pith. sign in

arxiv: 2606.13320 · v1 · pith:C5OXYC2Dnew · submitted 2026-06-11 · ⚛️ physics.optics

Learning light scattering from operator parameter spaces to Galerkin-consistent solution spaces

Lida Liu , Jingwei Wang , Wei Cao , Yang Zhang , Yuntian Chen This is my paper

Pith reviewed 2026-06-27 06:02 UTC · model grok-4.3

classification ⚛️ physics.optics
keywords finite element methodoperator learningnanophotonicslight scatteringGalerkin methodvariational formulationMaxwell equationsneural networks
0
0 comments X

The pith

FEMONet maps physical parameters of wave problems to finite-element coefficients that obey the variational weak form of the vector wave equations.

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

The paper presents FEMONet as a way to learn solutions to parameterized optical scattering problems without the usual choice between slow numerical solvers and less accurate neural approximations. It encodes the physical entities of a wave-equation problem in an operator parameter space and connects that space to finite-element solution coefficients through the variational weak form. The network therefore outputs expansion coefficients rather than raw field values, and the assembled stiffness matrices absorb all spatial derivatives so that the training loss never requires differentiating the network output with respect to coordinates. A reader would care because the resulting model promises both the generality of operator learning and the stability and accuracy guarantees that come from staying inside a Galerkin framework.

Core claim

FEMONet is the first Galerkin-consistent operator-learning framework for complex-valued optical scattering; it learns from an operator parameter space directly to a solution space of finite-element expansion coefficients by enforcing the variational weak form of the governing vector wave equations, absorbing spatial derivatives into pre-assembled stiffness matrices and load vectors, and thereby preserving compatible trial and test spaces during training.

What carries the argument

The Galerkin-consistent formulation that predicts finite-element expansion coefficients (rather than unconstrained field values) while absorbing spatial derivatives into assembled stiffness matrices and load vectors.

If this is right

  • Classical finite-element solvers can be extended from single instances to families of parameterized scattering problems.
  • Training becomes more efficient because the physics loss no longer requires coordinate derivatives of the network output.
  • The same framework achieves high accuracy on dielectric, metallic, arrayed, plasmonic, and fully three-dimensional nanophotonic structures.
  • Generalization holds across the range of structures without retraining for each new geometry or material.
  • The approach supplies a stable, physics-respecting forward model suitable for downstream inverse design tasks.

Where Pith is reading between the lines

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

  • The same finite-element-constrained operator idea could be applied to other linear wave equations such as acoustic or elastic scattering without changing the core architecture.
  • Because the stiffness matrices are assembled once, the method may combine naturally with existing finite-element libraries to produce hybrid simulation pipelines.
  • If the operator parameter space is expanded to include fabrication tolerances, the model could directly output statistics over ensembles of manufactured devices.
  • The separation between parameter space and solution space suggests a route to transfer learning: pre-train on simple dielectrics and fine-tune on plasmonic cases with far fewer samples.

Load-bearing premise

Predicting finite-element expansion coefficients instead of raw field values will automatically keep the learned solutions inside compatible trial and test spaces and produce stable training across all structure types.

What would settle it

A test case on a plasmonic or three-dimensional metallic scatterer in which the FEMONet coefficients produce a residual in the weak-form loss that grows with network depth or exceeds the residual obtained from a standard finite-element solver on the same mesh.

Figures

Figures reproduced from arXiv: 2606.13320 by Jingwei Wang, Lida Liu, Wei Cao, Yang Zhang, Yuntian Chen.

Figure 1
Figure 1. Figure 1: Operator-parameter-space-augmented MIONet architecture for optical scattering operator learning. (a) The [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Comparative ablation and generalization study on basic lossless scatterers. (a) Geometric structures of the [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: (a) Geometric structures of the scatterers. (b) Loss convergence curves. (c) MSE histograms with correspond [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: (a) Geometric structures of the scatterers. (b) Loss convergence curves. (c) MSE histograms with corre [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Heatmap analysis of log10(MSE) for different network hyperparameters. single isolated optimum. This indicates that FEMONet is not highly sensitive to a narrowly selected hyperparameter setting. The combination of stable convergence, accurate field reconstruction, and a broad low-MSE region demonstrates the numerical robustness of the FEM-constrained operator-learning framework. 2.5 Sparse-sample learning o… view at source ↗
Figure 6
Figure 6. Figure 6: (a) Geometric structures of the scatterers. (b) Loss convergence curves. (c) MSE histograms with correspond [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: (a) Geometric structures of the scatterers. (b) Loss convergence curves. (c) MSE histograms with corre [PITH_FULL_IMAGE:figures/full_fig_p010_7.png] view at source ↗
read the original abstract

