Microsoft Research Invests in the Future of Computing

REDMOND, Wash., May 25, 2005 — Among early-career professors in the computer-science arena, intriguing intellectual pursuits are rarely in short supply. When these young academicians aren’t in the classroom teaching, they can often be found in their university’s research lab, exploring the far reaches of image synthesis, complexity theory, language processing, swarm robotics or bioinformatics. But the area where newly minted academic researchers typically do encounter shortages is funding; without a track record, they can find it difficult to attract the financial support that discipline-stretching research demands.

The dilemma — faced by those entering the teaching and research ranks of universities everywhere — is precisely what Microsoft Research intends to address with its New Faculty Fellowship Awards Program. Just launched by the University Relations group at Microsoft Research, the program aims to support and recognize first-, second- and third-year faculty members who are pushing computing research in novel directions. The program has just announced its five winners for 2005 [see list], awarding US$200,000 cash grants to a handful of young professors who demonstrate exceptional creativity, the potential for high-impact, innovative research, and the likelihood of becoming leading thinkers in their field.

Frédo Durand, an assistant professor at the Massachusetts Institute of Technology, fits this bill. Along with teaching computer graphics and computer science at MIT, Durand pursues advanced research in realistic image synthesis and computational photography. As one of five recipients of a 2005 New Faculty Fellowship, Durand says he feels honored to have survived a rigorous selection process and the scrutiny of distinguished colleagues. He predicts that the prestige associated with the fellowship will help him recruit students and raise funding for his research.

“Microsoft Research’s support will provide more financial independence and allow me to support research that is more daring and might not have obvious short-term applications,” Durand says. “Blue-sky research is crucial for sustainable scientific and technological progress, and I am glad that Microsoft Research supports this view.”

Durand’s complex research is meaningful to him because it offers him a way to explore how humans perceive their visual environment, and what makes a picture compelling. Durand says he loves doing research in computer graphics because it’s an interdisciplinary field that provides a “wonderful opportunity” to broaden his horizons in math, physics, psychology and the visual arts. But he is quick to add that his research also promises value for a broader group.

“Beyond the computer graphics community, I think that the impact of my work will lie at the interface between vision and graphics,” Durand explains. “The techniques and theoretical models we build for image synthesis also shed insights on the corresponding inverse problems. Moreover, our work on computational photography is at the convergence of computer vision and computer graphics, with exciting cross-fertilization between these two fields.”

Similar excitement for advanced research spills out from Wei Wang, an assistant professor in the computer science department at the University of North Carolina, Chapel Hill. Another recipient of a Microsoft Research New Faculty Fellowship, Wang fills her UNC teaching schedule with graduate courses in datamining and bioinformatics; in fall 2005, she will teach a new undergraduate course on bioalgorithms as well.

But as with Durand, research is Wang’s true passion. One of her recent projects involves developing algorithms to find structural patterns in protein databases. Wang explains that current technology allows biologists to identify protein structures at an accelerating rate — far faster than it is possible to determine their function. Her research allows scientists to compare and identify common substructures of protein families, which are indicative of their function. These discovered patterns can then be used to classify and provide hypotheses for the functions of unknown proteins.

“Proteins are the building blocks of life,” Wang says. “Understanding their function is essential to curing diseases as well as understanding the many mysteries of life.”

Wang hopes that in the long term, her work will have a positive impact on the larger community of academic computer science.

“Among early-career professors in the computer-science arena, intriguing intellectual pursuits are rarely in short supply. When these young academicians aren’t in the classroom teaching, they can often be found in their university’s research lab, exploring the far reaches of image synthesis, complexity theory, language processing, swarm robotics or bioinformatics. But the area where newly minted academic researchers typically do encounter shortages is funding; without a track record, they can find it difficult to attract the financial support that discipline-stretching research demands.

The dilemma — faced by those entering the teaching and research ranks of universities everywhere — is precisely what Microsoft Research intends to address with its New Faculty Fellowship Awards Program. Just launched by the University Relations group at Microsoft Research, the program aims to support and recognize first-, second- and third-year faculty members who are pushing computing research in novel directions. The program has just announced its five winners for 2005 [see list], awarding US$200,000 cash grants to a handful of young professors who demonstrate exceptional creativity, the potential for high-impact, innovative research, and the likelihood of becoming leading thinkers in their field.

Frédo Durand, an assistant professor at the Massachusetts Institute of Technology, fits this bill. Along with teaching computer graphics and computer science at MIT, Durand pursues advanced research in realistic image synthesis and computational photography. As one of five recipients of a 2005 New Faculty Fellowship, Durand says he feels honored to have survived a rigorous selection process and the scrutiny of distinguished colleagues. He predicts that the prestige associated with the fellowship will help him recruit students and raise funding for his research.

“Microsoft Research’s support will provide more financial independence and allow me to support research that is more daring and might not have obvious short-term applications,” Durand says. “Blue-sky research is crucial for sustainable scientific and technological progress, and I am glad that Microsoft Research supports this view.”

