Pith. sign in

REVIEW 4 major objections 4 minor 55 references

A gradient-based and determinant-free framework for fully Bayesian Gaussian process regression

T0 review · 4 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Fully Bayesian Gaussian process regression can be sampled without determinant evaluations by adding an auxiliary Gaussian field and using gradient-based Hamiltonian Monte Carlo, giving quadratic per-step cost and linear memory.

desk verdict Solid gradient-based, determinant-free HMC for fully Bayesian GPR that deserves referee time, but the 'asymptotically unbiased' claim needs to be qualified for the finite pole expansion. read the letter →

arxiv 2412.20884 v2 pith:B6MQ73D3 submitted 2024-12-30 math.NA cs.NA

P. Michael Kielstra , Michael Lindsey This is my paper · ORCID
classification math.NAcs.NA MSC 65C4065F6062F15
keywords GaussianprocessregressionfullyBayesianinferenceHamiltonianMonteCarlodeterminant-freepseudofermionmatrix-freepoleexpansionkernelhyperparameters
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

Fully Bayesian Gaussian process regression treats kernel hyperparameters as unknown quantities to be inferred, but the posterior density contains the determinant of an $N\times N$ kernel matrix, whose repeated evaluation normally costs $O(N^3)$. This paper argues that the determinant can be removed entirely by rewriting its inverse square root as a Gaussian integral over an auxiliary variable, the pseudofermion trick, and sampling the joint distribution of hyperparameters and auxiliary variable. The authors couple this with Hamiltonian Monte Carlo, computing gradients through linear solves and automatic differentiation of quadratic forms, so each sampling step costs $O(N^2)$ operations and $O(N)$ memory in a matrix-free implementation. If the argument holds, fully Bayesian inference over many kernel hyperparameters becomes practical for large datasets, while avoiding the bias introduced by randomized log-determinant estimators.

What carries the argument

The load-bearing identity is the Gaussian integral $[\det A(\theta)]^{-1/2}\propto\int_{\mathbb{R}^N}\exp(-\tfrac{1}{2}\phi^\top A(\theta)\phi)\,d\phi$, which converts the determinant into an auxiliary Gaussian variable (the 'pseudofermion'). Two pieces of numerical machinery carry the argument: the pole expansion $A(\theta)^{-1/2}\approx\sum_{p=1}^{N_p} w_p(A(\theta)+\lambda_p I)^{-1}$ for drawing $\phi$, and the force formula in HMC that needs only linear solves plus automatic differentiation of quadratic forms. For the implicit-midpoint variant, Anderson-accelerated fixed-point iteration solves the integrator equations and the linear system simultaneously, linking its convergence to GMRES.

What would settle it

On a small problem where the exact posterior can be computed by dense quadrature, compare chains run with $N_p=5,15,30$ while holding every other setting fixed; if the estimated posterior shifts by more than the Monte Carlo standard error as $N_p$ increases, the finite pole expansion is biasing the target distribution.

Watch

Extended reading notes

Core claim

The paper's central claim is that the hyperparameter posterior $P(\theta)\propto |A(\theta)|^{-1/2}e^{-\frac{1}{2} y^\top A(\theta)^{-1}y}p(\theta)$ can be sampled without ever computing the determinant. The authors introduce an auxiliary Gaussian field $\phi\sim\mathcal{N}(0,A(\theta)^{-1})$ and target the joint density $P(\theta,\phi)\propto e^{-S(\theta)-\frac{1}{2}y^\top A(\theta)^{-1}y-\frac{1}{2}\phi^\top A(\theta)\phi}$, whose marginal over $\phi$ reproduces the determinant factor exactly. Conditioned on $\theta$, each $\phi$ update draws $\xi\sim\mathcal{N}(0,I_N)$ and computes $\phi\approx A(\theta)^{-1/2}\xi$ via a rapidly convergent pole expansion that requires only shifted positive-definite solves. Conditioned on $\phi$, each $\theta$ update uses random-walk Metropolis or HMC, with the force expressed as $\nabla_\theta[S(\theta)-\frac{1}{2}x^\top A(\theta)x+\frac{1}{2}\phi^\top A(\theta)\phi]$ where the solve $x=A(\theta)^{-1}y$ is held fixed. The resulting sampler is asymptotically unbiased HMC with $O(N^2)$ work per step and linear memory, provided the pole expansion is taken to convergence.

