Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

QMoE: A Quantum Mixture of Experts Framework for Scalable Quantum Neural Networks

T0 review · 4 major / 6 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 The paper claims that a mixture-of-experts quantum architecture, with a learnable quantum routing circuit selecting and aggregating specialized parameterized quantum experts per input, consistently beats standard quantum neural networks…

desk verdict Plausible architecture, but the headline accuracy gains look like a capacity effect: QMoE stacks a router plus four expert circuits against single-circuit baselines, so the central comparison needs resource-matched controls. read the letter →

arxiv 2507.05190 v1 pith:BUI2GTDH submitted 2025-07-07 quant-ph cs.CV

Hoang-Quan Nguyen , Xuan-Bac Nguyen , Sankalp Pandey , Samee U. Khan , Ilya Safro , Khoa Luu This is my paper · ORCID
classification quant-phcs.CV
keywords quantummixtureofexpertsmachinelearningparameterizedcircuitsneuralnetworksroutingstateaggregationimageclassificationNISQ
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 a quantum neural network organized as a mixture of experts can classify image data more accurately than standard quantum neural network baselines. The proposed QMoE architecture couples a learnable quantum routing circuit with several parameterized quantum expert circuits, so that each input is routed through a specialized combination of experts and the resulting quantum states are aggregated before measurement. On two-class and four-class tasks built from MNIST and Fashion-MNIST, the authors report consistent gains over the strongest quantum baselines, with the largest improvements on the four-class benchmarks. If the result holds under fair resource comparison, it suggests a modular path to more expressive quantum classifiers that does not require adding qubits or circuit depth to a single network.

What carries the argument

The load-bearing mechanism is the learned quantum routing circuit $G(\theta_G)$ acting together with controlled-expert operations and quantum state aggregation. After encoding, the routing circuit puts routing qubits into a learnable superposition over expert selections; controlled gates then activate the parameterized quantum experts $E_i(\theta_i)$ conditionally on that selection, and a final set of controlled operators aggregates the expert-transformed states into one measured output. Because the whole stack is differentiable and trained end-to-end, the network learns both which experts to create and when to invoke them.

What would settle it

Retrain every baseline and every QMoE configuration on the same four tasks with identical total parameter counts, circuit depth, optimizer settings, and random seeds; if the accuracy gaps reported in Table I shrink to within noise or reverse, the central claim is refuted.

Watch

Extended reading notes

Core claim

The paper's central claim is that applying the classical mixture-of-experts idea inside a quantum circuit yields a quantum classifier that consistently outperforms standard parameterized-quantum-circuit baselines. In the proposed design, a data encoding circuit $U(z)$ maps an input into quantum states shared by a learnable routing circuit $G(\theta_G)$ and $L$ parameterized experts $E_i(\theta_i)$; the routing circuit learns a superposition over expert-selection patterns, controlled operations activate the chosen experts, and a quantum state aggregation step combines their outputs before a projective measurement. The authors report QMoE accuracies of 98.62% on MNIST-2, 62.92% on MNIST-4, 84.46% on Fashion-2, and 63.81% on Fashion-4, compared with best baseline accuracies of 94.89%, 51.47%, 79.94%, and 53.65%, and read the larger four-class gains as evidence that routed expert specialization helps the model learn complex data patterns.

Load-bearing premise

The central comparison only supports the claim if the QMoE models and baseline quantum networks were trained under matched resource budgets—same total trainable parameters, circuit depth, epochs, optimizer settings, and seeds—and the paper does not report those budgets.

Editorial extensions

If this is right

  • If the reported gains persist under matched resource budgets, quantum machine learning can scale by adding specialized expert circuits rather than by deepening a single circuit, which is gentler on NISQ hardware limits.
  • The routing and aggregation loop stays on the quantum side, removing the need for classical control between the router and the experts in this design.
  • The ablation results show accuracy improving as the number of experts grows, making expert count a practical capacity dial for quantum classifiers.
  • Because the framework leaves each expert's internal circuit architecture open, it can wrap improved parameterized-quantum-circuit designs without changing the routing logic.

