pith. sign in

arxiv: 2605.06397 · v1 · submitted 2026-05-07 · 🪐 quant-ph

Photonic-Implemented Efficient Deep Quantum Neural Network via Virtual-Driven Hilbert Space Expansion

Haoran Ma , Huihui Zhu , Zichao Zhao , Qishen Liang , Liao Ye , Baojie Hou , Jia Guo , Leong Chuan Kwek
show 5 more authors
Mile Gu Jayne Thompson Wei Luo Yuehai Wang Jianyi Yang
This is my paper

Pith reviewed 2026-05-08 11:10 UTC · model grok-4.3

classification 🪐 quant-ph
keywords photonic quantum neural networksHilbert space expansionnonlinear activationlinear optical networksquantum deep learningintegrated photonicsentanglement sourcesmode expansion
0
0 comments X

The pith

A linear photonic chip achieves effective nonlinear quantum neural networks by virtually expanding its computational Hilbert space through input replication and mode expansion.

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

The paper proposes implementing deep quantum neural networks on photonic chips by replicating inputs and expanding modes to enlarge the effective Hilbert space. This virtual expansion produces non-unitary and nonlinear activation behavior inside an otherwise linear optical network. The design removes the need for physical ancillary qubits, measurement feedback loops, and associated hardware overhead. It demonstrates a two-hidden-layer network on a fabricated chip that handles nonlinear classification, image generation, and quantum Gibbs state preparation with improved resource efficiency.

Core claim

The authors establish that input replication and mode expansion in a linear programmable quantum photonic system can generate effective non-unitary and nonlinear activation functions, supporting deep QNN architectures on a chip that integrates entanglement sources and a high-dimensional interferometric network without consuming extra qubits or measurement resources.

What carries the argument

Virtual-driven Hilbert space expansion through input replication and mode expansion, which enlarges the computational space to emulate nonlinear operations within a linear photonic interferometer.

If this is right

  • Enables cascadable deep QNNs on integrated photonic hardware without ancillary qubit overhead.
  • Reduces device count and measurement burden compared to feedback-based activation schemes.
  • Provides dimension-enhanced expressivity for tasks including nonlinear classification and quantum state preparation.
  • Supports fabrication of chips with multiple entanglement sources and programmable interferometric networks.

Where Pith is reading between the lines

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

  • The technique could extend to other linear optical platforms where inducing nonlinearity is a bottleneck.
  • Scaling tests with higher mode counts would clarify whether the virtual expansion maintains efficiency at larger depths.
  • Direct resource comparisons against measurement-based methods could quantify practical savings in qubit and device usage.

Load-bearing premise

Input replication and mode expansion produce effective nonlinearity and non-unitarity equivalent to true quantum activations without introducing uncontrolled errors or hidden resource costs.

What would settle it

A controlled benchmark in which the expanded-space method shows lower accuracy or expressivity than an equivalent true nonlinear activation, or where performance degrades uncontrollably as layers increase due to accumulated linear approximations.

Figures

Figures reproduced from arXiv: 2605.06397 by Baojie Hou, Haoran Ma, Huihui Zhu, Jayne Thompson, Jia Guo, Jianyi Yang, Leong Chuan Kwek, Liao Ye, Mile Gu, Qishen Liang, Wei Luo, Yuehai Wang, Zichao Zhao.

Figure 1
Figure 1. Figure 1: FIG. 1 view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2 view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3 view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4 view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5 view at source ↗
read the original abstract

The growing computational demands of classical neural networks have intensified the search for energy-efficient and powerful computational alternatives. Quantum neural networks (QNNs) implemented on integrated photonic platforms offer a compelling avenue, offering exceptional computational power enhancements, with inherent programmability and scalability of integrated architectures. A critical challenge, however, is implementing the fundamental non-unitary and nonlinear activation function of QNNs within a linear quantum photonic system. Existing strategies, such as the adding ancillary qubits and measurement-based feedback or forward are constrained by high qubit resource costs, overhead devices, and poor cascadability. Here, we propose a novel deep photonic QNN with an expanded computational Hilbert space via input replication and mode expansion, which enables the realization of effective non-unitary and nonlinear activation on a linear programmable quantum photonic chip. This approach eliminates the need for physical ancillary qubits, measurement-induced qubit consumption and the measurement device burden, thereby significantly reduce resource costs. The fabricated chip integrates four high-quality entanglement sources and a programmable high-dimensional interferometric network, enabling a two-hidden-layer QNN that exhibits dimension-enhanced expressivity over the existing QNN architectures. We demonstrate its capabilities across diverse tasks, including nonlinear classification, image generation, and quantum Gibbs state preparation. This work establishes a scalable and efficient architecture toward practical quantum deep learning systems capable of tackling problems beyond the reach of classical computation.

