Pith. sign in

REVIEW 3 major objections 5 minor 62 references

Universal Quantum Error Mitigation via Random Inverse Depolarizing Approximation

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read The paper introduces RIDA, which estimates a circuit's global depolarization probability with a random inverse subcircuit and divides the noisy expectation value by one minus that probability.

desk verdict RIDA is a simple, well-benchmarked error-mitigation heuristic whose universality claim rests on a first-order full-depolarization assumption that the paper's own local-noise model contradicts. read the letter →

arxiv 2508.17513 v1 pith:RZFCDBNI submitted 2025-08-24 quant-ph

Alexander X. Miller , Micheline B. Soley This is my paper · ORCID
classification quant-ph
keywords quantumerrormitigationdepolarizingnoisemodelrandominversecircuitsexpectationvalueestimationzero-noiseextrapolationreadoutNISQrandomizedbenchmarking
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

This paper introduces RIDA, an error mitigation method that estimates the total depolarization probability of a target quantum circuit without knowing its ideal output. The estimate comes from a single randomly built identity circuit: half of the target's one- and two-qubit gates, balanced in type, followed by their inverses, so the noise-free expectation value is known to be one. The noisy expectation of that estimation circuit directly gives the depolarization probability, and dividing the target's noisy expectation value by one minus that probability yields an approximate error-free value. Across simulated incoherent and coherent noise, on circuits of four to seven qubits over a range of error rates and shot counts, the paper reports that RIDA has lower root-mean-square error than benchmark combinations of exponential zero-noise extrapolation with readout twirling and CNOT-only depolarization with quadratic extrapolation.

What carries the argument

The load-bearing object is the RIDA estimation circuit: from the target circuit, randomly select exactly half of the one-qubit gates and half of the two-qubit gates, excluding gates on terminal qubits that cannot affect the measured expectation value, then append the inverse of that selected half. The whole estimation circuit is the identity in the absence of noise, so its error-free expectation value is exactly $1$; measuring it under noise gives $p_0 \approx 1 - \langle O_0^{\mathrm{noisy}}\rangle$, which under the global depolarizing approximation estimates the target's $p$. Because each gate has a 50% chance of appearing once in each half, the expected composition matches the target, and fixing the count at half minimizes estimator variance. The corrected expectation is $\langle O \rangle = \langle O_{\mathrm{noisy}}\rangle / (1-p_0)$.

What would settle it

Construct a two-qubit target whose ideal expectation is known, and engineer two gate errors that partially cancel inside the estimation circuit, for example an over-rotation on a selected gate followed by an under-rotation of equal magnitude on its inverse. Compare the RIDA estimate $\hat p$ with the true depolarization probability obtained by fitting the noisy and ideal expectation values at high shot count; a disagreement larger than shot noise would show that the independence-and-irreversibility assumption is violated and the $1/(1-\hat p)$ correction is biased.

Watch

Extended reading notes

Core claim

RIDA's central claim is that the global depolarization probability of a target circuit can be extracted from a random identity circuit with the same gate composition, and that this estimate is accurate enough to correct expectation values by the depolarizing-model formula $\langle O \rangle = \langle O_{\mathrm{noisy}}\rangle/(1-p)$. The paper argues that selecting exactly half of the one-qubit gates and half of the two-qubit gates minimizes the variance of the estimate and makes the average estimation-circuit depolarization probability equal to the target's under a first-order independent-error model. The method is universal in the sense that it applies to any expectation-value-estimating circuit, requires no knowledge of individual error rates, and intrinsically handles measurement error; with twirled readout its measurement-error mitigation is shown to be mathematically equivalent to TREX. In numerical tests it outperforms exponential ZNE plus TREX and CNOT-only depolarization plus quadratic ZNE across all considered circuit sizes, error multipliers, shot numbers, and both incoherent and coherent error models.

Load-bearing premise

The estimate is unbiased only if gate errors are independent, irreversible, and small enough that the estimation circuit's depolarization probability is approximated by twice the sum of the individual gate error rates; under correlated, partially reversible, or large errors, the half-and-half selection is no longer guaranteed to match the target circuit's depolarization probability.

