Pith. sign in

REVIEW 4 major objections 8 minor 40 references

From Local Interactions to Global Operators: Scalable Gaussian Process Operator for Physical Systems

T0 review · 4 major / 8 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read This paper proposes LoGoS-GPO, a Gaussian Process Operator that combines K-nearest-neighbor spatial kernels, Kronecker factorization, and sparse variational inference to lower operator-learning inference cost from $O(N^3 d^3)$ to $O(M^3 +…

desk verdict A promising blend of sparsification tricks for GPOs, but the central complexity claim rests on an unproven and likely invalid assumption about the inverse of the KNN-truncated kernel. read the letter →

arxiv 2506.15906 v1 pith:5ZPUMY6O submitted 2025-06-18 stat.ML cs.LG

Sawan Kumar , Tapas Tripura , Rajdip Nayek , Souvik Chakraborty This is my paper · ORCID
classification stat.MLcs.LG
keywords GaussianprocessoperatorlearningKroneckerfactorizationK-nearest-neighborkernelsparsevariationalinferencewaveletneuraluncertaintyquantificationparametricPDEs
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

This paper argues that a Gaussian Process Operator—a probabilistic model for learning maps between function spaces—can be made scalable without sacrificing accuracy by combining three structural shortcuts: a Kronecker product splitting the kernel into parameter-space and spatial parts, a K-nearest-neighbor truncation of the spatial covariance, and sparse variational inference over inducing points in the parameter space. On Burgers, wave advection, Darcy flow, and Navier–Stokes benchmarks, the method reports relative L2 errors between 0.43% and 2.01%, lower than the original GPO and a standard sparse GP baseline. If the claim holds, it removes the cubic-complexity barrier that has kept probabilistic neural operators off high-resolution PDE problems, while preserving predictive uncertainty estimates. The paper also introduces an enhanced wavelet neural operator that supplies a non-zero prior mean and a latent-space kernel, which is what the authors credit for compensating the accuracy lost to local sparsification.

What carries the argument

The load-bearing object is the doubly sparse covariance $K = K_a(\Phi(A),\Phi(A')) \otimes K_x(x,x')$: a Kronecker product of a parameter-space kernel evaluated on neural-operator embeddings and a spatial kernel $K_x$ truncated so each row retains only its $K$ nearest neighbors. Kronecker structure lets matrix products and inverses act on the two factors separately, and sparse variational inference with $M$ inducing points replaces the full $Nd$-dimensional posterior with a low-rank one, giving the claimed $O(M^3 + M^2B + dK^2)$ cost. The non-zero mean $m(a,x)$, produced by an enhanced wavelet neural operator, is not decorative: it carries the global, operator-aware information that the localized spatial kernel discards, and the same neural operator defines the latent space $\Phi(\cdot)$ in which the parameter-space kernel is evaluated. The paper's formal route from function-valued GP to this finite model is the probabilistic currying equivalence $G(a)(x) = f(a,x)$.

What would settle it

Compute the smallest eigenvalue of the truncated matrix $K_x$ from Eq. (13) on a uniform 1D grid with an RBF kernel and small $K$; if any eigenvalue is negative for typical lengthscales, the matrix is not a valid covariance and the posterior variance in Eq. (23b) can go negative. Separately, benchmark the inversion step's wall-clock scaling in $d$ at fixed $K$: growth like $d^2$ or $d^3$ rather than $O(dK)$ would falsify the headline complexity claim.

Watch

Extended reading notes

Core claim

The central claim is that a doubly sparse Gaussian Process Operator can match or beat exact GPO accuracy at a fraction of the cost. The proposed model writes the covariance as $K = K_a(\Phi(A),\Phi(A')) \otimes K_x(x,x')$, keeps only the $K$ nearest neighbors of each spatial grid point in $K_x$, and treats the parameter-space factor $K_a$ with sparse variational inference using $M$ inducing points. The paper states that this combination reduces inference cost from $O(N^3 d^3)$ to $O(M^3 + M^2B + dK^2)$, where $N$ is the number of training samples, $d$ the discretization size, and $B$ the mini-batch size. On four nonlinear PDE benchmarks—1D Burgers, 1D wave advection, 2D Darcy flow on a notched triangular domain, and 2D Navier–Stokes—the reported relative L2 errors are 0.86%, 0.43%, 1.38%, and 2.01%, respectively, all below the corresponding GPO and SVGP errors. A separate claim is that the enhanced wavelet neural operator, which replaces the pointwise-multiplication approximation of wavelet convolution with a true convolution integral in the wavelet domain, provides the expressive mean and latent embedding that let the sparsified kernel stay accurate.

