Pith. sign in

REVIEW 5 major objections 4 minor 1 cited by

Correcting a noisy quantum computer using a quantum computer

T0 review · 5 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A trained quantum circuit replaces the classical error decoder

desk verdict A genuinely new decoder idea with a solid small-scale proof of concept, overreaching in its self-correcting claims and unsupported by hardware. read the letter →

arxiv 2506.08331 v1 pith:2UL4SGFO submitted 2025-06-10 quant-ph cond-mat.dis-nncond-mat.stat-mech

Pan Zhang This is my paper
classification quant-phcond-mat.dis-nncond-mat.stat-mech
keywords quantumerrorcorrectiondecodingcircuitvariationalsurfacecodecircuit-levelnoiseminimum-weightperfectmatchingreal-timeself-correctingcomputer
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Decoding in quantum error correction is usually a classical inference step: a computer reads the syndrome and decides which logical correction to apply. This paper proposes doing that step with a small variational quantum circuit instead. Given a noisy circuit $A$, a decoding circuit $B$ is trained on syndrome–logical-sector pairs so that its measurement distribution approximates the conditional probability of the logical sector given the syndrome; once trained, $B$ can be deployed on the same kind of hardware as $A$ and samples a correction at circuit speed. In numerical experiments on the surface-code memory under circuit-level depolarizing noise at distances 3, 5, and 7, the quantum decoder performs on par with minimum-weight perfect matching and crosses the break-even point at distances 5 and 7. If the scheme holds, the classical decoder and its latency disappear from the feedback loop, and a noisy quantum computer could in principle correct itself using its own operations.

What carries the argument

The central object is the variational decoding circuit: a small register of qubits with $B$ blocks of $X$- and $Y$-rotation gates whose angles are $\theta_{q,b}\cdot \gamma_i$ and $\phi_{q,b}\cdot \gamma_i$ for learnable parameters $\theta,\phi$ and syndrome bit $\gamma_i$, interleaved with fixed two-qubit gates, ending in a computational-basis measurement. Because the gate angle is zero when $\gamma_i=0$, trivial syndrome bits do nothing, so the active decoding time is proportional to the number of nontrivial syndrome bits. The circuit is trained by tensor-network simulation of its forward pass and backpropagation through the parameters, making the final measurement distribution approximate the logical-sector posterior, and the same syndrome-controlled rotation mechanism is what enables the proposed measurement-free self-correcting variant.

What would settle it

Simulate or run the decoder with the same depolarizing noise model applied to its own gates as to the data circuit, and check whether the combined logical error rate remains below break-even at distance 7; if it rises above break-even, the noiseless-decoder assumption is load-bearing. A separate check is end-to-end latency: if classical electronics can feed the syndrome back faster than the decoder can produce a corrected qubit, the speed argument loses its premise.

Watch

Extended reading notes

Core claim

The central claim is that decoding can be recast as a quantum sampling task. The decoder is a parameterized circuit whose single-qubit rotations are controlled by the syndrome: each syndrome bit enters the gate angle as a product with a learnable parameter, and a fixed set of two-qubit gates entangles the decoder qubits. Training minimizes the cross-entropy between the circuit's output distribution $q(\beta|\gamma)$ and the true conditional distribution $P(\beta|\gamma)$ over logical sectors, using data pairs from an error model. Once trained, running the circuit yields a sampled logical correction at the speed of the quantum operation being corrected. The paper demonstrates this on the surface-code memory in the Z-basis under depolarizing circuit-level noise for distances 3, 5, and 7, reporting logical error rates comparable to minimum-weight perfect matching and below break-even for the larger distances, and argues the construction extends to general codes with multiple logical qubits and to a 'self-correcting' circuit in which ancilla states, rather than classical measurement outcomes, control the decoding gates.

Load-bearing premise

The numerical demonstrations all assume a noiseless decoding circuit, so the central claim stands or falls on whether the decoder's own error probability stays negligible compared with the circuit it corrects.

Editorial extensions

If this is right

  • Decoding runs at the speed of the corrected circuit instead of the slower clock of classical control electronics, because the decoder is itself a quantum circuit with mostly single-qubit gates.
  • The same trained ansatz transfers across codes and noise models, since the circuit architecture does not depend on the code's topology.
  • For codes with many logical qubits, sampling a correction from the circuit's final state bypasses the sequential autoregressive sampling used by classical neural decoders, which the paper argues gives a quantum sampling advantage.
  • The syndrome-controlled rotation mechanism suggests a measurement-free 'self-correcting circuit' in which ancilla states directly control the decoding gates, eliminating classical electronics from the correction loop.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A direct test of the scheme would add depolarizing noise to the decoding circuit itself; the paper assumes it is noiseless, so the crossover where decoder noise overwhelms the correction is a natural next calculation.
  • The real-time speed claim assumes syndrome-to-parameter feedback within the operation time; a hardware experiment that measures end-to-end latency from syndrome to applied correction would settle whether the quantum decoder actually beats classical electronics.
  • The same syndrome-parameterized gate construction could serve other inference tasks that must act inside a quantum feedback loop, such as feedforward in magic-state distillation or adaptive circuits.
  • Figure 3's self-correcting example still shows ancilla measurements; constructing the measurement-free variant described in the text, e.g., with coherent controlled gates from ancilla to decoder, remains an open implementation step.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 4 minor

Summary. The paper proposes replacing the classical decoder in quantum error correction with a parameterized quantum circuit B. Given a syndrome γ, the single-qubit rotation angles in B are set as functions of γ via learnable parameters; training minimizes cross-entropy between B's output distribution q(β|γ) and labeled (γ, β) pairs generated from the noise model. Numerical simulations using Stim for surface codes of distance 3, 5, and 7 under circuit-level depolarizing noise at p = 0.001 show that a 3-qubit, 10-block ansatz reaches test logical error rates comparable to or better than MWPM. The paper then sketches a 'self-correcting' variant in which ancilla states directly control the decoding gates, claiming to eliminate classical measurements and electronics.

Significance. If the central result holds, the conceptual contribution is interesting: shifting decoding from classical post-processing to a quantum circuit could, in principle, align decoding latency with quantum gate times and enable real-time error correction. The numerical protocol is straightforward, the paper uses standard Stim-generated data, and the authors provide a public code repository. However, the present evidence supports only a limited claim: a small variational circuit, assumed noiseless and trained and tested on the same noise model, can approximate the conditional decoding distribution for the tested surface codes. The stronger claims in the abstract and title—deployment on a noisy device, real-time speed, and self-correction without classical devices or measurements—are not established by the experiments or analysis.

major comments (5)
  1. [Self-correcting quantum computer (Fig. 3)] The claim that decoding can be accomplished 'without classical devices or measurements' is internally inconsistent with the presented scheme. The text states that 'the ancilla qubits are measured' and that these ancilla qubits 'are directly utilized to govern the learned single-qubit gates,' yet it also states that 'it does not involve any classical operations' and that 'measurement outcomes are not utilized as classical inputs to the decoder.' A measured ancilla qubit produces a classical bit, and using that bit to set gate parameters requires classical feedforward. If the intended mechanism is coherent control by the unmeasured ancilla state, the circuit diagram must show controlled gates driven by the ancilla, not M (measurement) blocks, and the text must describe the coherent feedback path. As drawn, the scheme either resorts to classical measurement and feedforward, contradicting the abstract, or is incomplete.
  2. [Numerical experiments] The numerical experiments explicitly assume a perfectly noiseless decoding circuit ('we consider the decoding circuit to be noiseless'), yet the abstract and discussions claim deployment on a noisy quantum device. No analysis is given of how noise in the decoding circuit changes the logical error rate, and the assertion in Discussions that 'a noisy decoding circuit could also accomplish the decoding task, as the decoding process itself is probabilistic' is unsupported by any simulation or fault-tolerance argument. Because the decoding circuit would be part of the same noisy hardware, this assumption is load-bearing for the practical relevance of the method.
  3. [Decoding speed and real-time claims] The claim that 'the decoding speed matches the speed of the quantum circuits being corrected' is not justified by the arguments presented. The paper estimates gate operation times (e.g., 30 ns single-qubit gates) but does not model the latency of syndrome measurement, classical feedforward of measurement outcomes, control-electronics overhead, or the time needed to set gate parameters in response to a syndrome. In the non-self-correcting scheme, the syndrome must be known before the decoding gate parameters can be applied, and this measurement-to-parameter latency may dominate the total decoding time. A concrete latency model or end-to-end timing comparison is needed to support the real-time claim.
  4. [Training section (label generation)] The training section specifies that data consist of pairs of syndrome γ and logical sector β 'derived from an error model or gathered experimentally,' but it does not state how β is assigned for each syndrome. If β is obtained by sampling a single error from the noise model and taking its logical sector, then the cross-entropy loss is an unbiased estimator of the maximum-likelihood posterior only under specific sampling assumptions; if β instead comes from a specific decoder (e.g., MWPM or a lookup table), the comparison to MWPM is no longer an independent test. The paper claims to approximate maximum-likelihood decoding, so the exact label-generation procedure must be stated.
  5. [Fig. 2 and numerical comparison] The y-axis of Fig. 2 is labeled 'Logical Error Rate,' but the break-even line is placed at 0.999, while the text states 'break-even point with a 0.001 logical error rate.' Since the physical error rate is p = 0.001, the break-even logical error rate should be 0.001, not 0.999; the plotted quantity therefore appears to be the logical success rate (1 − logical error rate). This discrepancy makes the statements 'better than MWPM' and 'exceeds the break-even value' ambiguous. The axis label and the text should be reconciled so that the quantitative claims are unambiguous.
minor comments (4)
  1. [Equations (1) and (2)] The notation for the rotation parameters is unclear: the text says 'the parameters of the rotation gates are θ_i x_i' and later writes θ_{q,b}·γ_i, but the stated parameter count 2QBm implies a separate parameter for each gate and each syndrome bit. Please define θ_{q,b,i} and ϕ_{q,b,i} explicitly and use them consistently in the matrix definitions.
  2. [Numerical experiments (Fig. 2)] The training curves are shown without error bars or standard deviations, and the number of random seeds is not reported. Adding this information would strengthen the claim that the observed performance is reproducible.
  3. [Discussions (Quantum sampling advantage)] The paragraph on quantum sampling advantage is speculative and does not include a benchmark or resource estimate; it should be clearly labeled as a conjecture rather than a demonstrated advantage.
  4. [Abstract and title] The title and abstract claim that decoding can be accomplished 'without classical devices or measurements,' but the main numerical scheme explicitly uses syndrome measurements and classical parameter setting. The self-correcting scheme is presented only as an illustration with no numerical support. The wording should be tempered to match the evidence.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the decoding circuit is trained on one syndrome sample and evaluated on a held-out sample against an external MWPM benchmark; self-citations are not load-bearing.

full rationale

The central numerical claim is that a learned quantum decoding circuit performs on par with MWPM on surface codes under circuit-level noise. The paper trains the circuit on 200,000 syndromes and evaluates on 100,000 test syndromes generated with a different random seed from the same Stim error model. This is a standard supervised learning protocol with held-out generalization testing, not a fitted parameter being renamed as a prediction or a quantity that is forced by construction. The trained parameters are optimized to minimize cross-entropy on training pairs and then tested on unseen syndromes; the comparison to MWPM is an external benchmark. The self-citations [43, 44, 45, 46] are used to justify that neural-network decoders and tensor-network simulations are available tools, but the core decoding ansatz and its evaluation do not reduce to these citations. There is no uniqueness theorem imported from prior work, and no ansatz is smuggled in via citation. The manuscript's inconsistency concerning Fig. 3—where ancilla qubits are shown as measured despite the claim that the scheme works without classical measurements—is a correctness or support gap, not a circularity of the derivation. Similarly, the noiseless-decoding-circuit assumption is an idealization that limits the strength of the practical claim but does not make the argument circular. Overall, the derivation is self-contained as a proposal and numerical demonstration, and no circular step meeting the evidentiary bar is present.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on noiselessness of the decoder, a known noise model, sufficient expressivity of the small ansatz, and (for the self-correcting extension) coherent feedback without measurement. The first is explicitly acknowledged, the fourth is contradicted by the paper's own figure.

