Pith. sign in

REVIEW 4 major objections 3 minor 36 references

Analog rotation gates can be made to scale as O(p φ^1.5) instead of O(p φ) by fusing magic states in a balanced tree before repeat-until-success teleportation.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 09:28 UTC pith:GFDXOCWN

load-bearing objection Balanced fusion is a real advance in analog rotation error scaling, but the headline O(pφ^1.5) is conditional on an idealized noise model the paper does not verify. the 4 major comments →

arxiv 2607.20756 v1 pith:GFDXOCWN submitted 2026-07-22 quant-ph

Suppressing errors in analog logical rotation gates via balanced fusion

Sam McArdle , Alexander M. Dalzell , Fernando G.S.L. Brand\~ao This is my paper
classification quant-ph MSC 81P6868Q12 PACS 03.67.Pp
keywords balanced fusionanalog logical rotationsrepeat-until-successmagic stateserror scalingcoherent errorsTrotter simulationMixing Lemma
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper proposes and analyzes 'balanced fusion,' a method for preparing the large-angle magic states consumed by repeat-until-success (RUS) implementations of logical R_z(φ) rotations. Its central claim is that this method improves the expected error of the rotation from O(p φ) to O(p φ^1.5), where p is the physical error rate, without fallback synthesis. The improvement matters because it would lift the O(1) cap on idealized Trotter simulation time: with enough Trotter steps, both Trotter error and rotation error can be suppressed. The key analytic step is showing that coherent errors in the magic states are decohered by the teleportation process, so that only the second moment of the rotation-angle error accumulates, and that a balanced fusion tree keeps that accumulation sub-linear in the final angle. If the scaling holds, small-angle rotations become a more credible alternative to T-state cultivation plus synthesis in early fault-tolerant devices.

Core claim

The central claim is that a depth-t balanced fusion tree—building |M(2^m φ)> from roughly 2^{ceil(m/2)} copies of |M(2^{floor(m/2)} φ)>, rather than from 4^m copies at the base angle—reduces the accumulated error of the RUS rotation to O(p φ^1.5). The argument uses the Mixing Lemma to bound the distance between a mixture of rotation channels and the mean rotation by the variance of the angles; because RUS teleportation destroys the coherent (first-moment) part of the error, the error budget is set by E[Φ²]=O(p θ²) for each fusion input. For a general target angle, S-gate resets keep every trial angle below π/4, and the paper bounds total error by (5√π/2)p φ^1.5 with expected resource cost O(

What carries the argument

Balanced fusion is the central mechanism: a binary teleportation tree that fuses two |M(θ)> states into |M(2θ)> with success probability 1/2, discarding |+> on failure, so on average four inputs produce one output. Using inputs at half the target angle rather than at the base angle balances error growth against resource cost, since after t successful fusions the number of independent angle errors grows as 2^t while the angle doubles as 2^t. The Mixing Lemma is the analytic engine: for a mixture of diagonal rotations, the diamond distance to the mean rotation is bounded by the variance of the angles, and this converts the infidelity/trace-distance mismatch into an additive second-moment budge

Load-bearing premise

The entire error scaling rests on the assumption that for every angle θ, the black-box magic-state preparation has angle-error first moment O(pθ) and second moment O(pθ²), and that after t fusions the 2^t angle errors are independent so their variances add; the paper notes this is not derived from a concrete fusion circuit and needs full noisy simulation.

