Pith. sign in

REVIEW 4 major objections 3 minor 2 cited by

Quantum Relational Knowledge Distillation

T0 review · 4 major / 3 minor · reviewed 2026-08-05 · deepseek-v4-flash

Pith's one-line read QRKD replaces the pairwise distances and angles of classical relational distillation with quantum kernel values, and reports that students distilled this way outperform classical RKD on every benchmark tested.

desk verdict A plausible and practical distillation idea, but the quantum-specific claim needs a classical-kernel control before it earns the 'enhanced' label. read the letter →

arxiv 2508.13054 v1 pith:FDFDS7RI submitted 2025-08-18 quant-ph

classification quant-ph
keywords quantumkernelknowledgedistillationrelationalmachinelearningHilbertspacefeaturemapmodelcompressionvisiontaskslanguagemodels
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

Quantum Relational Knowledge Distillation (QRKD) is proposed as a way to make knowledge distillation more accurate by using quantum-computed relations between samples. Instead of aligning a student model's pairwise distances and angles to the teacher's, as classical RKD does, QRKD encodes classical features as quantum states and uses quantum kernel values as the relational signal. The paper claims that this signal captures richer inter-sample relationships and reports consistent student-model improvements over classical RKD on vision benchmarks (MNIST, CIFAR-10) and language benchmarks (WikiText-2, Penn Treebank, IMDB). The practical payoff is that both teacher and student remain classical models deployable on ordinary hardware, with quantum computation required only during training.

What carries the argument

The load-bearing object is the quantum kernel: for a pair of classical feature vectors, each is encoded into a quantum state by a chosen feature map and the inner product (fidelity) between the two states is computed. QRKD replaces the pairwise distances and angles of classical RKD with these quantum kernel values, so the student is trained to reproduce a teacher's quantum-relational structure. The claim is that this Hilbert-space relational structure carries information that a classical distance-angle relation does not.

What would settle it

Run QRKD against a classical-kernel version of the same loss on the same feature vectors and hyperparameters; if the classical-kernel student matches or beats QRKD's improvement over RKD, the claim that quantum relations are responsible would be refuted.

Watch

Extended reading notes

Core claim

The core discovery the authors set out to establish is that relational knowledge distilled through a quantum kernel improves a student network's performance on the tested tasks relative to classical relational distillation. Concretely, QRKD takes features from teacher and student, maps them into a high-dimensional space of quantum states via a quantum feature map, reads the quantum kernel—the inner product between encoded states—and uses those values as pairwise relational information in the training loss. The teacher and student are never executed as quantum circuits; only the kernel estimation requires quantum resources. The paper reports consistent gains on all five benchmarks and describ

Load-bearing premise

The results stand on the assumption that quantum kernel values computed from the chosen feature map capture relationships between samples that classical pairwise distances and angles miss; the paper does not test whether the same feature map with a classical kernel would give the same gain.

Editorial extensions

If this is right

  • A student distilled under QRKD can be served on standard hardware with no quantum runtime, so the reported accuracy gain comes at zero additional inference cost.
  • The method applies to both convolutional vision models and transformer-based language models, indicating the quantum relational signal is not tied to one architecture.
  • If the reported gains hold, model compression can borrow quantum expressiveness without requiring end users to own or operate quantum devices.
  • QRKD inherits the pairwise structure of RKD, so the added training cost is the quantum kernel estimation rather than a change to the distillation protocol itself.

Reading between the lines

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

  • My inference: the untested control that would settle the quantum-specific claim is a classical kernel evaluated on the same encoded features; the paper does not report that comparison.
  • My inference: since QRKD needs only pairwise kernel values, the kernels could be precomputed and cached once per dataset, which would remove quantum hardware from the iterative training loop entirely.
  • My inference: the same relational loss could be applied to intermediate feature layers rather than only final features, which might show whether the quantum relational signal compounds through network depth.
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 / 3 minor

Summary. The paper proposes Quantum Relational Knowledge Distillation (QRKD), an extension of Relational Knowledge Distillation (RKD) in which classical features are mapped to a Hilbert space, interpreted as quantum states, and used to compute quantum kernel values that serve as an additional relational distillation loss. The authors claim that across vision benchmarks (CNNs on MNIST and CIFAR-10) and language benchmarks (GPT-2 on WikiText-2, Penn Treebank, and IMDB), QRKD consistently improves student model performance over classical RKD. A key practical selling point is that teacher and student models remain classical and deployable on standard hardware, with quantum computation required only during training. The manuscript positions this as the first demonstration of quantum-enhanced knowledge distillation in a fully classical deployment setting.

Significance. If the empirical claim holds, the paper would offer a practical and broadly applicable recipe for incorporating quantum computation into model compression while keeping deployment classical. The evaluation spans both vision and language domains, which is broader than much of the distillation literature, and the framing of 'quantum only during training' is appealing. However, the paper's contribution is almost entirely empirical, and the manuscript as supplied cannot support the advertised conclusion: the full text is corrupted, there is no control against a classical kernel-based RKD using the same feature encoding, and the reported tables appear to lack statistical error bars. The quantum-specific attribution is therefore untested, and the central claim is not yet substantiated. The paper does not provide code or machine-checked proofs; its strongest assets are the scope of the benchmarks and the clarity of the deployment scenario.

major comments (4)
  1. [Entire full text (after abstract)] The supplied manuscript body is severely corrupted: the majority of the text after the abstract is mojibake, making the method, experimental setup, and results tables unreadable. Equations defining the quantum kernel and the objective cannot be verified, and the numerical values in the experimental tables are not reliably legible. This is not a minor presentation issue; it blocks any substantive assessment of the paper's central empirical claim. A clean, readable version must be provided before the paper can be properly reviewed.
  2. [Section 3 (method) and Tables 1–2] The experiments compare QRKD only against classical distance/angle RKD. There is no ablation in which the quantum kernel is replaced by a classical kernel (e.g., RBF/Gaussian or polynomial) built from the same feature encoding and evaluated with the same training protocol. Since any finite-dimensional quantum kernel equals a classical kernel, and since the small qubit counts used in such experiments are classically simulable, the observed gains could be caused by the choice of a nonlinear kernel or by the additional scalar loss weight lambda, rather than by any quantum resource. The abstract's claim that quantum kernel values capture 'richer inter-sample relationships' therefore rests on an untested attribution. Adding a classical-kernel RKD baseline is essential to isolate the quantum-specific contribution.
  3. [Tables 1–2 and abstract] The central claim that QRKD 'consistently improves student model performance' is not supported by statistical evidence. The tables, where legible, do not report standard deviations, standard errors, or the number of random seeds. Distillation gains on MNIST/CIFAR-10 are often within fractions of a percent, so the reported differences may be within run-to-run noise. The authors should report mean ± standard deviation over at least 3–5 seeds and, where possible, include a paired test or effect-size measure to substantiate 'consistent' improvement.
  4. [Section 3 (loss function) and experimental setup] The method introduces several free parameters: the QRKD loss weight lambda, the quantum feature map/encoding parameters, and the number of measurement shots. The manuscript does not specify how these were chosen, what ranges were searched, or whether the conclusions are sensitive to them. If lambda was tuned separately for QRKD and the classical RKD baseline, the comparison is not controlled. Please state the exact values used, the tuning protocol, and include a sensitivity analysis for lambda.
minor comments (3)
  1. [Abstract and Section 3] The phrase 'richer inter-sample relationships' is not defined operationally. Please specify the metric or property used to compare the quantum kernel relations with classical distance/angle relations, and clarify what 'richer' means in this context.
  2. [Section 3 (kernel definition)] If the quantum kernel is defined as the squared overlap |⟨φ(x_i)|φ(x_j)⟩|^2, please state this explicitly and discuss how finite-shot estimation affects the loss gradient and the training noise. This is important for reproducibility.
  3. [Related work / abstract] The claim of 'first demonstration of quantum-enhanced knowledge distillation' should be tempered or supported by a related-work comparison. The corrupted text makes it impossible to verify whether prior quantum-assisted distillation or kernel-based distillation methods are discussed.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: QRKD's relational loss is computed from quantum kernels, not fitted to target performance; the reported improvement is an external empirical comparison.

full rationale

The paper's derivation chain is: map classical teacher/student features into a Hilbert space; compute quantum kernel values k_Q(x_i,x_j)=|<phi(x_i)|phi(x_j)>|^2; add a quantum kernel consistency term to the classical RKD objective; train the classical student; measure test accuracy on held-out benchmarks. The claimed 'consistent improvement' is an experimental comparison against classical RKD, not a quantity forced by construction from the loss or from fitted parameters. The quantum kernel values are inputs to the training loss, not parameters fitted to the reported test accuracies. No uniqueness theorem is imported, no ansatz is justified solely by a self-citation, and no definitional equivalence between the prediction and the input is visible. The absence of a classical-kernel ablation using the same feature encoding is a real threat to the 'quantum-enhanced' attribution, but it is a control/comparison problem rather than a circularity: it does not show that the result reduces to its inputs by definition. The paper is self-contained against external benchmarks on the claims it actually evaluates, so the circularity score is 0.

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

The abstract does not introduce new physical entities. The main assumptions are about the usefulness of quantum kernels and the standardness of the benchmarks. Several hyperparameters are likely present in the full method but are not visible in the abstract.

free parameters (3)
  • QRKD loss weight lambda = not reported in abstract
    Balances the task loss and the quantum relational loss; standard in KD and typically tuned per task or dataset.
  • quantum feature map parameters = not reported in abstract
    The encoding determines the quantum kernel; whether the parameters are fixed or trained is not specified in the abstract.
  • number of measurement shots = not reported in abstract
    Affects the precision of estimated quantum kernel values and would be chosen by the experimenter.
assumptions (3)
  • domain assumption Classical features can be efficiently encoded as quantum states and the quantum kernel is computable
    The abstract assumes this mapping is feasible and that kernel values can be estimated as part of training.
  • domain assumption Quantum kernel values capture richer inter-sample relationships than classical distances and angles
    This is the central premise for QRKD's expected benefit, asserted in the abstract but not proven there.
  • domain assumption Standard training and evaluation protocols on MNIST, CIFAR-10, WikiText-2, Penn Treebank, and IMDB are used
    The validity of the comparison depends on standard data splits and training procedures, which are not detailed in the abstract.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Quantum Relational Knowledge Distillation." pith.science (2026). https://pith.science/paper/FDFDS7RI

@misc{pith2026250813054,
  author       = {Pith},
  title        = {Pith review of: Quantum Relational Knowledge Distillation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FDFDS7RI}},
  note         = {Machine review of arXiv:2508.13054}
}
read the original abstract

Knowledge distillation (KD) is a widely adopted technique for compressing large models into smaller, more efficient student models that can be deployed on devices with limited computational resources. Among various KD methods, Relational Knowledge Distillation (RKD) improves student performance by aligning relational structures in the feature space, such as pairwise distances and angles. In this work, we propose Quantum Relational Knowledge Distillation (QRKD), which extends RKD by incorporating quantum relational information. Specifically, we map classical features into a Hilbert space, interpret them as quantum states, and compute quantum kernel values to capture richer inter-sample relationships. These quantum-informed relations are then used to guide the distillation process. We evaluate QRKD on both vision and language tasks, including CNNs on MNIST and CIFAR-10, and GPT-2 on WikiText-2, Penn Treebank, and IMDB. Across all benchmarks, QRKD consistently improves student model performance compared to classical RKD. Importantly, both teacher and student models remain classical and deployable on standard hardware, with quantum computation required only during training. This work presents the first demonstration of quantum-enhanced knowledge distillation in a fully classical deployment setting.

Discussion (0). Sign in to comment.

Forward citations

Cited by 2 Pith papers

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

  1. MerLin: A Discovery Engine for Photonic and Hybrid Quantum Machine Learning

    cs.LG 2026-02 unverdicted novelty 7.0 of 10

    MerLin is a new open-source discovery engine for photonic and hybrid quantum machine learning that integrates circuit simulations into standard ML frameworks and reproduces 18 prior works as reusable benchmarks.

  2. Quantum-Gated Task-interaction Knowledge Distillation for Pre-trained Model-based Class-Incremental Learning

    cs.LG 2026-04 unverdicted novelty 6.0 of 10

    QKD applies quantum-gated modulation to task embeddings to enable inter-task knowledge distillation that bridges representation gaps and mitigates catastrophic forgetting in streaming class-incremental learning.

Reference graph

Works this paper leans on

76 extracted references · 11 linked inside Pith · cited by 2 Pith papers

  1. [1]

    A survey of convolutional neural networks: analysis, applications, and prospects

    Zewen Li, Fan Liu, Wenjie Yang, Shouheng Peng, and Jun Zhou. A survey of convolutional neural networks: analysis, applications, and prospects. IEEE transactions on neural networks and learning systems , 33(12):6999--7019, 2021

  2. [2]

    Language models are unsupervised multitask learners

    Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. Language models are unsupervised multitask learners. OpenAI blog , 1(8):9, 2019

  3. [3]

    Awq: Activation-aware weight quantization for on-device llm compression and acceleration

    Ji Lin, Jiaming Tang, Haotian Tang, Shang Yang, Wei-Ming Chen, Wei-Chen Wang, Guangxuan Xiao, Xingyu Dang, Chuang Gan, and Song Han. Awq: Activation-aware weight quantization for on-device llm compression and acceleration. Proceedings of machine learning and systems , 6:87--100, 2024

  4. [4]

    Distilling the knowledge in a neural network

    Geoffrey Hinton, Oriol Vinyals, and Jeff Dean. Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531 , 2015

  5. [5]

    Like what you like: Knowledge distill via neuron selectivity transfer

    Zehao Huang and Naiyan Wang. Like what you like: Knowledge distill via neuron selectivity transfer. arXiv preprint arXiv:1707.01219 , 2017

  6. [6]

    Data-free knowledge distillation for deep neural networks

    Raphael Gontijo Lopes, Stefano Fenu, and Thad Starner. Data-free knowledge distillation for deep neural networks. arXiv preprint arXiv:1710.07535 , 2017

  7. [7]

    Unifying distillation and privileged information

    David Lopez-Paz, L \'e on Bottou, Bernhard Sch \"o lkopf, and Vladimir Vapnik. Unifying distillation and privileged information. arXiv preprint arXiv:1511.03643 , 2015

  8. [8]

    Relational knowledge distillation

    Wonpyo Park, Dongju Kim, Yan Lu, and Minsu Cho. Relational knowledge distillation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 3967--3976, 2019

Show all 76 references
  1. [9]

    An introduction to quantum machine learning

    Maria Schuld, Ilya Sinayskiy, and Francesco Petruccione. An introduction to quantum machine learning. Contemporary Physics , 56(2):172–185, October 2014

  2. [10]

    Quantum machine learning

    Jacob Biamonte, Peter Wittek, Nicola Pancotti, Patrick Rebentrost, Nathan Wiebe, and Seth Lloyd. Quantum machine learning. Nature , 549(7671):195--202, 2017

  3. [11]

    Training classical neural networks by quantum machine learning

    Chen-Yu Liu, En-Jui Kuo, Chu-Hsuan Abraham Lin, Sean Chen, Jason Gemsun Young, Yeong-Jar Chang, and Min-Hsiu Hsieh. Training classical neural networks by quantum machine learning. In 2024 IEEE International Conference on Quantum Computing and Engineering (QCE) , volume 2, page...

  4. [12]

    A quantum circuit-based compression perspective for parameter-efficient learning

    Chen-Yu Liu, Chao-Han Huck Yang, Hsi-Sheng Goan, and Min-Hsiu Hsieh. A quantum circuit-based compression perspective for parameter-efficient learning. In The Thirteenth International Conference on Learning Representations , 2025

  5. [13]

    Quantum machine learning in feature hilbert spaces

    Maria Schuld and Nathan Killoran. Quantum machine learning in feature hilbert spaces. Physical review letters , 122(4):040504, 2019

  6. [14]

    Supervised learning with quantum-enhanced feature spaces

    Vojt e ch Havl \' c ek, Antonio D C \'o rcoles, Kristan Temme, Aram W Harrow, Abhinav Kandala, Jerry M Chow, and Jay M Gambetta. Supervised learning with quantum-enhanced feature spaces. Nature , 567(7747):209--212, 2019

  7. [15]

    Power of data in quantum machine learning

    Hsin-Yuan Huang, Michael Broughton, Masoud Mohseni, Ryan Babbush, Sergio Boixo, Hartmut Neven, and Jarrod R McClean. Power of data in quantum machine learning. Nature communications , 12(1):2631, 2021

  8. [16]

    Similarity of neural network representations revisited

    Simon Kornblith, Mohammad Norouzi, Honglak Lee, and Geoffrey Hinton. Similarity of neural network representations revisited. In International conference on machine learning , pages 3519--3529. PMLR, 2019

  9. [17]

    A kernel perspective on distillation-based collaborative learning

    Sejun Park, Kihun Hong, and Ganguk Hwang. A kernel perspective on distillation-based collaborative learning. Advances in Neural Information Processing Systems , 37:91827--91879, 2024

  10. [18]

