pith. sign in

arxiv: 2605.19611 · v1 · pith:FPRRGX6Nnew · submitted 2026-05-19 · 💻 cs.CV · cs.ET

Inverse Design of Metasurface based Absorbers using Physics Guided Conditional Diffusion Models

Vineetha Joy , Jamshed Palai , Satwik Sahoo , Anshuman Kumar , Amit Sethi , Hema Singh This is my paper

Pith reviewed 2026-05-20 05:22 UTC · model grok-4.3

classification 💻 cs.CV cs.ET
keywords inverse designmetasurface absorbersconditional diffusion modelsphysics-guided learningelectromagnetic simulationgenerative modelingsurrogate regularization
0
0 comments X

The pith

A physics-guided conditional diffusion model generates metasurface absorber designs that match target reflection spectra with spectral MSE of 0.0006 in about 30 seconds.

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

The paper establishes a generative framework for the inverse design of metasurface absorbers that produces geometries satisfying specified reflection characteristics across 2 to 18 GHz. It conditions a diffusion process on target spectra using feature-wise linear modulation and adds spectrum-level loss terms from a pre-trained surrogate electromagnetic simulator to enforce physical accuracy and manufacturability. A sympathetic reader would care because conventional optimization requires repeated full-wave simulations that take months, whereas this approach samples suitable designs rapidly and supplies multiple alternatives per target.

Core claim

The physics guided condition quality enhanced diffusion framework integrates target reflection characteristics via feature wise linear modulation and embeds a pre-trained surrogate EM simulator to introduce physics aware regularization through spectrum level loss functions, achieving an average spectral mean squared error of 0.0006 and band alignment accuracy of 0.958 while generating practically realizable metasurfaces.

What carries the argument

The conditional diffusion model with FiLM conditioning on target spectra and spectrum-level regularization from a surrogate EM simulator, which guides generation toward physically consistent and fabricable metasurface geometries.

If this is right

  • Multiple distinct yet valid geometries can be produced for any given target spectrum, giving engineers design choices.
  • Generation time drops to roughly 30 seconds under comparable resources, replacing months-long iterative optimization loops.
  • The approach has been shown to produce designs validated by experimental measurements in the 2 to 18 GHz band.

Where Pith is reading between the lines

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

  • Similar conditioning and surrogate-regularization techniques could be applied to inverse design of other electromagnetic devices such as antennas or filters.
  • The framework reduces dependence on repeated full-wave simulations during the design cycle, freeing computational resources for larger parameter spaces.
  • Diverse outputs per condition may enable secondary selection criteria such as fabrication tolerance or material constraints.

Load-bearing premise

The pre-trained surrogate electromagnetic simulator must accurately and without bias approximate the full-wave responses of the generated metasurface geometries.

What would settle it

Fabricate a generated metasurface design and measure its experimental reflection spectrum; a large mismatch between the measured spectrum and both the target and the surrogate prediction would falsify the claimed fidelity.

Figures

Figures reproduced from arXiv: 2605.19611 by Amit Sethi, Anshuman Kumar, Hema Singh, Jamshed Palai, Satwik Sahoo, Vineetha Joy.

Figure 1
Figure 1. Figure 1: Illustration of forward and inverse design of metasurface [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Forward and backward processes in a diffusion model. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Architecture of the proposed inverse design framework (a) Complete block diagram (b) A FiLM layer ( [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Configuration of metasurface based RAS included in dataset (a) [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Evolution of the meta-atom geometry from random noise during [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Comparison between target spectra and the spectra obtained on simulation of generated meta-atoms for different samples [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Diverse metasurface geometries generated by the proposed [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: (a) Target reflection spectra (b) Meta atom geometry [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Fabricated prototype of planar RAS (15cm x 15cm) (a) Top [PITH_FULL_IMAGE:figures/full_fig_p009_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: (a) Experimental set up for measurement of S [PITH_FULL_IMAGE:figures/full_fig_p009_10.png] view at source ↗
read the original abstract