What would settle it

Run a full noisy simulation of the actual fusion/RUS circuits and measure the distribution of the implemented rotation angle: if the variance of the total angle error after t fusions grows faster than 2^t (e.g., as 4^t), the O(p φ^1.5) bound is false. A second test is calibration sensitivity: set physical R_z overrotation λ=1+ξ and measure the infidelity of |M(φ)>; if it is dominated by O(d²ξ²φ²) rather than O(pφ²), then for small φ the claimed improvement disappears below a φ set by the calibration floor.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • In the idealized noise model, each R_z(φ) gate can be implemented with expected error O(p φ^1.5), an asymptotic improvement over the O(p φ) of earlier analog rotation proposals.
  • The Trotter simulation-time cap disappears: choosing r=Ω(T^3) Trotter steps makes the rotation-angle error scale as T^1.5/r^0.5, so both Trotter and logical errors can be driven arbitrarily low without bounded T.
  • Fallback to Clifford+T synthesis is not required for the improved scaling; where fallback is added, balanced fusion still gives a 3.8× error reduction at κ=6, φ=10^-4 and about 30× at κ=12, φ=10^-5 relative to baseline RUS with fallback.
  • Expected resource cost of the RUS process becomes O(log(1/φ)), rather than growing with the number of trials in a way that destroys the error improvement.
  • The trace distance and infidelity of the prepared magic state coincide to leading order, so error analysis can be done with the second moment of the angle error.

Where Pith is reading between the lines

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

  • The same balanced-tree construction should transfer to any probabilistic gate-teleportation cascade with commuting rotation errors; the argument only needs a mixture of rotations whose errors add in quadrature, so other early fault-tolerant gadgets may gain similar error suppression.
  • A useful next test is a full noisy simulation that tracks the joint distribution of angle errors through the fusion tree. If correlations among errors rather than independence develop, the O(p φ^1.5) promise should be revised.
  • Because overrotation of physical R_z gates by factor 1+ξ gives infidelity O(d²ξ²φ²) and trace distance O(dξφ), the practical regime of the method is bounded by calibration precision: at sufficiently small φ, the calibration floor—not p—will set the rotation error.
  • The O(p²φ) coherent residue suggests a hybrid regime: for tiny angles where φ≲p², the second-order term dominates, so a correction scheme that cancels the first moments without reintroducing a linear-in-φ error could extend the method further.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 3 minor

Summary. The manuscript introduces 'balanced fusion', a repeat-until-success (RUS) method for implementing analog logical rotation gates. In the baseline approach, the resource states |M(2^m φ)> are prepared directly, leading to an accumulated error O(pφ) after the RUS cascade. Balanced fusion instead builds each large-angle resource state from a shallow binary tree of smaller-angle states, balancing the error accumulation against the resource cost; the expected second-moment error of the fused state is claimed to scale as O(2^{1.5m} p φ²), which translates to an overall logical rotation error O(pφ^{1.5}) (with a subleading O(p²φ) term). The central technical tool is a Mixing-Lemma variance bound showing that, for mixtures of rotation channels, the diamond distance of the RUS channel is controlled by the second moment of the accumulated angle error. The paper also analyzes a fallback protocol to cultivation+synthesis and discusses implications for first-order Trotter simulation, where it claims the O(pφ^{1.5}) scaling removes the T=O(1) simulation-time cap in an idealized model.

Significance. If the claimed O(pφ^{1.5}) scaling survives realistic preparation noise, it is a meaningful improvement for early fault-tolerant circuits, and the variance-accumulation analysis in App. B is a clean and potentially reusable contribution. The paper is transparent about its idealizations: Sec. III explicitly lists full noisy simulation, calibration floors, and more realistic error models as open problems. The analytical derivation in App. B is internally coherent, and the numerical check in Fig. 3 supports the Mixing-Lemma bound for the base-state mixture. However, the central result is conditional on a black-box noise model whose key assumption—independent angle-error addition through fusions—is not derived from a concrete circuit. In addition, the advertised Trotter application contains an asymptotic error as written, and the calibration-floor discussion indicates a concrete linear-in-φ term that can dominate the claimed improvement. These issues are fixable but need to be addressed before the paper can be accepted.