Load-bearing premise

The load-bearing premise is that the K-nearest-neighbor-truncated spatial covariance matrix $K_x$ is still a valid covariance whose inverse can be computed in $O(dK)$ time; a sparse matrix with $K$ nonzeros per row is not generally positive definite, and its inverse is generally dense, so Eq. (20)'s use of $[K_x]^{-1}$ and the quoted $O(dK)$ inversion cost are not automatic.

Editorial extensions

If this is right

  • If the complexity claim holds, GPOs become trainable at grid resolutions and sample sizes that were previously out of reach, with per-epoch time and memory growing roughly linearly in the grid size $d$ rather than cubically.
  • The four benchmarks suggest that probabilistic operator learning does not have to trade uncertainty quantification for scalability: predictive intervals remain narrow on unseen inputs from the same distribution, including near the discontinuity in the wave advection case.
  • The non-zero neural-operator mean is a transferable recipe: any discretization-invariant operator could supply the mean and the latent embedding, turning a deterministic neural operator into a scalable Gaussian process with uncertainty estimates.
  • Replacing the pointwise-multiplication approximation in wavelet convolution with a genuine convolution integral in the wavelet domain is presented as an architectural fix that preserves multiscale locality while regaining global interactions.
  • If resolution independence transfers as reported, the model offers a route to training on coarse grids and predicting on finer grids with calibrated uncertainty, which is the practical setting for parametric PDE surrogates.

Reading between the lines

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

  • The step most worth stress-testing is the claimed $O(dK)$ inversion of the KNN-truncated spatial matrix: a sparse matrix with $K$ nonzeros per row is not guaranteed to be positive definite, and its inverse is generally dense, so the stated cost depends on an extra structural assumption the paper leaves implicit.
  • The accuracy comparisons are against GPO and SVGP; a sharper test of whether sparsification costs accuracy would compare against deterministic wavelet or Fourier neural operators at the same resolution and sample budget.
  • A natural extension is to make the neighbor set adaptive to kernel lengthscales or learned, which could recover long-range spatial interactions while retaining the sparse layout.
  • One could probe the role of the operator-aware mean directly by ablating it: if accuracy collapses when the mean is set to zero, the expressive prior mean is doing the work that the sparse kernel alone cannot.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 8 minor

Summary. This paper proposes LoGoS-GPO, a sparse variational Gaussian process operator that combines a Kronecker-factorized covariance over (input-feature, spatial-grid) pairs, K-nearest-neighbor truncation of the spatial covariance, and a neural-operator mean and latent embedding. The central scaling claim in Section 3.1 is that this reduces training cost to O(M^3 + M^2B + dK^2) from O(N^3d^3) for exact inference. The method is evaluated on Burgers, wave advection, Darcy, and Navier-Stokes benchmarks, reporting relative L2 errors of 0.43%-2.01%, which are better than GPO and SVGP, together with scaling experiments in grid size and sample count. An enhanced wavelet neural operator is introduced to provide the mean function and the latent kernel embedding.

Significance. If the scalability claim were established, LoGoS-GPO would be a practically useful step toward uncertainty quantification in operator learning, and the empirical comparison against SVGP and GPO on standard PDE benchmarks is a genuine strength. The paper makes a concrete, falsifiable algorithmic proposal (KNN sparsification plus Kronecker factorization plus sparse variational inference) and reports external benchmarks rather than constructing the evaluation from its own prior outputs. However, the advertised complexity reduction rests on unproved spectral and algorithmic properties of the KNN-truncated spatial matrix, and the current derivation does not support the central claim as written. The empirical results are plausible but cannot by themselves validate the complexity model.

