Pith. sign in

REVIEW 4 major objections 5 minor 58 references

Transversal Gates for Highly Asymmetric qLDPC Codes

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper claims the first qLDPC codes with transversal gates outside the Clifford group, including a balanced product code with transversal T and a direct construction with transversal P_q.

desk verdict First believable qLDPC codes with transversal non-Clifford gates; the main example checks out, but the verification is notebook-bound and the abstract overclaims the scaling. read the letter →

arxiv 2506.15905 v1 pith:TDOOMO2M submitted 2025-06-18 quant-ph

Heather Leitch , Alastair Kay This is my paper
classification quant-ph MSC 81P6881P70
keywords transversalgatesquantumLDPCcodesnon-CliffordphasebalancedproducthypergraphCSSfaulttolerance
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

Transversal gates are attractive for fault tolerance because they propagate errors minimally, but quantum LDPC codes were widely suspected to forbid non-Clifford examples. This paper claims the first constructions of qLDPC codes with transversal gates outside the Clifford group---phase gates $P_q$, including $S$ and $T$. The central example is a balanced product code $[[1080,232,3]]$ whose $16$-logical-qubit subsystem admits transversal $T$, plus a direct construction that yields $k$ logical qubits with transversal $P_q$ and $d_X$ almost linear at the price of $d_Z = O(1)$. The authors also show that distance balancing preserves transversality only when it increases the $X/Z$ asymmetry, and they assemble a toolbox that turns one transversal phase gate into individual phase and controlled-phase gates. If right, transversal non-Clifford gates are achievable in the qLDPC regime, opening a route to universal computation on high-rate codes despite strongly asymmetric distance.

What carries the argument

The engine is Lemma 1, a criterion for a CSS code to have a transversal $P_q$ gate: with the row-wise dot-product convention $x\cdot y$ that records positions where both bit strings have a $1$, one needs $|H_X^{\cdot i}\cdot L_X^{\cdot j}\cdot p| \equiv 0 \pmod{2^{q+1-i-j}}$ for all $i+j=1,\dots,q$ and $|L_X\cdot p| \equiv w \pmod{2^q}$, for some phase pattern $p\in[2q]^N$. Assumption 1, that the logical $X$ operators can be chosen with pairwise disjoint physical support, makes every $j\ge 2$ term vanish and so drives the whole verification; its price is the inequality $d_X\,k\le n$. The construction machinery is the transpose Tanner code (Tanner codes combine an expander graph with a small local code $C_0$): rather than use a Tanner matrix $A$ directly, the paper puts $A^T$ into $H_X$, so rows of $H_X$ are composed of blocks of rows of $C_0^T$, and any local code $C_0$ whose transpose satisfies the Lemma 1 parity conditions transplants a transversal phase gate into the whole qLDPC code. A secondary mechanism is the gate toolbox (Lemmas 2--5), which converts the single transversal $P_w$ into individual $P_{q-1}$ gates and controlled-phase gates.

What would settle it

Enumerate every logical basis state of the claimed $16$-qubit subsystem of the $[[1080,232,3]]$ code, apply the physical transversal $T$, and check that each state acquires exactly the phase prescribed by logical $T$; any mismatch shows the example is not a transversal $T$ gate. Equally decisive would be an explicit CSS qLDPC code satisfying the Lemma 1 conditions with both distances growing with $n$ and $k$ linear, which would break the paper's $O(1)$ $Z$-distance limitation.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that qLDPC codes with transversal non-Clifford gates exist, contradicting a prior expectation while sidestepping the known no-go result because that result applies only to robust hypergraph product codes and the new examples are not robust. The balanced product example is a $[[1080,232,3]]$ CSS code; restricting to a $16$-logical-qubit subsystem whose logical operators satisfy Lemma 1 with $q=3$, applying $T$ transversally at the physical level realizes a logical transversal $T$. The direct construction of Section VI instead parallels many copies of a small local code $C_0$ with the required parity properties, giving transversal $P_q$ and $k$ logical qubits; its bit-flip distance grows almost linearly with $n$ while its phase-flip distance is $O(1)$. The paper also establishes that known distance-balancing operations keep the transversal gate only in the direction that increases the asymmetry, and that from a single transversal $P_w$ one can derive individual logical phase gates and controlled-phase gates.

Load-bearing premise

Everything in the gate verification rests on Assumption 1: the logical $X$ operators can be written so that no two of them share a physical qubit, which kills the higher-order terms in Lemma 1 but also forces $d_X k \le n$.