Load-bearing premise

The sampler's unbiasedness is proven only in the limit as the number $N_p$ of pole terms tends to infinity; with the finite $N_p=15$ used in experiments, the paper gives no bound on the resulting bias in the posterior over hyperparameters.

Editorial extensions

If this is right

  • Fully Bayesian GPR becomes practicable for kernel matrices with tens of thousands of points and dozens of hyperparameters on a single GPU, because per-step work is $O(N^2)$ and memory is $O(N)$.
  • The sampler sidesteps the hard-to-control bias of randomized log-determinant estimators; the only approximation entering the target is the finite pole expansion for drawing $\phi$.
  • Per-step cost depends essentially on the linear solves, not on the number of hyperparameters, so flexible non-stationary kernels with many parameters can be inferred without a per-parameter slowdown.
  • The implicit-midpoint HMC variant maintains high acceptance probability as the number of hyperparameters grows, whereas leapfrog HMC and random-walk Metropolis acceptance falls in the reported experiments.
  • Because the core operations are matrix-free, the same sampler can exploit sparsity or fast matrix-vector products, potentially pushing the per-step cost below $O(N^2)$.

Reading between the lines

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

  • The same Gaussian-integral elimination of a determinant should transfer to other Bayesian models with a Gaussian prior over latent fields, wherever applying the covariance inverse square root by iterative solves is feasible.
  • A pragmatic diagnostic suggested by the finite-pole approximation is to run short chains at increasing $N_p$ and check stability of the posterior, since the pole count needed for fixed accuracy grows only logarithmically with the condition number.
  • The implicit integrator's stable acceptance at high parameter dimension hints that higher-order symplectic implicit integrators, or tighter coupling between integrator and linear solves, could extend the usable step size, though the nonlinear solve's convergence is the bottleneck.
  • Combining the approach with inducing-point or sparse kernel approximations would likely extend it to very large $N$, since the method's cost is dominated by kernel matrix-vector products.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper presents a fully Bayesian Gaussian process regression sampler that avoids log-determinant evaluation by augmenting the hyperparameter posterior with a Gaussian auxiliary field φ (the pseudofermion trick). The conditional refresh of φ is performed by approximating A(θ)^{-1/2}ξ with a pole expansion; θ is updated by RWM or HMC using a force computed by automatic differentiation of quadratic forms. The authors also propose an implicit-midpoint HMC variant solved by Anderson-accelerated fixed-point iteration. Numerical experiments verify against quadrature in a two-hyperparameter example, demonstrate near-O(N^2) scaling up to about 6000 points, and apply the method to 3000-point satellite CO2 data with up to 18 hyperparameters. The abstract and introduction claim asymptotically unbiased HMC with O(N^2) operations per step and O(N) memory.

Significance. The contribution is potentially significant: if the numerical approximations can be controlled, the framework gives a practical, flexible, GPU-friendly route to fully Bayesian GPR in high-dimensional hyperparameter spaces, a problem where determinant-based HMC is prohibitive. The algebraic core is sound: the Gaussian integral identity is exact, the force expression (3.6) is an algebraic identity that avoids differentiating through iterative solvers, and the verification with B=500 chains against exact quadrature in Section 4.1 is good evidence that the implemented pipeline works in a small well-conditioned setting. The authors also ship code, use external benchmarks, and report wall-clock scaling consistent with the claimed complexity. The main caveats are that the finite pole expansion changes the target distribution without a quantified bias bound, and the implicit integrator's nonlinear solve is only heuristic.

