Imagine a world where your local newscast can provide you with a forecast of the future, right alongside the daily weather forecast. Sound like science fiction? Well, thanks to remarkable advancements in the burgeoning field of Big Data processing, predicting the future could soon be as easy as predicting the weather, according to Phys.org

"Big Data" is a blanket term for any collection of data sets so large and complex that it becomes difficult to process. The amount of data that would be required to accurately predict the future, especially in terms of largescale human social behavior, certainly qualifies as big — humongous, in fact. Until recently, such a data set would have been considered impossible to process. 

"There is essentially more data being uploaded than what we're actually experiencing," explained Devi Parikh, who leads the Computer Vision lab at Virginia Tech. "Every second of life produces somewhere near two hours of content."

Parikh was a panelist at a recent gathering of the Early Model Based Event Recognition using Surrogates project, or EMBERS, an organization based at Virginia Tech that is attempting to harness the immensity of data being constantly uploaded worldwide with an aim of predicting the future with it. He spoke to a crowd of computer engineers, artificial intelligence experts, philosophers, and virtual reality gurus who had all amassed to hear about the latest in Big Data processing.

As Parikh and others explained at the symposium, what had once seemed impossible — predicting the future through Big Data processing — today seems inevitable. In the past year alone, the team at EMBERS has honed their methodology to correctly forecast some major events, such as the riots after Paraguay's president was impeached, the Hantavirus outbreaks in Argentina, and the recent mass protests in Brazil and Venezuela.

All of the event alerts generated by the system can be emailed to researchers in real time and are compared against a gold standard report organized by a third party. A report card can even be generated that grades the predictions. Tests which analyse alternative forecasts and determine what might have been overlooked are frequently performed to improve the predictions.

Although these advancements are exciting, they also raise a lot of ponderous philosophical questions. For instance, if we can predict the future, then will it also be possible to change the future? And if so, might the system one day perform alternative-future predictions so that we can choose the best possible future?

Many questions, however, also raise practical and ethical concerns. "Let's say you can predict the future and you intervene to prevent an uprising," said Parikh. "How can you be sure that an uprising was actually going to happen?"

In other words, how do you adequately test a theory unless a history is actually allowed to play out? Furthermore, if we're to learn anything from current debates about data mining, it's that the issue of privacy is also sure to be a concern. 

Even so, as computer networks become better equipped to handle vast amounts of data and our algorithms for interpreting the data become more sophisticated, predicting the future might be too tempting an application to avoid. After all, knowledge is power, and it's difficult to imagine a more powerful kind of knowledge than knowledge of the future.

Related on MNN: