Pith. sign in

REVIEW 6 major objections 6 minor 1 cited by

Distribution-Adaptive Dynamic Shot Optimization for Variational Quantum Algorithms

T0 review · 6 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A variational quantum algorithm can cut its total measurement shots by roughly half by deriving each round's shot budget from the entropy of the previous round's output distribution.

desk verdict A plausible entropy-feedback heuristic whose headline shot savings are fitted numbers, not predictions—worth a careful referee but not acceptance as is. read the letter →

arxiv 2412.17485 v1 pith:KZ5SU4WM submitted 2024-12-23 quant-ph

Youngmin Kim , Enhyeok Jang , Hyungseok Kim , Seungwoo Choi , Changhun Lee , Donghwi Kim , Woomin Kyoung , Kyujin Shin
show 1 more author
Won Woo Ro
This is my paper · ORCID
classification quant-ph
keywords variationalquantumalgorithmsdynamicshotallocationinformationentropyHellingerdistancemeasurementreductionQAOAVQEresourceestimation
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

The paper tries to establish a cheap feedback rule for variational quantum algorithms (VQAs): spend measurement shots according to how concentrated the circuit's output distribution has become, rather than fixing the same budget every round. It reports that the shots needed to approximate a distribution to a target Hellinger distance grow roughly exponentially with the distribution's information entropy. On that basis it proposes DDS, which sets the next round's shot count from the entropy of the previous round's measured outcomes, capped at the conventional 1,024-shot budget. In simulations of QAOA and VQE, the rule is claimed to cut total shots by roughly half while keeping final accuracy close to the fixed-shot baseline, and to beat tiered schedules that cut shots without looking at the state. The practical payoff, if the claim holds, is cheaper VQA training on near-term hardware without redesigning the optimizer.

What carries the argument

The load-bearing object is the mapping $S = k \times 2^{H}$, a calibration curve that the paper fits to static random-number-generation circuits: for a target Hellinger distance of 0.05, the required shots grow exponentially with the information entropy $H$ of the output distribution. In training, DDS estimates $H$ from the counts of the previous iteration and substitutes it into this curve, with $k$ chosen by qubit count and circuit complexity. The mechanism works because VQA output distributions typically start near-uniform, which means high entropy and many shots, and become concentrated on a few high-probability states as parameters converge, which means low entropy and few shots.

What would settle it

Run DDS on a VQA while also computing the exact output distribution by statevector simulation, and compare the entropy estimated from the finite shot sample with the true entropy at the low shot counts DDS actually assigns; if the estimated entropy systematically understates the true entropy, part of the reported saving is an artifact of the estimator's bias. The sharper test is to drive the same shot schedule with exact statevector entropy and check whether the roughly 50% shot saving and the accuracy changes survive.

Watch

Extended reading notes

Core claim

The paper's central claim is that the number of measurement shots a VQA needs in a given training round tracks the information entropy of the circuit's output distribution through an approximately exponential relation $S = k \times 2^{H}$, where $k$ is a per-circuit constant. DDS computes the empirical entropy of the previous round's measurement outcomes and feeds it into this relation to set the next round's shot count, starting from a default of 1,024 shots and never exceeding it. The reported results are that this rule cuts the total shot budget by 50.69% on average for QAOA and 47.09% for VQE, with the approximation-ratio gap only 6.03% higher than the fixed-shot baseline on QAOA and 24.39% lower on VQE; under simulated device noise the saving is about 31% with minimal accuracy loss. Compared with tiered schedules that simply reduce shots from 1,000 downward, DDS uses more shots but converges to substantially lower final costs, improving the approximation-ratio gap by 63.34% on QAOA and 37.78% on VQE.

Load-bearing premise

The method assumes that the entropy computed from one round's finite shot sample reliably predicts how many shots the next round genuinely needs, and that the exponential calibration measured on static random circuits carries over unchanged to VQA circuits.

Editorial extensions

If this is right

  • On four QAOA graph models, DDS cuts total shots by an average of 50.69% while increasing the approximation-ratio gap by only 6.03% relative to the 1,024-shot fixed baseline.
  • On VQE molecular benchmarks, DDS cuts total shots by 47.09% and lowers the approximation-ratio gap by 24.39% relative to the fixed baseline.
  • Under noise models with real-device gate-error rates, DDS still reduces total shots by about 31.15%, and with the shot cap removed it uses 4.30% fewer shots than the fixed baseline while improving accuracy by 18.05%.
  • DDS leaves the total number of training iterations essentially unchanged, so the savings come from per-iteration measurement cost rather than from stopping early.
  • Against tiered linear- and step-function schedules, DDS trades additional shots for much better final accuracy: 63.34% ARG improvement on QAOA, 37.78% on VQE, and roughly 68% in the noisy setting.

