A group of physicists has successfully developed an ion circuit, a processor based not on the movement of electrons in a solid semiconductor, but on the movement of charged atoms and molecules in an aqueous medium. solution. Because it’s closer to how the brain communicates information, they say the device they’ve developed could be the next step in brain-like computing.
“Ionic circuits in aqueous solution use ions as charge carriers for signal processing. We report a water ionic circuit… The presentation of a functional ionic circuit capable of analog computing is a step towards more sophisticated water ionics,” writes Harvard University physicist Woo-Bin Jung and his team in a new study.
Much of the signaling in the brain is the movement of charged molecules, called ions, in a fluid medium. Although it is extremely difficult to replicate the brain’s incredible processing power, scientists thought that a similar system could be used in computing: pushing ions through an aqueous solution. This will be slower than traditional silicon-based computing, but could have interesting benefits, ScienceAlert reports. differently. But first, scientists must show that it can work.
Jung and his colleagues are working on just that. The first step was the development of a functional ion transistor, a device that switches or amplifies a signal. In their latest development, hundreds of these transistors have been combined to work together in an ionic circuit.
Bull’s eye transistor electrode arrangement has a small disk-shaped electrode in the center, surrounded by two concentric ring electrodes. This is due to the aqueous solution of quinone molecules. A voltage applied to the central disk generates a current of hydrogen ions in the quinone solution. Meanwhile, the two ring electrodes change the pH value of the gate solution, increasing or decreasing the ion current.
This transistor physically increases the “weight” parameter determined by the gate of the ring. pair on the voltage on the plate, and how the ion current gives the answer. However, neural networks rely heavily on a mathematical operation called matrix multiplication, which means multiplication multiple times. So the team designed 16×16 transistor arrays, each capable of arithmetic multiplication, to create an ionic circuit capable of matrix multiplication.
“Matrix multiplication is the most common computation in artificial intelligence neural networks. Our ionic circuit performs matrix multiplication in water in a similar way, based entirely on the electrochemical mechanism,” says Jung.
The technology, of course, has significant limitations. The 16 currents could not be broken down individually, which meant that the operation had to be performed sequentially rather than simultaneously, significantly slowing down the already relatively slow technology. But success is a step towards more complex ionic computing: we can only find a solution if we see a problem.
The next step is to introduce a wider range of molecules into the system to see if they can, this allows the circuit to process more complex information. “Until now, we have only used 3-4 types of ions, such as hydrogen and quinone ions, to provide venting and ion transport in the water ion transistor. It will be very interesting to use more diverse types of ions and see how we can use them to enrich the information content to be processed,” Jung said. in the form of a hybrid technology that combines the capabilities of both
hardware, software, tests, interesting and colorful news from the IT world by clicking here!