Despite the fact that the term “neural network” can be applied to the anatomical structure and computing system, these neural networks, the differences are more than similarities. First and foremost, this is due to the extremely complicated structure of the neural plexus of the brain. But that could change thanks to the development of scientists from the National Institute of standards and technology, USA. Their invention can become a new stage in the development of the technology of construction of neural networks.
The fact that the human brain consists of billions of neurons each connected to tens of thousands of other neurons. This complex structure in a rather simplified form used as a basis for constructing artificial neural networks, with the only difference that the current electronics cannot cope with such a complex routing and number of links between elements of the neural network have to be reduced in the tens and hundreds of times, which affects performance. Of course, there are now projects aimed at emulation of the brain, but they run into the power limit of existing technologies.
Scheme of structure of the chip. The three-dimensional structure provides a complex routing scheme and the transmission speed signal is required to simulate the brain.
The team from the National Institute of standards and technology USA offers a slightly different approach. Scientists want to use light instead of electricity as a means of signal transmission. For this we developed a special chip which distributes the optical signals on a miniature grid in different directions. Through the use of a new type of signal, managed to overcome the problem of the connection elements of the neural network, vertically stacking two layers of photonic structures. They limit the light lines for directing optical signals approximately same as the wires transmit an electrical impulse. This approach allows you to create complex routing required to simulate the neural structures of the brain and increase as the speed of the signal and the number of links between network elements. In addition, the system is very well scalable and easy to expand in the future. And if “normal” neural networks are making progress in self-learning and pattern recognition, it is difficult to imagine what will be able to structure that simulates the brain.
Created a neural network which simulates the structure of brain
Vladimir Kuznetsov