Google managed to reduce error rates by 40 times with the magic state cultivation method it developed on the Willow quantum processor.
Google has taken an important step to reduce the high margin of error, one of the biggest obstacles to quantum computers. The company managed to significantly reduce error rates with a new method it tested on the Willow quantum processor.
The qubits used in quantum computers can easily make errors because they are very sensitive to environmental effects. Google researchers have developed a special technique called magic state cultivation to solve this problem.
High Accuracy with Magic State Cultivation
In order for quantum computers to perform complex operations, special quantum states protected by error correction systems are needed. These situations consume a lot of hardware and computing resources when prepared using traditional methods.
The traditional magic state distillation method focuses on obtaining fewer reliable results by processing a large number of erroneous states. This process can be so costly that it alone could consume 95 percent of the resources of future quantum computers.
The magic state cultivation method designed by Google researchers Craig Gidney, Noah Shutty and Cody Jones makes this process more efficient. The method is based on preparing a potentially faulty quantum state in a small area and checking it with a series of precise measurements.
While trials that fail the checks are eliminated, the quantum state obtained from successful ones can be expanded to a larger structure. Thus, high-accuracy results can be produced by working on a smaller system from the beginning.
This approach enables the production of high-fidelity cases directly, rather than processing a large number of imperfect cases and producing fewer quality results. In this way, the hardware and computing resources required by quantum computers are significantly reduced.
Overcoming Technical Challenges in the Willow Processor
In initial experiments led by researcher Emma Rosenfeld, noise and temperature changes from electrical equipment in the laboratory led to unexpected deviations. To solve this problem, the team improved the processor's control systems and developed a new measurement technique.
This technique, called kickback tomography, enabled a more accurate assessment of the reliability of quantum states by reducing the effect of noise. As a result of these studies, the error rate in quantum states prepared with the magic state cultivation method decreased approximately 40 times compared to the situation where the method was not used.
In the experiments, logical T states used in complex calculations were prepared with 99.99 percent accuracy. The team also demonstrated that they could transfer quantum states prepared using a technique called grafting between different error correction methods.
This transfer was achieved by switching from a color code with distance 3 to a surface code with distance 5. This development is important for the development of quantum computers that can use different error correction methods together in the future.
However, additional errors that emerged during the transition reveal that the method needs further development. These results obtained by Google prove that the magic state cultivation approach is applicable on real hardware.
