Optimal Calibration of Qubit Detuning and Crosstalk

Assaf Hamo; David Shnaiderov; Emanuele G. Dalla Torre; Eugene Demler; Matan Ben Dov; Yoav Woldiger

REVIEW 3 major objections 5 minor 25 references

Optimal Calibration of Qubit Detuning and Crosstalk

T0 review · 3 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 By measuring both X and Y quadratures of a Ramsey signal at a single time t = 1/γ, qubit detuning and dephasing can be estimated at the Cramér–Rao bound with up to 50% fewer shots than standard multi-time sweeps.

desk verdict A genuinely useful two-quadrature single-time Ramsey protocol and a clever crosstalk-decoupling scheme, but the exact optimal-time claim and the 50% shot-saving figure rest on an unexamined constant-variance approximation. read the letter →

arxiv 2507.10661 v2 pith:VUHWKIXU submitted 2025-07-14 quant-ph cond-mat.othercond-mat.stat-mechmath-phmath.MPphysics.data-an

David Shnaiderov , Matan Ben Dov , Yoav Woldiger , Assaf Hamo , Eugene Demler , Emanuele G. Dalla Torre This is my paper · ORCID

classification quant-phcond-mat.othercond-mat.stat-mechmath-phmath.MPphysics.data-an

keywords qubitcalibrationRamseyinterferometryFisherinformationCramér–RaobounddephasingcrosstalksuperconductingqubitsNVcenter

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 tries to establish that the standard way of calibrating qubit detunings — sweeping many Ramsey times and fitting a damped cosine — is wasteful. Using Fisher information and the Cramér–Rao bound, the authors show that for a single qubit the optimal strategy is to measure both quadratures X and Y at one time t = 1/γ, which yields the same precision with up to 50% fewer shots and is robust to parameter drift. For coupled qubits, they show that four strategically prepared Ramsey experiments reduce crosstalk calibration to independent single-qubit problems, with error essentially independent of the number of qubits. If right, this gives quantum processor calibration a concrete resource reduction and a scalable protocol for static ZZ crosstalk.

What carries the argument

The central object is the Fisher information matrix of the Ramsey measurement in the Gaussian limit (Eq. 5). For the two-quadrature strategy at one time, the matrix is I(t) = ($t^{2}$ $e^{{−2γt}}$/$σ^{2}$) times the 2×2 identity, because the sine and cosine derivatives combine through sin²(ωt) + cos²(ωt) = 1 and the off-diagonal terms cancel. Minimizing the resulting Cramér–Rao bound $σ²e^{{2γt}}$/t² with respect to t gives t = 1/γ, independent of ω. For coupled qubits, the machinery is a four-experiment state-preparation scheme: with neighbors fixed in |0⟩ the target Ramsey trace is a pure single-qubit cosine at ω, and with a neighbor in |1⟩ it is a cosine at ω + J, so crosstalk estimation reduces to decoupled single-qubit problems.

What would settle it

Compute the Fisher information for the X,Y strategy with the exact binomial variance of a ±1 measurement, Var = 1 − ⟨X(t)⟩², instead of the constant σ² assumed in Eq. (5); if the optimal time moves away from t = 1/γ or the predicted shot reduction drops below 50%, the optimization is an artifact of the constant-variance approximation.

Watch

Extended reading notes

Core claim

The paper establishes that, for a single qubit with Markovian dephasing, the optimal Ramsey calibration of detuning ω and dephasing rate γ is to measure both quadratures ⟨X(t)⟩ = cos(ωt)$e^{{−γt}}$ and ⟨Y(t)⟩ = sin(ωt)$e^{{−γt}}$ at a single time t = 1/γ. At that time the Fisher information matrix becomes proportional to the identity, so ω and γ are estimated independently; because Y is antisymmetric in ω, the sign of the detuning is also recoverable. The Cramér–Rao bound is reduced by a factor of about 0.7 compared with common many-time strategies, translating to up to 50% fewer shots at fixed precision. For coupled qubits, the paper reduces calibration to four Ramsey experiments in which only half the qubits are superposed while neighbors are fixed in |0⟩ or |1⟩, turning crosstalk estimation into decoupled single-qubit problems whose error is essentially independent of system size.

