arxiv: 2605.12046 · v1 · submitted 2026-05-12 · 🪐 quant-ph · cs.AI· cs.LG

Recognition: 2 theorem links

· Lean Theorem

Rethink the Role of Neural Decoders in Quantum Error Correction

Ge Yan , Shanchuan Li , Yuxuan Du

Authors on Pith no claims yet

Pith reviewed 2026-05-13 04:50 UTC · model grok-4.3

classification 🪐 quant-ph cs.AIcs.LG

keywords quantum error correctionneural decoderssurface codesFPGA deploymentINT4 quantizationinductive biasdata scaling

0 comments

The pith

Neural decoder performance in quantum error correction depends more on training data volume than on model architecture, with inductive bias and 4-bit quantization enabling microsecond FPGA latency.

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

The paper unifies representative neural decoders for surface-code decoding into five architectural paradigms and tests them with an end-to-end compression pipeline on FPGA hardware for codes up to distance 9. Experiments show that expanding the training dataset improves decoding accuracy more than increasing architectural complexity. Models with suitable inductive bias achieve higher accuracy, while INT4 quantization proves necessary to reach the microsecond-scale speeds required for practical real-time operation. These results address the accuracy-latency tradeoff that has limited deployment of neural decoders in near-term quantum error correction.

Core claim

Near-term decoding performance is driven more by data scale than architectural complexity; appropriate inductive bias is essential for achieving high decoding accuracy; and INT4 quantization is a prerequisite for meeting microsecond-scale latency requirements on FPGAs. The authors unify five paradigms and apply compression to evaluate deployability on hardware for codes up to d=9.

What carries the argument

The five unified architectural paradigms for neural decoders together with an end-to-end compression pipeline that enables FPGA evaluation under explicit accuracy-latency constraints.

If this is right

Increasing training data volume can raise decoding accuracy without requiring more elaborate network designs.
Embedding appropriate inductive biases into decoder models is required to reach high accuracy levels.
INT4 quantization becomes mandatory to satisfy the microsecond latency target on FPGA platforms.
The unified paradigms and pipeline supply concrete design rules for building scalable real-time neural QEC decoders.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Emphasis on data generation rather than architecture invention could become the dominant research direction for near-term neural decoders.
Validation on actual quantum hardware with device-specific noise will be needed before claiming readiness for production codes.
The same data-scale priority may extend to other machine-learning tasks in quantum control and calibration.

Load-bearing premise

The five paradigms and compression pipeline represent the broader space of neural decoders, and results on codes of distance at most 9 will generalize to larger distances and real-device noise without extra hardware effects.

What would settle it

A complex architecture trained on limited data outperforming a simple architecture trained on large data in decoding accuracy for d=9 surface codes, or INT4 quantization failing to meet microsecond latency on FPGA hardware.

Figures

Figures reproduced from arXiv: 2605.12046 by Ge Yan, Shanchuan Li, Yuxuan Du.

**Figure 1.** Figure 1: Schematic of a d = 3 rotated surface code. a) Data qubits (yellow) are located at the vertices, interleaving with two types of detector qubits (stabilizers) located on the faces. Red squares represent x-stabilizers (detecting z errors), and blue squares represent z-stabilizers (detecting x errors). b) As time goes on, the errors accumulate on the detectors. essential for meeting real-time latency constrain… view at source ↗

**Figure 2.** Figure 2: Schematic illustration of the entire pipeline. Left: Syndrome data is processed via a causal sliding window (r = d) to maintain fault-tolerant boundaries. Middle: A diverse spectrum of neural decoders is evaluated, ranging from MLPs and 3D-CNNs to temporal models, Transformers, and topology-aware GNNs. Right: To resolve the latency bottleneck, we employ an aggressive INT4 compression pipeline (combining Pr… view at source ↗

**Figure 3.** Figure 3: Neural decoder benchmark on the rotated surface code. (a) LER vs. training set size with model size “large”. The dashed gray line marks the MWPM baseline; the shaded region indicates LER below MWPM. Error bars show the min–max range over three independent runs. Inset (d=9): 3D-CNN outliers at 5 × 105 and 106 samples (15.6% and 6.5%) are shown separately to preserve the main axis scale. (b) Model comparison… view at source ↗