Reading between the lines

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

  • Editorial: reading the shot rule as a fixed oversampling factor over the distribution's effective support size suggests the same entropy feedback could apply to any distributional sampling task, not only VQA training.
  • Editorial: because the reported savings are averages over a small set of circuits and one classical optimizer, a sweep over seeds, optimizers, and larger qubit counts is the natural next test of whether the roughly 50% figure holds.
  • Editorial: the entropy that drives DDS is itself estimated from the shot sample DDS chose, so the schedule and its own estimator form a feedback loop; comparing DDS against a version driven by exact statevector entropy would separate the saving due to the entropy-shots relation from any bias of the finite-sample estimator.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

6 major / 6 minor

Summary. This manuscript proposes distribution-adaptive dynamic shot (DDS), a heuristic that sets the number of measurement shots for each VQA iteration to S = k * 2^H, where H is the Shannon entropy of the empirical output distribution from the previous iteration. The authors motivate the exponential form by simulations of random-number-generation circuits under a Hellinger-distance constraint (Section II.C), and evaluate DDS on QAOA (PL, BA, WS, SK graphs with 4, 8, and 12 qubits) and VQE (H2, LiH, BeH2) in noiseless and noise-injected simulators, comparing to fixed 1,024-shot, linear, and step baselines. They report an average shot reduction of about 50% for QAOA and 47% for VQE at roughly unchanged ARG, and large ARG improvements over tiered allocation baselines.

Significance. If the central claim were established, DDS would be an attractive plug-in shot-reduction scheme for VQAs because it is optimizer-agnostic and requires only the entropy of the previous iteration's counts. The paper also provides a useful empirical observation that required shot counts grow roughly exponentially with output-distribution entropy. However, the reported quantitative savings are not yet established as a predictive property: the constant k is selected per benchmark, the entropy feedback from finite samples is not analyzed, and the experimental record contains internal inconsistencies and at least one data error. The paper does not ship code, proofs, or a fixed decision rule that would allow an independent check of the headline numbers.

major comments (6)
  1. [Section IV.B.3] The text states that 'For each specific QAOA model and qubit count, we configured the constant k to identify an optimal value' using Figure 15. The headline 50.69% shot reduction and 6.03% ARG increase are therefore the best in-sample outcomes of a per-benchmark parameter search, not the behavior of a fixed adaptive rule. A valid test would fix k, or choose it on separate validation data, and report held-out results with multiple random seeds and graph instances.
  2. [Section III.B, Eq. (4)] The H used in Eq. (4) is the plug-in entropy of a finite-shot histogram whose size S was itself set by Eq. (4). Plug-in entropy is biased downward, and the bias grows as S shrinks; Eq. (4) exponentiates this bias, so the feedback loop can drive shot counts below what the true distribution would require. The paper provides no analysis of this loop, no quantification of the bias, and no comparison between the empirical H values observed during training and the exact-distribution calibration of Section II.C.
  3. [Section II.C and Section IV.B.3] The calibration in Section II.C targets a Hellinger distance of 0.05, but the k values actually used (64, 8, and 2 for 4, 8, and 12 qubits) are not connected to that target. For a 12-qubit uniform state, k=2 yields S=8192, whereas achieving Hd=0.05 for that distribution requires on the order of 2*10^5 shots. The paper does not explain why the Hd=0.05 calibration justifies the much smaller shot counts used in the experiments.
  4. [Table I] The rows for 4-qubit BA QAOA and 4-qubit WS QAOA are identical in Savg, Iterations, and Average Cuts. This is almost certainly a copy error; the affected entries cannot both be correct, and the aggregate 50.69% reduction figure depends on these rows.
  5. [Section IV.B.1 and Tables I-II] All results are single runs with no seeds, repetitions, or error bars, and the random graph ensembles are not specified. Given that COBYLA and shot noise are stochastic, the claimed 6.03% ARG difference from the standard method is within the range where run-to-run variation could change the conclusion.
  6. [Section III.B vs. Section IV.B.3] The parameter guidance is internally inconsistent: Section III.B says larger qubit counts should use a larger k, while Section IV.B.3 says larger entropy requires a smaller k, and the experiments use k=64, 8, and 2 for increasing qubit counts. The k values used for VQE are not reported at all, so the VQE results in Table II are not reproducible from the text.
minor comments (6)
  1. [Algorithm 1] The pseudocode initializes prev_entropy to 10, which exceeds the maximum entropy of small circuits, and it omits the 1,024-shot upper bound mentioned in Section IV.B.1; the algorithm as written is therefore not a complete specification of the evaluated method.
  2. [Appendix C] The formula Hrand = log2 n should be Hrand = n for an n-qubit uniform distribution, unless n is redefined as the number of outcomes rather than the number of qubits.
  3. [Section IV.B.2 and Conclusion] An increase in ARG of 26.92% is described in the conclusion as 'maintaining a slight increase in accuracy'; since ARG is an error metric, an increase in ARG is a decrease in accuracy, and the wording should be corrected.
  4. [Section IV.B.1] The text states that for VQE the DDS shot count increases as the entropy of the distribution grows, which contradicts the motivating claim in Section III.A that entropy generally decreases during VQA training; this tension needs to be resolved.
  5. [Section II.B, Eq. (2)] The Hellinger distance definition has typesetting errors in the PDF text; the formula should be checked so that the square root and normalization are unambiguous.
  6. [Section V] The statements that DDS is directly applicable to fault-tolerant quantum computers and that it will reduce costs in cloud-based quantum computing are speculative and not supported by the presented experiments.