Editorial extensions

If this is right

  • One estimation circuit, reused across a class of similar target circuits, provides both gate-error and measurement-error mitigation without a separate readout calibration step.
  • RIDA's sampling overhead matches the optimal unbiased-estimator scaling; under the paper's assumptions this is a cubic improvement over exponential ZNE and a quintic improvement over CNOT-only depolarization plus quadratic ZNE.
  • In the high-error regime where other methods break down, RIDA remains the only considered method with practically useful RMSE, and the analytic scaling indicates it needs far fewer shots than exponential ZNE to beat the unmitigated result.
  • Because the corrected error is low across the full range of error-free expectation values, RIDA is suited to estimating arbitrary Pauli-string observables, not only near-extremal ones.

Reading between the lines

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

  • A complementary-half test could expose the method's key assumption: generating a second estimation circuit from the unselected half of the gates should give the same $p$ up to shot noise; disagreement would signal correlated or reversible errors that the first-order model ignores.
  • Because the estimation circuit's identity structure does not depend on rotation angles, RIDA may serve as a cheap per-layer noise probe for parameterized circuits, giving a depolarization estimate that follows the circuit's gate skeleton rather than its specific parameters; the paper's reuse of one estimation circuit for a class of circuits points in this direction but does not prove it on general
  • If the TREX equivalence carries over to non-depolarizing measurement noise, RIDA could replace dedicated readout calibration inside existing ZNE pipelines, lowering overhead by one calibration stage; the paper proves the equivalence only under the depolarizing model with twirled readout.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes RIDA (Random Inverse Depolarizing Approximation), an error mitigation method built on the global depolarizing channel model. For a target circuit with observable O, RIDA constructs a 'depolarization estimation circuit' by randomly selecting half of the target's one-qubit gates and half of its two-qubit gates, followed by the inverses of the selected gates, so that the estimation circuit implements the identity and has a known noiseless expectation <O_0> = 1. Measurement of this circuit gives p_0 = 1 - <O_0_epsilon>, an estimate of the depolarization probability, and the target expectation value is corrected via <O> approximately <O_epsilon>/(1-p_0). The authors report numerical simulations on 4- to 7-qubit EfficientSU2 circuits under an IBM-Kingston-inspired local depolarizing noise model and a twirled coherent noise model, and claim that RIDA outperforms exponential ZNE + TREX and CNOT-only depolarization + quadratic ZNE across all tested error multipliers, shot numbers, and observable expectation values, with 16- to 47-fold reductions in shot/gate overhead relative to benchmarks and an analytical overhead that is cubic/quintic better than the benchmarks. The Supplemental Material provides proofs of the optimality of the 50/50 gate-selection rule, an equivalence between RIDA and TREX for measurement error, analytic overhead scalings, and extended numerical results.

Significance. If the claims were fully established, RIDA would be a significant practical tool: the method is simple, requires no noise characterization, uses a reusable estimation circuit, and the numerical evidence is extensive and favorable. The authors make the code publicly available and state precise, falsifiable predictions (depolarization-probability accuracy, RMSE as a function of shot number, improvement thresholds), which are genuine strengths. However, the load-bearing theoretical claims — universality over observables, unbiasedness and optimality of the 50/50 selection, and optimal shot-overhead scaling — are derived under an idealized per-gate global depolarizing model and are not reconciled with the local depolarizing noise model actually simulated. In particular, the estimation-circuit observable is left unspecified for target observables that are not Z-type Pauli strings, which affects the validity of Eq. (4) for a large fraction of the paper's own numerical tests. These are central-claim issues, not presentation points.