**Figure 4.** Figure 4: Impact of quantization and pruning on decoder fidelity. (a) Quantization sensitivity under PTQ and QAT. While 3D-CNN retains robustness at W8A8, TCN and Transformer exhibit severe degradation even at 8-bit precision, and all models fail at W4A4. However, QAT is able to recover the accuracy loss at INT4. (b) Pruning robustness based on the recovered QAT-W4A4 models. The curves track LER as unstructured spar… view at source ↗

**Figure 5.** Figure 5: Syndrome extraction circuits. Standard circuit-level implementation of weight-4 stabilizers. (a) The Z-stabilizer measures parity Z1Z2Z3Z4. (b) The X-stabilizer measures parity X1X2X3X4, utilizing Hadamard gates on the ancilla to facilitate measurement in the X-basis. D denotes data qubits, A denotes stabilizers, and R denotes the unexpected errors occur on data qubits. The circuit width is scaled to fit t… view at source ↗

**Figure 6.** Figure 6: Data scale triggers a memorization-to-generalization transition. Training dynamics for 3D-CNN and TCN at d=7 under two data regimes: 105 samples (left, data-scarce) vs 5×106 samples (right, data-abundant). With limited data, severe train-validation gaps emerge (0.56 for 3D-CNN, 0.17 for TCN), indicating that models memorize training examples rather than learning generalizable error patterns. Scaling to 5×1… view at source ↗

**Figure 7.** Figure 7: BCE loss is a reliable surrogate for logical error rate. Correlation between binary cross-entropy (BCE) loss and logical error rate (LER) during training for 3D-CNN and Transformer at d=7 with 5×106 samples. Both training loss (left) and validation loss (right) exhibit strong log-linear relationships with LER, confirming that minimizing the BCE objective effectively drives down the true decoding performanc… view at source ↗

**Figure 8.** Figure 8: Inductive bias determines architectural scalability to large code distances. Performance comparison of four neural decoder architectures across code distances d ∈ {3, 5, 7, 9}, evaluated on 5×106 training samples. Structure-aware architectures (3D-CNN, TCN, Transformer) consistently outperform the MWPM baseline across all distances, with error rates tracking the baseline trend and maintaining a stable perf… view at source ↗

**Figure 9.** Figure 9: Data scale cannot compensate for the absence of inductive bias. MLP performance as a function of training set size at large code distances (d=7 and d=9). At d=7 (left), MLP improves from 32% to 3.2% LER as data scales to 5×106 samples, yet remains 3.4× above the MWPM baseline (0.93%). At d=9 (right), performance plateaus near 40% until the largest dataset, then drops to 16.9%—still 26× worse than MWPM (0.6… view at source ↗

**Figure 10.** Figure 10: Training efficiency varies significantly across architectures. Training dynamics for 3D-CNN, TCN, and Transformer at d=9 with 2.5×107 samples. Left: Logical error rate (LER) vs wall-clock training time. Right: Validation loss vs training time. All three architectures achieve similar final performance (LER ≈0.47–0.49%, below MWPM baseline of 0.64%), but training time differs dramatically: 3D-CNN reaches th… view at source ↗

read the original abstract

Quantum error correction (QEC) is essential for enabling quantum advantages, with decoding as a central algorithmic primitive. Owing to its importance and intrinsic difficulty, substantial effort has been made to QEC decoder design, among which neural decoders have recently emerged as a promising data-driven paradigm. Despite this progress, practical deployment remains hindered by a fundamental accuracy-latency tradeoff, often on the microsecond timescale. To address this challenge, here we revisit neural decoders for surface-code decoding under explicit accuracy-latency constraints, considering code distances up to d=9 (161 physical qubits). We unify and redesign representative neural decoders into five architectural paradigms and develop an end-to-end compression pipeline to evaluate their deployability and performance on FPGA hardware. Through systematic experiments, we reveal several previously underexplored insights: (i) near-term decoding performance is driven more by data scale than architectural complexity; (ii) appropriate inductive bias is essential for achieving high decoding accuracy; and (iii) INT4 quantization is a prerequisite for meeting microsecond-scale latency requirements on FPGAs. Together, these findings provide concrete guidance toward scalable and real-time neural QEC decoding.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper's main finding is that training data volume and inductive bias matter more than architectural bells and whistles for neural surface-code decoders, with INT4 quantization required to hit microsecond FPGA latency.