    Rethinking centered kernel alignment in knowledge distillation

    Zikai Zhou, Yunhang Shen, Shitong Shao, Linrui Gong, and Shaohui Lin. Rethinking centered kernel alignment in knowledge distillation. In Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence , pages 5680--5688, 2024

  11. [19]

    Supervised quantum machine learning models are kernel methods

    Maria Schuld. Supervised quantum machine learning models are kernel methods. arXiv preprint arXiv:2101.11020 , 2021

  12. [20]

    Exponential concentration in quantum kernel methods

    Supanut Thanasilp, Samson Wang, Marco Cerezo, and Zo \"e Holmes. Exponential concentration in quantum kernel methods. Nature communications , 15(1):5200, 2024

  13. [21]

    Neural quantum kernels: Training quantum kernels with quantum neural networks

    Pablo Rodriguez-Grasa, Yue Ban, and Mikel Sanz. Neural quantum kernels: Training quantum kernels with quantum neural networks. Physical Review Research , 7(2):023269, 2025

  14. [22]

    Geometrical and statistical properties of systems of linear inequalities with applications in pattern recognition

    Thomas M Cover. Geometrical and statistical properties of systems of linear inequalities with applications in pattern recognition. IEEE Transactions on Electronic Computers , EC-14(3):326--334, 1965

  15. [23]

    A training algorithm for optimal margin classifiers

    Bernhard E Boser, Isabelle M Guyon, and Vladimir N Vapnik. A training algorithm for optimal margin classifiers. In Proceedings of the fifth annual workshop on Computational learning theory , pages 144--152, 1992

  16. [24]

    Pointer sentinel mixture models

    Stephen Merity, Caiming Xiong, James Bradbury, and Richard Socher. Pointer sentinel mixture models. arXiv preprint arXiv:1609.07843 , 2016

  17. [25]

    Building a large annotated corpus of english: The penn treebank

    Mary Ann Marcinkiewicz. Building a large annotated corpus of english: The penn treebank. Using Large Corpora , 273:31, 1994

  18. [26]

    Learning word vectors for sentiment analysis

    Andrew Maas, Raymond E Daly, Peter T Pham, Dan Huang, Andrew Y Ng, and Christopher Potts. Learning word vectors for sentiment analysis. In Proceedings of the 49th annual meeting of the association for computational linguistics: Human language technologies , pages 142--150, 2011

  19. [27]

    Quantum knowledge distillation for large language models

    Lingxiao Li, Yihao Wang, Jiacheng Fan, Jing Li, Sujuan Qin, Qiaoyan Wen, and Fei Gao. Quantum knowledge distillation for large language models. arXiv preprint arXiv:2505.13205 , 2025

  20. [28]

    Quantum computation and quantum information

    Michael A Nielsen and Isaac L Chuang. Quantum computation and quantum information . Cambridge university press, 2010

  21. [29]

    Juan Carrasquilla and Roger G. Melko. Machine learning phases of matter. Nature Physics , 13(5):431--434, 2017

  22. [30]

    Sebastian J. Wetzel. Unsupervised learning of phase transitions: From principal component analysis to variational autoencoders. Phys. Rev. E , 96:022140, Aug 2017

  23. [31]

    Evert P. L. van Nieuwenburg, Ye-Hua Liu, and Sebastian D. Huber. Learning phase transitions by confusion. Nature Physics , 13(5):435--439, 2017

  24. [32]

    Learning phase transitions from dynamics

    Evert van Nieuwenburg, Eyal Bairey, and Gil Refael. Learning phase transitions from dynamics. Phys. Rev. B , 98:060301, Aug 2018

  25. [33]

    Probing many-body localization with neural networks

    Frank Schindler, Nicolas Regnault, and Titus Neupert. Probing many-body localization with neural networks. Phys. Rev. B , 95:245134, Jun 2017

  26. [34]

    Machine learning the thermodynamic arrow of time

    Alireza Seif, Mohammad Hafezi, and Christopher Jarzynski. Machine learning the thermodynamic arrow of time. Nature Physics , 17(1):105--113, 2021

  27. [35]

    Random sampling neural network for quantum many-body problems

    Chen-Yu Liu and Daw-Wei Wang. Random sampling neural network for quantum many-body problems. Phys. Rev. B , 103:205107, May 2021

  28. [36]

    Classification with quantum neural networks on near term processors

    Edward Farhi and Hartmut Neven. Classification with quantum neural networks on near term processors. arXiv preprint arXiv:1802.06002 , 2018

  29. [37]

    Quantum machine learning: a classical perspective

    Carlo Ciliberto, Mark Herbster, Alessandro Davide Ialongo, Massimiliano Pontil, Andrea Rocchetto, Simone Severini, and Leonard Wossnig. Quantum machine learning: a classical perspective. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences , 474(...

