Despite the fact that we often talk about neural networks and artificial intelligence, is still what we call “electronic brain,” its structure is radically different from the structure of the human brain. But there are so-called neuromorphic chips, which are much faster than their electronic counterparts, but much less developed, which hampers the development of this technology. Recently, thanks to the joint work of scientists from the USA and France, managed to create a structure very similar to human synapse (the point of contact between two cells, serving for transmission of nerve impulses). This development could be the first step towards creating a true electronic brain.
Artificial synapse in its properties almost identical to natural. The main property that prevents the artificial brain to be like our, is “neuroplasticity”, i.e. the possibility to change under the influence of experience, as well as to restore after damage or in response to external stimuli. In other words, thanks to this, we can learn new information and gain experience.
When you design a new artificial synapse, the scientists created a precise mathematical model, based on which it was made detailed calculations of the shape and structure of a microscopic device. The artificial synapse is designed to develop Autonomous self-learning systems. On the basis of the synapses of a new type, as I hope the researchers can create a real electronic brain with the function of artificial intelligence at the hardware level and not software which is used now.
On the left — the principle of operation of the synapse human right — an artificial synapse
The basis of the neuroplasticity of the synapse is the tunneling on the basis of the ferroelectric material. The conductivity of the phase is incremented each time it passes through a pulse of electrical current. Different actions generate different kinds of impulses, which allows the synapse to remember recurring events and “gain experience”. The areas that are not used in conductivity with time are erased, “forgotten”. Now researchers are going to create the first prototype chip-on-chip which will be formed a complex neural network capable of self-instructive.
Artificial synapse, which can be a component of an electronic brain
Vladimir Kuznetsov