Efficient and generalizable full-wave simulation is essential for nanophotonic analysis and inverse design, yet existing methods face a tradeoff between the high computational cost of numerical solvers and the limited generalizability of neural operator models for complex optical scattering. Here, we introduce FEMONet, a finite-element-constrained operator-learning framework that learns light scattering from an operator parameter space to a Galerkin-consistent solution space. The operator parameter space encodes the physical entities defining a wave-equation problem, while the variational weak form links this space to the coordinate and physical solution spaces. Integrated with operator-learning networks, FEMONet extends classical solvers from isolated problem instances to parameterized scattering operators. To our knowledge, FEMONet represents the first Galerkin-consistent operator-learning framework for complex-valued optical scattering, grounded in the variational weak form of the governing vector wave equations. Finite-element discretization absorbs spatial derivatives into assembled stiffness matrices and load vectors, removing coordinate-based derivatives of the neural-network output from the physics loss and improving training efficiency. By predicting finite-element expansion coefficients rather than unconstrained field values, the Galerkin-consistent formulation preserves compatible trial and test spaces, achieving high accuracy, stable training, and generalization across dielectric, metallic, arrayed, plasmonic, and three-dimensional nanophotonic structures.

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 / 0 minor

Summary. The manuscript introduces FEMONet, a finite-element-constrained operator-learning framework that maps an operator parameter space (encoding physical entities in wave-equation problems) to a Galerkin-consistent solution space of finite-element expansion coefficients for complex-valued optical scattering. It grounds the approach in the variational weak form of the governing vector wave equations, absorbs spatial derivatives into pre-assembled stiffness matrices and load vectors, and claims this yields high accuracy, stable training, and generalization across dielectric, metallic, arrayed, plasmonic, and 3D nanophotonic structures while being the first such Galerkin-consistent operator-learning method.

Significance. If the central claims hold with supporting evidence, the work could meaningfully advance parameterized full-wave nanophotonic simulation by bridging classical FEM variational solvers with neural operators, reducing the cost of repeated solves for families of scattering problems. The emphasis on preserving compatible trial/test spaces via coefficient prediction rather than pointwise fields is a potentially useful architectural choice, though its practical impact remains to be quantified.

major comments (2)
  1. [Abstract] Abstract: The claim that 'by predicting finite-element expansion coefficients rather than unconstrained field values, the Galerkin-consistent formulation preserves compatible trial and test spaces' is load-bearing for the central contribution. This preservation requires (i) the network output to lie exactly in the span of the chosen FE basis for every parameter instance and (ii) the physics loss to be formed exclusively from pre-assembled stiffness/load operators without additional coordinate derivatives or penalty terms. The abstract supplies no indication that either condition is enforced by architecture or loss design.
  2. [Abstract] Abstract: The assertions of 'high accuracy, stable training, and generalization across dielectric, metallic, arrayed, plasmonic, and three-dimensional nanophotonic structures' constitute the primary performance claims, yet the abstract (and the provided text) contains no quantitative error metrics, baselines, ablation studies, or cross-regime results. Without such evidence the soundness of the Galerkin-consistency advantage cannot be evaluated.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments. We address each major comment below and will revise the abstract to improve clarity on the Galerkin-consistent formulation and to include key quantitative highlights from the results.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The claim that 'by predicting finite-element expansion coefficients rather than unconstrained field values, the Galerkin-consistent formulation preserves compatible trial and test spaces' is load-bearing for the central contribution. This preservation requires (i) the network output to lie exactly in the span of the chosen FE basis for every parameter instance and (ii) the physics loss to be formed exclusively from pre-assembled stiffness/load operators without additional coordinate derivatives or penalty terms. The abstract supplies no indication that either condition is enforced by architecture or loss design.