Circularity Check

1 steps flagged · score 6.0 of 10

The reported ~50% shot savings are in-sample k-tuning results, not predictions of a fixed entropy-based rule.

  1. fitted input called prediction [Section IV.B.3, 'Dynamic Shots on Entropy' (page 9); Eq. (4) in Section III.B]
    "In the QAOA experiments presented above, we set this constant to 64 for 4-qubit circuits, 8 for 8-qubit circuits, and 2 for 12-qubit circuits. The constant value can be adjusted depending on both the number of qubits and the complexity of the quantum circuit. For each specific QAOA model and qubit count, we configured the constant k to identify an optimal value."

    The central shot-reduction claim relies on Eq. (4), S = k * 2^H, which has a free constant k. In evaluation, k is not fixed by the entropy-calibration argument in Section II.C; instead it is tuned per benchmark by sweeping powers of two in Figure 15 and selecting an optimal value on the same QAOA models whose aggregate 50.69% shot reduction and 6.03% ARG increase are then reported as DDS's outcome. The headline savings are in-sample selected values from the k sweep, not predictions of a single fixed adaptive rule. The same tuned k values also determine the entropy-feedback trajectory in Algorithm 1, so the reported performance is forced by the manual per-benchmark choice of k rather than derived from the claimed entropy-shot relationship.

full rationale

The paper is self-contained against external benchmarks, and there are no load-bearing self-citations: the only same-author reference is an unrelated trapped-ion molecule simulation, and no uniqueness theorem or prior-work ansatz is invoked to justify DDS. The exponential relation in Eq. (4) is itself an empirical fit, which is a stated observation rather than a first-principles derivation; that alone is not circular. The circularity enters at evaluation: k is selected per QAOA model and qubit count by sweeping values and choosing an optimal configuration on the same benchmarks whose averaged 50.69% shot reduction is later presented as the central result. This is a fitted-input-called-prediction step because the reported savings are in-sample optima rather than out-of-sample predictions of a fixed method. I do not count the finite-sample entropy feedback loop (Algorithm 1 computes H from the previous iteration's shots) as circular, because it is a stability/bias concern rather than a definitional equivalence; similarly, the common final 1,024-shot evaluation for all methods is a fair-comparison design, not circularity. Overall score 6: one fitted step affects the central quantitive claim, but the method has independent content in its dynamic feedback and benchmark comparisons against fixed-shot and tiered baselines.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests on one per-benchmark fitted constant k, hand-set initial entropy and shot caps, and a calibration threshold that is not enforced in the deployed method. In addition, five domain assumptions about entropy as a feedback signal, temporal smoothness, transfer from random circuits, and unbiased entropy estimation are invoked without proof. No new physical entities are introduced.

