Abstract:
Google announced that the Google Quantum Artificial Intelligence team has successfully overcome a key bottleneck in quantum computing: it has demonstrated that the "magic state cultivation" solution is feasible on the Willow processor, and the error is reduced by about 40 times compared to not using this technology. This breakthrough result was published today in the journal Science and is a key step in promoting the large-scale practical application of quantum computing.
Magic Bottleneck
As early as 2004, long before quantum computing became a practical engineering direction, researchers predicted a thorny bottleneck.
Want to realize
universal quantum computing
, to run a program of practical value, it must be able to perform a specific set of quantum operations. This is accomplished by first preparing so-called "magic states," so named because these quantum states can generate quantum operations beyond those required for classical calculations.But there's a problem here: producing high-quality, error-corrected monsters is extremely resource-intensive. Some estimates show that a quantum computer may use up to 95% of all computing power to generate magic states, leaving very few resources for actual calculations. There is another difficulty: once the magic state is used, it cannot be reused and is a scarce resource.
Therefore, for the past two decades, scientific researchers have been continuously reducing the resource cost of magic state preparation.
How to achieve this breakthrough
In 2024, Google Quantum AI researchers Craig Gidney, Noah Schulti and Cody Jones borrowed ideas from the literature and optimized them to develop what is now called
Magic State Cultivation
technology. This program is expected to significantly reduce resource consumption. The core idea is: the cultivation process is like a filter. Noisy quantum states are injected into a small area of the quantum chip, and a series of highly coordinated measurements are then performed to troubleshoot errors and eliminate unqualified samples. Craig's team proved that with this method, it is possible to "cultivate" quantum states with extremely high purity and low error, and can simultaneously convert them into larger-scale and practical coding forms.This solution has excellent theoretical results. If it can be implemented on quantum hardware, it can significantly reduce the scale of quantum computers required for practical tasks.
Like many new technology research and development, project implementation requires a series of technological innovation and engineering research.
One example: When testing the cultivation protocol for the first time, Google quantum AI researcher Emma Rosenfield observed data that seemed to defy the laws of physics. After months of troubleshooting, simulations, peer discussions, and consistency checks, she discovered that the problem stemmed from incorrect assumptions in the team’s commonly used measurement methods. Specifically, a type of noise caused by drift and temperature fluctuations in laboratory electrical equipment interferes with the validity of benchmark test results. The team needs to simultaneously adjust the theoretical framework and engineering plans. The research team added additional control methods to the system to weaken the impact of the noise; it also created a set of benchmark testing methods to use the technology itself as a measurement tool to solve the problem from the root.
Relying on this achievement and more algorithm innovations and engineering solutions, Google has pushed quantum measurement technology to a new top level. On Google Quantum AI's Willow processor, this protocol achieves up to 40 times error compression compared to quantum states without cultivation processing; the resulting T-state logic fidelity reaches a record high of 99.99%.
This result has now been published in Science. It is the first time to demonstrate a feasible solution to one of the most serious resource bottlenecks in quantum computing, and proves on real hardware that Google can efficiently prepare this key magic state.
Research findings and future prospects
Google has also initially verified that the theory proposed by Craig and others can be adapted to its quantum computing architecture. The team realized the switch between "color code" (a new type of quantum error correction code, very suitable for the cultivation of magic states) and "surface code" (a more mature error correction scheme), grafting the color code quantum state with a distance of 3 to the surface code with a distance of 5. This grafting protocol first came from a theoretical paper in 2024. This experiment confirmed that the purified quantum state can survive stably and be transmitted smoothly in the hardware.
Comments