(ORDO NEWS) — The future of neural network computing may be a little more murky than we expected.
A group of physicists have successfully developed the ion circuit, a processor based on the movement of charged atoms and molecules in an aqueous solution, rather than electrons in a solid semiconductor.
Because it’s closer to how the brain communicates information, they say, their device could be the next step forward in brain-like computing.
“Ionic circuits in aqueous solutions tend to use ions as charge carriers for signal processing,” the team, led by physicist Woo-Bing Young of the Harvard John A. Paulson School of Engineering and Applied Sciences (SEAS), wrote in their report. new article.
“Here we report a water-ionic circuit… This demonstration of a functional ionic circuit capable of analog computing is a step towards more sophisticated water ionics.”
The main part of signal transmission in the brain is the movement of charged molecules, called ions, through the fluid. average.
Although the brain’s incredible processing power is extremely difficult to replicate, scientists believe a similar system could be used for computing: pushing ions through an aqueous solution.
It will be slower than regular silicon. based on calculations, but it may have some interesting advantages.
For example, ions can be created from a wide variety of molecules, each with different properties that can be used in different ways. But first scientists have to show that it can work.
This is what Jung and his colleagues have been working on. The first step was the development of a functional ion transistor, a device that switches or amplifies a signal.
Their latest achievement was to combine hundreds of such transistors to work together in the form of an ionic circuit.
The transistor consists of electrodes arranged in a bullseye pattern, with a small disk-shaped electrode in the center and two concentric ring electrodes around it. This interacts with an aqueous solution of quinone molecules.
A voltage applied to the central disk generates a current of hydrogen ions in the quinone solution. Meanwhile, two ring electrodes modulate the pH of the solution to a threshold by increasing or decreasing the ion current.
This transistor physically multiplies the “weight” parameter set by the gate of the ring pair with the voltage across the disk, resulting in an ion current. .
However, neural networks rely heavily on a mathematical operation called matrix multiplication, which involves multiple multiplications.
So the team designed 16-by-16 transistor arrays, each capable of arithmetic multiplication, to create an ionic circuit that can perform matrix multiplication.
“Matrix multiplication is the most common computation in neural networks for artificial intelligence,” Jung says. “Our ionic circuit performs matrix multiplication in water in an analog fashion, based entirely on an electrochemical mechanism.”
Of course, this technology has significant limitations. The 16 currents could not be resolved individually, which meant that the operation had to be performed sequentially rather than simultaneously, significantly slowing down an already relatively slow technology.
However, its success is a step towards more complex ionic computing: only by seeing the problem can we find solutions.
The next step is to introduce more molecules into the system to see if this allows the circuit to process more complex information.
“Until now, we have used only 3 to 4 ionic species, such as hydrogen and quinone ions, to provide ventilation and ion transport in a water ion transistor,” says Jung.
“It will be very interesting to use more diverse ionic particles and see how we can use them to enrich the content of the processed information.”
The ultimate goal, the team notes, is not to compete with or replace electronics with ionics, but to complement, perhaps in the form of a hybrid technology with the capability of both.
—
Online:
Contact us: [email protected]
Our Standards, Terms of Use: Standard Terms And Conditions.