Reading between the lines

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

  • The paper does not report whether QMoE and its baselines were trained with matching total trainable parameters, circuit depth, epochs, and seeds, so the headline accuracy gaps should be read as conditional on that unstated premise.
  • A direct test of the mechanism would fix total parameter count while varying the number of experts; if specialization drives the gain, accuracy should rise with expert count even at constant capacity.
  • Comparing QMoE with the classically controlled mixture of quantum experts cited as [35] on the same benchmarks would isolate what the quantum-native routing contributes.
  • Because controlled-expert operations and state aggregation are the most hardware-sensitive parts, simulating the framework under realistic noise models would show whether the advantage survives outside exact simulation.
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

4 major / 6 minor

Summary. The paper proposes QMoE, a hybrid quantum-classical architecture that combines a learnable quantum routing circuit G(θ_G) with L parameterized quantum experts E_i(θ_i) and a quantum-state aggregation step. The authors evaluate QMoE on MNIST and Fashion-MNIST binary and 4-class tasks using TorchQuantum simulation, reporting accuracies in Table I and ablations over gate operators and expert counts in Table II. The central claim is that QMoE consistently outperforms standard PQC-based quantum neural networks, with the best reported gains attributed to routing-based expert specialization.

Significance. If the comparative claim is supported, QMoE would be a useful modular design for quantum machine learning: it imports the mixture-of-experts idea into the parameterized-quantum-circuit setting, trains end-to-end with gradient methods, and the ablation over expert counts is a reasonable first study. The authors are to be credited for evaluating four benchmarks and four gate sets, and for making the architecture diagram explicit. However, the current evidence does not yet establish that the gains come from routing and expert specialization rather than from additional trainable parameters; the manuscript also omits the mathematical specification needed to reproduce the routing and aggregation steps. These gaps are fixable and do not invalidate the underlying idea.

major comments (4)
  1. [§V.B / Table I] The central comparison is not resource-matched. The QMoE model contains a routing circuit G(θ_G) plus L=4 expert circuits, while the baselines are single PQC circuits; yet no parameter counts, qubit counts, circuit depths, training epochs, or random seeds are reported. The accuracy gaps in Table I (e.g., 62.92% vs. 51.47% on MNIST-4) are therefore equally consistent with a capacity increase as with a benefit of the routing mixture. Table II reinforces this concern: accuracy rises monotonically with the number of experts under every operator set. Please add resource-matched baselines, such as a single PQC with a comparable total parameter count and circuit depth, and report the resource budgets for all models.
  2. [§IV.A–C / Fig. 2] The proposed mechanism is not specified mathematically. The classical MoE output is defined in Eq. (1), but there is no equivalent equation for the QMoE output: no expression for the routing circuit's effect, for the controlled-expert operations, or for the aggregation observable. The phrase 'quantum states aggregation' is described only verbally and via a figure. As a result, the claim that the measurement outcome is an input-dependent mixture of expert transformations is not derivable from the text, and the framework cannot be reimplemented unambiguously. Please provide the full state evolution, the definition of the measured observable(s), and any classical post-processing.
  3. [§IV.B / Abstract / §I] The paper claims sparsity and scalability, but no sparsity mechanism is described: the routing circuit places control qubits in a superposition and all L experts are applied as controlled operations in the circuit. Hence the quantum circuit must realize every expert, and the computational cost grows linearly with L; there is no conditional skip of expert execution comparable to the top-k routing in classical MoE. If the intended efficiency claim is about parameter count or inductive bias rather than execution cost, the manuscript should state that explicitly and support it with an appropriate resource analysis.
  4. [§V.A / Table I] No statistical reliability is reported. The experiments appear to be single runs, with no seeds, error bars, or repeated trials. Given the small accuracy differences among some configurations, e.g., 97.29% vs. 98.62% for two QMoE variants on MNIST-2, it is impossible to tell whether the reported improvements are significant. Please report results averaged over multiple random seeds with standard deviations, or otherwise justify that the single-run comparisons are stable.
minor comments (6)
  1. [§I] There is a missing word in 'a trainable quantum dynamically selects which experts to activate'; it should be 'a trainable quantum circuit' or 'a trainable quantum routing circuit'. The sentence in the contributions paragraph, 'we introduce a quantum routing circuit and parameterized quantum experts to determine the input data for the specialized experts', is also awkward and should be rephrased.
  2. [§II.B / Eq. (2)] The text says the state |ψ⟩ is transformed via V(θ) to a new state |ψ⟩, using the same symbol for the input and output states. Use different labels for clarity, e.g., |ψ_in⟩ and |ψ_out⟩.
  3. [§V.C / Table II] The ablation study is reported only on MNIST-4. Adding an ablation on a second benchmark, or on a 2-class task, would help establish that the observed scaling with the number of experts is not task-specific.
  4. [§III / Ref. [35]] The closely related prior work [35], a globally trained mixture of quantum experts, is cited but not compared experimentally or discussed in sufficient detail. The novelty of the quantum-native routing over that work should be stated explicitly.
  5. [§V.A] The learning rate is reported, but the number of training epochs, batch size, and convergence criterion are not. Please add these details to the implementation description.
  6. [§IV.C] The parameter-shift rule is mentioned without a citation. Please add a reference to the original parameter-shift rule or a standard textbook treatment.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the QMoE accuracy claim rests on external benchmark comparisons; self-citations are background, and the missing resource-match controls are a correctness concern, not logical circularity.