Load-bearing premise

The derivation treats each measured time point's mean as Gaussian with a constant variance σ², even though the true variance of a ±1 measurement is 1 − ⟨X(t)⟩² and depends on the unknown ω and γ.

Editorial extensions

If this is right

Single-qubit calibration can be reduced to one Ramsey time with two π/2 phases, cutting the shot budget by up to 50% at fixed precision.
Measuring Y alongside X also determines the sign of ω, which X-only schemes cannot.
The two-quadrature strategy is more robust than equally spaced sampling when the actual dephasing rate drifts from the assumed value.
Crosstalk calibration of a one-dimensional chain, a bipartite lattice, or a heavy-hex connectivity graph needs only four Ramsey experiments, with error essentially independent of qubit count.
Experimental data on a single NV center and two superconducting transmons reproduce the predicted ordering of the three strategies.

Reading between the lines

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

The paper leaves implicit that the constant-variance Gaussian approximation is the main quantitative assumption; redoing the optimization with the exact state-dependent variance 1 − ⟨X(t)⟩² is a direct next check and could shift the optimal time away from t = 1/γ.
Because the optimal time is independent of ω, the protocol can be run without first knowing the detuning; the paper demonstrates robustness to γ drift, and the same argument extends to ω drift.
The four-experiment decoupling for chains, bipartite lattices, and heavy-hex connectivity can be formulated as a graph-tiling task for arbitrary qubit connectivity; the paper mentions this tiling optimization but does not solve it.
For non-Markovian baths, the exponential decay is replaced by a function involving the bath correlation time; inserting that model into the same Fisher-information optimization should yield a new optimal time, a testable extension.

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.

Desk Editor's Note

read the letter

Short version: this paper is worth engaging. The two-quadrature single-time idea is genuinely nice, and the four-experiment crosstalk decoupling is a practical contribution. But the headline quantitative claims—that t=1/γ is optimal independent of ω, and that you save 50% of shots—are not actually proven. They rest on treating the measurement variance as constant when for ±1 measurements it is state-dependent. That doesn't kill the qualitative recommendation, but it does mean the exact optimal time and the 50% figure should be taken with salt.

What's new: the analytic Fisher-information computation for X and Y at a single time in SM4, the reduction of crosstalk calibration to four decoupled single-qubit experiments, and the experimental downsampling comparison on NV and transmon hardware. The derivation of the X,Y Fisher matrix at fixed time is clean, and the QuTiP simulation in Fig. 2 supports the ordering of the three strategies. The four-experiment protocol is clever and likely scalable; the heavy-hex tiling idea is plausible. The relative experimental comparisons are fair because they use long-measurement ground truth as a benchmark, and the raw data are shown in the SM.

Soft spots, in order of severity. First, the variance issue. Eq. (5) and SM2/SM4 model each time point's sample mean as Gaussian with variance σ², implicitly constant. But each Xi is the average of ±1 outcomes, so its variance is (1−⟨X(t)⟩²)/Nshots, which depends on ω and γ. With the correct heteroskedastic variance, the Fisher matrix is no longer proportional to identity, Iωγ does not vanish, and t=1/γ is not a stationary point of Tr[I⁻¹]; the optimum shifts and depends on ω (for ω=γ=1 it moves to about 0.95/γ). So the exact result in SM4 and the associated 0.7 CRB reduction / 50% shot-saving are approximate. The authors should either re-derive with the correct variance or justify why the approximation is good in the relevant regime. The Fig. 2 simulations likely can't distinguish t=1/γ from 0.95/γ at the shown shot counts, so they don't rescue the precise claim.

Second, the experimental validation is only relative. The NV data show a two-order-of-magnitude offset from the CRB, and the transmon crosstalk error saturates at large Ntot; the paper attributes these to SPAM and quasiparticles but doesn't model them, so the absolute predictive power of the CRB against real hardware remains untested. Third, the heavy-hex tiling claim in SM6 is asserted rather than demonstrated in the visible text; I would want the explicit tiling. Fourth, no code or data files are shipped, though the SM gives enough detail to reproduce the main simulation.

