Emil Miller
02-06-2012, 09:34 AM
The Science and Technology section of Le Monde has an interesting article headed Networks: the quest for the laws of the Web.
It explains how search engines and social networking companies employ mathematicians to discover ever-more powerful algorithms so they can use their networks to pinpoint individual traits of their users.
This is common knowledge but in a sub article, headed Big Brother and Big Data, it underlines a problem arising from this activity. It isn’t the risk of a Big Brother who knows everything about us, but the disappearance of individuals themselves, to be replaced by statistical profiles and their control by algorithms using databases to produce correlations between profiles. Thomas Berns of the free university of Brussels says: “ We are falling into the historic error of using statistics to describe the social and to process this mass of data, there’s no need for hypotheses; the correlations suffice. That’s why, with my colleague Antoinette Rouvois, we speak of government by algorithm.”
Chris Anderson, described as a numerical intellectual, has proclaimed the end of science under assault from the processing of ‘big data’. “The scientific approach with its hypotheses, models and tests is obsolete. Correlation has overtaken causality.”
The obvious question posed by this ‘substitute science’ is: are humans going to be increasingly reduced to mere numbers to be exploited by those who not only control large parts of the web but are actively engaged in attempting to discover or construct the laws that govern it?
It explains how search engines and social networking companies employ mathematicians to discover ever-more powerful algorithms so they can use their networks to pinpoint individual traits of their users.
This is common knowledge but in a sub article, headed Big Brother and Big Data, it underlines a problem arising from this activity. It isn’t the risk of a Big Brother who knows everything about us, but the disappearance of individuals themselves, to be replaced by statistical profiles and their control by algorithms using databases to produce correlations between profiles. Thomas Berns of the free university of Brussels says: “ We are falling into the historic error of using statistics to describe the social and to process this mass of data, there’s no need for hypotheses; the correlations suffice. That’s why, with my colleague Antoinette Rouvois, we speak of government by algorithm.”
Chris Anderson, described as a numerical intellectual, has proclaimed the end of science under assault from the processing of ‘big data’. “The scientific approach with its hypotheses, models and tests is obsolete. Correlation has overtaken causality.”
The obvious question posed by this ‘substitute science’ is: are humans going to be increasingly reduced to mere numbers to be exploited by those who not only control large parts of the web but are actively engaged in attempting to discover or construct the laws that govern it?