Pith. sign in

REVIEW 2 major objections 5 minor 44 references

Polarization and Orbital Angular Momentum Encoded Quantum Toffoli Gate Enabled by Diffractive Neural Networks

T0 review · 2 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read The paper reports a deterministic single-photon Toffoli gate encoded in polarization and orbital angular momentum, with process fidelity 94.05 ± 0.02%.

desk verdict Solid experimental Toffoli gate on a single photon, but the linear-scaling claim in the abstract is unsupported and should be cut or heavily qualified. read the letter →

arxiv 2411.17266 v1 pith:3UV6LFCR submitted 2024-11-26 quant-ph physics.optics

Qianke Wang , Dawei Lyu , Jun Liu , Jian Wang This is my paper · ORCID
classification quant-phphysics.optics MSC 81P68 PACS 03.67.Lx42.50.Tx
keywords quantumToffoligatesingle-photonqubitsorbitalangularmomentumpolarizationencodingdiffractiveneuralnetworkspatiallightmodulatorprocesstomographycontrolled-NOT
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that a three-qubit Toffoli gate can be implemented on a single photon by splitting the job between two optical properties: polarization carries the first control qubit, while the orbital angular momentum (the twist of the light's phase front) carries the second control and the target. A polarization-controlled diffractive neural network—trained phase patterns on a spatial light modulator—performs the required controlled-NOT on the orbital angular momentum modes only when the photon is horizontally polarized. The authors demonstrate this experimentally, reporting a truth table visibility of 97.27 ± 0.20% and a process fidelity of 94.05 ± 0.02% from quantum process tomography over 216 input states. The significance is that a direct three-qubit gate can replace networks of many two-qubit gates, and the optical hardware scales with a few phase planes rather than an exponentially growing stack of bulk optics.

What carries the argument

The load-bearing mechanism is the polarized diffractive neural network: an array of phase-only pixels on each of four layers acts as trainable neurons, connected by free-space diffraction modeled with the angular spectrum method and optimized by gradient descent to minimize the mean-squared error between the predicted and ideal output fields. When the network is displayed on a polarization-sensitive spatial light modulator aligned to horizontal polarization, only the $|H\rangle$ photon component traverses the diffractive layers and undergoes the OAM CNOT, while the $|V\rangle$ component passes unmodified; this realizes the controlled-CNOT decomposition of the Toffoli gate in a single compact optical path.

What would settle it

Send a vertically polarized photon carrying $|+3\rangle$ through the four phase planes and measure the output: if any population appears in horizontal-polarization modes or in OAM orders other than $\pm1$ and $\pm3$, the polarization-controlled assumption is violated. A cleaner test is to check that the device acts as the identity on every vertical-polarization input; a deviation larger than the reported process-fidelity uncertainty would refute the paper's central mechanism.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that the Toffoli operation $U_{\mathrm{Tof}} = |0\rangle\langle 0| \otimes I_4 + |1\rangle\langle 1| \otimes U_{\mathrm{CNOT}}$ can be realized by training a diffractive neural network to act as a polarization-controlled OAM CNOT gate. The eight computational basis states are encoded as $\{|V\rangle, |H\rangle\} \otimes \{|l|=1, |l|=3\} \otimes \{l<0, l>0\}$, and a set of four phase planes, loaded onto a polarization-sensitive spatial light modulator, transforms the horizontal component of the input while leaving the vertical component untouched. The authors report an average truth table visibility of $97.27\pm0.20\%$, output state fidelities between roughly $93\%$ and $98\%$ on probe states, and a quantum process fidelity of $94.05\pm0.02\%$, with simulations suggesting an upper bound of $99.09\%$ for the same design under ideal alignment.

Load-bearing premise

The whole construction rests on the assumption that the polarization-sensitive spatial light modulator applies the trained phase pattern to the horizontal polarization component and leaves the vertical component completely unaffected, while keeping all light inside the four encoded OAM modes; the measured 91.3% modulation efficiency for H relative to V shows that this separation is only approximately satisfied.

Editorial extensions

If this is right

  • A direct three-qubit Toffoli gate can be built from one polarization-sensitive spatial light modulator, a few wave plates, and holograms, eliminating the 23–29 discrete optical elements used in earlier single-photon OAM gates.
  • The same trained-network approach extends to other three-qubit controlled gates; simulations in the paper give process fidelities above 99% for CCH, Fredkin, and CCZ gates, up to 99.89%.
  • Because the gate is deterministic and not post-selected, it is compatible in principle with cascading into larger quantum circuits such as Grover search and Shor's algorithm.
  • Full quantum process tomography, not just truth-table checks, confirms the gate's quantum behavior with a process fidelity of 94.05 ± 0.02%.