major comments (4)
  1. [Section 3.1, Algorithm 1, Abstract/Introduction] The refresh step in Algorithm 1 (line 3) uses the finite pole expansion (3.2) with Np=15 to form φ ≈ A(θ)^{-1/2}ξ. This draw is not exactly distributed as N(0,A(θ)^{-1}) and it is not followed by a Metropolis correction, so the chain generated by Algorithm 1 is not stationary for the joint density (3.1), and the θ-marginal is not exactly P(θ) in (2.2). Consequently the 'asymptotically unbiased HMC' statements in the abstract and Section 1 are justified only in the limit Np → ∞ and iterative-solver tolerance → 0. Appendix A bounds the operator-norm error of the pole approximation, but no bound is given for the induced bias in the posterior distribution of θ, and the error depends on the conditioning of A(θ), which grows with N. This matters because Section 1 criticizes trace-estimator methods for introducing a posterior bias that is difficult to control a priori; the present method has the same structural issue at fixed Np. I recommend either bounding the posterior bias (e.g., via a perturbation bound on the conditional refresh) or explicitly qualifying the unbiasedness claims and adding a sensitivity study in Np.
  2. [Section 4.1] The correctness verification is carried out only for N=10 and n=2 hyperparameters (Section 4.1, Figures 4.1–4.2, Tables 1–2). Because the pole-expansion error scales with the condition number M/m, agreement at N=10 cannot calibrate the bias for the large, less well-conditioned problems targeted in Sections 4.2–4.3. A tractable intermediate check (for example, comparing moment estimates at Np = 15, 30, 60 on a problem with N ≈ 100–300 where the determinant can still be computed exactly) would provide evidence that the chosen Np is adequate in the regime where the method is claimed to be useful.
  3. [Section 3.2.3] The implicit-midpoint HMC variant rests on solving the nonlinear system (3.7) by Anderson-accelerated fixed-point iteration (3.8). The paper explicitly states that sufficient conditions for the full nonlinear iteration to converge are outside its scope, and the convergence results cited (Appendix B, [54]) are for the linear case. If the fixed-point iteration is stopped before convergence or diverges, the map used in the HMC proposal is not the symplectic implicit-midpoint flow, and the Metropolis-Hastings acceptance probability in Algorithm 1 is no longer a valid correction for the target density. The experiments do not report Anderson iteration counts, residuals, or failure rates. This gap does not affect the leapfrog-based method, but it is a load-bearing issue for one of the paper's two HMC proposals, especially because Figure 4.4 highlights the implicit integrator's stable acceptance in high dimensions.
  4. [Section 1 / Section 4.2] The complexity claim of O(N^2) operations per step is stated without accounting for the numerical parameters. If Np must grow with N to control the bias identified above, the cost per φ-refresh grows accordingly (logarithmically in condition number, by Appendix A, but the constant is not quantified). The empirical scaling in Figure 4.3 is consistent with O(N^2) at fixed Np=15 and fixed solver tolerances, but it does not establish O(N^2) at a controlled posterior bias as N grows. The claim should be qualified as O(N^2) at fixed numerical tolerances.
minor comments (4)
  1. [Section 3.1] The sentence after Eq. (3.2), 'using the the pole expansion,' contains a duplicated 'the'.
  2. [Section 3.2.2] In the leapfrog update equations, the second line is printed as π1 ← π1 − (Δt)F_φ(θ_{1/2}); it should be π1 ← π0 − (Δt)F_φ(θ_{1/2}), otherwise the scheme is undefined.
  3. [Section 4.3] In the IAT analysis, the default emcee criterion N ≥ 50τ is not met for the larger Ncheb cases (the text states N ≥ 10τ in the most complicated case), so the reported 95% confidence interval radii may be underestimated; a longer chain or an explicit caveat would be appropriate.
  4. [Section 4.2] The fixed step size Δt = 0.01 is used for all proposal mechanisms, but RWM and the implicit integrator are not otherwise tuned; the wall-clock comparison in Figure 4.3 may therefore favor leapfrog HMC more than a fully tuned comparison would.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the determinant-free derivation is self-contained and verified externally.

full rationale

The paper's central claim is not circular. The target posterior (2.2) is standard, and the determinant is eliminated by the exact Gaussian integral identity [det A(θ)]^{-1/2} ∝ ∫ e^{-(1/2)φ^T A(θ)φ} dφ, so the extended density (3.1) marginalizes exactly to (2.2). The Gibbs refresh φ = A(θ)^{-1/2} ξ is an exact sampling prescription, and replacing it with the Np-term pole expansion (3.2) is a numerical approximation with a quoted convergence rate, not a fitted parameter or a renamed input. The HMC force F_φ is obtained by direct differentiation of U_φ, and the reformulation via x_θ = A(θ)^{-1}y with the dependence of x on θ frozen is an algebraic identity, not an assumption that imports the conclusion. Hyperparameters are not fitted to the quantities being verified: the Section 4.1 check against dense quadrature is an external benchmark, and the CO2 experiment is an application. Self-citations ([23], [37], [41]) are used only for Nyström options, conditioning scalings, and the standard pseudofermion trick in physics; none carries a load-bearing uniqueness or ansatz claim, and none is the sole support for the framework. The finite Np=15 pole expansion and iterative-solver tolerances may introduce a bias in the implemented sampler, but this is an approximation-error or correctness concern about the precise meaning of 'asymptotically unbiased,' not circularity in which an output is equivalent to an input by construction.

Assumptions & free parameters 1 free parameters · 7 assumptions · 0 invented entities

The method rests on standard mathematical identities plus two practically motivated assumptions: the finite pole expansion is accurate enough that the posterior bias is negligible, and the nonlinear fixed-point iteration for the implicit integrator converges. The latter is explicitly unproved in the text.

free parameters (1)
  • Number of pole expansion terms Np = 15
    Chosen by hand; controls the accuracy of the approximate A^{-1/2} multiplication and therefore the bias in the phi update. The paper does not provide an automatic selection or a posterior bias bound.
assumptions (7)
  • standard math Gaussian integral identity: det(A)^(-1/2) ∝ ∫ e^{-1/2 φ^T A φ} dφ
    Exact identity used to remove the determinant; Section 3.
  • standard math Pole expansion for A^{-1/2} converges at the stated rate (Hale et al. 2008)
    Used to sample phi approximately; Appendix A.
  • standard math HMC with a symplectic integrator and Metropolis correction leaves the target P(θ|φ) invariant
    Standard MCMC theory; Section 3.2.
  • domain assumption A(θ) is positive definite for all θ in the support
    Required for the Gaussian integral, the pole expansion, and the linear solves; Section 2.3.
  • domain assumption The observation model (2.1) with Gaussian noise is correct
    Defines the posterior (2.2); used in all experiments.
  • ad hoc to paper Finite pole expansion (Np=15) yields a phi sample close enough that posterior bias is negligible
    The sampler runs with fixed Np; the resulting bias is not quantified.
  • ad hoc to paper Existence, uniqueness, and convergence of the Anderson-accelerated fixed-point iteration for the implicit midpoint system
    Required for the implicit HMC proposal; authors state the analysis is out of scope.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A gradient-based and determinant-free framework for fully Bayesian Gaussian process regression." pith.science (2026). https://pith.science/paper/B6MQ73D3

@misc{pith2026241220884,
  author       = {Pith},
  title        = {Pith review of: A gradient-based and determinant-free framework for fully Bayesian Gaussian process regression},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B6MQ73D3}},
  note         = {Machine review of arXiv:2412.20884}
}
read the original abstract