No circularity problem: the optimal times come from the model, not from fits. Citation of Ref. 16 (shared author) is for an optional non-Markovian extension, not the core derivation.

Who this is for: experimentalists doing routine qubit calibration, and anyone who wants a concrete worked example of CRB-based protocol design. It deserves a serious referee, but the authors should address the heteroskedastic variance before the quantitative claims are trusted. My recommendation: send to peer review, with a request for major revision.

Referee Report

3 major / 5 minor

Summary. This paper uses Fisher information and the Cramér–Rao bound to optimize Ramsey-based calibration of qubit detuning and dephasing. For a single qubit it claims that measuring both X and Y quadratures at a single time t = 1/γ minimizes the trace of the inverse Fisher information and yields up to a 50% reduction in the number of shots. The method is extended to crosstalk calibration by reducing coupled-qubit estimation to four decoupled single-qubit Ramsey experiments. The claims are supported by a QuTiP simulation that matches the authors' Gaussian-model CRB and by experiments on an NV center and superconducting transmons.

Significance. If the central optimality claim held, the work would be practically valuable for reducing calibration overhead, and the crosstalk-decoupling idea is elegant and potentially scalable. Strengths include an analytic derivation in SM4, numerical simulations that match the model used in the derivation, and experimental comparisons of three strategies on real devices. However, the headline quantitative claims rely on a homoskedastic Gaussian measurement model that does not describe projective ±1 measurements, so the universal optimal time t = 1/γ and the 50% shot-reduction figure are not established for the actual measurement statistics.

major comments (3)

[SM4, Eqs. (16)–(21)] The derivation of t = 1/γ assumes that the X and Y sample means have a common constant variance σ². For projective ±1 measurements the per-shot variance is 1 − ⟨X(t)⟩² = 1 − e^(−2γt) cos²(ωt) for X and 1 − e^(−2γt) sin²(ωt) for Y. Replacing σ² by these state-dependent variances makes the Fisher information no longer proportional to the identity and introduces a nonzero I_ωγ; consequently t = 1/γ is not a stationary point of Tr[I^(−1)] for the true measurement statistics. For example, at ω = γ = 1 the trace at t ≈ 0.95/γ is slightly smaller than at t = 1/γ. The claims that the optimal time is independent of ω and that the protocol gives a factor-0.7 CRB reduction, i.e. a 50% shot saving, therefore need to be re-derived under the exact Bernoulli likelihood or explicitly restricted to the homoskedastic model.
[Fig. 3(a) and note 22] The NV experimental errors in Fig. 3(a) lie almost two orders of magnitude above the theoretical CRB, and note 22 attributes this to weak spin-dependent fluorescence and photon collection efficiency. As a result, the statement that the experiments 'confirm our theoretical predictions' is supported only at the level of the relative ordering of the three strategies, not at the level of the quantitative CRB reduction or the 50% shot saving. To retain the quantitative claim, the detection inefficiency needs to be incorporated into the likelihood model, or the paper should be reframed as a relative comparison.
[Fig. 3(b)] For the two-transmon crosstalk experiment, Fig. 3(b) shows that for Ntot > 10^3 the one-time and two-time strategies saturate to a value different from the multi-time ground-truth value, which the text attributes to quasiparticle fluctuations. This means the experimental support for the multi-qubit version of the 50% reduction claim is confined to the low-Ntot regime. The main text should state this limitation explicitly in the experimental comparison, not only in the concluding discussion.

minor comments (5)

[SM4, after Eq. (21)] The text says the second derivative f''(t_opt) = 2e²/t_opt⁴ > 0 'indeed a maximum,' but a positive second derivative means f has a minimum there, which is the desired minimum of the CRB; the wording should be corrected.
[Main text, Eq. (2)] For a continuous noise process, F(t − t′) = γδ(t − t′) should refer to the Dirac delta function, not the Kronecker delta.
[Main text, after Eq. (4)] The parameter vector is written as θ = (ω, γ, t1, ..., tNtimes), but the times are optimization variables rather than estimated parameters; this should be clarified to avoid the impression that the CRB is taken with respect to the measurement times.
[Eq. (5) and SM2] Eq. (5) in the main text omits the 1/σ² factor that appears in SM2, and the treatment of Nshots differs between the two expressions; if σ² is chosen so that the two agree, that choice should be stated explicitly.
[SM1] The claim that the optimization 'converged to two measurement times only' is a numerical observation; if the two-time optimality is intended as a general result, a proof or a more systematic search over Ntimes should be provided.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the optimal Ramsey measurement strategy is derived from the stated model and benchmarked externally, not fitted and renamed as a prediction.