major comments (3)
  1. [Method, Eq. (4); SM I.C; SM II.A] The estimation circuit implements the identity on |0...0>, so the noiseless expectation <O_0> used in Eqs. (3)-(4) equals 1 only if O_0 is a Pauli string in the +1 stabilizer of |0...0>, i.e., a product of Z operators. For a weight-1 X or Y Pauli string, <O_0> = 0; under the depolarizing model Eq. (2) the noisy expectation is then identically zero, Eq. (4) gives p_0 = 1, and the correction factor 1/(1-p_0) is undefined. The manuscript never specifies how O_0 is realized for such target observables (for example, a measurement-basis rotation on the estimation circuit, or a different O_0 whose estimate is argued to be observable-independent), and this is in direct tension with the claim that RIDA estimation circuits 'involve the same measurements as the target circuit.' The numerical tests include random weight-1 Pauli strings (SM II.A), so as written either the implemented procedure differs from the described one or the method fails on two thirds of the test observables. The universality claim therefore requires a resolution, and the TREX-equivalence argument (SM I.C, Eqs. (29)-(38)), which presupposes +/-1 measurement outcomes on the identity circuit, must be reconciled with that resolution as well.
  2. [SM I.A, Eqs. (1)-(2) and (22)-(28)] The derivation p_0 = 1 - product(1-epsilon_i)^2 approximately 2*sum(epsilon_i) treats every gate error as contributing weight 1 to the measured depolarization probability, independent of the gate's location and of the measured observable. This is self-consistent only under a per-gate global depolarizing model. Under the paper's own simulation noise model — 'local, uncorrelated depolarizing noise channels' (SM II.B) — the first-order contribution of gate i to the measured p_0 is 2*epsilon_i*c_i, where c_i depends on whether the Heisenberg-evolved observable has non-identity support on gate i's qubits at that circuit position; the target circuit's effective polarization is correspondingly p = sum(epsilon_i*d_i) with generally different weights d_i. Equations (27)-(28) equalize only the unweighted numbers of one- and two-qubit gates, so they do not imply E[p_0] = E[p]; this requires a relation between the c_i and d_i that is not shown and is false for low-weight observables. Consequently, the unbiasedness of RIDA and the optimality of the 50/50 selection rule are not established under the noise model used in the numerics, and the same assumption propagates into the overhead (SM III.C-E) and improvement-threshold (SM III.B) derivations. The authors should re-derive E[p_0] under local depolarizing noise, e.g., using the scrambling of the propagated observable in random circuits, or explicitly state the global-depolarizing-model assumption and delimit its domain of validity.
  3. [SM III.C, Eqs. (101)-(103)] The sampling-overhead analysis and the claim that RIDA's shot overhead 'coincides with the optimal result for an unbiased estimator' treat p as a known constant. In the method, p is replaced by p_0, which is itself measured from estimation circuits with finite shot counts (10^7 total shots distributed over 50 circuits in the numerics, SM II.A). By the delta method, Var(<O_epsilon>/(1-p_0)) is approximately Var(<O_epsilon>)/(1-p)^2 + <O_epsilon>^2*Var(p_0)/(1-p)^4; the second term, omitted from Eqs. (101)-(103), is comparable to the first when the estimation and target shot budgets are equal (as in SM II.A). The optimal-scaling claim therefore holds only in the limit of negligible estimation noise; the analysis should include this term and quantify the estimation shot budget required to reach the quoted optimum.
minor comments (5)
  1. [Title] The title as displayed contains a typo ('Depolarizi ng'); please ensure the submitted file is clean.
  2. [SM II.A] Please clarify the shot accounting in the comparisons: the text says estimation circuits use 10^7 total shots and target circuits use 'the same number of total shots,' but it is not explicit whether a method's total budget includes its estimation/calibration circuits; this matters for the claim of lower baseline shot overhead.
  3. [SM I.C] The TREX-equivalence proof is written for weight-1 Pauli strings with +/-1 outcomes; the extension to higher-weight strings and to linear combinations follows by linearity but should be stated explicitly.
  4. [Fig. 2] The caption of Fig. 2(e) refers to a 'fixed shot number' without giving its value; please state the shot number in the caption.
  5. [SM II.F] The exponential ZNE implementation relies on heuristic fallbacks (Eqs. (73)-(76)) when the three-point data are non-monotonic; because these fallbacks affect the benchmark exactly in the high-error regime where RIDA's reported advantage is largest, the authors should justify or cite them.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: RIDA's depolarization probability is measured from a separate identity circuit, not fitted to or defined by the target expectation value.