free parameters (3)
  • Decoder weights theta, phi = not reported (optimized)
    The 2QBm rotation angles are fitted to minimize cross-entropy on 200,000 training syndrome-correction pairs; the reported logical error rates depend on these fitted values.
  • Ansatz size: Q (decoding qubits), B (blocks) = Q=3, B=10
    Chosen by hand; no ablation or scaling study is provided to justify sufficiency for larger codes.
  • Training hyperparameters = 10,000 epochs; learning rate not stated; 200,000 training samples
    Selected by hand; the final test performance in Fig. 2 is a function of the epoch count.
assumptions (4)
  • domain assumption The decoding circuit is noiseless and fault-tolerant in all numerical experiments.
    Stated in 'Numerical experiments': 'we consider the decoding circuit to be noiseless.' The paper argues a real decoding circuit would be 'much less noisy' than the corrected circuit, but provides no noise model.
  • domain assumption The error model is known, and Stim-generated data faithfully represents it.
    Both training and test sets are generated with Stim from a depolarizing circuit-level noise model; the decoder cannot be better than the model it was trained on.
  • ad hoc to paper The 3-qubit, 10-block variational circuit can approximate the decoding map P(beta|gamma).
    No expressivity analysis is given; the paper assumes the ansatz is sufficient based on numerical success at d<=7 and p=0.001.
  • ad hoc to paper Ancilla qubits can directly control decoding gates without measurement, enabling the self-correcting scheme.
    Invoked in the 'Self-correcting quantum computer' section; the accompanying Fig. 3 shows measurements, so this assumption is unbacked and internally contradicted.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Correcting a noisy quantum computer using a quantum computer." pith.science (2026). https://pith.science/paper/2UL4SGFO

@misc{pith2026250608331,
  author       = {Pith},
  title        = {Pith review of: Correcting a noisy quantum computer using a quantum computer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2UL4SGFO}},
  note         = {Machine review of arXiv:2506.08331}
}
abstract

Quantum computers require error correction to achieve universal quantum computing. However, current decoding of quantum error-correcting codes relies on classical computation, which is slower than quantum operations in superconducting qubits. This discrepancy makes the practical implementation of real-time quantum error correction challenging. In this work, we propose a decoding scheme that leverages the operations of the quantum circuit itself. Given a noisy quantum circuit $A$, we train a decoding quantum circuit $B$ using syndrome measurements to identify the logical operators needed to correct errors in circuit $A$. The trained quantum circuit $B$ can be deployed on quantum devices, such as superconducting qubits, to perform real-time decoding and error correction. Our approach is applicable to general quantum codes with multiple logical qubits and operates efficiently under various noise conditions, and the decoding speed matches the speed of the quantum circuits being corrected. We have conducted numerical experiments using surface codes up to distance 7 under circuit-level noise, demonstrating performance on par with the classical minimum-weight perfect matching algorithm. Interestingly, our method reveals that the traditionally classical task of decoding error-correcting codes can be accomplished without classical devices or measurements. This insight paves the way for the development of self-correcting quantum computers.

Figures

Figures reproduced from arXiv: 2506.08331 by the authors.

Figure 1
Figure 1. FIG. 1 [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Numerical results for our quantum circuit decoder on the surface code quantum memory in the [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Illustration of the self-correcting repetition-code quantum memory with distance 3. The blue shaded [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Superior resilience to poisoning and amenability to unlearning in quantum machine learning

    quant-ph 2025-08 conditional novelty 5.0 of 10

    A simulator study reports that QNNs hold accuracy under label flipping better than a large MLP and unlearn faster, but the claimed fundamental advantage is not established without regularized classical baselines.

