Researchers Cut Quantum Error Rates with Optimised Atom Loss

A fidelity improvement of 5.3x has been achieved by optimising atom loss in neutral-atom quantum error correction systems. Compilers previously focused on minimising total atom loss, however Xinyi Li at Stevens Institute of Technology and colleagues developed CAST, a new compiler optimisation pass that instead minimises ‘decoder-weighted harm’ by strategically managing where atoms are lost within the quantum code. This approach recognises that deposition location key impacts logical error rates, lowering them in 35 of 48 tested settings. An optimisation method for neutral-atom quantum computers considers not only minimising qubit loss but also precisely where those losses occur within the system. Strategically managing atom deposition reduces errors; essentially directing where atoms are lost during computation to improve performance. By focusing on error ‘deposition’ than total loss rates, Xinyi Li and colleagues achieved greater stability and accuracy in simulated tests across various configurations. Xinyi Li at Stevens Institute of Technology and colleagues have created an optimisation technique for neutral-atom quantum computers which considers not just minimising qubit loss but also where within the system those losses occur, this is termed ‘deposition’. Total atom loss alone provides incomplete information because different deposition patterns can induce substantially different error rates even with identical expected loss budgets.

The team formalised this concept as decoder-aware risk deposition and created CAST, a compiler set of tools that optimises placement based on code sensitivity maps. Across numerous tests including surface-code memory simulations, CAST lowered logical error rates relative to standard methods in 35 out of 48 settings. This resulted in improvements up to 5.3x. Decoder-aware atom loss mitigates logical error rates in neutral-atom qubits A fidelity improvement of up to 5.3x in logical error rates resulted from optimising atom-loss deposition within neutral-atom quantum computers. Earlier systems lacked the ability to deliberately manage qubit losses during computation, meaning performance was limited by total loss alone. Researchers at Stevens Institute of Technology developed CAST, a new compiler pass which formalises decoder-aware risk deposition and minimises ‘decoder-weighted harm’ through intelligent direction of unavoidable qubit losses towards less sensitive areas of code. Substantial reductions in errors are now possible across various surface-code memory models, particularly when routing has Slack and exposure is heterogeneous; these conditions previously presented challenges for effective optimisation. Through CAST, a novel compiler pass that formalises decoder-aware risk deposition and reduces ‘decoder-weighted harm’, the team achieved up to a 5·3x fidelity improvement in logical error rates compared to systems focused solely on total atom loss. The evaluation spanned diverse scenarios including lattice surgery simulations, variations in physical architecture, and different surface-code memory models, consistently exceeding topology-blind exposure minimisation in 35 of 48 tested settings. Gains were most pronounced where initial high exposure coincided with low decoder sensitivity, redirecting potential errors from critical code areas; furthermore, analysis revealed routing slack and heterogeneous exposure patterns enhanced optimisation opportunities. Decoder optimisation surpasses simple qubit loss mitigation for strong error correction Increasingly sophisticated compiler strategies are essential as the pursuit of stable qubits continues relentlessly. Simply minimising qubit losses is insufficient for effective quantum error correction, a fact demonstrated by researchers at Stevens Institute of Technology. Their new approach, decoder-aware risk deposition implemented through CAST, highlights that while optimising atom loss location demonstrably improves performance over existing topology-blind methods, ultimate validation requires evaluation via the decoder itself. Careful consideration of how errors are detected and corrected within system architecture is necessary to achieve substantial gains. This work establishes a proactive compiler strategy, namely decoder-aware risk deposition, which demonstrably enhances performance across simulated quantum systems even before decoding begins. By focusing on ‘deposition’, the location of atom loss, rather than minimising total losses, scientists refined how compilers manage qubit failure in neutral-atom arrays; this recognises where atoms are lost significantly impacts logical error rates during computation. CAST, a new compiler pass overlaying code sensitivity with exposure tracking, facilitates strategic redirection of unavoidable errors towards less critical areas of the quantum code structure. The research demonstrated that optimising for the location of qubit loss, rather than simply reducing overall loss, improves quantum error correction within neutral-atom array systems. This is because where an atom is lost affects logical error rates differently depending on its position relative to the decoding process. Researchers developed CAST, a compiler tool which minimises harm by strategically directing potential errors toward lower-impact regions of the quantum code; it achieved improvements in 35 out of 48 tested scenarios using surface-code memory models and physical architecture simulations. The findings suggest careful consideration of decoder sensitivity during compilation can enhance performance before even beginning error correction. 👉 More information🗞 Where Atom Loss Lands Matters: Decoder-Aware Risk Deposition in Neutral-Atom QEC✍️ Xinyi Li, Yifeng Peng and Ying Wang🧠 ArXiv: https://arxiv.org/abs/2608.17913 More like thisQuantum AlgorithmsFire Opal and Black Opal combine for quantum finance learningQuantum AlgorithmsResearchers Achieve 7.44e-9 Fidelity for 200-Qubit StatesQuantum Research NewsGerman scientists cut Toffoli gate count for sparse quantum statesQuantum Research NewsQuantum circuits scale linearly with system size, research confirmsStay currentSee today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals. Tags: