Pith. sign in

REVIEW 5 major objections 7 minor 92 references

Quantum backend quality is workload-dependent: the paper's common UQ pipeline ranks Brisbane first on QSVT spectral reliability, Osaka first on aggregate VQA quality, and splits the ten VQA workload wins across all four backends.

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

2026-08-02 02:10 UTC pith:34DMDRFV

load-bearing objection A coherent and honest framework for application-level backend benchmarking, but the specific backend rankings are statistically thin and should be treated as illustrative, not conclusive. the 5 major comments →

arxiv 2607.14392 v1 pith:34DMDRFV submitted 2026-07-15 cs.ET cs.CEphysics.comp-phquant-ph

Unified Uncertainty Quantification Framework Bridging Noisy Quantum Backends Across Variational Quantum Algorithms and Quantum Signal Processing

Priyabrata Senapati , Vibin Abraham , Qiang Guan , Bo Peng This is my paper
classification cs.ET cs.CEphysics.comp-phquant-ph
keywords uncertainty quantificationquantum backend benchmarkingvariational quantum algorithmsquantum singular value transformationGreen's function reconstructionBayesian optimizationGaussian process surrogatebackend ranking
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 tries to establish that application-level uncertainty quantification can serve as a common benchmark language for very different quantum workloads: ten families of variational quantum algorithms and a quantum signal processing reconstruction of a molecular Green's function. It claims the same statistical machinery—Gaussian-process surrogate, Bayesian optimization with posterior refinement, sensitivity analysis, density estimation, backend ranking, and resource costing—faithfully characterizes both branches through a single tuple interface. If right, backend quality is not a single number but a workload-dependent pattern: on the four simulated noise models studied, Brisbane is most reliable on QSVT spectral reconstruction, Osaka best on aggregate variational quality, Kyoto reaches the spectral target late, and Kawasaki wins specific variational tasks. The implications are concrete for benchmarking practice: rankings change with metric and workload, and no single benchmark should be used as a universal proxy.

Core claim

The central claim is that noisy quantum backends can be characterized by how reliably they reach useful task-level behavior, not by their best achieved objective, and that this reliability is workload-dependent. The paper demonstrates this by collecting every backend evaluation as a tuple (xi, y), treating y as a backend-conditioned task output corrupted by per-evaluation shot noise, and then running one offline pipeline—surrogate-guided search, posterior refinement, elementary-effect sensitivity analysis, density-level robust regions, hit-rate/time-to-good ranking, and routed resource cost—on the recorded histories for both the VQA and QSVT branches. On the QSVT branch, the target is recove

What carries the argument

The load-bearing object is the tuple interface y_n = g_b(xi_n) + eps_n, where xi is the control vector (variational parameters or QSVT phase angles), g_b is the expected task outcome induced by backend b's noise channel, and eps_n is the per-evaluation fluctuation modeled as heteroscedastic Gaussian shot noise. This interface lets the same Gaussian-process surrogate, Bayesian optimization acquisition, variational/MCMC posterior refinement, elementary-effect sensitivity fingerprint, density-level robust-region estimate, and hit-rate/time-to-good ranking act on both the ten-family VQA branch and the 27-dimensional QSVT phase-vector branch without branch-specific modifications. On the QSVT side

Load-bearing premise

All reported backend rankings come from static simulated noise models, not live hardware, so the rankings reflect one calibration snapshot and may change with calibration drift, queue delays, or real device non-stationarity.

What would settle it

Run the same 100-evaluation guided optimization protocol on a live backend across several calibration windows and check whether the hit-rate ordering Brisbane > Osaka > Kawasaki > Kyoto and the VQA workload-winner pattern reproduce; if live-device rankings flip with calibration state or differ from simulated ones, the static-snapshot claim is falsified.

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

If this is right

  • Backend quality reports should include reliability metrics (hit rate, time-to-good, robust-region geometry) alongside best-value metrics, because best value alone ties backends that differ strongly in consistency.
  • A backend's rank can change when the workload changes, so cross-workload rankings must be reported per task family rather than as one universal score.
  • The same UQ pipeline transfers from variational to non-variational workloads, suggesting a common application-level benchmark template for future quantum algorithms.
  • Routed compilation cost must be read together with reliability: QSVT routed depth inflates by roughly 5.6-5.8x and total gates by 7.4-7.7x over ideal, changing the practical meaning of a backend's spectral success.
  • Robust parameter regions are backend-specific in location and geometry, so noise compensation learned on one backend does not transfer unchanged to another.

Where Pith is reading between the lines

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

  • If the same pipeline were run on live, drift-prone hardware, the rankings might reorder with calibration state; the paper's static-snapshot results give a lower bound on time-varying uncertainty, not a prediction about live-device rankings.
  • The finding that sensitivity mass concentrates in a backend-specific low-dimensional subset of the 27 QSVT phases suggests a practical calibration protocol: tune only the top sensitivity coordinates per backend and fix the rest, potentially cutting optimization cost substantially.
  • The LiH projection implies a gate-volume multiplier of roughly 4^6, suggesting the benchmark's real near-term value is comparative hardware characterization rather than chemistry scale-up; the same tuple pipeline could be extended to other matrix-function workloads to test whether the workload-dependence pattern is general.
  • Because VQAMET stays shallow and nearly backend-invariant while VQCFE is highly selective, workload difficulty and workload cost are separate axes; benchmark designers could stratify suites by both dimensions to avoid conflating 'hard' with 'expensive'.

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

5 major / 7 minor

Summary. The paper presents a unified uncertainty quantification framework for application-level benchmarking of noisy quantum backends, applied to two workload classes: ten variational quantum algorithm (VQA) families and a Quantum Singular Value Transformation (QSVT) construction of the H2 Green's function. Both branches share the tuple model y_n = g_b(xi_n) + eps_n (Eq. 11) and one offline pipeline (GP surrogate, Bayesian optimization with posterior refinement, Morris/robust-density analysis, backend ranking, resource estimation). The framework is instantiated on four IBM fake backends (Brisbane, Kawasaki, Kyoto, Osaka). The headline results are workload-dependent VQA backend preferences (Brisbane best mean rank 2.38, Osaka best mean quality) and a QSVT reliability ordering Brisbane > Osaka > Kawasaki > Kyoto, based on hit rates 51/48/44/44 out of 100 evaluations and a Time-to-Good tie-break. The paper also reports sensitivity fingerprints, robust-region geometry, and transpiled resource costs, and projects the QSVT pipeline to LiH.

