Quantinuum New Quantum Flag Fault-Tolerant State Preparation Improves Error Rate to

Ben Criger and colleagues from Quantinuum achieved a logical block error rate of approximately 0.00014 using a method for quantum error correction on Quantinuum’s System Model H2 computer. The researchers optimized circuits designed to prepare stabilizer states using integer linear programming, a mathematical technique applied to the construction of circuits for fault-tolerant state preparation. This approach allows for circuits with comparable or fewer gates than existing methods while also detecting up to three errors during computation. Testing involved 10,000 shots, with roughly 1.6% post-selected due to weight-two errors, representing a trade-off between performance and data yield. The work demonstrates a step toward more reliable quantum computation by reducing gate counts and improving error detection.

The researchers optimized circuits that prepare stabilizer states, a crucial step in both initializing logical qubits and building error correction systems. This optimization also enables detection of up to three errors within the quantum system, enhancing the robustness of calculations. The researchers tested their approach with 10,000 shots on the System Model H2, acknowledging a trade-off inherent in current quantum systems; roughly 1.6% of shots were post-selected due to weight-two errors. This post-selection process, while reducing the yield of usable data, contributes to improved error correction performance. The team specifically applied this method to derive a Steane error correction gadget for the two-block group algebra code, demonstrating its practical application. The authors state in their paper that reducing the number of gates needed to prepare a stabilizer state fault-tolerantly can simultaneously reduce computation time and increase reliability, highlighting the dual benefits of their optimization strategy.