Pith. sign in

REVIEW 21 cited by

Operator Learning: Algorithms and Analysis

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2402.15715 v1 pith:T7FUZL4N submitted 2024-02-24 cs.LG cs.NAmath.NA

Nikola B. Kovachki , Samuel Lanthaler , Andrew M. Stuart This is my paper · ORCID
classification cs.LGcs.NAmath.NA
keywords operatorsneurallearningapproximateapproximationfunctionsmodelsoperator
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

Operator learning refers to the application of ideas from machine learning to approximate (typically nonlinear) operators mapping between Banach spaces of functions. Such operators often arise from physical models expressed in terms of partial differential equations (PDEs). In this context, such approximate operators hold great potential as efficient surrogate models to complement traditional numerical methods in many-query tasks. Being data-driven, they also enable model discovery when a mathematical description in terms of a PDE is not available. This review focuses primarily on neural operators, built on the success of deep neural networks in the approximation of functions defined on finite dimensional Euclidean spaces. Empirically, neural operators have shown success in a variety of applications, but our theoretical understanding remains incomplete. This review article summarizes recent progress and the current state of our theoretical understanding of neural operators, focusing on an approximation theoretic point of view.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 21 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Quasi-optimal hierarchically semi-separable matrix approximation

    math.NA 2025-05 conditional novelty 8.0 of 10

    A randomized algorithm produces an HSS approximation with expected error at most O(log(N/k)) times optimal, using O(k log(N/k)) matrix-vector products.

  2. Optimizing the Preconditioner: A Black-box Online-to-Nonconvex Conversion with Static Regret Minimization Oracles

    cs.LG 2026-07 conditional novelty 7.0 of 10

    An OCO algorithm with only O(√T) static regret, pluggable as a preconditioner selector, recovers the classical O(1/√T) stationarity rate on smooth stochastic nonconvex problems and the O(T^{-2/7}) rate on nonsmooth ones.

  3. Computational Math with Neural Networks is Hard

    math.NA 2025-05 conditional novelty 7.0 of 10

    Under SETH, approximating integrals, Poisson solutions, or matrix-vector products for neural network inputs requires runtime at least accuracy^{-1+o(1)}.

  4. Probabilistic operator learning: generative modeling and uncertainty quantification for foundation models of differential equations

    stat.ML 2025-09 conditional novelty 6.0 of 10

    ICON is shown to compute the posterior predictive mean of differential equation solutions, and a generative extension, GenICON, provides samples from this distribution for uncertainty quantification.

  5. Contextual Online Decision Making with Infinite-Dimensional Functional Regression

    stat.ML 2025-01 reject novelty 6.0 of 10

    A unified online decision-making framework that learns context-dependent CDFs via infinite-dimensional functional regression, with regret controlled by the eigenvalue decay of a design integral operator.

  6. LazyDINO: Fast, scalable, and efficiently amortized Bayesian inversion via structure-exploiting and surrogate-driven measure transport

    math.NA 2024-11 conditional novelty 6.0 of 10

    A new amortized Bayesian inversion method trains a derivative-informed neural surrogate of the parameter-to-observable map and then uses it to optimize a lazy transport map in a low-dimensional latent space.

  7. Shopping Companion: Benchmarking and Training LLM Agents for Long-Horizon Preference-Grounded E-Commerce Tasks

    cs.CL 2026-03 unverdicted novelty 5.0 of 10

    A new long-horizon preference-grounded shopping benchmark shows SOTA LLMs below 70% success, while a 4B model fine-tuned with tool-wise process rewards beats stronger baselines.

  8. Diffeomorphic Neural Operator Learning

    math.NA 2025-08 unverdicted novelty 5.0 of 10

    A neural operator that evolves fields by composing learned diffeomorphisms, enforcing relabeling symmetry and targeting conservative, non-diffusive turbulent forecasts.

  9. Mondrian: Transformer Operators via Domain Decomposition

    cs.LG 2025-06 conditional novelty 5.0 of 10

    Mondrian applies transformer attention to subdomain-restricted functions, decoupling the model from the grid resolution.