Significance. If the empirical claims were secured, this would be a genuinely useful contribution: it provides a common statistical language for variational and non-variational workloads; the ten VQA instances are concretely specified (Tables 5-7); the QSVT branch includes block-encoding details and a real-time segmentation error bound (App. B); the resource layer separates deep-single-circuit from many-shallow-circuits cost regimes; and the LiH projection (Eqs. 19-24) is a quantitative, falsifiable scaling prediction. The paper is also unusually explicit about scope (Sec. 5, App. J) and the workflow design (Algorithm 1, saved records, offline analysis) is reproducible in structure. However, the central empirical conclusions currently outrun the evidence: the QSVT good-tuple classifier is under-specified, the hit-rate ranking has no confidence intervals or seed variation, the VQA quality metric is relative by construction, and the sensitivity fingerprints rest on unvalidated surrogates. These are fixable within the manuscript's scope, which is why I am not recommending rejection.

major comments (5)
  1. [§2.4, Eq. (14), Table 8, Fig. 3] The binary good/bad classification behind every QSVT reliability number is not reproducible. A 'good tuple' is defined as one in which the guided peak loss of Eq. (14) 'reaches its lower target,' but the target value is never given; the per-peak weights w_{j,r} are never specified; and the peak distance ℓ(·,·) is only described as 'position and height aware.' Fig. 3's caption fixes only the sign symmetry min(|ω−ω_ideal|,|ω+ω_ideal|), not the tolerance in position or height. Since the hit rates in Table 8 (51/48/44/44) and the resulting ranking Brisbane > Osaka > Kawasaki > Kyoto derive entirely from this classifier, the central QSVT result cannot be independently verified. Please report the exact threshold(s) and weights, and test the stability of the ranking to variation in these choices.
  2. [§4.2, Table 8, App. C] The QSVT ranking rests on a single 100-evaluation adaptive trajectory per backend with no attached uncertainty. At 44/100 the binomial standard error is about 5 percentage points, so the 51/48/44/44 spread is within roughly one standard error of a common rate; the Kawasaki/Kyoto tie-break (Time-to-Good 1 vs 15) is a single-run minimum and is not a stable discriminator. Because the trajectory is adaptive, hit rate conflates backend noise with GP/acquisition/VI behavior. App. C itself notes that a beta-binomial ranking model could have been used but was not, and Fig. 17(b)'s 'pairwise win probabilities' have no stated definition or error bars. Please add seed variation, confidence intervals, and/or a model-based comparison, and re-frame the ordering as provisional.
  3. [§2.4, §4.1, Fig. 7, §6] quality_norm is a per-workload min-max rescaling of the best and worst observed values across the four backends, which makes all VQA aggregate rankings relative to the tested backend set and to noisy trajectory extrema. Adding or removing a backend would rescale every score, and a workload in which all four backends fail would still yield a 1.0 for the least-bad backend. The definition is stated, but the abstract's claim that the framework measures 'how reliably each backend reaches useful task level behavior' and the conclusion's 'concrete backend conclusions' overstate what this metric provides. The aggregate gaps are also small (mean ranks 2.38-2.61 over ten workloads) with no reported uncertainty. Please scope the VQA claims to relative performance within the tested set and provide bootstrap-type uncertainty over workloads and refinement settings.
  4. [§4.3, App. D] The Morris/Sobol/SHAP sensitivity fingerprints and robust-region densities are computed 'on surrogate models fitted to the backend evaluation records' (App. D), but no validation of the GP surrogates is reported—no cross-validated error, no comparison against direct evaluations—and 100 adaptive samples in a 27-dimensional phase space is a thin basis for global sensitivity analysis. Claims that the top-five phase block carries roughly two-thirds of the Morris mass and that the robust-region geometry is backend-specific therefore inherit unquantified surrogate error. Please report surrogate predictive accuracy and confirm the Morris ordering with a direct sampling-based estimate or a stability analysis across surrogate fits.
  5. [§5, App. E, §4.6, §6] The paper is appropriately explicit in §5 that all results use IBM fake-backend noise models and Aer execution rather than live QPUs, and I credit that. However, the abstract and conclusion present 'backend rankings' (e.g., 'Brisbane is the strongest backend') without re-qualifying that these are statements about static calibration snapshots. The 'transpile once, bind phases' cache of App. E is noise-faithful only because the fake backends are static; live hardware would require a refresh policy (as App. E acknowledges) and would invalidate the single-snapshot noise model. Please carry the §5 qualification into the abstract and conclusion, and state clearly that the framework has not been tested on non-stationary hardware.
minor comments (7)
  1. [§2.4, Table 3] Mean Eval95 is defined as reaching '95% of the per workload best value,' but the sign convention is unclear for minimization workloads where y = -f: if the best value is negative, 95% of it is a worse threshold. Please state the convention explicitly and note how negative optima are handled.
  2. [§2.2, §3.2] The 27-dimensional phase vector is the concatenation of the cosine-branch (13) and sine-branch (14) QSP phase vectors of the degree pair (12,13). Eq. (7) presents a single (d+1)-dimensional vector; state the concatenation explicitly so the dimension is checkable.
  3. [Fig. 17(b)] The 'pairwise win probabilities' have no definition in the text or caption. Specify what is being compared (evaluations? runs?) and how the probability is estimated.
  4. [App. J] Appendix J lists only two threats to validity (branch complementarity and breadth/tradeoff). The statistical issues affecting Table 8—single trajectory per backend, no confidence intervals, and the unspecified good-tuple threshold—are the most serious threats to the central claim and should be acknowledged there.
  5. [App. C / data availability] The paper states that offline analysis can be reproduced from saved evaluation records, but no data or code availability statement is provided. For a benchmarking paper, releasing the evaluation logs (or at least the aggregate tables with uncertainty) would substantially strengthen the reproducibility case.
  6. [§4.5, Eq. (24)] The LiH projection relies on 'the H2 fit gives roughly 2.5×10^8 routed gates at λt=5,' but this fit is not linked to any figure or table in the manuscript, so the projection is not traceable. Please state where this number comes from.
  7. [Table 9, §4.3] Table 9 has no paired checkpoint sets for VQLS and VQAPDE, yet §4.3 claims that VQAPDE and VQLS 'distribute their sensitivity more broadly' without noting this incompleteness. Add a cross-reference to Table 9's dashes when making sensitivity claims about those workloads.

Circularity Check

0 steps flagged

No significant circularity: all benchmark metrics are explicitly defined from recorded evaluations, and the under-specified QSVT good-tuple threshold is a reproducibility/correctness issue, not a circular reduction.

full rationale

The paper's derivation chain is a benchmarking methodology, not a predictive theory whose outputs are fed back into its definitions. The shared tuple model y_n = g_b(xi_n) + eps_n (Eq. 11) and the GP update (Eq. 16) are standard statistical machinery; the VQA and QSVT objectives are defined independently (Eqs. 1, 14). The quality_norm metric is explicitly a per-workload min-max rescaling across the four backends, so the resulting rankings are by construction relative to the tested set, but this is stated openly in Section 2.4 rather than hidden, and no claim is made that it provides an absolute external scale. The QSVT hit-rate ranking depends on the unquantified 'lower target' and per-peak weights in Eq. (14); however, the paper does not say or imply that these were fitted to produce the backend ordering, and the appendix reports raw hit rates and Time-to-Good values. An unspecified threshold is a reproducibility limitation (as the skeptic notes), not a circular step by the paper's own equations. Self-citations such as [1], [72], and [74] appear, but the central QSVT construction is also anchored in standard external references [29, 47, 48, 53] and many-body texts [27, 52], and no load-bearing uniqueness theorem is imported from the authors' prior work. The stated limitations (fake-backend noise models, qualitative noise correlations) further scope the claims rather than smuggling in conclusions. No prediction in the paper reduces to a fitted parameter or to a self-citation chain by construction, so the appropriate circularity score is 0.

Axiom & Free-Parameter Ledger

6 free parameters · 5 axioms · 1 invented entities

The paper introduces no new physical entities; the load-bearing free parameters are the benchmark-design choices (loss weights, eta, good-tuple target, level-set threshold, min-max normalization) and the fitted QSVT phase centers. The key domain assumption is the faithfulness of fake-backend snapshots and the transpile-once cache. The LiH gate-count projection is an explicit linear extrapolation.

free parameters (6)
  • QSVT phase vector theta (27 dimensions) = backend-specific centers reported in Fig. 16 (figure absent)
    The 27 phase entries are the control variables optimized on each backend; the paper reports backend-specific robust-region centers, so they are fitted parameters, not derived constants.
  • Per-peak weights w_{j,r} in guided peak loss Eq. (14) = not stated
    The guided peak loss sums weighted position/height-aware peak distances; the weights are chosen by hand and are not reported, yet they define what counts as a good tuple and hence the hit rates.
  • Broadening parameter eta = 0.15 a.u. = 0.15
    Chosen for the H2 spectral benchmark; controls peak width and affects whether a reconstructed peak is classified as good.
  • Good-tuple lower target for Eq. (14) = not stated
    The definition of good versus bad tuples depends on a lower target value for the guided peak loss that is never quantified, yet hit rate and Time-to-Good are the QSVT ranking metrics.
  • alpha=0.9 level-set threshold for robust region = 0.9
    Convention chosen for density level sets in Sect. 2.4; affects reported robust-region volumes.
  • Per-workload min-max normalization range = [0,1] with 1=best, 0=worst across four backends
    quality_norm rescales each workload by the best/worst achieved on the four backends, so every quality value is relative to the tested set rather than to an absolute physical reference.
axioms (5)
  • domain assumption Aer fake-backend noise models are faithful proxies for the four named IBM backends.
    All reported rankings are generated from fake-backend snapshots; Sect. 5 explicitly limits the claims to static snapshots, and the noise-faithfulness of the transpile-once cache (Appendix E) depends on this.
  • standard math The QSVT polynomial approximation error bound (Eq. 31) with segmentation r=ceil(tau/tau_max) is valid.
    Uses the telescoping identity for segmented evolution error; this is standard QSP/QSVT theory (Gilyén et al.), though the specific segmentation parameters tau_max=1 are chosen here.
  • domain assumption The H2 four-orbital Hamiltonian block encoding with three QSVT ancillas and the Hadamard-test wrapper correctly implements the retarded Green's function of Eqs. (2)-(5).
    The circuit-level construction in Appendix B is asserted to realize the hole-sector Green's function estimator; no independent verification (e.g., comparison to exact diagonalization) is shown in the reviewed text.
  • standard math CLT justifies Gaussian per-evaluation noise in Eq. (11).
    The paper explicitly states that the Gaussian model is a regression-side approximation for shot-averaged estimates, not a claim about hardware noise; this is a standard modeling assumption for GP surrogates.
  • domain assumption The LiH projection of Sect. 4.5 assumes the H2 routed-gate empirical trend is a linear proxy for LiH.
    Equation (24) uses G_LiH ~ 4.1e3 * G_H2 as a projection; the paper itself labels it as a projection, not a measured circuit. This assumption is load-bearing for the resource-scaling claim.
invented entities (1)
  • No new physical entities introduced. no independent evidence
    purpose: None
    The QSVT phases, GP surrogate, and benchmark metrics are methodological constructs, not new physical objects.

pith-pipeline@v1.3.0-alltime-deepseek · 36627 in / 9163 out tokens · 72652 ms · 2026-08-02T02:10:51.810602+00:00 · methodology

0 comments
read the original abstract

We present an uncertainty quantification (UQ) framework for application level benchmarking and characterization of noisy quantum backends. The framework compares two workload classes under one statistical pipeline: noisy intermediate scale quantum (NISQ) variational quantum algorithms (VQAs) and Quantum Singular Value Transformation (QSVT) based Green's function reconstruction. For the VQA branch, we evaluate ten benchmark families spanning chemistry, optimization, simulation, compiling, linear solving, partial differential equations, metrology, error correction, tomography, and channel fidelity estimation. For the QSVT branch, we reconstruct orbital resolved Green's functions and spectral peaks from a block encoded real time propagator. The workflow combines Bayesian optimization, posterior distribution refinement, sensitivity analysis, robust parameter density estimation, backend ranking, noise correlation, and resource estimation analysis. Instead of reporting only one best parameter vector, the framework identifies robust parameter regions, residual gaps to ideal behavior, backend specific failure modes, and calibration sensitive uncertainty. The result is a common benchmark for variational and non-variational workloads that measures how reliably each backend reaches useful task level behavior.

Figures

Figures reproduced from arXiv: 2607.14392 by Bo Peng, Priyabrata Senapati, Qiang Guan, Vibin Abraham.

