mlsanityData to 14 Oct 2026

Green-Routed Neural Operators: Physics Determines Where the Network Reads

arXiv 2610.05337 ↗First posted 4 Oct 2026cs.LG cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

We identify a mismatch between the physical role of transport fields in many PDEs and their usual role in neural operators: PDEs use them to select read coordinates, whereas neural operators typically treat them only as input values. We address this mismatch with the Green-Routed Neural Operator (GRNO), which uses the governing equation to determine where latent features are sampled. A parameter-free equation adapter evaluates the diagnostic relation and constructs a departure map whose values are the read coordinates. A multiscale encoder-decoder combines centered and routed reads of latent features to learn the complete finite-time update. Across five two- and three-dimensional PDE systems, GRNO achieves the lowest mean final relative L2 error on four under 40-step autoregressive evaluation and remains competitive on Keller-Segel. Fixed-weight route interventions reveal strong dependence on direction and spatial alignment in four systems, with weak dependence in Keller-Segel. In independently trained ablations, GRNO achieves lower mean errors than variants that supply the transport field only as an input feature, substitute a learned displacement for the equation-specified route, or apply the route with a spatial misalignment, across all five systems. It also substantially outperforms directly advecting the physical state and learning the remaining update, indicating that equation-specified read coordinates provide an effective structural prior for long-horizon PDE forecasting.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 32 of its 43 references that are in the map. First posted on arXiv 4 Oct 2026.

Builds on

32 of 43 references.

References · 32

  1. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  2. 2022A ConvNet for the 2020scited by 3,579
  3. 2020Fourier Neural Operator for Parametric Partial Differential Equationscited by 2,908
  4. 2019DeepONet: Learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operatorscited by 2,018
  5. 2017Deformable Convolutional Networkscited by 1,488
  6. 2015Spatial Transformer Networkscited by 1,429
  7. 2020Neural Operator: Graph Kernel Network for Partial Differential Equationscited by 734
  8. 2021Machine learning accelerated computational fluid dynamicscited by 471
  9. 2022Fourier Neural Operator with Learned Deformations for PDEs on General Geometriescited by 427
  10. 2022Message Passing Neural PDE Solverscited by 386
  11. 2020Multipole Graph Neural Operator for Parametric Partial Differential Equationscited by 349
  12. 2021Choose a Transformer: Fourier or Galerkincited by 343
20 more
  1. 2024Transolver: A Fast Transformer Solver for PDEs on General Geometriescited by 326
  2. 2023Geometry-Informed Neural Operator for Large-Scale 3D PDEscited by 303
  3. 2021Multiwavelet-based Operator Learning for Differential Equationscited by 234
  4. 2023Spherical Fourier Neural Operators: Learning Stable Dynamics on the Spherecited by 233
  5. 2022U-NO: U-shaped Neural Operatorscited by 230
  6. 2023PDE-Refiner: Achieving Accurate Long Rollouts with Neural PDE Solverscited by 216
  7. 2023Convolutional Neural Operators for robust and accurate learning of PDEscited by 214
  8. 2020Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solverscited by 205
  9. 2024CViT: Continuous Vision Transformer for Operator Learningcited by 51
  10. 2023Group Equivariant Fourier Neural Operators for Partial Differential Equationscited by 45
  11. 2021Data-driven discovery of Green's functions with human-understandable deep learningcited by 44
  12. 2024Differentiability in Unrolled Training of Neural Physics Simulators on Transient Dynamicscited by 27
  13. 2022Learning Green's functions associated with time-dependent partial differential equationscited by 26
  14. 2024BENO: Boundary-embedded Neural Operators for Elliptic PDEscited by 24
  15. 2024Neural Green's Operators for Parametric Partial Differential Equationscited by 11
  16. 2026Learning to Advect: A Neural Semi-Lagrangian Architecture for Weather Forecastingcited by 4
  17. 2025Generalized Spherical Neural Operators: Green's Function Formulationcited by 3
  18. 2025Neural Interpretable PDEs: Harmonizing Fourier Insights with Attention for Scalable and Interpretable Physics Discoverycited by 2
  19. 2026DGNet: Discrete Green Networks for Data-Efficient Learning of Spatiotemporal PDEscited by 2
  20. 2026Flowers: A Warp Drive for Neural PDE Solverscited by 2

Built on by

Not cited yet.