the book of pub stories huh. Dave’s customers go through exactly the same ritual as I had you imagine at the beginning, but when it comes to that phone call, his customers have slightly different expectations. Read more

human translation vs machine translation

A Translation Showdown: Man vs Machine Translation

Computer scientists began trying to solve the problem of machine translation in the 1950s.  Since then, both the availability and quality of machine translation have improved tremendously. But in the battle of human translation vs machine translation, are humans now expendable?

Some scientists working on machine translation claim that with recent improvements, algorithms are almost as good at translation as humans.  And when the subject of “jobs that will soon be taken over by robots” comes up, futurists almost always put “translation” in the crosshairs.

But what happens when machines take on human translators? Earlier this month, Sejong Cyber University and the International Interpretation and Translation Association of Korea decided to find out. 3 machine translation programs went up against a group of human translators. It was a translation showdown: human translation vs machine translation.

Man versus machine, the translation industry’s version of the famous contest between John Henry and the steam-powered hammer  Guess who won? Read more

Machine Translation Versus Human Translation: A Professional Weighs In

Which is better, machine-powered translation or human-powered translation? In this weekend’s New York Times, David Bellos, the director of the Program in Translation and Intercultural Communication at Princeton University, weighs in.

Bellos points out that  both machines and humans make mistakes in translation. While we like to joke about the fallacy of relying on a machine to translate all the different nuances of language, translators who are poorly trained or are working too hard make errors, too.

Machine Translation

Bellos notes that machine translation is well-suited to situations where there are not enough translators or interpreters available and translations don’t have to be perfect to be usable. For example, machine translation was extremely helpful for emergency personnel on the ground in Haiti.

However, machine translation relies on either a dictionary of words or their meanings combined with the rules of grammar that can be used to combine them or on comparing the text being translated to other, similar texts that have already been translated. Neither method is 100% accurate, especially when it comes to translating literature, creative writing and figures of speech.  Even Google Translate, which has access to all of the literature contained in Google’s considerable library of books, often comes up with gibberish when faced with literary translations.

Sure, computers don’t get tired, and they don’t base their performance on whether or not they are being paid a decent salary. But, as Bellos notes:

“Machine translation is not conceived or programmed to take into account the purpose, real-world context or style of any utterance. “

In situations in which an accurate translation is a must, a qualified human translator who is familiar with the nuances of both languages and cultures will beat a machine every time-even if the machine has all the power of Google behind it.

Microsoft Releases Online Hmong Translator

In February, Microsoft announced that they had added Hmong to the list of languages translated by their search engine, Bing.  Last week, Microsoft announced the release of a new, free online Hmong/English translator for smart phones, chats and websites. What’s with the focus on Hmong? It’s actually part of an initiative to help preserve the language within the Hmong community in the US.

The Hmong currently living in the US are mainly refugees from the Indochina Wars, as many Hmong sided with America in both Vietnam and in the “Secret War” in Laos. In the US, they are safe from the persecution they faced in their homeland, but their culture is at risk as their children often neglect Hmong in favor of speaking only English.

As Microsoft program manager Will Lewis explained to Business Week,

“All these years, the language has been preserved, despite efforts to eradicate it. Now, the irony is that in the United States, a country where they’re free to speak it, the thing that never happened in Hmong history is happening; some children are not learning Hmong.”

So, Hmong community leaders decided to partner with Microsoft to bring the language to the place kids spend most of their time these days: the Internet. Though the online translator will also help older Hmong refugees who can’t speak English, the big hope is that it will also help keep the language relevant for the younger generation. As Fresno State University outreach counselor Phong Yang told a local news affiliate, “Without language, a culture will disappear.”

Also, the technology used to create the translator holds the promise of being able to do the same thing for the thousands of other threatened minority languages scattered across the globe. Using dictionary entries and documents in both the original language and English, the program is able to “learn” what words are likely to mean by using context clues.

Learning the language of their grandparents along with English allows kids to take pride in their heritage. Need proof? The Business Week article quotes US-born Joshua Lor. As a young boy, Lor told his mother that he didn’t want to be Hmong any more. Lor said that learning the language was key to changing that perspective:

“My grandpa told me stories about the Hmong, about how he served in the war, and how they moved from Laos to Thailand to America. The language opened my eyes to the history of Hmong culture. It’s exciting that the translator can help kids do that.”

Hmong is actually a macrolanguage with numerous closely-related dialects. Currently, the translator only works with Hmong White, though the team is working on one for Hmong Green, the other major dialect spoken by Hmong in the United States.

Photo Credit: Attribution Some rights reserved by Arian Zwegers

The dangers of online machine translation extend beyond quality

The Dangers of Online Machine Translation Extend Beyond Quality

Data privacy and data security have become two increasingly hot topics in recent years. As technology grows rapidly in its scope and capabilities, it seems that everyone from Google to the government is keen to glean all they can from our personal data.