free parameters (4)
  • k (shot-scaling constant in S = k * 2^H) = 64 (4-qubit QAOA), 8 (8-qubit QAOA), 2 (12-qubit QAOA); VQE values not reported
    Chosen per benchmark to 'identify an optimal value' (Section IV.B.3), so reported shot savings are conditional on per-benchmark fitting.
  • Initial entropy value in Algorithm 1 = 10
    Hand-set default in Algorithm 1, line 2; affects only the first iteration and is capped at the maximum shot bound.
  • Maximum shot bound = 1,024 (DDS) or 100,000 (DDS M)
    Implementation cap that directly bounds DDS shot usage; the DDS M variant shows results depend on this choice.
  • Target Hellinger distance threshold = 0.05
    Used to generate the calibration curve in Figure 3; DDS never enforces this threshold, and the k values used imply much larger Hellinger distances, so the calibration is not actually applied.
assumptions (5)
  • standard math Shannon entropy and Hellinger distance are suitable measures of distribution accuracy for VQA measurement.
    Definitions in Sections II.A and II.B; these are standard mathematical tools, but the choice of Hellinger distance to govern cost-function accuracy is an application choice.
  • domain assumption Entropy of the output distribution is a sufficient feedback signal for setting the next iteration's shot count.
    Stated in Sections I and III as the basis of DDS; the paper gives no derivation that entropy captures all sampling difficulty relevant to VQA cost estimation.
  • domain assumption Consecutive VQA iterations have similar output distributions, so the previous iteration's entropy predicts the current one.
    Section III.B: 'consecutive iterations typically exhibit similar distributions'; smoothness is assumed, not verified per benchmark.
  • domain assumption The entropy-to-shot calibration measured on random number generation circuits transfers to QAOA and VQE circuits and to cost estimation.
    Section II.C uses random circuits (QASMBench) to establish the exponential relation; Section III applies it to VQA without a transfer argument.
  • domain assumption Empirical entropy computed from a finite shot sample is a reliable estimate of the true output-distribution entropy.
    Used throughout Section III.B; finite-sample entropy is biased downward, especially at low shot counts, and the paper does not correct or analyze this.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Distribution-Adaptive Dynamic Shot Optimization for Variational Quantum Algorithms." pith.science (2026). https://pith.science/paper/KZ5SU4WM

@misc{pith2026241217485,
  author       = {Pith},
  title        = {Pith review of: Distribution-Adaptive Dynamic Shot Optimization for Variational Quantum Algorithms},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KZ5SU4WM}},
  note         = {Machine review of arXiv:2412.17485}
}
read the original abstract

Variational quantum algorithms (VQAs) have attracted remarkable interest over the past few years because of their potential computational advantages on near-term quantum devices. They leverage a hybrid approach that integrates classical and quantum computing resources to solve high-dimensional problems that are challenging for classical approaches alone. In the training process of variational circuits, constructing an accurate probability distribution for each epoch is not always necessary, creating opportunities to reduce computational costs through shot reduction. However, existing shot-allocation methods that capitalize on this potential often lack adaptive feedback or are tied to specific classical optimizers, which limits their applicability to common VQAs and broader optimization techniques. Our observations indicate that the information entropy of a quantum circuit's output distribution exhibits an approximately exponential relationship with the number of shots needed to achieve a target Hellinger distance. In this work, we propose a distribution-adaptive dynamic shot (DDS) framework that efficiently adjusts the number of shots per iteration in VQAs using the entropy distribution from the prior training epoch. Our results demonstrate that the DDS framework sustains inference accuracy while achieving a ~50% reduction in average shot count compared to fixed-shot training, and ~60% higher accuracy than recently proposed tiered shot allocation methods. Furthermore, in noisy simulations that reflect the error rates of actual IBM quantum systems, DDS achieves approximately a ~30% reduction in the total number of shots compared to the fixed-shot method with minimal degradation in accuracy, and offers about ~70% higher computational accuracy than tiered shot allocation methods.

Figures

Figures reproduced from arXiv: 2412.17485 by the authors.

