Google Tests a Way to Cut Errors in a Key Quantum Computing Resource
Published on · Joaquín Dati
Quantum computers need to do more than perform calculations: they also have to keep errors under control. Google has demonstrated a technique on its Willow processor that cuts errors roughly fortyfold when preparing special quantum states. The experiment tackles a problem that could otherwise consume much of a quantum computer’s capacity.
A resource that can only be used once
These resources are called “magic states.” They enable operations needed for general-purpose quantum computing, but producing high-quality versions takes substantial resources. Each state can only be used once. Some estimates cited by Google suggest that making them could occupy up to 95% of a quantum computer’s capacity.
The new approach acts as a filter. It starts with an error-prone state, runs a coordinated series of measurements and rejects faulty attempts. On Willow, the team produced one type of magic state with 99.99% fidelity, a measure of how closely it matches the intended state.
Published in Science, the work turns a theoretical proposal into a demonstration on real hardware. The researchers still need to show that the technique works at a larger scale, and other engineering challenges remain before it can support practical quantum computing.