Inverse design of metasurfaces for specific electromagnetic responses requires generating geometries that satisfy stringent spectral constraints while maintaining manufacturability. Conventional design methodologies rely on iterative optimization routines using full wave simulations, which become extremely time consuming and computationally intensive for large design spaces. In addition, commonly employed generative approaches often exhibit limited conditional fidelity and the generated designs often contain fine or irregular features that are impractical to fabricate. In this regard, we propose a physics guided condition quality enhanced diffusion framework for the inverse design of metasurface based absorbers. Here, the conditioning information consisting of target reflection characteristics is integrated into the model using feature wise linear modulation (FiLM). Furthermore, to enforce adherence to target spectra, a pre trained surrogate EM simulator is embedded into the framework introducing physics aware regularization through spectrum level loss functions. The efficiency of the proposed model is demonstrated by generating practically realizable metasurfaces for different types of reflection characteristics in the frequency range of 2 to 18 GHz. The proposed framework achieves an average spectral mean squared error of 0.0006 and band alignment accuracy of 0.958 between the target spectra and the spectra produced by the generated designs, demonstrating high conditional accuracy. In addition, the model generates multiple geometries for the same condition, thereby providing diverse design alternatives to the engineer. The proposed model produces the suitable design in approximately 30 seconds, whereas the conventional approach can take several months under comparable computational resources. The efficiency of the model is also established via experimental measurements.

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 / 2 minor

Summary. The manuscript proposes a physics-guided conditional diffusion model for the inverse design of metasurface absorbers in the 2–18 GHz range. Target reflection spectra are injected via FiLM conditioning, and a pre-trained surrogate electromagnetic simulator supplies spectrum-level loss terms to enforce physical fidelity during training and sampling. The framework is reported to produce manufacturable geometries with an average spectral MSE of 0.0006 and band-alignment accuracy of 0.958, generate multiple design alternatives per target, and complete inference in approximately 30 seconds versus months for conventional optimization. Experimental measurements on fabricated samples are cited in support of the results.

Significance. If the surrogate-to-full-wave discrepancy remains small on the generated geometries, the approach would offer a practical acceleration of metasurface absorber design while preserving spectral fidelity and manufacturability. The combination of conditional diffusion with embedded physics regularization and the provision of diverse solutions per target are potentially useful contributions to the inverse-design literature.

major comments (2)
  1. [Results / Experimental Validation] Results section (performance metrics): the reported spectral MSE of 0.0006 and band-alignment accuracy of 0.958 are obtained by comparing target spectra against spectra produced by the surrogate simulator on the generated geometries. Because the diffusion model is trained with spectrum-level losses derived from the same surrogate, these figures risk being partly tautological; a table comparing surrogate predictions, independent full-wave simulations, and measured data on the identical fabricated samples is required to demonstrate that the low error reflects Maxwell fidelity rather than surrogate fidelity.
  2. [Methods / Surrogate EM Simulator] Methods (surrogate training): the manuscript does not report the surrogate’s own validation accuracy, training-set size, or out-of-distribution error on geometries similar to those produced by the diffusion model. Without these quantities it is difficult to bound the systematic bias that could be introduced into the physics-aware loss.
minor comments (2)
  1. [Figures] Figure captions should explicitly state whether the plotted spectra are surrogate-evaluated or full-wave simulated.
  2. [Abstract] The abstract claims “experimental validation” but provides no quantitative comparison (e.g., measured vs. target MSE); this detail should be added or the claim qualified.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed comments. We address each major comment below and will revise the manuscript to strengthen the validation of our results and the description of the surrogate model.

read point-by-point responses

  1. Referee: [Results / Experimental Validation] Results section (performance metrics): the reported spectral MSE of 0.0006 and band-alignment accuracy of 0.958 are obtained by comparing target spectra against spectra produced by the surrogate simulator on the generated geometries. Because the diffusion model is trained with spectrum-level losses derived from the same surrogate, these figures risk being partly tautological; a table comparing surrogate predictions, independent full-wave simulations, and measured data on the identical fabricated samples is required to demonstrate that the low error reflects Maxwell fidelity rather than surrogate fidelity.

