A simple artificial intelligence (AI) system that consists of a jelly-like material and is connected to electrodes can 'learn' to play the classic video game pong and improve over time, so a study published today 1 .

The results are a first step to show that synthetic materials can use a fundamental form of 'memory' to increase performance, says Brett Kagan, Chief Scientific Officer at Cortical Labs in Melbourne, Australia. "The system shows in a similar way memory how a river bed records the memory of a river," he says.

In 2022, KAGAN and his colleagues 2 href = "https://www.nature.com/articles/d41586-03229-Y" Data-Track = "Click" Data-Label = "https://www.nature.com/articles/d41586-03229-Y" Data-track-category = "body Link "> Neurons in a bowl - known as Dishbrain - can learn to play the table tennis -like video game through electrical stimulation. Inspired by this work, Yoshikatsu Hayashi, a biomedical engineer at the University of Reading, Great Britain, and his colleagues, and his colleagues, whether a non-biological material could also master.

Hayashi and his colleagues turned to hydrogels - jelly -like materials that are used for various applications such as components for soft robots - and contain the charged particles called ions. When this hydrogel is electrically stimulated, the ions move through the material and pull water molecules with it, which changes the hydrogel. This change in the distribution of the ions influences the next arrangements of particles, says Hayashi.

"It's like a physical memory."

animated sequence of a computer, the hydrogel plays the video game pong.

to test whether this 'memory' could enable the hydrogel to play pong , the researchers used electrodes to connect the material to the game on a computer. The game was divided into a grid of six squares that corresponded to six couples of electrodes. Every time the ball went through one of the squares, the corresponding electrodes sent an electrical signal to the hydrogel, which changed the ion position. Then sensor electrodes measured the electrical current of the surrounded ions and returned this information back to the computer, which interpreted it as an command to move the game racket into a new position. Over time, this formed into a fundamental 'memory', since the movements of the ions were influenced by their previous shifts.

Fast learner

At the beginning, the hydrogel hit the ball about half the time, but increased its hit rate to 60% in about 24 minutes, which indicates that the material updates its 'memory' of the movements of the ball using the ion pattern. The improved performance also led to longer rallies - the times when the ball is involved.

The researchers carried out control experiments in which the hydrogel received incorrect information about the position of the ball or was 'blind' operated by not stimulating it at all. This meant that the positions of the ions of the gel did not exactly reflect the screen game. Under these conditions, the pong game of the hydrogel showed no improvements, which indicates that it only gets better if the correct information is added.

The hydrogel dominated pong not as quickly as Dishbrain, which needed less than 20 minutes to do its best. "Hydrogels are a much simpler system," says Hayashi. But he adds that the results indicate that hydrogels have further arithmetic skills that researchers could support in developing more efficient algorithms.

"The authors have followed a creative approach to transfer concepts from neuroscience to a more physical system," says Kagan. But more work has to be done to show that hydrogels can actually 'learn', he adds.