Figure 1
Figure 1. FIG. 1. The required number of shots to obtain the probabil [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Hellinger distance of random number generation [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Required number of shots with Hellinger distance [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (16 more)
Figure 5
Figure 5. Figure 5: FIG. 5. The number of shots and average cuts per iteration [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. The number of shots and average cuts per iteration [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 4
Figure 4. Figure 4: FIG. 4. The number of shots and average cuts per iteration [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 8
Figure 8. Figure 8: FIG. 8. Total shot counts required for the training of various [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]
Figure 7
Figure 7. Figure 7: FIG. 7. The number of shots and average cuts per iteration [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 11
Figure 11. Figure 11: FIG. 11. Total shot counts required for the training of various [PITH_FULL_IMAGE:figures/full_fig_p008_11.png]
Figure 10
Figure 10. Figure 10: FIG. 10. The number of shots and the energy of each iteration [PITH_FULL_IMAGE:figures/full_fig_p008_10.png]
Figure 13
Figure 13. Figure 13: FIG. 13. Total shot counts required for the training of Sher [PITH_FULL_IMAGE:figures/full_fig_p009_13.png]
Figure 14
Figure 14. Figure 14: FIG. 14. The ARG values for the training of Sherring [PITH_FULL_IMAGE:figures/full_fig_p009_14.png]
Figure 15
Figure 15. Figure 15: FIG. 15. The average cuts and total shots for the QAOA [PITH_FULL_IMAGE:figures/full_fig_p010_15.png]
Figure 16
Figure 16. Figure 16: FIG. 16. The average cuts error per iteration against standard [PITH_FULL_IMAGE:figures/full_fig_p015_16.png]
Figure 17
Figure 17. Figure 17: FIG. 17. The average cuts error per iteration against stan [PITH_FULL_IMAGE:figures/full_fig_p015_17.png]
Figure 20
Figure 20. Figure 20: FIG. 20. The energy error per iteration against standard shot [PITH_FULL_IMAGE:figures/full_fig_p016_20.png]
Figure 22
Figure 22. Figure 22: FIG. 22. The average cuts error per iteration against the [PITH_FULL_IMAGE:figures/full_fig_p016_22.png]
Figure 23
Figure 23. Figure 23: FIG. 23. The number of shots and average cuts per itera [PITH_FULL_IMAGE:figures/full_fig_p018_23.png]
Figure 24
Figure 24. Figure 24: FIG. 24. The average cuts error per iteration against the [PITH_FULL_IMAGE:figures/full_fig_p018_24.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Shot-Efficient ADAPT-VQE via Reused Pauli Measurements and Variance-Based Shot Allocation

    quant-ph 2025-07 conditional novelty 4.0 of 10

    A shot-efficient ADAPT-VQE variant that reuses grouped Pauli measurements from VQE optimization for gradient estimation and adds variance-based shot allocation reaches chemical accuracy with fewer measurements in smal...

Reference graph

Works this paper leans on

58 extracted references · 28 canonical work pages · cited by 1 Pith paper

  1. [1]

    The linear shot func- tion [30], starting from 1,000 shots (close to the de- fault of 1,024), gradually reduces the shot count over training iterations

    Scenarios The fixed-shot method uses 1,024 shots as its config- uration, aligning with the default setting in many exist- ing quantum computing systems. The linear shot func- tion [30], starting from 1,000 shots (close to the de- fault of 1,024), gradually reduces the shot count over training iterations. This function is defined as Si = max(20, Sbegin − l...

    work page

  2. [2]

    Benchmarks and Backends To evaluate the performance of the DDS approach, quantum circuits for both the Quantum Approxi- mate Optimization Algorithm (QAOA) and the Vari- ational Quantum Eigensolver (VQE) were constructed. For QAOA, we selected target graphs commonly used in network science, specifically Power-Law (PL), Barab´ asi–Albert (BA), Watts–Strogat...

    work page 2024

  3. [3]

    (5) Eideal is the ideal expectation value of the quantum state, and Ereal is the calculated expectation value

    Metrics To measure the fidelity of VQA circuits, Approxima- tion Ratio Gap (ARG) [31] was used, defined as follows: ARG = 100 × Eideal − Ereal Eideal . (5) Eideal is the ideal expectation value of the quantum state, and Ereal is the calculated expectation value. When the 6 fidelity is high, the measured expectation value is closer to the ideal value, resu...

    work page

  4. [4]

    lin fn” and “step fn

    Evaluation on Scalable VQAs Dynamic shot allocation was applied to both QAOA and VQE models for several molecules. We tested four QAOA models and three VQE models, running each with the standard fixed shot method of 1,024 shots, as well as with two existing shot reduction methods: the lin- ear function and step function-based shot allocation ap- proaches ...

    work page

  5. [5]

    However, due to the pres- ence of noise in real quantum hardware, the number of shots is sometimes increased to ensure more reliable re- sults [37]

    Evaluation under Noise Condition On actual quantum devices, experiments are typically conducted with a default configuration of 1,024 shots [36], similar to quantum simulators. However, due to the pres- ence of noise in real quantum hardware, the number of shots is sometimes increased to ensure more reliable re- sults [37]. To simulate experiments under n...

    work page

  6. [6]

    As the number of qubits increases, entropy also increases, often requiring a smaller constant value to maintain the desired shot count

    Dynamic Shots on Entropy As discussed in Section-II C, determining the number of shots based on entropy requires a constant parameter. As the number of qubits increases, entropy also increases, often requiring a smaller constant value to maintain the desired shot count. In the QAOA experiments presented above, we set this constant to 64 for 4-qubit circui...

    work page

  7. [7]

    P. W. Shor, Polynomial-time algorithms for prime factor- ization and discrete logarithms on a quantum computer, SIAM review 41, 303 (1999)

    work page 1999

  8. [8]

    L. K. Grover, A fast quantum mechanical algorithm for database search, in Proceedings of the twenty-eighth an- nual ACM symposium on Theory of computing (1996) pp. 212–219

    1996

Show all 58 references
  1. [9]

    Arute, K

    F. Arute, K. Arya, R. Babbush, D. Bacon, J. C. Bardin, R. Barends, R. Biswas, S. Boixo, F. G. Brandao, D. A. Buell, et al., Quantum supremacy using a programmable superconducting processor, Nature 574, 505 (2019)

    2019

  2. [10]

    M. A. Nielsen and I. L. Chuang, Quantum computation and quantum information (Cambridge university press, 2010)

    2010

  3. [11]

    Farhi and A

    E. Farhi and A. W. Harrow, Quantum supremacy through the quantum approximate optimization algo- rithm, arXiv preprint arXiv:1602.07674 (2016)

    2016 arXiv

  4. [12]

    Jozsa, Quantum algorithms and the fourier transform, Proceedings of the Royal Society of London

    R. Jozsa, Quantum algorithms and the fourier transform, Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences 454, 323 (1998)

    1998

  5. [13]

    Cerezo, A

    M. Cerezo, A. Arrasmith, R. Babbush, S. C. Benjamin, S. Endo, K. Fujii, J. R. McClean, K. Mitarai, X. Yuan, L. Cincio, et al., Variational quantum algorithms, Nature Reviews Physics 3, 625 (2021)

    2021

  6. [14]

    Biamonte, P

    J. Biamonte, P. Wittek, N. Pancotti, P. Rebentrost, N. Wiebe, and S. Lloyd, Quantum machine learning, Na- ture 549, 195 (2017)

    2017

  7. [15]

    Rebentrost, M

    P. Rebentrost, M. Mohseni, and S. Lloyd, Quantum sup- port vector machine for big data classification, Physical Review Letters 113, 130503 (2014)

    2014

  8. [16]

    J. R. McClean, J. Romero, R. Babbush, and A. Aspuru- Guzik, The theory of variational hybrid quantum- classical algorithms, New Journal of Physics 18, 023023 (2016)

    2016

  9. [17]

    Havl ´ ıˇ cek, A

    V. Havl ´ ıˇ cek, A. D. C´ orcoles, K. Temme, A. W. Harrow, A. Kandala, J. M. Chow, and J. M. Gambetta, Super- vised learning with quantum-enhanced feature spaces, Nature 567, 209 (2019)

    2019

  10. [18]

    Preskill, Quantum computing in the nisq era and be- yond, Quantum 2, 79 (2018)

    J. Preskill, Quantum computing in the nisq era and be- yond, Quantum 2, 79 (2018)

    2018

  11. [19]

    Tilly, H

    J. Tilly, H. Chen, S. Cao, D. Picozzi, K. Setia, Y. Li, E. Grant, L. Wossnig, I. Rungger, G. H. Booth, et al., The variational quantum eigensolver: a review of meth- ods and best practices, Physics Reports 986, 1 (2022)

    2022

  12. [20]

    Ito, Latency-aware adaptive shot allocation for run- time efficient variational quantum algorithms, arXiv preprint arXiv:2302.04422 (2023)

    K. Ito, Latency-aware adaptive shot allocation for run- time efficient variational quantum algorithms, arXiv preprint arXiv:2302.04422 (2023)

    2023 arXiv

  13. [21]

    Ito and K

    K. Ito and K. Fujii, Santaqlaus: A resource-efficient method to leverage quantum shot-noise for optimiza- tion of variational quantum algorithms, arXiv preprint arXiv:2312.15791 (2023)

    2023 arXiv

  14. [22]

    D. Luo, J. Shen, R. Dangovski, and M. Soljacic, Koopman operator learning for accelerating quantum optimization and machine learning, arXiv preprint arXiv:2211.01365 (2022)

    2022 arXiv

  15. [23]

    J. M. K¨ ubler, A. Arrasmith, L. Cincio, and P. J. Coles, An adaptive optimizer for measurement-frugal varia- tional algorithms, Quantum 4, 263 (2020)

    2020

  16. [24]

    Arrasmith, L

    A. Arrasmith, L. Cincio, R. D. Somma, and P. J. Coles, Operator sampling for shot-frugal optimization in variational algorithms, arXiv preprint arXiv:2004.06252 (2020)

    2020 arXiv

  17. [25]

    A. Gu, A. Lowe, P. A. Dub, P. J. Coles, and A. Arrasmith, Adaptive shot allocation for fast conver- gence in variational quantum algorithms, arXiv preprint arXiv:2108.10434 (2021)