    Authors: We agree that the current metrics rely on the surrogate and could appear circular. To demonstrate fidelity to Maxwell's equations and experimental reality, we will add a new table in the revised Results section. The table will compare target spectra against surrogate predictions, independent full-wave simulations performed on the generated geometries, and measured data from the fabricated samples already referenced in the manuscript. We will perform the additional full-wave simulations on the same fabricated designs to provide this direct multi-way comparison. revision: yes

  2. Referee: [Methods / Surrogate EM Simulator] Methods (surrogate training): the manuscript does not report the surrogate’s own validation accuracy, training-set size, or out-of-distribution error on geometries similar to those produced by the diffusion model. Without these quantities it is difficult to bound the systematic bias that could be introduced into the physics-aware loss.

    Authors: We concur that these details are necessary to evaluate potential bias. In the revised Methods section we will report the surrogate’s validation accuracy on its held-out test set, the size of the training dataset used to train the surrogate, and its error on out-of-distribution geometries that resemble those produced by the diffusion model. This information will allow readers to assess the reliability of the spectrum-level loss terms. revision: yes

Circularity Check

0 steps flagged

No circularity: metrics benchmarked against independent targets and experiments

full rationale

The paper reports performance via spectral MSE (0.0006) and band alignment (0.958) between target spectra and spectra from generated designs, plus experimental measurements. No equations, self-citations, or fitted parameters are shown that reduce these claims to quantities defined by the model's own inputs or prior self-work. The pre-trained surrogate supplies training loss but the final claims remain externally falsifiable against targets and fabricated samples, keeping the derivation self-contained.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The framework depends on standard assumptions of diffusion models plus the domain assumption that the surrogate simulator is accurate enough for regularization; no new entities are postulated.

free parameters (1)
  • diffusion model hyperparameters and loss weights
    Typical trainable or tuned parameters in ML frameworks; not enumerated in abstract but required for training the conditional model and physics loss.
axioms (1)
  • domain assumption The surrogate EM simulator accurately approximates full-wave responses for the purpose of spectrum-level regularization.
    Invoked to justify embedding the simulator for physics-aware loss functions.

pith-pipeline@v0.9.0 · 5814 in / 1328 out tokens · 43403 ms · 2026-05-20T05:22:29.603994+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

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

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

65 extracted references · 65 canonical work pages · 1 internal anchor

  1. [1]

    An overview of the theory and applications of metasurfaces: The two-dimensional equivalents of metamaterials,

    C. L. Holloway, E. F. Kuester, J. A. Gordon, J. O. Hara, J. Booth, and D. R. Smith, “An overview of the theory and applications of metasurfaces: The two-dimensional equivalents of metamaterials,” IEEE Antennas and Propagation Magazine , vol. 54, no. 2, pp. 10– 35, Apr. 2012

    work page 2012

  2. [2]

    Light propagation with phase discontinuities: Generalized laws of reflection and refraction,

    N. Yu, P. Genevet, M. A. Kats, F. Aieta, J. Tetienne, F. Capasso, and Z. Gaburro, “Light propagation with phase discontinuities: Generalized laws of reflection and refraction,” Science, vol. 334, no. 6054, pp. 333–337, Oct. 2011

    work page 2011

  3. [3]

    Flat optics with designer metasurfaces,

    N. Yu and F. Capasso, “Flat optics with designer metasurfaces,” Nature Materials, vol. 13, no. 2, pp. 139–150, Feb. 2014

    work page 2014

  4. [4]

    Metasurfaces for stealth applications: A comprehensive review,

    V. Joy, A. Dileep, P. V. Abhilash, R. U. Nair, H. Singh, “Metasurfaces for stealth applications: A comprehensive review,” Journal of Electronics Materials, vol. 50, no. 6, pp. 3129–3148, Jun. 2021

    work page 2021

  5. [5]

    Small-Size Broadband Coding Metasurface for RCS Reduction Based on Particle Swarm Optimization Algorithm,

    H. Hao, S. Du, and T. Zhang, “Small-Size Broadband Coding Metasurface for RCS Reduction Based on Particle Swarm Optimization Algorithm,” Progress in Electromagnetics Research M, vol. 81, pp. 97–105, 2019

    work page 2019

  6. [6]

    Customized inverse design of metamaterial absorber based on target-driven deep learning method,

    J. Hou, H. Lin, W. Xu, X. Shi, R. Tang, L. Chen, and Y. Tian, “Customized inverse design of metamaterial absorber based on target-driven deep learning method,” IEEE Access , vol. 8, pp. 211849–211859, 2020

    work page 2020

  7. [7]