major comments (4)
  1. [Sec. II.A] The claim that choosing r=Ω(T^3) makes the balanced-fusion logical error arbitrarily small is incorrect as stated. With φ∝T/r and r=Θ(T^3), the logical error scales as r φ^{1.5} = T^{1.5}/r^{0.5} = Θ(1). To make this vanish as T→∞ one needs r=ω(T^3), e.g. r=Θ(T^{3+ε}). Please correct the asymptotic statement (and the associated sentence 'we can arbitrarily suppress both Trotter and logical errors').
  2. [App. B, Eqs. (B30)-(B34)] The O(pφ^{1.5}) result rests entirely on the assumption that after t successful fusions the error channel is E^{2^t}, i.e. that the 2^t angle errors add with independent increments. No explicit fusion circuit is given; the sentence 'When one successfully fuses two states ρ, one arrives at E(E(|M(2θ)⟩⟨M(2θ)|))' is an assertion. The manuscript itself lists full noisy simulation as future work (Sec. III). Please either derive the fusion error model from a concrete circuit (e.g., the dephasing model of App. A), or explicitly state the main theorem as conditional on this assumption and adjust the abstract and title accordingly.
  3. [Sec. III, calibration floor] The paper acknowledges that overrotation R_z(x)→R_z(λx) with λ=1+ξ gives trace distance O(dξφ). Feeding this through the same variance-accumulation argument (App. B, Eqs. (B27)-(B33)) produces an O(dξφ) contribution to the final error, from the E[Φ]^2 term in Eq. (B33). For fixed ξ and sufficiently small φ the term dξφ dominates the advertised O(pφ^{1.5}), so this is not merely an open question: it is a concrete floor on the scaling unless dξ ≲ p φ^{1/2}. The paper should state the final scaling as O(pφ^{1.5}+dξφ), or otherwise delimit the parameter regime in which the improvement holds.
  4. [App. B, Eqs. (B30)-(B31)] The moment assumptions E[Φ]=O(pθ) and E[Φ²]=O(pθ²) are introduced without derivation from the App. A density matrix. App. A gives an infidelity O(p φ^{2(1-1/d)}) and a trace distance O(pφ), not these moments directly, and the first moment E[Φ] is never explicitly bounded. Since E[Φ] appears in the cross term of Eq. (B33) and contributes to the p²φ term in Eq. (B37), please derive both moment bounds from the App. A mixture, or state them as formal assumptions together with the precise parameter regime (including the App. A restriction φ < p d < φ^{2/d}).
minor comments (3)
  1. [Fig. 3 and App. C.2] The fitted exponent in Fig. 3 (labeled 1.90 for d=13) does not match the analytic exponent 2(1−1/d)≈1.846, nor the text 'O(dpφ^{1−1/d})' in App. C.2, which is itself missing a factor of 2. Please reconcile the label, the fit, and the analytic expression.
  2. [Sec. II.B, Eq. (4)] The comparison with cultivation+synthesis uses α≈1 without a sensitivity analysis. Since α is asserted to depend on the code and circuit, please give the crossover as a function of α or state explicitly which values of α can reverse the 15× advantage.
  3. [App. C.3, Eqs. (C21)-(C25)] The derivation of the off-diagonal terms in the twirled and untwirled circuits is terse; in particular, the appearance of Z_1 Z_2 in Eq. (C22) is easy to misread. A one-line explanation that the operators are supported on different registers would improve readability.

Circularity Check

0 steps flagged

No significant circularity: the O(p φ^{1.5}) scaling is derived from external magic-state error assumptions plus a new variance analysis, not from fitted data or self-citation.

full rationale

The central claim is derived in App. B from two ingredients: (i) the Mixing Lemma, an external mathematical result, and (ii) the assumed first/second moments E[Φ]=O(pθ), E[Φ²]=O(pθ²) for base analog-rotation magic states, quoted from Refs. [6,7], which are not the present authors' work. The balanced-fusion construction then computes E[Φ²_tot] by i.i.d. variance addition (Eqs. B32–B35) and combines it with the RUS trial probabilities (Eqs. 2–3) to obtain O(p φ^{1.5}). None of this reduces to fitting the target quantity; the only least-squares fit (Fig. 3) is an illustrative line through an analytically computed curve for the Mixing Lemma bound, and it does not set any constant in the paper's results. The paper's own Sec. III explicitly flags the idealized noise model and finite-R_z calibration as open questions requiring full noisy simulation, so the main result is conditional, but conditionality is not circularity. No load-bearing self-citation or definitional identification was found.

Axiom & Free-Parameter Ledger

3 free parameters · 6 axioms · 0 invented entities

The paper contributes no new free parameters in the asymptotic scaling: O(pφ^1.5) follows from combining the prior O(pθ²) magic-state error with a new variance-accumulation calculation. The real burden is the unverified fusion error-channel model (E^{2^t} with i.i.d. angle errors) and the idealized no-calibration-floor assumption. Constants α and κ only enter the numerical comparison, and the Fig. 3 fit is illustrative.

free parameters (3)
  • resource-state error prefactor α = ≈1 (taken from Ref. [7]'s d/30 estimate; not fitted here)
    Introduced in Eq. (4) for the fallback error comparison; the O(pφ^1.5) scaling claim is independent of its value.
  • fallback threshold κ = 6, 8, or 12 in the comparison examples
    User-chosen protocol parameter for the fallback variant; not part of the no-fallback scaling claim and not optimized in the paper.
  • Fig. 3 fit constants a,b = a≈1.38, b≈1.90 (least-squares to own analytical curve)
    Illustrative scaling fit for the Mixing-Lemma analysis; not used to set the constants in the central O(pφ^1.5) result.
axioms (6)
  • domain assumption Black-box preparation of |M(θ)> with E[Φ]=O(pθ), E[Φ²]=O(pθ²) for all θ; state error is a mixture of rotation channels.
    Central input imported from Refs. [6,7] (App. A); App. B §3 uses it as the starting point for fusion error.
  • ad hoc to paper Fusion teleportation succeeds with probability 1/2; on failure the state is discarded; a depth-t fusion tree has error channel E^{2^t} with i.i.d. angle errors.
    The heart of balanced fusion; asserted in Sec. I and App. B §3 without an explicit circuit-level derivation.
  • domain assumption RUS teleportation decoheres coherent errors, so total error is controlled by the variance (second moment) of angle errors; correction channels commute and the Mixing Lemma bound applies.
    Proved via the Mixing Lemma in App. B §1–2; relies on non-adaptive mixtures of rotations and noiseless Clifford corrections.
  • domain assumption Clifford gates, memory, and stabilizer measurements are error-free; only magic-state preparation errors and RUS teleportation errors contribute.
    Stated explicitly in Sec. II.B and App. A; the paper says memory/Clifford errors set an error floor ignored in the ideal analysis.
  • domain assumption No physical R_z calibration floor; overrotation λ=1+ξ is not part of the main result.
    Discussed as an open question in Sec. III; the authors show it would worsen infidelity to O(d²ξ²φ²) and trace distance to O(dξφ).
  • domain assumption CSS code distance d odd; d≡1 mod4 or a Clifford fix; tan φ = tan^d x and x can be small while φ small.
    From App. A, the transversal rotation construction; needed for the infidelity O(pφ^{2(1-1/d)}) and approximations ≈2.

pith-pipeline@v1.3.0-alltime-deepseek · 18183 in / 17459 out tokens · 140806 ms · 2026-08-01T09:28:33.894365+00:00 · methodology

0 comments
read the original abstract

There have been a number of recent proposals to use analog logical rotations in early fault-tolerant quantum algorithms. Existing proposals implement a logical rotation by angle $\phi$, with error $O(p\phi)$, where $p$ is the physical error rate. While this is not fault-tolerant, if $\phi$ is sufficiently small the logical error rate can be suppressed. In this work, we introduce and analyze the `balanced fusion' technique for improving the error scaling of analog logical rotations to $O(p \phi^{1.5})$, without having to resort to fallback synthesis. Balanced fusion is enabled by an improved analysis of the accumulation of coherent error terms in analog logical rotations. Our techniques improve the viability of analog logical rotations for small rotation angles as an alternative to cultivation-powered rotation synthesis.

Figures

Figures reproduced from arXiv: 2607.20756 by Alexander M. Dalzell, Fernando G.S.L. Brand\~ao, Sam McArdle.