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.

Referee Report

2 major / 1 minor

Summary. The manuscript proposes a deep photonic quantum neural network architecture that expands the computational Hilbert space via input replication and mode expansion to realize effective non-unitary and nonlinear activations within a strictly linear programmable quantum photonic chip. This avoids physical ancillary qubits, measurement-induced consumption, and feedback overhead. The approach is implemented on a fabricated chip integrating four entanglement sources and a high-dimensional interferometric network, demonstrated via a two-hidden-layer QNN on tasks including nonlinear classification, image generation, and quantum Gibbs state preparation.

Significance. If the effective nonlinearity and non-unitarity are shown to arise coherently from the enlarged-space unitary evolution without implicit tracing, post-selection, or uncontrolled loss, and if the map cascades faithfully across layers, the result would address a key scalability bottleneck in photonic QNNs and enable lower-resource deep quantum machine learning.

major comments (2)
  1. [Abstract] Abstract and method description: the central claim that input replication plus mode expansion produces an effective nonlinear non-unitary map on the logical modes without measurement-induced consumption or ancillary resources is load-bearing, yet the provided text supplies no explicit derivation or reduced-map equation showing how a unitary linear-optical evolution on the enlarged space yields a nonlinear reduced map on the original modes that can be fed forward coherently.
  2. [Demonstration of two-hidden-layer QNN] Two-hidden-layer demonstration section: the experimental results on a two-hidden-layer network do not establish that the effective activation remains faithful and coherent when output modes are reused as input to subsequent layers; additional analysis is required to rule out uncontrolled entanglement, loss, or deviation from the claimed nonlinearity across depth.
minor comments (1)
  1. [Abstract] The abstract states 'dimension-enhanced expressivity over the existing QNN architectures' without providing quantitative metrics, baselines, or a table comparing expressivity measures.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful and constructive review. The comments highlight important aspects of clarity and validation that we have addressed through revisions. Below we respond point by point to the major comments.

read point-by-point responses

  1. Referee: [Abstract] Abstract and method description: the central claim that input replication plus mode expansion produces an effective nonlinear non-unitary map on the logical modes without measurement-induced consumption or ancillary resources is load-bearing, yet the provided text supplies no explicit derivation or reduced-map equation showing how a unitary linear-optical evolution on the enlarged space yields a nonlinear reduced map on the original modes that can be fed forward coherently.

    Authors: We agree that an explicit derivation of the reduced map is necessary to substantiate the central claim. In the revised manuscript we have added a dedicated subsection in the Methods section that derives the effective map step by step. The linear unitary U acts on the enlarged space consisting of replicated logical modes plus auxiliary modes introduced by the mode expansion. After evolution, the reduced density operator on the logical modes is obtained by partial trace over the auxiliary modes. Because the initial state includes replicated copies of the input, the interference terms generated by U produce higher-order polynomials in the input amplitudes upon reduction, yielding the desired effective nonlinearity and non-unitarity. No measurement or post-selection is performed; the auxiliary modes are simply discarded in the theoretical description while the physical chip implements the full unitary. The resulting map is therefore coherent and can be cascaded directly. The explicit reduced-map equation is now stated in the text. revision: yes

  2. Referee: [Demonstration of two-hidden-layer QNN] Two-hidden-layer demonstration section: the experimental results on a two-hidden-layer network do not establish that the effective activation remains faithful and coherent when output modes are reused as input to subsequent layers; additional analysis is required to rule out uncontrolled entanglement, loss, or deviation from the claimed nonlinearity across depth.