Figure 1
Figure 1. Figure 1: Shared UQ workflow for application level benchmarking across VQAs and QSVT Green’s function [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Four backends VQA benchmark views built from §3.1: each of the ten VQA workloads is run on [PITH_FULL_IMAGE:figures/full_fig_p013_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: QSVT spectral function reconstructions across the four backends, built from the [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: VQA resource summary built from the same runs as Figure 2. Every circuit produced during the [PITH_FULL_IMAGE:figures/full_fig_p017_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: QSVT resource summary built from the QSVT block encoding circuit of §2.2 evaluated across a sweep [PITH_FULL_IMAGE:figures/full_fig_p018_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Compact QSVT multimetric resource summary for the four routed backends and the ideal reference [PITH_FULL_IMAGE:figures/full_fig_p019_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Cross paradigm backend summary panel for the four backends of §3. The five columns are computed [PITH_FULL_IMAGE:figures/full_fig_p020_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Shared variational workload template. Each VQA begins from a task specific preparation or encoding [PITH_FULL_IMAGE:figures/full_fig_p027_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Green’s function circuit overview and component breakdown. Panels (a) and (b) compare the active [PITH_FULL_IMAGE:figures/full_fig_p030_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Shared UQ workflow used throughout this work. The same closed loop generates posterior updates, [PITH_FULL_IMAGE:figures/full_fig_p034_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Degree sweep summary for the QSVT propagator and Green’s function circuits. Width is structural, [PITH_FULL_IMAGE:figures/full_fig_p036_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Time sweep summary at fixed degree (12, 13). Panels (a) and (b) share identical linear axis ranges and report the metric values directly in millions, with 𝐺(𝑡) solid, 𝐸QSVT(𝑡) dashed, black lines for the ideal references, and the legend of panel (a) applying to both. Panel (c) shows the segment count staircase of Eq. (47). The fixed degree resource curves are close to linear in 𝑡 because each additional s… view at source ↗
Figure 13
Figure 13. Figure 13: Supplemental VQA resource view. Every workload reports the same width of [PITH_FULL_IMAGE:figures/full_fig_p038_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Unsuccessful tuple spectral diagnostics. Gray curves denote individual unsuccessful reconstructions, [PITH_FULL_IMAGE:figures/full_fig_p040_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Cross backend sensitivity summaries. The four heatmaps share one row ordering, sorted by the mean [PITH_FULL_IMAGE:figures/full_fig_p041_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Density model summaries. Backend robust regions have different centers and partial overlap rather [PITH_FULL_IMAGE:figures/full_fig_p042_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Application level Green’s/QSVT benchmark views supporting Table 8. Brisbane is strongest for this [PITH_FULL_IMAGE:figures/full_fig_p043_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: VI based VQA parameter distribution summaries. Panel (a) reports the backend local good rate under [PITH_FULL_IMAGE:figures/full_fig_p044_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Backend level summaries for the VQA branch. The regret panel provides the aggregate optimization [PITH_FULL_IMAGE:figures/full_fig_p045_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: Objective distributions for the VQA workloads under the VI posterior refinement baseline. Each [PITH_FULL_IMAGE:figures/full_fig_p046_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: Application level VQA backend benchmarks. Brisbane has the best mean rank, while Osaka gives the [PITH_FULL_IMAGE:figures/full_fig_p047_21.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

92 extracted references · 23 canonical work pages

  1. [1]

    Vibin Abraham, Priyabrata Senapati, Himadri Pathak, and Bo Peng. 2026. Elucidating Many-Body Effects in Molecular Core Spectra through Real-Time Approaches: Efficient Classical Approximations and a Quantum Perspective.The Journal of Chemical Physics164, 10 (2026), 104113. doi:10.1063/5.0313721

    work page doi:10.1063/5.0313721 2026

  2. [2]

    Thomas E. Baker. 2021. Lanczos Recursion on a Quantum Computer for the Green’s Function and Ground State. Physical Review A103, 3 (2021), 032404. doi:10.1103/PhysRevA.103.032404

    work page doi:10.1103/physreva.103.032404 2021

  3. [3]

    Jiang, Samuel Daulton, Benjamin Letham, Andrew Gordon Wilson, and Eytan Bakshy

    Maximilian Balandat, Brian Karrer, Daniel R. Jiang, Samuel Daulton, Benjamin Letham, Andrew Gordon Wilson, and Eytan Bakshy. 2020. BoTorch: A Framework for Efficient Monte-Carlo Bayesian Optimization. InAdvances in Neural Information Processing Systems, Vol. 33. 21524–21538. arXiv:1910.06403

    Pith/arXiv arXiv 2020

  4. [4]

    Sohaib Alam, Guillermo Alonso-Linaje, B

    Ville Bergholm, Josh Izaac, Maria Schuld, Christian Gogolin, Shahnawaz Ahmed, Vishnu Ajith, M. Sohaib Alam, Guillermo Alonso-Linaje, B. AkashNarayanan, Ali Asadi, et al. 2018. PennyLane: Automatic Differentiation of Hybrid Quantum-Classical Computations. arXiv:1811.04968 [quant-ph] doi:10.48550/arXiv.1811.04968

    Pith/arXiv arXiv doi:10.48550/arxiv.1811.04968 2018

  5. [5]

    Lennart Bittel and Martin Kliesch. 2021. Training Variational Quantum Algorithms Is NP-Hard.Physical Review Letters 127, 12 (2021), 120502. doi:10.1103/PhysRevLett.127.120502

    doi:10.1103/physrevlett.127.120502 2021

  6. [6]

    Tizian Blatz, Joyce Kwan, Julian Léonard, and Annabelle Bohrdt. 2024. Bayesian Optimization for Robust State Preparation in Quantum Many-Body Systems.Quantum8 (2024), 1388. doi:10.22331/q-2024-06-27-1388

    work page doi:10.22331/q-2024-06-27-1388 2024

  7. [7]

    Blei, Alp Kucukelbir, and Jon D

    David M. Blei, Alp Kucukelbir, and Jon D. McAuliffe. 2017. Variational Inference: A Review for Statisticians.J. Amer. Statist. Assoc.112, 518 (2017), 859–877. doi:10.1080/01621459.2017.1285773

    arXiv 2017

  8. [8]

    Robin Blume-Kohout, Timothy Proctor, and Kevin Young. 2025. Quantum Characterization, Verification, and Validation. arXiv:2503.16383 doi:10.48550/arXiv.2503.16383

    Pith/arXiv arXiv doi:10.48550/arxiv.2503.16383 2025

  9. [9]

    Robin Blume-Kohout and Kevin C. Young. 2020. A Volumetric Framework for Quantum Computer Benchmarks. Quantum4 (2020), 362. doi:10.22331/q-2020-11-15-362

    doi:10.22331/q-2020-11-15-362 2020

  10. [10]

    Cerezo, Yigit Subasi, Lukasz Cincio, and Patrick J

    Carlos Bravo-Prieto, Ryan LaRose, M. Cerezo, Yigit Subasi, Lukasz Cincio, and Patrick J. Coles. 2023. Variational Quantum Linear Solver.Quantum7 (2023), 1188. doi:10.22331/q-2023-11-22-1188

    doi:10.22331/q-2023-11-22-1188 2023

  11. [11]