major comments (4)
  1. [§3.1, Eq. (13) and Eq. (20)] The KNN-truncated matrix K_N^x defined in Eq. (13) is not guaranteed to be symmetric, because the K-nearest-neighbor relation is not symmetric, and it is not guaranteed to be positive definite. Even for a symmetrized banded version, truncating an RBF kernel on a fine uniform grid gives a tridiagonal section whose smallest eigenvalue is approximately 1 - 2ρ, which is negative whenever the adjacent-point correlation ρ exceeds 0.5. Since Eq. (20) uses (K_N^x)^{-1} in a Kronecker inverse and Eqs. (21)-(23) treat K_N^x as a covariance, the variational objective and predictive variance are not valid as stated. The construction needs to be repaired, for example by symmetrization with a diagonal jitter or by replacing the KNN truncation with a Vecchia-style sparse precision that is guaranteed positive definite.
  2. [§3.1, Algorithm 1 and complexity statement] Algorithm 1 computes all pairwise distances for each grid point, so constructing K_N^x costs O(d^2) kernel evaluations and O(d^2 log d) sort time, which contradicts the stated O(d log N + dK) construction and inversion cost. Furthermore, the inverse of a sparse matrix is generally dense, and no argument is given that K_N^x admits a bounded-fill-in Cholesky factorization or that its inverse can be applied in O(dK). The overall complexity claim O(M^3 + M^2B + dK^2) is therefore not established; please specify the exact linear algebra (bandwidth, fill-in, or iterative solver) and derive the cost from the actual algorithm.
  3. [§3.2, Eqs. (27)-(31)] The derivation of the enhanced wavelet convolution is not self-contained. Equation (28) gives a convolution in the wavelet translation variable t, but Eq. (29) replaces it with F^{-1}[R_φ · F(Wv_j)(s,ω)] without specifying the domain of the Fourier transform, the relation between R_φ and K_φ, or how the discrete wavelet transform and FFT implementations compose. Please provide a precise statement of the transforms and variables, or give the actual computational formula used in the experiments; without this, the claimed improvement over the original WNO cannot be verified.
  4. [§3.1, Eq. (12)] Equation (12) asserts the separable Kronecker decomposition K = K_a ⊗ K_x, but this does not follow from the latent-space kernel in Eq. (9) unless Φ and k_ax are assumed to factorize over the input-feature and spatial coordinates. The paper should state this explicitly as a modeling assumption and indicate what class of kernels is covered; as written, the decomposition is introduced without derivation or justification.
minor comments (8)
  1. [§3.2, Eq. (25)] Equation (25) uses g*(ω) for the Fourier transform and integrates C_g over Ω with respect to ω; please standardize the notation (e.g., \hat g or \mathcal{F}g) and state the integration domain as the frequency axis.
  2. [§4.2, Eq. (34)] The periodic boundary condition u(x - π) = u(x + π) on x ∈ (0,1) is inconsistent with the stated domain; it should presumably read u(0,t) = u(1,t).
  3. [Table 2] The learning rates are reported in an ambiguous product notation such as '58×10^-3' and '32×10^-2'; please use standard scientific notation.
  4. [Algorithm 2] Algorithm 2 reuses the symbol B for both the mini-batch subset and the mini-batch size; rename one of them to avoid confusion.
  5. [§4.4] The text says Fig. 8 investigates sample-size variation, but Fig. 8 shows model predictions; the reference should probably be to Fig. 9b.
  6. [Table 3] The relative L2 error used in Table 3 is not defined; please provide the formula used to compute it.
  7. [§3.1 vs. §2.2] The notation for kernel hyperparameters is inconsistent: Section 2.2 defines θk = {θ_{k1}, θ_{k2}}, while Section 3.1 uses θk = {θ_{k3}, θa, θx}; please unify the notation.
  8. [§4, enhanced WNO] The paper claims the enhanced WNO as a contribution, but no experiment compares LoGoS-GPO using the enhanced WNO with LoGoS-GPO using the original WNO; without such an ablation, the incremental benefit of the enhanced WNO is not demonstrated.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the core empirical claims are tested on external PDE benchmarks against independent baselines; the sparse-kernel scalability concern is a correctness gap, not a circular reduction.