    Authors: We agree the abstract should more explicitly connect the design choices to enforcement of the conditions. The manuscript specifies that the network outputs finite-element expansion coefficients (ensuring outputs lie exactly in the chosen basis for every parameter instance) and that the physics loss is assembled exclusively from pre-computed stiffness matrices and load vectors (removing coordinate derivatives and penalty terms). We will revise the abstract to state these enforcement mechanisms directly. revision: yes

  2. Referee: [Abstract] Abstract: The assertions of 'high accuracy, stable training, and generalization across dielectric, metallic, arrayed, plasmonic, and three-dimensional nanophotonic structures' constitute the primary performance claims, yet the abstract (and the provided text) contains no quantitative error metrics, baselines, ablation studies, or cross-regime results. Without such evidence the soundness of the Galerkin-consistency advantage cannot be evaluated.

    Authors: Quantitative error metrics, baselines, ablation studies, and cross-regime results appear in the results section of the manuscript. To strengthen the abstract we will add concise representative metrics (e.g., relative L2 errors on test sets) and explicit mention of generalization across the listed structure classes while preserving abstract length. revision: yes

Circularity Check

0 steps flagged

No circularity: derivation rests on standard variational FEM without reduction to inputs

full rationale

The paper defines FEMONet by choosing to output finite-element expansion coefficients and to form the physics loss exclusively from pre-assembled stiffness/load operators. This choice directly inherits the Galerkin structure from the classical weak form; it does not derive any new result that is then fed back as a fitted parameter or self-defined quantity. No self-citation chain is load-bearing for the central claim, and the abstract provides no equations that equate a prediction to its own training target by construction. The framework is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The framework rests on the standard variational weak form of the vector wave equation and the assumption that finite-element discretization removes the need for coordinate derivatives in the loss; no free parameters or invented entities are described in the abstract.

axioms (1)
  • domain assumption The variational weak form of the governing vector wave equations links the operator parameter space to the coordinate and physical solution spaces.
    Invoked in the abstract as the grounding for the Galerkin-consistent formulation.

pith-pipeline@v0.9.1-grok · 5761 in / 1069 out tokens · 22053 ms · 2026-06-27T06:02:47.787593+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

44 extracted references · 2 canonical work pages · 1 internal anchor

  1. [1]

    John Wiley & Sons, 2015

    Jian-Ming Jin.The finite element method in electromagnetics. John Wiley & Sons, 2015

    2015

  2. [2]

    Finite-difference time-domain methods.Nature Reviews Methods Primers, 3(1):75, 2023

    FL Teixeira, C Sarris, Y Zhang, D-Y Na, J-P Berenger, Y Su, M Okoniewski, WC Chew, V Backman, and Jamesina J Simpson. Finite-difference time-domain methods.Nature Reviews Methods Primers, 3(1):75, 2023

    2023

  3. [3]

    Artech House, 2009

    Giuseppe Pelosi, Roberto Coccioli, and Stefano Selleri.Quick finite elements for electromagnetic waves. Artech House, 2009

    2009

  4. [4]

    Wiley-IEEE Press, 1993

    Roger F Harrington.Field computation by moment methods. Wiley-IEEE Press, 1993

    1993

  5. [5]

    Springer, 2007

    Stefan A Maier et al.Plasmonics: fundamentals and applications, volume 1. Springer, 2007

    2007

  6. [6]

    Modes and mode volumes of leaky optical cavities and plasmonic nanoresonators.ACS Photonics, 1(1):2–10, 2014

    Philip Trøst Kristensen and Stephen Hughes. Modes and mode volumes of leaky optical cavities and plasmonic nanoresonators.ACS Photonics, 1(1):2–10, 2014

    2014

  7. [7]

    Theory of the spontaneous optical emission of nanosize photonic and plasmon resonators.Physical Review Letters, 2013

    P Lalanne. Theory of the spontaneous optical emission of nanosize photonic and plasmon resonators.Physical Review Letters, 2013

    2013

  8. [8]

    Light propagation with phase discontinuities: generalized laws of reflection and refraction.science, 334(6054):333–337, 2011

