Study boosts quantum computing magic state efficiency
Cette publication existe aussi en Français
Scientists from Alice & Bob and Inria, France’s national institute for research in digital science and technology, have submitted a study for peer review illustrating the most hardware-efficient method to date for producing magic states on superconducting quantum computers. This method is a critical step toward realising practical quantum computing.
This work makes magic state preparation, a resource-intensive process, cheaper in terms of qubit requirements and considerably faster, consolidating the roadmap to the universal fault-tolerant gate set needed to run practical quantum computing applications at scale.
The new code developed was inspired by existing methods and tailored to work with the company’s cat qubits, which inherently protect against bit flip errors in quantum computations.
To achieve universal quantum computation with the ability to run any possible quantum algorithm, including those with known speedups over classical ones like Shor’s and Grover’s algorithms, a quantum computer must support a complete set of specific fundamental operations or gates. Scientists can implement some of these gates directly. However, other gates require the use of special quantum resources known as magic states. These states must be carefully prepared and manipulated during computation to enable “non-trivial” operations that cannot otherwise be executed directly, completing the universal gate set. In other words, magic states enable the last gates required to complete the set of operations needed to perform any possible quantum algorithm.
Magic state research
Optimal magic state distillation would require complex 3D arrangements of qubits, leading to architectures too hard to engineer in solid-state QPUs. Recent research from Google scientists theorised that magic state production with low error rates would require just 463 qubits in a novel 2D architecture, combining decades of quantum error correction research and overcoming the 3D architecture challenge to reduce the qubit needs significantly.
The researchers at Alice & Bob and Inria have managed to “unfold” the 3D code, much like flattening a box, into an even more practical 2D layout. The unfolded code for magic state preparation, internally nicknamed the “Heart Code” for its distinctive shape, is only possible thanks to noise-biased qubits, such as the cat qubits developed by Alice & Bob.
The unfolded code simplifies overall operations and reduces overhead, requiring only 53 qubits to produce one magic state. The result is a remarkable 8.7× reduction in qubit requirements compared to the leading approaches while requiring 5× fewer quantum error correction cycles, making it approximately 5× faster than state-of-the-art superconducting platforms for the same error rate of less than 1 in a million.
“The most advanced players in quantum have lined up breakthrough after breakthrough to reduce the costs of magic state preparation, and it’s exciting to see how our work further improves the state of the art on noise-biased qubits,” said Diego Ruiz, author of the paper and a Ph.D. student at Alice & Bob and Inria.
Importantly, this magic state protocol only uses components that are already required for Alice & Bob’s quantum error correction architecture, including both the fundamental operations and the physical cat qubits. Hence, no new developments are needed, and the company can focus on reaching the targeted performance.
This research highlights the promising advantages of cat qubits’ noise bias to reduce the hardware overhead needed for practical quantum computing, not only when combined with advanced error codes like LDPC outlined in another recent paper by Alice & Bob researchers, but now also in the context of magic state preparation.
“The community is finally solving this looming obstacle to useful quantum computers, with some players even achieving the first proof-of-concept magic state preparation in experimental settings,” said Théau Peronnin, CEO of Alice & Bob. “Capitalising on the cat qubits, this work further de-risks our roadmap while showcasing the long-term advantages of our universal platform.”
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