Figure 1
Figure 1. Figure 1: FIG. 1. Schematic of the balanced fusion RUS process for implementing a single-qubit rotation [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Error accumulation and angle growth during the RUS process with [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Computing the error using the Mixing Lemma, Eq. ( [PITH_FULL_IMAGE:figures/full_fig_p016_3.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

36 extracted references · 11 linked inside Pith

  1. [1]

    Eastin and E

    B. Eastin and E. Knill, Restrictions on transversal en- coded quantum gate sets, Physical review letters102, 110502 (2009)

    2009

  2. [2]

    C. M. Dawson and M. A. Nielsen, The solovay-kitaev algorithm, arXiv preprint quant-ph/0505030 (2005)

    Pith/arXiv arXiv 2005

  3. [3]

    Litinski, Magic state distillation: Not as costly as you think, Quantum3, 205 (2019)

    D. Litinski, Magic state distillation: Not as costly as you think, Quantum3, 205 (2019)

    2019

  4. [4]

    Gidney, N

    C. Gidney, N. Shutty, and C. Jones, Magic state culti- vation: growing t states as cheap as cnot gates, arXiv preprint arXiv:2409.17595 (2024)

    Pith/arXiv arXiv 2024

  5. [5]

    N. J. Ross and P. Selinger, Optimal ancilla-free clifford+ t approximation of z-rotations., Quantum Inf. Comput. 16, 901 (2016)

    2016

  6. [6]

    H. Choi, F. T. Chong, D. Englund, and Y. Ding, Fault tolerant non-clifford state preparation for arbitrary rota- tions, arXiv preprint arXiv:2303.17380 (2023)

    Pith/arXiv arXiv 2023

  7. [7]

    Toshio, Y

    R. Toshio, Y. Akahoshi, J. Fujisaki, H. Oshima, S. Sato, and K. Fujii, Practical quantum advantage on partially fault-tolerant quantum computer, Phys. Rev. X15, 021057 (2025)

    2025

  8. [8]

    Zhang, T.-H

    J. Zhang, T.-H. Wei, X.-N. Zhuang, Z.-Y. Chen, W.- C. Kong, Y.-C. Wu, and G.-P. Guo, Low-overhead and high-fidelity preparation of logical non-clifford states with multi-level transversal injection, arXiv preprint arXiv:2509.23642 (2025)

    arXiv 2025

  9. [9]

    Huang, P.-G

    E. Huang, P.-G. Rozon, A. Dua, S. Gopalakrishnan, and M. J. Gullans, A robust phase of continuous transver- sal gates in quantum stabilizer codes, arXiv preprint arXiv:2510.01319 (2025)

    arXiv 2025

  10. [10]

    Yoshioka, A

    N. Yoshioka, A. Seif, A. Cross, and A. Javadi- Abhari, Transversal gates for probabilistic implemen- tation of multi-qubit pauli rotations, arXiv preprint arXiv:2510.08290 (2025)

    Pith/arXiv arXiv 2025

  11. [11]

    Toshio, S

    R. Toshio, S. Kanasugi, J. Fujisaki, H. Oshima, S. Sato, and K. Fujii, Star-magic mutation: Even more efficient analog rotation gates for early fault-tolerant quantum computer, arXiv preprint arXiv:2603.22891 (2026)

    arXiv 2026

  12. [12]

    Campbell, Shorter gate sequences for quantum com- puting by mixing unitaries, Phys

    E. Campbell, Shorter gate sequences for quantum com- puting by mixing unitaries, Phys. Rev. A95, 042306 (2017)

    2017

  13. [13]

    M. B. Hastings, Turning gate synthesis errors into inco- herent errors, arXiv preprint arXiv:1612.01011 (2016)

    Pith/arXiv arXiv 2016

  14. [14]

    Akahoshi, R

    Y. Akahoshi, R. Toshio, J. Fujisaki, H. Oshima, S. Sato, and K. Fujii, Compilation of trotter-based time evolution for partially fault-tolerant quantum computing architec- ture, arXiv preprint arXiv:2408.14929 (2024)