    Nanfang Yu, Patrice Genevet, Mikhail A Kats, Francesco Aieta, Jean-Philippe Tetienne, Federico Capasso, and Zeno Gaburro. Light propagation with phase discontinuities: generalized laws of reflection and refraction.science, 334(6054):333–337, 2011

    2011

  9. [9]

    Inverse design in nanophotonics.Nature photonics, 12(11):659–670, 2018

    Sean Molesky, Zin Lin, Alexander Y Piggott, Weiliang Jin, Jelena Vuckovi´c, and Alejandro W Rodriguez. Inverse design in nanophotonics.Nature photonics, 12(11):659–670, 2018

    2018

  10. [10]

    Deep neural networks for the evaluation and design of photonic devices.Nature Reviews Materials, 6(8):679–700, 2021

    Jiaqi Jiang, Mingkun Chen, and Jonathan A Fan. Deep neural networks for the evaluation and design of photonic devices.Nature Reviews Materials, 6(8):679–700, 2021

    2021

  11. [11]

    Maziar Raissi, Paris Perdikaris, and George E Karniadakis. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations.Journal of Computational physics, 378:686–707, 2019

    2019

  12. [12]

    Physics-informed neural networks for inverse problems in nano-optics and metamaterials.Optics express, 28(8):11618–11633, 2020

    Yuyao Chen, Lu Lu, George Em Karniadakis, and Luca Dal Negro. Physics-informed neural networks for inverse problems in nano-optics and metamaterials.Optics express, 28(8):11618–11633, 2020

    2020

  13. [13]

    Maxwellnet: Physics-driven deep neural network training based on maxwell’s equations.Apl Photonics, 7(1), 2022

    Joowon Lim and Demetri Psaltis. Maxwellnet: Physics-driven deep neural network training based on maxwell’s equations.Apl Photonics, 7(1), 2022

    2022

  14. [14]

    arXiv preprint arXiv:1912.00873 , year=

    Ehsan Kharazmi, Zhongqiang Zhang, and George Em Karniadakis. Variational physics-informed neural networks for solving partial differential equations.arXiv preprint arXiv:1912.00873, 2019

    work page arXiv 1912

  15. [15]

    hp-vpinns: Variational physics-informed neural networks with domain decomposition.Computer Methods in Applied Mechanics and Engineering, 374:113547, 2021

    Ehsan Kharazmi, Zhongqiang Zhang, and George Em Karniadakis. hp-vpinns: Variational physics-informed neural networks with domain decomposition.Computer Methods in Applied Mechanics and Engineering, 374:113547, 2021

    2021

  16. [16]

    Learning nonlinear operators via deeponet based on the universal approximation theorem of operators.Nature machine intelligence, 3(3):218–229, 2021

    Lu Lu, Pengzhan Jin, Guofei Pang, Zhongqiang Zhang, and George Em Karniadakis. Learning nonlinear operators via deeponet based on the universal approximation theorem of operators.Nature machine intelligence, 3(3):218–229, 2021

    2021

  17. [17]

    Learning the solution operator of parametric partial differential equations with physics-informed deeponets.Science advances, 7(40):eabi8605, 2021

    Sifan Wang, Hanwen Wang, and Paris Perdikaris. Learning the solution operator of parametric partial differential equations with physics-informed deeponets.Science advances, 7(40):eabi8605, 2021

    2021

  18. [18]

    Fourier Neural Operator for Parametric Partial Differential Equations

    Zongyi Li, Nikola Kovachki, Kamyar Azizzadenesheli, Burigede Liu, Kaushik Bhattacharya, Andrew Stuart, and Anima Anandkumar. Fourier neural operator for parametric partial differential equations.arXiv preprint arXiv:2010.08895, 2020

    work page internal anchor Pith review Pith/arXiv arXiv 2010

  19. [19]

    Understanding and mitigating gradient flow pathologies in physics-informed neural networks.SIAM Journal on Scientific Computing, 43(5):A3055–A3081, 2021

    Sifan Wang, Yujun Teng, and Paris Perdikaris. Understanding and mitigating gradient flow pathologies in physics-informed neural networks.SIAM Journal on Scientific Computing, 43(5):A3055–A3081, 2021

    2021

  20. [20]

