By Rob Knies
October 16, 2007 9:30 PM PT
Challenges abound. We all encounter challenges—to make grades, to care for our children, to put food on the table. Challenge is one of those elemental components that define life itself.
For researchers, though, challenges are to be embraced, not merely surmounted. Identify a problem, devise a hypothetical solution, test, refine—this is the scientific method; this is the essence of research.
It comes as little surprise, then, that when TechVista, Microsoft Research India’s annual research symposium, was held in New Delhi on Oct. 13, the recurrent theme, the lingua franca that resonated through the varied presentations, was challenge. No self-congratulations here; the most important hurdles, it was made clear, are those to come, not those already cleared.
A TechVista audience of students, faculty members, and influentials from government and industry gazed upon a stage in the Royal Ballroom of the InterContinental Eros flanked by a pair of message boards that read simply: Inspire. Innovate. Invent. Between these twin triads of mantra-like imperatives strode a succession of luminaries to share a kaleidoscope of perspectives on the benefits to be derived from confronting computer science’s looming challenges.
“Research,” P. Anandan, managing director of Microsoft Research India, told the assemblage by way of introduction, “is a pleasure. Researchers are, by nature, people who are open, creative, and not bound by too many rules. And when you’re not bound by rules, you are looking to have fun.”
TechVista attendees enjoyed their share. They got a glimpse of the computational future, both long-term and on the immediate horizon. They got a grounding in the theoretical nature of the intriguing challenges the field provides. And they heard the hopes that Microsoft has for the potential of technology, and the dreams that could be actuated—for India and for the world beyond.
THE FUTURE WRIT LARGE
In a talk entitled 10 Years into the Future, Rick Rashid, senior vice president of Microsoft Research, sketched some of the breakthroughs within the grasp of researchers tackling today’s challenges.
He cited the likelihood of a 2017 in which input/output technology would be ubiquitous, on virtually any surface available to us. He offered software engineers the prospect of provable systems to ensure that their technology will work as intended. He suggested the enormous implications of a world in which nearly unlimited storage would provide information access to a lifetime of experiences, in which breakthrough imagery could supply heretofore-unimaginable views of the environment around us, in which streaming intelligence could enable us to anticipate the unexpected.
“Predicting is hard,” Rashid said. “Technology does not move in a straight line. The things I spoke to are things that are in the laboratory. You will be able to do them. What I can’t say is what society will do with them. There are always new approaches that take you by surprise. The key thing is that it’s hard to be able to predict what people will do with technology, what cultures do with technology.”
Craig Mundie, Microsoft chief research and strategy officer, and Tony Hey, Microsoft corporate vice president of External Research, provided a look at the challenges the computing industry and the scientific community face as they enter critical evolutionary junctures.
With physical limitations to multiprocessor production ushering in an era of multicore chips, the days of sequential programming, Mundie said, are giving way to the necessity for parallel computing.
“We’ve talked about this at the research level,” he noted, “as almost a holy grail for many years.” To attain that goal, he added, the computing industry must overcome the dual challenges of concurrency and complexity.
“Our business, our engineering, in my mind, has not yet achieved the same level of formalism in composition that allows us to build giant buildings or build big bridges,” Mundie said. “Other engineering disciplines have been forced to create a more rigorous, controlled composition system in order to deal with scale and complexity. The reality is that, in software, there’s still too much art and too little engineering.”
Society depends upon further advances in computing, he observed, offering the hope that better use of computers’ currently underutilized capabilities could enable the next stage of computational advances.
“I want to challenge you to think more broadly about how … we can create a world of next-generation applications, a world of fully productive computation,” Mundie concluded. “It’s the ability to bring these things together and to think at a higher conceptual level, using more sophisticated tools and formal levels of composing reliable software systems, that I think is the opportunity for all us in the years ahead.”
A DELUGE OF DATA
Tracing the history of scientific methodology from the experimental to the theoretical to the computational, Hey said today’s scientist is working in a Web-based, data-centric realm, one he terms eResearch.
“In the last five or six years,” he said, “I think we’re seeing the nature of science being transformed.”
Hey cited a number of attributes he foresees for eResearch:
“What is really interesting,” Hey said, “is trying to use the technology from industry—from Microsoft and others—to help scientists solve their problems.”