read the letter

The central takeaway is that near-term neural decoding performance on surface codes comes mostly from scaling up the training data rather than from more complex network designs, and that INT4 quantization is necessary to meet microsecond latency targets on FPGAs. The authors unify five decoder families, run controlled experiments varying data size and bias on codes up to distance 9, and close the loop with an end-to-end FPGA compression pipeline. That hardware evaluation is the part that stands out; most earlier neural-decoder papers stop at simulation accuracy and never measure deployable latency under realistic constraints. The systematic variation of training-set size and the direct comparison of inductive-bias choices give the three headline claims a reasonably solid empirical footing. The work is therefore useful for anyone who actually needs to put a decoder on hardware in the next few years. The main limitations are the usual ones for this style of study: results stop at d=9 and use simulated noise, so generalization to larger distances or real-device noise remains untested. The unification covers only the five chosen paradigms, and it is possible that an unexamined architecture could shift the relative importance of data versus bias. The abstract is thin on exact baseline definitions and statistical procedures, so the full paper must demonstrate that the controls are tight enough to support the ordering of effects. Readers working on practical QEC implementations will find the guidance on data scale and quantization directly actionable. The paper is coherent on its own terms and addresses a concrete deployment bottleneck, so it is worth sending out for peer review even if the claims need tightening on the experimental details.

Referee Report

0 major / 1 minor

Summary. The paper unifies five neural decoder architectures for surface-code QEC (d≤9), introduces an end-to-end compression pipeline, and reports FPGA evaluations showing that training-data scale dominates architectural complexity for near-term accuracy, that inductive bias is required for high performance, and that INT4 quantization is necessary to reach microsecond latency.

Significance. If the controlled comparisons hold, the work supplies actionable guidance for hardware-constrained QEC deployment by prioritizing data volume and quantization over model sophistication. The unified paradigms and measured FPGA latencies constitute a useful benchmark for the community.

minor comments (1)

[Abstract] Abstract: the description of the systematic experiments omits the number of independent trials, the precise baseline decoders and training-set sizes used for the data-scale vs. architecture comparison, and any statistical tests supporting the ordering of effects.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive assessment of our work and the recommendation for minor revision. We are pleased that the controlled comparisons, unified architectural paradigms, and FPGA latency measurements are viewed as providing actionable guidance for hardware-constrained QEC deployment.

Circularity Check

0 steps flagged

No significant circularity in empirical evaluation

full rationale

This is an empirical hardware-evaluation study that unifies five decoder architectures, varies training-set sizes, and measures FPGA latency after INT4 quantization on surface codes up to d=9. No derivation chain, first-principles prediction, or load-bearing claim reduces to a fitted parameter or self-citation by construction; all headline insights rest on direct experimental comparisons and hardware benchmarks that remain independent of the paper's own outputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

This is an empirical study of neural-network performance on quantum decoding tasks. No mathematical axioms, free parameters, or invented physical entities are introduced or required by the central claims in the abstract.

pith-pipeline@v0.9.0 · 5505 in / 1096 out tokens · 56359 ms · 2026-05-13T04:50:17.634045+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean (J-uniqueness, Aczél classification) washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

We unify and redesign representative neural decoders into five architectural paradigms and develop an end-to-end compression pipeline... near-term decoding performance is driven more by data scale than architectural complexity; appropriate inductive bias is essential...
IndisputableMonolith/Foundation/AlexanderDuality.lean alexander_duality_circle_linking unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Rotated surface code... d=9 (161 physical qubits)... spatiotemporal syndrome volume