Editorial extensions

If this is right

  • Transversal $T$ (or any $P_q$) can be applied directly to these qLDPC codes, so fault-tolerant circuits built on them would not need to distill the non-Clifford gate.
  • The direct construction gives $k \sim n$ logical qubits with transversal $P_q$ and $d_X$ almost linear, so the rate stays high even with the gate advantage.
  • The $d_Z=O(1)$ bound means these codes are best used as inner components in a concatenated or asymmetric-noise setting, not as standalone memories.
  • Because distance balancing preserves transversality only when increasing asymmetry, any attempt to symmetrize these codes must abandon this balancing method or lose the gate.
  • From a single transversal $P_w$ with odd $w$, the toolbox produces localised $S$ gates and controlled-phase gates within one block, so a single seed gate generates a much wider gate set.

Reading between the lines

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

  • The authors leave implicit that Assumption 1 is the real bottleneck; if it can be relaxed, the same transpose-Tanner idea might yield transversal non-Clifford gates with non-constant $Z$ distance, and the $d_X k \le n$ inequality hints at a trade-off theorem rather than a construction artifact.
  • Because the direct construction is essentially $k$ parallel copies of a small local code, any local code whose rows satisfy the same dot-product parity conditions (for instance the 7-bit code used for $S$ or the 15-bit code used for $T$) should seed an infinite family with the same transversal gate, independent of the expander graph's expansion.
  • A concrete applied test not performed here is to concatenate these codes with an outer code that corrects $Z$ errors; if the inner code's transversal $T$ survives the concatenation, the usual $T$-magic-state distillation overhead could be avoided.
  • The toolbox suggests the design target worth optimising is a single transversal $P_4$ (logical $\sqrt{T}$), since it localises to $T$ without ancillas; instances at high distance would make the practical payoff much larger than the particular examples here.
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

4 major / 5 minor

Summary. The paper claims to give the first constructions of qLDPC codes with transversal non-Clifford gates, contradicting the earlier Burton–Browne no-go for robust hypergraph product codes. The main ingredients are a restatement of known necessary and sufficient conditions for transversal phase gates (Lemma 1), a disjoint-support assumption on logical X operators (Assumption 1), and explicit constructions based on transpose Tanner codes embedded in balanced product and hypergraph product codes. The paper presents a [[140,16,3,5]] example with transversal S, a [[1080,232,3]] example with a 16-logical-qubit subsystem claimed to have transversal T, and a direct construction that amounts to parallel copies of a local code. It also collects standard circuit identities into a 'toolbox' for localizing phase gates and producing controlled-phase gates, and it shows that one direction of distance balancing does not preserve transversality.

Significance. If the central examples verify, the paper would be a genuine counterexample to the expectation that qLDPC codes cannot host transversal non-Clifford gates, and the toolbox for localizing and combining such gates is useful for subsequent work. The paper is commendably explicit about the weaknesses of its constructions, including the O(1) Z distance and the restrictive Assumption 1, and it ships a Mathematica notebook for the numerical examples. The main risk is that the crucial verification of the transversal T example is delegated to the notebook and is not present in the text; the abstract also overstates the achievable code parameters relative to the paper's own bounds.

