Cognitive systems is an exploratory focus of IBM Research and may help transform the way computers are used in the future. The most well known display of this work was through Watson and his Jeopardy smackdown in 2011 of the best human contestants Jeopardy could find.
While traditional systems today provide “if then what” equations, cognitive systems take an adaptive approach in hypothesizing best answers and are refined through learning over time. As Dharmendra Modha, the manager of Cognitive Computing at IBM, states, “We’re not trying to build the brain, we’re trying to draw inspiration from the brain.” For systems to operate in this fashion IBM highlights four necessary characteristics:
- They are data-centric
- They are designed for statistical analytics
- These systems “scale-in”
- They automate system and workload management
It is interesting to think of the possibilities this may provide to the future of business analytics or even the added dependency this could cause for computers in decision making. (i.e. the occasional, “Why don’t I just Google that before I process my question and spare myself the effort”) Either way, it is undeniably an interesting topic.
For more great insight into cognitive systems read this recently published article on the IBM Research site, Cognitive systems: A new era of computing.