Reference graph

Works this paper leans on

49 extracted references · 23 canonical work pages · cited by 1 Pith paper

  1. [1]

    P. W. Shor, Polynomial-time algorithms for prime factorization and discrete logarithms on a quantum computer, SIAM review41, 303 (1999)

    1999

  2. [2]

    Gidney and M

    C. Gidney and M. Ekerå, How to factor 2048 bit rsa integers in 8 hours using 20 million noisy qubits, Quantum5, 433 (2021)

    2021

  3. [3]

    Preskill, Quantum computing in the nisq era and beyond, Quantum2, 79 (2018)

    J. Preskill, Quantum computing in the nisq era and beyond, Quantum2, 79 (2018)

    2018

  4. [4]

    A. G. Fowler, M. Mariantoni, J. M. Martinis, and A. N. Cleland, Surface codes: Towards practical large-scale quantum computation, Physical Review A86, 032324 (2012)

    2012

  5. [5]

    Horsman, A

    D. Horsman, A. G. Fowler, S. Devitt, and R. Van Meter, Surface code quantum computing by lattice surgery, New Journal of Physics14, 123011 (2012)

    2012

  6. [6]

    E. T. Campbell, A theory of single-shot error correction for adversarial noise, Quantum Science and Technology4, 025006 (2019)

    work page 2019

  7. [7]

    J. P. Bonilla Ataides, D. K. Tuckett, S. D. Bartlett, S. T. Flammia, and B. J. Brown, The XZZX surface code, Nature communications12, 2172 (2021)

    work page 2021

  8. [8]

    Bravyi, A

    S. Bravyi, A. W. Cross, J. M. Gambetta, D. Maslov, P. Rall, and T. J. Yoder, High-threshold and low-overhead fault-tolerant quantum memory, Nature627, 778 (2024)

    2024

Show all 49 references
  1. [9]

    M. B. Hastings and J. Haah, Dynamically generated logical qubits, Quantum5, 564 (2021)

    2021

  2. [10]

    Q. Xu, J. P. Bonilla Ataides, C. A. Pattison, N. Raveendran, D. Bluvstein, J. Wurtz, B. Vasi ´c, M. D. Lukin, L. Jiang, and H. Zhou, Constant-overhead fault-tolerant quantum computation with reconfig- urable atom arrays, Nature Physics20, 1084–1090 (2024)

    2024

  3. [11]

    A. J. Landahl, J. T. Anderson, and P. R. Rice, Fault-tolerant quantum computing with color codes, arXiv preprint arXiv:1108.5738 (2011)

    2011 arXiv

  4. [12]

    A. G. Fowler, 2-d color code quantum computation, Physical Review A83, 042310 (2011)

    2011

  5. [13]

    Bomb ´ın, Topological subsystem codes, Physical Review A81, 032301 (2010)

    H. Bomb ´ın, Topological subsystem codes, Physical Review A81, 032301 (2010)

    2010

  6. [14]

    Duclos-Cianci and D

    G. Duclos-Cianci and D. Poulin, A renormalization group decoding algorithm for topological quantum codes, in2010 IEEE Information Theory Workshop(2010) pp. 1–5

    2010

  7. [15]

    Bravyi, M

    S. Bravyi, M. Suchara, and A. Vargo, Efficient algorithms for maximum likelihood decoding in the surface code, Physical Review A90, 032326 (2014)

    2014

  8. [16]

    Bombin, R

    H. Bombin, R. S. Andrist, M. Ohzeki, H. G. Katzgraber, and M. A. Martin-Delgado, Strong resilience 10 of topological codes to depolarization, Physical Review X2, 10.1103/PhysRevX.2.021004 (2012)

    2012 doi

  9. [17]

    Bombin and M

    H. Bombin and M. A. Martin-Delgado, Optimal resources for topological two-dimensional stabilizer codes: Comparative study, Physical Review A76, 10.1103/PhysRevA.76.012305 (2007)

    2007 doi

  10. [18]

    Bacon, S

    D. Bacon, S. T. Flammia, A. W. Harrow, and J. Shi, Sparse quantum codes from quantum circuits, IEEE Transactions on Information Theory63, 2464 (2017), arXiv:1411.3334 [quant-ph]

    2017 arXiv

  11. [19]

    Honciuc Menendez, A

    D. Honciuc Menendez, A. Ray, and M. Vasmer, Implementing fault-tolerant non-clifford gates using the [[8, 3, 2]] color code, Physical Review A109, 062438 (2024)

