[Published in the March 2001 issue of CRN as part of our series on member labs and centers]

Microsoft Research

By Daniel T. Ling
Daniel Ling is vice-president of Microsoft Research Redmond at Microsoft Corp. He was one of the founders of the Redmond laboratory, and served as its Director from 1995 to 2000. Dan Ling holds a B.S., M.S. and Ph.D. in Electrical Engineering from Stanford University. Microsoft Research is a sponsor of CRA's Outstanding Undergraduate Award program.

In 1991, when Microsoft was still a fairly small company and many businesses were scaling back their research efforts, Microsoft made a strong commitment to basic research in computer science by establishing the first research lab by a software company.

It reflected the recognition that the age of the microprocessor had only just begun and that improvements in silicon, storage, and communications technology, combined with the right software, would continue to dramatically transform every aspect of our lives. Over the past ten years, the lab has grown substantially and has contributed to almost every major product at Microsoft. We now have more than 500 researchers working on a wide range of projects–from quantum computing to operating systems, from programming languages and tools to signal- processing and speech recognition.

Microsoft Research (MSR) maintains four laboratories scattered across the globe: our main facility near the Microsoft corporate campus in Redmond, Washington, and facilities in San Francisco, Cambridge (United Kingdom), and Beijing. The Bay Area Research Center (BARC) in San Francisco was started in 1995. An exciting project from BARC is TerraServer (see http://www.terraserver.microsoft.com), one of the world’s largest online databases containing about 3TB of satellite, aerial photography, and topographic maps from around the world. This effort involved close cooperation with the U.S. Geological Survey (USGS), which provided the majority of the data for the United States. For their work, Jim Gray and Tom Barclay were recognized by USGS with the John Wesley Powell award.

Two years later, in 1997, we started our lab in Cambridge, which maintains close ties with Cambridge University. The Cambridge Lab conducts research in a number of areas, including programming languages, computer vision, learning, and systems. Their work on information retrieval has made important contributions to Microsoft products.

In 1998 we founded the lab in Beijing, which collaborates with many universities and research institutions in China and elsewhere in Asia. The Beijing lab focuses on two main areas: multimedia, and making computers vastly simpler for speakers of Asian languages. An exciting research result to come out of Beijing is a fundamental understanding of the sampling issues in the new image-based rendering approach to computer graphics.

MSR has a goal to research technologies that we believe will drive changes in computing over the next decade. As an industrial laboratory, our key focus is to work closely with the Microsoft development groups and support the corporate vision of "empowering people through great software anytime, any place, and on any device." Over the past decade we have made many contributions to Microsoft products and services.

Our research in data-mining and databases has resulted in a number of contributions to SQL Server, including novel self-tuning components in SQL Server that allow SQL to adapt automatically to changing query workloads. The system uses the actual workload and the query optimizer to perform "what-if" analysis on new configurations, suggests changes to the user, and can implement them automatically if desired. Our work with the SQL Server team also resulted in new data-mining features built into the database, including algorithms from MSR, a query language for data-mining, and a programming interface for third-party components.

We are also making significant contributions in the natural language area, working closely with our product group partners. This has resulted in the grammar-checking capabilities that first appeared in Word95, as well as new features in the upcoming version of Office. Karen Jensen and George Heidorn, who founded the Natural Language Processing group, were among the earliest researchers to join MSR, even before its official creation. The group is now undertaking an ambitious project to create a rich semantic network to support more complex natural-language-understanding tasks such as machine translation.

Two years ago, MSR started the Programmer Productivity Research Center led by Amitabh Srivastava. We recognized the need for additional research on the technology, tools, and processes for software development to improve the quality of software and the productivity of programmers. This group is now a major focus of our work. Tools from this group have been deployed throughout the company and are being used routinely by the major development groups.

Recently, Microsoft announced a new strategic direction called .NET, which combines several exciting directions. First it envisions a new generation of the Web, turning it into a distributed programming environment. Websites and web servers will no longer just provide data to a browser, but will provide services to distributed applications via XML. .NET also emphasizes new application areas, including rich support for communities, communications, and collaboration. Finally .NET involves providing access to services and information from a whole range of devices, from the PC and PDAs to game devices and cellular phones.

MSR was involved in defining .NET and providing important technology in several areas such as communications and collaboration. Eric Horvitz, one of our senior researchers in the Adaptive Systems and Interaction group, has been particularly interested in building intelligent systems that acknowledge the scarcity of human attention. All of us suffer from information overload from a variety of communications and information sources, such as e-mail, voice-mail, and computer-generated notifications. If the system models the urgency and importance of the notification as well as the context of the user (is the user busy, on the road, in a meeting, etc.), the system could intelligently prioritize notifications and choose the delivery mechanism¾ instant messaging, e-mail, or cellular phone, for example.

We are also working on understanding how to improve telepresentations, and perhaps make them even superior to live lectures. The ability to provide effective distance learning, offered anytime and any place, is clearly of great interest to educators. We would like to transmit presentations to a remote audience while still engaging the audience, and give the remote audience the ability to participate with each other and the lecturer. We are also working on capturing the presentation for later viewing, but having it behave as a living document that continues to evolve via annotations posted by all the viewers. To determine the effectiveness of a variety of techniques, we have deployed prototype software for internal Microsoft training programs.

Another researcher, Marc Smith, is a sociologist who is studying what he calls "social cyberspace." He defines this as the place where people interact. He is trying to understand the social protocols in online communities, and how social mechanisms can be supported with technology in order to create an effective and self-sustaining cyber-community.

In addition to contributing to Microsoft, another important goal for MSR is to be an active participant in the international research community. To this end, our researchers are encouraged to serve as editors of journals, to serve on national and international panels and advisory boards, to work on conference program committees, and to give talks throughout the world. We also work closely with the university community. One of our researchers, David Salesin, serves as an adjunct faculty member in computer graphics at the University of Washington where he teaches and supervises students. He has done landmark work in the area of non-photorealistic rendering, including his computer-generated pen-and-ink illustrations and computer-generated watercolors. David received SIGGRAPH’s Computer Graphics Achievement Award last year.

We have also worked with other research centers and universities on a project called the Multi University Research Laboratory (MURL—http://murl.microsoft.com). This is a publicly accessible website that features computer science lectures on demand. These lectures were captured and contributed by each of the partner institutions: Carnegie Mellon University, the Massachusetts Institute of Technology, Stanford University, the University of Washington, Microsoft Research, and the Xerox Palo Alto Research Center.

Microsoft gave us a unique opportunity to create a new computer science research lab from the ground up. Given this opportunity, we focused on attracting outstanding people, creating an open and collaborative research environment with a research agenda driven by the researchers, and providing them with the opportunity to improve the lives of millions of people through Microsoft products. This has proved to be an incredibly exciting and stimulating combination.

<< back to Labs and Centers Articles homepage