    2021 arXiv

  18. [26]

    S. Choi, I. Loaiza, and A. F. Izmaylov, Fluid fermionic fragments for optimizing quantum measurements of elec- tronic hamiltonians in the variational quantum eigen- solver, Quantum 7, 889 (2023)

    2023

  19. [27]

    T.-C. Yen, A. Ganeshram, and A. F. Izmaylov, Deter- ministic improvements of quantum measurements with grouping of compatible operators, non-local transforma- tions, and covariance estimates, npj Quantum Informa- tion 9, 14 (2023)

    2023

  20. [28]

    Choi, T.-C

    S. Choi, T.-C. Yen, and A. F. Izmaylov, Improving quan- tum measurements by introducing “ghost” pauli prod- ucts, Journal of Chemical Theory and Computation 18, 7394 (2022)

    2022

  21. [29]

    Wecker, M

    D. Wecker, M. B. Hastings, and M. Troyer, Progress to- wards practical quantum variational algorithms, Physical Review A 92, 042303 (2015)

    2015

  22. [30]

    Crawford, B

    O. Crawford, B. van Straaten, D. Wang, T. Parks, E. Campbell, and S. Brierley, Efficient quantum measure- ment of pauli operators in the presence of finite sampling error, Quantum 5, 385 (2021)

    2021

  23. [31]

    S. M. Mniszewski, P. A. Dub, S. Tretiak, P. M. Anisimov, Y. Zhang, and C. F. Negre, Reduction of the molecular hamiltonian matrix using quantum community detection, Scientific Reports 11, 4099 (2021)

    2021

  24. [32]

    Zhang, K

    Z.-J. Zhang, K. Nakaji, M. Choi, and A. Aspuru-Guzik, A composite measurement scheme for efficient quantum observable estimation, arXiv preprint arXiv:2305.02439 (2023)

    2023 arXiv

  25. [33]

    L. Zhu, S. Liang, C. Yang, and X. Li, Optimizing shot assignment in variational quantum eigensolver measure- ment, Journal of Chemical Theory and Computation 20, 2390 (2024)

    2024

  26. [34]

    Liang, L

    S. Liang, L. Zhu, X. Liu, C. Yang, and X. Li, Artificial- 12 intelligence-driven shot reduction in quantum measure- ment, arXiv preprint arXiv:2405.02493 (2024)

    2024 arXiv

  27. [35]

    S. S. Kahani and A. Nobakhti, A novel framework for shot number minimization in quantum variational algo- rithms, arXiv preprint arXiv:2307.04035 (2023)

    2023 arXiv

  28. [36]

    Phalak and S

    K. Phalak and S. Ghosh, Shot optimization in quan- tum machine learning architectures to accelerate train- ing, IEEE Access 11, 41514 (2023)

    2023

  29. [37]

    Ayanzadeh, N

    R. Ayanzadeh, N. Alavisamani, P. Das, and M. Qureshi, Frozenqubits: Boosting fidelity of qaoa by skipping hotspot nodes, in Proceedings of the 28th ACM Inter- national Conference on Architectural Support for Pro- gramming Languages and Operating Systems, Volume 2 (2023) pp. 311–324

    2023

  30. [38]

    C. E. Shannon, A mathematical theory of communica- tion, The Bell system technical journal 27, 379 (1948)

    1948

  31. [39]

    D. M. Greenberger, M. A. Horne, and A. Zeilinger, Going beyond bell’s theorem, in Bell’s theorem, quantum theory and conceptions of the universe(Springer, 1989) pp. 69– 72

    1989

  32. [40]

    Hellinger, Neue begr¨ undung der theorie quadratischer formen von unendlichvielen ver¨ anderlichen., Journal f¨ ur die reine und angewandte Mathematik 1909, 210 (1909)

    E. Hellinger, Neue begr¨ undung der theorie quadratischer formen von unendlichvielen ver¨ anderlichen., Journal f¨ ur die reine und angewandte Mathematik 1909, 210 (1909)

    1909

  33. [41]

    A. Li, S. Stein, S. Krishnamoorthy, and J. Ang, Qasm- bench: A low-level quantum benchmark suite for nisq evaluation and simulation, ACM Transactions on Quan- tum Computing 4, 1 (2023)

    2023

  34. [42]

    IBM, Ibm quantum, https://quantum- computing.ibm.com/ (2024)

    2024

  35. [43]

    T. Miki, D. Tsukayama, R. Okita, M. Shimada, and J. Shirakashi, Quantum-classical variational approaches with single-qubit operation on near-term quantum pro- cessors, in 2022 IEEE 22nd International Conference on Nanotechnology (NANO)(IEEE, 2022) pp. 303–306