Hackers, too, are eager to get their hands on our data, whether it be personal account and credit card details or log-ins and passwords to company accounting systems. Indeed, company data is the holy grail for many of those who use the internet with nefarious purposes in mind.

This makes the recently revealed privacy breach at translate.com all the more alarming. In this case, hackers had no need to resort to phishing tactics or man-in-the-middle attacks in order to gain access to company data – the information was freely available on the internet for all to see. Read more

Old England

Oldest English Words

According to the BBC Reading University researchers have identified some of the oldest English words in the languages history.

‘I’, ‘We’, ‘Two’ and ‘Three’ are among the oldest known words, which could be thousands of years old.

The Researchers have created a computer model, which can analyse the rate of change of words. It can also predict which words will become extinct.

They believe “squeeze”, “guts”, “stick” and “bad” could become obsolete first.

The computer programme is designed to log a timeline showing how modern Indo-European words have changed over time. Students can use the software to look up any date and they can see which words were used at that time.
The researchers using the computer programme found that common words or words with precise meanings were more inclined to be the oldest and most long standing.

Basically, if you were able to go back in time (if you can that’s amazing you should tell someone about that!) Reading University could provide you with a pocket guide to the language of that time. This would enable you to communicate with English speakers throughout the ages.

This amazing piece of software can also travel forward in time and predict how words are likely to change in the future.

Does the T.A.R.D.I.S have this facility the Doctor might find this tool very useful.
Basically these guys have too much time on their hands; if they invented a tool to actually travel in time then I’d be impressed.

Words will change over time, it is inevitable. Kids make up words all the time, some stick, some don’t. New words are added to our dictionary now and again, but I don’t think old ones really disappear or become extinct. They will always be remembered somehow in books or multimedia programmes.

I guess words go out of fashion. It’s all swings and roundabouts really.

The Future of Translation

Ray Kurzweil is a pioneer in technological fields including speech recognition and text-to-speech synthesis. He also believes one day soon, computers will develop their own consciousness and superhuman levels of intelligence. In fact, in an interview with the Huffington Post, Kurzweil told industry guru Nataly Kelly that he believes computers will be able to translate as well as human translators by 2029.

At the moment, it’s obvious that machine translation has a way to go. But it’s improving all the time,and what if Kurzweil is right? Is there a future for translators in the brave new world he predicts? Fortunately, the answer is yes.  Kurzweil says that:

“These technologies don’t replace whole fields, in general. What they do is replace a certain way of applying them.”

The translation field is always changing, but translation companies that are willing to change with the times should have no problem thriving, according to Kurzweil:

“These tools are going to increase our ability to use, create, understand, manipulate and translate language. The idea is not to resist the tools, but to use them to do more.”

Read more

Google Translate

Why Machine Translation is Not Good Enough

Machine translation that’s good enough to substitute for human interpreters is like the great white whale, sought by science fiction writers, businesses and militaries alike. However, despite all the hype about the latest iPhone translation app and the ubiquity of Google Translate, nobody has yet managed to produce an algorithm that does the job as well as a bilingual human.

A recent article in Slate on the efforts of the US military to develop a machine translation device to substitute for human interpreters in Afghanistan is a case in point. The article describes the results of a 5 year research effort funded by DARPA. The snippet below shows just how well the device performed in place of a Pashto interpreter:

Rachel asked: “Would you introduce me to him?” Aziz failed to understand the machine’s translation (though he does speak English), so she asked again: “Could you introduce me to the village elder?” This time, there was success, after a fashion. Aziz, via the device, replied: “Yes, I can introduce myself to you.”

Unfortunately, Aziz was not the village elder in question. C3PO, where are you when we need you?  DARPA’s speech-to-speech translation system, called TransTac, achieved an 80% accuracy rate by the end of the research project. Obviously, these are not the droids we were looking for.

The problem, as Slate points out, is that computers are great at storing knowledge and making calculations, but they lack the key ingredient of a successful interpreter: understanding. Attempts to add this essential human ingredient by comparing machine translations to human-created translations and by having real people rate translations for quality also tend to fall short. Plus, it’s slow and expensive. As Slate writer Konstantin Kakaes put it:

The difficulty of knowing if a translation is good is not just a technical one: It’s fundamental. The only durable way to judge the faith of a translation is to decide if meaning was conveyed. If you have an algorithm that can make that judgment, you’ve solved a very hard problem indeed.

We couldn’t agree more, and that’s why it’s so important to use trained human translators and interpreters for important communications.

We recently published a further article regarding the pitfalls of MT, you can view it here: