Science: with and without the Internet, Pt I

Michael Nielsen has a mission: to open up science. A pioneer in the field of quantum computing (a quick Google scholar search shows his top publication to have over 13,000 citations), Nielsen pretty much dropped out of academia to pursue his quest to change science. Like myself he is an Australian living in Toronto. But this week he has released a popular book that explains his quest: Reinventing Discovery: The New Era of Networked Science.

I read the book last month and it represents an optimistic journey. To anyone interested in science and the digital world it is a must read. But what does it all mean? How could one really open up science?

It turns out that opening up science has a number of dimensions but they can be broadly compartmentalized into the creation and dissemination of knowledge but, as we will see, these things are tied together. The most familiar form of ‘open science’ relates purely to dissemination and I’ll focus on it in this post leaving creation for a future one. As an example of dissemination, Google provides information on searches related to the flu to search as an early warning system for outbreaks. And there is a longstanding notion that publicly funded research should be publicly available. But as Fiona Murray and I recently pointed out, there are some trade-offs there that may mean that common sense notions of fairness may not equate with the efficient provision of incentives. Nonetheless, there is an important sense in which having knowledge accessible to generate useful applications is what it’s all about and something that we should err on the side of doing.

Where Nielsen spends more time is on the importance of open access to knowledge in creating new knowledge. He considers a retired information scientist, Don Swanson, who drew from data on Medline to find hidden patterns across millions of scientific papers and how openly available data from the Sloan Digital Sky Survey has enabled large-scale searches to identify astronomical phenomena like black holes. But surely it isn’t too surprising that when we take data and make it easy for experts to access it, new knowledge should come of it.

Again there is a sense that leveraging digital technologies to make large data-sets available is a no brainer. Moreover, precisely because we can leverage digital technologies to exploit those data-sets, raises the returns to investing in them.

Where things become more radical — in the sense, of previously thought of as unconventional — is with regard to who might play a role in science. Our conventional mindset is that scientific creation is the preview of the expert, those with years of training or at least those who aspire to be trained and play the scientific game. But Nielsen brings several counter-examples to our attention. One is Galaxy Zoo, a website that offers anyone the chance to classify galaxies — a task that still relies on humans rather than computers. While, for the most part, that endeavour is just part of getting an important task done, on occasion, as Nielsen documents, it has actually allowed people to be associated with significant breakthroughs. Astronomy has always had an amateur role though. Where things get more interesting is the notion that this type of activity — which requires cognitive processes not significantly above that of FarmVille — could tap into leisure activities. A primary case in point is Foldit — a game that allows molecular structures to be analyzed — and which recently led to a breakthrough in AIDS research.

The opportunities in this type of ‘citizen science’ are twofold. First, there is a notion of community — that individuals can be part of scientific endeavours. Second, there is the notion of connection. By having participation in this way, the scientific mission may not be so distant and so may ultimately face less resistance. But I must admit that I remain to be convinced that there are significant breakthroughs to be had beyond the important notion that mass collaboration by people can outperform similar tasks that may be ascribed to machines. None of that should diminish the efforts but it should give us pause about their rate of return.

What is more interesting is where Nielsen starts his book — how the Internet can facilitate collaborations amongst scientists. But I’ll leave that web of issues for next time. Meanwhile, take a look at Nielsen’s compelling TED talk and you can help but hope there are more like him in the scientific community.