Reading between the lines

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

  • A testable consequence the authors leave implicit: the gap between the measured 94.05% process fidelity and the simulated 99.09% upper bound is largely attributable to polarization crosstalk and phase-plane misalignment, so improving the 91.3% H/V modulation efficiency should push the fidelity toward that bound.
  • The same training procedure could be applied to larger OAM alphabets, such as $|l| = 1, 3, 5$, to encode more qubits per photon without adding phase planes, although the paper does not demonstrate this.
  • Because all three qubits live in the same photon, the 'entanglement' used here is a correlation between internal degrees of freedom rather than spatially separated particles; extending the scheme to genuinely nonlocal entanglement would require a device that modulates polarization and OAM jointly, which the paper notes as a future need.
  • Manuscript-internal note: the text cites a reference as '[42?]' when discussing iterative optimization of the DNN, leaving one supporting citation unresolved; this does not affect the reported experimental numbers but leaves that particular route to improving the theoretical upper bound without a complete citation.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 5 minor

Summary. The paper reports an experimental implementation of a three-qubit quantum Toffoli gate encoded in the polarization and orbital angular momentum (OAM) degrees of freedom of a single photon. The gate is realized with a polarization-sensitive spatial light modulator loaded with four phase patterns trained as a diffractive neural network (DNN), which implements a polarization-controlled CNOT between the amplitude and sign OAM qubits. The authors characterize the gate with an 8-state truth table (mean visibility 97.27±0.20%), quantum state tomography on 216 probe states (state fidelities approximately 93–98%), and full three-qubit quantum process tomography with maximum-likelihood reconstruction, reporting a process fidelity of 94.05±0.02%. They also present a simulated 'upper bound' performance of 99.09% process fidelity and simulated extensions to other three-qubit gates. The central claim is that this DNN-based approach realizes a compact Toffoli gate without requiring exponential optical elements.

Significance. If the experimental result is taken at face value, this is a credible demonstration of a deterministic single-photon three-qubit gate using a programmable diffractive network, and the characterization is unusually thorough: a full truth table, tomography on 216 inputs, and complete quantum process tomography with Monte Carlo uncertainties. This is a meaningful step for OAM-based photonic quantum logic, showing that trained phase masks can replace bulk interferometers for a fixed small Hilbert space. However, the advertised scalability advantage—'without requiring exponential optical elements' and the linear vs. exponential comparison in Fig. 1(c)—is not established and, as argued below, requires a quantitative resource analysis that accounts for the space–bandwidth product of the phase planes. The experimental demonstration itself does not depend on the scalability claim, so the core result is defensible once the overreach is removed.

major comments (2)
  1. [Introduction; Fig. 1(c); Abstract] The claim that the DNN approach realizes multi-qubit gates 'without requiring exponential optical elements' and with optical resources scaling linearly with qubit number (Fig. 1(c), abstract, and conclusion) is not supported by the manuscript. Each diffractive layer is a phase-only mask with a finite pixel count P, and the forward model in S1 samples the field on an n×n grid; a single layer can independently act on at most O(P) transverse spatial modes. Encoding n qubits in the OAM degree of freedom requires 2^n mutually orthogonal modes, so P must grow at least as 2^n just to represent the input and output basis, independent of the number of layers. Thus the total number of programmable degrees of freedom scales exponentially with qubit number. The citation to Ref. [35] concerns the classification capacity of diffractive surfaces, not unitary synthesis in an exponentially large mode space, and the present experiment uses only a fixed four-mode OAM subspace (|l|=1 and |l|=3, with sign). No analysis or data is provided for n>3. Please either retract the linear-scaling claim or replace it with a resource count that includes the space–bandwidth product of the phase planes; otherwise the abstract and Fig. 1(c) overstate the method's scalability.
  2. [S5; main text 'simulated upper bound'] The phrase 'theoretical upper bound performance' for the simulation in S5 is misleading. The simulation evaluates the trained DNN model against the ideal Toffoli transformation, so it gives the performance of one particular trained network, not an upper bound over all possible designs or a fundamental limit. The discrepancy with the experimental fidelity is then attributed to apparatus imperfections, which is reasonable, but the '99.09%' should be described as the ideal-model fidelity of the trained phase patterns. This is a load-bearing point only insofar as the paper presents the simulation as a property of the design rather than as a trained-instance benchmark; it should be reworded for accuracy.
minor comments (5)
  1. [S1, Eq. (S15)] The text calls Eq. (S15) a 'mean squared error (MSE)' loss, but the formula L = (1/n^2) Σ |g - ĝ| is a mean absolute error. Either the formula should include a square (or the square root), or the text should say 'mean absolute error'.
  2. [S2] The first sentence of S2 contains a typo: 'depcited' should be 'depicted'.
  3. [S3, Fig. S4] The text 'visible in Fig. R2(b)' should refer to Fig. S4(b); the 'R2' label appears to be an editing artifact.
  4. [References] Reference [42] is written as '[42? ]' with a stray question mark; please clean this up.
  5. [Fig. 1(c) caption] The caption says 'Exponential and linear demands for optics related to the number of qubits' but does not define what counts as an 'optic' in each scheme; adding the definition (e.g., discrete bulk elements vs. phase-mask pixels vs. phase-mask count) would help the reader evaluate the comparison.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the experimental Toffoli gate characterization is measured independently of the DNN training objective; the only caveat is a non-load-bearing simulated 'upper bound' that simply re-evaluates the trained network.

