Give me data and I’ll change your world.
Such claims are commonplace in the world of business, governance, healthcare, elections—any walk of life you may think of. The catch words remind us of late John Martin’s ‘Give me half a tanker of iron and I’ll give you an ice age’. One of the catchiest lines ever spoken by an oceanographer, Martin’s idea of iron fertilisation of oceans, to trigger plankton blooms that would act as carbon sinks, became popular through the 1990s when global warming was becoming evident. Martin may have said this in half-seriousness, but big data fans, today, are very serious about changing the status quo. Sandy Pentland of Massachusetts Institute of Technology has had ambitions of reinventing human society on a large database platform.
Now, designers of algorithms, which play a critical role in data generation and analysis, are getting serious. Because algorithms are proving to be discriminatory, reinforcing human prejudices, and generally not as objective as they are widely thought to be. And since the march of algorithms is not going to be stopped—India plans to use it for policing—the next best thing to do is to create algorithmic checks and balances that can raise the right flags.
Nearly half of India’s 50 or so algorithm designers, theoretical computer scientists to be precise, gathered in Mysore last weekend to brainstorm on such issues. To bring computational lens, a term coined by four Berkeley computer scientists, on other sciences. It doesn’t mean applying computer science to a problem, rather it’s a way of looking at the processes of nature (or other disciplines) and giving them a new algorithmic perspective.
Nisheeth Vishnoi is leading the charge. His own work is a prime example of such an approach. Among other things, he has shown a connection between the dynamical systems in signal processing and the single-celled organism, slime mould. Currently, at the Swiss Ècole Polytechnique Fédérale De Lausanne, Vishnoi is exploring if more computer scientists in India could use this lens and if a new Institute of Computation and Human Sciences would make sense.
“We don’t want to do big data. We want to understand how we are generating the data. Unless we understand what is special about the data, we don’t have any hope of making leapfrog progress,” says Vishnoi.
It’s no surprise then that Kris Gopalakrishnan, Infosys co-founder, who has supported a few computational science research initiatives in India, has shown deep interest in backing Vishnoi’s idea. At the Infosys Leadership Institute in Mysore, Gopalakrishnan kicked off the idea exploration with a fairly broad and ambitious call.