A group of scientists led by Google announced a breakthrough in the field of quantum computing. They again demonstrated quantum supremacy – the ability of a quantum computer to perform calculations that a classical one cannot – but this time they focused on the accuracy of the calculations. Scientists have also shown that there are phase transitions in computational processes, which opens the way to the further development of quantum technologies.
Image source: Google, Nature
Back in 2019, Google announced that it had achieved quantum supremacy, causing heated debate in the scientific community. IBM then questioned this result, arguing that classical algorithms could be optimized to solve similar problems. In a new paper published in the journal Nature, the scientists described an experiment using Random Circuit Sampling (RCS), during which the 67-qubit system performed 32 calculation cycles. The emphasis is not on quantum superiority, but on the fact that even in the presence of noise—the main limitation for quantum processors and the main cause of computational errors—it is possible to achieve computational advances that exceed the capabilities of classical systems. This proves that quantum computing is approaching the practical application phase.
The term “quantum supremacy” causes some controversy in the scientific community. Some researchers prefer to use the terms Quantum Utility or Quantum Advantage. The latter term implies not only the theoretical superiority of quantum devices, but also their practical benefits. Unlike quantum supremacy, which is not related to actual utility for tasks, quantum advantage involves performing tasks faster and more efficiently than classical computers.
Quantum processors, despite their potential, remain extremely sensitive to external noise, such as temperature fluctuations, magnetic fields or even cosmic radiation. These noises can significantly reduce the accuracy of calculations. In the Google study, scientists studied the effect of noise on the operation of quantum devices and conducted an experiment that allowed them to study two key phase transitions: a dynamic transition that depends on the number of cycles, and a quantum phase transition that affects the error rate. The results showed that even in the presence of noise, quantum systems of the NISQ era can achieve computational complexity beyond the reach of classical systems.
Phase transitions in random circuit sampling (RCS). The graph illustrates two phase transitions. The first is from a concentrated distribution of bit strings over a small number of cycles to a wide or anti-concentrated distribution. The second is a transition under noise conditions, in which a high error per cycle leads to a transition from a system with complete correlation to a representation in the form of several unrelated subsystems
The random circuit sampling (RCS) method used in the experiment has previously been criticized for its simplicity and apparent uselessness. However, Google emphasizes that RCS is a key method for moving towards problems that cannot be solved on classical computers. This method optimizes quantum correlations using iSWAP-type operations, which prevents the simplification of classical emulations. Thanks to this approach, Google was able to clearly define the boundaries of the capabilities of quantum systems, stimulating competition between quantum and classical computing platforms.
The study also examines the prospects for the practical use of quantum processors. One of the first examples would be certified generation of truly random numbers, which requires high computational complexity and robustness to noise. Sergio Boixo, head of quantum research at Google, said in an interview with Nature: “If quantum devices can’t demonstrate an advantage with RCS, the simplest use case, then it’s unlikely they can do so in other problems.”
Google Quantum Computing Roadmap
Google’s work represents a significant contribution to the development of quantum technologies. While the practical application of quantum devices remains challenging, areas such as certified random number generation could be the first step towards their commercial use. Despite the difficulties associated with noise, Google’s experiments show that the transition from theoretical research to practical applications of quantum devices is becoming increasingly possible.