Microsoft Research Redmond: 2012 in Review

Microsoft Research Redmond: 2012 in Review

  • Comments 1
  • Likes

Posted by Eric Horvitz and Yi-Min Wang, managing co-directors of Microsoft Research Redmond

Microsoft Research 2012 in Review logo 

As we look back on the year at Microsoft Research Redmond, a flood of creative efforts and achievements come to mind. These include mission-focused pursuits aimed at solving urgent challenges, the pursuit of new understandings at the foundations of computer science, and blue-sky initiatives exploring new possibilities. Notable developments, honors, and influences are far too numerous to include in a short blog post, so we can touch on only a small subset of representative milestones.

On the foundations front, a stunning set of experiments provided evidence for an elusive particle named the Majorana fermion. A team at the Delft University of Technology, led by Leo Kouwenhoven, used an experimental setup proposed and funded by our Station Q. Majoranas have been proposed as central in enabling an approach to quantum computing being pursued at Station Q.

Moving from quantum computing to the more familiar world of Turing and von Neumann, Flat Datacenter Storage won the MinuteSort record hands-down. Stunning advances in real-world program synthesis for end users were revealed via a feature in Excel named FlashFill, heralded by CNN Money as “Excel 2013’s coolest new feature.” Moving beyond the excitement of a feature in Office, the work reveals only a glimmer of a visionary path forward for achieving “programming by demonstration.” In the realm of enterprises and data centers, efforts led to the fielding of Windows Server 2012 Primary Data Dedup and Hekaton for SQL Server. A new approach to erasure coding called Local Reconstruction Codes won a best-paper award and went into production for Windows Azure, significantly reducing storage costs. New platforms developed at our lab for performing machine learning and fielding predictive models matured into tools now used across Microsoft. And code-verification tools from our Research in Software Engineering teams continued to see use throughout Microsoft, including application in verifying the ARM code in Windows RT.

Tentacles of innovation from Microsoft Research Redmond reached into multiple user experiences, with some recent advances now in the hands of consumers. For example, designs for sensing orientation from Microsoft Research enabled the Forza Horizon Wireless Speed Wheel for Xbox. Predictive methods for assisting with text input called Word Flow have been incorporated into Windows Phone 8. Numerous other developments explored new directions in computing, from compelling to exotic, including efforts in the Humantenna project, the debut of the Printing Dress during CES and SXSW, and the release of Cliplets. Many visionary prototypes were developed, including Illumishare and Holoflector.

Yi-Min WangNumerous projects addressed societal, health, and environmental challenges. As an example, MT Hub, a collaborative effort from our Microsoft Research Connections team, provides a crowdsourced approach to translation, with the ability to capture and preserve rare languages such as Hmong Daw and Mayan tongues. Predictive models for challenging patient outcomes developed by Microsoft Research were deployed in hospitals throughout the world via Caradigm, a joint venture in health-care computing between Microsoft and GE Healthcare, created with guidance from Microsoft Research.

Outreach, collaboration with colleagues, and sharing continued on multiple fronts. In one development, Microsoft Research’s release of the source code for the Z3 automated theorem prover was greeted with great enthusiasm by the computing community. Researchers across Microsoft Research labs continued intensive engagements in organizing conferences and workshops, as well as with dedicated service with panels and committees for NGOs and government agencies. As an example, Peter Lee of Microsoft Research Redmond  co-authored the 2012 President's Council of Advisors on Science and Technology review of Federal investments in networking and IT research, which included an effort on updating the “tire tracks” visualization of the influence of computer-science research on industry and society. Multiple awards and citations celebrated the long-term contributions by Microsoft researchers and engineers, and the lab received numerous best-paper awards and two test-of-time awards for contributions that shine from a 10-year perspective.

On another front, we worked to add new sparkle to our physical surroundings, lighting up the atrium in Building 99 with artistry and technology. Our gallery team created a persistent art space in our lobby featuring the creative talents of people at Microsoft Research. And several prototypes harnessing machine perception and learning were deployed into operational service within the building.

Eric HorvitzIt was a great year for recruiting top computer-science talent at Microsoft Research Redmond and, more broadly, at Microsoft. New hires included strong talent in the algorithms and systems areas. In a collaborative effort, the Redmond lab worked closely with its sister labs to found Microsoft Research New York City, attracting key new talent to Microsoft Research at a new hub of activity in a bustling city.

At the end of 2012, we had a change in leadership at the Redmond lab. After two years of inspirational service as lab director, Peter Lee moved into a new role as head of Microsoft Research USA, overseeing the U.S. labs. Deputy managing directors Eric Horvitz and Yi-Min Wang were asked to serve as managing co-directors of Microsoft Research Redmond.

Looking forward to 2013, we see strong innovation continuing on a portfolio of research efforts aimed at achieving results along a spectrum of time horizons. We’ve talked about Kinect as an example of the flow of ideas from long-term, open research and mission-focused goals coming together to create magic. Another example is the cross-lab collaboration on a rich synthesis of efforts in machine learning, speech recognition, machine translation, and text-to-speech that enabled the awe-inspiring demonstration of speech-to-speech translation in Tianjin, China in October—with Rick Rashid, Microsoft chief research officer, speaking in English and the audience hearing his words almost flawlessly translated in his own voice into Mandarin. We’ll continue to embrace a diversity of approaches to ideas and innovation as a strategy for advancing the frontiers of computer science and for enhancing Microsoft’s products and services. We know that 2013 will be even more exciting, and we can’t wait to see where things head!

Your comment has been posted.   Close
Thank you, your comment requires moderation so it may take a while to appear.   Close
Leave a Comment
  • Very good, ideas and innovation as a strategy for education.