full rationale

The central claim of the paper is an empirical accuracy comparison (Section V.B, Table I) between QMoE and standard PQC baselines on MNIST and Fashion-MNIST. No quantity in that comparison is defined in terms of another, no fitted parameter is relabeled as a prediction, and no uniqueness claim is imported from prior work. The many self-citations (references 8, 9, 13, 14, 15, 16, 17) appear in the background and related-work sections and are not used to establish the QMoE advantage; the advantage is claimed directly from the Table I numbers. Table II's monotonic accuracy improvement with expert count is an ablation result, not a circular construction: it does not redefine the reported test accuracy in terms of the number of experts by construction. The absence of reported parameter counts, circuit depths, epochs, and seeds in Section V.A is a potentially serious experimental-control concern, since the observed gains could be capacity effects rather than effects of the routing mechanism, but that is a correctness and comparability issue, not logical circularity. The comparison is not equivalent to its own inputs by definition, and the paper is self-contained against external benchmarks. Therefore no circular step is identified.

Assumptions & free parameters 3 free parameters · 2 assumptions · 0 invented entities

QMoE is an empirical architecture paper, so the ledger is light. Trainable network parameters are learned from data and are not fitted constants for a derived law. The key unproved premises are that parameter-shift gradients work through the controlled routing structure and that the controlled aggregation implements a learnable weighted mixture. No new physical entities are introduced.

free parameters (3)
  • Number of experts L = 4 (main results); 2, 3 in ablations
    Chosen by hand; Table II shows accuracy improves with L, so the main comparison uses this design choice.
  • Learning rate for Adam = 2e-3
    Standard Adam hyperparameter stated in Section V.A; not fitted to the data.
  • Image resolution = 8x8
    Images resized to 8x8 before encoding; this preprocessing choice sets the task difficulty.
assumptions (2)
  • domain assumption Parameter-shift rule gives unbiased gradients for QMoE and gradients propagate through the controlled routing and aggregation.
    Invoked in Section IV.B ('updated via gradient-based methods, such as the parameter-shift method') and Section IV.C for end-to-end training; standard in variational QML but not proved here.
  • ad hoc to paper The controlled-expert aggregation produces a measurement outcome that is a learnable, input-dependent mixture of expert transformations.
    Section IV.C describes aggregation only as 'a set of simple controlled operators'; no equation or proof is given that the final observable corresponds to a weighted mixture.

how reviews work

0 comments
Cite this review

Pith. "Pith review of QMoE: A Quantum Mixture of Experts Framework for Scalable Quantum Neural Networks." pith.science (2026). https://pith.science/paper/BUI2GTDH

@misc{pith2026250705190,
  author       = {Pith},
  title        = {Pith review of: QMoE: A Quantum Mixture of Experts Framework for Scalable Quantum Neural Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BUI2GTDH}},
  note         = {Machine review of arXiv:2507.05190}
}
read the original abstract

Quantum machine learning (QML) has emerged as a promising direction in the noisy intermediate-scale quantum (NISQ) era, offering computational and memory advantages by harnessing superposition and entanglement. However, QML models often face challenges in scalability and expressiveness due to hardware constraints. In this paper, we propose quantum mixture of experts (QMoE), a novel quantum architecture that integrates the mixture of experts (MoE) paradigm into the QML setting. QMoE comprises multiple parameterized quantum circuits serving as expert models, along with a learnable quantum routing mechanism that selects and aggregates specialized quantum experts per input. The empirical results from the proposed QMoE on quantum classification tasks demonstrate that it consistently outperforms standard quantum neural networks, highlighting its effectiveness in learning complex data patterns. Our work paves the way for scalable and interpretable quantum learning frameworks.