    Pith/arXiv arXiv 2024

  15. [15]

    Chung, A

    M.-Z. Chung, A. H. Kavaki, A. Scherer, A. Khalid, X. Kong, T. Kawakubo, N. Anand, G. A. Dagnew, Z. Webb, A. Silva,et al., Partially fault-tolerant quantum computation for megaquop applications, arXiv preprint arXiv:2603.13093 (2026)

    arXiv 2026

  16. [16]

    P. Zeng, G. Zheng, Q. Xu, and L. Jiang, Error-structure- tailored early fault-tolerant quantum computing, arXiv preprint arXiv:2511.19983 (2025)

    arXiv 2025

  17. [17]

    A. M. Childs, Y. Su, M. C. Tran, N. Wiebe, and S. Zhu, Theory of trotter error with commutator scaling, Physi- cal Review X11, 011020 (2021)

    2021

  18. [18]

    Kliuchnikov, K

    V. Kliuchnikov, K. Lauter, R. Minko, A. Paetznick, and C. Petit, Shorter quantum circuits via single-qubit gate approximation, Quantum7, 1208 (2023)

    2023

  19. [19]

    Kubica, A

    A. Kubica, A. Haim, Y. Vaknin, H. Levine, F. Brand˜ ao, and A. Retzker, Erasure qubits: Overcoming the t 1 limit in superconducting circuits, Physical Review X13, 041022 (2023)

    2023

  20. [20]

    Jacoby, Y

    S. Jacoby, Y. Vaknin, A. Retzker, and A. L. Grimsmo, Magic state injection with erasure qubits, PRX Quantum 6, 040323 (2025)

    2025

  21. [21]

    Vaknin, S

    Y. Vaknin, S. Jacoby, A. Grimsmo, and A. Retzker, Ef- ficient magic state cultivation on the surface code, arXiv preprint arXiv:2502.01743 (2025)

    arXiv 2025

  22. [22]

    Bothe, C

    M. Bothe, C. S¨ underhauf, M. J. Witham, E. T. Camp- bell, and N. S. Blunt, More efficient clifford+ t synthesis for small-angle rotations and application to trotteriza- tion, arXiv preprint arXiv:2605.31544 (2026)

    Pith/arXiv arXiv 2026

  23. [23]

    H. Zhao, A. Zlokapa, H. Neven, R. Babbush, J. Preskill, J. R. McClean, and H.-Y. Huang, Exponential quan- tum advantage in processing massive classical data, arXiv preprint arXiv:2604.07639 (2026)

    Pith/arXiv arXiv 2026

  24. [24]

    Ismail, I

    R. Ismail, I. Chen, C. Zhao, R. Weiss, F. Liu, H. Zhou, S.-T. Wang, A. Sornborger, and M. Kornjaˇ ca, Transver- sal star architecture for megaquop-scale quantum simula- tion with neutral atoms, arXiv preprint arXiv:2509.18294 (2025)

    Pith/arXiv arXiv 2025

  25. [25]

    Sethi, S

    S. Sethi, S. Khan, A. Awasthi, A. Anand, and J. M. Baker, Injeqt: Improved magic-state injection protocol for fault-tolerant quantum extractor architectures, arXiv preprint arXiv:2604.25094 (2026)

    Pith/arXiv arXiv 2026

  26. [26]

    T. J. Yoder, E. Schoute, P. Rall, E. Pritchett, J. M. Gam- betta, A. W. Cross, M. Carroll, and M. E. Beverland, Tour de gross: A modular quantum computer based on bivariate bicycle codes, arXiv preprint arXiv:2506.03094 (2025). 7 A. Overview of analog logical rotations We will discuss the approach for preparing analog logical rotation magic states introdu...

    Pith/arXiv arXiv 2025

  27. [27]

    [6] and the target magic state|M(ϕ)⟩, under an idealized noise model

    Calculation of the infidelity in the small logical angle limit In this appendix, we will compute the infidelity between the output density matrix of the analog rotations protocol of Ref. [6] and the target magic state|M(ϕ)⟩, under an idealized noise model. The calculation is straightforward if one makes the small physical rotations angles approximation (s...

