It came as no surprise to us when the BBC ‘broke’ the news last week that supercomputers could help predict major world events. Supercomputing is a topic close to our geeky hearts at Catalysis – we’ve written about it for the past ten years and part of our daily mission is to ’make servers sexy’. So the latest US research, which shows that feeding a computer with millions of articles can allow us to chart deteriorating national sentiment ahead of revolutions, drew a few knowing nods around the office.
The US study retrospectively charted deteriorating national sentiment ahead of the recent revolutions in Libya and Egypt. Now, “hindsight is a great thing!” I hear you clamour. This is true, but the scientists conducting the study firmly believe that this process could be used to anticipate upcoming conflict. Following the recent riots across the UK, you can’t help but wonder whether supercomputers could be giving us some direction.
How does a supercomputer do it?
We have been lucky in the past to host a number of press visits to CERN (the European Laboratory for Particle Physics), helping us to understand this technology better ourselves while illustrating the powers of a supercomputer to others. CERN is at the apex of scientific research and relies on highly advanced data centre to power its programme so we have a good idea of what supercomputers can achieve.
In the recent US-based study, which looked at national sentiment, led by Mr. Leetaru, the process began with feeding millions of articles into the supercomputer. The articles were then analysed for two main types of information: mood and location. And so you start to build up a picture of good news, versus bad news and where it is all happening. When it comes to predicting trouble, these results are all aggregated and begin to show notable dips – or highs- when it comes to sentiment both inside, and outside, a country.
For Egypt, the BBC reports that the tone of media coverage in the month before President Hosni Mubarak’s resignation had fallen to a low only seen twice before in the preceding thirty years. Other previous dips in sentiment include the low felt in 2003 prior to the US invasion of Iraq. Mr Leetaru, leading the study, claimed that his system appeared to generate better intelligence than the US government was working with at the time.
When it comes to the UK, the papers are still full of articles trying to analyse why the recent riots took place. This research suggests that supercomputers could help us by identifying pressure points and making sure we are in the right place at the right time to either prevent – or deal with – them. The government would need to decide whether this has a place in its social strategy and whether we want to direct significant resources towards this. What do you think?