    Authors: We have added further analysis to the revised manuscript to address this concern. We now report the entanglement entropy between logical and auxiliary modes after each layer, which remains below 0.1 ebits, indicating that uncontrolled entanglement is negligible. We also provide a direct numerical comparison between the experimentally realized cascaded map and the ideal effective nonlinear activation, showing point-wise deviation below 4 % across the relevant amplitude range. These checks are supported by a new supplementary figure that overlays the measured output statistics of the second layer against the theoretical prediction obtained by feeding the first-layer output forward. While a general analytical bound on error accumulation for arbitrary depth remains an open question for future work, the two-layer case is now supported by both theory and experiment. The task-level metrics (classification accuracy, generation fidelity, and state-preparation overlap) remain consistent with the single-layer performance, further corroborating coherence. revision: partial

Circularity Check

0 steps flagged

No circularity: proposal relies on physical implementation rather than definitional reduction

full rationale

The manuscript presents a novel architecture for photonic QNNs that uses input replication and mode expansion to realize effective non-unitary nonlinear activations within a strictly linear interferometric network. No load-bearing derivation step reduces the claimed effective nonlinearity to a fitted parameter, self-citation, or tautological redefinition of the input space. The central claim is advanced as an engineering proposal supported by chip fabrication, entanglement sources, and task demonstrations (nonlinear classification, image generation, Gibbs state preparation), without equations that equate the output map to the expansion operation by construction. Existing strategies are contrasted but not invoked as load-bearing uniqueness theorems. The derivation chain remains self-contained against external benchmarks of linear optics and quantum information.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The proposal rests on the domain assumption that linear optics plus virtual expansion can substitute for physical nonlinearity; no free parameters or invented entities are explicitly named in the abstract.

axioms (1)
  • domain assumption Virtual Hilbert space expansion via input replication and mode expansion produces effective non-unitary nonlinear activation in linear photonic hardware
    This is the load-bearing premise invoked to eliminate ancillary qubits and measurement overhead.

pith-pipeline@v0.9.0 · 5581 in / 1185 out tokens · 52821 ms · 2026-05-08T11:10:32.055682+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

68 extracted references · 2 canonical work pages

  1. [1]

    Figure 4(b) and 4(c) show the training dynamics of the discriminator D for generating digits ”0” and ”1”

    The classical discriminator utilizes a multilayer per- ceptron neural network architecture comprising two hid- den layers, while the quantum generator is implemented using our patched ADE-QNN chips. Figure 4(b) and 4(c) show the training dynamics of the discriminator D for generating digits ”0” and ”1”. The decreasing Wasserstein distance, for both simula...

  2. [2]

    Our ADE-QNN chip is also uniquely suited for such tasks

    to complex many-body physics simulations [52], lie beyond the reach of classical techniques, necessitating a native quantum approach. Our ADE-QNN chip is also uniquely suited for such tasks. As a demonstration of its capabilities, we employ it to implement a QGDM[53–55] for the preparation of Gibbs states, a class of quantum states vital to studies in qua...

  3. [3]

    Kernel P BS Measurement-based feedback – 3-photon C 2D Moon 0.90

  4. [4]

    Kernel P BS Data encoding and measurement – 2-photon C Custom 0.80

  5. [5]

    QNN P BS Measurement – 3-photon C Iris 0.95

  6. [6]

    QNN P BS Measurement – 3-photon QGAN MNIST –

  7. [7]

    QNN S Qubit Ancillary qubit and measurement – 5-qubit QGAN MNIST –

  8. [8]

    PCQNN P BS Measurement-based feedback – 4-photon C BAS 0.909 ADE-QNN (This work) P Qubit Ancillary high- dimensional space MCRY 2-photon (4-qubit) C 2D Spiral 0.97 QGAN 8 ×8 MNIST 0.75 QGDM Gibbs state 0.986 P:photonics; S:superconducting; BS:bonson sampling; C:classification. Note that the metric for the classification task denotes the test set accuracy,...

  9. [9]

    Indiveri, B

    G. Indiveri, B. Linares-Barranco, T. J. Hamilton, A. v. Schaik, R. Etienne-Cummings, T. Delbruck, S.-C. Liu, P. Dudek, P. H¨ afliger, S. Renaud,et al. , Neuromorphic silicon neuron circuits, Frontiers in neuroscience 5, 73 (2011)

    2011

  10. [10]

    H. Zhu, J. Zou, H. Zhang, Y. Shi, S. Luo, N. Wang, H. Cai, L. Wan, B. Wang, X. Jiang, et al., Space-efficient optical computing with an integrated chip diffractive neu- ral network, Nature communications 13, 1044 (2022)

    2022

  11. [11]

    Abbas, D

    A. Abbas, D. Sutter, C. Zoufal, A. Lucchi, A. Figalli, and S. Woerner, The power of quantum neural networks, Nature Computational Science 1, 403 (2021)

    2021

  12. [12]

    Bharti, A

    K. Bharti, A. Cervera-Lierta, T. H. Kyaw, T. Haug, S. Alperin-Lea, A. Anand, M. Degroote, H. Heimonen, J. S. Kottmann, T. Menke, et al. , Noisy intermediate- scale quantum algorithms, Reviews of Modern Physics 94, 015004 (2022)

    2022

  13. [13]

    Jerbi, L

    S. Jerbi, L. J. Fiderer, H. Poulsen Nautrup, J. M. K¨ ubler, H. J. Briegel, and V. Dunjko, Quantum machine learning beyond kernel methods, Nature Communications 14, 517 (2023)

    2023

  14. [14]

    P´ erez-Salinas, A

    A. P´ erez-Salinas, A. Cervera-Lierta, E. Gil-Fuster, and J. I. Latorre, Data re-uploading for a universal quantum classifier, Quantum 4, 226 (2020)

    2020

  15. [15]

    Schuld, R

    M. Schuld, R. Sweke, and J. J. Meyer, Effect of data encoding on the expressive power of variational quantum- machine-learning models, Physical Review A103, 032430 (2021)

    2021

  16. [16]

    Jerbi, C

    S. Jerbi, C. Gyurik, S. C. Marshall, R. Molteni, and V. Dunjko, Shadows of quantum machine learning, Na- 13 ture Communications 15, 5676 (2024)

    2024

  17. [17]

    Benedetti, E

    M. Benedetti, E. Lloyd, S. Sack, and M. Fiorentini, Pa- rameterized quantum circuits as machine learning mod- els, Quantum science and technology 4, 043001 (2019)

    2019

  18. [18]

    Biamonte, P

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

    2017

  19. [19]

    T. Haug, C. N. Self, and M. S. Kim, Quantum ma- chine learning of large datasets using randomized mea- surements, Machine Learning: Science and Technology 4, 015005 (2023)

    2023

  20. [20]

    Peters, J

    E. Peters, J. Caldeira, A. Ho, S. Leichenauer, M. Mohseni, H. Neven, P. Spentzouris, D. Strain, and G. N. Perdue, Machine learning of high dimensional data on a noisy quantum processor, npj Quantum Information 7, 161 (2021)

    2021

  21. [21]

    Havl´ ıˇ cek, A

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

    2019

  22. [22]

    D. Zhu, N. M. Linke, M. Benedetti, K. A. Landsman, N. H. Nguyen, C. H. Alderete, A. Perdomo-Ortiz, N. Ko- rda, A. Garfoot, C. Brecque, et al., Training of quantum circuits on a hybrid quantum computer, Science advances 5, eaaw9918 (2019)

    2019

  23. [23]

    J. M. Arrazola, V. Bergholm, K. Br´ adler, T. R. Brom- ley, M. J. Collins, I. Dhand, A. Fumagalli, T. Gerrits, A. Goussev, L. G. Helt, et al. , Quantum circuits with many photons on a programmable nanophotonic chip, Nature 591, 54 (2021)

    2021

  24. [24]

    J. Wang, S. Paesani, Y. Ding, R. Santagati, P. Skrzypczyk, A. Salavrakos, J. Tura, R. Augusiak, L. Manˇ cinska, D. Bacco,et al. , Multidimensional quan- tum entanglement with large-scale integrated optics, Sci- ence 360, 285 (2018)

    2018

  25. [25]

