Development of artificial neural network 'true' by memory resonance

An artificial neural network built on "memory resistors," plastics memristors with brainstorming " structure " has been successfully developed by Italian and Russian scientists, promising to replace it. For traditional silicon semiconductors, the idea of computer simulating human brain and artificial intelligence comes closer to reality.

Many projects have already been developed to develop artificial neural networks with self-learning capabilities like the human brain. The biggest problem, however, has not been overcome yet, they all use outdated silicon semiconductors and have certain limits. So how to create "brain-like" chips? For that purpose, Italian and Russian researchers created a neural network with "memory resistors , " a type of resistor capable of remembering its previous impedance.

Italian and Russian researchers have created a neural network with "memory resistors."

This new generation of resistors is said to be capable of functioning as neural circuits and are ideal components for the development of neural " neural " neural networks, where the process of transmitting the effect will be longer. And importantly, this chip generation has the advantage of materials and technology that allows it to be made in very small dimensions (theoretically 10nm) without the use of too complex substances. In this way, you can design a neural net neuron as small as a regular chip without requiring additional new tech support.

Still, scientists admit that these are only the first steps in this technology. Currently a prototype of the chip has been created, but the size is quite large, to 1mm and it only learned the most basic tasks. However, its potential is enormous. In addition to creating neural networks that act like the real brain, it also allows the creation of thought-provoking and analytical robots. This is a promising technology and if It may soon change the face of the computer industry in the future.