A mobile app developed by a pair of former UCSF medical students could help doctors collect more accurate medical histories from patients, even if they don’t speak the same language. Called MediBabble, the free, donation-supported iPhone app lets doctors play common medical questions aloud in five languages: Cantonese, Haitian Creole, Mandaran, Russian, and Spanish. Patients can respond via yes/no answers, gestures, or scrolling on the device. Alex Blau, who began developin the app three years ago with fellow medical student Brad Cohn, says the app can help doctors get through the crucial medical history portion of a patient encounter faster and more accurately despite language barriers. “Ninety percent of diagnoses come from the patient’s self-reported medical history, so the ability to communicate is critical,” Blau says in a feature posted this week at UCSF’s website. “Time is not an asset doctors or patients have. You need that information when you need it,”

Comments | Reprints | Share:          

When science fiction fans think about language translation, they have two main reference points. One is the Universal Translator, software built into the communicators used by Star Trek crews for simultaneous, two-way translation of alien languages. The other is the Babel fish from Douglas Adams’ The Hitchhiker’s Guide to the Galaxy, which did the same thing from its home in the listener’s auditory canal.

When AltaVista named its Web-based text translation service after the Babel fish in 1997, it was a bit of a stretch: the tool’s translations were often hilariously bad. For a while, in fact, it seemed that the predictions of the Star Trek writers—that the Universal Translator would be invented sometime around the year 2150—might be accurate.

But the once-infant field of machine translation has grown up quite a bit in the last half-decade. It’s been nourished by the same three trends that I wrote about on Monday in the first part of this week’s series about Google’s vision of “augmented humanity.” One is the gradual displacement of rules-based approaches to processing speech and language by statistical, data-driven approaches, which have proved far more effective. Another is the creation of a distributed cloud-computing infrastructure capable of holding the statistical models in active memory and crunching the numbers on a massive scale. Third, and just as important, has been the profusion of real-world data for the models to learn from.

In machine translation, just as in speech recognition, Google has unique assets in all three of these areas—assets that are allowing it to build a product-development lead that may become more and more difficult for competitors to surmount. Already, the search giant offers a “Google Translate” app that lets an Android user speak to his phone in one language and hear speech-synthesized translations in a range of languages almost instantly. In on-stage previews, Google has been showing off “conversation-mode” version of the app that does the same thing for two people. (Check out Google employees Hugo Barra and Kay Oberbeck carrying out a conversation in English and German in this section of a Google presentation in Berlin last September.)

While still experimental, the conversation app is eerily reminiscent of the fictional Universal Translator. Suddenly, the day seems much closer when anyone with an Internet-connected smartphone will be able to make their way through a foreign city without knowing a word of the local language.

In October, I met with Franz Josef Och, the head of Google’s machine translation research effort behind the Translate app, and learned quite a bit about how Google approaches translation. Och’s long-term vision is similar to that of Michael Cohen, who leads Google’s efforts in speech recognition. Cohen wants to eliminate the speech-text dichotomy as an impediment, so that it’s easier to communicate with and through our mobile devices; Och wants to take away the problem of language incomprehension. “The goal right from the beginning was to say, what can we do to break down the language barrier wherever it appears,” Och says.

This barrier is obviously higher for many Americans than it is for others, present company included—I’m functionally monolingual despite years of Russian, French, and Spanish classes. (“It’s always a shock to Americans,” Google CEO Eric Schmidt quipped during the Berlin presentation, but “people actually don’t all speak English.”) So a Babel fish in my ear—or in my phone, at any rate—would definitely count as a step toward the augmented existence Schmidt describes.

But in the big picture, Google’s machine translation work is really just a subset of its larger effort to make the world’s information “universally accessible and useful.” After all, quite a bit of this information is in languages other than those you or I may understand.

