Meet a new contender in the world of video games: a non-living hydrogel.
This seemingly simple material has surprised scientists by demonstrating its ability to learn and improve its gameplay.
A team of robotics engineers led by Vincent Strong of the University of Reading carried out this Pong gaming experiment using hydrogels.
“Hydrogels are polymer structures that retain water, much like gelatine deserts or seaweed carrageen. The hydrogel that we used was ionized polyacrylamide, which is more resistant to heat and electricity allowing us to interface it with the Pong game within the computer via our multi-electrode array,” Strong explained to Interesting Engineering (IE).
Pong is a classic game featuring two paddles and a ball. The goal is to keep the ball from going past the paddle.
Hydrogel electrically stimulated
Researchers connected a hydrogel to a virtual Pong environment and stimulated it electrically for its movement to play this game.
The “electro-active polymer” hydrogel was used for this experiment. This type of hydrogel reacts to electrical stimulation due to the presence of ions in the surrounding medium.
The hydrogel’s paddle was controlled by the flow of ions inside its structure, while the ball’s position was encoded via electrical stimulation.
“The key element is that the hydrogel we use is ionized, meaning it contains free-floating charged ions that can be influenced by an electric field, causing them to migrate and pull water molecules with them, which results in shape changes within the hydrogel,” Strong added.
The researchers measured the hydrogel’s improvement in gameplay by tracking its score during the game. The game’s scoring system provided points for successful paddle strikes while resetting the score to zero for missed ball shots.
The accuracy increased over time
Over time, the hydrogel was able to adapt and increase ball striking accuracy by up to 10%. Interestingly, this enabled the hydrogel to play longer rallies.
The study team highlights that the experiment demonstrates how non-living objects may use “memory” to learn about their surroundings.
“We attributed this improvement to the hydrogel’s ‘memory’ of ball motion, essentially as the game goes on the hydrogel gains experience about how the ball moves and can better intercept it with the paddle. The ions and water molecules moving within the hydrogel form a sort of abstract representation of how the ball moves,” Strong told IE.
Strong went on to describe what is known as the hydrogel’s “memory.” This is attributed to “hysteresis” in ion motion.
When electrically excited, the hydrogel expands and takes longer to shrink again. “This means that there is a form of hysteresis in the ion motion, as each consecutive stimulation moves the ions less and less as they gather. This acts as a form of memory, as the result of each stimulation on the ion’s motion is directly influenced by previous stimulations and ion motions,” he added.
This behavior is analogous to a memory function, in which past experiences influence future actions.
Only took 20 minutes to reach peak hits
This experiment was inspired by previous studies, which showcased that electrically activated brain cells can learn to play Pong.
“The basic principle in both neurons and hydrogels is that ion migration and distributions can work as a memory function that can correlate with sensory-motor loops in the Pong world. In neurons, ions run within the cells; in the gel, they run outside,” said Yoshikatsu Hayashi of the University of Reading in a press release.
RECOMMENDED ARTICLES
While neurons playing Pong acquired peak ball skills in around 10 minutes, the hydrogel required closer to 20 minutes to reach its full Pong capability.
Neural networks serve as the foundation for the majority of modern AI algorithms. With this new study, the researchers have demonstrated that hydrogels embody a different type of “intelligence.” They believe this could be employed to create less complex algorithms in the future.
Moving forward, the researchers plan to explore the hydrogel’s “memory” further by examining its mechanisms and testing its ability to accomplish other tasks.
Mrigakshi Dixit Mrigakshi is a science journalist who enjoys writing about space exploration, biology, and technological innovations. Her work has been featured in well-known publications including Nature India, Supercluster, The Weather Channel and Astronomy magazine. If you have pitches in mind, please do not hesitate to email her.
0