    Daan Camps and Roel Van Beeumen. 2022. FABLE: Fast Approximate Quantum Circuits for Block-Encodings. InProceedings of the 2022 IEEE International Conference on Quantum Computing and Engineering (QCE). 104–113. doi:10.1109/QCE53715.2022.00029

    arXiv 2022

  12. [12]

    Olson, Matthias Degroote, Peter D

    Yudong Cao, Jonathan Romero, Jonathan P. Olson, Matthias Degroote, Peter D. Johnson, Mária Kieferová, Ian D. Kivlichan, Tim Menke, Borja Peropadre, Nicolas P. D. Sawaya, Sukin Sim, Libor Veis, and Alán Aspuru-Guzik

  13. [13]

    Cerezo, Andrew Arrasmith, Ryan Babbush, Simon C

    M. Cerezo, Andrew Arrasmith, Ryan Babbush, Simon C. Benjamin, Suguru Endo, Keisuke Fujii, Jarrod R. McClean, Kosuke Mitarai, Xiao Yuan, Lukasz Cincio, and Patrick J. Coles. 2021. Variational Quantum Algorithms.Nature Reviews Physics3 (2021), 625–644. doi:10.1038/s42254-021-00348-9

    doi:10.1038/s42254-021-00348-9 2021

  14. [14]

    Cerezo, Alexander Poremba, Lukasz Cincio, and Patrick J

    M. Cerezo, Alexander Poremba, Lukasz Cincio, and Patrick J. Coles. 2020. Variational Quantum Fidelity Estimation. Quantum4 (2020), 248. doi:10.22331/q-2020-03-26-248

    doi:10.22331/q-2020-03-26-248 2020

  15. [15]

    Cerezo, Akira Sone, Tyler Volkoff, Lukasz Cincio, and Patrick J

    M. Cerezo, Akira Sone, Tyler Volkoff, Lukasz Cincio, and Patrick J. Coles. 2021. Cost Function Dependent Barren Plateaus in Shallow Parametrized Quantum Circuits.Nature Communications12 (2021), 1791. doi:10.1038/s41467-021- 21728-w

    doi:10.1038/s41467-021- 2021

  16. [16]

    Raj Chakrabarti and Herschel Rabitz. 2007. Quantum control landscapes.International Reviews in Physical Chemistry 26, 4 (2007), 671–735. doi:10.1080/01442350701633300

    work page doi:10.1080/01442350701633300 2007

  17. [17]

    Childs and Nathan Wiebe

    Andrew M. Childs and Nathan Wiebe. 2012. Hamiltonian Simulation Using Linear Combinations of Unitary Operations. Quantum Information and Computation12, 11–12 (2012), 901–924. doi:10.26421/QIC12.11-12-1

    doi:10.26421/qic12.11-12-1 2012

  18. [18]

    Cross, Lev S

    Andrew W. Cross, Lev S. Bishop, Sarah Sheldon, Paul D. Nation, and Jay M. Gambetta. 2019. Validating Quantum Computers Using Randomized Model Circuits.Physical Review A100, 3 (2019), 032328. doi:10.1103/PhysRevA.100. 032328 23

    doi:10.1103/physreva.100 2019

  19. [19]

    Diksha Dhawan, Dominika Zgid, and Mario Motta. 2024. Quantum Algorithm for Imaginary-Time Green’s Functions. Journal of Chemical Theory and Computation20, 11 (2024), 4629–4638. doi:10.1021/acs.jctc.4c00241

    work page doi:10.1021/acs.jctc.4c00241 2024

  20. [20]

    Daoyi Dong and Ian R Petersen. 2010. Quantum control theory and applications: a survey.IET Control Theory & Applications4, 12 (2010), 2651–2671. doi:10.1049/iet-cta.2009.0508

    arXiv 2010

  21. [21]

    Yulong Dong, Lin Lin, and Yu Tong. 2022. Ground-State Preparation and Energy Estimation on Early Fault-Tolerant Quantum Computers via Quantum Eigenvalue Transformation of Unitary Matrices.PRX Quantum3, 4 (2022), 040305. doi:10.1103/PRXQuantum.3.040305

    doi:10.1103/prxquantum.3.040305 2022

  22. [22]

    Eisert, D

    J. Eisert, D. Hangleiter, N. Walk, I. Roth, D. Markham, R. Parekh, U. Chabaud, and E. Kashefi. 2020. Quantum Certification and Benchmarking.Nature Reviews Physics2, 7 (2020), 382–390. doi:10.1038/s42254-020-0186-4

    doi:10.1038/s42254-020-0186-4 2020

  23. [23]

    Nakagawa

    Suguru Endo, Ikko Kurata, and Yuya O. Nakagawa. 2020. Calculation of the Green’s Function on Near-Term Quantum Computers.Physical Review Research2, 3 (2020), 033281. doi:10.1103/PhysRevResearch.2.033281

    work page doi:10.1103/physrevresearch.2.033281 2020

  24. [24]

    Wallman, Lukas Postler, Michael Meth, Roman Stricker, Esteban A

    Alexander Erhard, Joel J. Wallman, Lukas Postler, Michael Meth, Roman Stricker, Esteban A. Martinez, Philipp Schindler, Thomas Monz, Joseph Emerson, and Rainer Blatt. 2019. Characterizing Large-Scale Quantum Computers via Cycle Benchmarking.Nature Communications10 (2019), 5347. doi:10.1038/s41467-019-13068-7

    doi:10.1038/s41467-019-13068-7 2019

  25. [25]

    Turner, and Matthias Poloczek

    David Eriksson, Michael Pearce, Jacob Gardner, Ryan D. Turner, and Matthias Poloczek. 2019. Scalable Global Optimization via Local Bayesian Optimization. InAdvances in Neural Information Processing Systems

    2019

  26. [26]

    Edward Farhi, Jeffrey Goldstone, and Sam Gutmann. 2014. A Quantum Approximate Optimization Algorithm. arXiv:1411.4028. doi:10.48550/arXiv.1411.4028

    Pith/arXiv arXiv doi:10.48550/arxiv.1411.4028 2014

  27. [27]

    Fetter and John D

    Alexander L. Fetter and John D. Walecka. 2003.Quantum Theory of Many-Particle Systems. Dover

    2003

  28. [28]

    Stepan Fomichev, Pablo A. M. Casares, Jay Soni, Utkarsh Azad, Alexandr Kunitsa, Arne-Christian Voigt, Jonathan E. Mueller, and Juan Miguel Arrazola. 2025. Fast Simulations of X-Ray Absorption Spectroscopy for Battery Materials on a Quantum Computer. doi:10.48550/arXiv.2506.15784