The Magic Is in the Data

Given the importance of language understanding in military affairs, from intelligence-gathering to communicating with local citizens in conflict zones, it isn’t surprising that Och, like Cohen, found his way to Google by way of the U.S. Defense Advanced Research Projects Agency (DARPA). The German native, who had done masters work in statistical machine translation at the University of Nuremberg and PhD work at the University of Aachen, spent the early 2000s doing DARPA-funded research at USC’s Information Sciences Institute. His work there focused on … Next Page »

Comments (1) | Reprints | Share:          

Livemocha is ready to kick some language learning butt—at least according to chief executive Michael Schutzler. The Bellevue, WA-based startup is already the largest online language learning community in the world, with over six million members from over 200 countries actively studying some 38 languages—not bad numbers for a 3-year-old company. But this isn’t enough for Schutzler, who has his eyes set on beating popular language learning software company Rosetta Stone.

“Our mission is to do this for every language, and our goal is to get every single person on the planet to converse in multiple languages,” he says.

While this may seem like a pretty lofty ambition, given that there are literally thousands of languages spoken throughout the world, Livemocha does have one advantage—it already has an established global community. Even though the company is based here in the states, this isn’t a U.S.-centric company. Ninety percent of Livemocha users, according to Schutzler, are based outside of the U.S.

Livemocha is structured quite differently from its rival, Arlington, VA-based Rosetta Stone (NYSE: RST). Rosetta Stone offers CD-ROM courses in 29 languages (and a few more in varying dialects)—most of which come in multiple levels, and run anywhere from $219 to $249 a pop. Instead of selling retail software, Livemocha offers a series of free and paid lessons accessible online, where it connects learners with native speakers around the world in real time—what Schutzler calls the centerpiece that differentiates it from the competition.

Every month Livemocha’s 6 million members help to build out its community by creating more than a million speaking and writing activities for each other, reviewing each others’ work. Its users generate an average of 35,000 exercise reviews every day—and they practice real time conversation skills with other users in the network.

“At any moment in time half the community is contributing in an editorial fashion, and the other half of the community is learning,” Schutzler says.”We’ve created this transformative language experience that makes conversation the real language.”

Rosetta Stone’s CD-ROM can’t do that, he says.

Michael Schutzler

“The investment in Rosetta Stone is massive, but the average time its users spend on it is only 2.5 hours, over 6 months,” he says.

Why is this? Because learning a foreign language is “really hard work,” especially for adults. So when the going gets tough, many give up. Livemocha has found a way to marry language learning course curriculum with community interaction, and this, he says, is what makes the Livemocha system more effective. While Schutzler declined to comment on the retention levels of Livemocha’s users, he did say that the engagement is much more pronounced because of the social nature of the community. He himself logs on to tutor others in the community in English and German, and is currently learning Spanish, and “refreshing” his Arabic.

“At the heart of what we do is conversation practice, and the only hope we ever have of learning and mastering a language is talking with a conversation,” Schutzler says. The company even has a growing number of members who are starting to make a living on Livemocha, through points earned tutoring other members, grading coursework, and creating new practice exercises.

Livemocha currently has 38 free languages available on its site—everything from commonly spoken languages such as Spanish, English, and Hindi, to the more obscure Catalan, Icelandic, and Latvian. And yesterday it rolled out its more advanced paid service, Livemocha Active Courses, in five … Next Page »

Comments (1) | Reprints | Share:          

Seattle-based Amazon.com announced today that its Kindle Digital Text Platform can now be used by authors and publishers to upload their electronic books in Spanish, Portuguese, and Italian to the Kindle Store. The service is already available for English, French, and German books. Amazon (NASDAQ: AMZN) also said yesterday that its Kindle application is now available as a free download for BlackBerry devices. That extends the reach of Kindle e-books beyond Kindle devices, iPhones, iPod Touch, and PCs.