Durand’s complex research is meaningful to him because it offers him a way to explore how humans perceive their visual environment, and what makes a picture compelling. Durand says he loves doing research in computer graphics because it’s an interdisciplinary field that provides a “wonderful opportunity” to broaden his horizons in math, physics, psychology and the visual arts. But he is quick to add that his research also promises value for a broader group.

“Beyond the computer graphics community, I think that the impact of my work will lie at the interface between vision and graphics,” Durand explains. “The techniques and theoretical models we build for image synthesis also shed insights on the corresponding inverse problems. Moreover, our work on computational photography is at the convergence of computer vision and computer graphics, with exciting cross-fertilization between these two fields.”



Fredo Durand

Similar excitement for advanced research spills out from Wei Wang, an assistant professor in the computer science department at the University of North Carolina, Chapel Hill. Another recipient of a Microsoft Research New Faculty Fellowship, Wang fills her UNC teaching schedule with graduate courses in datamining and bioinformatics; in fall 2005, she will teach a new undergraduate course on bioalgorithms as well.

But as with Durand, research is Wang’s true passion. One of her recent projects involves developing algorithms to find structural patterns in protein databases. Wang explains that current technology allows biologists to identify protein structures at an accelerating rate — far faster than it is possible to determine their function. Her research allows scientists to compare and identify common substructures of protein families, which are indicative of their function. These discovered patterns can then be used to classify and provide hypotheses for the functions of unknown proteins.



Wei Wang

“Proteins are the building blocks of life,” Wang says. “Understanding their function is essential to curing diseases as well as understanding the many mysteries of life.”

Wang hopes that in the long term, her work will have a positive impact on the larger community of academic computer science.

“A personal goal of my research and teaching is to excite future computer scientists about the immense opportunities made available by applying their talents to biological problems,” she says. “Similarly, I hope to provide biologists with an appreciation for the incredible problem-solving advantages provided by computation methods.”

To Wang, winning a New Faculty Fellowship means the freedom to pursue her research on a significantly larger scale than would otherwise have been possible. She predicts that the Microsoft Research fellowship will also allow her to pursue non-traditional collaborations with researchers in genetics, biochemistry and pharmacology.

Collaboration is an equally appealing prospect for Subhash Khot, another winner of a Microsoft Research New Faculty Fellowship. Khot is a first-year faculty member in the College of Computing at Georgia Institute of Technology, where he teaches courses in algorithms and computational complexity.

“The fellowship gives me an opportunity to collaborate with the world-class research groups at Microsoft Research,” Khot says. Khot looks forward to visiting Microsoft Research and furthering his work in theoretical computer science. “Though a computer has become a part of our everyday life, we are far from understanding the fundamental nature of computation,” he explains. “I like to address this question through my research.”

Benefits Include Cash and Collaboration



Subhash Khot

According to Rick Rashid, senior vice president of Microsoft Research, Wang and Khot weren’t the only New Faculty Fellowship Award recipients who cited collaboration as a key part of the prize.

“All of the finalists stated that a relationship with Microsoft Research has tremendous value for them,” Rashid says. “The potential for collaboration has mutual benefit for us. We see huge value in finding fresh new perspective and people who can interact with our researchers.”

Sailesh Chutani, director of University Relations at Microsoft Research, adds that the New Faculty Fellowship Award Program was founded on the concept of novel approaches and unorthodox thinking.

“We don’t know where the next great idea is going to come from,” Chutani says. “By giving people the intellectual and creative freedom to pursue their vision, who knows where we can push the boundaries?”

The New Faculty Fellowship Award Program grew out of conversations among Microsoft Research, its faculty advisory board members and others in academia, and is modeled after a similar Microsoft Research program that awards fellowships to university graduate students in computer science, mathematics and electrical engineering. The new program was also prompted by the recent decline in research funding, which makes it especially difficult to secure backing from traditional resources for more “speculative” work.

Maria Klawe, dean of Engineering and Applied Science at Princeton University, recognizes the need to fill this gap. Klawe, who helped judge the Microsoft Research New Faculty Fellowship Awards, credits the program for supporting young faculty members who have an exciting research vision. She notes that financial support in the field of computing is especially crucial today as government and industry cut back on their funding of fundamental research.

“These awards will be of immense value to the candidates and to the field,” says fellow judge Robert Constable, Dean of the Faculty of Computing and Information Science at Cornell University. “The candidates will receive more attention to their needs and increased visibility for their results. The quality of their work will reflect well on computer science as a discipline.”

Microsoft Research plans to award five new faculty fellowships annually. To address the under-representation of women in computing, the program has set a goal of awarding at least two of the New Faculty Fellowships per year to qualified female applicants.