full rationale

No circular step could be exhibited with the required reduction. The construction builds on the authors' own GPO [21] and WNO [4] works, but these are prior building blocks, not the evidence for the new claims. The central scalability claim (Section 3.1) follows, at least in intention, from KNN sparsification, Kronecker structure, and sparse variational inference, and the accuracy claims are measured on held-out test functions for externally defined PDE benchmarks (e.g., the Burgers data setup of Li et al. [3], the Navier-Stokes forcing-to-vorticity task) with comparison to the independent SVGP baseline of Hensman et al. [22]. The reported L2 errors are therefore not equivalent to any fitted quantity by construction. The only self-referential element worth noting is that the 'enhanced WNO' is not ablated against the original WNO, so the marginal value of that component is not isolated; this is a missing baseline, not circularity. Similarly, the KNN-truncated K_x from Eq. (13) is not guaranteed positive definite and its inverse is generally dense, which casts doubt on the claimed O(dK)/O(dK^2) inversion cost, but an unjustified complexity claim is a correctness risk rather than a circularity. No equation in the paper reduces to its own input, and no load-bearing uniqueness or prior-result citation is used to force the conclusion.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central contribution rests on standard GP machinery plus several unproved structural assumptions: separability of the kernel, positive-definiteness and invertibility of the sparse KNN matrix, and the FFT-based wavelet convolution. The two most important experimental hyperparameters, K and M, are not reported. No code is available. These are the main barriers to trusting the complexity and accuracy claims.

free parameters (4)
  • Number of nearest neighbors K
    Controls spatial sparsity in Eq. (13) and the claimed O(dK^2) complexity; not reported in Table 2 for any experiment.
  • Number of inducing points M
    Controls the O(M^3) variational cost and approximation fidelity; never specified in the experimental setup.
  • Kernel hyperparameters theta_k (lengthscales, signal variances, wavelet filter parameters)
    Learned by maximizing the ELBO but values are not reported, so the kernel and posterior cannot be reconstructed.
  • Neural operator mean parameters theta_m and latent embedding parameters theta_phi
    Optimized end-to-end; architecture widths, depths, and final parameter values are not given.
assumptions (5)
  • standard math The probabilistic currying theorem [33] equates the function-valued GP over operators with a vector-valued GP over augmented inputs (Eq. 6).
    Used to justify the finite-dimensional GP formulation; cited from [33] and not proved in this paper.
  • domain assumption The K-nearest-neighbor truncated kernel matrix K_N^x in Eq. (13) remains a valid positive definite covariance whose inverse is available at O(dK) cost.
    No proof is given; truncation can destroy positive-definiteness and the inverse of a sparse matrix is generally dense, yet Eq. (20) uses [K_x]^{-1}.
  • standard math Fubini's theorem applies to the wavelet-convolution interchange in Eq. (28), requiring square-integrable functions and finite energy.
    The authors state the regularity assumption and invoke Fubini; standard, but the conditions are not verified for the learned features.
  • domain assumption Fourier-domain elementwise multiplication implements the wavelet-domain convolution in Eqs. (29)-(30).
    The paper asserts this equivalence to define the enhanced WNO, but does not derive it rigorously; the symbols s and omega are also mixed in Eq. (29).
  • domain assumption The separable Kronecker structure K = K_a ⊗ K_x in Eq. (12) adequately captures the covariance of each PDE dataset.
    This separability is the basis of the computational acceleration but is not validated against a non-separable kernel.

how reviews work

0 comments
Cite this review

Pith. "Pith review of From Local Interactions to Global Operators: Scalable Gaussian Process Operator for Physical Systems." pith.science (2026). https://pith.science/paper/5ZPUMY6O

@misc{pith2026250615906,
  author       = {Pith},
  title        = {Pith review of: From Local Interactions to Global Operators: Scalable Gaussian Process Operator for Physical Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5ZPUMY6O}},
  note         = {Machine review of arXiv:2506.15906}
}
read the original abstract