  30. [38]

    Quantum machine learning in feature hilbert spaces

    Maria Schuld and Nathan Killoran. Quantum machine learning in feature hilbert spaces. Physical Review Letters , 122(4), February 2019

  31. [39]

    Lov K. Grover. A fast quantum mechanical algorithm for database search. ACM Symposium on Theory of Computing (STOC) , 28:212--219, 1996

  32. [40]

    A grover-search based quantum learning scheme for classification

    Yuxuan Du, Min-Hsiu Hsieh, Tongliang Liu, and Dacheng Tao. A grover-search based quantum learning scheme for classification. New Journal of Physics , 23(2):023020, feb 2021

  33. [41]

    Olson, Matthias Degroote, Peter D

    Yudong Cao, Jonathan Romero, Jonathan P. Olson, Matthias Degroote, Peter D. Johnson, M \'a ria Kieferov \'a , Ian D. Kivlichan, Tim Menke, Borja Peropadre, Nicolas P. D. Sawaya, Sukin Sim, Libor Veis, and Al \'a n Aspuru-Guzik. Quantum chemistry in the age of quantum computing...

  34. [42]

    Quantum-enhanced support vector machine for large-scale multi-class stellar classification

    Kuan-Cheng Chen, Xiaotian Xu, Henry Makhanov, Hui-Hsuan Chung, and Chen-Yu Liu. Quantum-enhanced support vector machine for large-scale multi-class stellar classification. In Advanced Intelligent Computing Technology and Applications: 20th International Conference, ICIC 2024, ...

  35. [43]

    Multi-gpu-enabled hybrid quantum-classical workflow in quantum-hpc middleware: Applications in quantum simulations

    Kuan-Cheng Chen, Xiaoren Li, Xiaotian Xu, Yun-Yuan Wang, and Chen-Yu Liu. Multi-gpu-enabled hybrid quantum-classical workflow in quantum-hpc middleware: Applications in quantum simulations. arXiv preprint arXiv:2403.05828 , 2024

  36. [44]

    Reinforcement learning quantum local search

    Chen-Yu Liu and Hsi-Sheng Goan. Reinforcement learning quantum local search. In 2023 IEEE International Conference on Quantum Computing and Engineering (QCE) , volume 2, pages 246--247. IEEE, 2023

  37. [45]

    Practical quantum search by variational quantum eigensolver on noisy intermediate-scale quantum hardware

    Chen-Yu Liu. Practical quantum search by variational quantum eigensolver on noisy intermediate-scale quantum hardware. In 2023 International Conference on Computational Science and Computational Intelligence (CSCI) , pages 397--403. IEEE, 2023

  38. [46]

    Quantum natural language processing on near-term quantum computers

    Konstantinos Meichanetzidis, Stefano Gogioso, Giovanni de Felice, Nicolò Chiappori, Alexis Toumi, and Bob Coecke. Quantum natural language processing on near-term quantum computers. Electronic Proceedings in Theoretical Computer Science , 340:213–229, September 2021

  39. [47]

    Implementation of trained factorization machine recommendation system on quantum annealer

    Chen-Yu Liu, Hsin-Yu Wang, Pei-Yen Liao, Ching-Jui Lai, and Min-Hsiu Hsieh. Implementation of trained factorization machine recommendation system on quantum annealer. In 2024 International Joint Conference on Neural Networks (IJCNN) , pages 1--8. IEEE, 2024

  40. [48]

    Quantum recommendation systems

    Iordanis Kerenidis and Anupam Prakash. Quantum recommendation systems. arXiv preprint arXiv:1603.08675 , 2016

  41. [49]

    Quantum generative adversarial networks

    Pierre-Luc Dallaire-Demers and Nathan Killoran. Quantum generative adversarial networks. Phys. Rev. A , 98:012324, Jul 2018

  42. [50]

    Learnability of quantum neural networks

    Yuxuan Du, Min-Hsiu Hsieh, Tongliang Liu, Shan You, and Dacheng Tao. Learnability of quantum neural networks. PRX Quantum , 2:040337, Nov 2021