full rationale

The derivation chain is self-contained and does not reduce to its own inputs. The depolarization probability p0 is obtained by measuring a separately constructed identity estimation circuit with known error-free expectation value hO0i = 1, and the target correction hOi = hO_noisyi/(1 - p0) is then applied through the global depolarizing model. There is no fitting of p to the target data, no parameter renamed as a prediction, and no load-bearing self-citation chain: the only self-references are to the public GitHub code repository, which is not part of the analytic argument. The statistical optimality proof in SM Sec. I A derives the half-gate selection rule from stated assumptions about independent, irreversible gate errors, rather than tuning to the target circuit. The main identified weakness is that the first-order formula p0 ≈ 2 sum epsilon_i may fail under the paper's own local depolarizing simulation model, because whether a given gate error contributes to p0 can depend on commutation with the measured observable and on gate order. That is a model-misspecification and correctness risk, not a circular reduction: the argument's assumptions are explicit and externally checkable, and the numerical benchmarks use an independent IBM Kingston-like local depolarizing simulator. Accordingly, no specific circular step can be quoted or exhibited.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No continuous free parameters are fitted to data: the depolarization probability is measured from a separate identity circuit, and the 1/2 selection fraction is derived from a first-order variance argument rather than tuned on the target data. The method nevertheless rests on the global depolarizing approximation, a statistical model of independent gate errors, and an implicit assumption about the measured estimation-circuit observable, all listed as axioms.

assumptions (5)
  • domain assumption The noise on any circuit is a global depolarizing channel (1-p)rho + p I/2^n with a single depolarization probability p.
    Used in Eq. (1) and everywhere in the correction formula Eq. (3); if the real noise is far from this model, the correction factor is wrong.
  • domain assumption Depolarization probability of a circuit is well approximated by the sum of individual gate error probabilities, with independent, irreversible single-gate errors; in an estimation circuit each selected gate's error is counted twice.
    SM Sec. I A Eqs. (1)-(4); this is the basis of the half-gate optimality proof and of the claim E[p0]=p.
  • domain assumption For every target circuit there is an estimation-circuit observable O0 with trace zero and noiseless expectation <O0>=1, so that p0 = 1 - <O0_noisy>.
    Eq. (4) in the main text; the paper asserts this for circuits that are a circuit followed by its inverse but does not specify how it holds for arbitrary target Pauli observables.
  • domain assumption Noise statistics are stationary while target and estimation circuits are run, so a p0 measured on one estimation circuit transfers to the target circuit class.
    Main text: results can be reused for similarly constructed circuits 'with the exception of periodic updates to capture non-stationary noise fluctuations'.
  • domain assumption Omitting terminal-qubit gates and appending the companion circuit does not change the noiseless identity property of the estimation circuit.
    SM Sec. I B argues this via decomposition into controlled gates; it is necessary for p0 to retain its interpretation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Universal Quantum Error Mitigation via Random Inverse Depolarizing Approximation." pith.science (2026). https://pith.science/paper/RZFCDBNI

@misc{pith2026250817513,
  author       = {Pith},
  title        = {Pith review of: Universal Quantum Error Mitigation via Random Inverse Depolarizing Approximation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RZFCDBNI}},
  note         = {Machine review of arXiv:2508.17513}
}
read the original abstract

Given the severity of noise in near-term quantum computing, error mitigation is essential to reduce error in quantum-computer-generated expectation values. We introduce RIDA (Random Inverse Depolarizing Approximation), a simple universal method that harnesses randomly generated circuits to estimate a given circuit's global depolarization probability and corresponding error-free expectation value. Numerical tests indicate RIDA outperforms key benchmarks, suggestive of significant accuracy improvements for applications of quantum computing across fields including physics and chemistry.

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

62 extracted references · 40 canonical work pages

  1. [1]

    L. K. Grover, in Proceedings of the Twenty-Eighth Annual ACM Symposium on Theory of Computing , STOC ’96 (Association for Computing Machinery, New York, NY, USA, 1996) p. 212{219

    work page 1996

  2. [2]

    P. W. Shor, SIAM J. Comput. 26, 1484 (1997)

    1997

  3. [3]

    Katabarwa, K

    A. Katabarwa, K. Gratsea, A. Caesura, and P. D. John- son, PRX Quantum 5, 020101 (2024)

    2024

  4. [4]

    E. T. Campbell, B. M. Terhal, and C. Vuillot, Nature 549, 172 (2017)

    2017

  5. [5]

    Harper and S

    R. Harper and S. T. Flammia, Phys. Rev. Lett. 122, 080504 (2019)

    2019

  6. [6]