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

65 extracted references · 65 canonical work pages

[1]

Physical Review Letters , volume=

Establishing a new benchmark in quantum computational advantage with 105-qubit zuchongzhi 3.0 processor , author=. Physical Review Letters , volume=. 2025 , publisher=

work page 2025
[2]

Multi-Scale Context Aggregation by Dilated Convolutions

Multi-scale context aggregation by dilated convolutions , author=. arXiv preprint arXiv:1511.07122 , year=

work page Pith review arXiv
[3]

Quantum , volume=

Machine-learning-assisted correction of correlated qubit errors in a topological code , author=. Quantum , volume=. 2018 , publisher=

work page 2018
[4]

Quantum , volume=

A fault-tolerant honeycomb memory , author=. Quantum , volume=. 2021 , publisher=

work page 2021
[5]

Quantum Science and Technology , volume=

Decoding small surface codes with feedforward neural networks , author=. Quantum Science and Technology , volume=. 2017 , publisher=

work page 2017
[6]

Physical Review Research , volume=

Neural network decoder for near-term surface-code experiments , author=. Physical Review Research , volume=. 2025 , publisher=

work page 2025
[7]

arXiv preprint arXiv:2404.11197 , year=

Benchmarking Machine Learning Models for Quantum Error Correction , author=. arXiv preprint arXiv:2404.11197 , year=

work page arXiv
[8]

Nature (Simulated) , year=

AlphaQubit 2: Scaling Transformers for Fault-Tolerant Quantum Computing , author=. Nature (Simulated) , year=

work page
[9]

Quantum , volume=

Almost-linear time decoding algorithm for topological codes , author=. Quantum , volume=

work page
[10]

MICRO , year=

Iceberg: An FPGA-based microarchitecture for real-time quantum error correction , author=. MICRO , year=

work page
[11]

IEEE Transactions on Quantum Engineering , year=

Helios: High-throughput union-find decoder on FPGA , author=. IEEE Transactions on Quantum Engineering , year=

work page
[12]

Available: https://arxiv.org/abs/2103.13630

A survey of quantization methods for efficient neural network inference , author=. arXiv preprint arXiv:2103.13630 , year=

work page arXiv
[13]

arXiv preprint arXiv:2412.20380 , year=

Artificial intelligence for quantum error correction: A comprehensive review , author=. arXiv preprint arXiv:2412.20380 , year=

work page arXiv
[14]

Quantum , volume=

Decoding algorithms for surface codes , author=. Quantum , volume=. 2024 , publisher=

work page 2024
[15]

arXiv preprint arXiv:2511.01062 , year=

ECCentric: An Empirical Analysis of Quantum Error Correction Codes , author=. arXiv preprint arXiv:2511.01062 , year=

work page arXiv
[16]

Physical review letters , volume=

Neural belief-propagation decoders for quantum error-correcting codes , author=. Physical review letters , volume=. 2019 , publisher=

work page 2019
[17]

Science Advances , volume=

High-performance fault-tolerant quantum computing with many-hypercube codes , author=. Science Advances , volume=. 2024 , publisher=

work page 2024
[18]

Reviews of Modern Physics , volume=

Quantum error correction for quantum memories , author=. Reviews of Modern Physics , volume=. 2015 , publisher=

work page 2015
[19]

Physical Review A—Atomic, Molecular, and Optical Physics , volume=

Surface codes: Towards practical large-scale quantum computation , author=. Physical Review A—Atomic, Molecular, and Optical Physics , volume=. 2012 , publisher=

work page 2012
[20]

Physical Review A—Atomic, Molecular, and Optical Physics , volume=

Surface code quantum computing with error rates over 1\ author=. Physical Review A—Atomic, Molecular, and Optical Physics , volume=. 2011 , publisher=

work page 2011
[21]

2025 , publisher=

Quantum error correction below the surface code threshold , journal=. 2025 , publisher=

work page 2025
[22]

Proceedings of the 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems , pages=

