arxiv: 2604.01616 · v4 · submitted 2026-04-02 · 🪐 quant-ph

Recognition: 2 theorem links

· Lean Theorem

Quantum-Enhanced Processing with Tensor-Network Frontends for Privacy-Aware Federated Medical Diagnosis

Hiroshi Yamauchi , Anders Peter Kragh Dalskov , Hideaki Kawaguchi , Rodney Van Meter

Authors on Pith no claims yet

Pith reviewed 2026-05-13 21:48 UTC · model grok-4.3

classification 🪐 quant-ph

keywords tensor networksfederated learningquantum computingmedical image classificationprivacy preservationmulti-party computation

0 comments

The pith

Tensor-network frontends compress medical images so a small quantum processor can refine privacy-protected federated classifications.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper builds a hybrid system in which client devices run tensor networks to shrink high-dimensional medical scans into small latent vectors. These vectors are aggregated under multi-party computation for privacy, then passed to a quantum-enhanced processor that embeds them as quantum states and reads out class predictions via observables. On PneumoniaMNIST the tree tensor network plus quantum step gives the most stable accuracy, provided the number of qubits roughly equals the latent dimension; noisy qubits reduce the gain. The same compression step also lowers the communication volume required for secure aggregation, showing that representation size simultaneously governs both quantum feasibility and privacy overhead.

Core claim

Client-side tensor networks produce compact latent representations whose dimension sets both the communication cost of MPC-secured aggregation and the qubit count needed for post-aggregation quantum refinement; when the latent dimension is matched to available qubits the tree tensor network frontend combined with the quantum-enhanced processor yields the most balanced accuracy profile across noiseless and noisy conditions.

What carries the argument

The Quantum-Enhanced Processor (QEP), which maps the aggregated low-dimensional latent vector into a quantum state and extracts diagnostic information from expectation values of chosen observables.

If this is right

Communication volume in the secure-aggregation stage scales directly with the dimension chosen for the tensor-network latent space.
Quantum refinement remains stable only when qubit number is chosen to match the latent dimension produced by the frontend.
Noise on the quantum device reduces classification performance relative to the noiseless case, so error mitigation becomes necessary for practical use.
Representation compression, secure aggregation, and quantum readout must be co-designed rather than optimized separately.

Where Pith is reading between the lines

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

The same latent-dimension matching rule could guide qubit allocation in other distributed sensing tasks that must stay within small quantum registers.
Tensor-network compression may serve as a general bridge that lets classical federated pipelines hand off to quantum processors without requiring full-image quantum hardware.
If the observed stability advantage of the tree tensor network holds across datasets, it could become a default frontend choice for hybrid quantum-classical medical pipelines.

Load-bearing premise

The quantum refinement step produces a genuine, reproducible accuracy gain over a purely classical classifier applied to the same aggregated latent features.

What would settle it

A side-by-side test on identical tensor-network latents showing that replacing the quantum readout head with a classical neural-network head yields no statistically significant accuracy difference under matched noise levels.

Figures

Figures reproduced from arXiv: 2604.01616 by Anders Peter Kragh Dalskov, Hideaki Kawaguchi, Hiroshi Yamauchi, Rodney Van Meter.

**Figure 1.** Figure 1: Proposed TN+MPC+QEP pipeline. Client-side tensor-network encoders produce latent features, MPC-secured aggrega [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗

**Figure 2.** Figure 2: Parameterized quantum circuit used in the QEP. Each [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** Figure 3: Example samples from PneumoniaMNIST. The top row [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

**Figure 6.** Figure 6: TTN-based QEP: (a) test-accuracy distributions under [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗

**Figure 5.** Figure 5: Internal behavior of the QEP across MPS, TTN, and [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗

**Figure 7.** Figure 7: Modeled communication overhead as a function of [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗

read the original abstract

We propose a privacy-aware hybrid framework for federated medical image classification that combines tensor-network representation learning, MPC-secured aggregation, and post-aggregation quantum refinement. The framework is motivated by two practical constraints in privacy-aware federated learning: MPC can introduce substantial communication overhead, and direct quantum processing of high-dimensional medical images is unrealistic with a small number of qubits. To address both constraints within a single architecture, client-side tensor-network frontends, Matrix Product State (MPS), Tree Tensor Network (TTN), and Multi-scale Entanglement Renormalization Ansatz (MERA), compress local inputs into compact latent representations, after which a Quantum-Enhanced Processor (QEP) refines the aggregated latent feature through quantum-state embedding and observable-based readout. Experiments on PneumoniaMNIST show that the effect of the QEP is frontend-dependent rather than uniform across architectures. In the present setting, the TTN+QEP combination exhibits the most balanced overall profile. The results also suggest that the QEP behaves more stably when the qubit count is sufficiently matched to the latent dimension, while noisy conditions degrade performance relative to the noiseless setting. The MPC benchmark further shows that communication cost is governed primarily by the dimension of the protected latent representation. This indicates that tensor-network compression plays a dual role: it enables small-qubit quantum processing on compressed latent features and reduces the communication overhead associated with secure aggregation. Taken together, these results support a co-design perspective in which representation compression, post-aggregation quantum refinement, and privacy-aware deployment should be optimized jointly.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper sketches a workable hybrid pipeline that compresses medical images with tensor networks, secures aggregation via MPC, and adds a small quantum refinement step, but the reported gains from the quantum part are not cleanly separated from the rest of the design.

read the letter

The core contribution is a concrete architecture that puts MPS, TTN, or MERA frontends on the client to shrink images into low-dimensional latents, runs MPC on those latents, and then feeds the result into a Quantum-Enhanced Processor that embeds the latent into a quantum state and reads out via observables. The experiments on PneumoniaMNIST indicate that the quantum step helps more with some frontends than others, that TTN plus QEP gives the most even performance, and that communication cost drops mainly because the latent dimension is small. That last point is useful: tensor compression is doing double duty by enabling both the small quantum device and cheaper secure aggregation. The motivation section is clear about why full quantum processing of raw images is unrealistic and why MPC alone is expensive, so the co-design framing makes sense on paper. The results also note that performance drops under noise and that qubit count should roughly match latent dimension for stability. Those are practical observations worth having. The main weakness is that the abstract and summary give no classical baseline that matches parameter count or architecture, no error bars, and no mention of statistical tests or seed variance. Without those controls it is difficult to know whether the reported edge for TTN+QEP comes from the quantum embedding itself or from differences in how the tensor networks compress the data or how the downstream classifier is tuned. The paper is aimed at researchers who already work on tensor-network machine learning or small-scale quantum-classical hybrids for privacy-sensitive domains. A reader looking for an end-to-end example of how these pieces can be wired together will find the architecture description and the communication-cost measurements useful. I would send it to peer review because the setup is specific enough to be checked and the practical constraints it addresses are real, even though the current evidence for the quantum contribution needs tightening.

Referee Report

3 major / 2 minor

Summary. The manuscript proposes a hybrid privacy-aware federated learning framework for medical image classification. Client-side tensor-network frontends (MPS, TTN, MERA) compress high-dimensional inputs into compact latent representations; these are aggregated via MPC and then refined by a Quantum-Enhanced Processor (QEP) that performs quantum-state embedding and observable readout. Experiments on PneumoniaMNIST indicate that QEP effects are frontend-dependent, with the TTN+QEP combination showing the most balanced performance profile, improved stability when qubit count matches latent dimension, and overall communication savings from tensor-network compression.

Significance. If the empirical claims hold under proper controls, the work would provide a concrete example of co-design among tensor-network representation learning, secure aggregation, and small-qubit quantum post-processing for privacy-sensitive medical applications. The dual role of compression in enabling both quantum feasibility and reduced MPC overhead is a potentially useful insight for resource-constrained federated settings.

major comments (3)

[Experimental results] Experimental results (abstract and §5): the central claim that the QEP delivers a genuine, frontend-dependent improvement rests on comparisons whose classical baseline architecture (linear readout, MLP, or kernel method), parameter count, and hyperparameter tuning protocol are not specified, preventing isolation of any quantum contribution from the tensor-network compression itself.
[Experimental results] Experimental results (abstract and §5): no error bars, standard deviations across random seeds, or statistical significance tests are reported for the accuracy figures or stability claims, despite explicit statements about noiseless vs. noisy degradation and qubit-latent dimension matching.
[QEP description] QEP description (§3.3): the quantum-state embedding procedure, circuit ansatz, and choice of observable for readout are described only at a high level, making it impossible to assess whether the reported gains are reproducible or attributable to quantum mechanics rather than classical post-processing of the same latent vectors.