    Gottesman, arXiv:2210.15844 [quant-ph] (2022)

    D. Gottesman, arXiv:2210.15844 [quant-ph] (2022)

    arXiv 2022

  7. [7]

    Farhi, J

    E. Farhi, J. Goldstone, and S. Gutmann, arXiv:1411.4028 [quant-ph] (2014)

    arXiv 2014

  8. [8]

    Peruzzo, J

    A. Peruzzo, J. McClean, P. Shadbolt, M.-H. Yung, X.-Q. Zhou, P. J. Love, A. Aspuru-Guzik, and J. L. O’Brien, Nat. Commun. 5, 4213 (2014)

    2014

Show all 62 references
  1. [9]

    Bharti, A

    K. Bharti, A. Cervera-Lierta, T. H. Kyaw, T. Haug, S. Alperin-Lea, A. Anand, M. Degroote, H. Heimonen, J. S. Kottmann, T. Menke, et al. , Rev. Mod. Phys. 94, 015004 (2022)

    2022

  2. [10]

    Tilly, H

    J. Tilly, H. Chen, S. Cao, D. Picozzi, K. Setia, Y. Li, E. Grant, L. Wossnig, I. Rungger, G. H. Booth, and J. Tennyson, Phys. Rep. 986, 1{128 (2022)

    2022

  3. [11]

    Blekos, D

    K. Blekos, D. Brand, A. Ceschini, C.-H. Chou, R.-H. Li, K. Pandya, and A. Summer, Phys. Rep. 1068, 1{66 (2024)

    2024

  4. [12]

    Z. Cai, R. Babbush, S. C. Benjamin, S. Endo, W. J. Huggins, Y. Li, J. R. McClean, and T. E. O’Brien, Rev. Mod. Phys. 95, 045005 (2023)

    2023

  5. [13]

    Zimbor as, B

    Z. Zimbor as, B. Koczor, Z. Holmes, E.-M. Borrelli, A. Gily en, H.-Y. Huang, Z. Cai, A. Ac n, L. Aolita, L. Banchi, et al. , arXiv:2501.05694 [quant-ph] (2025)

    2025 arXiv

  6. [14]

    P. D. Nation and M. Treinish, PRX Quantum 4, 010327 (2023)

    2023

  7. [15]

    Z. Hu, Y. Lin, Q. Guan, and W. Jiang, in 2023 60th ACM/IEEE Design Automation Conference (DAC) (2023) pp. 1{6

    2023

  8. [16]

    Blume-Kohout, K

    R. Blume-Kohout, K. Rudinger, E. Nielsen, T. Proctor, and K. Young, arXiv:2012.12231 [quant-ph] (2020)

    2020 arXiv

  9. [17]

    Temme, S

    K. Temme, S. Bravyi, and J. M. Gambetta, Phys. Rev. Lett. 119, 180509 (2017)

    2017

  10. [18]

    S. Endo, S. C. Benjamin, and Y. Li, Phys. Rev. X 8, 031027 (2018)

    2018

  11. [19]

    Kandala, K

    A. Kandala, K. Temme, A. D. C orcoles, A. Mezzacapo, J. M. Chow, and J. M. Gambetta, Nature 567, 491 (2019)

    2019

  12. [20]

    Giurgica-Tiron, Y

    T. Giurgica-Tiron, Y. Hindy, R. LaRose, A. Mari, and W. J. Zeng, in 2020 IEEE International Conference on Quantum Computing and Engineering (QCE) (IEEE,

    2020

  13. [22]

    van den Berg, Z

    E. van den Berg, Z. K. Minev, A. Kandala, and K. Temme, Nat. Phys. 19, 1116 (2023)

    2023

  14. [23]

    R. S. Gupta, E. van den Berg, M. Takita, D. Rist e, K. Temme, and A. Kandala, Phys. Rev. A 109, 062617 (2024)

    2024

  15. [24]

    Ma and M

    Y. Ma and M. S. Kim, Phys. Rev. A 109, 012431 (2024)

    2024

  16. [25]

    M. A. Nielsen and I. L. Chuang, Quantum Computa- tion and Quantum Information: 10th Anniversary Edi- tion (Cambridge University Press, 2010)

