High-Rate Concatenated Quantum Error-Correcting Codes for Qudits

arXiv:2610.11225 · quant-ph · Submitted 2026-10-08 · Read on arXiv

Listen

Transcript

Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.

Kai: I'm Kai, and with me are Mira and Lev, guest researcher.

Mira: Today's paper: "High-Rate Concatenated Quantum Error-Correcting Codes for Qudits".

Kai: The gist High-rate concatenated quantum codes for qudits extend many-hypercube (MHC) codes from qubits to qudits of prime dimension q,

Mira: First, who's behind it and why it matters.

Title and authors: Kai: So we're looking at this paper, "High-Rate Concatenated Quantum Error-Correcting Codes for Qudits," and it’s about taking those standard many-hypercube codes we use for qubits and extending them to qudits, which are quantum systems with a dimension q that isn't just two.

Mira: That’s right, and the authors are Takanori Nishi and Hayato Goto from RIKEN Center for Quantum Computing. It immediately tells you they're looking at how these codes work when you move from qubits to higher dimensional systems, which is a big step for building quantum computers.

Lev: From my side, I’m interested in seeing if this extension actually translates into something practical, like codes that can run on actual hardware and handle real noise levels.

Kai: Exactly Lev; the paper isn't just doing the math on paper, it’s proposing a way to use these codes for qudits because they offer a larger local Hilbert space than qubits.

Mira: The authors are extending many-hypercube codes, or MHC codes, from qubits to qudits of prime dimension q, and they construct both level-two and level-three MHC codes with specific parameter sets like six/q squared, four/q squared, and two/q squared for the level one cases <ref:2610.11225#pg1,codes, from qubits to qudits of prime dimension $q>.

Lev: I see them defining the stabilizers at level one like SZ(one) = Z 1Z 2Z 3Z 4Z 5Z six and SX(one) = X oneX twoX threeX fourXfiveX six which is the starting point for everything else <ref:2610.11225#pg1>.

Kai: And then they generalize the level-by-level minimum-distance decoder, which is a strategy originally developed for qubit MHC codes, to work with these qudit structures.

Mira: The title itself points toward high rate concatenated quantum codes, meaning they're trying to build codes that are efficient enough for practical use while still being robust against errors.

Lev: It seems like the immediate goal is showing how this generalization works formally before we worry about experimental noise models yet.

The paper's summary: Kai: Now, looking at the summary of "High-Rate Concatenated Quantum Error-Correcting Codes for Qudits," they are essentially taking those MHC codes and making them work for qudits of prime dimension q, which is a key area because qudits are seen as building blocks for efficient fault-tolerant quantum computation, or FTQC.

Mira: They construct the level-two and level-three MHC codes with parameters like six/q squared, four/q squared, and two/q squared for the first levels, and they define specific stabilizer generators and logical Pauli gates for those levels to make the code structure concrete <ref:2610.11225#pg1,construct the level-2 and level-3 MHC codes with parameters>.

Lev: They then move on to a decoding strategy called the level-by-level minimum-distance decoder, which is adapted from qubit MHC codes, but now applied to these qudit structures.

Kai: The core finding they present is that when you test their qudit MHC codes under an X-error model, their code capacity performance improves as the local dimension q increases.

Mira: Specifically, they find that the error threshold for the level-three case increases monotonically with q, reaching ten point zero percent for q = thirteen which is nearly double the qubit threshold of five point one percent <ref:2610.11225#pg1,q$, reaching 10.0% for $q = 13>.

Lev: That doubling of the threshold is significant because it means their code can tolerate a much higher physical error rate before the logical errors become dominant, which is exactly what we want for a fault-tolerant system.

Kai: They also noted that for sufficiently large local dimensions, they observe what they call a pronounced waterfall regime in qudit MHC codes, where the logical error rate drops much faster than you'd expect just from the code distance.

The paper's improvements: Mira: What’s interesting is how they interpret these results using entropy-based analysis of the underlying q-ary noise channel, which reveals a competition between two effects that are happening at once.

Kai: They found that increasing q enriches the syndrome alphabet and makes it easier to distinguish errors, but at fixed channel entropy, larger values of q also correspond to higher physical error rates and thus bigger typical error weights.

Lev: So the paper suggests a trade-off: you get better syndrome information with higher q, but you're fighting against larger physical noise effects that grow with it.

Mira: They quantified this competition using the ratio xi(L), which measures how close the typical error weight is to the weight of the lowest-weight uncorrectable error, and they showed this ratio actually increases with q at a fixed H q.

Kai: This helps explain why you see different behaviors in different regimes, like the waterfall regime versus when things are operating under lower noise conditions.