  28. [28]

    1 +iδ+ Z δ 0 ds Z s 0 ds′(−eis′ ) # Taylor expansion with remainder (B4) = 2 iE[δ]|{z} 0 +E

    The Mixing Lemma As shown in Eq. (A8), the protocols for preparing analog logical rotation magic states output a mixture of magic states for rotations by different angles. Teleporting in a state of this form (for use in RUS, or balanced fusion), will implement a channel representing a mixture of rotations. We can use the Mixing Lemma [12, 13] to bound the...

  29. [29]

    Let the rotation channel by angleθbe denotedR z(θ) := Rz(θ)[·]Rz(θ)†

    RUS teleportation In RUS, ideally on trialiwe prepare the magic state|M(ϕ (i))⟩, whereϕ (1) =ϕand fori≥2 ϕ(i) = ( 2ϕ(i−1) ifϕ (i−1) < π/8 2ϕ(i−1) −π/4 otherwise (B9) that is, we reset so thatϕ (i) < π/4 always holds. Let the rotation channel by angleθbe denotedR z(θ) := Rz(θ)[·]Rz(θ)†. Using perfect resource states, the channel implemented on trialiof RUS...

  30. [30]

    In general, after depthtof successful fusions, one arrives at the stateE 2t (|M(2 tθ)⟩ ⟨M(2tθ)|)

    The second moment in balanced fusion For any angleθ, the analog rotations method allows the preparation of base statesρ=E(|M(θ)⟩ ⟨M(θ)|), where Eis a noise channel which is close to the identity, and equal to a mixture of rotation angles with random angle Φ satisfying E[Φ] =O(pθ) (B30) E[Φ2] =O(pθ 2) (B31) When one successfully fuses two statesρ, one arri...

  31. [31]

    A, and shows that the infidelity of the magic state scales asO(pϕ 2)

    A short summary of recent progress Analog logical rotations have been investigated in a number of recent publications, as we discuss below: •[6]: This paper introduces them= 1 logical analog rotations method described in App. A, and shows that the infidelity of the magic state scales asO(pϕ 2). •[7]: This work makes a number of contributions:

  32. [32]

    A, which has a higher success probability and better error scaling than them= 1 case

    Introduces them >1 logical analog rotations method described in App. A, which has a higher success probability and better error scaling than them= 1 case

  33. [33]

    Observes that while the infidelity of the magic state scales asO(pϕ 2), the trace distance scales asO(pϕ) due to coherent terms

  34. [34]

    C 2) to effectively cancel the coherent error

    Uses the probabilistic rotations (see App. C 2) to effectively cancel the coherent error. However, due to error accumulation in the RUS process (Eq. (1) in our main text), the overall error in the gate is still O(pϕ). •[14]: This paper applies the methods of Ref. [7] to detailed resource estimates for QPE. •[24]: This paper applies the analog logical rota...

  35. [35]

    [7], a probabilistic rotation cancellation scheme was developed

    Probabilistic rotations In Ref. [7], a probabilistic rotation cancellation scheme was developed. It is observed that to leading order, the noise channel from successfully teleporting the state in Eq. (A8) is Eϕ(ρ) = (1−q 1)ρ+q 1Rϕ1−ϕ(ρ),(C1) whereR θ is the channel representation of the rotationR z(θ). Ref. [7] then defines a corrective channel Cϕ(ρ) = (1...

  36. [36]

    [8] discussed a twirled version of the RUS teleportation circuit, designed to eliminate the coherent term in the density matrix

    A revised analysis of twirling, and decoherence by RUS teleportation Ref. [8] discussed a twirled version of the RUS teleportation circuit, designed to eliminate the coherent term in the density matrix. However, we show through more careful analysis that it is the teleportation process, rather than the 16 10-6 10-5 10-4 10-3 10-2 Target logical angle φ 10...