Operator learning offers a powerful paradigm for solving parametric partial differential equations (PDEs), but scaling probabilistic neural operators such as the recently proposed Gaussian Processes Operators (GPOs) to high-dimensional, data-intensive regimes remains a significant challenge. In this work, we introduce a novel, scalable GPO, which capitalizes on sparsity, locality, and structural information through judicious kernel design. Addressing the fundamental limitation of cubic computational complexity, our method leverages nearest-neighbor-based local kernel approximations in the spatial domain, sparse kernel approximation in the parameter space, and structured Kronecker factorizations to enable tractable inference on large-scale datasets and high-dimensional input. While local approximations often introduce accuracy trade-offs due to limited kernel interactions, we overcome this by embedding operator-aware kernel structures and employing expressive, task-informed mean functions derived from neural operator architectures. Through extensive evaluations on a broad class of nonlinear PDEs - including Navier-Stokes, wave advection, Darcy flow, and Burgers' equations - we demonstrate that our framework consistently achieves high accuracy across varying discretization scales. These results underscore the potential of our approach to bridge the gap between scalability and fidelity in GPO, offering a compelling foundation for uncertainty-aware modeling in complex physical systems.

Figures

Figures reproduced from arXiv: 2506.15906 by the authors.

Figure 1
Figure 1. LoGoS-GPO architecture. Schematic representation of the proposed architecture for the LoGoS-GPO. The input field a(x) is processed through two distinct uplift projection layers, denoted by P1 and P2, which map the original input into high-dimensional latent representations suitable for the GP mean and kernel modeling paths, respectively. These projected embeddings are then passed through enhanced Wavelet Neural Oper… view at source ↗
Figure 2
Figure 2. 1D Burger. Figure illustrates the predictive performance of the proposed framework on three representative test inputs, each shown in a column. The first row displays the test input u in(x, 0) over a spatial grid of resolution 1024. The second row compares the predicted solution uˆ pred(x, 1) against the ground truth, with shaded regions indicating the 95% confidence interval. The third row presents the pointwise pr… view at source ↗
Figure 3
Figure 3. Scalability and performance comparison for 1D Burger Equation. Figs. 3a and 3b compare LoGoS-GPO and GPO across varying grid resolutions and training sample sizes. LoGoS-GPO demonstrates improved efficiency in accuracy, wall clock time, and memory usage, demonstrating its suitability for large-scale problem settings. steady-state Darcy flow is given by: −∇ · (a(x, y)∇u(x, y)) = f(x, y), (x, y) ∈ Ω, u(x, y) [PITH_FU… view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: 1D Wave advection equation. Figure shows the prediction obtained from the proposed framework LoGoS￾GPO. The first row shows one of the three representative test inputs on a spatial resolution of 200. The second row shows the corresponding ground truth and the mean pred…
Figure 5
Figure 5. Figure 5: Scalability and performance analysis for 1D wave advection equation. We evaluate the LoGoS-GPO and GPO across varying grid sizes (Fig. 5a) and training sample sizes (Fig. 5b). LoGoS-GPO consistently demonstrates lower error, faster training time, and reduced memory foo…
Figure 6
Figure 6. Figure 6: 2D Darcy flow with a notch in the triangular domain. Figure illustrates the predictive performance of the proposed framework on a representative test sample over a spatial domain discretized with a 51 × 51 grid. Each column represents a different subplot: Input shows t…
Figure 7
Figure 7. Figure 7: Scalability and performance comparison for 2D Darcy flow equation with a notch in a triangular domain. Figs. 7a and 7b compare LoGoS-GPO and GPO across varying grid resolutions and training sample sizes. LoGoS-GPO demonstrates improved efficiency in accuracy, wall cloc…
Figure 8
Figure 8. Figure 8: 2D Navier stokes equation. Figure illustrates the predictive performance of the proposed framework on a representative test sample over a spatial domain discretized with a 64 × 64 grid. Each column represents a different subplot: Input shows the one forcing term; Truth…
Figure 9
Figure 9. Figure 9: Scalability and performance comparison for 2D Navier Stokes equation: (a) Grid configuration; (b) Sampling points for training. Figs. (a) and (b) compare LoGoS-GPO and vanilla GPO across varying spatial resolutions and sample sizes for the 2D Navier-Stokes problem. LoG…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 31 canonical work pages

  1. [1]

    Deeponet: Learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operators.arXiv preprint arXiv:1910.03193, 2019

    Lu Lu, Pengzhan Jin, and George Em Karniadakis. Deeponet: Learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operators.arXiv preprint arXiv:1910.03193, 2019

    arXiv 1910

  2. [2]

    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

  3. [3]

