For many years, the development of artificial intelligence has followed a simple rule: make it bigger with more groups, more connections, more computing power. However, a new study suggests otherwise.
Instead of going big, the study’s authors built a very small object – a quantum system with nine spins interacting with an atom – and asked it to take on problems that normally require much larger machines.
The result was unexpected. This little system did not stand still; outperformed traditional machine learning methods with thousands of nodes in tasks such as predicting temperature patterns over several days.
“This represents the first experimental demonstration of quantum machine learning over large-scale classical models of real-world tasks,” the study authors note.
So does this mean that scientists have been approaching quantum computing the wrong way all along?
Letting the system think for itself
One of the biggest challenges in quantum computing is control. Many methods rely on carefully designed quantum circuits, where each step must be performed precisely.
However, today’s quantum devices have small disturbances (noise) from the environment, which can quickly destroy these calculations. This is one of the reasons why real-world applications remain out of reach.
The researchers went back and tried something different. They borrowed a concept from machine learning called reservoir computing.
In this way, you do not control the system. You feed in data, let the system do the math, and read the results. This intelligence comes from the way the system works naturally and reprograms it.
“Quantum reservoir computing offers high potential for machine learning,” the study authors say.
To do this, the team used nuclear energy techniques to manipulate the nine spins of the atom — essentially tiny magnets at the quantum level. These spins interact, creating an internal state that is constantly changing. When data is entered into this system, it is not static. It spreads, mixes, and changes in complex ways.
This is where quantum physics makes a difference. The system can exist in many places at the same time and create strong internal relationships. As a result, even a small number of elements can produce very good behavior patterns.
Instead of orchestrating each step, researchers let these forces unfold and extract actionable information from the results.
To change the defect in the structure
In many quantum experiments, dissipation (the process by which a system loses energy) is the problem of elimination. It deletes data and generates errors, but here, it is used on purpose.
Why? Because predictive functions rely on memory. To predict what will come next, the system must retain traces of what came before—but not too much. If it remembers everything equally, it gets stuck. If it is forgotten quickly, it loses its meaning.
Dissipation provided a natural way to balance this. It gradually removed the old information while allowing the latest information to influence the system more strongly. In other words, what is usually thought of as noise became a mind control tool.
From standards to real weather
To check if their method works, the researchers first turned to a standard test called NARMA, which is often used to test weather forecasting methods. The quantum setup gave its first big result here, reducing prediction errors by one to two orders of magnitude compared to earlier experimental methods.
However, benchmark tests are one thing, real world data is another. So the study authors turned to weather forecasting, focusing on temperature conditions over several days. Despite its simplicity, the nine-spin system was able to follow these patterns with remarkable accuracy.
The most striking comparison was when they pitted it against an older model known as an echo state network—a well-known way of using storage computing. While the old system scaled up to thousands of nodes, the much smaller quantum system still performed well on multi-day forecasts.
“In long-term weather forecasting, our quantum pool achieves higher accuracy than traditional pools with thousands of nodes, suggesting that practical advantages in real-time forecasting can be achieved with modern quantum devices,” the authors of the study said.
Thinking on the way to functional quantum mechanics
This work marks a turning point in how quantum computing can develop. Instead of waiting for large, perfectly controlled machines, researchers may be able to extract value from small, imperfect systems right now—by harnessing their inherent strengths rather than fighting them.
“We present a quantum reservoir computing method based on quantum spin systems, using many-body interactions to produce energy storage, thus avoiding the problems of deep circuits,” the authors of the study added.
That said, the process is still in its infancy. The current system is limited in scope and has only been tested on certain types of problems. It’s not a general-purpose computer, and scaling it up will bring new challenges.
However, the study provides a very important lesson that progress does not come by adding more. It comes from using what you already have in a smart way.
The study was published in the journal Physical Examination Letters.
#quantum #system #atoms #beats #network #thousands #nodes