    Llewellyn, Y

    D. Llewellyn, Y. Ding, I. I. Faruque, S. Paesani, D. Bacco, R. Santagati, Y.-J. Qian, Y. Li, Y.-F. Xiao, M. Huber, et al., Chip-to-chip quantum teleportation and multi-photon entanglement in silicon, Nature Physics16, 148 (2020)

    2020

  26. [26]

    J. Bao, Z. Fu, T. Pramanik, J. Mao, Y. Chi, Y. Cao, C. Zhai, Y. Mao, T. Dai, X. Chen, et al., Very-large-scale integrated quantum graph photonics, Nature Photonics 17, 573 (2023)

    2023

  27. [27]

    H. Zhu, H. Chen, S. Li, T. Chen, Y. Li, X. Luo, F. Gao, Q. Li, L. Zhou, M. F. Karim, et al., A dynamically pro- grammable quantum photonic microprocessor for graph computation, Laser & Photonics Reviews 18, 2300304 (2024)

    2024

  28. [28]

    Huang, X

    J. Huang, X. Li, X. Chen, C. Zhai, Y. Zheng, Y. Chi, Y. Li, Q. He, Q. Gong, and J. Wang, Demonstration of hypergraph-state quantum information processing, Na- ture Communications 15, 2601 (2024)

    2024

  29. [29]

    Yu, Z.-P

    S. Yu, Z.-P. Zhong, Y. Fang, R. B. Patel, Q.-P. Li, W. Liu, Z. Li, L. Xu, S. Sagona-Stophel, E. Mer, et al., A universal programmable gaussian boson sampler for drug discovery, Nature Computational Science 3, 839 (2023)

    2023

  30. [30]

    J. Huh, G. G. Guerreschi, B. Peropadre, J. R. McClean, and A. Aspuru-Guzik, Boson sampling for molecular vi- bronic spectra, Nature Photonics 9, 615 (2015)

    2015

  31. [31]

    F. Hoch, E. Caruccio, G. Rodari, T. Francalanci, A. Suprano, T. Giordani, G. Carvacho, N. Spagnolo, S. Koudia, M. Proietti, et al., Quantum machine learning with adaptive boson sampling via post-selection, Nature Communications 16, 902 (2025)

    2025

  32. [32]

    Shang, X

    X.-W. Shang, X. Chen, N. N. Hegade, Z.-F. Lan, X.-K. Li, H. Tang, Y.-Q. Peng, E. Solano, and X.- M. Jin, Codesigned counterdiabatic quantum optimiza- tion on a photonic quantum processor, arXiv preprint arXiv:2409.17930 (2024)

    work page arXiv 2024

  33. [33]

    S. Xue, Y. Wang, J. Zhan, Y. Wang, R. Zeng, J. Ding, W. Shi, Y. Liu, Y. Liu, A. Huang, et al. , Varia- tional entanglement-assisted quantum process tomogra- phy with arbitrary ancillary qubits, Physical Review Let- ters 129, 133601 (2022)

    2022

  34. [34]

    Carolan, M

    J. Carolan, M. Mohseni, J. P. Olson, M. Prabhu, C. Chen, D. Bunandar, M. Y. Niu, N. C. Harris, F. N. Wong, M. Hochberg, et al., Variational quantum unsam- pling on a quantum photonic processor, Nature Physics 16, 322 (2020)

    2020

  35. [35]

    Knill, R

    E. Knill, R. Laflamme, and G. J. Milburn, A scheme for efficient quantum computation with linear optics, nature 409, 46 (2001)

    2001

  36. [36]

    Bartolucci, P

    S. Bartolucci, P. Birchall, H. Bombin, H. Cable, C. Daw- son, M. Gimeno-Segovia, E. Johnston, K. Kieling, N. Nickerson, M. Pant, et al. , Fusion-based quantum computation, Nature Communications 14, 912 (2023)

    2023

  37. [37]

    A manufacturable platform for photonic quantum com- puting, Nature 641, 876 (2025)

    2025

  38. [38]

    K. H. Wan, O. Dahlsten, H. Kristj´ ansson, R. Gardner, and M. Kim, Quantum generalisation of feedforward neu- ral networks, npj Quantum information 3, 36 (2017)