Besides the cash grant of $200,000 (issued in two $100,000 yearly installments) and the opportunity to collaborate with Microsoft Research scientists who work in similar areas, fellows will be invited to give presentations at this year’s Microsoft Research Faculty Summit in July. Fellowship recipients also get extensive access to other resources, such as Microsoft-sponsored events for skills development, MSDN (Microsoft Developer Network) subscriptions and technical publications.

Rigorous Selection Process

New faculty members who vied for the inaugural Microsoft Research fellowships found their work cut out for them. The call for nominations yielded 110 applications, spanning a range of technologies that in many cases, but not all, mapped to the 55 research areas Microsoft invests in. Chutani notes that these applicants were already among “the best and the brightest,” because the program stipulated that each university could nominate only one candidate through its provost’s office.

The 110 applications were reviewed by a pool of 30 Microsoft Researchers, who weighed nominees against criteria such as creativity, innovation, stimulating work and the novel nature of their research to produce 20 semifinalists. The semifinalists then underwent review by a prestigious group of second-round evaluators, including Princeton’s Klawe and Cornell’s Constable. The 11 finalists selected by this panel then came to Microsoft headquarters in Redmond for in-person interviews and crisp five-minute presentations of their work. By the time this process yielded the five winners, Klawe and Constable found themselves impressed by the outstanding research accomplishments of the candidates and their ability to articulate a compelling vision for their future work.

“It is clear that some of these candidates have already had a major impact on their fields in just one or two years beyond their Ph.D. theses,” Constable notes. “This is extraordinary. All of them have done exceptional research.”

Describing the five winners specifically, Klawe says, “They are truly inspired and inspiring about their work. They have a strong vision for what they can do to revolutionize their field … and the potential to succeed.”

And the Other Winners Are…



Radhika Nagpal

Radhika Nagpal, who began teaching at Harvard University in September 2004, was among those who rose to the top in this rigorous selection process. Nagpal teaches an undergraduate programming course as well as a graduate course on biologically-inspired approaches to distributed systems and multi-agent systems. The latter corresponds to her research interest, which lies in engineering self-organizing and self-repairing systems, and in better understanding robust collective behavior in biological systems.

Nagpal explains that the current focus of her research is on programming paradigms for embedded homogeneous distributed systems, such as sensor-actuator networks, self-reconfigurable modular robots and swarm robotics. The techniques that her research group employs draw on concepts from embryo development, which were proposed by biologists to explain how globally robust behavior can emerge from the decentralized interactions of less reliable cells. More recently, her group has begun branching out into modeling of multi-cellular behavior during embryogenesis.

“I have always been fascinated by the ability of biological systems to achieve amazingly complex structures and complex functions as a whole, even though the constituent parts are constantly being replaced,” Nagpal says. “From cells to social insects, there seems to be a tantalizing idea that one could achieve reliability from cheap, mass-produced parts with limited life and limited reliability. I believe that by trying to replicate such robustness, we will learn new ways of programming systems.”

Nagpal characterizes the Microsoft Research fellowship she received as both an honor and an opportunity.

“For a first-year faculty like myself, it’s a huge boon to be able to pursue long-term and more interdisciplinary research goals,” Nagpal says. “One of my goals is to use some of the grant to support students who are interested in computational/systems biology and want to do both theoretical and experimental work.”



Dan Klein

Rounding out the group of five fellowship winners is Dan Klein, who began his first year on the faculty at the University of California at Berkeley by teaching a new graduate course on statistical natural language processing. Klein is considered an expert in natural language processing, which involves using computers to analyze and understand human languages. His research focuses on designing systems that learn language in an unsupervised way, by automatically detecting linguistic structure. He explains that this approach is much like how human children learn language, in other words, by exposure. In contrast, most existing work on language processing has taken the much easier supervised learning approach, in which the details are more instructed than discovered.

An assistant professor in Electrical Engineering and Computer Sciences department of the University of California at Berkeley, Klein chose his field based on a lifelong love of both languages and computers. He believes it was “just luck on my part” that their intersection was also an important area of study.

“Taking linguistics classes in college, I was fascinated by how complex language was on one hand, and how we learn and use it so effortlessly on the other,” Klein says. “It’s exciting to build systems that are capable of automatically detecting and analyzing some of that complexity. It’s also rewarding to work on hard practical problems, like text comprehension, information extraction and machine translation, which can improve our own ability to process and understand information.”

Because so much of the digital information available today is in text form, natural language processing is relevant to many areas of computer science, Klein says, from interface design to databases to general machine learning and artificial intelligence. His own interests lie in applications ranging from text understanding to machine translation to speech processing.

Klein predicts that the financial component of the Microsoft Research New Faculty Fellowship will go a long way toward supporting the next generation of top students as they pursue important projects, especially projects that are more speculative in nature.

“It’s always easier to get funding for areas of current expertise, but this kind of award provides the freedom to branch out and explore ideas in new areas,” he says. “In addition, I hope this fellowship will facilitate collaboration with Microsoft Researchers, who are very active in my field.”