Following the first wave of computing, essentially comprised of basic calculators, we are presently in the second era of computing, characterized by programmable hardware and software. But whereas computers currently function according to the prewritten “if A, then B” models developed by mathematician John von Neumann in the 1940s, cognitive systems—dubbed the third era of computing—will program themselves to adapt to complex circumstances. The revolution, then, lies in the development of every Sci-Fi lover’s dream: the ability for computers to not only perceive and integrate sensory stimuli, but to also learn from these experiences. “The structure of this machine is entirely different from today’s commercial computers. The memory and processing elements are built close together,” explains Dimitri Kanevsky, IBM Master of Invention. “Operations are asynchronous and event driven; that is, they have no predetermined order or schedule. And instead of being programmed, they learn. Just like us.” Whereas computers have traditionally been programmed to make rapid calculations, cognitive systems function according to analytics. These systems are furthermore capable of processing unstructured data (such as images and video) and natural language. By integrating the hardware circuitry involved in the systems’ memory and processing, cognitive computers imitate pathways in the brain, allowing data processing and memory to make connections in a manner that mimics your brain’s neurons. “Using advanced algorithms and silicon circuitry, cognitive computers learn through experiences, find correlations, create hypotheses, and remember—and learn from—the outcomes,” Kanevsky says. Just as psychology and neuroscience professors repeat the time-tested phrase “cells that fire together wire together,” a paraphrase of Donald Hebb’s theory of associative learning from 1949, the algorithms used in Artificial Intelligence (AI) will respond to the computer’s environment and experiences. “The important characteristic of cognitive computing is that it is implemented as a distributive system of small computing elements (nodes) that are interconnected between themselves,” Kanevsky says. “These nodes perform hierarchical transformations, sending data from one layer of nodes to other layers. Each time that data is sent to a new level of layers its representation is improved.” (via John Hagel)
Posted on: Wed, 25 Sep 2013 20:30:49 +0000
Trending Topics
Recently Viewed Topics
© 2015