    An ultrathin invisibility skin cloak for visible light,

    X. Ni, Z. J. Wong, M. Mrejen, Y. Wang, and X. Zhang, “An ultrathin invisibility skin cloak for visible light,” Science, vol. 349, no. 6254, pp. 1310–1314, Sep. 2015

    work page 2015

  8. [8]

    Polarization- insensitive and transparent frequency-selective surface for dual-band GSM shielding,

    Y. Yang, W. Li, K. N. Salama, and A. Shamim, “Polarization- insensitive and transparent frequency-selective surface for dual-band GSM shielding,” IEEE Transactions on Antennas and Propagation, vol. 69, no. 5, pp. 2779–2789, May 2021

    work page 2021

  9. [9]

    Metasurface holograms for visible light,

    X. Ni, A. Kildishev, and V. Shalaev, “Metasurface holograms for visible light,” Nature Comm. vol. 4, Art. no. 2807, Nov. 2013

    work page 2013

  10. [10]

    3D-integrated metasurfaces for full-colour holography,

    Y. Hu, X. Luo, Y. Chen, Q. Liu, X. Liu, Y. Wang, N. Liu and H. Duan, “3D-integrated metasurfaces for full-colour holography,” Light: Science and Applications, vol. 8, Art. no. 86, Sep. 2019

    work page 2019

  11. [11]

    Reflective metasurface for 5G and beyond wireless communications,

    M. Selvaraj, R. Vijay, and R. Anbazhagan, “Reflective metasurface for 5G and beyond wireless communications,” Scientific Reports , vol. 15, Art. no. 126, 2025

    work page 2025

  12. [12]

    Amplitude–phase independently encoding space-division multiplexed wireless communication using beamforming reconfigurable metasurfaces,

    Q. Xiong, Z. Zhang, C. Huang, M. Pu, J. Luo, Y. Guo, J. Ye, W. Pan, X. Ma, L. Yan, and X. Luo, “Amplitude–phase independently encoding space-division multiplexed wireless communication using beamforming reconfigurable metasurfaces,” Advanced Optical Materials, vol. 12, no. 28, 2024

    work page 2024

  13. [13]

    A broadband achromatic metalens for focusing and imaging in the visible,

    W. T. Chen, A. Y. Zhu, V. Sanjeev, M. Khorasaninejad, Z. Shi, E. Lee, and F. Capasso, “A broadband achromatic metalens for focusing and imaging in the visible,” Nature Nanotechnology, vol. 13, no. 3, pp. 220–226, Mar. 2018

    work page 2018

  14. [14]

    High-efficiency metasurfaces: Principles, realizations, and applications,

    Q. He, S. Sun, S. Xiao, and L. Zhou, “High-efficiency metasurfaces: Principles, realizations, and applications,” Advanced Optical Materials, vol. 6, no. 19, Art. no. 1800415, 2018

    work page 2018

  15. [15]

    Metasurface- empowered optical multiplexing and multifunction,

    S. Chen, W. Liu, Z. Li, H. Cheng, and J. Tian, “Metasurface- empowered optical multiplexing and multifunction,” Advanced Materials, vol. 32, no. 3, Art. no. 1805912, 2020

    work page 2020

  16. [16]

    Broadband multifunctional efficient meta-gratings based on dielectric waveguide phase shifters,

    M. Khorasaninejad and F. Capasso, “Broadband multifunctional efficient meta-gratings based on dielectric waveguide phase shifters,” Nano Letters, vol. 15, no. 10, pp. 6709–6715, Oct. 2015

    work page 2015

  17. [17]

    Structured transparent low- emissivity coatings with high microwave transmission,

    O. Bouvard, M. Lanini, and L. Burnier, “Structured transparent low- emissivity coatings with high microwave transmission,” Applied Physics A, vol. 123, no. 1, Art. no. 66, 2017

    work page 2017

  18. [18]

    Loss-assisted metasurface at an exceptional point,

    S. Dong, G. Hu, Q. Wang, Y. Jia , Q. Zhang, G. Cao, J. Wang, S. Chen, D. Fan, W. Jiang, Y. Li, A. Alu, and C.-W. Qiu, “Loss-assisted metasurface at an exceptional point,” ACS Photonics, vol. 7, no. 12, pp. 3321–3327, Dec. 2020

    work page 2020

  19. [19]