    2022

  36. [44]

    L. Zhao, Q. Wang, J. J. Goings, K. Shin, W. Kyoung, S. Noh, Y. M. Rhee, and K. Kim, Enhancing the electron pair approximation with measurements on trapped-ion quantum computers, npj Quantum Information 10, 76 (2024)

    2024

  37. [45]

    Eddins, M

    A. Eddins, M. Motta, T. P. Gujarati, S. Bravyi, A. Mez- zacapo, C. Hadfield, and S. Sheldon, Doubling the size of quantum simulators by entanglement forging, PRX Quantum 3, 010309 (2022)

    2022

  38. [46]

    W. J. Huggins, B. A. O’Gorman, N. C. Rubin, D. R. Reichman, R. Babbush, and J. Lee, Unbiasing fermionic quantum monte carlo with a quantum computer, Nature 603, 416 (2022)

    2022

  39. [47]

    T. P. Gujarati, M. Motta, T. N. Friedhoff, J. E. Rice, N. Nguyen, P. K. Barkoutsos, R. J. Thompson, T. Smith, M. Kagele, M. Brei, et al., Quantum computation of re- actions on surfaces using local embedding, npj Quantum Information 9, 88 (2023)

    2023

  40. [48]

    D. Wu, R. Rossi, F. Vicentini, N. Astrakhantsev, F. Becca, X. Cao, J. Carrasquilla, F. Ferrari, A. Georges, M. Hibat-Allah, et al., Variational benchmarks for quan- tum many-body problems, Science 386, 296 (2024)

    2024

  41. [49]

    Cimini, M

    V. Cimini, M. Valeri, S. Piacentini, F. Ceccarelli, G. Cor- rielli, R. Osellame, N. Spagnolo, and F. Sciarrino, Varia- tional quantum algorithm for experimental photonic mul- tiparameter estimation, npj Quantum Information 10, 26 (2024)

    2024

  42. [50]

    M. A. Rahman, M. S. Akter, E. Miller, B. Timofti, H. Shahriar, M. Masum, and F. Wu, Fine-tuned vari- ational quantum classifiers for cyber attacks detection based on parameterized quantum circuits and optimiz- ers, in 2024 IEEE 48th Annual Computers, Software, and Applications...

    2024

  43. [51]

    M. Wang, P. Das, and P. J. Nair, Qoncord: A multi- device job scheduling framework for variational quantum algorithms, in 2024 57th IEEE/ACM International Sym- posium on Microarchitecture (MICRO)(IEEE, 2024) pp. 735–749

    2024

  44. [52]

    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, A variational eigenvalue solver on a photonic quantum processor, Nature Communications 5, 4213 (2014)

    2014

  45. [53]

    Blekos, D

    K. Blekos, D. Brand, A. Ceschini, C.-H. Chou, R.-H. Li, K. Pandya, and A. Summer, A review on quan- tum approximate optimization algorithm and its vari- ants, Physics Reports 1068, 1 (2024)

    2024

  46. [54]

    Zhou, S.-T

    L. Zhou, S.-T. Wang, S. Choi, H. Pichler, and M. D. Lukin, Quantum approximate optimization algorithm: Performance, mechanism, and implementation on near- term devices, Physical Review X 10, 021067 (2020)

    2020

  47. [55]

    Kandala, A

    A. Kandala, A. Mezzacapo, K. Temme, M. Takita, M. Brink, J. M. Chow, and J. M. Gambetta, Hardware- efficient variational quantum eigensolver for small molecules and quantum magnets, Nature 549, 242 (2017)

    2017

  48. [56]

    H. R. Grimsely, S. E. Economou, E. Barnes, and N. J. Mayhall, An adaptive variational algorithm for exact molecular simulations on a quantum computer., Nature Communications 10 (2019)

    2019

  49. [57]

    Std.” method represents a standard shot approach, maintaining a constant 1,024 shots per iteration without adjustments. The “lin fn

    X. Liu, A. Angone, R. Shaydulin, I. Safro, Y. Alexeev, and L. Cincio, Layer vqe: A variational approach for combinatorial optimization on noisy quantum comput- ers, IEEE Transactions on Quantum Engineering 3, 1–20 (2022). Appendix A: Measurement in Quantum Computing In quantum...

    2022

  50. [58]

    For the VQE models, molecular systems including H2, LiH, and BeH 2 were simulated. Across all VQA applications and varying qubit numbers, DDS demon- strated a significant reduction in the required number of shots compared to the standard method, while maintain- ing a similar t...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.