Figures

Figures reproduced from arXiv: 2507.05190 by the authors.

Figure 1
Figure 1. A conventional framework of classical mixture of experts. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The overall framework of the proposed quantum mixture of experts. The framework includes a classical data encoding circuit [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. , which is applied to all routing and expert qubits in parallel. Following this encoding, a parameterized quantum routing circuit G(θG) is applied to the routing. This circuit learns a superposition over expert selection patterns through its trainable parameters θG. The output of the routing circuit determines which subset of experts Ei(θi) is activated via controlled operations. B. Parameterized Quantum Experts The… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: A conventional architecture of parameterized quantum circuits. [PITH_FULL_IMAGE:figures/full_fig_p003_4.png]
Figure 5
Figure 5. Figure 5: Data samples of the MNIST dataset. TABLE I EXPERIMENTAL ACCURACIES (%) ON 2- AND 4-CLASS BENCHMARKS. Compute Type Method Operators MNIST-4 MNIST-2 Fashion-4 Fashion-2 Quantum Baseline RX 49.87 92.99 49.53 74.60 Quantum Baseline RY 49.27 93.04 49.13 75.35 Quantum Baseli…

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. Image classification via a quantum-inspired strategy involving a mixture of experts

    cs.LG 2026-07 conditional novelty 5.0 of 10

    Joint multi-expert unitary convolution plus stabilizer-syndrome features roughly halves MNIST/Fashion-MNIST failure rates versus classical diffusion pooling, at moderate GPU overhead.

Reference graph

Works this paper leans on

39 extracted references · 20 canonical work pages · cited by 1 Pith paper

  1. [35]

    Solving mnist with a globally trained mixture of quantum experts,

    P. A. X. Tognini, L. Banchi, and G. De Palma, “Solving mnist with a globally trained mixture of quantum experts,” arXiv preprint arXiv:2505.14789, 2025

    arXiv 2025

  2. [1]

    Quantum computing and the entanglement frontier,

    J. Preskill, “Quantum computing and the entanglement frontier,” arXiv preprint arXiv:1203.5813, 2012

    arXiv 2012

  3. [2]

    Quantum computing in the nisq era and beyond,

    ——, “Quantum computing in the nisq era and beyond,” Quantum, vol. 2, p. 79, 2018

    work page 2018

  4. [3]

    Characterizing quantum supremacy in near-term devices,

    S. Boixo, S. V . Isakov, V . N. Smelyanskiy, R. Babbush, N. Ding, Z. Jiang, M. J. Bremner, J. M. Martinis, and H. Neven, “Characterizing quantum supremacy in near-term devices,”Nature Physics, vol. 14, no. 6, pp. 595–600, 2018

    2018

  5. [4]

    Quantum k-nearest neighbors algorithm,

    A. Basheer, A. Afham, and S. K. Goyal, “Quantum k-nearest neighbors algorithm,” arXiv preprint arXiv:2003.09187 , 2020

    arXiv 2003

  6. [5]

    Quantum support vector machine for big data classification,

    P. Rebentrost, M. Mohseni, and S. Lloyd, “Quantum support vector machine for big data classification,” Physical review letters , vol. 113, no. 13, p. 130503, 2014

    2014

  7. [6]

    The method of quantum clustering,

    D. Horn and A. Gottlieb, “The method of quantum clustering,” Advances in neural information processing systems , vol. 14, 2001

    work page 2001

  8. [7]

    Algorithm for data clustering in pattern recognition problems based on quantum mechanics,

    ——, “Algorithm for data clustering in pattern recognition problems based on quantum mechanics,” Physical review letters , vol. 88, no. 1, p. 018702, 2001

    work page 2001

Show all 39 references
  1. [8]

    Quantum vision clustering,

    X. B. Nguyen, H. Churchill, K. Luu, and S. U. Khan, “Quantum vision clustering,” arXiv preprint arXiv:2309.09907 , 2023

    2023 arXiv

  2. [9]

    Qclusformer: A quantum transformer-based framework for unsupervised visual clustering,

    X.-B. Nguyen, H.-Q. Nguyen, S. Y .-C. Chen, S. U. Khan, H. Churchill, and K. Luu, “Qclusformer: A quantum transformer-based framework for unsupervised visual clustering,” in 2024 IEEE International Conference on Quantum Computing and Engineering (QCE) , vol. 2. IEEE, 2024, pp. 347–352