    Wide band-pass frequency selective surface with reduced periodicity for energy-efficient windows at higher frequencies,

    J. Fleury, M. Lanini, and C. Pose, “Wide band-pass frequency selective surface with reduced periodicity for energy-efficient windows at higher frequencies,” Applied Physics A, vol. 126, no. 6, Art. no. 417, 2020

    work page 2020

  20. [20]

    Review of numerical optimization techniques for meta-device design,

    S. D. Campbell, D. Sell, R.P. Jenkins, E. B. Whiting, J. A. Fan and D.H. Werner, “Review of numerical optimization techniques for meta-device design,” Optical Materials Express , vol. 9, no. 4, pp. 1842–1863, 2019

    work page 2019

  21. [21]

    Topology optimization for nanophotonics,

    J. S. Jensen and O. Sigmund, “Topology optimization for nanophotonics,” Laser & Photonics Reviews, vol. 5, no. 2, pp. 308– 321, Mar. 2011

    work page 2011

  22. [22]

    Topology optimization of photonic and phononic crystals and metamaterials: A review,

    W. Li, F. Meng, Y. Chen, Y. Li, and X. Huang, “Topology optimization of photonic and phononic crystals and metamaterials: A review,” Advanced Theory and Simulations , vol. 2, Art. no. 1900017, 2019

    work page 2019

  23. [23]

    Near-ideal optical metamaterial absorbers with super-octave bandwidth,

    J. A. Bossard, L. Lin, S. Yun, L. Liu, D. H. Werner, and T. S. Mayer, “Near-ideal optical metamaterial absorbers with super-octave bandwidth,” ACS Nano, vol. 8, no. 2, pp. 1517–1524, Feb. 2014

    work page 2014

  24. [24]

    Adaptive genetic algorithm for optical metasurfaces design,

    S. Jafar-Zanjani, S. Inampudi, and H. Mosallaei, “Adaptive genetic algorithm for optical metasurfaces design,” Scientific Reports, vol. 8, no. 1, Art. no. 11040, 2018

    work page 2018

  25. [25]

    Application of a microgenetic algorithm (MGA) to the design of broadband microwave absorbers using multiple frequency selective surface screens buried in dielectrics,

    S. Chakravarty, R. Mittra, and N. R. Williams, “Application of a microgenetic algorithm (MGA) to the design of broadband microwave absorbers using multiple frequency selective surface screens buried in dielectrics,” IEEE Transactions on Antennas and Propagation, vol. 50, no. 3, pp. 284–296, Mar. 2002

    work page 2002

  26. [26]

    Synthesis of low/zero index of refraction metamaterials from frequency selective surfaces using genetic algorithms,

    M. A. Gingrich and D. H. Werner, “Synthesis of low/zero index of refraction metamaterials from frequency selective surfaces using genetic algorithms,” Electronics Letters, vol. 41, no. 23, pp. 1266– 1267, Nov. 2005

    work page 2005

  27. [27]

    Optimising efficiency and gain of small meander line RFID antennas using ant colony system,

    A. Lewis, G. Weis, M. Randall, A. Galehdar and D. Thiel, “Optimising efficiency and gain of small meander line RFID antennas using ant colony system,” in Proceedings of the IEEE Congress on Evolutionary Computation , Norway, 2009, pp. 1486– 1492

    work page 2009

  28. [28]

    Design and optimization of 3-D frequency-selective surfaces based on a multiobjective lazy ant colony optimization algorithm,

    D. Z. Zhu, P. L. Werner, and D. H. Werner, “Design and optimization of 3-D frequency-selective surfaces based on a multiobjective lazy ant colony optimization algorithm,” IEEE Transactions on Antennas and Propagation , vol. 65, no. 12, pp. 7137–7149, Dec. 2017

    work page 2017

  29. [29]

    Fabrication and characterization of multiband polarization- independent 3-D printed frequency selective structures with ultra- wide fields of view,