    2024

  12. [20]

    B. W. Reichardt, D. Aasen, R. Chao, A. Chernoguzov, W. van Dam, J. P. Gaebler, D. Gresh, D. Luc- chetti, M. Mills, S. A. Moses,et al., Demonstration of quantum computation and error correction with a tesseract code, arXiv preprint arXiv:2409.04628 (2024)

    2024 arXiv

  13. [21]

    Y . Hong, E. Durso-Sabina, D. Hayes, and A. Lucas, Entangling four logical qubits beyond break-even in a nonlocal code, Physical Review Letters133, 180601 (2024)

    2024

  14. [22]

    Da Silva, C

    M. Da Silva, C. Ryan-Anderson, J. Bello-Rivas, A. Chernoguzov, J. Dreiling, C. Foltz, J. Gaebler, T. Gatterman, D. Hayes, N. Hewitt,et al., Demonstration of logical qubits and repeated error correc- tion with better-than-physical error rates, arXiv preprint arXiv:2404.02280 (2024)

    2024 arXiv

  15. [23]

    Mayer, C

    K. Mayer, C. Ryan-Anderson, N. Brown, E. Durso-Sabina, C. H. Baldwin, D. Hayes, J. M. Dreil- ing, C. Foltz, J. P. Gaebler, T. M. Gatterman,et al., Benchmarking logical three-qubit quantum fourier transform encoded in the steane code on a trapped-ion quantum computer, arXiv pre...

    2024 arXiv

  16. [24]

    Ryan-Anderson, J

    C. Ryan-Anderson, J. G. Bohnet, K. Lee, D. Gresh, A. Hankin, J. Gaebler, D. Francois, A. Chernogu- zov, D. Lucchetti, N. C. Brown,et al., Realization of real-time fault-tolerant quantum error correction, Physical Review X11, 041058 (2021)

    2021

  17. [25]

    Y . Zhao, Y . Ye, H.-L. Huang, Y . Zhang, D. Wu, H. Guan, Q. Zhu, Z. Wei, T. He, S. Cao,et al., Realization of an error-correcting surface code with superconducting qubits, Physical Review Letters 129, 030501 (2022)

    2022

  18. [26]

    Bluvstein, S

    D. Bluvstein, S. J. Evered, A. A. Geim, S. H. Li, H. Zhou, T. Manovitz, S. Ebadi, M. Cain, M. Kali- nowski, D. Hangleiter,et al., Logical quantum processor based on reconfigurable atom arrays, Nature 626, 58 (2024)

    2024

  19. [27]

    Google Quantum AI, Suppressing quantum errors by scaling a surface code logical qubit, Nature614, 676 (2023)

    2023

  20. [28]

    Google Quantum AI and collaborators, Quantum error correction below the surface code threshold, 11 Nature638, 920 (2025), arXiv:2408.13687 [quant-ph]

    2025 arXiv

  21. [29]

    Google Quantum AI, Exponential suppression of bit or phase errors with cyclic error correction, Na- ture595, 383 (2021)

    2021

  22. [30]

    Dennis, A

    E. Dennis, A. Kitaev, A. Landahl, and J. Preskill, Topological quantum memory, Journal of Mathe- matical Physics43, 4452 (2002)

    2002

  23. [31]

    Higgott, Pymatching: A python package for decoding quantum codes with minimum-weight per- fect matching, ACM Transactions on Quantum Computing3, 1 (2022)

    O. Higgott, Pymatching: A python package for decoding quantum codes with minimum-weight per- fect matching, ACM Transactions on Quantum Computing3, 1 (2022)

    2022

  24. [32]

    Higgott and C

    O. Higgott and C. Gidney, Sparse blossom: correcting a million errors per core second with minimum- weight matching, arXiv preprint arXiv:2303.15933 (2023)

    2023 arXiv

  25. [33]

    Wu and L

    Y . Wu and L. Zhong, Fusion blossom: Fast mwpm decoders for qec, in2023 IEEE International Conference on Quantum Computing and Engineering (QCE), V ol. 1 (IEEE, 2023) pp. 928–938