major comments (4)
  1. [Abstract and Section I (Assumption 1)] The abstract claims codes with k growing linearly in n and bit-flip distance growing '(almost) linearly' in n, but Assumption 1 in Section I gives d_X k ≤ n, so these two scalings cannot hold simultaneously. Section VII similarly reports d_X∼O(l) and k∼O(sm/l) for the balanced-product family, giving k d_X = O(n). Please correct the abstract and specify which asymptotic regime (constant l, growing l, or the Section VI trade-off) is claimed for each property.
  2. [Section IV.B] The central T-gate example [[1080,232,3]] is not verified in the text. The sentence 'These satisfy all the requirements of Lemma 1 for q = 3' is an assertion, not a derivation. In particular, the conditions with i+j=2 and i+j=3 (for example |H_X·L_X·p| mod 4 and |H_X·H_X·L_X·p| mod 2) involve overlaps between rows of A^T and logical operators supported on full R-orbits; these overlaps can have support outside a single C0 block, so the block-wise reduction stated earlier in the section is not sufficient as written. Provide the explicit p vector, the 16 logical operators, and the computed dot products, or include the full notebook verification as an appendix, so the claim is checkable from the manuscript alone.
  3. [Section IV.A] The same verification gap occurs for the [[140,16,dZ=3,dX=5]] example with transversal S: the text states 'We can verify that |H_X·p|≡0 mod 4 and |H_X·H_X·p|≡0 mod 2' without giving p or the logical operators used, and the claim that the logical X operators act on distinct sets of qubits is stated without proof. Since this example motivates the entire transpose-Tanner approach, please include the explicit p and a table of the verification, or move the notebook computation into the main text or a self-contained appendix.
  4. [Section VI and Section VII] The Section VI 'direct construction' is k parallel copies of a single-qubit storage code (AT = 1_{k dX}⊗C0), so it reproduces, without expansion, the transversality of the local code. This does not yield codes with both k linear and d_X growing; it is precisely a trade-off between k and d_X. The Section VII summary claims 'first constructions of qLDPC codes that have transversal gates which are not Clifford'; given that the direct construction is parallel repetition and the balanced-product example has transversal T on only a 16-qubit subsystem (with dZ=3), the strength of this claim needs to be qualified clearly in the text.
minor comments (5)
  1. [Section IV.B] The term 'subsystem code' is used without definition; please clarify whether the 16 logical operators are a genuine subsystem and how the remaining logical qubits behave under the transversal gate.
  2. [Section III.C] The Tanner construction allows arbitrary column permutations of C0; state explicitly which ordering is used in the examples and whether the claimed transversality verification depends on that ordering.
  3. [Section I] The notation [2q]^n for p is defined, but Eq. (1) uses P_pi^q without explicitly saying that the scalar p_i is taken modulo 2q; please make the modular arithmetic of the physical phases explicit.
  4. [Sections IV.A and IV.B] Reference [39] (the notebook) is cited only at the end of Section IV.A; please cite it at the point of each numerical verification, including the [[1080,232,3]] example.
  5. [Section V] The distance-balancing discussion states that rebalancing to increase asymmetry preserves transversality, but the text could make this positive direction more explicit before discussing the negative direction.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity found: transversality is lifted from explicit local codes through balanced/hypergraph product blocks with algebraic verification, and the paper's self-citations are not load-bearing.

full rationale

The paper's derivation chain is not circular. Lemma 1 is an external restated criterion for transversal phase gates; the authors then construct explicit parity-check matrices and logical operators for balanced and hypergraph product codes, verifying the Lemma 1 dot-product conditions algebraically. The local code C0 (Hamming or Reed-Muller) is chosen because it already satisfies the single-code transversality conditions, and the transpose-Tanner block structure ensures these conditions lift to the large code; this is ordinary code composition, not a fitted prediction or a definitional equivalence. The central Assumption 1 (disjoint logical-X supports) is stated explicitly and its distance cost dX·k ≤ n is quantified, so it is not smuggled in. Self-citations [30] and [32] are present but are not load-bearing: [30] only notes that the balanced-product construction need not use Tanner codes, and [32] supplies a side remark about minimal q-orthogonal codes. The most complex example's verification is delegated to a supporting notebook [39]; that is a reproducibility and completeness concern, not a circular argument. Even the Section VI direct construction, which the authors themselves describe as 'literally just k parallel copies of a single-qubit storage option,' is an honest explicit construction whose claimed transversal gate is established by inspection of the explicit matrices rather than by fitting. No step reduces to its own inputs by construction, and no fitted parameter is renamed as a prediction.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The construction is an engineering of transversality from a small local code, with the cost paid in Z distance. No fitted numerical parameters appear; the free parameters are combinatorial design choices (the local code and the graph symmetry). The key structural assumption, Assumption 1, is a restrictive design choice that bounds the distance trade-off. No new physical entities are postulated.