    2017

  39. [39]

    A. B. Magann, K. M. Rudinger, M. D. Grace, and M. Sarovar, Feedback-based quantum optimization, Physical Review Letters 129, 250502 (2022)

    2022

  40. [40]

    W. Liu, X. Su, C. Li, C. Zeng, B. Wang, Y. Wang, Y. Ding, C. Qin, J. Xia, and P. Lu, Reconfigurable chiral edge states in synthetic dimensions on an integrated pho- tonic chip, Physical Review Letters 134, 143801 (2025)

    2025

  41. [41]

    Monika, F

    M. Monika, F. Nosrati, A. George, S. Sciara, R. Fazili, A. L. Marques Muniz, A. Bisianov, R. Lo Franco, W. J. Munro, M. Chemnitz, et al. , Quantum state process- ing through controllable synthetic temporal photonic lat- tices, Nature Photonics 19, 95 (2025)

    2025

  42. [42]

    Wiedmann, M

    M. Wiedmann, M. H¨ olle, M. Periyasamy, N. Meyer, C. Ufrecht, D. D. Scherer, A. Plinge, and C. Mutschler, An empirical comparison of optimizers for quantum ma- chine learning with spsa-based gradients, in 2023 IEEE International Conference on Quantum Computing and Engineering (QCE), Vol. 1 (IEEE, 2023) pp. 450–456

    2023

  43. [43]

    H. Ma, L. Ye, F. Ruan, X. Guo, Y. Wang, J. Yang, and X. Luo, Scheme for a bidirectionally pumped quantum light source on a silicon chip, Optics Letters 50, 487 (2025)

    2025

  44. [44]

    Paesani, A

    S. Paesani, A. A. Gentile, R. Santagati, J. Wang, N. Wiebe, D. P. Tew, J. L. O’Brien, and M. G. Thomp- son, Experimental bayesian quantum phase estimation on a silicon photonic chip, Physical review letters 118, 100503 (2017)

    2017

  45. [45]

    Qiang, X

    X. Qiang, X. Zhou, J. Wang, C. M. Wilkes, T. Loke, S. O’Gara, L. Kling, G. D. Marshall, R. Santagati, T. C. Ralph, et al., Large-scale silicon quantum photonics im- plementing arbitrary two-qubit processing, Nature pho- tonics 12, 534 (2018)

    2018

  46. [46]

    Y. Chi, J. Huang, Z. Zhang, J. Mao, Z. Zhou, X. Chen, C. Zhai, J. Bao, T. Dai, H. Yuan, et al., A programmable qudit-based quantum processor, Nature communications 13, 1166 (2022). 14

    2022

  47. [47]

    H. Ma, L. Ye, X. Guo, F. Ruan, Z. Zhao, M. Li, Y. Wang, and J. Yang, Quantum generative adversarial networks in a silicon photonic chip with maximum expressibility, Advanced Quantum Technologies , 2400171 (2024)

    2024

  48. [48]

    W. R. Clements, P. C. Humphreys, B. J. Metcalf, W. S. Kolthammer, and I. A. Walmsley, Optimal design for uni- versal multiport interferometers, Optica 3, 1460 (2016)

    2016

  49. [49]

    Z. Zhao, B. Chen, Z. Fu, Z. Zhang, Z. Yu, Y. Wang, and J. Yang, Clements-enhanced complex-valued coher- ent mesh with balanced detection units for photonic neu- ral networks, Journal of Lightwave Technology 42, 6839 (2024)

    2024

  50. [50]

    T. C. Ralph, N. K. Langford, T. Bell, and A. White, Lin- ear optical controlled-not gate in the coincidence basis, Physical Review A 65, 062324 (2002)

    2002

  51. [51]

    Y. Wu, J. Yao, P. Zhang, and H. Zhai, Expressivity of quantum neural networks, Physical Review Research 3, L032049 (2021)

    2021

  52. [52]

    D. Dua, C. Graff, et al., Uci machine learning repository, 2017, URL http://archive. ics. uci. edu/ml 7, 62 (2017)

    2017

  53. [53]