  43. [51]

    Mahdi Soltanolkotabi, Adel Javanmard, and Jason D. Lee. Theoretical insights into the optimization landscape of over-parameterized shallow neural networks. IEEE Transactions on Information Theory , 65(2):742--769, 2019

  44. [52]

    Expressive power of parametrized quantum circuits

    Yuxuan Du, Min-Hsiu Hsieh, Tongliang Liu, and Dacheng Tao. Expressive power of parametrized quantum circuits. Phys. Rev. Research , 2:033125, Jul 2020

  45. [53]

    Learning quantum phase estimation by variational quantum circuits

    Chen-Yu Liu, Kuan-Cheng Chen, and Chu-Hsuan Abraham Lin. Learning quantum phase estimation by variational quantum circuits. In 2024 International Joint Conference on Neural Networks (IJCNN) , pages 1--6. IEEE, 2024

  46. [54]

    Escaping from the barren plateau via gaussian initializations in deep variational quantum circuits

    Kaining Zhang, Liu Liu, Min-Hsiu Hsieh, and Dacheng Tao. Escaping from the barren plateau via gaussian initializations in deep variational quantum circuits. Advances in Neural Information Processing Systems , 35:18612--18627, 2022

  47. [55]

    Quantum gram-schmidt processes and their application to efficient state readout for quantum algorithms

    Kaining Zhang, Min-Hsiu Hsieh, Liu Liu, and Dacheng Tao. Quantum gram-schmidt processes and their application to efficient state readout for quantum algorithms. Phys. Rev. Research , 3:043095, Nov 2021

  48. [56]

    Quantum differentially private sparse regression learning

    Yuxuan Du, Min-Hsiu Hsieh, Tongliang Liu, Shan You, and Dacheng Tao. Quantum differentially private sparse regression learning. IEEE Transactions on Information Theory , 68(8):5217–5233, August 2022

  49. [57]

    Quantum noise protects quantum classifiers against adversaries

    Yuxuan Du, Min-Hsiu Hsieh, Tongliang Liu, Dacheng Tao, and Nana Liu. Quantum noise protects quantum classifiers against adversaries. Phys. Rev. Research , 3:023153, May 2021

  50. [58]

    The complexity of nisq

    Sitan Chen, Jordan Cotler, Hsin-Yuan Huang, and Jerry Li. The complexity of nisq. Nature Communications , 14(1):6001, 2023

  51. [59]

    Bromley, Josh Izaac, Maria Schuld, and Nathan Killoran

    Andrea Mari, Thomas R. Bromley, Josh Izaac, Maria Schuld, and Nathan Killoran. Transfer learning in hybrid classical-quantum neural networks. Quantum , 4:340, October 2020

  52. [60]

    Quantum federated learning with quantum networks

    Tyler Wang, Huan-Hsin Tseng, and Shinjae Yoo. Quantum federated learning with quantum networks. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2024:13401--13405, 2024

  53. [61]

    Parametrized quantum policies for reinforcement learning

    Sofiene Jerbi, Casper Gyurik, Simon Marshall, Hans Briegel, and Vedran Dunjko. Parametrized quantum policies for reinforcement learning. Advances in Neural Information Processing Systems , 34:28362--28375, 2021

  54. [62]

    Quantum agents in the gym: a variational quantum algorithm for deep q-learning

    Andrea Skolik, Sofiene Jerbi, and Vedran Dunjko. Quantum agents in the gym: a variational quantum algorithm for deep q-learning. Quantum , 6:720, 2022

  55. [63]

    Variational quantum circuits for deep reinforcement learning

    Samuel Yen-Chi Chen, Chao-Han Huck Yang, Jun Qi, Pin-Yu Chen, Xiaoli Ma, and Hsi-Sheng Goan. Variational quantum circuits for deep reinforcement learning. IEEE access , 8:141007--141024, 2020

  56. [64]

    Quantum-train: Rethinking hybrid quantum-classical machine learning in the model compression perspective

    Chen-Yu Liu, En-Jui Kuo, Chu-Hsuan Abraham Lin, Jason Gemsun Young, Yeong-Jar Chang, Min-Hsiu Hsieh, and Hsi-Sheng Goan. Quantum-train: Rethinking hybrid quantum-classical machine learning in the model compression perspective. arXiv preprint arXiv:2405.11304 , 2024