free parameters (2)
  • Local code C0 = [7,4,3] Hamming code (S example) or 15-bit Hamming code (T example)
    Chosen by hand to satisfy the q-orthogonal condition |C0^{·i}| = 0 mod 2^{q+1-i}; this is what gives the large code its transversal gate. The Z distance is then capped at d(C0).
  • Graph and symmetry order l = Kirkman (15,3,1) design with l=5; complete graph on 16 vertices with l=3
    Chosen to provide the LDPC structure and the symmetry whose orbits define the 'good' logical qubits (one per orbit); the X distance scales as O(l) and k as O(m/l).
assumptions (4)
  • standard math Lemma 1 (transversal phase gate conditions) from Webster, Quintavalle and Bartlett [31]
    The paper restates and relies on the necessary and sufficient conditions for a CSS code to have a transversal P_q gate; the proof is not reproduced, only cited.
  • ad hoc to paper Assumption 1: logical X operators act on disjoint supports
    Invoked in Section I to ignore all terms with j≥2 in Lemma 1's conditions. The paper admits it limits dX·k ≤ n, so code distance against X errors is bounded by n/k.
  • domain assumption The local code C0 satisfies |C0^{·i}| = 0 mod 2^{q+1-i}
    This is the mechanism that transfers transversality from C0 to the transpose Tanner code (Section IV). It forces C0 to have a null vector (all-ones), which caps the Z distance at d(C0)=O(1).
  • domain assumption The graph has a symmetry R A = A C^T with all orbits of length l (balanced product)
    Required for the balanced product code commutation and for the orbit structure that yields the logical operators in Eq. (7) satisfying Assumption 1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Transversal Gates for Highly Asymmetric qLDPC Codes." pith.science (2026). https://pith.science/paper/TDOOMO2M

@misc{pith2026250615905,
  author       = {Pith},
  title        = {Pith review of: Transversal Gates for Highly Asymmetric qLDPC Codes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TDOOMO2M}},
  note         = {Machine review of arXiv:2506.15905}
}
abstract

Transversal gates are the ideal gates in a fault-tolerant scenario; relatively easy to implement, and minimally error propagating. Their availability will maximise fault tolerant thresholds, enabling universal quantum computation in a wider range of noisy hardware. Transversal gates in quantum low density parity check (qLDPC) codes are largely unstudied, with the early results of Burton & Browne suggesting that transversal non-Clifford gates may be impossible. In this paper, we contradict this expectation with constructions for both hypergraph product codes and balanced product codes, although these first examples have weak properties. We find qLDPC codes with transversal phase gates that have a number of logical qubits that grows linearly with $n$, the number of physical qubits. The distance is highly asymmetric; while the distance against bit flip errors also grows (almost) linearly in $n$, the distance against phase flip errors is limited to $O(1)$. Moreover, existing distance rebalancing techniques are one-sided; we show that they only preserve the transversality when rebalancing to increase the asymmetry, not decrease it. We also collate a toolbox of techniques that identify a single transversal gate from which we can individuate a large range of different transversal gates. This is critical when addressing the question of what a transversal phase gate truly means when there are many logical qubits in the system.

Figures

Figures reproduced from arXiv: 2506.15905 by the authors.

Figure 1
Figure 1. In a gate sequence P w L,q, XL, P −w L,q , XL, those logical qubits (top row) which do not have the logical X applied cancel, and yield an 1. Those on which XL is applied (bottom row) create a localised logical gate of twice the phase. Assumption 1. There exists a representation of the generators of the logical X operators LX which all act on distinct sets of physical qubits, i.e. l1 · l2 = 0 for all l1 ̸= l2 ∈ LX. … view at source ↗
Figure 2
Figure 2. In a gate sequence of controlled-not (transversal, between two different code blocks), PL, controlled-not, P † L , then those qubits that are not acted upon by PL (top row) are unchanged. Those acted upon by PL (bottom row) gain a net controlled-phase action. 6. Apply any necessary phase PL,q−1 correction based on measurement results, as specified in the follow￾ing circuits. (We have these by Lemma 2.) At the end of… view at source ↗
Figure 4
Figure 4. Circuit for performing controlled-phase between [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

58 extracted references · 37 canonical work pages

  1. [1]

    1 applies, giving a net phase gate of P 2wi q

    For logical qubit i, the bottom row of Fig. 1 applies, giving a net phase gate of P 2wi q . Note that P 2 q =Pq−1. We can now apply the circuit identities and conclude that if Pw L,q = Pp q , then Pwiei L,q = Pp·li q−1 where li is the logical X operator corresponding to logical qubit i. Thus, with q = 4, we can convert P 1,1,...,1 L,4 , transversal logica...

    work page

  2. [2]

    Apply Pw L,q to the ancilla block

    work page

  3. [3]

    Prepare an ancilla block in |+⟩⊗k byX basis mea- surement

    work page

  4. [4]

    Apply transversal c- not controlled by the data block and targetting the ancillas

    work page

  5. [5]

    Perform an X basis measurement on every logical qubit of the ancilla block except i

    work page

  6. [6]

