The world of big data is nothing without achieving results, a well known axiom in the tech world whose implications aren’t so obvious to the layman. That’s the job of Kira Radinsky, the newly dubbed director of data science at eBay, who opened
Her work started by correlating events as they are mentioned in the media. Over time, her work has been able to develop predictive models called event chains, noticing that were one event to follow another, the dominoes would fall in just the right order to lead to an outbreak of a disease.
Without the ability to abstract multiple possible conclusions, event chains might be too rigid. The main issue in building predictive algorithms based on these things still depends on our preexisting knowledge of what actually causes certain results, whether it be an economic downturn or epidemic. Abstracting a result will allow machines to draw conclusions not already fleshed out in extant data sources.
“The only way to address causality is perform a controlled experiment,” which might not sound very exciting for people expecting fast results, but the hypothesis-observation-conclusion scientific method of old is already speeding up thanks to the very technology that she hopes will accelerate new medical and economic discoveries.
“I envision a system that still has those predictive data pools. It looks at the different data you obtain that different labs are giving all the time,” she highlights, saying it is easier than ever to request data from laboratories. From there, you can see more “automated processes that can lead to those types of discoveries.”
In her mind the result of all these abilities is inescapable: the hastening of new discoveries in the hands not of humans but the fingers of digital.
“If you ask me what the future is, the future is making discoveries in an automated way,” going from a long-term timetable of 5-10 years to make new medical discoveries, but we could bring that timetable to mere days, if not seconds.