Gaussian Process Regression (GPR) is widely used for inferring functions from noisy data. GPR crucially relies on the choice of a kernel, which might be specified in terms of a collection of hyperparameters that must be chosen or learned. Fully Bayesian GPR seeks to infer these kernel hyperparameters in a Bayesian sense, and the key computational challenge in sampling from their posterior distribution is the need for frequent determinant evaluations of large kernel matrices. This paper introduces a gradient-based, determinant-free approach for fully Bayesian GPR that combines a Gaussian integration trick for avoiding the determinant with Hamiltonian Monte Carlo (HMC) sampling. Our framework permits a matrix-free formulation and reduces the difficulty of dealing with hyperparameter gradients to a simple automatic differentiation. Our implementation is highly flexible and leverages GPU acceleration with linear-scaling memory footprint. Numerical experiments demonstrate the method's ability to scale gracefully to both high-dimensional hyperparameter spaces and large kernel matrices.

Figures

Figures reproduced from arXiv: 2412.20884 by the authors.

Figure 4.1
Figure 4.1. Uniform norm error of the empirical marginal CDF for the two components [PITH_FULL_IMAGE:figures/full_fig_p013_4_1.png] view at source ↗
Figure 4.2
Figure 4.2. Error of the empirical means for the two components [PITH_FULL_IMAGE:figures/full_fig_p015_4_2.png] view at source ↗
Figure 4.3
Figure 4.3. Computational time per sampling step of various methods on an A100 GPU. [PITH_FULL_IMAGE:figures/full_fig_p016_4_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4.4
Figure 4.4. Figure 4.4: Acceptance probability (solid lines) and wall clock (dots) scaling with number of [PITH_FULL_IMAGE:figures/full_fig_p017_4_4.png]
Figure 4.5
Figure 4.5. Figure 4.5: Inferred GPR means for satellite data. on a 50 × 50 grid, following [46]. Then we empirically average over the hyperparameter uncertainty using our samples and plot the resulting function in [PITH_FULL_IMAGE:figures/full_fig_p018_4_5.png]
Figure 4.6
Figure 4.6. Figure 4.6: Inferred GPR standard deviations for satellite data. [PITH_FULL_IMAGE:figures/full_fig_p019_4_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 51 canonical work pages

  1. [54]

    effectively independent

    Homer F. Walker and Peng Ni. Anderson acceleration for fixed-point iterations.SIAM Journal on Numerical Analysis, 49(4):1715–1735, January 2011. A Pole expansion for the inverse matrix square root Following [31], we note that, for an analytic matrix functionf(A), a generalized form of the Cauchy integral formula allows us to write f(A) = A 2πi Z Γ z−1f(z)...

    work page 2011

  2. [1]

    Hogg, and Michael O’Neil

    Sivaram Ambikasaran, Daniel Foreman-Mackey, Leslie Greengard, David W. Hogg, and Michael O’Neil. Fast direct methods for Gaussian processes.IEEE Transactions on Pattern Analysis and Machine Intelligence, 38(2):252–265, February 2016

    work page 2016

  3. [2]

    Anderson

    Donald G. Anderson. Iterative procedures for nonlinear integral equations.J. ACM, 12(4):547– 560, October 1965

    work page 1965

  4. [3]

    Uniform approximation of common Gaussian process kernels using equispaced Fourier grids.Applied and Computational Harmonic Analysis, 71:101640, July 2024

    Alex Barnett, Philip Greengard, and Manas Rachh. Uniform approximation of common Gaussian process kernels using equispaced Fourier grids.Applied and Computational Harmonic Analysis, 71:101640, July 2024

    work page 2024

  5. [4]

    A conceptual introduction to Hamiltonian Monte Carlo

    Michael Betancourt. A conceptual introduction to Hamiltonian Monte Carlo. (arXiv:1701.02434), July 2018. arXiv:1701.02434 [stat]

    arXiv 2018

  6. [5]

