Download Research Tools
As the saying goes, “Seeing is believing.” But with computers, that’s only half the story. Cameras are becoming an ever-present part of our world. They are built into cell phones and laptops, and dot the landscape in storefronts and on street corners. Their pervasive images present us with a wealth of information. So how do we extract information from these images and use it?
One hundred excited students from across Russia and the Commonwealth of Independent States (CIS) converged at Moscow State University for the 3rd Annual Microsoft Research Summer School.
That question set the scene for 100 excited students from across Russia and the Commonwealth of Independent States (CIS), a quarter of whom were women, who converged on Moscow State University (MSU) for the third annual Microsoft Research Summer School. This year’s session focused on the intricacies of computer vision, with activities led by Microsoft Research experts and leading European academics.
The summer session began with a special welcome from Nikolay Pryanishnikov, president of Microsoft Russia. “Supporting young talent is traditionally one of our key strategic priorities,” Pryanishnikov told the students. “We are confident that, with the help of events like this Microsoft Research Summer School, our young specialists will be able to realize their ideas, reach new peaks, and increase the innovation potential of the Russian economy.”
The students were busy throughout the week; each day was packed with intensive academic talks, demonstrations, and hands-on laboratory sessions that were designed to educate attendees about fundamental and state-of-the-art techniques in computer vision. Andrew Fitzgibbon, a principal researcher at Microsoft Research Cambridge, gave a detailed description of how decades of computer vision research, along with ground-breaking ideas from Microsoft Research, came together to make Kinect technology a reality. The summer session also featured industry talks: Aram Pakhchanian of ABBYY, a Moscow-based company that specializes in optical character recognition, and Michael Nikonov of iPi Soft, a company that specializes in motion capture technology, talked about how to create a startup company in computer vision.
Andrew Blake, managing director of Microsoft Research Cambridge, was delighted to lecture and talk to the enthusiastic students. “It was clearly a splendidly vibrant event, with tremendous enthusiasm from the students,” he said. “This really is a landmark event for Microsoft in Russia. It marks a milestone in the maturity of the developing links between Microsoft Research, Microsoft Research Connections, and Moscow State University.”
Anton Konushin, head of the Vision Group at MSU, hopes that others can benefit from the Summer School. “Our school was truly a most selective one, with only one out of five students was accepted to the school. But with video lectures available online soon, we hope that this 400 students who hadn't made it to the event, can also become familiar with materials. We plan to make the influence of the school to Russian computer vision community a long-lasting one."
At the end of the week, students departed the summer school filled with enthusiasm and a deeper insight into how computer vision can change our world.
—Fabrizio Gagliardi, Director, Microsoft Research Connections EMEA (Europe, the Middle East, and Africa)
It was barely a year ago that European scientific and industry leaders came together with the goal of developing, testing, and deploying a high-quality, interoperable cloud platform for industry and research. The result was VENUS-C, which stands for Virtual multidisciplinary EnviroNments USing Cloud infrastructures.
Jointly sponsored by the European Commission and a consortium of 14 partners, among them Microsoft Research, VENUS-C was conceived to meet the needs of seven different research and commercial areas: bioinformatics, systems biology, drug discovery, civil engineering, civil protection, civil emergencies, and marine biodiversity. VENUS-C has since developed into a functional, operational platform, and is being used for 15 new pilots that received seed funds after an open call elicited 60 proposals from across Europe.
Conceived to meet the needs of seven research and commercial areas, VENUS-C is now a functional platform that is being used for 15 new pilot programs in various fields.
The success of VENUS-C is hardly surprising, since it offers powerful computing resources, open solutions, and a user-centric focus—all without the upfront costs of expensive IT installations. What’s more, VENUS-C’s massive computing power helps expedite research, speeding the time from hypothesis to result.
Vladimir Sykora, co-founder of Molplex, a small U.K. company working on drug discovery, and recipient of one of the 15 pilot grants, remarks, “Thanks to the VENUS-C platform, we will be able to do in a few weeks molecular computations that would have taken a year to complete on our own servers. This application allows us to quickly estimate the activity in the human body of new chemical compounds.”
VENUS-C also offers the cloud advantages of scalability, providing resources as and when needed. This is enormously valuable in many areas of scientific research when peak computing needs occur sporadically and often unpredictably.