  57. [65]

    Qtrl: Toward practical quantum reinforcement learning via quantum-train

    Chen-Yu Liu, Chu-Hsuan Abraham Lin, Chao-Han Huck Yang, Kuan-Cheng Chen, and Min-Hsiu Hsieh. Qtrl: Toward practical quantum reinforcement learning via quantum-train. In 2024 IEEE International Conference on Quantum Computing and Engineering (QCE) , volume 2, pages 317--322. IEEE, 2024

  58. [66]

    Quantum-train long short-term memory: Application on flood prediction problem

    Chu-Hsuan Abraham Lin, Chen-Yu Liu, and Kuan-Cheng Chen. Quantum-train long short-term memory: Application on flood prediction problem. In 2024 IEEE International Conference on Quantum Computing and Engineering (QCE) , volume 2, pages 268--273. IEEE, 2024

  59. [67]

    Federated quantum-train with batched parameter generation

    Chen-Yu Liu and Samuel Yen-Chi Chen. Federated quantum-train with batched parameter generation. In 2024 15th International Conference on Information and Communication Technology Convergence (ICTC) , pages 1133--1138. IEEE, 2024

  60. [68]

    Quantum-train with tensor network mapping model and distributed circuit ansatz

    Chen-Yu Liu, Chu-Hsuan Abraham Lin, and Kuan-Cheng Chen. Quantum-train with tensor network mapping model and distributed circuit ansatz. In ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages 1--4. IEEE, 2025

  61. [69]

    Quantum-trained convolutional neural network for deepfake audio detection

    Chu-Hsuan Abraham Lin, Chen-Yu Liu, Samuel Yen-Chi Chen, and Kuan-Cheng Chen. Quantum-trained convolutional neural network for deepfake audio detection. In 2025 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW) , pages 1--5. IEEE, 2025

  62. [70]

    Quantum-train-based distributed multi-agent reinforcement learning

    Kuan-Cheng Chen, Samuel Yen-Chi Chen, Chen-Yu Liu, and Kin K Leung. Quantum-train-based distributed multi-agent reinforcement learning. In 2025 IEEE Symposium for Multidisciplinary Computational Intelligence Incubators (MCII Companion) , pages 1--5. IEEE, 2025

  63. [71]

    Programming variational quantum circuits with quantum-train agent

    Chen-Yu Liu, Samuel Yen-Chi Chen, Kuan-Cheng Chen, Wei-Jia Huang, and Yen-Jui Chang. Programming variational quantum circuits with quantum-train agent. In 2025 International Conference on Quantum Communications, Networking, and Computing (QCNC) , pages 544--548. IEEE, 2025

  64. [72]

    Introduction to quantum-train toolkit

    Chen-Yu Liu, Chu-Hsuan Abraham Lin, Wei-Jia Huang, and Min-Hsiu Hsieh. Introduction to quantum-train toolkit. In 2024 IEEE International Conference on Quantum Computing and Engineering (QCE) , volume 2, pages 456--457. IEEE, 2024

  65. [73]

    Frame generation in hilbert space: Generative interpolation of measurement data for quantum parameter adaptation

    Chen-Yu Liu, Kuan-Cheng Chen, Samuel Yen-Chi Chen, Huang wei hao, Wei-Jia Huang, and Yen Jui Chang. Frame generation in hilbert space: Generative interpolation of measurement data for quantum parameter adaptation. In ICLR 2025 Workshop on Deep Generative Model in Machine Learn...

  66. [74]

    Quantum enhanced knowledge distillation

    Simone Piperno, Leonardo Lavagna, Francesca De Falco, Andrea Ceschini, Antonello Rosato, David Windridge, Massimo Panella, et al. Quantum enhanced knowledge distillation. In Proceedings of Quantum Techniques in Machine Learning (QTML 2024) , pages 1--3. 2024

  67. [75]

    Tensor networks for complex quantum systems

    Rom \'a n Or \'u s. Tensor networks for complex quantum systems. Nature Reviews Physics , 1(9):538--550, 2019

  68. [76]

    Neural-network quantum states, string-bond states, and chiral topological states

    Ivan Glasser, Nicola Pancotti, Moritz August, Ivan D Rodriguez, and J Ignacio Cirac. Neural-network quantum states, string-bond states, and chiral topological states. Physical Review X , 8(1):011006, 2018

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.