Definition of Cognitive Computing

The raw computing power of a computer is getting closer, every day, to the one presented by a human brain, so at some point, someone it should have happened to him: what if we take advantage of this to imitate with a computer how a brain works human?

Cognitive computing is a computing discipline that takes as its guide the way the human brain works to solve computational problems.

While the artificial intelligence to which we are more accustomed does not stop consisting of a series of algorithms that are executed on conventional computers (very powerful, yes), in the case of the computing cognitive we are going to refer to a set of hardware and software that imitates, as a whole, the functioning of the brain as an organ, that is, how neurons interrelate.

If humans have senses (such as sight, touch, hearing) that allow us to receive information from the world for further processing, cognitive computers also have their "senses," in the form of cameras, microphones, and sensors.

This allows them to capture information from the outside. The processing of this is carried out by high-powered computers whose software works with natural language.

With this, cognitive computing makes use of various sciences within computing: recognition of images, sound, intelligence artificial, reasoning probabilistic, or learning automated, among others.

An indispensable condition that any cognitive computer must fulfill is learning.

An artificial intelligence must be able to face problems for which it has not been prepared or programmed. Only then can we say that it is "intelligent", and it becomes more so as time passes and we fill it with more "experiences".

Take, for example, the concept of a tree; a computer, by itself, is incapable of knowing what a tree is, although we can "teach" it. In this case, we would do it by image recognition.

And we should do it with the more images the better, so that in the future, and when capturing the image of a tree that there is no been processed previously, by comparison you will be able to “discern” that this is the same that a thousand have already taught you. times.

Humans, in fact, learn that way; Perhaps when we first see a new species of tree, we do not know what it is exactly, but even so, we will know that it is a tree. The process by which our brain reaches this conclusion it is the one that seeks to imitate cognitive computing.

To meet this challenge, the cognitive computer must be able to cope with the discipline of unstructured data analysis.

And what can be the applications of cognitive computing? Well, because it allows the computer to "understand" the communication and its context (let's put the non-verbal language of people), a cognitive system could perform the function of a multi-language translator in time real, being able to distinguish aspects of language (such as double meanings or jokes) and transmit them in their translations or, at least explain them and clarify them by interlocutor so there are no misunderstandings.

Photo: Fotolia - itskatjas