    (We have these by Lemma 2.) At the end of step 3, every logical qubit of the ancilla block is in state |+⟩ except for qubit i, which is in the statePwi q |+⟩

    Apply any necessary phase PL,q−1 correction based on measurement results, as specified in the follow- ing circuits. (We have these by Lemma 2.) At the end of step 3, every logical qubit of the ancilla block is in state |+⟩ except for qubit i, which is in the statePwi q |+⟩. For all logical qubits except i, the ensuing circuit is thus |ψ⟩ |ψ⟩ |+⟩ meaning t...

    work page

  7. [7]

    5 = = P P† P 2 Figure 2

    Measure the ancilla block in the Z basis. 5 = = P P† P 2 Figure 2. In a gate sequence of controlled- not (transversal, between two different code blocks), PL, controlled- not, P† L, then those qubits that are not acted upon by PL (top row) are unchanged. Those acted upon by PL (bottom row) gain a net controlled-phase action

    work page

  8. [8]

    Eastin and E

    B. Eastin and E. Knill, Restrictions on Transversal En- coded Quantum Gate Sets, Phys. Rev. Lett. 102, 110502 (2009)

    work page 2009

Show all 58 references
  1. [9]

    Apply a transversal controlled- not between two control blocks

  2. [10]

    On the target block, apply Pei L,2

  3. [11]

    intra-block gates inter-block gates [[n,k,d ]] code block Figure 3

    Apply the same transversal controlled-not between the two control blocks. intra-block gates inter-block gates [[n,k,d ]] code block Figure 3. For multiple instances (“blocks”) of a qLDPC code, transversal gates can be applied in two ways. Inter-block copies interact different ...

  4. [12]

    local code

    On the target block, apply P−ei L,2 . On the logical qubits j̸=i (Fig. 2, top row, taking P = S), the only action is two controlled- nots, which cancel each other. On logical qubits i (Fig. 2, bottom row), the net effect is controlled-phase between that pair of logical qubits ...

  5. [13]

    Bravyi and R

    S. Bravyi and R. Koenig, Classification of topologically protected gates for local stabilizer codes, Phys. Rev. Lett. 110, 170503 (2013), arXiv:1206.1609 [quant-ph]

    2013 arXiv

  6. [14]

    V. V. Sivak, A. Eickbusch, B. Royer, S. Singh, I. Tsiout- sios, S. Ganjam, A. Miano, B. L. Brock, A. Z. Ding, L. Frunzio, S. M. Girvin, R. J. Schoelkopf, and M. H. De- voret, Real-time quantum error correction beyond break- even, Nature 616, 50 (2023)

    2023

  7. [15]

    Aliferis, D

    P. Aliferis, D. Gottesman, and J. Preskill, Quantum accu- racy threshold for concatenated distance-3 codes, Quant. Inf. Comput. 6, 97 (2006)

    2006

  8. [16]

    Aliferis and A

    P. Aliferis and A. W. Cross, Subsystem fault tolerance with the Bacon-Shor code, Phys. Rev. Lett. 98, 220502 (2007), arXiv:quant-ph/0610063

    2007 arXiv

  9. [17]

    Paetznick and B

    A. Paetznick and B. W. Reichardt, Fault-tolerant ancilla preparation and noise threshold lower bounds for the 23- qubit Golay code (2013), arXiv:1106.2190

    2013 arXiv

  10. [18]

    Knill, Quantum computing with realistically noisy de- vices, Nature 434, 39 (2005)

    E. Knill, Quantum computing with realistically noisy de- vices, Nature 434, 39 (2005)

    2005

  11. [19]

    Dennis, A

    E. Dennis, A. Kitaev, A. Landahl, and J. Preskill, Topo- logical quantum memory, J. Math. Phys.43, 4452 (2002)

    2002

  12. [20]

    N. P. Breuckmann, K. Duivenvoorden, D. Michels, and B. M. Terhal, Local Decoders for the 2D and 4D Toric Code (2016), arXiv:1609.00510

    2016 arXiv

  13. [21]

    Even at more modest sizes, the properties are promising [22]

    and surface codes. Even at more modest sizes, the properties are promising [22]. To date, little is known about the ability to perform fault-tolerant computation on qLDPC codes, other than some very general results which lack the improvements that can be yielded by specialisat...