    2010

  17. [26]

    A. M. Dalzell, N. Hunter-Jones, and F. G. S. L. Brand~ ao, PRX Quantum 3, 010333 (2022)

    2022

  18. [27]

    Deshpande, P

    A. Deshpande, P. Niroula, O. Shtanko, A. V. Gorshkov, B. Fe erman, and M. J. Gullans, PRX Quantum 3, 040329 (2022)

    2022

  19. [28]

    A. M. Dalzell, N. Hunter-Jones, and F. G. S. L. Brand~ ao, arXiv:2111.14907 [quant-ph] (2021)

    2021 arXiv

  20. [29]

    Urbanek, B

    M. Urbanek, B. Nachman, V. R. Pascuzzi, A. He, C. W. Bauer, and W. A. de Jong, Phys. Rev. Lett. 127, 270502 (2021)

    2021

  21. [30]

    Czarnik, A

    P. Czarnik, A. Arrasmith, P. J. Coles, and L. Cincio, Quantum 5, 592 (2021)

    2021

  22. [31]

    Vovrosh, K

    J. Vovrosh, K. E. Khosla, S. Greenaway, C. Self, M. S. Kim, and J. Knolle, Phys. Rev. E 104, 035309 (2021)

    2021

  23. [32]

    S. Choi, T. A. Chowdhury, and K. Yu, arXiv:2506.20125 [quant-ph] (2025)

    2025 arXiv

  24. [33]

    Knill, D

    E. Knill, D. Leibfried, R. Reichle, J. Britton, R. B. Blakestad, J. D. Jost, C. Langer, R. Ozeri, S. Seidelin, and D. J. Wineland, Phys. Rev. A 77, 012307 (2008)

    2008

  25. [34]

    Arute, K

    F. Arute, K. Arya, R. Babbush, D. Bacon, J. C. Bardin, R. Barends, R. Biswas, S. Boixo, F. G. S. L. Brandao, D. A. Buell, et al. , Nature 574, 505 (2019)

    2019

  26. [35]

    J. J. Wallman and S. T. Flammia, New J. Phys. 16, 103032 (2014)

    2014

  27. [36]

    Harper, I

    R. Harper, I. Hincks, C. Ferrie, S. T. Flammia, and J. J. Wallman, Phys. Rev. A 99, 052350 (2019)

    2019

  28. [37]

    Mills, A

    J. Mills, A. Sireesh, D. Leichtle, J. Ro e, and E. Kashe , arXiv:2508.05523 [quant-ph] (2025)

    2025 arXiv

  29. [38]

    The error-free expectation value for the target circuit then immediately follows from the relationship Eq

    indicate RIDA inherently includes measurement er- ror mitigation that is both procedurally and mathemat- ically equivalent to that of state-of-the-art TREX [21] when accompanied by measurement twirling, such that RIDA requires no further measurement error mitigation. The error...

  30. [39]

    See Supplemental Material for detailed methods and re- sults, which includes Ref. [53]

  31. [40]

    A. X. Miller and M. B. Soley, Random Inverse De- polarizing Approximation, https://github.com/ Alexander-Miller0/RIDA (2025)

    2025

  32. [41]

    Tsubouchi, T

    K. Tsubouchi, T. Sagawa, and N. Yoshioka, Phys. Rev. Lett. 131, 210601 (2023)

    2023

  33. [42]

    S. J. Devitt, W. J. Munro, and K. Nemoto, Rep. Prog. Phys. 76, 076001 (2013)

    2013

  34. [43]

    B. M. Terhal, Rev. Mod. Phys. 87, 307 (2015)

    2015

  35. [44]

    Stricker, D

    R. Stricker, D. Vodola, A. Erhard, L. Postler, M. Meth, M. Ringbauer, P. Schindler, T. Monz, M. M uller, and 6 R. Blatt, Nature 585, 207 (2020)

    2020

  36. [45]

    S. A. Moses, C. H. Baldwin, M. S. Allman, R. Ancona, L. Ascarrunz, C. Barnes, J. Bartolotta, B. Bjork, P. Blan- chard, M. Bohn, et al. , Phys. Rev. X 13, 041052 (2023)

    2023

  37. [46]

