Showing posts with label Emerging Tech. Show all posts
Showing posts with label Emerging Tech. Show all posts

Tuesday, May 6, 2014

How the 'One Microsoft' mission is changing Microsoft Research

Summary: Some of the sweeping cultural changes affecting Microsoft are having an impact on the company's research arm, as well, says MSR chief Peter Lee.


Cultural changes at Microsoft over the past year-plus are reverberating throughout the company, including across the company's 1,100-plus-strong Microsoft Research arm.

Since last July's "One Microsoft" reorganization," things have gotten friendlier" between different teams at the company, including Microsoft Research (MSR) said Corporate Vice President and MSR chief Peter Lee. "There's more of a communal feel."

I had a chance to chat with Lee during an open house at Microsoft Research New York's new digs on May 5. The open house was held in honor of the two-year anniversary of the New York lab, which Microsoft created around a core of former Yahoo researchers that the company hired.

The MSR New York office is now home to about 20 researchers. MSR New York has a particular focus on social media and social science research, along with machine learning and big data.

In the not-so-distant past, some product teams at Microsoft had to be coaxed to consider incorporating MSR-developed technologies and techniques in their commercial products. TechFest, Microsoft's annual internal research fair, had become a place for researchers to try to pitch product teams to pick up their technologies.

Things are different now, said Lee, who became the head of MSR worldwide last July. Many product groups inside the company are clamoring to adopt technologies pioneered inside MSR. The Windows Phone team and Xbox teams, in particular, have been at the forefront of taking MSR technologies in building them into commercialized entities, such as the Cortana digital assistant in Windows Phone 8.1 and the Kinect sensor. MSR also has contributed to more enterprise-focused products, including the PowerBI business-intelligence and Oslo Office 365 social-feed app from the Office team.

When I asked whether Microsoft CEO Satya Nadella's "mobile first/cloud first" mantra was influencing the types of research projects MSR is undertaking, Lee was quick to note that "MSR management doesn't tell (researchers) what to do."
That said, "We all want to be part of a winning team," Lee acknowledged. And the company's product teams, with their mobile first/cloud first marching orders, "are welcoming research with open arms," these days, he added.

More "open arms" seemingly has translated into more Microsoft research projects becoming commercialized and/or part of commercialized products. But Lee said he doesn't want to err on the side of MSR becoming nothing but an extension of the company's product groups. 

"We still need to be wackier and more forward-looking.... We are not just about product development. We are looking beyond the headlights," he said. 

An example of MSR embracing a longer-term view is Microsoft's recent hire of Norman Whitaker, who most recently was the Deputy Director of the Information Innovation Office at DARPA, the U.S. Department of Defense's research arm. Whitaker previously was the Program Manager for the DARPA Urban Challenge program, which encompasses self-driving vehicles, as well as other kind of automated bots. (Lee himself formerly worked at DARPA before joining Microsoft.) At Microsoft, he will be heading up a new "Special Projects" team in MSR focusing on so-far unnamed disruptive technologies. MSR Special Projects is Microsoft's version of Google X, sources of mine have said. 

Lee is also personally committed to being more open and forthcoming about MSR's work, he said. That's another shift from the philosophy of some former members of MSR management. 

In the new One Microsoft, "we now have to work with everyone. If there are secrets (between product teams) at Microsoft, it makes things harder for us," Lee said. 

He also noted that 85 percent of MSR's work is done in conjunction with academic researchers who are intent on seeing their research published. That requires a level of openness and transparency. (MSR doesn't have to reveal all its secrets to us blogging sleuths and/or competitors. For example, MSR may opt to publish information about a newly developed algorithm without disclosing how, when and where it will be used by a Microsoft product group.) 

One more cultural change impacting MSR that I found interesting is on the cross-platform front. In recent months, Microsoft has been releasing more and more software and services aimed to work not just with Windows, but also iOS, Android and Linux. Lee acknowledged that the cross-platform emphasis is also happening in MSR, too. 

"We're looking at more cross-platform, and embracing competing OSes," he said. "There's more of an interest in exposing our services on other operating systems" these days. The MSR "Drawbridge" library OS work is just one example of investigations in this vein. 

The faster delivery cadence adopted by the Windows, Office 365 and other Microsoft teams is something in which MSR is involved, too. The idea of "test flighting" multiple features and updating services even more frequently than once daily are areas the implications of which are on the radar screen of research, too, Lee said.