full rationale

The paper's derivation chain is self-contained. Equation (2) gives the closed-form Ramsey signal from the stated Markovian dephasing model; Eq. (5) and SM2 define the Fisher information under an explicit Gaussian/constant-variance approximation; SM4 analytically optimizes the resulting Cramer-Rao bound and obtains t = 1/gamma; and the coupled-qubit protocol is constructed from the model Hamiltonian Eq. (6), then tested against numerical simulation and experiment. No parameter is fitted and then renamed as a prediction: the 'ground truth' values from long measurements are used only as fixed benchmarks for downsampled experimental comparisons, so the optimality claims are not statistically forced by the benchmark procedure. The only self-citation, Ref. 16, shares an author but is proposed as an optional non-Markovian extension and carries no load in the central optimality proof. Ref. 18 and Ref. 19 are external prior works used for comparison/context. The constant-variance approximation in Eq. (5) and SM Eq. (19) is a genuine modeling assumption--the true Bernoulli variance is state-dependent--and may shift the optimal time quantitatively, but this is a correctness/robustness concern rather than circularity, because the optimum is derived from the approximate model rather than fitted to its outputs. No equation in the paper reduces to its own inputs by construction, and no load-bearing claim imports a uniqueness theorem from overlapping authors.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central claims rest on a few modeling choices: Markovian dephasing, constant-variance Gaussian shot noise, and a static ZZ crosstalk model. These are standard domain assumptions, but the constant-variance approximation is not exact for binary measurements and is not validated against the state-dependent variance case.

free parameters (1)

Quadrature measurement variance sigma^2 = 1 (implicit)
Eq. (5) and SM2 set the variance of each quadrature sample mean to a constant sigma^2, effectively 1, independent of the signal. The true variance for +/-1 outcomes is 1 - <X>^2, so this choice affects the CRB magnitude and the optimal time t = 1/gamma.

assumptions (4)

domain assumption Short-memory (Markovian) dephasing with F(t-t') = gamma*delta(t-t'), giving <X(t)> = cos(omega*t)e^{-gamma*t}
Invoked after Eq. (1) to derive Eq. (2); all Fisher information calculations in the main text and SM use this closed form.
domain assumption Sample means of quadrature measurements are Gaussian with constant variance sigma^2
SM2 derives Eq. (5) from a Gaussian likelihood; the main-text Eq. (5) omits sigma^2, effectively setting it to 1.
domain assumption Shot noise dominates, while SPAM and pulse errors are negligible in the relevant regime
The paper states SPAM errors limit precision to about 10^-3 and are independent of Ntimes, so the optimal resource allocation is analyzed for shot-noise domination.
domain assumption Crosstalk is a static ZZ coupling (J_i/4)(1-Z_i)(1-Z_{i+1})
Eq. (6) defines the coupled-qubit model; the four-experiment protocol and the tiling claims assume this exact form.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Optimal Calibration of Qubit Detuning and Crosstalk." pith.science (2026). https://pith.science/paper/VUHWKIXU

@misc{pith2026250710661,
  author       = {Pith},
  title        = {Pith review of: Optimal Calibration of Qubit Detuning and Crosstalk},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VUHWKIXU}},
  note         = {Machine review of arXiv:2507.10661}
}

read the original abstract

Characterizing and calibrating physical qubits is essential for maintaining the performance of quantum processors. A key challenge in this process is the presence of crosstalk that complicates the estimation of individual qubit detunings. In this work, we derive optimal strategies for estimating detuning and crosstalk parameters by optimizing Ramsey interference experiments using Fisher information and the Cramer-Rao bound. We compare several calibration protocols, including measurements of a single quadrature at multiple times and of two quadratures at a single time, for a fixed number of total measurements. Our results predict that the latter approach yields the highest precision and robustness in both cases of isolated and coupled qubits. We validate experimentally our approach using a single NV center as well as superconducting transmons. Our approach enables accurate parameter extraction with significantly fewer measurements, resulting in up to a 50% reduction in calibration time while maintaining estimation accuracy.