full rationale

The central claim is an experimental demonstration, not a first-principles prediction. The phase planes are trained with a gradient-descent MSE loss (Supplemental S1, Eq. S15) to approximate the CNOT, and the paper then independently measures the truth table (97.27±0.20%), quantum state tomography on 216 states, and quantum process tomography (process fidelity 94.05±0.02%) against the ideal Toffoli unitary. These experimental quantities are not read off the training loss; they are separate measurement data, so the central result is self-contained. The simulated 'theoretical upper bound' in the main text and Supplemental S5 is explicitly computed from 'output states ... as obtained during the DNN training process,' so it is an evaluation of the fitted design rather than an independent derivation; however, it is presented only as an idealized performance estimate and does not carry the experimental claim. The scalability assertion ('without requiring exponential optical elements') relies on an external citation [35] about diffractive-network capacity and is an extrapolation whose correctness is a separate concern, not a circular reduction to the paper's own inputs. No load-bearing self-citations or imported uniqueness theorems appear.

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

The central claim does not rest on fitted parameters; the DNN phase values are trained to realize the ideal CNOT and are implementation details. The physical assumptions are the standard angular spectrum diffraction model, the polarization-selective behavior of the SLM, and the isolation of the OAM computational subspace. No new entities are introduced.

assumptions (4)
  • standard math The angular spectrum method accurately models free-space diffraction between phase planes.
    Used in S1 forward propagation model (Eqs S12-S14).
  • domain assumption The liquid-crystal SLM modulates only the horizontal polarization component and leaves vertical component unaffected.
    Central to polarization-controlled operation (Fig 1e, main text); deviations contribute to error, measured as 91.3% relative efficiency.
  • domain assumption The chosen OAM modes (|l|=1,3; positive/negative sign) form an isolated 4D computational subspace with negligible crosstalk to other OAM orders.
    Required for the amplitude/sign qubit encoding (Eq 2); leakage would cause errors.
  • domain assumption Gradient descent training of the DNN converges to a physically realizable phase pattern implementing the CNOT.
    Verified in simulation (S5) and indirectly by experimental fidelities.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Polarization and Orbital Angular Momentum Encoded Quantum Toffoli Gate Enabled by Diffractive Neural Networks." pith.science (2026). https://pith.science/paper/3UV6LFCR

@misc{pith2026241117266,
  author       = {Pith},
  title        = {Pith review of: Polarization and Orbital Angular Momentum Encoded Quantum Toffoli Gate Enabled by Diffractive Neural Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3UV6LFCR}},
  note         = {Machine review of arXiv:2411.17266}
}
abstract

Controlled quantum gates play a crucial role in enabling quantum universal operations by facilitating interactions between qubits. Direct implementation of three-qubit gates simplifies the design of quantum circuits, thereby being conducive to performing complex quantum algorithms. Here, we propose and present an experimental demonstration of a quantum Toffoli gate fully exploiting the polarization and orbital angular momentum of a single photon. The Toffoli gate is implemented using the polarized diffractive neural networks scheme, achieving a mean truth table visibility of $97.27\pm0.20\%$. We characterize the gate's performance through quantum state tomography on 216 different input states and quantum process tomography, which yields a process fidelity of $94.05\pm 0.02\%$. Our method offers a novel approach for realizing the Toffoli gate without requiring exponential optical elements while maintaining extensibility to the implementation of other three-qubit gates.