Lev: For someone trying to run this on hardware, knowing that the LLMD decoder is showing this kind of suppression of high-weight logical failures is important because it suggests a much better decoding approach than just relying on the code distance alone.

Conclusion: Kai: So to wrap up, the paper "High-Rate Concatenated Quantum Error-Correcting Codes for Qudits" shows that higher dimensional quantum systems can substantially improve high-rate concatenated codes, especially when you use advanced decoding techniques like the LLMD decoder.

Mira: The key point is that by using qudits and analyzing the entropy of the noise channel, they show a competition between getting richer syndrome information and dealing with growing physical error weights at fixed entropy.

Lev: I’d say what this means practically is that for qudit FTQC architectures, combining these high-rate concatenated codes with higher dimensions and these specific decoding methods might be a promising route toward low-overhead systems.

Kai: And they flag a point about the waterfall regime, which indicates that the LLMD decoder strongly suppresses certain types of high-weight logical failures when q is large.

Mira: They also suggest that while q thirteen was chosen for tractability, exploring larger dimensions like experimentally demonstrated q = twenty-five systems is an interesting direction for future work.

Lev: I agree; the LLMD decoder's performance under these models is what gives us confidence in using these qudit-based approaches, and we need to see how it handles noise models that are even more realistic than the X-error model they used here.

RIKEN Center for Quantum Computing (RQC) · Toshiba Corporation

quant-ph

Submitted: 2026-10-08

Updated: 2026-10-08

Comments: 11 pages, 2 figures

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 89/100

The gist: The gist High-rate concatenated quantum codes for qudits extend many-hypercube (MHC) codes from qubits to qudits of prime dimension q, finding that performance improves with increasing local

Key concepts

MHC Codes: Many-Hypercube codes are quantum error-correcting codes designed to protect quantum information. This paper extends these codes from qubits to qudits (systems with more than two states). The extension involves constructing level-2 and level-3 versions of these codes using specific stabilizer generators and logical gates tailored for qudits.
Concatenated Codes: Concatenated codes are a type of quantum error correction where multiple layers of codes are nested together. This paper uses high-rate concatenated structures, meaning the overall code is efficient in terms of the number of physical qubits needed to encode one logical qubit. It involves recursive construction and decoding across different levels.
Syndrome Information Enrichment: This refers to how the syndrome measurements (the results from error checks) help distinguish between different types of errors. Increasing the dimension q enriches this information, making it better at identifying which specific error occurred. This enrichment competes with other factors in determining overall code performance.
Waterfall Regime: The waterfall regime is an observation in qudit MHC codes where the logical error rate drops much faster than predicted by standard distance-based estimates, especially for large local dimensions. This phenomenon is explained by the interplay between syndrome information and the growth of typical physical error weights at fixed noise entropy.

Terminology

Hamming Distance — 7× in this paper · explained in 2 episodes
Distillation — 5× in this paper · explained in 2 episodes
Hilbert Space — 4× in this paper · explained in 2 episodes
Entanglement — 2× in this paper · explained in 4 episodes
Anyons — 1× in this paper · explained in 2 episodes
Computational Tractability — 1× in this paper · explained in 2 episodes
Consistency — 1× in this paper · explained in 3 episodes
Decode — 1× in this paper · explained in 2 episodes
Generalization — 1× in this paper · explained in 13 episodes

Summary

The gist High-rate concatenated quantum codes for qudits extend many-hypercube (MHC) codes from qubits to qudits of prime dimension q, finding that performance improves with increasing local dimension q and revealing a competition between syndrome information enrichment and typical physical error weight growth

Code Construction and Generalization

The work extends the MHC code family from qubits to qudits of prime dimension q, constructing the level-2 and level-3 MHC codes with parameters 6/q2, 4/q2, 2/q2 and 6/3, 4/3, 2/3 q, respectively The encoding rate is defined as R = k/n for a [[n, k, d]]q code. At level-1, the stabilizer generators are chosen as SZ(1) = Z1Z2Z3Z4Z5Z6 and SX (1) = X1X†2X3X†4X5X†6. The logical Pauli gates are defined as Z(1) 1 = Z1Z2, X(1) 1 = X2X†3, Z(1) 2 = Z2Z3, and X(1) 2 = X†1X2. For the level-3 case, the stabilizers SZ(3)i′j′,k′ and Pauli operators Z(3)i′j′,k′, X(3)i′j′,k' are derived by decomposing Pauli operators recursively.