    Fourier neural operator for parametric partial differential equations.arXiv preprint arXiv:2010.08895, 2020

    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

    arXiv 2010

  4. [4]

    Tapas Tripura and Souvik Chakraborty. Wavelet neural operator for solving parametric partial differential equations in computational mechanics problems.Computer Methods in Applied Mechanics and Engineering, 404:115783, 2023

    2023

  5. [5]

    Physics informed WNO.Computer Methods in Applied Mechanics and Engineering, 418:116546, 2024

    Navaneeth N., Tapas Tripura, and Souvik Chakraborty. Physics informed WNO.Computer Methods in Applied Mechanics and Engineering, 418:116546, 2024

    work page 2024

  6. [6]

    Multi-fidelity wavelet neural operator surrogate model for time-independent and time-dependent reliability analysis.Probabilistic Engineering Mechanics, 77:103672, 2024

    Tapas Tripura, Akshay Thakur, and Souvik Chakraborty. Multi-fidelity wavelet neural operator surrogate model for time-independent and time-dependent reliability analysis.Probabilistic Engineering Mechanics, 77:103672, 2024

    work page 2024

  7. [7]

    A foundational neural operator that continuously learns without forgetting, 2023

    Tapas Tripura and Souvik Chakraborty. A foundational neural operator that continuously learns without forgetting, 2023

    work page 2023

  8. [8]

    LNO: Laplace Neural Operator for Solving Differential Equations, 2023

    Qianying Cao, Somdatta Goswami, and George Em Karniadakis. LNO: Laplace Neural Operator for Solving Differential Equations, 2023

    work page 2023

Show all 40 references
  1. [9]

    Variational physics-informed neural operator (vino) for solving partial differential equations.Computer Methods in Applied Mechanics and Engineering, 437:117785, 2025

    Mohammad Sadegh Eshaghi, Cosmin Anitescu, Manish Thombre, Yizheng Wang, Xiaoying Zhuang, and Timon Rabczuk. Variational physics-informed neural operator (vino) for solving partial differential equations.Computer Methods in Applied Mechanics and Engineering, 437:117785, 2025

    2025

  2. [10]

    A wavelet neural operator based elastography for localization and quantification of tumors.Computer Methods and Programs in Biomedicine, 232:107436, 2023

    Tapas Tripura, Abhilash Awasthi, Sitikantha Roy, and Souvik Chakraborty. A wavelet neural operator based elastography for localization and quantification of tumors.Computer Methods and Programs in Biomedicine, 232:107436, 2023. 17 Scalable Gaussian Process Operator for Physica...

    2023

  3. [11]

    A physics-informed variational deeponet for predicting crack path in quasi-brittle materials.Computer Methods in Applied Mechanics and Engineering, 391:114587, March 2022

    Somdatta Goswami, Minglang Yin, Yue Yu, and George Em Karniadakis. A physics-informed variational deeponet for predicting crack path in quasi-brittle materials.Computer Methods in Applied Mechanics and Engineering, 391:114587, March 2022

    2022

  4. [12]

    Predicting crack nucleation and propagation in brittle materials using deep operator networks with diverse trunk architectures

    Elham Kiyani, Manav Manav, Nikhil Kadivar, Laura De Lorenzis, and George Em Karniadakis. Predicting crack nucleation and propagation in brittle materials using deep operator networks with diverse trunk architectures. Computer Methods in Applied Mechanics and Engineering, 441:1...

    2025

  5. [13]

    Learning bias corrections for climate models using deep neural operators, 2023

    Aniruddha Bora, Khemraj Shukla, Shixuan Zhang, Bryce Harrop, Ruby Leung, and George Em Karniadakis. Learning bias corrections for climate models using deep neural operators, 2023

    2023

  6. [14]

    Fourcastnet: A global data-driven high-resolution weather model using adaptive fourier neural operators, 2022

    Jaideep Pathak, Shashank Subramanian, Peter Harrington, Sanjeev Raja, Ashesh Chattopadhyay, Morteza Mardani, Thorsten Kurth, David Hall, Zongyi Li, Kamyar Azizzadenesheli, Pedram Hassanzadeh, Karthik Kashinath, and Animashree Anandkumar. Fourcastnet: A global data-driven high-...