As Costas Papadachos of the Geophysical Laboratory at Aristotle University notes, “Geoscientists involved in the difficult territory of earthquake impact assessment have much to gain from initiatives like VENUS-C. Our involvement offers a prime opportunity to access unprecedented resources, only when and where necessary for earthquake impact estimation and related information dissemination, without worrying how to build and maintain the corresponding infrastructures and operational tools.” Papadachos heads one of the 15 new pilot experiments on the VENUS-C cloud platform.
Cloud computing can provide new approaches to data collection and management, too. Using the VENUS-C platform, Collaboratorio, an Italian-based micro-enterprise that is managing one of the civil engineering scenarios, is collecting data on the performance of new buildings, creating a database of qualitative and statistical information that can be used to find the designs that best fit specific environmental and urban contexts. The cloud will help Collaboratorio’s researchers mine the data to identify trends, perform extrapolation studies, and address common challenges that are related to building and environmental impact.
As VENUS-C embarks on its second year, we look forward to the platform performing even bigger and more complex applications. This is a very precocious one-year-old!
—Fabrizio Gagliardi, Director, Microsoft Research Connections EMEA
Collaboration can be a great catalyst for new ideas. Whether working with colleagues from down the hall or a team from another continent, we have found that working together strengthens our ideas. A prime example is the Barcelona Supercomputing Centre – Microsoft Research Centre in Barcelona, Spain. Microsoft Research Cambridge began collaborating with the Barcelona Supercomputing Center (BSC) in 2006. We formalized the relationship with the establishment of the BSC – Microsoft Research Centre in January 2008. The Centre focuses on the design and interaction of future microprocessors and software for the mobile and desktop market segments.
The BSC – Microsoft Research Centre is home to a talented group of students who are working towards their PhDs and who bring their creativity and enthusiasm to tackle tomorrow’s problems. “I am very happy that the Centre is a model of open research,” said Centre director Mateo Valero. “We share our findings with the community and all of our software and applications are available for download at our website.”
The program has an extremely young team with more than 15 PhD candidates, Valero explained. Leading the student group was Ferad Zyulkyarov, who is at the forefront of Transactional Memory (TM) research. Working under the supervision of Valero, and his colleagues Osman Unsal and Adrián Cristal, Zyulkyarov investigated how this new approach to multi-core programming could make software development much easier for future computer architectures.
Ferad Zyulkyarov defends his thesis in Barcelona
A Different Point of View
Previous TM research had focused on evaluating and improving TM implementations. Zyulkyarov took a unique approach to the problem, looking at it from the programmer’s point of view. As part of his thesis, Zyulkyarov developed one of the first real-world TM applications: a rewrite of the Quake Game Server that replaced traditional memory locks with TM atomic blocks. This makes life much easier for the programmer, potentially transforming multi-core software development for the future.
Zyulkyarov encountered some obstacles during his project. For example, he had to develop a better debugger and profiling support, neither of which existed before he created them. When he reviewed the performance of the core server code, Zyulkyarov could see the potential for TM. There is still some optimization work to be done, but the potential is there.
During the project, Zyulkyarov collaborated closely with Tim Harris, senior researcher, Systems and Networking Group, Microsoft Research Cambridge. Harris is proud of the work Zyulkyarov accomplished during their time together. “It’s great to see Ferad’s work come to fruition,” Harris said. “He’s made substantial contributions to the development of programming tools for using TM, and I hope that we’ll now be able to apply these ideas to other parts of the multi-core challenge.”
The First of Many PhDs from Barcelona
The first of the 15 students to receive his PhD, and now at Intel, Zyulkyarov is just one example of the young talent being fostered through the BSC – Microsoft Research Centre, driving the industry to tackle some of its most challenging problems. “In the five years since we have started, the Centre has matured quite a lot, and this is the first fruit of the collaboration with BSC and Microsoft Research,” Valero said, adding he is especially grateful to Harris for serving as Ferad’s mentor. “I know that more [success stories] will follow soon,” he added.
I am very glad—thinking back to my first visit to BSC five years ago—in seeing how far we came. This is the result of all the energy and enthusiasm we have all put together in the enterprise. This is only the first of a successful series of PhD awards, which we will see taking place in the next few years.