minor comments (2)

[Abstract] Abstract: the phrase 'noiseless vs. noisy comparisons' is used without naming the noise model (depolarizing, amplitude damping, etc.) or the simulator backend.
Notation: bond dimensions and latent-vector sizes for the three tensor-network frontends are mentioned but not tabulated or related explicitly to the qubit count used in the QEP.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript. We address each major comment below and will revise the manuscript to incorporate the suggested clarifications and additions.

read point-by-point responses

Referee: [Experimental results] Experimental results (abstract and §5): the central claim that the QEP delivers a genuine, frontend-dependent improvement rests on comparisons whose classical baseline architecture (linear readout, MLP, or kernel method), parameter count, and hyperparameter tuning protocol are not specified, preventing isolation of any quantum contribution from the tensor-network compression itself.

Authors: We agree that explicit specification of the classical baselines is required to isolate any quantum contribution. In the revised manuscript we will add a dedicated subsection in §5 describing the classical post-processing pipelines (linear readout, MLP with parameter count matched to the QEP, and kernel SVM), the exact hyperparameter search protocol (grid search over learning rate, hidden-layer sizes, and regularization), and the training schedule used for all methods. This will make the frontend-dependent gains attributable to the QEP rather than to differences in classical readout capacity. revision: yes
Referee: [Experimental results] Experimental results (abstract and §5): no error bars, standard deviations across random seeds, or statistical significance tests are reported for the accuracy figures or stability claims, despite explicit statements about noiseless vs. noisy degradation and qubit-latent dimension matching.

Authors: We acknowledge the absence of statistical reporting. The revised version will rerun all experiments over 10 independent random seeds, report mean accuracy with standard deviation, add error bars to all figures, and include paired t-tests (with p-values) for the noiseless-versus-noisy and qubit-dimension-matching comparisons. These additions will directly support the stability claims made in the text. revision: yes
Referee: [QEP description] QEP description (§3.3): the quantum-state embedding procedure, circuit ansatz, and choice of observable for readout are described only at a high level, making it impossible to assess whether the reported gains are reproducible or attributable to quantum mechanics rather than classical post-processing of the same latent vectors.

Authors: We will expand §3.3 with the concrete implementation details: angle embedding of each latent coordinate into an RY rotation on a dedicated qubit, a hardware-efficient ansatz consisting of L layers of RY rotations followed by nearest-neighbor CZ entangling gates, and readout via the expectation value of the Pauli-Z operator on the first qubit. These specifications will be accompanied by pseudocode and a circuit diagram to ensure reproducibility and to distinguish the quantum circuit from classical post-processing of the identical latent vectors. revision: yes

Circularity Check

0 steps flagged

No significant circularity; claims are empirical observations

full rationale

The paper proposes a hybrid framework using tensor-network frontends (MPS, TTN, MERA) for compression followed by MPC aggregation and QEP refinement, then reports experimental results on PneumoniaMNIST. Key statements such as 'the TTN+QEP combination exhibits the most balanced overall profile' and 'the QEP behaves more stably when the qubit count is sufficiently matched to the latent dimension' are presented as direct outcomes of architecture comparisons under noiseless and noisy conditions. No equations, fitted parameters, or first-principles derivations are shown that reduce these performance claims to self-definitions or internal inputs by construction. No self-citation load-bearing steps, uniqueness theorems, or ansatzes appear in the provided text. The derivation chain consists of a proposed architecture tested empirically, with communication cost tied to latent dimension as an observed property rather than a circular prediction.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available; no explicit free parameters, axioms, or invented entities are stated. The framework implicitly assumes tensor networks preserve diagnostic information in the latent space and that quantum observables on that space yield classification gains.

pith-pipeline@v0.9.0 · 5593 in / 1171 out tokens · 34131 ms · 2026-05-13T21:48:17.726382+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

client-side tensor-network frontends... Quantum-Enhanced Processor (QEP) refines the aggregated latent feature through quantum-state embedding and observable-based readout
IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Experiments on PneumoniaMNIST show that the effect of the QEP is frontend-dependent

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

45 extracted references · 45 canonical work pages

[1]