    Ghosh, A

    J. Ghosh, A. G. Fowler, J. M. Martinis, and M. R. Geller, Phys. Rev. A 88, 062329 (2013)

    2013

  38. [47]

    J. D. Teoh, P. Winkel, H. K. Babla, B. J. Chapman, J. Claes, S. J. de Graaf, J. W. Garmon, W. D. Kalfus, Y. Lu, A. Maiti, et al. , Proc. Natl. Acad. Sci. 120, e2221736120 (2023)

    2023

  39. [48]

    J. D. Teoh, P. Winkel, H. K. Babla, B. J. Chapman, J. Claes, S. J. de Graaf, J. W. Garmon, W. D. Kalfus, Y. Lu, A. Maiti, et al. , Proc. Natl. Acad. Sci. 120, e2320127120 (2023)

    2023

  40. [49]

    K. S. Chou, T. Shemma, H. McCarrick, T.-C. Chien, J. D. Teoh, P. Winkel, A. Anderson, J. Chen, J. C. Cur- tis, S. J. de Graaf, et al. , Nat. Phys. 20, 1454 (2024)

    2024

  41. [50]

    Mehta, J

    N. Mehta, J. D. Teoh, T. Noh, A. Agrawal, A. Ander- son, B. Birdsall, A. Brahmbhatt, W. Byrd, M. Cacioppo, A. Cabrera, et al. , arXiv:2503.10935 [quant-ph] (2025)

    2025 arXiv

  42. [51]

    A. G. Fowler, Phys. Rev. A 88, 042308 (2013)

    2013

  43. [52]

    Suchara, A

    M. Suchara, A. W. Cross, and J. M. Gambetta, in 2015 IEEE International Symposium on Information Theory (ISIT) (IEEE, 2015) pp. 1119{1123

    2015

  44. [53]

    N. C. Brown and K. R. Brown, Phys. Rev. A 100, 032325 (2019)

    2019

  45. [54]

    J. J. Wallman and J. Emerson, Phys. Rev. A 94, 052325 (2016). Supplemental Material: Universal Quantum Error Mitigation via Rando m Inverse Depolarizing Approximation Alexander X. Miller and Micheline B. Soley 1, 2, 3 1���������� �� ���������� ���������� �� ������������������ ...

    2016

  46. [55]

    (52) and depolarizing model Eq

    The exponential extrapolation function Eq. (52) and depolarizing model Eq. (40) constitute accurate represen- tations of the error

  47. [56]

    The shot error is su ciently low such that the error-free estimat ion value � 0 can be approximated linearly as a function of � 1, � 3, and � 5. Use of these assumptions is substantiated by numerical simulations giv en the agreement between the numerical mean squared error and...

    2000

  48. [57]

    van den Berg, Z

    E. van den Berg, Z. K. Minev, and K. Temme, Phys. Rev. A 105, 032620 (2022)

    2022

  49. [58]

    Urbanek, B

    M. Urbanek, B. Nachman, V. R. Pascuzzi, A. He, C. W. Bauer, an d W. A. de Jong, Phys. Rev. Lett. 127, 270502 (2021)

    2021

  50. [59]

    Nachman, M

    B. Nachman, M. Urbanek, W. A. de Jong, and C. W. Bauer, npj Qua ntum Inf. 6, 84 (2020)

    2020

  51. [60]

    J. J. Wallman and J. Emerson, Phys. Rev. A 94, 052325 (2016)

    2016

  52. [61]

    Knill, D

    E. Knill, D. Leibfried, R. Reichle, J. Britton, R. B. Blakesta d, J. D. Jost, C. Langer, R. Ozeri, S. Seidelin, and D. J. Wineland, Phys. Rev. A 77, 012307 (2008)

    2008

  53. [62]

    Tsubouchi, T

    K. Tsubouchi, T. Sagawa, and N. Yoshioka, Phys. Rev. Lett . 131, 210601 (2023)

    2023

  54. [63]

    A. X. Miller and M. B. Soley, Random Inverse Depolarizing App roximation, https://github.com/Alexander-Miller0/RIDA (2025). 8 It is interesting to note that the unmitigated, exponential ZNE + TREX, and CNOT-only depolarization + quadratic ZNE res ults exhibit to di ering exten...

    2025

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.