Figures

Figures reproduced from arXiv: 2411.17266 by the authors.

Figure 1
Figure 1. FIG. 1. Concept of the polarization and orbital angular mo [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Experimental setup. (a) Scheme of the setup. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 4
Figure 4. FIG. 4. Reconstructed density matrices of (a) [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: FIG. 5. Quantum process tomography of the Toffoli gate. [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 32 canonical work pages

  1. [35]

    Kulce, D

    O. Kulce, D. Mengu, Y. Rivenson, and A. Ozcan, All- optical information-processing capacity of diffractive sur- faces, Light: Science & Applications 10, 25 (2021)

    work page 2021

  2. [1]

    Barenco, C

    A. Barenco, C. H. Bennett, R. Cleve, D. P. DiVincenzo, N. Margolus, P. Shor, T. Sleator, J. A. Smolin, and H. Weinfurter, Elementary gates for quantum computa- tion, Phys. Rev. A 52, 3457 (1995)

    1995

  3. [2]

    Bravyi and A

    S. Bravyi and A. Kitaev, Universal quantum computa- tion with ideal clifford gates and noisy ancillas, Phys. Rev. A 71, 022316 (2005)

    2005

  4. [3]

    J. E. Bourassa, R. N. Alexander, M. Vasmer, A. Patil, I. Tzitrin, T. Matsuura, D. Su, B. Q. Baragiola, S. Guha, G. Dauphinais, K. K. Sabapathy, N. C. Menicucci, and I. Dhand, Blueprint for a Scalable Photonic Fault- Tolerant Quantum Computer, Quantum 5, 392 (2021)

    2021

  5. [4]

    Knill, R

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

    2001

  6. [5]

    Boixo, S

    S. Boixo, S. V. Isakov, V. N. Smelyanskiy, R. Babbush, N. Ding, Z. Jiang, M. J. Bremner, J. M. Martinis, and H. Neven, Characterizing quantum supremacy in near- term devices, Nat. Phys. 14, 595 (2018)

    2018

  7. [6]

    D. Qin, Y. Chen, and Y. Li, Error statistics and scalabil- ity of quantum error mitigation formulas, npj Quantum Inf. 9, 35 (2023)

    work page 2023

  8. [7]

    B. P. Lanyon, M. Barbieri, M. P. Almeida, T. Jennewein, T. C. Ralph, K. J. Resch, G. J. Pryde, J. L. O’Brien, A. Gilchrist, and A. G. White, Simplifying quantum logic using higher-dimensional hilbert spaces, Nat. Phys. 5, 134 (2009)

    work page 2009

Show all 44 references
  1. [8]

    Takeda, A

    K. Takeda, A. Noiri, T. Nakajima, T. Kobayashi, and 6 S. Tarucha, Quantum error correction with silicon spin qubits, Nature 608, 682 (2022)

    2022

  2. [9]

    Figgatt, D

    C. Figgatt, D. Maslov, K. A. Landsman, N. M. Linke, S. Debnath, and C. Monroe, Complete 3-qubit grover search on a programmable quantum computer, Nat. Commun. 8, 1918 (2017)

    2017

  3. [10]

    J. Chu, X. He, Y. Zhou, J. Yuan, L. Zhang, Q. Guo, Y. Hai, Z. Han, C.-K. Hu, W. Huang, H. Jia, D. Jiao, S. Li, Y. Liu, Z. Ni, L. Nie, X. Pan, J. Qiu, W. Wei, W. Nuerbolati, Z. Yang, J. Zhang, Z. Zhang, W. Zou, Y. Chen, X. Deng, X. Deng, L. Hu, J. Li, S. Liu, Y. Lu, J. Niu, D. ...

