Quantum Computers: 2.5% Error Threshold Reached

Julio C. Magdalena de la Fuente and colleagues at Massachusetts Institute of Technology have created a non-Pauli stabilizer code that completes a universal gate set on established topological codes within a strictly two-dimensional framework. The code exhibits a key decoding threshold of approximately 2.5% under realistic error conditions, approaching the performance of more complex decoders. Their analysis confirms exponential suppression of logical error rates, suggesting that incorporating non-Clifford logic into two-dimensional codes could achieve performance comparable to dedicated quantum memory systems.

Near-threshold decoding enables two-dimensional fault-tolerant quantum computation

The new just-in-time decoder reaches a threshold of approximately 2.5%, a substantial improvement over the 1.8% previously possible with naive decoders. This result nears the 2.9% threshold of decoders utilising full syndrome history. This breakthrough unlocks fault-tolerant quantum computation in two dimensions using non-Pauli stabilizer codes, previously hampered by the need for complex distillation processes. Topological codes, such as surface and toric codes, are promising candidates for fault-tolerant quantum computation due to their inherent robustness against local errors. However, implementing a universal set of quantum gates, meaning the ability to perform any quantum computation, within these codes is challenging. Standard topological codes primarily support Clifford gates, which are relatively easy to implement, but lack the power to perform non-Clifford gates, essential for universal computation. These non-Clifford gates, like the Toffoli gate, require complex operations such as magic state distillation, which introduces significant overhead in terms of qubit requirements and operational complexity. The development of this non-Pauli stabilizer code offers a potential solution by enabling the implementation of non-Clifford gates directly within the two-dimensional code structure, reducing the need for extensive distillation. The decoder reliably determines mid-circuit X corrections, essential for these codes, and demonstrates comparable performance to standard decoding methods.

Exponential suppression of logical error rates was confirmed through finite-size scaling analysis, suggesting this approach could rival the performance of two-dimensional quantum memory and broaden the applicability of CSS codes. Finite-size scaling is a crucial technique in quantum error correction research, allowing researchers to extrapolate the performance of the code as the system size increases. By analysing how the logical error rate decreases with increasing code distance, a measure of the code’s ability to protect information, researchers can determine the code’s asymptotic performance and assess its viability for large-scale quantum computation. The observed exponential suppression indicates that the code effectively protects quantum information from errors, even as the system grows in complexity. CSS codes, a class of quantum error-correcting codes, benefit from this advancement, potentially expanding their utility in various quantum computing architectures. Knowledge of X corrections improved Z error decoding, raising the threshold to around 2.2%. This synergistic effect highlights the potential for optimising decoding strategies by leveraging information gained from correcting one type of error to enhance the correction of others. The improvement from 2.5% to 2.2% demonstrates that a holistic approach to decoding can yield significant performance gains. Achieving strong error correction represents a monumental step towards practical quantum computers, although a performance bottleneck arises from reliance on a secondary decoding stage for ‘Z’ errors. While the initial focus has been on efficiently correcting X errors, the subsequent decoding of Z errors remains a critical challenge. This secondary stage introduces additional complexity and can limit the overall performance of the error correction scheme. The research team acknowledges that refining ‘heralding strategies’ could yield further improvements. Heralding strategies involve measuring ancillary qubits to gain information about the errors that have occurred, allowing for more informed decoding decisions. Optimising these strategies could potentially reduce the overhead associated with the secondary decoding stage and further enhance the overall performance of the error correction scheme.

Topological codes face a challenge in achieving a low-overhead universal gate set with limited connectivity. A non-Pauli stabilizer code can complete a universal gate set on topological toric and surface codes in two dimensions. Toric and surface codes are particularly attractive due to their relatively simple structure and high threshold for error correction. However, implementing a universal gate set requires extending the capabilities of these codes beyond Clifford gates. This new non-Pauli stabilizer code provides a mechanism for achieving this extension without significantly increasing the complexity of the code or the overhead associated with its implementation. Unlike conventional Pauli codes, fault-tolerant syndrome extraction using this code requires mid-circuit X corrections. Syndrome extraction is the process of identifying errors that have occurred in the quantum system without directly measuring the encoded quantum information. In conventional Pauli codes, this process relies on measuring Pauli operators (X, Y, and Z). However, the non-Pauli stabilizer code requires measuring mid-circuit X corrections, which introduces additional challenges in terms of implementation and decoding. A just-in-time matching decoder was constructed and benchmarked to reliably determine these corrections, achieving a threshold of approximately 2.5%, similar to a decoder with full syndrome access. The just-in-time decoder is designed to process the syndrome information as it becomes available, rather than storing the entire syndrome history. This approach reduces the memory requirements and computational complexity of the decoding process, making it more practical for large-scale quantum systems.

Non-Clifford logic in 2D codes may perform comparably to 2D quantum memory. A pathway to universal quantum gates is now established within existing topological code frameworks, offering a complementary approach to qubit reduction. Qubit reduction techniques aim to minimise the number of physical qubits required to implement a given quantum algorithm. This new approach provides an alternative strategy by enabling the implementation of universal gates within existing topological code frameworks, potentially reducing the need for extensive qubit reduction. This formalism also extends to other code families, broadening its potential impact on fault-tolerant quantum computation and offering adaptability for use with different quantum architectures. The principles underlying this non-Pauli stabilizer code are not limited to toric and surface codes; they can be applied to other topological code families, potentially expanding its impact on the field of quantum error correction. Furthermore, the adaptability of this approach allows it to be used with different quantum architectures, such as superconducting qubits, trapped ions, and photonic qubits. By developing a just-in-time decoding strategy for non-Pauli stabilizer codes, error correction thresholds approaching those of conventional methods were achieved, sidestepping the need for complex code distillation. This represents a significant advancement in the field of quantum error correction, as it demonstrates that it is possible to achieve high performance without relying on computationally expensive code distillation techniques. Incorporating knowledge of these initial corrections improves the decoding of subsequent errors, raising the potential for performance comparable to dedicated quantum memory architectures, and ultimately establishing a pathway to implement universal quantum gates within the established framework of two-dimensional topological codes.

The research demonstrated a method for performing universal quantum gates within two-dimensional topological codes, achieving a threshold of approximately 2.2 per cent. This is significant because it offers an alternative to reducing the number of qubits needed for quantum computation, instead focusing on improving gate implementation within existing code structures. Researchers developed a just-in-time decoding strategy that allows for error correction thresholds comparable to conventional methods, without the need for complex code distillation. The findings suggest that non-Clifford logic in these 2D codes could perform similarly to 2D quantum memory, and the principles are applicable to other topological code families.