Decoding Strategies

Two decoding strategies are considered: the hard-decision (HD) decoder inspired by Knill’s concatenated decoding scheme and the level-by-level minimum-distance (LLMD) decoder originally developed for the qubit MHC codes The HD decoder recursively checks stabilizer consistency at each concatenation level, where a block is accepted if its syndrome is consistent with the code constraints. In contrast, the LLMD decoder retains all encoded strings that achieve the minimum Hamming distance from the observed measurement outcomes, propagating these candidates recursively while enforcing stabilizer constraints between neighboring blocks.

Performance Analysis and Thresholds

Numerical simulations evaluate code capacity under a qudit X-error model, finding that the error threshold increases monotonically with q for the LLMD decoder, reaching 10.0% for q = 13, nearly double the qubit threshold of 5.1%. The performance of the LLMD decoder proves with increasing q, with the threshold increasing monotonically and nearly doubling from 5.1% for q = 2 to 10.0% for q = 13. Furthermore, a pronounced waterfall regime in qudit MHC codes is observed for sufficiently large local dimensions, where the logical error rate decreases substantially faster than expected from conventional distance-based estimates.

Entropy-Based Interpretation

The analysis of decoding performance in terms of the entropy of the underlying qary noise channel reveals a competition between two effects: increasing q enriches the syndrome alphabet and improves distinguishability, while at fixed channel entropy, larger values of q correspond to higher physical error rates and hence larger typical error weights. This competition naturally explains the different q dependences in the waterfall and low-error regimes. The ratio ξ(L) ≡ n(L)ε/(d(L)/2) quantifies the proximity of the typical error weight to the weight of the lowest-weight uncorrectable error, and this ratio increases with q at fixed Hq.

Conclusion

The results demonstrate that higher-dimensional quantum systems can substantially improve the performance of high-rate concatenated quantum codes and motivate further development of quditbased FTQC architectures. The paper suggests that combining high-rate concatenated code architectures with higher-dimensional quantum systems and advanced decoding techniques may offer a promising route toward low-overhead FTQC. The emergence of a waterfall regime for sufficiently large values of q indicates that the LLMD decoder strongly suppresses classes of high-weight logical failures. This phenomenon is explained by the competition between syndrome information enrichment and the growth of typical physical error weights at fixed entropy. The results show that higher-dimensional quantum systems can significantly improve the performance of high-rate concatenated quantum codes when the available syndrome information is utilized effectively. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 reflects a compromise between computational tractability and experimental relevance. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work. The choice of q ≤ 13 is not intended to represent a fundamental experimental limit. Exploring still larger dimensions, including experimentally demonstrated q = 25 systems, is an interesting direction for future work. The paper leaves the extension of the LLMD decoder to such noise models for future work.

Improvements for AI systems

Improved QEC performance for higher-dimensional quantum systems: The system can achieve a logical error rate decrease substantially faster than expected from conventional distance-based estimates in high-rate concatenated qudit codes due to a pronounced waterfall regime, as observed in the paper's results.
Enhanced fault tolerance via decoding strategy: The AI system can utilize the Level-by-Level Minimum-Distance (LLMD) decoder, which is shown to substantially outperform the Hard-Decision (HD) decoder, leading to a threshold that increases monotonically with q, reaching 10.0% for q=13.
Entropy-aware noise channel analysis: The system can perform decoding by analyzing the qary entropy of the noise channel, which reveals a competition between the increased syndrome information available at larger q and the simultaneous growth of the typical physical error weight at fixed entropy.

Abstract

Qudits provide a larger local Hilbert space than qubits and have recently attracted interest as building blocks for efficient fault-tolerant quantum computation (FTQC). In this work, we extend many-hypercube (MHC) codes, high-rate concatenated quantum codes, from qubits to qudits of prime dimension q. We construct the level-2 and level-3 MHC codes with parameters [! [6 squared,4 squared,2 2]!] q and [! [6 cubed,4 cubed,2 3]!] q, respectively, and generalize the level-by-level minimum-distance (LLMD) decoder originally proposed for the qubit MHC codes. In a qudit X-error model, we evaluate the code-capacity performance of the qudit MHC codes for dimensions up to q=13. We find that the error threshold increases monotonically with the local dimension q, reaching 10.0% for q=13, nearly double the value of 5.1% for qubits. Furthermore, for sufficiently large local dimensions, we observe a pronounced waterfall regime in which the logical error rate decreases substantially faster than expected from conventional distance-based estimates. To understand the origin of the observed qudit advantage, we analyze the decoding performance in terms of the entropy of the underlying q-ary noise channel. This analysis reveals a competition between the increased syndrome information available at larger q and the simultaneous growth of the typical physical error weight at fixed entropy. These results demonstrate that higher-dimensional quantum systems can substantially improve the performance of high-rate concatenated quantum codes and motivate further development of qudit-based FTQC architectures.