MITS: A Quantum Sorcerer's Stone for Designing Surface Codes

Area of Science:

• Quantum Computing
• Quantum Information Science

Background:

• Quantum Error Correction (QEC) parameter optimization is vital due to diverse physical noise in quantum computers.
• Traditional forward simulation methods for deriving logical error rates can be resource-intensive.
• Manual adjustment of QEC parameters is inefficient due to daily quantum error rate fluctuations.

Purpose of the Study:

• To introduce MITS, a novel reverse engineering tool for STIM.
• To automate the determination of optimal QEC parameters based on specific quantum hardware noise models and target logical error rates.
• To minimize qubit and gate utilization by aligning QEC settings with hardware constraints.

Main Methods:

• Developed MITS, a reverse engineering tool that interfaces with STIM.
• Investigated various heuristic and machine learning models for parameter optimization.
• Utilized XGBoost and Random Forest regression models.

Main Results:

• MITS automatically determines optimal QEC settings for given noise models and target error rates.
• XGBoost and Random Forest models demonstrated high effectiveness, achieving Pearson correlation coefficients of 0.98 and 0.96, respectively.
• The approach minimizes qubit and gate usage by precisely matching logical error rates with hardware constraints.

Conclusions:

• MITS offers an efficient solution for optimizing QEC parameters in quantum computing.
• Automated parameter tuning using machine learning significantly improves resource efficiency.
• This method is crucial for advancing practical quantum error correction strategies.