Neuromorphic engineering, what’s that?
You’ve never heard of neuromorphic engineering? The truth is that it is very probable, and it is that although this discipline is not recent (it is more than three decades old), it is quite little known outside the academic and research fields.
It is thought that the inventor was Carver Mead, an American electrical engineer in the late 1960s. This technology arose from imitating the behavior of the nervous system of animals.
More than half a century ago, this advanced person had the ability to intuit that computers could at some point be inspired by the human brain to develop the capabilities of computer algorithms.
From this idea, over time research groups emerged, even causing some companies to start up different neuro-morphic engineering projects.
What is neuromorphic engineering?
Neuromorphic engineering or also called “neuromorphic computing” is a technology that is inspired by the way nature solves different problems.
Although many times it goes unnoticed by us, many things in our daily lives are inspired by nature, among others, the shape of some high-speed trains is inspired by the kingfisher.
The truth is that so many technological innovations have been achieved by following this line of thought that a whole discipline has emerged that is called “biomimetics”. This science takes as a reference the way in which nature itself has solved different problems over the years.
The truth is that neuro-morphic engineering is one of the most complex, and in recent years it has become an interdisciplinary subject that combines:
- Physics.
- Microelectronics.
- Biology.
- Mathematics.
- Computer science.
Thanks to these technological advances, similarities were seen between the behaviour of transistors and the way in which neurons communicate with each other by means of electrical impulses. And from that came the idea of designing a kind of hybrid architecture that combines digital and analog components. In other words, the idea was to build an artificial electronic brain that could process information in a similar way to an organic brain.
However, it can be said that at present, it has not yet been possible to create a neuro-morphic system that equals the human brain (the capacity of the brain of small mammals has been equalled).
What is the purpose of neuromorphic computing
Today’s neuromorphic chips are capable of incorporating tens of thousands of artificial neurons. Making them also capable of communicating with each other in a similar way as neurons do.
The way of processing information from neuromorphic chips is totally different from the way computers, mobile phones, tablets etc. are used. The truth is that the potential of this type of system lies in its ability to solve problems in a much more agile way than a conventional computer. How? By imitating the way the brain uses to solve problems.
Currently, neuro-morphic systems are mainly used for research since they manage to solve problems up to a thousand times faster than a classic microprocessor and also use up to ten thousand times less energy. The numbers make an impact, don’t they? However, despite these good results, neuro-morphic computing is not the best solution to all problems.
One of the technologies that is having a major influence on neuromorphic computing is artificial intelligence. Currently, algorithms are being designed that are capable of processing unstructured data in real time, and which also have a large amount of noise. In this way, neuromorphic engineering can be perfectly applied to: optimization problems, automatic learning, pattern recognition, satisfaction of limitations among many other things.