    D. Z. Zhu, M. D. Gregory, P. L. Werner, and D. H. Werner, “Fabrication and characterization of multiband polarization- independent 3-D printed frequency selective structures with ultra- wide fields of view,” IEEE Transactions on Antennas and Propagation, vol. 66, no. 11, pp. 6096–6105, Nov. 2018

    work page 2018

  30. [30]

    Particle swarm optimization in electromagnetics,

    J. Robinson and Y. Rahmat-Samii, “Particle swarm optimization in electromagnetics,” IEEE Transactions on Antennas and Propagation, vol. 52, no. 2, pp. 397–407, Feb. 2004

    work page 2004

  31. [31]

    Stochastic optimization of low-loss optical negative-index metamaterial,

    A. V. Kildishev, U. K. Chettiar, Z. Liu, V. M. Shalaev, D.-H. Kwon, Z. Bayraktar, and D. H. Werner, “Stochastic optimization of low-loss optical negative-index metamaterial,” Journal of the Optical Society of America B, vol. 24, no. 10, pp. A34–A39, Oct. 2007

    work page 2007

  32. [32]

    Deep learning enabled inverse design in nanophotonics,

    S. So, T. Badloe, J. Noh, J. Bravo-Abad, and J. Rho, “Deep learning enabled inverse design in nanophotonics,” Nanophotonics, vol. 9, no. 5, pp. 1041–1057, 2020

    work page 2020

  33. [33]

    Deep learning for the design of photonic structures,

    W. Ma, Z. Liu, Z. A. Kudyshev, A. Boltasseva, W. Cai, and Y. Liu, “Deep learning for the design of photonic structures,” Nature Photonics, vol. 15, pp. 77–90, Jan. 2021

    work page 2021

  34. [34]

    Nanophotonic particle simulation and inverse design using artificial neural networks,

    J. Peurifoy, Y. Shen, L. Jing, Y. Yang, F. Cano-Renteria, B. Delacy, J. D. Joannopoulos, M. Tegmark, and M. Solja čić, “Nanophotonic particle simulation and inverse design using artificial neural networks,” Science Advances, vol. 4, Art. no. eaar4206, 2018

    work page 2018

  35. [35]

    Simultaneous inverse design of materials and structures via deep learning: Demonstration of dipole resonance engineering using core–shell nanoparticles,

    S. So, J. Mun, and J. Rho, “Simultaneous inverse design of materials and structures via deep learning: Demonstration of dipole resonance engineering using core–shell nanoparticles,” ACS Applied Materials & Interfaces, vol. 11, no. 27, pp. 24264–24268, 2019

    work page 2019

  36. [36]

    Plasmonic nanostructure design and characterization via deep learning,

    I. Malkiel, M. Mrejen, A. Nagler, U. Arieli, L. Wolf, and H. Suchowski, “Plasmonic nanostructure design and characterization via deep learning,” Light: Science and Applications , vol. 7, . 60, 2018

    work page 2018

  37. [37]

    Training deep neural networks for the inverse design of nanophotonic structures,

    D. Liu, Y. Tan, E. Khoram, and Z. Yu, “Training deep neural networks for the inverse design of nanophotonic structures,” ACS Photonics, vol. 5, no. 4, pp. 1365–1369, 2018

    work page 2018

  38. [38]

    Constrained tandem neural network assisted inverse design of metasurfaces for microwave absorption,

    X. He, X. Cui, and C. T. Chan, “Constrained tandem neural network assisted inverse design of metasurfaces for microwave absorption,” Optics Express, vol. 31, no. 25, pp. 40969–40979, 2023

    work page 2023

  39. [39]

    A deep learning-driven forward and inverse cooperative network for circular dichroism in chiral metasurfaces,

    Z. Hu, W. Su, K. Hu, and B. Tang, “ A deep learning-driven forward and inverse cooperative network for circular dichroism in chiral metasurfaces,” Journal of Materials Chemistry C, vol. 13, pp. 9717– 9723, 2025

    work page 2025

  40. [40]

    Generative model for the inverse design of metasurfaces,

    Z. Liu, D. Zhu, K. Lee, A. S. Kim, and J. R. W. Mark, “Generative model for the inverse design of metasurfaces,” ACS Nano, vol. 12, no. 7, pp. 7073–7083, 2018

    work page 2018

  41. [41]