Comments (1) | Reprints | Share:          

Oren Etzioni, a computer science and engineering professor at the University of Washington, has certainly ventured out of the ivory tower since he first came to Seattle 18 years ago. The Israel-born computer scientist founded three startup companies out of UW: Netbot, a comparison shopping agent acquired by Excite in 1997, Clearforest, a text-miner acquired by Reuters in 2007, and Farecast, an airlines fare prediction tool acquired by Microsoft just last year.

He’s also a venture partner at Madrona Venture Group, a Seattle VC firm that funds technology startups. And he’s been involved as a consultant or advisor to numerous other startups and local businesses, most recently Eggsprout, a new Seattle startup by UW alums that combines social networking with job hunting.

I sat down with Etzioni in his office on the UW campus to chat about his philosophies on technology, startups, and investing in a tricky economy. He also showed me his latest projects—two new software technologies that search the Web in innovative ways.

PanImages, an image search tool, mines Google Images and Flickr for pictures—but the twist is, it works in hundreds of different languages (online translation tools currently don’t work for many of these). So you can not only perform image searches if you happen to speak a language that doesn’t have its own Google Images page, but you can also type in a query in one language and see its translations and corresponding images in other languages. I typed in “shoes,” and we looked at the top image hits for shoes in Italian (pointy), English (sporty), and Serbian (boots).

His other new software application, TextRunner, searches 500 million Web pages for relationships between words. You can type in a question like “What kills bacteria?”, and it finds everything in these 500 million pages that has the relationship “kill” to bacteria, returning answers ranked by the number of hits. So, antibiotics are at the top of the list with 304 hits, but down the list you find that “garlic” came back with seven hits. (More on these projects below.)

Here are edited excerpts from the rest of our conversation:

Xconomy: What do you think academia can bring to business?

Oren Etzioni: The primary mission of a professor is teaching and research, but I want to maximize my positive impact on the world, and I’ve learned over the years that there are lots of ways of doing that. I started out focused on research and writing research papers,and then found out that few people read those. Then I found that mentoring a graduate student can be a very meaningful way to have an impact. Another way to have an impact is that some research ideas have practical applicability, and the best way to get them out to the real world is through commercialization.

To give a concrete example, one of my graduate students, Erik Selberg, and I built the first meta-search engine, MetaCrawler. We were running it here and it became very popular. After a while, the head of the lab staff came to me and said, ‘This MetaCrawler thing is generating so much traffic on our network that people can’t get access to their homework, you have to get rid of it.’ So now I have this problem, I’ve created this thing that people love, so now what? The ideas were already out there in research papers, but we didn’t want to kill the application. We found that by licensing it to a startup company, we gave it some life and millions of people were using it every day at its peak. So it was very gratifying.

X: How do you decide if a research project you’re working on will have commercial applications?

OE: The first question I ask myself is, do people really badly want to use this? And I like to see evidence of that. In both the cases of MetaCrawler and Farecast, I was surprised how much … Next Page »

Comments (3) | Reprints | Share:          

“Language lessons,” says the Korean sportscaster in his Howard Cosell voice at the end of Better Off Dead. It’s what allowed Lane Meyer (played by John Cusack) to ski the K-12 on one ski, and 23 years later, it’s still the key to building relationships in global business. So when we heard on Friday that Livemocha, a Bellevue, WA-based online language-learning startup, had won the NWSource People’s Picks 2008 award for “favorite startup,” we were eager to find out more.

Livemocha beat out 10 other fellow finalists including AdaQuest, Evri, and Xconomy (yes, really—we have no idea who voted for us, but thank you). Last year’s winner was Zillow, so Livemocha is in good company. The NWSource writeup cites Livemocha’s “model of diversity, with employees from India, China, Japan, Iran and the Ukraine, to name a few,” and its “very robust Ping-Pong culture.” Evri, for its part, has a Ping-Pong table, but not a Ping-Pong culture, according to one employee. Here at Xconomy, we have none of the above, but we’re working on it.

