Searcharxiv⌕ Search
Home/Archive/arXiv

arXiv · 2610.28748

CoDecoder: Make Quantum Decoders Up to 100x Faster and Even More Accurate

Yicheng Guang·Yuliang Xiong·Xun Gao·Yueqi Chen
View original source ↗Read source document ↗Export metadata ↓

Abstract

Fault-tolerant quantum computers need fast and accurate error decoding. We present CoDecoder, a small neural frontend that composes with an existing decoder to improve accuracy and reduce mean decoding time. It predicts each shot's logical outcome, retains confident predictions and routes the remaining shots to the unchanged fallback. A logical-operator-aware output head scores shifted logical operators, combining equivalent logical lines on surface codes and deciding logical qubits jointly on bivariate-bicycle (BB) codes. Fused GPU inference and efficient routing reduce the overhead paid on every shot. In registered tests under circuit-level noise, composition reduces failures by 6.4-31.4% relative to correlated matching on surface codes of distance 7-17 and by 22-50% relative to full-DEM Relay-BP on two BB codes, with 2.3-124-fold speed-ups across these comparisons. With our re-implementations of the neural decoder Cascade, composition reduces decoding time in all four settings. A joint decision of both decoders on the routed shots keeps this speed and makes the composition more accurate than Cascade in all four, with 14-18% fewer failures on the [[72,12,6]] code. On the [[72,12,6]] BB code, a conventional head with independent logical outputs fails 3.8-5.0 times as often as our joint class score under matched training conditions. Per-shot analysis shows how composition gains accuracy. On retained shots, the frontend corrects more fallback errors than it introduces, even when it is less accurate overall.

Explore related subjects

quant-ph

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Yicheng Guang, Yuliang Xiong, Xun Gao, Yueqi Chen. 2026-10-20. CoDecoder: Make Quantum Decoders Up to 100x Faster and Even More Accurate. https://arxiv.org/abs/2610.28748

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Everlasting secure key agreement from a quantum computational hybrid security model

We introduce the Quantum Computational Timelock (QCT) model, combining time-lock encryption with a quantitative noisy-quantum-storage assumption. Time-lock encryption gives Bob immediate access to the encoding seed and gives Eve the same final information as access to that seed only after a protected interval. Eve may then process her retained information without computational restriction. We state this delayed-access encryption assumption explicitly and construct MUB-QCT from a complete set of mutually unbiased bases in dimension $d$ and an affine pair-wise independent permutation family. For single-copy encoding, every seed-independent immediate measurement has trace-distance advantage at most $1/(2\sqrt{d+1})$ and mutual information at most $1/(d+1)$. Effective residual storage at diamond distance at most $δ$ from replacement adds at most $δ$ to the individual-attack distinguishing bound. Under i.i.d. individual attacks and entropy-preserving acceptance, we obtain an asymptotic key-rate bound and formulate final-key secrecy in composable trace distance. For $m$ independently randomized signals sharing one bit and seed, the optimal leakage under non-adaptive product measurements is $\min\{1,m/(d+1)\}$, allowing a linear-in-$d$ detection enhancement in that attack model. We also construct a seed-independent joint measurement that recovers the bit after disclosure with error at most $\varepsilon$ from $O(\log d+\log(1/\varepsilon))$ signals, retaining only classical information after measurement. This gives a quantitative separation between product and joint measurements. Bob uses two logical detection outcomes independently of $d$.

quant-ph↗
arXiv2026

Linear optics and the problem of Bell-like state discrimination

A linear optics-based scheme to implement various quantum information processing tasks is of paramount importance due to its ease of implementation and low noise. Many information-theoretic tasks depend on the successful discrimination of Bell states. A no-go theorem states that it is not possible to perfectly discriminate among the four Bell states when the measurement apparatus is restricted to linear optical elements. The success probability is only $50\%$. Using extra resources such as hyperentanglement, ancillary entanglement, and even a minimum amount of non-linearity, complete Bell-state discrimination can be achieved. The success probability for Bell-like state discrimination is only $25\%$. Bell-like states are non-maximally entangled two-qubit states that resemble Bell states in structure but have unequal Schmidt coefficients. We find that this can be boosted to $50\%$ using hyperentanglement, where extra ancillary degrees of freedom (DOF) of photons, such as polarization, momentum, or OAM, are also entangled along with the system DOF. This contrasts with Bell-state discrimination, where $100\%$ can be achieved. Furthermore, we find that using correlations in the photons' arrival times, all four Bell states can be distinguished with $100\%$ success probability, while for Bell-like state discrimination, it increases only to $37.5\%$ from $25\%$. We observe a similar contrast when using ancillary entangled photons. While the success probability for discriminating all four Bell states increases as $1-\frac{1}{2^N}$, where $N$ is the number of ancillary photon pairs, for Bell-like states, it depends on the state parameters and can be less than $25\%$ in some cases. Also, adding further ancillary photons decreases the success probability. We then show that using non-linear gadgets, namely SFG, $100\%$ success probability can be achieved even for Bell-like state discrimination.

quant-ph↗
arXiv2026

Memory-Augmented Hybrid Quantum Reservoir Computing

Reservoir computing (RC) is an effective method for predicting chaotic systems by using a high-dimensional dynamic reservoir with fixed internal weights, while keeping the learning phase linear, which simplifies training and reduces computational complexity compared to fully trained recurrent neural networks (RNNs). Quantum reservoir computing (QRC) uses the exponential growth of Hilbert spaces in quantum systems, allowing for greater information processing, memory capacity, and computational power. However, the original QRC proposal requires coherent injection of inputs multiple times, complicating practical implementation. We present a hybrid quantum-classical approach that implements memory through classical post-processing of quantum measurements. This approach avoids the need for multiple coherent input injections and is evaluated on benchmark tasks, including the chaotic Mackey-Glass time series prediction. We tested our model on two physical platforms: a fully connected Ising model and a Rydberg atom array. The optimized model demonstrates promising predictive capabilities, achieving a higher number of steps compared to previously reported approaches.

quant-ph↗