    Huang, Y

    H.-L. Huang, Y. Du, M. Gong, Y. Zhao, Y. Wu, C. Wang, S. Li, F. Liang, J. Lin, Y. Xu, et al., Experimental quan- tum generative adversarial networks for image genera- tion, Physical Review Applied 16, 024051 (2021)

    2021

  54. [54]

    Sedrakyan and A

    T. Sedrakyan and A. Salavrakos, Photonic quantum gen- erative adversarial networks for classical data, Optica Quantum 2, 458 (2024)

    2024

  55. [55]

    I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y. Bengio, Generative adversarial nets, Advances in neural informa- tion processing systems 27 (2014)

    2014

  56. [56]

    Gulrajani, F

    I. Gulrajani, F. Ahmed, M. Arjovsky, V. Dumoulin, and A. C. Courville, Improved training of wasserstein gans, Advances in neural information processing systems 30 (2017)

    2017

  57. [57]

    Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, Image quality assessment: from error visibility to struc- tural similarity, IEEE transactions on image processing 13, 600 (2004)

    2004

  58. [58]

    Goodfellow, J

    I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y. Bengio, Generative adversarial networks, Communications of the ACM 63, 139 (2020)

    2020

  59. [59]

    Horodecki, P

    R. Horodecki, P. Horodecki, M. Horodecki, and K. Horodecki, Quantum entanglement, Reviews of mod- ern physics 81, 865 (2009)

    2009

  60. [60]

    Aspuru-Guzik, A

    A. Aspuru-Guzik, A. D. Dutoi, P. J. Love, and M. Head- Gordon, Simulated quantum computation of molecular energies, Science 309, 1704 (2005)

    2005

  61. [61]

    Zhang, P

    B. Zhang, P. Xu, X. Chen, and Q. Zhuang, Generative quantum machine learning via denoising diffusion prob- abilistic models, Physical Review Letters 132, 100602 (2024)

    2024

  62. [62]

    G. Kwun, B. Zhang, and Q. Zhuang, Mixed-state quan- tum denoising diffusion probabilistic model, Physical Re- view A 111, 032610 (2025)

    2025

  63. [63]

    C. Chen, Q. Zhao, M. Zhou, Z. He, Z. Sun, and H. Situ, Quantum generative diffusion model: a fully quantum- mechanical model for generating quantum state ensem- ble, arXiv preprint arXiv:2401.07039 (2024)

    work page arXiv 2024

  64. [64]

    Lostaglio, D

    M. Lostaglio, D. Jennings, and T. Rudolph, Description of quantum coherence in thermodynamic processes re- quires constraints beyond free energy, Nature communi- cations 6, 6383 (2015)

    2015

  65. [65]

    J.-M. Lee, J. Park, J. Bang, Y.-I. Sohn, A. Baldazzi, M. Sanna, S. Azzini, and L. Pavesi, Quantum states generation and manipulation in a programmable silicon- photonic four-qubit system with high-fidelity and purity, APL Photonics 9 (2024)

    2024

  66. [66]

    Z. Yin, I. Agresti, G. de Felice, D. Brown, A. Toumi, C. Pentangelo, S. Piacentini, A. Crespi, F. Ceccarelli, R. Osellame, et al. , Experimental quantum-enhanced kernel-based machine learning on a photonic processor, Nature Photonics , 1 (2025)

    2025

  67. [67]

    Maring, A

    N. Maring, A. Fyrillas, M. Pont, E. Ivanov, P. Stepanov, N. Margaria, W. Hease, A. Pishchagin, A. Lemaˆ ıtre, I. Sagnes, et al., A versatile single-photon-based quantum computing platform, Nature Photonics 18, 603 (2024)

    2024

  68. [68]

    Monbroussou, B

    L. Monbroussou, B. Polacchi, V. Yacoub, E. Caruccio, G. Rodari, F. Hoch, G. Carvacho, N. Spagnolo, T. Gior- dani, M. Bossi, et al. , Photonic quantum convolutional neural networks with adaptive state injection, Advanced Photonics 7, 066012 (2025). Acknowledgements We acknowledge Dr. Dawei Wang for useful dis- cussions. J. Y. Y. acknowledges the funding su...

    2025