LIFE, SIMULATED AND SUSTAINED
TechVista also featured a number of fascinating presentations of futuristic work being performed today by researchers based both in India and abroad.
“What are the challenges?” he asked. “The first, natural question is: How do we scale it? The second challenge is how to make the output useful for naïve users. The third is how to ensure availability.
“This is where research comes in.”
The aAqua ecosystem uses an e-library of advice, including sections termed Crop Doctor and Crop Recommendations, to deliver information to farmers’ cellphones, utilizing multilingual machine translation. The project faces plenty of challenges—but also offers much promise.
“ICT for social economic development,” Ramamritham asserted, “has come of age.”
Using what he jocularly referred to as “X-ray vision,” he described the creation of surgical visualizations to enable computer-assisted surgery, in which exacting analysis and color coding of various brain structures can enable doctors to make precise maneuvers during delicate operations.
“This is bringing together a sequence of tools,” Grimson explained, “a sequence of machine-vision tools, a sequence of machine-learning tools, a sequence of graphics tools, to let surgeons build patient-specific models that they can use for surgical guidance and navigation, and letting clinicians build patient-specific models that actually see how structures change as a function of disease, in response to pharmaceuticals, or as a function of genetic effects, in order to let computational tools change the face of medicine.”
“Computer animation has a bunch of very creative challenges,” he said. “What you want to do is create totally believable clothing.”
Those challenges include creating ordinary garments on ordinary, human-like characters, good enough for viewers to suspend their disbelief while also being fast and robust enough to be practical for film.
The measure of Kass’ success in doing so can be measured by box-office receipts. His list of credits includes such silver-screen smashes as A Bug’s Life;Toy Story 2;Monsters, Inc.;Finding Nemo; and The Incredibles.
Veloso’s project featured autonomous robots using perception and cognition to produce action. Her initial video showed round, wheeled, sensor-equipped robots interpreting multisensory information to achieve centralized, distributed learning resulting in actual robotic teamwork.
A second video had Sony AIBO dog robots communicating collaboratively to share information not available to all robots, such as telling a teammate where a ball is even when it is blocked from the teammate’s view.
The research challenge, Veloso explained, is how to combine a local view with communicated information to create a coherent world model.
“My work,” she said, “is trying to bring robots to life.”
‘THE NEW INDIA’
The approximately 650 persons attending TechVista also accorded warm responses to a pair of non-researchers, Kapil Sibal, union minister for Science & Technology and Earth Sciences for the Indian government, and Arun Shourie, member of the Indian parliament and former minister of Disinvestment, Communications and Information Technology.
In his keynote address, Sibal talked of the contributions computer-science research could bring to areas of concern in India, such as in agriculture, healthcare, energy, and the environment.
“It’s mind-boggling,” he stated, “what computer science can do.”
Sibal challenged Microsoft Research India to provide help for a few ambitious projects: bringing 3-D computer representations of Delhi’s monuments and cultural sites to students across the nation, producing a virtual science museum accessible to Web users, establishing a sensor network to monitor water usage in India.
“Research by itself is not a solution,” he observed. “You need to move beyond research. Behind research is the face of an ordinary man living an ordinary life dealing with ordinary problems.”
Shourie, a crusading journalist who gave TechVista’s closing address, expressed his appreciation to organizations such as Microsoft Research India for helping to bolster Indian business, expertise, and creativity.
“You are the new India,” he said. “It is because of persons like you that the world’s perception has changed, and therefore India’s perception of itself has changed.”
“India is very unique for us at Microsoft,” Koppolu said. “This is where we have the presence of all the business of Microsoft in one particular country. We have research here, we have product development, we have sales and marketing. We have the Microsoft IT group here, we have technical support, and, of course, entertainment and devices. The whole spectrum of Microsoft is present.”
As for Microsoft Research, Rashid said the mission of its India lab remains unchanged: to continue to collaborate with government, industry, and academia in an effort to extend the state of the art in computer science.
“Why do you make investments in basic research?” he asked. “One of the things we believe in is that fundamental research can change the world.”
For the scientific community, in India and beyond, that, Sibal observed, makes a big difference.
“If you go into the discipline of science,” he said, “research is at the heart of everything.”