    2025 arXiv

  14. [22]

    C. M. Dawson and M. A. Nielsen, The Solovay-Kitaev algorithm (2005), arXiv:quant-ph/0505030

    2005 arXiv

  15. [23]

    Kliuchnikov, D

    V. Kliuchnikov, D. Maslov, and M. Mosca, Fast and ef- ficient exact synthesis of single qubit unitaries generated by Clifford and T gates (2012), arXiv:1206.5236

    2012 arXiv

  16. [24]

    Kliuchnikov, A

    V. Kliuchnikov, A. Bocharov, M. Roetteler, and J. Yard, A Framework for Approximating Qubit Unitaries (2015), arXiv:1510.03888

    2015 arXiv

  17. [25]

    A. Y. Kitaev, Quantum computations: Algorithms and error correction, Russ. Math. Surv. 52, 1191 (1997)

    1997

  18. [26]

    For the transpose Tanner codes of Section IV, assume that we have some initial code A0 which is an m×n matrix with row totals s and column totals c, meaning that ms = nc

    does not apply. For the transpose Tanner codes of Section IV, assume that we have some initial code A0 which is an m×n matrix with row totals s and column totals c, meaning that ms = nc. Now take the Tanner code of it with a matrix C0 which iss×s. This yields an ms×n parity-ch...

  19. [27]

    Bravyi, D

    S. Bravyi, D. Poulin, and B. Terhal, Tradeoffs for reliable quantum information storage in 2D systems, Phys. Rev. Lett. 104, 050503 (2010), arXiv:0909.5200 [quant-ph]

    2010 arXiv

  20. [28]

    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, J. P. Bonilla Ataides, N. Maskara, I. Cong, X. Gao, P. Sales Rodriguez, T. Karolyshyn, G. Semeghini, M. J. Gullans, M. Greiner, V. Vuleti´ c, and M. D. Lu...

    2024

  21. [29]

    Tillich and G

    J.-P. Tillich and G. Zemor, Quantum LDPC codes with positive rate and minimum distance proportional to n 1 2 , IEEE Trans. Inform. Theory 60, 1193 (2014), arXiv:0903.0566 [quant-ph]

    2014 arXiv

  22. [30]

    N. P. Breuckmann and J. N. Eberhardt, Balanced Prod- uct Quantum Codes, IEEE Transactions on Information Theory 67, 6653 (2021)

    2021

  23. [31]

    Panteleev and G

    P. Panteleev and G. Kalachev, Asymptotically good Quantum and locally testable classical LDPC codes, in Proceedings of the 54th Annual ACM SIGACT Sympo- sium on Theory of Computing , STOC 2022 (Association for Computing Machinery, New York, NY, USA, 2022) pp. 375–388, arXiv:21...

    2022 arXiv

  24. [32]

    Panteleev and G

    P. Panteleev and G. Kalachev, Quantum LDPC Codes with Almost Linear Minimum Distance, IEEE Trans. In- form. Theory 68, 213 (2022), arXiv:2012.04068 [quant- ph]

    2022 arXiv

  25. [33]

    Dinur, M.-H

    I. Dinur, M.-H. Hsieh, T.-C. Lin, and T. Vidick, Good Quantum LDPC Codes with Linear Time Decoders (2022), arXiv:2206.07750 [quant-ph]

    2022 arXiv

  26. [34]

    Without further improvements, such as [45]

  27. [35]

    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, Nature 627, 778 (2024)

    2024

  28. [36]

    Gottesman, Fault-Tolerant Quantum Computation with Constant Overhead (2014), arXiv:1310.2984 [quant- ph]

    D. Gottesman, Fault-Tolerant Quantum Computation with Constant Overhead (2014), arXiv:1310.2984 [quant- ph]

    2014 arXiv

  29. [37]

    A. A. Kovalev and L. P. Pryadko, Fault-Tolerance of ”Bad” Quantum Low-Density Parity Check Codes, Phys. Rev. A 87, 020304 (2013), arXiv:1208.2317 [quant-ph]

    2013 arXiv

  30. [38]

    Fawzi, A

    O. Fawzi, A. Grospellier, and A. Leverrier, Constant overhead quantum fault-tolerance with quantum ex- pander codes, in 2018 IEEE 59th Annual Symposium on Foundations of Computer Science (FOCS) (2018) pp. 743–754, arXiv:1808.03821 [quant-ph]

    2018 arXiv

  31. [39]