    2023

  26. [34]

    A. G. Fowler, Optimal complexity correction of correlated errors in the surface code, arXiv preprint arXiv:1310.0863 (2013)

    2013 arXiv

  27. [35]

    Higgott, T

    O. Higgott, T. C. Bohdanowicz, A. Kubica, S. T. Flammia, and E. T. Campbell, Improved decoding of circuit noise and fragile boundaries of tailored surface codes, Physical Review X13, 031007 (2023)

    2023

  28. [36]

    Roffe, D

    J. Roffe, D. R. White, S. Burton, and E. Campbell, Decoding across the quantum low-density parity- check code landscape, Physical Review Research2, 10.1103/PhysRevResearch.2.043423 (2020)

    2020 doi

  29. [37]

    Panteleev and G

    P. Panteleev and G. Kalachev, Degenerate Quantum LDPC Codes With Good Finite Length Perfor- mance, Quantum5, 585 (2021)

    2021

  30. [38]

    H. Cao, S. Zhao, D. Feng, Z. Shen, H. Yan, T. Su, W. Sun, H. Xu, F. Pan, H. Yu,et al., Exact decoding of quantum error-correcting codes, Physical Review Letters134, 190603 (2025)

    2025

  31. [39]

    Fossorier and S

    M. Fossorier and S. Lin, Soft-decision decoding of linear block codes based on ordered statistics, IEEE Transactions on Information Theory41, 1379 (1995)

    1995

  32. [40]

    Fossorier, Iterative reliability-based decoding of low-density parity check codes, IEEE Journal on Selected Areas in Communications19, 908 (2001)

    M. Fossorier, Iterative reliability-based decoding of low-density parity check codes, IEEE Journal on Selected Areas in Communications19, 908 (2001)

    2001

  33. [41]

    C. T. Chubb, General tensor network decoding of 2d pauli codes, arXiv preprint arXiv:2101.04125 (2021)

    2021 arXiv

  34. [42]

    Bausch, A

    J. Bausch, A. W. Senior, F. J. Heras, T. Edlich, A. Davies, M. Newman, C. Jones, K. Satzinger, M. Y . Niu, S. Blackwell,et al., Learning high-accuracy error decoding for quantum processors, Nature635, 834 (2024)

    2024

  35. [43]

    H. Cao, F. Pan, D. Feng, Y . Wang, and P. Zhang, Generative decoding for quantum error-correcting 12 codes, arXiv preprint arXiv:2503.21374 (2025)

    2025 arXiv

  36. [44]

    H. Cao, F. Pan, Y . Wang, and P. Zhang, qecgpt: decoding quantum error-correcting codes with gener- ative pre-trained transformers, arXiv preprint arXiv:2307.09025 (2023)

    2023 arXiv

  37. [45]

    Pan and P

    F. Pan and P. Zhang, Simulation of quantum circuits using the big-batch tensor network method, Physical Review Letters128, 030501 (2022)

    2022

  38. [46]

    F. Pan, K. Chen, and P. Zhang, Solving the Sampling Problem of the Sycamore Quantum Circuits, Physical Review Letters129, 090502 (2022), publisher: APS

    2022

  39. [47]

    Gidney, Stim: A fast stabilizer circuit simulator, Quantum5, 497 (2021), 2103.02202 [quant-ph]

    C. Gidney, Stim: A fast stabilizer circuit simulator, Quantum5, 497 (2021), 2103.02202 [quant-ph]

    2021 arXiv

  40. [48]

    Arute, K

    F. Arute, K. Arya, R. Babbush, D. Bacon, J. C. Bardin, R. Barends, R. Biswas, S. Boixo, F. G. Brandao, D. A. Buell,et al., Quantum supremacy using a programmable superconducting processor, Nature574, 505 (2019)

    2019

  41. [49]

    D. Gao, D. Fan, C. Zha, J. Bei, G. Cai, J. Cai, S. Cao, F. Chen, J. Chen, K. Chen,et al., Establishing a new benchmark in quantum computational advantage with 105-qubit zuchongzhi 3.0 processor, Physical Review Letters134, 090601 (2025). 13

    2025

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.