Lilliput: a lightweight low-latency lookup-table decoder for near-term quantum error correction , author=. Proceedings of the 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems , pages=

work page
[23]

Journal of Mathematical Physics , volume=

Topological quantum memory , author=. Journal of Mathematical Physics , volume=. 2002 , publisher=

work page 2002
[24]

IEEE Transactions on Information Theory , volume=

Quantum LDPC codes with almost linear minimum distance , author=. IEEE Transactions on Information Theory , volume=. 2021 , publisher=

work page 2021
[25]

Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

Quantization and training of neural networks for efficient integer-arithmetic-only inference , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

work page
[26]

International conference on machine learning , pages=

Up or down? adaptive rounding for post-training quantization , author=. International conference on machine learning , pages=. 2020 , organization=

work page 2020
[27]

Proceedings of the 54th annual ACM SIGACT symposium on theory of computing , pages=

Asymptotically good quantum and locally testable classical LDPC codes , author=. Proceedings of the 54th annual ACM SIGACT symposium on theory of computing , pages=

work page
[28]

Nature , volume=

High-threshold and low-overhead fault-tolerant quantum memory , author=. Nature , volume=. 2024 , publisher=

work page 2024
[29]

1997 , publisher=

Stabilizer codes and quantum error correction , author=. 1997 , publisher=

work page 1997
[30]

IEEE Transactions on information theory , volume=

A recursive approach to low complexity codes , author=. IEEE Transactions on information theory , volume=. 2003 , publisher=

work page 2003
[31]

Quantum , volume=

Sparse blossom: correcting a million errors per core second with minimum-weight matching , author=. Quantum , volume=. 2025 , publisher=

work page 2025
[32]

Nature , volume=

Learning high-accuracy error decoding for quantum processors , author=. Nature , volume=. 2024 , publisher=

work page 2024
[33]

12.0 , author=

Xilinx/brevitas: Release v0. 12.0 , author=. Zenodo , year=

work page
[34]

Canadian Journal of mathematics , volume=

Paths, trees, and flowers , author=. Canadian Journal of mathematics , volume=. 1965 , publisher=

work page 1965
[35]

Quantum , volume=

Degenerate quantum LDPC codes with good finite length performance , author=. Quantum , volume=. 2021 , publisher=

work page 2021
[36]

Physical Review A , volume=

Fault-tolerant weighted union-find decoding on the toric code , author=. Physical Review A , volume=. 2020 , publisher=

work page 2020
[37]

npj Quantum Information , volume=

Machine learning message-passing for the scalable decoding of QLDPC codes , author=. npj Quantum Information , volume=. 2025 , publisher=

work page 2025
[38]

arXiv preprint arXiv:2307.09025 , year=

qecGPT: decoding quantum error-correcting codes with generative pre-trained transformers , author=. arXiv preprint arXiv:2307.09025 , year=

work page arXiv
[39]

arXiv preprint arXiv:2510.06257 , year=

Toward Uncertainty-Aware and Generalizable Neural Decoding for Quantum LDPC Codes , author=. arXiv preprint arXiv:2510.06257 , year=

work page arXiv
[40]

arXiv preprint arXiv:2510.22724 , year=

Scalable Neural Decoders for Practical Real-Time Quantum Error Correction , author=. arXiv preprint arXiv:2510.22724 , year=

work page arXiv
[41]

Scientific reports , volume=

Deep neural network probabilistic decoder for stabilizer codes , author=. Scientific reports , volume=. 2017 , publisher=

work page 2017
[42]

Quantum Science and Technology , volume=

Deep neural decoders for near term fault-tolerant experiments , author=. Quantum Science and Technology , volume=. 2018 , publisher=

work page 2018
[43]

Physical review letters , volume=

Neural decoder for topological codes , author=. Physical review letters , volume=. 2017 , publisher=

work page 2017
[44]

IEEE Transactions on Quantum Engineering , year=

FPGA-based distributed union-find decoder for surface codes , author=. IEEE Transactions on Quantum Engineering , year=

work page
[45]

arXiv preprint arXiv:2510.21600 , year=

Real-time decoding of the gross code memory with FPGAs , author=. arXiv preprint arXiv:2510.21600 , year=

work page arXiv
[46]