    Pith/arXiv arXiv doi:10.48550/arxiv.2506.15784 2025

  29. [29]

    András Gilyén, Yuan Su, Guang Hao Low, and Nathan Wiebe. 2019. Quantum Singular Value Transformation and Beyond: Exponential Improvements for Quantum Matrix Arithmetics. InProceedings of the 51st Annual ACM SIGACT Symposium on Theory of Computing. 193–204. doi:10.1145/3313276.3316366

    arXiv 2019

  30. [30]

    Williams-Young, and Wibe A

    Niladri Gomes, David B. Williams-Young, and Wibe A. de Jong. 2023. Computing the Many-Body Green’s Function with Adaptive Variational Quantum Dynamics.Journal of Chemical Theory and Computation19, 11 (2023), 3313–3323. doi:10.1021/acs.jctc.3c00150

    work page doi:10.1021/acs.jctc.3c00150 2023

  31. [31]

    Gabriel Greene-Diniz, David Zsolt Manrique, Kentaro Yamamoto, Evgeny Plekhanov, Nathan Fitzpatrick, Michal Krompiec, Rei Sakuma, and David Muñoz Ramo. 2024. Quantum Computed Green’s Functions Using a Cumulant Expansion of the Lanczos Method.Quantum8 (2024), 1383. doi:10.22331/q-2024-06-20-1383

    work page doi:10.22331/q-2024-06-20-1383 2024

  32. [32]

    Nguyen, Noah Goss, Brian Marinelli, Ravi K

    Akel Hashim, Long B. Nguyen, Noah Goss, Brian Marinelli, Ravi K. Naik, Trevor Chistolini, Jordan Hines, J. P. Marceaux, Yosep Kim, Pranav Gokhale, Teague Tomesh, Senrui Chen, Liang Jiang, Samuele Ferracin, Kenneth Rudinger, Timothy Proctor, Kevin C. Young, Irfan Siddiqi, and Robin Blume-Kohout. 2025. Practical Introduction to Benchmarking and Characteriza...

    doi:10.1103/prxquantum.6.030202 2025

  33. [33]

    Cerezo, and Patrick J

    Zoë Holmes, Kunal Sharma, M. Cerezo, and Patrick J. Coles. 2022. Connecting Ansatz Expressibility to Gradient Magnitudes and Barren Plateaus.PRX Quantum3, 1 (2022), 010313. doi:10.1103/PRXQuantum.3.010313

    doi:10.1103/prxquantum.3.010313 2022

  34. [34]

    Wood, Jake Lishman, Julien Gacon, Simon Martiel, Paul D

    Ali Javadi-Abhari, Matthew Treinish, Kevin Krsulich, Christopher J. Wood, Jake Lishman, Julien Gacon, Simon Martiel, Paul D. Nation, Lev S. Bishop, Andrew W. Cross, Blake R. Johnson, and Jay M. Gambetta. 2024. Quantum Computing with Qiskit. arXiv:2405.08810 doi:10.48550/arXiv.2405.08810

    Pith/arXiv arXiv doi:10.48550/arxiv.2405.08810 2024

  35. [35]

    Jones, Matthias Schonlau, and William J

    Donald R. Jones, Matthias Schonlau, and William J. Welch. 1998. Efficient Global Optimization of Expensive Black-Box Functions.Journal of Global Optimization13, 4 (1998), 455–492. doi:10.1023/A:1008306431147

    doi:10.1023/a:1008306431147 1998

  36. [36]

    Nakagawa, Kazunori Maruyama, Hirotaka Oshima, and Shintaro Sato

    Shota Kanasugi, Shoichiro Tsutsui, Yuya O. Nakagawa, Kazunori Maruyama, Hirotaka Oshima, and Shintaro Sato

  37. [37]

    Chow, and Jay M

    Abhinav Kandala, Antonio Mezzacapo, Kristan Temme, Maika Takita, Markus Brink, Jerry M. Chow, and Jay M. Gambetta. 2017. Hardware-Efficient Variational Quantum Eigensolver for Small Molecules and Quantum Magnets. Nature549, 7671 (2017), 242–246. doi:10.1038/nature23879

    doi:10.1038/nature23879 2017

  38. [38]

    Trevor Keen, Bo Peng, Karol Kowalski, Pavel Lougovski, and Steven Johnston. 2022. Hybrid Quantum-Classical Approach for Coupled-Cluster Green’s Function Theory.Quantum6 (2022), 675. doi:10.22331/q-2022-03-30-675

    work page doi:10.22331/q-2022-03-30-675 2022

  39. [39]

    Sornborger, and Patrick J

    Sumeet Khatri, Ryan LaRose, Alexander Poremba, Lukasz Cincio, Andrew T. Sornborger, and Patrick J. Coles. 2019. Quantum-Assisted Quantum Compiling.Quantum3 (2019), 140. doi:10.22331/q-2019-05-13-140

    doi:10.22331/q-2019-05-13-140 2019

  40. [40]

    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

  41. [41]

    Taichi Kosugi and Yu-ichiro Matsushita. 2020. Construction of Green’s Functions on a Quantum Computer: Quasipar- ticle Spectra of Molecules.Physical Review A101, 1 (2020), 012330. doi:10.1103/PhysRevA.101.012330 24 Senapati et al

    doi:10.1103/physreva.101.012330 2020

  42. [42]

    Bauman, Guang Hao Low, Martin Roetteler, John J

    Karol Kowalski, Nicholas P. Bauman, Guang Hao Low, Martin Roetteler, John J. Rehr, and Fernando D. Vila. 2024. Capturing Many-Body Correlation Effects with Quantum and Classical Computing.Physical Review A110, 6 (2024), L060401. doi:10.1103/PhysRevA.110.L060401

    work page doi:10.1103/physreva.110.l060401 2024

  43. [43]

    Alexandr Kunitsa, Diksha Dhawan, Stepan Fomichev, Juan Miguel Arrazola, Minghao Zhang, and Torin F. Stetina

  44. [44]

    Coles, Lukasz Cincio, Jarrod R

    Martín Larocca, Supanut Thanasilp, Samson Wang, Kunal Sharma, Jacob Biamonte, Patrick J. Coles, Lukasz Cincio, Jarrod R. McClean, Zoë Holmes, and M. Cerezo. 2025. Barren Plateaus in Variational Quantum Computing.Nature Reviews Physics(2025). doi:10.1038/s42254-025-00813-9

    doi:10.1038/s42254-025-00813-9 2025

  45. [45]