    Burton and D

    S. Burton and D. Browne, Limitations on Transver- sal Gates for Hypergraph Product Codes, IEEE Transactions on Information Theory 68, 1772 (2022), 15 arXiv:2012.05842 [quant-ph]

    2022 arXiv

  32. [40]

    A. O. Quintavalle, P. Webster, and M. Vasmer, Parti- tioning qubits in hypergraph product codes to implement logical gates, Quantum 7, 1153 (2023)

    2023

  33. [41]

    Krishna and D

    A. Krishna and D. Poulin, Fault-Tolerant Gates on Hypergraph Product Codes, Phys. Rev. X 11, 011023 (2021)

    2021

  34. [42]

    Banfield and A

    D. Banfield and A. Kay, Implementing Logical Operators using Code Rewiring (2022), arXiv:2210.14074 [quant- ph]

    2022 arXiv

  35. [43]

    Leitch and A

    H. Leitch and A. Kay, Balanced Product Codes, Demys- tified (2025), arXiv:2505.13679 [quant-ph]

    2025 arXiv

  36. [44]

    M. A. Webster, A. O. Quintavalle, and S. D. Bartlett, Transversal Diagonal Logical Operators for Stabiliser Codes (2023), arXiv:2303.15615 [quant-ph]

    2023 arXiv

  37. [45]

    Koutsioumpas, D

    S. Koutsioumpas, D. Banfield, and A. Kay, The Small- est Code with Transversal T (2022), arXiv:2210.14066 [quant-ph]

    2022 arXiv

  38. [46]

    A. M. Steane, Simple quantum error-correcting codes, Phys. Rev. A 54, 4741 (1996)

    1996

  39. [47]

    Raussendorf and H

    R. Raussendorf and H. J. Briegel, A One-Way Quantum Computer, Phys. Rev. Lett. 86, 5188 (2001)

    2001

  40. [48]

    Sometimes the definition of A being expanding can be given as bothA andAT being expanding, in terms of the definition we have just given

  41. [49]

    Tanner, A recursive approach to low complexity codes, IEEE Transactions on Information Theory27, 533 (1981)

    R. Tanner, A recursive approach to low complexity codes, IEEE Transactions on Information Theory27, 533 (1981)

    1981

  42. [50]

    They could turn out just to be stabilizers and of no concern, but we will see this is not they case

    We have not, at this point, proven that these indeed cor- respond to logical operators of the quantum code, not yet having introduced a quantum code. They could turn out just to be stabilizers and of no concern, but we will see this is not they case

  43. [51]

    F. N. Cole, Kirkman parades, Bull. Amer. Math. Soc. 28, 435 (1922)

    1922

  44. [52]

    Kay and H

    A. Kay and H. Leitch, Supporting Code for Transver- sal Gates for Highly Asymmetric qLDPC Codes, https://doi.org/10.17637/rh.29313224 (2025)

    2025 doi

  45. [53]

    This can be quite restrictive

    Recall that we require a permutation of both the vertices and the edges with the same order, such that all orbits have the same length l. This can be quite restrictive

  46. [54]

    Lubotzky, R

    A. Lubotzky, R. Phillips, and P. Sarnak, Ramanujan graphs, Combinatorica 8, 261 (1988)

    1988

  47. [55]

    S. Evra, T. Kaufman, and G. Z´ emor, Decodable Quan- tum LDPC Codes beyond the √n Distance Barrier Us- ing High-Dimensional Expanders, SIAM J. Comput. , FOCS20 (2022)

    2022

  48. [56]

    There are many more logical operators, just not ones we choose to keep for the transversal gates. In particular, the Js in LZ may be replaced by a full set of null vectors of C0, which is how we see that d(C0) upper bounds the Z distance, whileJs inLX may be replaced by unit v...

  49. [57]

    It was proven in [32] that the smallest q-orthogonal code with distance at least 3 has s = 2q+1− 1 columns (such cases have r = q + 1 rows), and our conditions for transversality, Lemma 1, include the q-orthogonal con- ditions as a subset. There could be slightly larger codes ...

  50. [58]

    Yamasaki and M

    H. Yamasaki and M. Koashi, Time-Efficient Constant- Space-Overhead Fault-Tolerant Quantum Computation, Nat. Phys. 20, 247 (2024). Appendix A: Incidence Matrix for Kirkman’s Schoolgirl Problem I0 =   1 0 0 0 1 1 0 0 1 0 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 ...

    2024

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.