Figures

Figures reproduced from arXiv: 2507.10661 by the authors.

**Figure 2.** FIG. 2. Numerical simulation of three strategies for estimat [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 4.** FIG. 4. Schematic representation of the four experiments [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗

Figures from the paper (4 more)

**Figure 6.** Figure 6: FIG. 6. Cram´er-Rao bound as a function of the measurement times in (a) the theoretical calculation and (b) IQCC experiment [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

**Figure 7.** Figure 7: FIG. 7. Ratio of the number of shots between the first and second measurements, as a function of [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]

**Figure 8.** Figure 8: FIG. 8. Raw Ramsey-interference data (crosses) and least-squares fits to [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]

**Figure 9.** Figure 9: FIG. 9. IBM tiling for probing crosstalk effects. The red nodes represent qubits in the [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 22 canonical work pages

[1]

title About calibration jobs

author IBM Quantum Administration . title About calibration jobs . ://docs.quantum.ibm.com/admin/calibration-jobs

work page
[2]

title Private communications

author Google Quantum, Rigetti Computing, IQCC, and others . title Private communications

work page
[3]

title Engineering high-coherence superconducting qubits

author Siddiqi, I. title Engineering high-coherence superconducting qubits . journal Nature Reviews Materials volume 6 , pages 875--891 ( year 2021 )

work page 2021
[4]

author Chatterjee, A. et al. title Semiconductor qubits in practice . journal Nature Reviews Physics volume 3 , pages 157--177 ( year 2021 )

work page 2021
[5]

author Kjaergaard, M. et al. title Superconducting qubits: Current state of play . journal Annual Review of Condensed Matter Physics volume 11 , pages 369--395 ( year 2020 )

work page 2020
[6]

, author Zhang, G

author Mundada, P. , author Zhang, G. , author Hazard, T. & author Houck, A. title Suppression of qubit crosstalk in a tunable coupling superconducting circuit . journal Physical Review Applied volume 12 , pages 054023 ( year 2019 )

work page 2019
[7]

author Ni, Z. et al. title Scalable method for eliminating residual zz interaction between superconducting qubits . journal Physical review letters volume 129 , pages 040502 ( year 2022 )

work page 2022
[8]

author Xie, L. et al. title Suppressing zz crosstalk of quantum computers through pulse and scheduling co-optimization . In booktitle Proceedings of the 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems , pages 499--513 ( year 2022 )

work page 2022

Show all 25 references

[9]

, author Go, M

author Heng, S. , author Go, M. & author Han, Y. title Estimating the effect of crosstalk error on circuit fidelity using noisy intermediate-scale quantum devices . journal arXiv preprint arXiv:2402.06952 ( year 2024 )

2024 arXiv
[10]

author Fors, S. P. , author Fern \'a ndez-Pend \'a s, J. & author Kockum, A. F. title Comprehensive explanation of zz coupling in superconducting qubits . journal arXiv preprint arXiv:2408.15402 ( year 2024 )

2024 arXiv
[11]

author Sarovar, M. et al. title Detecting crosstalk errors in quantum information processors . journal Quantum volume 4 , pages 321 ( year 2020 )

2020
[12]

& author Petruccione, F

author Breuer, H.-P. & author Petruccione, F. title The theory of open quantum systems ( publisher OUP Oxford , year 2002 )

2002
[13]

& author Paris, M

author Benedetti, C. & author Paris, M. G. title Effective dephasing for a qubit interacting with a transverse classical field . journal International Journal of Quantum Information volume 12 , pages 1461004 ( year 2014 )

2014
[14]

, author Misguich, G

author Shirizly, L. , author Misguich, G. & author Landa, H. title Dissipative dynamics of graph-state stabilizers with superconducting qubits . journal Physical Review Letters volume 132 , pages 010601 ( year 2024 )