But seriously, Livemocha has been steadily building its Web 2.0 e-learning service, which lets users around the world chat in their native language online and by video with people who want to learn foreign-language skills. The year-old startup, led by co-founder and CEO Shirish Nadkarni, has garnered lots of media attention and surpassed one million registered members in September. Last January, Nadkarni announced Livemocha had closed a $6 million venture round led by Seattle-based Maveron. (By coincidence, another investor is Sunny Gupta, whose new company Apptio we profiled today.) It all sounds like a perfect application for social networks, as long as the user interface is fun, reliable, and easy to use.

No word yet on whether I can find someone to teach me that Howard Cosell accent, but there are at least 11 major languages on offer…

Comments (1) | Reprints | Share:          

Being the leader of the pride, it seems, doesn’t guarantee you a nice meal everyday. Waltham, MA-based Lionbridge Technologies (NASDAQ: LIOX) may be the world’s largest provider of localization services, helping hundreds of other companies from Microsoft to Merrill Lynch to Pfizer go global by translating their websites, product manuals, software programs, and drug warnings into other languages. But it sure isn’t getting rich doing it: On May 6 the company reported a net loss of $4.4 million for the first quarter—its third straight quarterly loss—and last Friday Lionbridge stock dipped to $2.37 per share, its lowest point in more than five years, and down 61 percent compared to a year ago.

That’s lower than even the most pessimistic analysts were predicting at the beginning of 2008. As the financial website 24/7 Wall Street put it in a catty post recently, “It looks like the lion’s roar is a meow, at best.”

It’s a frustrating reversal for a company whose CEO, Rory Cowan, was being lionized (that’s the last cat pun, I promise) just three years ago for putting together a $180 million deal to buy one of its main competitors, the Global Solutions unit of New York-based Bowne & Company. That acquisition vaulted Lionbridge into the ranks of Boston’s 60 largest companies. And it’s a puzzling state of affairs, given that the company’s revenues are large and growing—a record $117 million in the first quarter, up more than 8 percent from a year earlier. But no matter how high the company’s revenues go, its expenses seem to go higher.

One answer to the puzzle seems to be that Lionbridge’s expansion has put it at the mercy of the same twin forces—technology and globalization—that drive demand for its services.

Analysts who follow localization services—a $12 billion industry in 2007—point out that translation is still a labor-intensive process, and that Lionbridge’s strategy of employing hundreds of managers, engineers, and translators in expensive regions like Europe may be backfiring with the weakness of the U.S. dollar against the Euro and many other currencies.

And while Lionbridge has made a significant investment in technology—especially on work-flow automation and memory systems that save labor by identifying material that’s already been translated—it hasn’t really profited from that spending. “What has happened is that the benefits of the efficiency gain that Lionbridge has been able to produce through the use of technology have been handed right to the client” in the form of prices that have stayed flat despite global inflation, says Ben Sargent, content globalization strategist at Common Sense Advisory, a localization market research firm in Lowell, MA.

It wasn’t supposed to be this way. Lionbridge, founded in 1996, has spent hundreds of millions of dollars on acquisitions and other growth strategies, and has the client list (Bayer, Cisco, DuPont, GE, Google, IBM, Merck, Merrill Lynch, Microsoft, Morgan Stanley, Nokia, Pfizer, Sony, and Wal-Mart are a few of the big names) and revenues ($452 million in 2007) to show for it. With 4,600 employees in 45 offices around the world and a network of 25,000 freelance translators skilled in more than 200 languages, the company is three times the size of … Next Page »

Comments (11) | Reprints | Share:          

Lionbridge (NASDAQ: LIOX), a Waltham, MA, “localization” company that translates English-language software, software manuals, online support materials, websites, and other product documentation into other languages, said today that it plans to buy back $12 million worth of its own common stock in 2008. That’s on top of $12 million in stock already purchased since the company’s board authorized the repurchase program in September 2007.

Comments | Reprints | Share: