Pith. sign in

REVIEW 2 major objections 6 minor 29 references

A flexible framework for large-scale FDTD simulations: open-source inverse design for 3D nanostructures

T0 review · 2 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read An open-source FDTD package makes gradient-based inverse design of 3D nanostructures practical by reconstructing earlier fields through time-reversed Maxwell updates, needing only six 2D boundary slices per step for memory.

desk verdict A genuinely useful open-source GPU FDTD inverse-design tool whose time-reversal gradient is unverified and whose speed claims compare unlike hardware; fixable, worth reviewing. read the letter →

arxiv 2412.12360 v2 pith:6C3XRAJB submitted 2024-12-16 physics.optics physics.comp-ph

Yannik Mahlau , Frederik Schubert , Konrad Bethmann , Reinhard Caspary , Antonio Calà Lesina , Marco Munderloh , Jörn Ostermann , Bodo Rosenhahn This is my paper · ORCID
classification physics.opticsphysics.comp-ph
keywords FDTDinversedesignautomaticdifferentiationtimereversalnanophotonicsJAXGPUphotonicintegratedcircuits
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 introduces FDTDX, an open-source Python package that performs inverse design of three-dimensional photonic nanostructures by differentiating through Finite-Difference Time-Domain (FDTD) simulations. Its central claim is that Maxwell's equations' time reversibility makes gradient computation memory-efficient: rather than storing the full 3D electromagnetic field at every time step, the backward pass reconstructs earlier fields with an inverse update, needing only six two-dimensional slices at the PML boundaries. This reduces memory enough that devices with millions of design parameters can be optimized on a single consumer GPU, and the package scales from one to multiple GPUs for simulations with billions of grid cells. The authors demonstrate the approach by optimizing a silicon waveguide bend to 92% efficiency (1.6 µm² footprint) and a two-photon-polymerization stitching element robust to random offsets up to 2 µm. If correct, this makes large-scale photonic inverse design accessible to researchers without commercial software licenses.

What carries the argument

The central object is the inverse FDTD update step: because the discretized Maxwell curl updates are invertible in the interior of the simulation domain, the automatic differentiation backward pass can reconstruct earlier field states instead of storing them. Only six two-dimensional slices of field data at the convolutional PML boundaries are checkpointed per time step, cutting memory from a full 3D volume per step to a small boundary set; this is implemented via reverse-mode automatic differentiation in JAX, with the Yee grid providing the staggered discretization that makes the curl operator local and reversible.

What would settle it

Run a small 3D FDTD device (e.g., the 3-million-cell bend) and compare the reverse-mode gradient from FDTDX against a central finite-difference gradient or an independent adjoint implementation; if the normalized gradient error exceeds a few percent or the reverse pass is unstable, the time-reversal memory saving is not accurate enough to rely on.

Watch

Extended reading notes

Core claim

FDTDX claims that gradient-based inverse design for 3D FDTD can be made practical by replacing checkpointed field storage with an inverse time-stepping pass. The electric and magnetic field updates on the Yee grid are reversible for linear materials, so the backward pass recomputes the fields from time step t+1 to t, and only the six 2D slices at the PML boundary need to be saved because the PML itself is not invertible. The authors show this memory-efficient differentiation works within the JAX framework, supports multi-GPU execution, and produces optimized devices: a silicon corner bend with -0.36 dB attenuation (92% efficiency) over 3 million grid cells, and a 3D polymer stitching device that keeps attenuation between -1 dB and -5 dB under random translations up to 2 µm, versus -35 dB for a standard waveguide. They also report speed comparisons where FDTDX outperforms other open-source FDTD solvers and approaches the speed of a commercial cloud solver at moderate resolutions.

Load-bearing premise

The load-bearing premise is that saving only six two-dimensional field slices at the PML boundaries each time step is enough for the inverse time-stepping pass to reconstruct all earlier fields exactly, so the computed gradients match the true objective; the paper does not include a numerical gradient-accuracy check.

Editorial extensions

If this is right

  • A single consumer GPU with roughly 20 GB of memory can run forward and backward passes for a 3-million-cell inverse design, making large-scale optimization accessible without specialized hardware.
  • The same solver scales to 2.3 billion grid cells by distributing across multiple GPUs, which the authors demonstrate up to four H100 cards.
  • The time-reversal gradient approach incurs only a small runtime overhead compared to checkpointing, because fields are recomputed in the backward pass rather than over many forward sweeps.
  • The relational object-constraint API lets users specify simulation scenes in relative coordinates, which the authors claim is new among FDTD tools and reduces setup errors.
  • Devices optimized with this framework achieve concrete performance targets: a 92% efficient silicon bend and a 2PP stitching element that outperforms a plain waveguide for misalignments above 0.5 µm in the y-direction.

Reading between the lines

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

  • A direct numerical gradient check (e.g., against finite differences or an adjoint solver) would confirm whether the inverse time-reversal pass is exact; the paper reports no such check, so the 92% bend and stitching results rest on the assumption that the saved PML slices fully capture the reverse dynamics.
  • The same time-reversal trick should extend to other reversible update schemes beyond FDTD, such as pseudospectral time-domain methods, whenever the interior update is invertible and the boundary condition is the only lossy element.
  • The stitching-device demonstration suggests a design principle for 2PP fabrication: deliberately insert a collimation and gathering stage so that translational stitching errors are converted into small coupling losses rather than mode mismatch; optimizing for robustness in z-translation is the natural next step the authors name.
  • Because the framework is open source, the community can implement custom CUDA kernels or alternative boundary conditions without needing a proprietary solver, which may shift the standard practice for large-scale photonic inverse design toward in-house GPU clusters.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 6 minor

Summary. The paper presents FDTDX, an open-source JAX-based FDTD simulation and inverse-design framework. The central methodological claim is a memory-efficient reverse-mode automatic differentiation that exploits the time-reversibility of Maxwell's equations, storing only six two-dimensional slices at the PML boundaries instead of full three-dimensional field histories. The paper also introduces a relational object API for scene specification and demonstrates multi-GPU scaling. Results include a speed comparison against Meep, Ceviche, OmniSim, and Tidy3D; an optimized silicon waveguide bend with reported 92% efficiency; and a polymer waveguide stitching device robust to translations up to 2 µm. The forward solver is validated against Meep in Appendix A.

Significance. If the gradient computation is correct, FDTDX is a potentially valuable open-source contribution: it provides native GPU execution, automatic differentiation, multi-GPU scaling, and a user-friendly constraint API, all of which address real bottlenecks in photonic inverse design. The forward-solver comparison against Meep in Appendix A is a positive external check, and the two-stage constrained optimization and stitching-robustness experiments are nontrivial demonstrations. However, the central numerical claims depend on the correctness of the reverse time-stepping gradients, which the paper does not validate, and the speed comparison is not a controlled software comparison. These issues must be addressed before the paper's main claims can be accepted.

major comments (2)
  1. [Section 2.2] The reverse time-stepping gradient is never numerically validated. Appendix A validates only the forward solver against Meep, whereas the optimized efficiencies reported in Sections 3.2 and 3.3 are produced by gradients computed through the inverse update. No finite-difference comparison, no independent adjoint comparison, and no reconstruction-error test for the inverse update are reported. This is load-bearing because Eqs. (1)-(2) describe a lossless, sourceless continuum system, while the actual discrete update includes TFSF source injection, CPML, and material averaging/interpolation, any of which can break exact time reversibility. Please add a small-scale gradient accuracy test (e.g., against finite differences or an independent adjoint) and a test that the six saved 2D PML-boundary slices are sufficient to reconstruct the interior fields in the backward pass, with a quantitative error measure.
  2. [Section 3.1, Table 2] The speed comparison is not a controlled software comparison. FDTDX is run on one or four NVIDIA H100 GPUs, Meep and Ceviche on a 16-core AMD EPYC CPU cluster, OmniSim on a consumer laptop CPU, and Tidy3D on undisclosed remote hardware. The text's claim that "FDTDX outperforms all other open-source simulation software by a large margin" and the reported roughly 10x speedup over Meep therefore conflate hardware differences with software performance. Please either benchmark the open-source solvers on comparable hardware (including a CPU-only FDTDX configuration, if necessary) or explicitly restrict the claim to "on the hardware used in this study" and remove the general outperformance claim.
minor comments (6)
  1. [Abstract] The abstract contains a duplicated word: "random translational offsets offsets of up to 2 µm."
  2. [Section 3.2, Figure 4] The y-axis label in Fig. 4(a) reads "Attenuation [dB]" with positive plotted values, while the text reports "an attenuation of -0.36dB, or equivalently an efficiency of 92%." Please clarify the sign convention (insertion loss versus transmission in dB) so the reader can interpret the optimization curves correctly.
  3. [Table 2] The table uses "×" entries without a caption explanation. Please state explicitly that "×" means the simulation was not run or could not be run on the available hardware/resource limits.
  4. [Section 2.3] There is a typo: "the algorithms returns an error message" should be "the algorithm returns an error message."
  5. [Appendix A] The Meep validation is presented only as field-distribution images; no quantitative error metric (e.g., maximum or mean normalized error across the planes) is reported. Adding a single error number would make the validation more convincing.
  6. [References] Reference [13] is cited as "Quantized inverse design for photonic integrated circuits" (2024) without a venue, arXiv identifier, or DOI. Please provide a complete citation so readers can locate the prior implementation of the inverse time-stepping method.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: forward solver is benchmarked against Meep, and the self-citation for the AD implementation is not load-bearing.

full rationale

The paper's central derivation chain is the FDTD forward simulation and the reverse-mode automatic differentiation via time reversal. The forward solver is independently validated against Meep in Appendix A on random scattering objects, which gives an external check on the simulator that produces the reported device efficiencies. The AD method is explicitly credited to the external work of Tang et al. [12], and the JAX implementation is attributed to the authors' own Ref. [13]. This self-citation is a pointer to prior implementation work rather than the sole justification for the paper's claims; the optimization curves in Fig. 4 provide internal evidence that the gradients are useful, and the open-source code is publicly available for reproduction. The reported 92% bend efficiency is the value of the optimized objective—this is what inverse design means—and it is not a definitional identity because the optimization could have failed; the simulator's accuracy is independently supported. The lack of a numerical gradient check (e.g., against finite differences) is a correctness and validation risk, not a circularity, because no equation or fitted parameter is being renamed as a prediction. There is no self-definitional reduction, no fitted input called prediction, no uniqueness theorem imported from the authors, and no ansatz smuggled in via citation. The only mild concern is the self-citation for implementation details, which is not load-bearing and does not reduce the results to their inputs.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on standard FDTD theory, the time-reversibility assumption for linear media, and the sufficiency of saving PML boundary slices for backward reconstruction. No fitted constants are involved; the listed free parameters are scenario settings chosen by the authors. The framework introduces no new physical entities. The main load-bearing assumptions are the gradient accuracy of the reverse-time update and the adequacy of interpolation-free gradients during optimization, neither of which is independently verified in the paper.

free parameters (3)
  • Courant factor = 0.99
    Standard CFL stability factor for FDTD; chosen for all experiments, a numerical setting rather than a fitted parameter.
  • Stitching offset bound = 2 micrometers
    Assumed maximum translational fabrication error in x and y for the stitching device; set by the authors as the target robustness requirement.
  • Minimum feature size brush = 100 nanometers
    Foundry minimum feature size constraint used in the silicon waveguide bend optimization; chosen to match realistic fabrication limits.
assumptions (4)
  • domain assumption Maxwell's equations are time-reversible for linear, non-dispersive materials in the interior of the simulation domain
    Used in Section 2.2 to justify the inverse FDTD update for gradient computation; excludes PML regions and nonlinear or dispersive media.
  • domain assumption Saving six 2D slices of the PML-interior boundary at each time step is sufficient to reconstruct earlier fields during the backward pass
    Stated in Section 2.2 as the memory-saving mechanism; the correctness rests on the invertibility of the interior update and the containment of irreversibility in the PML.
  • ad hoc to paper Gradients computed with interpolation disabled are adequate to guide optimization, with exact interpolation used only for final evaluation
    Section 2.1 states interpolation is disabled by default during optimization; the optimized designs are then evaluated with interpolation enabled. The paper does not test whether this mismatch biases the gradient path.
  • standard math Standard FDTD discretization on a Yee grid with CPML boundary conditions is a valid approximation of Maxwell's equations
    The framework is built on the Yee grid and CPML methods cited in Refs. 6 and 7; this is standard numerical electromagnetic theory.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A flexible framework for large-scale FDTD simulations: open-source inverse design for 3D nanostructures." pith.science (2026). https://pith.science/paper/6C3XRAJB

@misc{pith2026241212360,
  author       = {Pith},
  title        = {Pith review of: A flexible framework for large-scale FDTD simulations: open-source inverse design for 3D nanostructures},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6C3XRAJB}},
  note         = {Machine review of arXiv:2412.12360}
}
read the original abstract

We introduce an efficient open-source python package for the inverse design of three-dimensional photonic nanostructures using the Finite-Difference Time-Domain (FDTD) method. Leveraging a flexible reverse-mode automatic differentiation implementation, our software enables gradient-based optimization over large simulation volumes. Gradient computation is implemented within the JAX framework and based on the property of time reversibility in Maxwell's equations. This approach significantly reduces computational time and memory requirements compared to traditional FDTD methods. Gradient-based optimization facilitates the automatic creation of intricate three-dimensional structures with millions of design parameters, which would be infeasible to design manually. We demonstrate the scalability of the solver from single to multiple GPUs through several inverse design examples, highlighting its robustness and performance in large-scale photonic simulations. In addition, the package features an object-oriented and user-friendly API that simplifies the specification of materials, sources, and constraints. Specifically, it allows for intuitive positioning and sizing of objects in absolute or relative coordinates within the simulation scene. By rapid specification of the desired design properties and rapid optimization within the given user constraints, this open-source framework aims to accelerate innovation in photonic inverse design. It yields a powerful and accessible computational tool for researchers, applicable in a wide range of use cases, including but not limited to photonic waveguides, active devices, and photonic integrated circuits.

Figures

Figures reproduced from arXiv: 2412.12360 by the authors.

Figure 1
Figure 1. Relational size constraints across multiple axes and with different proportions. Axis 0 is the x-axis [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Relational positional constraints across multiple axes and with offsets. Axis 0 is the x-axis (horizontal) [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Simulation setup for a waveguide bend. The substrate consistis of silica with a refractive index of 1 [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Optimization of a waveguide bend. In (a), optimization results are presented using either strict [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Simulation scene for a stitching element robust to random translations. The mode source inserts light [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Analysis of waveguide stitching under random translation in the x- and y-axis. In (a), the coupling [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: (a) Energy distribution of the optimized design and a large offset of 2 [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: Distribution of the electric field component in x-direction measured in the XY, XZ and YZ plane at [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 28 canonical work pages

  1. [1]

    Meep: A flexible free-software package for electromagnetic simulations by the fdtd method,

    Oskooi, A. F., Roundy, D., Ibanescu, M., Bermel, P., Joannopoulos, J., and Johnson, S. G., “Meep: A flexible free-software package for electromagnetic simulations by the fdtd method,” Computer Physics Com- munications 181(3), 687–702 (2010)

    work page 2010

  2. [2]

    Tidy3d: hardware-accelerated electromagnetic solver for fast simulations at scale

    Flexcompute, “Tidy3d: hardware-accelerated electromagnetic solver for fast simulations at scale.” https://www.flexcompute.com/download-whitepaper/ (2022)

    work page 2022

  3. [3]

    Forward-mode differentiation of maxwell’s equations,

    Hughes, T. W., Williamson, I. A., Minkov, M., and Fan, S., “Forward-mode differentiation of maxwell’s equations,” ACS Photonics 6(11), 3010–3016 (2019)

    work page 2019

  4. [4]

    FDTD solution, a commercial professional software

    Lumerical Solution Inc., “FDTD solution, a commercial professional software.” https://www.lumerical.com (2024)

    work page 2024

  5. [5]

    Omnisim: Omni-directional photonic simulations

    Photon Design, “Omnisim: Omni-directional photonic simulations.” https://www.photond.com (2024)

    work page 2024

  6. [6]

    Numerical solution of initial boundary value problems involving maxwell’s equations in isotropic media,

    Kane Yee, “Numerical solution of initial boundary value problems involving maxwell’s equations in isotropic media,” IEEE Transactions on Antennas and Propagation 14, 302–307 (May 1966)

    work page 1966

  7. [7]

    Convolution pml (cpml): An efficient fdtd implementation of the cfs–pml for arbitrary media,

    Roden, J. A. and Gedney, S. D., “Convolution pml (cpml): An efficient fdtd implementation of the cfs–pml for arbitrary media,” Microwave and Optical Technology Letters 27(5), 334–339 (2000)

    work page 2000

  8. [8]

    and Hagness, S

    Taflove, A. and Hagness, S. C., [ Computational electrodynamics: the finite-difference time-domain method ], Artech House, Norwood, 3rd ed. (2005)

    work page 2005

Show all 29 references
  1. [9]

    A mathematical view of automatic differentiation,

    Griewank, A. and Walther, A., “A mathematical view of automatic differentiation,” Acta Numerica 12, 321–398 (2003)

    2003

  2. [10]

    and Walther, A., [ Evaluating Derivatives ], Society for Industrial and Applied Mathematics, second ed

    Griewank, A. and Walther, A., [ Evaluating Derivatives ], Society for Industrial and Applied Mathematics, second ed. (2008)

    2008

  3. [11]

    Achieving logarithmic growth of temporal and spatial complexity in reverse automatic differentiation,

    Griewank, A., “Achieving logarithmic growth of temporal and spatial complexity in reverse automatic differentiation,” Optimization Methods and Software 1(1), 35–54 (1992)

    1992

  4. [12]

    Time reversal differentiation of fdtd for photonic inverse design,

    Tang, R. J., Lim, S. W. D., Ossiander, M., Yin, X., and Capasso, F., “Time reversal differentiation of fdtd for photonic inverse design,” ACS Photonics 10(12), 4140–4150 (2023)

    2023

  5. [13]

    Quantized inverse design for photonic integrated circuits,

    Schubert, F., Mahlau, Y., Bethmann, K., Hartmann, F., Caspary, R., Munderloh, M., Ostermann, J., and Rosenhahn, B., “Quantized inverse design for photonic integrated circuits,” (2024)

    2024

  6. [14]

    Integrated metamaterials for efficient and compact free- space-to-waveguide coupling,

    Shen, B., Wang, P., Polson, R., and Menon, R., “Integrated metamaterials for efficient and compact free- space-to-waveguide coupling,” Opt. Express 22, 27175–27182 (Nov 2014)

    2014

  7. [15]

    ¨Uber die partiellen differenzengleichungen der mathematischen physik,

    Courant, R., Friedrichs, K., and Lewy, H., “ ¨Uber die partiellen differenzengleichungen der mathematischen physik,” Mathematische Annalen 100, 32–74 (1928)

    1928

  8. [16]

    Inverse design of photonic devices with strict foundry fabrication constraints,

    Schubert, M. F., Cheung, A. K. C., Williamson, I. A. D., Spyra, A., and Alexander, D. H., “Inverse design of photonic devices with strict foundry fabrication constraints,” ACS Photonics 9(7), 2327–2336 (2022)

    2022

  9. [17]

    Micro 3d printing by two-photon polymerization: Configurations and parameters for the nanoscribe system,

    Bunea, A.-I., del Castillo Iniesta, N., Droumpali, A., Wetzel, A. E., Engay, E., and Taboryski, R., “Micro 3d printing by two-photon polymerization: Configurations and parameters for the nanoscribe system,” Micro 1(2), 164–180 (2021)

    2021

  10. [18]

    openems - open electromagnetic field solver

    Liebig, T., “openems - open electromagnetic field solver.” https://www.openEMS.de (2024)

    2024

  11. [19]

    Automatic differentiation accelerated shape optimization approaches to photonic inverse design on rectilinear simulation grids,

    Hooten, S., Sun, P., Gantz, L., Fiorentino, M., Beausoleil, R. G., and Vaerenbergh, T. V., “Automatic differentiation accelerated shape optimization approaches to photonic inverse design on rectilinear simulation grids,” (2023)

    2023

  12. [20]

    Array programming with NumPy,

    Harris, C. R., Millman, K. J., van der Walt, S. J., Gommers, R., Virtanen, P., Cournapeau, D., Wieser, E., Taylor, J., Berg, S., Smith, N. J., Kern, R., Picus, M., Hoyer, S., van Kerkwijk, M. H., Brett, M., Haldane, A., del R ´ ıo, J. F., Wiebe, M., Peterson, P., G´ erard-Marc...

    2020

  13. [21]

    Performance of parallel fdtd method for shared- and distributed-memory architectures: Application tobioelectromagnet- ics,

    N., M. R.-C., Abal ¸enkovs, M., Angulo, L. M. D., Sanchez, C. C., Moglie, F., and Garcia, S. G., “Performance of parallel fdtd method for shared- and distributed-memory architectures: Application tobioelectromagnet- ics,” PLOS ONE 15, 1–16 (September 2020)

    2020

  14. [22]

    Nanophotonic inverse design with spins: Software architecture and practical considerations,

    Su, L., Vercruysse, D., Skarda, J., Sapra, N. V., Petykiewicz, J. A., and Vuˇ ckovi´ c, J., “Nanophotonic inverse design with spins: Software architecture and practical considerations,” (2019)

    2019

  15. [23]

    Python 3d fdtd simulator

    Laporte, F., “Python 3d fdtd simulator.” https://github.com/flaport/fdtd (2024)

    2024

  16. [24]

    Pytorch: An imperative style, high-performance deep learning library,

    Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., Desmaison, A., K¨ opf, A., Yang, E., DeVito, Z., Raison, M., Tejani, A., Chilamkurthy, S., Steiner, B., Fang, L., Bai, J., and Chintala, S., “Pytorch: An im...

    2019 arXiv

  17. [25]

    Luminescent ai

    Shen, P., “Luminescent ai.” https://github.com/paulxshen/Luminescent.jl (2024)

    2024

  18. [26]

    Khronos

    Hammond, A., “Khronos.” https://github.com/facebookresearch/Khronos.jl (2024)

    2024

  19. [27]

    Near-field optical microscopy simulations using graphics processing units,

    Klapetek, P. and Valtr, M., “Near-field optical microscopy simulations using graphics processing units,” Surface and Interface Analysis 42, 1109 – 1113 (06 2010)

    2010

  20. [28]

    fdtd-z : A systolic scheme for gpu-accelerated nanophotonic simulation

    Lu, J. and Vuˇ ckovi´ c, J., “fdtd-z : A systolic scheme for gpu-accelerated nanophotonic simulation.” https://github.com/spinsphotonics/fdtdz (2024)

    2024

  21. [29]

    Neural operator-based surrogate solver for free-form elec- tromagnetic inverse design,

    Augenstein, Y., Rep¨ an, T., and Rockstuhl, C., “Neural operator-based surrogate solver for free-form elec- tromagnetic inverse design,” ACS Photonics 10(5), 1547–1557 (2023)

    2023

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.