Microsoft’s Future Factory Shows Off its Latest Ideas

In Rick Rashid’s eyes, the biggest reason to have a huge corporate research division isn’t to feed a conveyer belt with fancy new products. Instead, the head of Microsoft Research says, basic research serves as more of an insurance policy, something to keep a company alive in a constantly evolving industry.

“We’re really here to make sure that Microsoft will be here 10 years from now, 15 years from now, or 20 years from now. If you think back 20 years ago when Microsoft Research was started, very few of the companies that were Microsoft’s peers at that time still exist today,” Rashid says. “You need to constantly be able to change, and having a fundamental, basic research group like Microsoft Research gives Microsoft that ability to be agile.”

It’s pretty easy to dispute the idea that Microsoft has been particularly nimble in dealing with disruptions in the tech industry. It’s presently spending billions in an attempt to thwart Google’s dominance in search and other Web-based services, and—like most other big companies—didn’t have the foresight to jump on the mobile computing revolution that turbocharged Apple into a mind-bogglingly valuable company.

But that doesn’t mean Rashid’s argument for investing in corporate research is wrong, necessarily. The Kinect, for example, is rightly touted as a home run that sprung out of Microsoft’s investment in R&D—just ask any other video-game console maker if they’d like to have that one back.

On Tuesday, I visited Microsoft’s campus to get a glimpse of some more wave-of-the-future projects being envisioned by Microsoft researchers from around the world. It’s part of a multi-day internal trade show called TechFest, which lets outsiders and Microsofties in to tour a long line of booths showing off cool new ideas.

One thing you noticed right away was the prevalence, at least in this preview screening, of devices using some form of the Kinect motion-sensor—Microsoft has clearly been pushing hard for new uses and ideas to combine one of its most notable successes. But more broadly, there were interesting ideas that blend several disciplines of computer science and engineering, and several that sought to make the tools or results more human in nature.

“Researchers tend to focus in a particular area, and people who are good at doing [user interface] design tend not to be the people who are good at doing back-end cluster computing. There’s just different specialties,” says Roy Levin, director of Microsoft’s Silicon Valley research center. “And yet, these fields have sort of independently advanced enough that people can see how to reach across and work together, and bring them together.

There were far too many cool projects to list here, but I’ve gathered are a few highlights from the day, ranging from ready-to-use consumer toys to heavy-duty research tools.

Cliplets blends digital videos with still photos, freezing some parts of the image as background while enabling other chunks to move around through the static landscape. The effect is surprisingly quite compelling—the resulting images can range from fun and amusing to eerie, dreamy, and cinematic—kind of like a memory.

They reminded me of like those ghostly paintings in the “Harry Potter” movies where characters in the portrait come alive. The service is available now, but only works on Windows 7, at least presently—I couldn’t try it out on my Macbook, sadly. Check out the Microsoft Research project page to get the full effect, which includes looped versions that make the grandma never stop sipping her tea, for instance.

—The Microsoft Translator Hub combines machine learning and crowdsourcing to help decode human languages. Researchers or businesses can use the system to build their own translation service, which Microsoft hopes will help ramp up the number of languages that are able to be translated with software.

The Translator Hub needs documents that are already written in two languages to properly start learning how to translate—Microsoft researchers say it takes around 100,000 sentences to train the software, which also can get a handle on industry-specific jargon, for instance.

Once those documents are all processed and matched up by the software, human users can take over. They review its accuracy and add to the machine learning by cleaning up and correcting suggested translations in a Web interface.

Rashid described the vision for the Translator Hub as leading to “a democratic and egalitarian world out there, where there isn’t just one organization deciding which languages get translated, but in fact a large community of people.”

There are around 40 languages available in the system already, which leaves more than a thousand possibilities. Using the Translator Hub is free. [Updated 3/9/12 to clarify that use of the Hub is free.]

—Under the banner of new search experiences, researchers from Cambridge, England are thinking of how to show users how a body of information online grows and changes over time. They’ve picked some organic metaphors, “seeds” and “branches.”

These new search metaphors instead display results on a given topic that someone might monitor as a beautiful, changing, plantlike image that sprouts new additions when more information enters the discussion.

The idea is that many Web users aren’t looking to access the information online in the way it’s commonly presented—a long list of text links, with some images. Instead, we tend to monitor topics we’re really interested in nearly constantly. And, unless it’s hot news, fresh information might not always find its way to the top of search results because it doesn’t rank particularly well yet.

“Web use is much more about doing particular tasks,” Microsoft’s Sian Lindley says. “So, often it’s about warming up in the morning by checking the news, and reading email, and settling in at work. Or it might be about taking quick breaks from what you’re doing at the moment.”

This is still very much in the development phase (so don’t look for it on Bing anytime soon). But the idea seems quite appealing—instead of getting pinged with an e-mail alert or clicking through headlines in an RSS reader, you could take a look at your digital news plant, and see what new stuff has bloomed.

Author: Curt Woodward

Curt covered technology and innovation in the Boston area for Xconomy. He previously worked in Xconomy’s Seattle bureau and continued some coverage of Seattle-area tech companies, including Amazon and Microsoft. Curt joined Xconomy in February 2011 after nearly nine years with The Associated Press, the world's largest news organization. He worked in three states and covered a wide variety of beats for the AP, including business, law, politics, government, and general mayhem. A native Washingtonian, Curt earned a bachelor's degree in journalism from Western Washington University in Bellingham, WA. As a past president of the state's Capitol Correspondents Association, he led efforts to expand statehouse press credentialing to online news outlets for the first time.