    When and why pinns fail to train: A neural tangent kernel perspective.Journal of Computational Physics, 449:110768, 2022

    Sifan Wang, Xinling Yu, and Paris Perdikaris. When and why pinns fail to train: A neural tangent kernel perspective.Journal of Computational Physics, 449:110768, 2022

    2022

  21. [21]

    Deep learning for the design of photonic structures.Nature photonics, 15(2):77–90, 2021

    Wei Ma, Zhaocheng Liu, Zhaxylyk A Kudyshev, Alexandra Boltasseva, Wenshan Cai, and Yongmin Liu. Deep learning for the design of photonic structures.Nature photonics, 15(2):77–90, 2021

    2021

  22. [22]

    Deep learning in nano-photonics: inverse design and beyond.Photonics research, 9(5):B182–B200, 2021

    Peter R Wiecha, Arnaud Arbouet, Christian Girard, and Otto L Muskens. Deep learning in nano-photonics: inverse design and beyond.Photonics research, 9(5):B182–B200, 2021

    2021

  23. [23]

    Neural operator: Learning maps between function spaces with applications to pdes.Journal of Machine Learning Research, 24(89):1–97, 2023

    Nikola Kovachki, Zongyi Li, Burigede Liu, Kamyar Azizzadenesheli, Kaushik Bhattacharya, Andrew Stuart, and Anima Anandkumar. Neural operator: Learning maps between function spaces with applications to pdes.Journal of Machine Learning Research, 24(89):1–97, 2023. 25 Learning light scattering from operator parameter spaces to Galerkin-consistent solution spaces

    2023

  24. [24]

    Machine prediction of topological transitions in photonic crystals.Physical Review Applied, 14(4):044032, 2020

    Bei Wu, Kun Ding, Che Ting Chan, and Yuntian Chen. Machine prediction of topological transitions in photonic crystals.Physical Review Applied, 14(4):044032, 2020

    2020

  25. [25]

    Springer, 2004

    Alexandre Ern and Jean-Luc Guermond.Theory and practice of finite elements, volume 159. Springer, 2004

    2004

  26. [26]

    Springer, 2013

    Daniele Boffi, Franco Brezzi, Michel Fortin, et al.Mixed finite element methods and applications, volume 44. Springer, 2013

    2013

  27. [27]

    The inf–sup condition and its evaluation for mixed finite element methods.Computers & structures, 79(2):243–252, 2001

    Klaus-Jürgen Bathe. The inf–sup condition and its evaluation for mixed finite element methods.Computers & structures, 79(2):243–252, 2001

    2001

  28. [28]

    Robust variational physics-informed neural networks.Computer Methods in Applied Mechanics and Engineering, 425:116904, 2024

    Sergio Rojas, Paweł Maczuga, Judit Muñoz-Matute, David Pardo, and Maciej Paszy ´nski. Robust variational physics-informed neural networks.Computer Methods in Applied Mechanics and Engineering, 425:116904, 2024

    2024

  29. [29]

    Temperature-dependent dark-field scattering of single plasmonic nanocavity.Nanophotonics, 9(10):3347–3356, 2020

    Wei Jiang, Huatian Hu, Qian Deng, Shunping Zhang, and Hongxing Xu. Temperature-dependent dark-field scattering of single plasmonic nanocavity.Nanophotonics, 9(10):3347–3356, 2020

    2020

  30. [30]

    Surface plasmon polariton–enhanced upconversion luminescence for biosensing applications.Nanophotonics, 13(21):3995–4006, 2024

    Duc Le, Marjut Kreivi, Sanna Aikio, Noora Heinilehto, Teemu Sipola, Jarno Petäjä, Tian-Long Guo, Matthieu Roussey, and Jussi Hiltunen. Surface plasmon polariton–enhanced upconversion luminescence for biosensing applications.Nanophotonics, 13(21):3995–4006, 2024

    2024

  31. [31]

    Visible light focusing flat lenses based on hybrid dielectric-metal metasurface reflector-arrays.Scientific reports, 7(1):45044, 2017

    Qingbin Fan, Pengcheng Huo, Daopeng Wang, Yuzhang Liang, Feng Yan, and Ting Xu. Visible light focusing flat lenses based on hybrid dielectric-metal metasurface reflector-arrays.Scientific reports, 7(1):45044, 2017

