Computer science can play a major role in fuelling scientific breakthroughs in other fields, but only if it can meet its own grand challenges, said Craig Mundie, chief research and strategy officer of Microsoft at a symposium in Brussels debating how best to use EU-funded research programmes to solve grand challenges.
Mundie spells out the challenges
The first of these challenges is to bring about a fundamental change in the way that software is written to make it reliable enough to take full control of life-critical systems, such as power grids or hospital equipment. Mundie noted that one way to tackle climate change would be to introduce a highly distributed power generation system consisting of millions of small electricity generators fuelled by renewable energy.
But the variable and unpredictable nature of wind and solar power would require the system to be run by a smart grid capable of balancing supply and demand. “If we go for a highly distributed system, we will find ourselves in a situation where the load will have to be controlled dynamically,” Mundie said. “It would be great to find a way to buffer electricity, but it doesn’t seem likely. However, it is conceivable that we can develop a closed-loop control system that extends all the way down to the devices, but it will involve a lot of work – a big research challenge.”
The way software is written today means it isn’t reliable enough to take on the challenge of managing such a closed-loop control system, he said.
The Microsoft research chief defined the second grand challenge for the software industry as achieving a major leap forward in what computers are capable of through the introduction of quantum computing. Mundie said that major advances in some key scientific fields, such as biology and electric propulsion, may depend on the use of quantum information systems which are many more times more powerful than today’s classical computing systems.
And the half challenge? That, he said, is for computer scientists to collaborate much more closely with scientists from other disciplines, so that software and computation can play a much more fundamental role in early-stage research. Mundie believes that many scientists don’t consider the potentially transformative role of computation early enough in the development of a research project, thinking of it more like electricity. He said the gap between the most “avant-garde” people in computer science and information technology, and their peers in other scientific disciplines is widening, meaning that most fundamental research projects fail to harness the potential of the latest computing technology to support breakthroughs in other fields.
“Everything is very incremental,” Mundie said. “One of the great challenges is to break this incrementalism.”
He described how his former colleagues Bill Gates and Nathan Myhrvold have invested in TerraPower, which is developing a new form of nuclear fusion that could help the world address the challenge of generating much more electricity without causing an environmental disaster. “They were able to consider entering into a new form of nuclear power because they can now buy, for a relatively modest sum, the kind of computing power that used to only be available to governments, and put it in their basements,” he said. “Many problems that were thought not to be economically computable, now are.”
Author: David Pringle
Photographer: Thierry Monasse