The future of digital health with federated learning,

N. Rieke, J. Hancox, W. Li, F. Milletari, H. R. Roth, and et al., “The future of digital health with federated learning,”npj Digital Medicine, vol. 3, no. 119, 2020

work page 2020
[2]

Feder- ated learning in medicine: Facilitating multi-institutional collaborations without sharing patient data,

M. J. Sheller, G. A. Reina, B. Edwards, J. Martin, and S. Bakas, “Feder- ated learning in medicine: Facilitating multi-institutional collaborations without sharing patient data,”Scientific Reports, vol. 10, no. 12598, 2020

work page 2020
[3]

Regulation (eu) 2016/679 (general data protection regulation),

European Parliament and Council, “Regulation (eu) 2016/679 (general data protection regulation),” 2016, official Journal of the European Union

work page 2016
[4]

Health insurance portability and accountability act of 1996,

U.S. Congress, “Health insurance portability and accountability act of 1996,” 1996

work page 1996
[5]

Act on the protection of personal information (japan),

Government of Japan, “Act on the protection of personal information (japan),” 2003, amended 2022

work page 2003
[6]

Act on anonymized medical data that are meant to contribute to research and development in the medical field,

——, “Act on anonymized medical data that are meant to contribute to research and development in the medical field,” 2017, act No. 28 of 2017

work page 2017
[7]

Proposal for a regulation on the european health data space,

European Commission, “Proposal for a regulation on the european health data space,” 2022, cOM(2022) 197 final

work page 2022
[8]

Communication-efficient learning of deep networks from decentralized data,

B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, “Communication-efficient learning of deep networks from decentralized data,”Proceedings of AISTATS, 2017

work page 2017
[9]

Advances and open problems in federated learning,

P. Kairouz and et al., “Advances and open problems in federated learning,”Foundations and Trends in Machine Learning, vol. 14, no. 1-2, pp. 1–210, 2021

work page 2021
[10]

Deep leakage from gradients,

L. Zhu, Z. Liu, and S. Han, “Deep leakage from gradients,” inNeurIPS, 2019

work page 2019
[11]

Membership inference attacks against machine learning models,

R. Shokri, M. Stronati, C. Song, and V . Shmatikov, “Membership inference attacks against machine learning models,” inIEEE S&P, 2017

work page 2017
[12]

Practical secure aggregation for privacy- preserving machine learning,

K. Bonawitz and et al., “Practical secure aggregation for privacy- preserving machine learning,” inProceedings of CCS, 2017

work page 2017
[13]

An invitation to distributed quantum neural networks,

L. Pira and C. Ferrie, “An invitation to distributed quantum neural networks,”Quantum Machine Intelligence, vol. 5, p. 23, 2023

work page 2023
[14]

Communication-efficient quantum algorithm for distributed machine learning,

H. Tang, B. Li, G. Wang, H. Xu, C. Li, A. Barr, P. Cappellaro, and J. Li, “Communication-efficient quantum algorithm for distributed machine learning,”Physical Review Letters, vol. 130, p. 150602, 2023

work page 2023
[15]

Review of distributed quantum computing: From single qpu to high performance quantum computing,

D. Barral, F. J. Cardama, G. D ´ıaz-Camacho, D. Failde, I. F. Llovo, M. Miras-Su ´arez, J. V ´azquez-P´erez, J. Villaluso, C. Pi ˜neiro, N. Costas, J. C. Piel, T. F. Pena, and A. G ´omez, “Review of distributed quantum computing: From single qpu to high performance quantum computing,” Computer Science Review, vol. 57, p. 100747, 2025

work page 2025
[16]

A design framework for the simulation of distributed quantum computing,

D. Ferrari and M. Amoretti, “A design framework for the simulation of distributed quantum computing,” inProceedings of the 2024 Workshop on High Performance and Quantum Computing Integration, ser. HPCQI ’24, 2024, pp. 4–10

work page 2024
[17]

Hybrid quantum-classical classifier based on tensor network and variational quantum circuit,

S. Y .-C. Chen, C.-M. Huang, C.-W. Hsing, and Y .-J. Kao, “Hybrid quantum-classical classifier based on tensor network and variational quantum circuit,”arXiv:2011.14651, 2020

work page arXiv 2011
[18]