    2022

  7. [15]

    Spherical neural operator network for global weather prediction.IEEE Transactions on Circuits and Systems for Video Technology, 2023

    Kenghong Lin, Xutao Li, Yunming Ye, Shanshan Feng, Baoquan Zhang, Guangning Xu, and Ziye Wang. Spherical neural operator network for global weather prediction.IEEE Transactions on Circuits and Systems for Video Technology, 2023

    2023

  8. [16]

    Variational Bayes Deep Operator Network: A data-driven Bayesian solver for parametric differential equations, 2022

    Shailesh Garg and Souvik Chakraborty. Variational Bayes Deep Operator Network: A data-driven Bayesian solver for parametric differential equations, 2022

    2022

  9. [17]

    Randomized prior wavelet neural operator for uncertainty quantification, 2023

    Shailesh Garg and Souvik Chakraborty. Randomized prior wavelet neural operator for uncertainty quantification, 2023

    2023

  10. [18]

    Approximate Bayesian neural operators: Uncertainty quantification for parametric PDEs.ArXiv, abs/2208.01565, 2022

    Emilia Magnani, Nicholas Kramer, Runa Eschenhagen, Lorenzo Rosasco, and Philipp Hennig. Approximate Bayesian neural operators: Uncertainty quantification for parametric PDEs.ArXiv, abs/2208.01565, 2022

    2022 arXiv

  11. [19]

    Kernel methods are competitive for operator learning, 2023

    Pau Batlle, Matthieu Darcy, Bamdad Hosseini, and Houman Owhadi. Kernel methods are competitive for operator learning, 2023

    2023

  12. [20]

    B-pinns: Bayesian physics-informed neural networks for forward and inverse pde problems with noisy data.Journal of Computational Physics, 425:109913, January 2021

    Liu Yang, Xuhui Meng, and George Em Karniadakis. B-pinns: Bayesian physics-informed neural networks for forward and inverse pde problems with noisy data.Journal of Computational Physics, 425:109913, January 2021

    2021

  13. [21]

    Sawan Kumar, Rajdip Nayek, and Souvik Chakraborty. Towards gaussian process for operator learning: An uncertainty aware resolution independent operator learning algorithm for computational mechanics.Computer Methods in Applied Mechanics and Engineering, 435:117664, 2025

    2025

  14. [22]

    Gaussian processes for big data.arXiv preprint arXiv:1309.6835, 2013

    James Hensman, Nicolo Fusi, and Neil D Lawrence. Gaussian processes for big data.arXiv preprint arXiv:1309.6835, 2013

    2013 arXiv

  15. [23]

    Hierarchical nearest-neighbor gaussian process models for large geostatistical datasets.Journal of the American Statistical Association, 111(514):800–812, 2016

    Abhirup Datta, Sudipto Banerjee, Andrew O Finley, and Alan E Gelfand. Hierarchical nearest-neighbor gaussian process models for large geostatistical datasets.Journal of the American Statistical Association, 111(514):800–812, 2016

    2016

  16. [24]

    Variational nearest neighbor gaussian process

    Luhuan Wu, Geoff Pleiss, and John P Cunningham. Variational nearest neighbor gaussian process. InInternational Conference on Machine Learning, pages 24114–24130. PMLR, 2022

    2022

  17. [25]

    Actually sparse variational gaussian processes

    Harry Jake Cunningham, Daniel Augusto de Souza, So Takao, Mark van der Wilk, and Marc Peter Deisenroth. Actually sparse variational gaussian processes. InInternational Conference on Artificial Intelligence and Statistics, pages 10395–10408. PMLR, 2023

    2023

  18. [26]

    Variational fourier features for gaussian processes.Journal of Machine Learning Research, 18(151):1–52, 2018

    James Hensman, Nicolas Durrande, and Arno Solin. Variational fourier features for gaussian processes.Journal of Machine Learning Research, 18(151):1–52, 2018

    2018

  19. [27]

    Kernel interpolation for scalable structured gaussian processes (kiss-gp)

    Andrew Wilson and Hannes Nickisch. Kernel interpolation for scalable structured gaussian processes (kiss-gp). InInternational conference on machine learning, pages 1775–1784. PMLR, 2015

    2015

  20. [28]