    Conditional generative adversarial networks for inverse design of multifunctional microwave metasurfaces,

    M. Kiani and J. Kiani, “Conditional generative adversarial networks for inverse design of multifunctional microwave metasurfaces,” Advanced Photonics Research., vol. 3, Art. no. 202200110, 2022

    work page 2022

  42. [42]

    Synthesis of multi-band reflective polarizing metasurfaces using a generative adversarial network,

    P. Naseri, G. Goussetis, N. J. G. Fonseca, and S. V. Hum, “Synthesis of multi-band reflective polarizing metasurfaces using a generative adversarial network,” Scientific Reports., vol. 12, 17006, 2022

    work page 2022

  43. [43]

    Prediction of multi- layer metasurface design using conditional deep convolutional generative adversarial networks,

    A. Nezaratizadeh, S. M. Hashemi, and M. Bod., “Prediction of multi- layer metasurface design using conditional deep convolutional generative adversarial networks,” International Journal for Light and Electron Optics, vol. 313, p. 172005, 2024

    work page 2024

  44. [44]

    Anchor-controlled generative adversarial network for high-fidelity electromagnetic and structurally diverse metasurface design,

    Y. Zeng, H. Cao, and X. Jin, “Anchor-controlled generative adversarial network for high-fidelity electromagnetic and structurally diverse metasurface design,” arXiv preprint arXiv:2408.16231v2, 2024

    work page arXiv 2024

  45. [45]

    A cyclical deep learning based framework for simultaneous inverse and forward design of nanophotonic metasurfaces,

    A. Mall, A. Patil, A. Sethi & A. Kumar, “A cyclical deep learning based framework for simultaneous inverse and forward design of nanophotonic metasurfaces,” Scientific Reports , Article number: 19427, 2020

    work page 2020

  46. [46]

    Intelligent coding metasurface holograms by physics-assisted unsupervised generative adversarial network,

    C. Liu, W. M. Yu, Q. Ma, L. Li, and T. J. Cui, “Intelligent coding metasurface holograms by physics-assisted unsupervised generative adversarial network,” Photonics Research, vol. 9, no. 4, pp. B159– B167, Apr. 2021

    work page 2021

  47. [47]

    Single-layer metasurface design with symmetrical free-form patterns using generative adversarial networks,

    G. Dai, H. Li, X. Zhang, and Y. Liu, “Single-layer metasurface design with symmetrical free-form patterns using generative adversarial networks,” Applied Soft Computing , vol. 127, Art. no. 109353, 2022

    work page 2022

  48. [48]

    Inverse design of metasurfaces with customized transmission characteristics of frequency band based on generative adversarial networks,

    H. P. Wang, D. M. Cao, X. Y. Pang, X. H. Zhang, S. Y. Wang, W. Y. Hou, C. C. Nie, and Y. B. Li, “Inverse design of metasurfaces with customized transmission characteristics of frequency band based on generative adversarial networks,” Optics Express, vol. 31, no. 23, pp. 37763–37783, 2023

    work page 2023

  49. [49]

    Fast design of plasmonic metasurfaces enabled by deep learning,

    A. Mall, A. Patil, D. Tamboli, A. Sethi and A. Kumar, “Fast design of plasmonic metasurfaces enabled by deep learning,” Journal of Physics D: Applied Physics, vol. 53, no. 49, October 2020

    work page 2020

  50. [50]

    A Machine learning-based approach to synthesize multilayer metasurfaces,

    P. Naseri and S. V. Hum, “A Machine learning-based approach to synthesize multilayer metasurfaces,” 2020 IEEE International Symposium on Antennas and Propagation and North American Radio Science Meeting, Montreal, QC, Canada, 2020, pp. 933-934

    work page 2020

  51. [51]

    A generative machine learning-based approach for inverse design of multilayer metasurfaces,

    P. Naseri and S. V. Hum, “A generative machine learning-based approach for inverse design of multilayer metasurfaces,” IEEE Transactions on Antennas and Propagation, vol. 69, no. 9, pp. 5725– 5739, Sep. 2021

    work page 2021

  52. [52]

    An Optimization Approach Based on Separated Latent Space for Inverse Design of Metasurfaces,