    Marjuka F Lazin, Christian R Shelton, Simon N Sandhofer, and Bryan M Wong. 2023. High-dimensional multi-fidelity Bayesian optimization for quantum control.Mach. Learn.: Sci. Technol.4, 4 (2023), 045014. doi:10.1088/2632-2153/ad0100

    work page doi:10.1088/2632-2153/ad0100 2023

  46. [46]

    Francesco Libbi, Jacopo Rizzo, Francesco Tacchino, Nicola Marzari, and Ivano Tavernelli. 2022. Effective Calculation of the Green’s Function in the Time Domain on Near-Term Quantum Processors.Physical Review Research4, 4 (2022), 043038. doi:10.1103/PhysRevResearch.4.043038

    work page doi:10.1103/physrevresearch.4.043038 2022

  47. [47]

    Guang Hao Low and Isaac L. Chuang. 2017. Optimal Hamiltonian Simulation by Quantum Signal Processing.Physical Review Letters118, 1 (2017), 010501. doi:10.1103/PhysRevLett.118.010501

    doi:10.1103/physrevlett.118.010501 2017

  48. [48]

    Guang Hao Low and Isaac L. Chuang. 2019. Hamiltonian Simulation by Qubitization.Quantum3 (2019), 163. doi:10.22331/q-2019-07-12-163

    doi:10.22331/q-2019-07-12-163 2019

  49. [49]

    Baldwin, Karl Mayer, and Timothy Proctor

    Thomas Lubinski, Sonika Johri, Paul Varosy, Jeremiah Coleman, Luning Zhao, Jason Necaise, Charles H. Baldwin, Karl Mayer, and Timothy Proctor. 2023. Application-Oriented Performance Benchmarks for Quantum Computing.IEEE Transactions on Quantum Engineering4 (2023), 1–35. doi:10.1109/TQE.2023.3253761

    arXiv 2023

  50. [50]

    Scott M Lundberg and Su-In Lee. 2017. A unified approach to interpreting model predictions.Advances in neural information processing systems30 (2017)

    2017

  51. [51]

    Gambetta, and Joseph Emerson

    Easwar Magesan, Jay M. Gambetta, and Joseph Emerson. 2011. Scalable and Robust Randomized Benchmarking of Quantum Processes.Physical Review Letters106, 18 (2011), 180504. doi:10.1103/PhysRevLett.106.180504

    doi:10.1103/physrevlett.106.180504 2011

  52. [52]

    Gerald D. Mahan. 2000.Many-Particle Physics(3 ed.). Springer

    2000

  53. [53]

    Martyn, Zane M

    John M. Martyn, Zane M. Rossi, Andrew K. Tan, and Isaac L. Chuang. 2021. Grand Unification of Quantum Algorithms. PRX Quantum2, 4 (2021), 040203. doi:10.1103/PRXQuantum.2.040203

    doi:10.1103/prxquantum.2.040203 2021

  54. [54]

    McClean, Jonathan Romero, Ryan Babbush, and Alán Aspuru-Guzik

    Jarrod R. McClean, Jonathan Romero, Ryan Babbush, and Alán Aspuru-Guzik. 2016. The Theory of Variational Hybrid Quantum-Classical Algorithms.New Journal of Physics18, 2 (2016), 023023. doi:10.1088/1367-2630/18/2/023023

    doi:10.1088/1367-2630/18/2/023023 2016

  55. [55]

    Johannes Jakob Meyer, Johannes Borregaard, and Jens Eisert. 2021. A Variational Toolbox for Quantum Multi-Parameter Estimation.npj Quantum Information7 (2021), 89. doi:10.1038/s41534-021-00425-y

    work page doi:10.1038/s41534-021-00425-y 2021

  56. [56]

    Mitarai, M

    K. Mitarai, M. Negoro, M. Kitagawa, and K. Fujii. 2018. Quantum Circuit Learning.Physical Review A98, 3 (2018), 032309. doi:10.1103/PhysRevA.98.032309

    doi:10.1103/physreva.98.032309 2018

  57. [57]

    Mooney, Gregory A

    Gary J. Mooney, Gregory A. L. White, Charles D. Hill, and Lloyd C. L. Hollenberg. 2021. Whole-Device Entanglement in a 65-Qubit Superconducting Quantum Computer.Advanced Quantum Technologies4, 7 (2021), 2100061. doi:10.1002/ qute.202100061

    2021

  58. [58]

    Max D. Morris. 1991. Factorial Sampling Plans for Preliminary Computational Experiments.Technometrics33, 2 (1991), 161–174. doi:10.1080/00401706.1991.10484804

    arXiv 1991

  59. [59]

    Rick Mukherjee, Frédéric Sauvage, Harry Xie, Robert Löw, and Florian Mintert. 2020. Preparation of ordered states in ultra-cold gases using Bayesian optimization.New Journal of Physics22, 7 (2020), 075001. doi:10.1088/1367-2630/ab8677

    work page doi:10.1088/1367-2630/ab8677 2020

  60. [60]

    Erik Nielsen, John King Gamble, Kenneth Rudinger, Travis Scholten, Kevin Young, and Robin Blume-Kohout. 2021. Gate Set Tomography.Quantum5 (2021), 557. doi:10.22331/q-2021-10-05-557

    doi:10.22331/q-2021-10-05-557 2021

  61. [61]

    Love, Alán Aspuru-Guzik, and Jeremy L

    Alberto Peruzzo, Jarrod McClean, Peter Shadbolt, Man-Hong Yung, Xiao-Qi Zhou, Peter J. Love, Alán Aspuru-Guzik, and Jeremy L. O’Brien. 2014. A Variational Eigenvalue Solver on a Photonic Quantum Processor.Nature Communications 5 (2014), 4213. doi:10.1038/ncomms5213

    doi:10.1038/ncomms5213 2014

  62. [62]

    Timothy Proctor, Stefan Seritan, Kenneth Rudinger, Erik Nielsen, Robin Blume-Kohout, and Kevin Young. 2022. Scalable Randomized Benchmarking of Quantum Computers Using Mirror Circuits.Physical Review Letters129, 15 (2022), 150502. doi:10.1103/PhysRevLett.129.150502

    work page doi:10.1103/physrevlett.129.150502 2022

  63. [63]

    Carl Edward Rasmussen and Christopher K. I. Williams. 2006.Gaussian Processes for Machine Learning. MIT Press

    2006

  64. [64]

    Ollitrault, Nicola Marzari, and Ivano Tavernelli

    Jacopo Rizzo, Francesco Libbi, Francesco Tacchino, Pauline J. Ollitrault, Nicola Marzari, and Ivano Tavernelli. 2022. One-Particle Green’s Functions from the Quantum Equation of Motion Algorithm.Physical Review Research4, 4 (2022), 043011. doi:10.1103/PhysRevResearch.4.043011

    work page doi:10.1103/physrevresearch.4.043011 2022

  65. [65]