    Scalable variational Gaussian process classifica- tion

    James Hensman, Alexander Matthews, and Zoubin Ghahramani. Scalable variational Gaussian process classifica- tion. InArtificial Intelligence and Statistics, pages 351–360. PMLR, 2015

    2015

  21. [29]

    Sparse orthogonal variational inference for gaussian processes

    Jiaxin Shi, Michalis Titsias, and Andriy Mnih. Sparse orthogonal variational inference for gaussian processes. In International Conference on Artificial Intelligence and Statistics, pages 1932–1942. PMLR, 2020

    1932

  22. [30]

    Stochastic gradient descent for gaussian processes done right.arXiv preprint arXiv:2310.20581, 2023

    Jihao Andreas Lin, Shreyas Padhy, Javier Antorán, Austin Tripp, Alexander Terenin, Csaba Szepesvári, José Miguel Hernández-Lobato, and David Janz. Stochastic gradient descent for gaussian processes done right.arXiv preprint arXiv:2310.20581, 2023

    arXiv

  23. [31]

    Operator-valued kernels for learning from functional response data.Journal of Machine Learning Research, 17(20):1–54, 2016

    Hachem Kadri, Emmanuel Duflos, Philippe Preux, Stéphane Canu, Alain Rakotomamonjy, and Julien Audiffren. Operator-valued kernels for learning from functional response data.Journal of Machine Learning Research, 17(20):1–54, 2016. 18 Scalable Gaussian Process Operator for Physic...

    2016

  24. [32]

    Operator-valued bochner theorem, fourier feature maps for operator-valued kernels, and vector-valued learning.arXiv preprint arXiv:1608.05639, 2016

    Ha Quang Minh. Operator-valued bochner theorem, fourier feature maps for operator-valued kernels, and vector-valued learning.arXiv preprint arXiv:1608.05639, 2016

    2016 arXiv

  25. [33]

    Linearization turns neural operators into function-valued gaussian processes, 2024

    Emilia Magnani, Marvin Pförtner, Tobias Weber, and Philipp Hennig. Linearization turns neural operators into function-valued gaussian processes, 2024

    2024

  26. [34]

    Multi-outputs gaussian process for predicting burkina faso covid-19 spread using correlations from the weather parameters.Infectious Disease Modelling, 2022

    Souleymane Zio, Bernard Lamien, Sibiri Tiemounou, Yoda Adaman, Inoussa Tougri, Mohamed Beidari, and Ouedraogo WYS Boris. Multi-outputs gaussian process for predicting burkina faso covid-19 spread using correlations from the weather parameters.Infectious Disease Modelling, 2022

    2022

  27. [35]

    Derivation of output correlation inferences for multi-output (aka multi-task) gaussian process, 2025

    Shuhei Watanabe. Derivation of output correlation inferences for multi-output (aka multi-task) gaussian process, 2025

    2025

  28. [36]

    Alvarez, Lorenzo Rosasco, and Neil D

    Mauricio A. Alvarez, Lorenzo Rosasco, and Neil D. Lawrence. Kernels for vector-valued functions: a review, 2012

    2012

  29. [37]

    Geostatistical space-time models, stationarity, separability, and full symmetry.Monographs On Statistics and Applied Probability, 107:151, 2006

    Tilmann Gneiting, Marc G Genton, and Peter Guttorp. Geostatistical space-time models, stationarity, separability, and full symmetry.Monographs On Statistics and Applied Probability, 107:151, 2006

    2006

  30. [38]

    Springer Berlin Heidelberg, Berlin, Heidelberg, 2004

    Carl Edward Rasmussen.Gaussian Processes in Machine Learning, pages 63–71. Springer Berlin Heidelberg, Berlin, Heidelberg, 2004

    2004

  31. [39]

    Neural operator induced Gaussian process framework for probabilistic solution of parametric partial differential equations, 2024

    Sawan Kumar, Rajdip Nayek, and Souvik Chakraborty. Neural operator induced Gaussian process framework for probabilistic solution of parametric partial differential equations, 2024

    2024

  32. [40]

    Springer, 2015

    Lokenath Debnath and Firdous Ahmad Shah.Wavelet transforms and their applications. Springer, 2015. 19

    2015

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.