    2024

  3. [10]

    Quantum neural networks,

    A. A. Ezhov and D. Ventura, “Quantum neural networks,” in Future Directions for Intelligent Systems and Information Sciences: The Future of Speech and Image Technologies, Brain Computers, WWW, and Bioinformatics. Springer, 2000, pp. 213–235

    2000

  4. [11]

    Quantum neural network for quantum neural computing,

    M.-G. Zhou, Z.-P. Liu, H.-L. Yin, C.-L. Li, T.-K. Xu, and Z.-B. Chen, “Quantum neural network for quantum neural computing,” Research, vol. 6, p. 0134, 2023

    2023

  5. [12]

    Quantum neural network (qnn) research a scientometrics assessment of global publications during 1990-2019

    B. Gupta and S. Dhawan, “Quantum neural network (qnn) research a scientometrics assessment of global publications during 1990-2019.” In- ternational Journal of Information Dissemination & Technology , vol. 10, no. 3, 2020

    1990

  6. [13]

    Defining quantum neural networks via quantum time evolution,

    A. Dendukuri, B. Keeling, A. Fereidouni, J. Burbridge, K. Luu, and H. Churchill, “Defining quantum neural networks via quantum time evolution,” arXiv preprint arXiv:1905.10912 , 2019

    1905 arXiv

  7. [14]

    Image processing in quantum computers,

    A. Dendukuri and K. Luu, “Image processing in quantum computers,” arXiv preprint arXiv:1812.11042 , 2018

    2018 arXiv

  8. [15]

    Hierarchical quantum control gates for functional mri understanding,

    X.-B. Nguyen, H.-Q. Nguyen, H. Churchill, S. U. Khan, and K. Luu, “Hierarchical quantum control gates for functional mri understanding,” in 2024 IEEE Workshop on Signal Processing Systems (SiPS) . IEEE, 2024, pp. 159–164

    2024

  9. [16]

    Diffusion-inspired quantum noise mitigation in parameterized quantum circuits,

    H.-Q. Nguyen, X. B. Nguyen, S. Y .-C. Chen, H. Churchill, N. Borys, S. U. Khan, and K. Luu, “Diffusion-inspired quantum noise mitigation in parameterized quantum circuits,” Quantum Machine Intelligence , vol. 7, no. 1, p. 55, 2025

    2025

  10. [17]

    Quantum-brain: Quantum-inspired neu- ral network approach to vision-brain understanding,

    H.-Q. Nguyen, X.-B. Nguyen, H. Churchill, A. K. Choudhary, P. Sinha, S. U. Khan, and K. Luu, “Quantum-brain: Quantum-inspired neu- ral network approach to vision-brain understanding,” arXiv preprint arXiv:2411.13378, 2024

    2024 arXiv

  11. [18]

    Quantum machine learning,

    J. Biamonte, P. Wittek, N. Pancotti, P. Rebentrost, N. Wiebe, and S. Lloyd, “Quantum machine learning,” Nature, vol. 549, no. 7671, pp. 195–202, 2017

    2017

  12. [19]

    Expressive power of parametrized quantum circuits,

    Y . Du, M.-H. Hsieh, T. Liu, and D. Tao, “Expressive power of parametrized quantum circuits,” Physical Review Research, vol. 2, no. 3, p. 033125, 2020

    2020

  13. [20]

    Outrageously large neural networks: The sparsely-gated mixture-of-experts layer,

    N. Shazeer, A. Mirhoseini, K. Maziarz, A. Davis, Q. Le, G. Hinton, and J. Dean, “Outrageously large neural networks: The sparsely-gated mixture-of-experts layer,” arXiv preprint arXiv:1701.06538 , 2017

    2017 arXiv

  14. [21]

    Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity,

    W. Fedus, B. Zoph, and N. Shazeer, “Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity,” Journal of Machine Learning Research , vol. 23, no. 120, pp. 1–39, 2022

    2022

  15. [22]

    Adaptive mixtures of local experts,

    R. A. Jacobs, M. I. Jordan, S. J. Nowlan, and G. E. Hinton, “Adaptive mixtures of local experts,” Neural computation, vol. 3, no. 1, pp. 79–87, 1991

    1991

  16. [23]

