mlsanityData to 14 Oct 2026

Beyond Simulation: Retain-and-Repair Neural Operators for Real-World Adaptation

arXiv 2609.39387 ↗First posted 30 Sep 2026cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Neural operators increasingly benefit from pretraining on numerical simulations, yet adapting them for real-world prediction remains challenging. We introduce the Retain-and-Repair Neural Operator (R2NO), a framework for adapting simulation-pretrained operators to real-world data while retaining useful pretrained structure. The pretrained operator is first finetuned on real data and then frozen to provide a source prediction, and a shared repair module learns a sequence of refinements from the same observations. Using orthogonal Fourier projections, a spectral ensemble fits a small ridge regression within each cell of the Fourier domain and combines the refinements by weights fitted on a held-out split of the real data. The cells are defined jointly by radial ranges, angular sectors, and measured channels, allowing refinement depth to vary with frequency magnitude, with orientation, and across channels. Including the source prediction as a candidate makes retention available in every cell, and independently trained repair modules enter the same combination as additional candidates. On all RealPDEBench systems and six backbones, R2NO consistently outperforms full finetuning and iterative refinement. The framework treats adaptation depth as a cell-specific choice learned from real data.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 36 of its 39 references that are in the map. First posted on arXiv 30 Sep 2026.

Builds on

36 of 39 references.

References · 36

  1. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  2. 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,926
  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. 2017Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real Worldcited by 1,498
  6. 2022FourCastNet: A Global Data-driven High-resolution Weather Model using Adaptive Fourier Neural Operatorscited by 776
  7. 2017Sim-to-Real Transfer of Robotic Control with Dynamics Randomizationcited by 672
  8. 2019Deep Equilibrium Modelscited by 615
  9. 2016Adaptive Computation Time for Recurrent Neural Networkscited by 506
  10. 2022PDEBENCH: An Extensive Benchmark for Scientific Machine Learningcited by 413
  11. 2017Snapshot Ensembles: Train 1, get M for freecited by 350
  12. 2024Transolver: A Fast Transformer Solver for PDEs on General Geometriescited by 326
24 more
  1. 2025Reasoning with Latent Thoughts: On the Power of Looped Transformerscited by 240
  2. 2023Looped Transformers as Programmable Computerscited by 237
  3. 2019Depth-Adaptive Transformercited by 236
  4. 2023Spherical Fourier Neural Operators: Learning Stable Dynamics on the Spherecited by 233
  5. 2018Closing the Sim-to-Real Loop: Adapting Simulation Randomization with Real World Experiencecited by 217
  6. 2023Convolutional Neural Operators for robust and accurate learning of PDEscited by 214
  7. 2024Poseidon: Efficient Foundation Models for PDEscited by 200
  8. 2024DPOT: Auto-Regressive Denoising Operator Transformer for Large-Scale PDE Pre-Trainingcited by 151
  9. 2019TossingBot: Learning to Throw Arbitrary Objects with Residual Physicscited by 137
  10. 2025Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Computationcited by 133
  11. 2019Learning Neural PDE Solvers with Convergence Guaranteescited by 77
  12. 2022Learning Deep Implicit Fourier Neural Operators (IFNOs) with Applications to Heterogeneous Material Modelingcited by 65
  13. 2025FourCastNet 3: A geometric approach to probabilistic machine-learning weather forecasting at scalecited by 62
  14. 2026LoopFormer: Elastic-Depth Looped Transformers for Latent Reasoning via Shortcut Modulationcited by 61
  15. 2022Real-time high-resolution CO2 geological storage prediction using nested Fourier neural operatorscited by 47
  16. 2023Hypernetwork-based Meta-Learning for Low-Rank Physics-Informed Neural Networkscited by 44
  17. 2024Parameterized Physics-informed Neural Networks for Parameterized PDEscited by 38
  18. 2018Augmenting Physical Simulators with Stochastic Neural Networks: Case Study of Planar Pushing and Bouncingcited by 36
  19. 2026Fixed-Point Reasoners: Stable and Adaptive Deep Looped Transformerscited by 28
  20. 2023Deep Equilibrium Based Neural Operators for Steady-State PDEscited by 19
  21. 2026RealPDEBench: A Benchmark for Complex Physical Systems with Real-World Datacited by 16
  22. 2025Fourier Neural Operator based surrogates for CO2 storage in realistic geologiescited by 3
  23. 2026Iterative Refinement Neural Operators are Learned Fixed-Point Solvers: A Principled Approach to Spectral Bias Mitigationcited by 2
  24. 2026FLAME: Physics-Guided Neural Operators for Onboard Satellite Methane Detection in Hyperspectral Imagerycited by 1

Built on by

Not cited yet.