Do you know Neuromorphic technology? This new computing platform integrates physics, math, computing and neuroscience to create computers inspired by the human brain, simulating artificial synapses that surpass the world’s fastest supercomputers!
The world’s fastest supercomputer, the Sunway TaihuLight, performs more calculations per second than a human brain, but consumes about 800.00 times more energy. That is to say, a human brain is not faster than a supercomputer – in fact, it is much slower. But it has the advantage of having a network of neurons that works simultaneously to solve a large number of problems at the same time.
The artificial neuron consists of a superconductive wire embedded in an LED, the two elements act as detector and signal transmitter, thus generating a “synapse”.

The contribution of a Brazilian.
A few days ago, a team from Stanford University announced the creation of an artificial organic synapse that can enable neuromorphic processors. The creation of this component, called enode (Electrochemical Neuromorphic Organic Device), was attended by the Brazilian Gregório Couto Faria, from the Institute of Physics of USP in São Carlos (IFSC-USP).
According to Gregory, Enode presents a typical property of neural structures. Each structure of our brain manifests about 100 states of potentiation, which correspond to different levels of memory. In this experiment, it is possible to change the conductivity of the active material to then access differentiated patterns of memory.

As in a human brain, a conjugated polymer is capable of conducting not only electrons, but also ions, making it possible to act as an electric current transducer for ionic current. When one neuron interacts with another, it opens channels of ion passage, concentrating the surrounding environment. In reproducing this function, an interface between artificial systems and living systems is established.

Today computers and cell phones have a machine-learning technique that works through trial and error. It was these techniques that enabled programs to defeat world chess champion Garry Kasparov and allow the Watson supercomputer to defeat both champions of the Jeopardy program in 2011. Currently, in 2016 the AlphaGo program defeated Go champion Lee Sedol.
What we are looking for now is another type of learning, rather than an accumulation of bytes in gigantic machines, multilevel processing units capable of resembling the flexibility of the human brain.

Did you already know this technology? Tell us in the comments section!