Quantum simulation with hybrid tensor networks,

X.-Z. Luo, Z. Li, and J. Li, “Quantum simulation with hybrid tensor networks,”arXiv:2007.00958, 2020

work page arXiv 2007
[19]

Quantum-classical computing via tensor networks,

N. Tornow, C. B. Mendl, and P. Bhatotia, “Quantum-classical computing via tensor networks,”arXiv:2410.15080, 2024

work page arXiv 2024
[20]

Ted-q: a tensor network enhanced distributed hybrid quantum machine learning framework,

Y . Chen, C.-Y . Kuo, Y . Du, D. Tao, and X. Wu, “Ted-q: a tensor network enhanced distributed hybrid quantum machine learning framework,” arXiv:2301.05451, 2023

work page arXiv 2023
[21]

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

C.-Y . Liu, C.-H. A. Lin, and K.-C. Chen, “Quantum-train with tensor network mapping model and distributed circuit ansatz,” inICASSP 2025. IEEE, 2025

work page 2025
[22]

Tensor quantum programming,

A. Ternovaya, A. Melnikov, V . Mamenchikov, N. Belokonev, S. Dolgov, A. Berezutskii, R. Ellerbrock, A. Mansell, and M. Perelshtein, “Tensor quantum programming,”arXiv:2403.13486, 2024

work page arXiv 2024
[23]

Privacy- preserving quantum federated learning via gradient hiding,

C. Li, N. Kumar, Z. Song, S. Chakrabarti, and M. Pistoia, “Privacy- preserving quantum federated learning via gradient hiding,”Quantum Science and Technology, vol. 9, no. 3, 2024

work page 2024
[24]

Quantum-secure multiparty deep learning,

K. Sulimany, S. K. Vadlamani, R. Hamerly, P. Iyengar, and D. Englund, “Quantum-secure multiparty deep learning,”Physical Review X, vol. 15, no. 4, p. 041056, 2025

work page 2025
[25]

Quantum machine learning tensor network states,

A. Kardashin, A. Uvarov, and J. Biamonte, “Quantum machine learning tensor network states,”Frontiers in Physics, vol. 8, 2020

work page 2020
[26]

Spectral tensor train parameterization of deep learning layers,

A. Obukhov, M. Rakhuba, A. Liniger, Z. Huang, S. Georgoulis, D. Dai, and L. Van Gool, “Spectral tensor train parameterization of deep learning layers,” inProceedings of The 24th International Conference on Artificial Intelligence and Statistics, ser. Proceedings of Machine Learning Research, vol. 130, 2021, pp. 3547–3555

work page 2021
[27]

Residual tensor train: A quantum- inspired approach for learning multiple multilinear correlations,

Y . Chen, Y . Pan, and D. Dong, “Residual tensor train: A quantum- inspired approach for learning multiple multilinear correlations,”IEEE Transactions on Artificial Intelligence, vol. 4, pp. 1101–1113, Oct. 2023

work page 2023
[28]

Matrix product state pre-training for quantum machine learning,

J. Dborin, F. Barratt, V . Wimalaweera, L. Wright, and A. G. Green, “Matrix product state pre-training for quantum machine learning,” Quantum Science and Technology, vol. 7, no. 3, p. 035014, 2022

work page 2022
[29]

A matrix product state model for simultaneous classification and generation,

A. Mossi, B. ˇZunkovi´c, and K. Flouris, “A matrix product state model for simultaneous classification and generation,”Quantum Machine In- telligence, vol. 7, 2025

work page 2025
[30]

Quantum-inspired event reconstruction with tensor networks: Matrix product states,

J. Y . Araz and M. Spannowsky, “Quantum-inspired event reconstruction with tensor networks: Matrix product states,”Journal of High Energy Physics, no. 08, p. 112, 2021

work page 2021
[31]

Tensor-train methods for sequential state and parameter learning in state-space models,

Y . Zhao and T. Cui, “Tensor-train methods for sequential state and parameter learning in state-space models,”Journal of Machine Learning Research, vol. 25, pp. 1–51, 2024

work page 2024
[32]

Federated learning using coupled tensor train decomposition,

X. Zhang, E. Kofidis, C. Zhu, L. Zhang, and Y . Liu, “Federated learning using coupled tensor train decomposition,”arXiv:2403.02898, 2024

work page arXiv 2024
[33]