  10. Convergent Operator-Splitting Scheme for Viscosity Solutions: A Foundation for Learning Domain-to-Solution Maps

    math.NA 2025-05 reject novelty 5.0 of 10

    A splitting FEM for constrained viscosity solutions is claimed to be convergent and to serve as a blueprint for a curse-of-dimensionality-free neural operator, but key proof steps are unsubstantiated.

  11. Graph-Based Operator Learning from Limited Data on Irregular Domains

    cs.LG 2025-05 reject novelty 5.0 of 10

    GOLA combines attention-based graph message passing with a learnable Fourier encoder and reports lower relative L2 error than GKN on four 2D PDE benchmarks, especially with few training samples.

  12. Learning convolution operators on compact Abelian groups

    cs.LG 2025-01 conditional novelty 5.0 of 10

    Ridge regression in translation-invariant Hilbert spaces learns convolution operators on compact Abelian groups at standard optimal rates, with source and capacity conditions reinterpreted as space versus frequency lo...

  13. Approximation Rates in Fr\'echet Metrics: Barron Spaces, Paley-Wiener Spaces, and Fourier Multipliers

    math.NA 2024-12 conditional novelty 5.0 of 10

    Two theorems give sufficient shallow-network width to reach a prescribed error in a Fréchet metric of semi-norms, applied to exponential spectral Barron, Gelfand-Shilov, and bandlimited (Paley-Wiener type) symbol classes.

  14. Convergence analysis of wide shallow neural operators within the framework of Neural Tangent Kernel

    cs.LG 2024-12 conditional novelty 5.0 of 10

    Over-parameterized shallow neural operators trained by gradient descent converge linearly to the global minimum of the empirical loss under mild sample conditions.

  15. Loss Terms and Operator Forms of Koopman Autoencoders

    cs.LG 2024-12 conditional novelty 5.0 of 10

    A benchmark study recommends full accuracy loss, reconstruction loss, unitary loss, and tridiagonal operator form for Koopman autoencoders.

  16. DPNO: A Dual Path Architecture For Neural Operator

    math.NA 2025-07 conditional novelty 4.0 of 10

    Applying a ResNet-like plus DenseNet-like dual path to DeepONet and FNO reduces relative L2 error on Burgers, Darcy flow, and 2D Navier-Stokes benchmarks compared with the original single-path models.

  17. An AI Approach for Learning the Spectrum of the Laplace-Beltrami Operator

    cs.CV 2025-07 conditional novelty 4.0 of 10

    A graph neural network predicts eigenvalues 2-50 of the Laplace-Beltrami operator on mechanical CAD meshes about 5 times faster than FEM and is claimed accurate on 99.3% of held-out parts.

  18. Principled Approaches for Extending Neural Architectures to Function Spaces for Operator Learning

    cs.LG 2025-06 conditional novelty 4.0 of 10

    A practical recipe to convert common neural architectures into discretization-agnostic neural operators, validated by Navier-Stokes experiments showing cross-resolution generalization of FNO-style models.

  19. Adversarial Autoencoders in Operator Learning

    cs.LG 2024-12 conditional novelty 4.0 of 10

    Adding adversarial autoencoder training improves DeepONet and Koopman autoencoder accuracy by 4% to 27% on five small-data benchmarks.

  20. Some Best Practices in Operator Learning

    cs.LG 2024-12 reject novelty 4.0 of 10

    Across six operator-learning tasks, GELU was the best activation, dropout always hurt, stochastic weight averaging helped only for small learning rates, and a learning rate finder was reported but not shown.

  21. STONet: A neural operator for modeling solute transport in micro-cracked reservoirs

    cs.LG 2024-12 conditional novelty 4.0 of 10

    STONet, a neural operator combining DeepONet with transformer-style attention, predicts contaminant concentration fields in micro-cracked reservoirs with reported relative errors below 1% against finite-element simula...

Pith tools