    Boerner, Stephen Deems, Thomas R

    Timothy J. Boerner, Stephen Deems, Thomas R. Furlani, Shelley L. Knuth, and John Towns. ACCESS: Advancing innovation: NSF’s advanced cyberinfrastructure coordination ecosystem: Services & support. InPractice and Experience in Advanced Research Computing 2023: Computing for the Common Good, PEARC ’23, pages 173–176, New York, NY, USA, September

    work page 2023

  7. [6]

    A randomized algorithm for approximating the log determinant of a symmetric positive definite matrix.Linear Algebra and its Applications, 533:95–117, November 2017

    Christos Boutsidis, Petros Drineas, Prabhanjan Kambadur, Eugenia-Maria Kontopoulou, and Anastasios Zouzias. A randomized algorithm for approximating the log determinant of a symmetric positive definite matrix.Linear Algebra and its Applications, 533:95–117, November 2017

    work page 2017

  8. [7]

    Kernel operations on the GPU, with Autodiff, without memory overflows.Journal of Machine Learning Research, 22(74):1–6, 2021

    BenjaminCharlier, JeanFeydy, JoanAlexisGlaunès, François-DavidCollin, andGhislainDurif. Kernel operations on the GPU, with Autodiff, without memory overflows.Journal of Machine Learning Research, 22(74):1–6, 2021

    work page 2021

Show all 55 references
  1. [8]

    Non-stationary anderson acceleration with optimized damping.Journal of Computational and Applied Mathematics, 451:116077, 2024

    Kewang Chen and Cornelis Vuik. Non-stationary anderson acceleration with optimized damping.Journal of Computational and Applied Mathematics, 451:116077, 2024

    2024

  2. [9]

    Stochastic gradient Hamiltonian Monte Carlo

    Tianqi Chen, Emily Fox, and Carlos Guestrin. Stochastic gradient Hamiltonian Monte Carlo. In Eric P. Xing and Tony Jebara, editors,Proceedings of the 31st International Conference on Machine Learning, volume32(2)ofProceedings of Machine Learning Research, pages1683–1691, Bejin...

    2014

  3. [10]

    Epperly, Joel A

    Yifan Chen, Ethan N. Epperly, Joel A. Tropp, and Robert J. Webber. Randomly pivoted cholesky: Practical approximation of a kernel matrix with few entry evaluations. Communications on Pure and Applied Mathematics, 2023

    2023

  4. [11]

    Forthcoming, November 2024

    Sinho Chewi.Log-Concave Sampling. Forthcoming, November 2024. 20

    2024

  5. [12]

    M. A. Clark and A. D. Kennedy. Accelerating dynamical-fermion computations using the rational hybrid Monte Carlo algorithm with multiple pseudofermion fields.Physical Review Letters, 98(5), January 2007

    2007

  6. [13]

    On randomized trace estimates for indefinite matrices with an application to determinants.Foundations of Computational Mathematics, 22(3):875– 903, June 2022

    Alice Cortinovis and Daniel Kressner. On randomized trace estimates for indefinite matrices with an application to determinants.Foundations of Computational Mathematics, 22(3):875– 903, June 2022

    2022

  7. [14]

    S. L. Cotter, G. O. Roberts, A. M. Stuart, and D. White. MCMC methods for functions: Modifying old algorithms to make them faster.Statistical Science, 28(3):424–446, August 2013

    2013

  8. [15]

    Mission CO2ntrol: A statistical scientist’s role in remote sensing of atmospheric carbon dioxide.Journal of the American Statistical Association, 113(521):152–168, January 2018

    Noel Cressie. Mission CO2ntrol: A statistical scientist’s role in remote sensing of atmospheric carbon dioxide.Journal of the American Statistical Association, 113(521):152–168, January 2018

    2018

  9. [16]

    Scalable log determinants for Gaussian process kernel learning

    Kun Dong, David Eriksson, Hannes Nickisch, David Bindel, and Andrew Gordon Wilson. Scalable log determinants for Gaussian process kernel learning. InProceedings of the 31st International Conference on Neural Information Processing Systems, NIPS’17, pages 6330– 6340, Red Hook, ...