    2023

  4. [11]

    Gidney and M

    C. Gidney and M. Eker ˚ a, How to factor 2048 bit RSA in- tegers in 8 hours using 20 million noisy qubits, Quantum 5, 433 (2021)

    2021

  5. [12]

    N. Yu, R. Duan, and M. Ying, Five two-qubit gates are necessary for implementing the toffoli gate, Phys. Rev. A 88, 010304 (2013)

    2013

  6. [13]

    T. Monz, K. Kim, W. H¨ ansel, M. Riebe, A. S. Villar, P. Schindler, M. Chwalla, M. Hennrich, and R. Blatt, Realization of the Quantum Toffoli Gate with Trapped Ions, Phys. Rev. Lett. 102, 040501 (2009)

    2009

  7. [14]

    J. D. Arias Espinoza, K. Groenland, M. Mazzanti, K. Schoutens, and R. Gerritsma, High-fidelity method for a single-step n-bit toffoli gate in trapped ions, Phys. Rev. A 103, 052437 (2021)

    2021

  8. [15]

    Levine, A

    H. Levine, A. Keesling, G. Semeghini, A. Omran, T. T. Wang, S. Ebadi, H. Bernien, M. Greiner, V. Vuleti´ c, H. Pichler, and M. D. Lukin, Parallel implementation of high-fidelity multiqubit gates with neutral atoms, Phys. Rev. Lett. 123, 170503 (2019)

    2019

  9. [16]

    Fedorov, L

    A. Fedorov, L. Steffen, M. Baur, M. P. da Silva, and A. Wallraff, Implementation of a toffoli gate with super- conducting circuits, Nature 481, 170 (2012)

    2012

  10. [17]

    A. D. Hill, M. J. Hodson, N. Didier, and M. J. Reagor, Realization of arbitrary doubly-controlled quan- tum phase gates (2021), arXiv:2108.01652 [quant-ph]

    2021 arXiv

  11. [18]

    Y. Kim, A. Morvan, L. B. Nguyen, R. K. Naik, C. J √ˇlnger, L. Chen, J. M. Kreikebaum, D. I. Santiago, and I. Siddiqi, High-fidelity three-qubit itoffoli gate for fixed-frequency superconducting qubits, Nat. Phys. 18, 783 (2022)

    2022

  12. [19]

    Carolan, C

    J. Carolan, C. Harrold, C. Sparrow, E. Mart ´ ın-L´ opez, N. J. Russell, J. W. Silverstone, P. J. Shadbolt, N. Mat- suda, M. Oguma, M. Itoh, G. D. Marshall, M. G. Thomp- son, J. C. F. Matthews, T. Hashimoto, J. L. O’Brien, and A. Laing, Universal linear optics, Science 349, 711 (2015)

    2015

  13. [20]

    Miˇ cuda, M

    M. Miˇ cuda, M. Sedl´ ak, I. Straka, M. Mikov´ a, M. Duˇ sek, M. Jeˇ zek, and J. Fiur´ aˇ sek, Efficient experimental esti- mation of fidelity of linear optical quantum toffoli gate, Phys. Rev. Lett. 111, 160407 (2013)

    2013

  14. [21]

    St´ arek, M

    R. St´ arek, M. Miˇ cuda, M. Mikov´ a, I. Straka, M. Duˇ sek, M. Jeˇ zek, and J. Fiur´ aˇ sek, Experimental investigation of a four-qubit linear-optical quantum logic circuit, Sci. Rep. 6, 33475 (2016)

    2016

  15. [22]

    M. Li, C. Li, Y. Chen, L.-T. Feng, L. Yan, Q. Zhang, J. Bao, B.-H. Liu, X.-F. Ren, J. Wang, S. Wang, Y. Gao, X. Hu, Q. Gong, and Y. Li, On-chip path encoded pho- tonic quantum toffoli gate, Photon. Res. 10, 1533 (2022)

    2022

  16. [23]

    Y. Li, L. Wan, H. Zhang, H. Zhu, Y. Shi, L. K. Chin, X. Zhou, L. C. Kwek, and A. Q. Liu, Quantum fredkin and toffoli gates on a versatile programmable silicon pho- tonic chip, npj Quantum Inf. 8, 112 (2022)

    2022

  17. [24]

    Liu and H.-R

    W.-Q. Liu and H.-R. Wei, Linear optical universal quan- tum gates with higher success probabilities, Advanced Quantum Technologies 6, 2300009 (2023)

    2023

  18. [25]

    S. Ru, Y. Wang, M. An, F. Wang, P. Zhang, and F. Li, Realization of a deterministic quantum Toffoli gate with a single photon, Phys. Rev. A 103, 022606 (2021)

    2021

  19. [26]

    Pan, Z.-B

    J.-W. Pan, Z.-B. Chen, C.-Y. Lu, H. Weinfurter, A. Zeilinger, and M. ˙Zukowski, Multiphoton entangle- ment and interferometry, Rev. Mod. Phys. 84, 777 (2012)