    Andrea Saltelli, Paola Annoni, Ivano Azzini, Francesca Campolongo, Marco Ratto, and Stefano Tarantola. 2010. Variance Based Sensitivity Analysis of Model Output. Design and Estimator for the Total Sensitivity Index.Computer Physics Communications181, 2 (2010), 259–270. doi:10.1016/j.cpc.2009.09.018 25

    doi:10.1016/j.cpc.2009.09.018 2010

  66. [66]

    Yuki Sato, Ruho Kondo, Satoshi Koide, Hideki Takamatsu, and Nobuyuki Imoto. 2021. Variational Quantum Algorithm Based on the Minimum Potential Energy for Solving the Poisson Equation.Physical Review A104, 5 (2021), 052409. doi:10.1103/PhysRevA.104.052409

    doi:10.1103/physreva.104.052409 2021

  67. [67]

    Frédéric Sauvage and Florian Mintert. 2020. Optimal Quantum Control with Poor Statistics.PRX Quantum1, 2 (2020), 020322. doi:10.1103/PRXQuantum.1.020322

    work page doi:10.1103/prxquantum.1.020322 2020

  68. [68]

    Maria Schuld, Ryan Sweke, and Johannes Jakob Meyer. 2021. Effect of Data Encoding on the Expressive Power of Variational Quantum-Machine-Learning Models.Physical Review A103, 3 (2021), 032430. doi:10.1103/PhysRevA.103. 032430

    doi:10.1103/physreva.103 2021

  69. [69]

    Athawale, David Pugmire, and Qiang Guan

    Priyabrata Senapati, Tushar M. Athawale, David Pugmire, and Qiang Guan. 2023. Advancing Comprehension of Quantum Application Outputs: A Visualization Technique. InQCCC 2023 - Proceedings of the 2023 International Workshop on Quantum Classical Cooperative Computing. 25–28. doi:10.1145/3588983.3596689

    arXiv 2023

  70. [70]

    Athawale, Ang Li, Weiwen Jiang, Zhepeng Wang, Qiang Guan, and Cheng Chang Lu

    Priyabrata Senapati, Samuel Yen-Chi Chen, Bo Fang, Tushar M. Athawale, Ang Li, Weiwen Jiang, Zhepeng Wang, Qiang Guan, and Cheng Chang Lu. 2024. PQML: Enabling the Predictive Reproducibility on NISQ Machines for Quantum ML Applications. InProceedings of the 2024 IEEE International Conference on Quantum Computing and Engineering (QCE). 1413–1424. doi:10.11...

    arXiv 2024

  71. [71]

    Athawale

    Priyabrata Senapati, Qiang Guan, David Pugmire, Cheng Chang Lu, and Tushar M. Athawale. 2025. Visualization of Noisy and Less Noisy Computational Basis States in Quantum Computing. InProceedings of the 2025 IEEE International Conference on Quantum Software (QSW). 12–21. doi:10.1109/QSW67625.2025.00011

    arXiv 2025

  72. [72]

    Priyabrata Senapati, Waylon Luo, Bo Peng, and Qiang Guan. 2026. Uncertainty Quantification Driven Benchmarking and Characterization of Noisy Quantum Backends: A VQE Case Study. InProceedings of the Great Lakes Symposium on VLSI 2026 (GLSVLSI ’26). Association for Computing Machinery, New York, NY, USA, 830–835. doi:10.1145/3787109. 3816378

    work page doi:10.1145/3787109 2026

  73. [73]

    Humble, Bo Fang, Shuai Xu, and Qiang Guan

    Priyabrata Senapati, Zhepeng Wang, Weiwen Jiang, Travis S. Humble, Bo Fang, Shuai Xu, and Qiang Guan. 2023. Towards Redefining the Reproducibility in Quantum Computing: A Data Analysis Approach on NISQ Devices. InProceedings of the 2023 IEEE International Conference on Quantum Computing and Engineering (QCE). 468–474. doi:10.1109/QCE57702.2023.00060

    arXiv 2023

  74. [74]

    Priyabrata Senapati, Shengye Zhu, Bo Peng, Bo Fang, and Qiang Guan. 2025. UQ-VarQA: Benchmarking and Charac- terizing NISQ Computers through Uncertainty Quantification of Variational Quantum Algorithms. InProceedings of the 2025 IEEE 43rd International Conference on Computer Design (ICCD). 356–363. doi:10.1109/ICCD65941.2025.00058

    arXiv 2025

  75. [75]

    Adams, and Nando de Freitas

    Bobak Shahriari, Kevin Swersky, Ziyu Wang, Ryan P. Adams, and Nando de Freitas. 2016. Taking the Human Out of the Loop: A Review of Bayesian Optimization.Proc. IEEE104, 1 (2016), 148–175. doi:10.1109/JPROC.2015.2494218

    arXiv 2016

  76. [76]

    B. W. Silverman. 1986.Density Estimation for Statistics and Data Analysis. Chapman & Hall

    1986

  77. [77]

    Jasper Snoek, Hugo Larochelle, and Ryan P. Adams. 2012. Practical Bayesian Optimization of Machine Learning Algorithms. InAdvances in Neural Information Processing Systems

    2012

  78. [78]

    Sullivan

    Timothy J. Sullivan. 2015.Introduction to Uncertainty Quantification. Springer

    2015

  79. [79]

    Booth, and Jonathan Tennyson

    Jules Tilly, Hongxiang Chen, Shuxiang Cao, Dario Picozzi, Kanav Setia, Ying Li, Edward Grant, Leonard Wossnig, Ivan Rungger, George H. Booth, and Jonathan Tennyson. 2022. The Variational Quantum Eigensolver: A Review of Methods and Best Practices.Physics Reports986 (2022), 1–128. doi:10.1016/j.physrep.2022.08.003

    doi:10.1016/j.physrep.2022.08.003 2022

  80. [80]

    Smith, Joshua Viszlai, Xin Chuan Wu, Nikos Hardavellas, Margaret R

    Teague Tomesh, Pranav Gokhale, Victory Omole, Gokul Subramanian Ravi, Kaitlin N. Smith, Joshua Viszlai, Xin Chuan Wu, Nikos Hardavellas, Margaret R. Martonosi, and Frederic T. Chong. 2022. SupermarQ: A Scalable Quantum Benchmark Suite. InProceedings of the 2022 IEEE International Symposium on High-Performance Computer Architecture (HPCA). 587–603. doi:10....

    arXiv 2022

Showing first 80 references.