Nature Electronics , volume=

A real-time, scalable, fast and resource-efficient decoder for a quantum computer , author=. Nature Electronics , volume=. 2025 , publisher=

work page 2025
[47]

arXiv e-prints , pages=

Pinball: A Cryogenic Predecoder for Surface Code Decoding Under Circuit-Level Noise , author=. arXiv e-prints , pages=

work page
[48]

Nano Futures , volume=

Real-time decoding for fault-tolerant quantum computing: Progress, challenges and outlook , author=. Nano Futures , volume=. 2023 , publisher=

work page 2023
[49]

Physical Review Research , volume=

Decoding across the quantum low-density parity-check code landscape , author=. Physical Review Research , volume=. 2020 , publisher=

work page 2020
[50]

Proceedings of the 28th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2 , pages=

Better than worst-case decoding for quantum error correction , author=. Proceedings of the 28th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2 , pages=

work page
[51]

arXiv preprint arXiv:2512.07737 , year=

A scalable and real-time neural decoder for topological quantum codes , author=. arXiv preprint arXiv:2512.07737 , year=

work page arXiv
[52]

arXiv preprint arXiv:2509.03954 , year=

LATTE: A Decoding Architecture for Quantum Computing with Temporal and Spatial Scalability , author=. arXiv preprint arXiv:2509.03954 , year=

work page arXiv
[53]

Proceedings of the 30th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2 , pages=

Micro blossom: Accelerated minimum-weight perfect matching decoding for quantum error correction , author=. Proceedings of the 30th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2 , pages=

work page
[54]

Physical review letters , volume=

Topological quantum distillation , author=. Physical review letters , volume=. 2006 , publisher=

work page 2006
[55]

Nature , pages=

Scaling and logic in the color code on a superconducting quantum processor , author=. Nature , pages=. 2025 , publisher=

work page 2025
[56]

IEEE Transactions on Computers , volume=

Comparing neural network based decoders for the surface code , author=. IEEE Transactions on Computers , volume=. 2019 , publisher=

work page 2019
[57]

arXiv preprint arXiv:2511.01741 , year=

HyperNQ: A Hypergraph Neural Network Decoder for Quantum LDPC Codes , author=. arXiv preprint arXiv:2511.01741 , year=

work page arXiv
[58]

Quantum , volume=

Stim: a fast stabilizer circuit simulator , author=. Quantum , volume=. 2021 , publisher=

work page 2021
[59]

IEEE Transactions on Information Theory , volume=

Hardness of decoding quantum stabilizer codes , author=. IEEE Transactions on Information Theory , volume=. 2015 , publisher=

work page 2015
[60]

Nature Communications , volume=

Artificial intelligence for quantum computing , author=. Nature Communications , volume=. 2025 , publisher=

work page 2025
[61]

IEEE Transactions on Quantum Engineering , volume=

Neural-network decoders for quantum error correction using surface codes: A space exploration of the hardware cost-performance tradeoffs , author=. IEEE Transactions on Quantum Engineering , volume=. 2022 , publisher=

work page 2022
[62]

arXiv preprint arXiv:0801.1241 , year=

On the iterative decoding of sparse quantum codes , author=. arXiv preprint arXiv:0801.1241 , year=

work page arXiv
[63]

An Introduction to Quantum Error Correction and Fault-Tolerant Quantum Computation

An introduction to quantum error correction and fault-tolerant quantum computation , author=. arXiv preprint arXiv:0904.2557 , year=

work page Pith review arXiv
[64]

2023 , publisher=

Suppressing quantum errors by scaling a surface code logical qubit , journal=. 2023 , publisher=

work page 2023
[65]

ACM SIGARCH Computer Architecture News , volume=

EIE: Efficient inference engine on compressed deep neural network , author=. ACM SIGARCH Computer Architecture News , volume=. 2016 , publisher=

work page 2016