    2012

  20. [27]

    Flamini, N

    F. Flamini, N. Spagnolo, and F. Sciarrino, Photonic quantum inf. process.: a review, Rep. Prog. Phys. 82, 016001 (2018)

    2018

  21. [28]

    Brandt, M

    F. Brandt, M. Hiekkam¨ aki, F. Bouchard, M. Huber, and R. Fickler, High-dimensional quantum gates using full- field spatial modes of photons, Optica 7, 98 (2020)

    2020

  22. [29]

    Babazadeh, M

    A. Babazadeh, M. Erhard, F. Wang, M. Malik, R. Nouroozi, M. Krenn, and A. Zeilinger, High- Dimensional Single-Photon Quantum Gates: Concepts and Experiments, Phys. Rev. Lett. 119, 180510 (2017)

    2017

  23. [30]

    Y. Wang, S. Ru, F. Wang, P. Zhang, and F. Li, Ex- perimental demonstration of efficient high-dimensional quantum gates with orbital angular momentum, Quan- tum Science & Technology 7, 015016 (2022)

    2022

  24. [31]

    F. Wang, S. Ru, Y. Wang, M. An, P. Zhang, and F. Li, Experimental demonstration of a quantum controlled- SW AP gate with multiple degrees of freedom of a sin- gle photon, Quantum Science & Technology 6, 035005 (2021)

    2021

  25. [32]

    Q. Zeng, T. Li, X. Song, and X. Zhang, Realization of optimized quantum controlled-logic gate based on the or- bital angular momentum of light, Opt. Express 24, 8186 (2016)

    2016

  26. [33]

    K. H. Kagalwala, G. D. Giuseppe, A. F. Abouraddy, and B. E. A. Saleh, Single-photon three-qubit quantum logic using spatial light modulators, Nat. Commun. 8, 739 (2017)

    2017

  27. [34]

    X. Lin, Y. Rivenson, N. T. Yardimci, M. Veli, Y. Luo, M. Jarrahi, and A. Ozcan, All-optical machine learning using diffractive deep neural networks, Science 361, 1004 (2018)

    2018

  28. [36]

    Bolduc, N

    E. Bolduc, N. Bent, E. Santamato, E. Karimi, and R. W. Boyd, Exact solution to simultaneous intensity and phase encryption with a single phase-only hologram, Opt. Lett. 38, 3546 (2013)

    2013

  29. [37]

    D. F. V. James, P. G. Kwiat, W. J. Munro, and A. G. White, Measurement of qubits, Phys. Rev. A 64, 052312 (2001)

    2001

  30. [38]

    Jeˇ zek, J

    M. Jeˇ zek, J. Fiur´ aˇ sek, and Z. c. v. Hradil, Quantum in- ference of states and processes, Phys. Rev. A 68, 012305 (2003)

    2003

  31. [39]

    Fiurasek and Z

    J. Fiurasek and Z. Hradil, Maximum-likelihood estima- tion of quantum processes, Physical Review A63, 020101 (2001)

    2001

  32. [40]

    Gilchrist, N

    A. Gilchrist, N. K. Langford, and M. A. Nielsen, Distance measures to compare real and ideal quantum processes, Phys. Rev. A 71, 062310 (2005)

    2005

  33. [41]

    Mengu, Y

    D. Mengu, Y. Zhao, N. T. Yardimci, Y. Rivenson, M. Jar- 7 rahi, and A. Ozcan, Misalignment resilient diffractive op- tical networks, Nanophotonics 9, 4207 (2020)

    2020

  34. [42]

    M. S. S. Rahman, J. Li, D. Mengu, Y. Rivenson, and A. Ozcan, Ensemble learning of diffractive optical net- works, Light: Science & Applications 10, 14 (2021)

    2021

  35. [43]

    X. Luo, Y. Hu, X. Ou, X. Li, J. Lai, N. Liu, X. Cheng, A. Pan, and H. Duan, Metasurface-enabled on-chip mul- tiplexed diffractive neural networks in the visible, Light: Science & Applications 11, 158 (2022)

    2022

  36. [44]

    Zhang, Z

    J. Zhang, Z. Ye, J. Yin, L. Lang, and S. Jiao, Polarized deep diffractive neural network for sorting, generation, multiplexing, and de-multiplexing of orbital angular mo- mentum modes, Opt. Express 30, 26728 (2022). Appendix S1: F orward propagation model, loss function and tr...

    2022

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.