2024
[15]

title T2* ramsey characterization

author Qiskit community . title T2* ramsey characterization . ://qiskit-community.github.io/qiskit-experiments/manuals/characterization/t2ramsey.html
[16]

author Curtis, J. B. , author Yacoby, A. & author Demler, E. title Non-gaussian noise magnetometry using local spin qubits . journal arXiv preprint arXiv:2505.03877 ( year 2025 )

2025 arXiv
[17]

, author Marsman, M

author Ly, A. , author Marsman, M. , author Verhagen, J. , author Grasman, R. P. & author Wagenmakers, E.-J. title A tutorial on fisher information . journal Journal of Mathematical Psychology volume 80 , pages 40--55 ( year 2017 )

2017
[18]

, author Hodgkinson, P

author Jones, J. , author Hodgkinson, P. , author Barker, A. & author Hore, P. title Optimal sampling strategies for the measurement of spin--spin relaxation times . journal Journal of Magnetic Resonance, Series B volume 113 , pages 25--34 ( year 1996 )

1996
[19]

author Zohar, I. et al. title Real-time frequency estimation of a qubit without single-shot-readout . journal Quantum Science and Technology volume 8 , pages 035017 ( year 2023 )

2023
[20]

author Johansson, J. R. , author Nation, P. D. & author Nori, F. title Qutip: An open-source python framework for the dynamics of open quantum systems . journal Computer physics communications volume 183 , pages 1760--1772 ( year 2012 )

2012
[21]

, author Aspect, A

author Grynberg, G. , author Aspect, A. & author Fabre, C. title Introduction to quantum optics: from the semi-classical approach to quantized light ( publisher Cambridge university press , year 2010 )

2010
[22]

note Note that the experimental measurements are of a statistical nature as (i) they rely on a weak spin-dependent fluorescence difference of about 30\
[23]

title The future of high-performance quantum computing ( year 2025 )

author Israel Quantum Computing Center . title The future of high-performance quantum computing ( year 2025 ). ://i-qcc.com/

2025
[24]

author Serniak, K. et al. title Hot nonequilibrium quasiparticles in transmon qubits . journal Physical review letters volume 121 , pages 157701 ( year 2018 )

2018
[25]

& author Misguich, G

author Landa, H. & author Misguich, G. title Nonlocal correlations in noisy multiqubit systems simulated using matrix product operators . journal SciPost Physics Core volume 6 , pages 037 ( year 2023 )

2023

Pith tools

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

[1] [1]

title About calibration jobs

author IBM Quantum Administration . title About calibration jobs . ://docs.quantum.ibm.com/admin/calibration-jobs

work page

[2] [2]

title Private communications

author Google Quantum, Rigetti Computing, IQCC, and others . title Private communications

work page

[3] [3]

title Engineering high-coherence superconducting qubits

author Siddiqi, I. title Engineering high-coherence superconducting qubits . journal Nature Reviews Materials volume 6 , pages 875--891 ( year 2021 )

work page 2021

[4] [4]

author Chatterjee, A. et al. title Semiconductor qubits in practice . journal Nature Reviews Physics volume 3 , pages 157--177 ( year 2021 )

work page 2021

[5] [5]

author Kjaergaard, M. et al. title Superconducting qubits: Current state of play . journal Annual Review of Condensed Matter Physics volume 11 , pages 369--395 ( year 2020 )

work page 2020

[6] [6]

, author Zhang, G

author Mundada, P. , author Zhang, G. , author Hazard, T. & author Houck, A. title Suppression of qubit crosstalk in a tunable coupling superconducting circuit . journal Physical Review Applied volume 12 , pages 054023 ( year 2019 )

work page 2019

[7] [7]

author Ni, Z. et al. title Scalable method for eliminating residual zz interaction between superconducting qubits . journal Physical review letters volume 129 , pages 040502 ( year 2022 )

work page 2022

[8] [8]

author Xie, L. et al. title Suppressing zz crosstalk of quantum computers through pulse and scheduling co-optimization . In booktitle Proceedings of the 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems , pages 499--513 ( year 2022 )

work page 2022