    Parameterized quantum circuits as machine learning models,

    M. Benedetti, E. Lloyd, S. Sack, and M. Fiorentini, “Parameterized quantum circuits as machine learning models,” Quantum Science and Technology, vol. 4, no. 4, p. 043001, 2019

    2019

  17. [24]

    Quantum circuit learning,

    K. Mitarai, M. Negoro, M. Kitagawa, and K. Fujii, “Quantum circuit learning,” Physical Review A , vol. 98, no. 3, p. 032309, 2018

    2018

  18. [25]

    Variational quantum reinforcement learning via evolutionary optimiza- tion,

    S. Y .-C. Chen, C.-M. Huang, C.-W. Hsing, H.-S. Goan, and Y .-J. Kao, “Variational quantum reinforcement learning via evolutionary optimiza- tion,” Machine Learning: Science and Technology , vol. 3, no. 1, p. 015025, 2022

    2022

  19. [26]

    Quantum algorithms for supervised and unsupervised machine learning,

    S. Lloyd, M. Mohseni, and P. Rebentrost, “Quantum algorithms for supervised and unsupervised machine learning,” arXiv preprint arXiv:1307.0411, 2013

    2013 arXiv

  20. [27]

    Quantum principal component analysis,

    ——, “Quantum principal component analysis,” Nature Physics, vol. 10, no. 9, pp. 631–633, 2014

    2014

  21. [28]

    Prediction by linear regression on a quantum computer,

    M. Schuld, I. Sinayskiy, and F. Petruccione, “Prediction by linear regression on a quantum computer,” Physical Review A , vol. 94, no. 2, p. 022342, 2016

    2016

  22. [29]

    Quantum gradient descent for linear systems and least squares,

    I. Kerenidis and A. Prakash, “Quantum gradient descent for linear systems and least squares,” Physical Review A , vol. 101, no. 2, p. 022316, 2020

    2020

  23. [30]

    Neural networks with quantum architec- ture and quantum learning,

    M. Panella and G. Martinelli, “Neural networks with quantum architec- ture and quantum learning,” International Journal of Circuit Theory and Applications, vol. 39, no. 1, pp. 61–77, 2011

    2011

  24. [31]

    Quantum convolutional neural networks,

    I. Cong, S. Choi, and M. D. Lukin, “Quantum convolutional neural networks,” Nature Physics, vol. 15, no. 12, pp. 1273–1278, 2019

    2019

  25. [32]

    Recurrent quantum neural networks,

    J. Bausch, “Recurrent quantum neural networks,” Advances in neural information processing systems , vol. 33, pp. 1368–1379, 2020

    2020

  26. [33]

    Experimental quantum generative adversarial networks for image generation,

    H.-L. Huang, Y . Du, M. Gong, Y . Zhao, Y . Wu, C. Wang, S. Li, F. Liang, J. Lin, Y . Xu et al. , “Experimental quantum generative adversarial networks for image generation,” Physical Review Applied, vol. 16, no. 2, p. 024051, 2021

    2021

  27. [34]

    Quantum autoencoders for efficient compression of quantum data,

    J. Romero, J. P. Olson, and A. Aspuru-Guzik, “Quantum autoencoders for efficient compression of quantum data,” Quantum Science and Technology, vol. 2, no. 4, p. 045001, 2017

    2017

  28. [36]

    Mnist handwritten digit database,

    Y . LeCun, C. Cortes, and C. Burges, “Mnist handwritten digit database,” ATT Labs [Online]. Available: http://yann.lecun.com/exdb/mnist , vol. 2, 2010

    2010

  29. [37]

    Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms,

    H. Xiao, K. Rasul, and R. V ollgraf, “Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms,” arXiv preprint arXiv:1708.07747, 2017

    2017 arXiv

  30. [38]

    Torchquantum case study for robust quantum circuits,

    H. Wang, Z. Liang, J. Gu, Z. Li, Y . Ding, W. Jiang, Y . Shi, D. Z. Pan, F. T. Chong, and S. Han, “Torchquantum case study for robust quantum circuits,” in Proceedings of the 41st IEEE/ACM International Conference on Computer-Aided Design , 2022, pp. 1–9

    2022

  31. [39]

    Adam: A method for stochastic optimization,

    D. P. Kingma, “Adam: A method for stochastic optimization,” arXiv preprint arXiv:1412.6980, 2014

    2014 arXiv

Pith tools

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