The human brain is still the most powerful organic processor. It's no surprise that artificial intelligence creators use it as an example. The neural networks they create perceive information in a very similar way. Now scientists at Stanford University have taken this one step further by developing an organic artificial synapse - a link between two neighboring neurons.
AI based on neural networks is already able to classify images and sounds, play chess and other intellectual games on a par with humans, create works of art, and drive a car. The main problem with intelligent software systems is that they are very energy intensive.
Stanford scientists managed to create an artificial synapse that mimics the process of simultaneous processing and storage of information. An artificial synapse built on the basis of a battery works like a transistor. The device consists of two thin films and three electrodes, where salt water is the electrolyte between them. Electrical signals are transmitted between two of them, and a control signal is applied to the third.
While conventional transistors have only two states, an artificial synapse is capable of accepting up to 500 different programmed states, which exponentially increases processing power.
When switching between states, some of the energy is borrowed from adjacent systems, thus significantly reducing consumption. However, an artificial synapse still uses 10, 000 times more energy than an organic one. But even so, the research results represent a step in the right direction, which will enable the creation of new generation computing devices in the future.