Federated hierarchical tensor networks: A collaborative learning quantum ai-driven framework for healthcare,

A. S. Bhatia and D. E. Bernal Neira, “Federated hierarchical tensor networks: A collaborative learning quantum ai-driven framework for healthcare,”arXiv:2405.07735, 2024

work page arXiv 2024
[34]

An efficient passive-to-active compiler for honest-majority MPC over rings,

M. Abspoel, A. P. K. Dalskov, D. Escudero, and A. Nof, “An efficient passive-to-active compiler for honest-majority MPC over rings,” in Applied Cryptography and Network Security - 19th International Con- ference, ACNS 2021, Kamakura, Japan, June 21-24, 2021, Proceedings, Part II, ser. Lecture Notes in Computer Science, vol. 12727. Springer, 2021, pp. 122–152

work page 2021
[35]

New primitives for actively-secure MPC over rings with applications to private machine learning,

I. Damg ˚ard, D. Escudero, T. K. Frederiksen, M. Keller, P. Scholl, and N. V olgushev, “New primitives for actively-secure MPC over rings with applications to private machine learning,” in2019 IEEE Symposium on Security and Privacy, SP 2019, San Francisco, CA, USA, May 19-23,

work page 2019
[36]

1102–1120

IEEE, 2019, pp. 1102–1120

work page 2019
[37]

Secure computation with fixed-point num- bers,

O. Catrina and A. Saxena, “Secure computation with fixed-point num- bers,” inFinancial Cryptography and Data Security, 14th International Conference, FC 2010, Tenerife, Canary Islands, Spain, January 25-28, 2010, Revised Selected Papers, ser. Lecture Notes in Computer Science, vol. 6052. Springer, 2010, pp. 35–50

work page 2010
[38]

Medmnist v2: A large-scale lightweight benchmark for 2d and 3d biomedical image classification,

J. Yanget al., “Medmnist v2: A large-scale lightweight benchmark for 2d and 3d biomedical image classification,”Scientific Data, 2023

work page 2023
[39]

Sirnn: A math library for secure RNN inference,

D. Rathee, M. Rathee, R. K. K. Goli, D. Gupta, R. Sharma, N. Chandran, and A. Rastogi, “Sirnn: A math library for secure RNN inference,” in 42nd IEEE Symposium on Security and Privacy, SP 2021, San Francisco, CA, USA, 24-27 May 2021. IEEE, 2021, pp. 1003–1020

work page 2021
[40]

Barren plateaus in variational quantum computing,

M. Larocca, S. Thanasilp, S. Wang, K. Sharma, J. Biamonte, P. J. Coles, L. Cincio, J. R. McClean, Z. Holmes, and M. Cerezo, “Barren plateaus in variational quantum computing,”Nature Reviews Physics, vol. 7, pp. 174–189, 2025

work page 2025
[41]

Exponential concentration in quantum kernel methods,

S. Thanasilp, S. Wang, M. Cerezo, and Z. Holmes, “Exponential concentration in quantum kernel methods,”Nature Communications, vol. 15, p. 5200, 2024

work page 2024
[42]

Nurdin, and Naoki Yamamoto

T. Yasuda, Y . Suzuki, T. Kubota, K. Nakajima, Q. Gao, W. Zhang, S. Shimono, H. I. Nurdin, and N. Yamamoto, “Quantum reservoir computing with repeated measurements on superconducting devices,” arXiv preprint arXiv:2310.06706, 2023

work page arXiv 2023
[43]

Universal blind quantum computation,

A. Broadbent, J. Fitzsimons, and E. Kashefi, “Universal blind quantum computation,” inProceedings of the 50th Annual IEEE Symposium on Foundations of Computer Science (FOCS 2009), 2009, pp. 517–526

work page 2009
[44]

Private quantum computation: an introduction to blind quantum computing and related protocols,

J. F. Fitzsimons, “Private quantum computation: an introduction to blind quantum computing and related protocols,”npj Quantum Information, vol. 3, p. 23, 2017

work page 2017
[45]

Qenclave - a practical solution for secure quantum cloud computing,

Y . Ma, E. Kashefi, M. Arapinis, K. Chakraborty, and M. Kaplan, “Qenclave - a practical solution for secure quantum cloud computing,” npj Quantum Information, vol. 8, p. 128, 2022

work page 2022