    2017

  32. [32]

    A high-efficient hybrid physics-informed neural networks based on convolutional neural network

    Zhiwei Fang. A high-efficient hybrid physics-informed neural networks based on convolutional neural network. IEEE Transactions on Neural Networks and Learning Systems, 33(10):5514–5526, 2021

    2021

  33. [33]

    An algorithm for the machine calculation of complex fourier series

    James W Cooley and John W Tukey. An algorithm for the machine calculation of complex fourier series. Mathematics of computation, 19(90):297–301, 1965

    1965

  34. [34]

    Fourier features let networks learn high frequency functions in low dimensional domains.Advances in neural information processing systems, 33:7537–7547, 2020

    Matthew Tancik, Pratul Srinivasan, Ben Mildenhall, Sara Fridovich-Keil, Nithin Raghavan, Utkarsh Singhal, Ravi Ramamoorthi, Jonathan Barron, and Ren Ng. Fourier features let networks learn high frequency functions in low dimensional domains.Advances in neural information processing systems, 33:7537–7547, 2020

    2020

  35. [35]

    Gradient-based learning applied to document recognition.Proceedings of the IEEE, 86(11):2278–2324, 1998

    Yann LeCun, Léon Bottou, Yoshua Bengio, and Patrick Haffner. Gradient-based learning applied to document recognition.Proceedings of the IEEE, 86(11):2278–2324, 1998

    1998

  36. [36]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016

    2016

  37. [37]

    Learning spatiotemporal features with 3d convolutional networks

    Du Tran, Lubomir Bourdev, Rob Fergus, Lorenzo Torresani, and Manohar Paluri. Learning spatiotemporal features with 3d convolutional networks. InProceedings of the IEEE international conference on computer vision, pages 4489–4497, 2015

    2015

  38. [38]

    Language modeling with gated convolutional networks

    Yann N Dauphin, Angela Fan, Michael Auli, and David Grangier. Language modeling with gated convolutional networks. InInternational conference on machine learning, pages 933–941. PMLR, 2017

    2017

  39. [39]

    Pay attention to mlps.Advances in neural information processing systems, 34:9204–9215, 2021

    Hanxiao Liu, Zihang Dai, David So, and Quoc V Le. Pay attention to mlps.Advances in neural information processing systems, 34:9204–9215, 2021

    2021

  40. [40]

    Femonet: Robust and universal operator learning for optical scattering problem via mionet

    Lida Liu. Femonet: Robust and universal operator learning for optical scattering problem via mionet. https://github.com/HUST-CPO/FEMONet-Robust-and-universal-Operator-learning-for-optical-scattering- problem-via-MIONet, 2026

    2026

  41. [41]

    The deep ritz method: A deep learning-based numerical algorithm for solving variational problems

    Yu B EW. The deep ritz method: A deep learning-based numerical algorithm for solving variational problems. Commun Math Stat, 6(1):1–12, 2018

    2018

  42. [42]

    Shahed Rezaei, Reza Najian Asl, Shirko Faroughi, Mahdi Asgharzadeh, Ali Harandi, Rasoul Najafi Koopas, Gottfried Laschet, Stefanie Reese, and Markus Apel. A finite operator learning technique for mapping the elastic properties of microstructures to their mechanical deformations.International Journal for Numerical Methods in Engineering, 126(1):e7637, 2025

    2025

  43. [43]

    Scattering activities bounded by reciprocity and parity conservation.Physical Review Research, 2(1):013277, 2020

    Weijin Chen, Qingdong Yang, Yuntian Chen, and Wei Liu. Scattering activities bounded by reciprocity and parity conservation.Physical Review Research, 2(1):013277, 2020

    2020

  44. [44]

    Arbitrary polarization-independent backscattering or reflection by rotationally symmetric reciprocal structures.Physical Review B, 103(4):045422, 2021

    Weijin Chen, Qingdong Yang, Yuntian Chen, and Wei Liu. Arbitrary polarization-independent backscattering or reflection by rotationally symmetric reciprocal structures.Physical Review B, 103(4):045422, 2021. 26

    2021