Friday 31 December 2010

Living Earth Simulator aims to predict everything that's happening on our planet

I thought I had stumbled across an old Goodies script, or a pilot episode for Dr Who, but no, this is real. As ambitous projects go, this one must be right up there alongside making Heathrow a world-class airport.



The Living Earth Simulator will collect all the data in the entire world, to predict everything from the next major disease outbreak to the next financial crisis. And taking a leaf from the Large Hadron Collder - we now have the concept of a "knowledge accelerator that can collide different fields of knowledge" - i.e. a "knowledge collider". (Note to self: must add that term to my repetoir of knowledge management jargon).



Call me an old sceptic, but I've yet to see any evidence that collecting more and more data makes us better at predicting the future, or that it is even possible to accurately capture social trends as a mathematical model.



With fairly limited data at my disposal, I'm willing to predict that the UK will come to a grinding halt with the first flurry of snow and ice next December 2011. Of course....lessons will be learnt (maybe from LES?)!

Amplify’d from io9.com

Living Earth Simulator aims to predict everything that's happening on our planetThe Living Earth Simulator is quite possibly the most ambitious computer project ever undertaken. This all-encompassing simulation will collect all the data in the entire world, to predict everything from the next major disease outbreak to the next financial crisis.

The Living Earth Simulator could do for our modern world what the Large Hadron Collider has done for the early universe, says project chair Dr. Dirk Helbing. He calls the LES a "knowledge accelerator" that can collide different fields of knowledge to produce a far greater understanding of what's going on in the world around us.

Such a program, he says, could help show us the next epidemic before it starts, illuminate better ways to deal with climate change, and predict when the next recession will hit. According to Dr. Helbing, the answers to all these mysteries can be found by examining the sum total of human activity:

"Many problems we have today - including social and economic instabilities, wars, disease spreading - are related to human behaviour, but there is apparently a serious lack of understanding regarding how society and the economy work. Revealing the hidden laws and processes underlying societies constitutes the most pressing scientific grand challenge of our century."

So where would they get all the data from? Lots of different organizations are already compiling massive amounts of data, and these would help feed into the Living Earth Simulator. Possible sources would include NASA's Planetary Skin project, which tracks climate data on every corner of the globe, as well as more everyday sites like Google Maps and, yes, Wikipedia. Helbing and his team also plan to incorporate medical records, the latest financial information, and, most frighteningly of all, everything that's going on in the world of social media.

Of course, once all that data is together, there's still the question of what to do with any of it. Helbing says this will require cooperation between social scientists and computer scientists to create the rules and programming that the LES needs to interpret the data and create an accurate model of the Earth as it is today. We've only now got the technology advanced enough to pull off such an endeavor, and it will still be very tricky.

Part of the solution, Dr. Helbing explains, is the rise of semantic web technology. This simple but powerful concept makes a computer see information not just as a set of numbers but as specific data in a specific context, meaning computers will be able to tell the difference between the seemingly random numbers making up, say, financial markets and weather reports in much the same way humans can.

An obvious question to ask is just how much the LES will be able to learn about particular people. On this point, Helbing argues that the vastness of the project should protect everyone's privacy, as the LES's aggregative strips out all individual data in an effort to create an overall picture.

Once you collect all the data and program the simulator, actually running the LES is relatively simple. Yes, the project will need huge banks of supercomputers to run the entire program, but the processing power required isn't beyond what we're currently capable of. Computer expert Pete Warden says that, in all probability, we do have the processing power to handle what the LES requires. That said, he's skeptical about whether the LES could actually produce useful results:

"Economics and sociology have consistently failed to produce theories with strong predictive powers over the last century, despite lots of data gathering. I'm sceptical that larger data sets will mark a big change. It's not that we don't know enough about a lot of the problems the world faces, from climate change to extreme poverty, it's that we don't take any action on the information we do have."

To this point, Dr. Helbing argues that the LES will offer predictive far in advance of our previous models, as it would be able to see global recessions and disease outbreaks coming before they really get started. It's a bold claim, and we won't know for sure what the real capabilities of the LES are until the day that it's up and running.

Read more at io9.com
 

No comments: