Cognitive Computer: IBM Researchers Ready for the New Age of Cognitive Computing
by Kimberly Carver
Is a cognitive computer the future of computing? While the computer has come a long way in the last few years, researchers at IBM say it hasn’t come far enough. Research in cognitive computing could lead to smart computers that compose, create, and digest via cognitive learning, a process that might seem a bit intimidating and complex at first, but is now being called the future of computers, and expected to be introduced into the market in as little as five to ten years.
Using the Senses to Create Smarter Computers
A cognitive computer would have to possess the senses that living things do in order to hear, see, smell, and taste, without those senses, the computer is little more than a ‘large calculator’ according to scientists at IBM. The cognitive senses include a variety of memory, learning, language, reasoning, and problem solving skills based on experiences of touch, taste, sound, sight, and smell.
In his prediction of new technology coming to the world, IBM Chief Innovative Officer Bernard Meyerson writes about the rapid change in computer technology in the last 60 years alone, and the adaption from a ‘fast calculator’ into a ‘cognitive computer’. He goes on to write about the ‘shift’ in technology, with new advances in science now available to give computers cognitive abilities including the five senses of sight, touch, smell, taste, and hearing. All of which is possible for computers now. The concept of the cognitive computer is part of IBM’s ‘Smarter Planet’ initiative, and has been talked about for four years, with some examples including biometric technology, and driverless cars already beginning to surface in the market.
Cognitive Computer System
In what the scientists at IBM are calling the ‘Cognitive Systems Era’ computers are becoming more advanced and being equipped with cognitive senses. These cognitive computers are embedded with data encryption and analytics, systems in which data storage is being merged with data processing, a system very much like the human system of learning.
The current computer can already hear and see, but doesn’t do anything more than record what it hears. A cognitive computer would think and intelligently decipher what it sees and hears in order to do something. For example, a dictation tool such as iPhone’s Siri actually intelligently listens to what is being said.
Essentially, cognition, or the ability to learn and make decisions could allow a cognitive computer to aid the human brain, to help make decisions, and to be a better assistant and tool for the human mind. But IBM scientists don’t wish to replicate the human brain, instead, they wish to create artificial intelligence that can keep up with, but not duplicate human intelligence.
What’s in a Cognitive System?
A Cognitive computer would display four basic characteristics including being data-centric, scale in architecture, automated system and workload management, and statistical analysis. Each of these factors would combine to allow the computer to ‘learn’ as it experiences new things. Data centric computers use stored data, shared databases, and storing procedures that do not rely on built-in memory housing, which often runs out or is filled to quickly to work as a standard. Scale in architecture in computers is the process of scaling down items such as memory and performance in order to perform more quickly and logically, analytics in computers would allow the computer to analyze data as it comes in, even having the ability to engage in ‘dialogue’ with people.
New Technologies that Make Cognitive Computation Possible
Cognitive computation isn’t the thing of the next century, or even of the next generation. Scientists at IBM predict it will be here within 5 years, mostly due to the many new technologies coming up in science. IBM is currently researching computer processors at an atomic level using as few as 12 magnetic atoms to store data. By comparison, a disc drive today uses more than a million of the same atoms. This system is called a quibit bit system, and could potentially factor a 3,000 digit number 1040 times faster than a system today.
At the same time, the SyNAPSE in collaboration with DARPA, have produced a chip that virtually emulates the neurons and spikes of the human brain. Called True North, two prototypes of the chip are already undergoing testing.
With new technology and new applications appearing nearly every day, the Cognitive Systems Era is all but here. While a cognitive computer might not be affordable for many years to come, the technology is developing now, and nearly ready for use.