Microsoft looks to future of scientific computing

Microsoft has stepped up its courtship of the international research community with the publication of an investigation of how computer science will underpin scientific discovery in the next 15 years.

2020 Science, a report based on the prognostications of 34 eminent practitioners brought together by the company, concludes that brute force computing is no longer sufficient to deal with the data deluge, particularly in the life sciences. Instead, the methods and models of computer science must be applied to drive and inform natural sciences in the same way as mathematics supports advances in physics or astronomy.

“This is the first-ever complete analysis undertaken by an international group of distinguished scientists that looks at the impact of computing and computer science over the next 15 years,” said Stephen Emmott, director of Microsoft’s scientific research programmes in Europe, launching the report in London on 21 March.

The study is the latest in a series of Microsoft initiatives to get better connected with the international scientific world – partly to build up its own R&D operations, on which it spends about 15% of revenues, and partly with an eye towards winning more business from research institutions. At present, while Microsoft products are common in universities, some of its competitors – such as IBM, Apple and Sun – have long had a stronger foothold in that marketplace. And recently, the stakes for Microsoft in academia have risen, as universities have become important battlefields in the competition between Microsoft’s Windows and the “open-source” Linux computer operating system.

Microsoft’s engagement in this arena comes on top of its European Science Initiative set up last year to work with research groups across Europe. Partnerships formed under its auspices include a collaboration with Trento University in computational biological research and the creation of a joint laboratory near Paris to be called the Microsoft-INRIA Institute of Computational Science.

This does not signal a direct move into the sciences market, Andrew Herbert, managing director of Microsoft Research Cambridge told Science|Business, but he noted: “The needs of scientists are impacting on our current products.” He said it was possible to envisage the development of ‘Office for Scientists’, providing the ability to visualise and analyse complex data sets, construct and manage models, and support collaboration between scientists within their own laboratory, or externally.

It is not clear if this would be a discrete new software product or built on existing platforms. There is already a high performance version of Windows, for example.

Emmott said the impact of computer science in the life sciences will be far greater than generally accepted. “It forms the foundation of nothing less than a scientific revolution, and is not simply about faster computers, but a means of transforming the way science is done.”

The insights will revolutionise the way some of the greatest challenges of the 21st century – from the risks to the biosphere, global pandemics of malaria and HIV, cancer and the search for new energy sources – are confronted.

“The essence of our findings is that the ability to tackle these [problems] is about to be transformed by new approaches to computing and computing science,” Emmott said.

Examples include the emergence of data cubes to deal with exponential data growth and complexity, software models of biological processes, algorithms that can take the data deluge and use it to make predictions on biological systems and processes, artificial/robot scientists that are capable of conducting experiments that humans can’t, or of speeding up ones that are tedious and time consuming, and molecular machines.

To publicise the issue, Microsoft is sponsoring the free dissemination on the Web of a number of articles in the 23 March edition of science journal Nature discussing the future of computing in science.