    J. -H. Kim and I. -P. Hong, "An Optimization Approach Based on Separated Latent Space for Inverse Design of Metasurfaces," in IEEE Antennas and Wireless Propagation Letters, vol. 23, no. 7, pp. 2135-2139, July 2024

    work page 2024

  53. [53]

    A generative meta-atom model for metasurface-based absorber designs,

    W. Ding, J. Chen, and R.-X. Wu, “A generative meta-atom model for metasurface-based absorber designs,” Advanced Optical Materials, vol. 11, no. 2, Art. no. 2201959, 2022

    work page 2022

  54. [54]

    Design of Metamaterials for Absorbers Based on Variational Autoencoder,

    Q. Li, J. Wang, T. Lei, T. Xiang, C. Qin and M. Yang, “Design of Metamaterials for Absorbers Based on Variational Autoencoder,” IEEE Access, vol. 12, pp. 92328-92336, 2024

    work page 2024

  55. [55]

    Novel method of performance-optimized metastructure design for electromagnetic wave absorption in specific band using deep learning,

    H.-I. On, L. Jeong, T.-M. Seo, Y. Jo, W. Choi, D.-J. Kang, J.-H. Park, and H.-J. Lee, “Novel method of performance-optimized metastructure design for electromagnetic wave absorption in specific band using deep learning,” Engineering Applications of Artificial Intelligence, vol. 137, p. 109274, 2024

    work page 2024

  56. [56]

    Modified variational autoencoder for inversely predicting plasmonic nanofeatures for generating structural color,

    P. Pillai, B. RaI, and P. Pal, “Modified variational autoencoder for inversely predicting plasmonic nanofeatures for generating structural color,” Scientific Reports, vol. 13, Art. no. 3536, 2023

    work page 2023

  57. [57]

    Hybrid quantum generative adversarial networks to inverse design metasurfaces for incident angle-independent unidirectional transmission,

    S. Warrier, J. Dontabhaktuni, “Hybrid quantum generative adversarial networks to inverse design metasurfaces for incident angle-independent unidirectional transmission,” arXiv preprint arXiv:2507.03518, 2025

    work page arXiv 2025

  58. [58]

    Continuous conditional generative adversarial networks: Novel empirical losses and label input mechanisms,

    X. Ding, Y. Wang, Z. Xu, W. J. Welch, and Z. J. Wang, “Continuous conditional generative adversarial networks: Novel empirical losses and label input mechanisms,” IEEE Trans.Pattern Analysis and Machine Intelligence, vol. 45, no. 7, pp. 8143-8158, Jul. 2023

    work page 2023

  59. [59]

    PaDGAN: A generative adversarial network for performance augmented diverse designs,

    W. Chen and F. Ahmed, “PaDGAN: A generative adversarial network for performance augmented diverse designs,” arXiv preprint arXiv:2002.11304, 2020

    work page arXiv 2002

  60. [60]

    Image super-resolution via iterative refinement,

    C. Saharia, J. Ho, W. Chan, T. Salimans, D. J. Fleet and M. Norouzi, “Image super-resolution via iterative refinement,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 4, pp. 4713-4726, April 2023

    work page 2023

  61. [61]

    Optimizing metamaterial inverse design with 3D conditional diffusion model and data augmentation,

    X. Zheng, J. Shiomi and T. Yamada, “Optimizing metamaterial inverse design with 3D conditional diffusion model and data augmentation,” Advanced Materials Tech., vol. 10, 10p., 2025

    work page 2025

  62. [62]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” in Proceedings of the 34th International Conference on Neural Information Processing Systems (NeurIPS), Vancouver, BC, Canada, 2020, pp. 6840–6851

    work page 2020

  63. [63]

    Classifier-Free Diffusion Guidance

    J. Ho and T. Salimans, “Classifier-free diffusion guidance,” arXiv preprint arXiv:2207.12598, 2022

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  64. [64]

    FiLM: Visual reasoning with a general conditioning layer,

    E. Perez, F. Strub, H. de Vries, V. Dumoulin, and A. Courville, “FiLM: Visual reasoning with a general conditioning layer,” in Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence, Article No.: 483, pp. 3942 - 3951, 2018

    work page 2018

  65. [65]

    CST Studio Suite®, CST AG, Germany, www.cst.com

    work page