    2017

  10. [17]

    Simon Duane, A. D. Kennedy, Brian J. Pendleton, and Duncan Roweth. Hybrid Monte Carlo. Physics Letters B, 195(2):216–222, September 1987

    1987

  11. [18]

    A determinant-free method to simulate the parameters of large Gaussian fields.Stat, 6(1):271–281, 2017

    Louis Ellam, Heiko Strathmann, Mark Girolami, and Iain Murray. A determinant-free method to simulate the parameters of large Gaussian fields.Stat, 6(1):271–281, 2017

    2017

  12. [19]

    Rebholz, and Mengying Xiao

    Claire Evans, Sara Pollock, Leo G. Rebholz, and Mengying Xiao. A proof that anderson acceleration improves the convergence rate in linearly converging fixed-point methods (but not in those converging quadratically).SIAM Journal on Numerical Analysis, 58(1):788–810, January 2020

    2020

  13. [20]

    Paul Laiu, and Thomas Strohmer

    Xue Feng, M. Paul Laiu, and Thomas Strohmer. Convergence analysis of the alternating Anderson-Picard method for nonlinear fixed-point problems, 2024

    2024

  14. [21]

    Paul Laiu, and Thomas Strohmer

    Xue Feng, M. Paul Laiu, and Thomas Strohmer. Convergence analysis of the alternating anderson-picard method for nonlinear fixed-point problems. (arXiv:2407.10472), July 2024. arXiv:2407.10472

    2024 arXiv

  15. [22]

    Hogg, Dustin Lang, and Jonathan Goodman

    Daniel Foreman-Mackey, David W. Hogg, Dustin Lang, and Jonathan Goodman. emcee: The MCMC hammer.Publications of the Astronomical Society of the Pacific, 125(925):306–312, March 2013. arXiv:1202.3665 [astro-ph, physics:physics, stat]

    2013 arXiv

  16. [23]

    Column and row subset selection using nuclear scores: algorithms and theory for Nyström approximation, CUR decomposition, and graph Laplacian reduction, 2024

    Mark Fornace and Michael Lindsey. Column and row subset selection using nuclear scores: algorithms and theory for Nyström approximation, CUR decomposition, and graph Laplacian reduction, 2024

    2024

  17. [24]

    Tropp, and Madeleine Udell

    Zachary Frangella, Joel A. Tropp, and Madeleine Udell. Randomized Nyström preconditioning. SIAM Journal on Matrix Analysis and Applications, 44(2):718–752, 2023

    2023

  18. [25]

    Fucito, E

    F. Fucito, E. Marinari, G. Parisi, and C. Rebbi. A proposal for Monte Carlo simulations of fermionic systems.Nuclear Physics B, 180(3):369–377, May 1981. 21

    1981

  19. [26]

    Lang.Quantum Chromodynamics on the Lattice: An Introductory Presentation, volume 788 ofLecture Notes in Physics

    Christof Gattringer and Christian B. Lang.Quantum Chromodynamics on the Lattice: An Introductory Presentation, volume 788 ofLecture Notes in Physics. Springer, Berlin, Heidelberg, 2010

    2010

  20. [27]

    Riemann manifold Langevin and Hamiltonian Monte Carlo methods.Journal of the Royal Statistical Society Series B: Statistical Methodology, 73(2):123– 214, March 2011

    Mark Girolami and Ben Calderhead. Riemann manifold Langevin and Hamiltonian Monte Carlo methods.Journal of the Royal Statistical Society Series B: Statistical Methodology, 73(2):123– 214, March 2011

    2011

  21. [28]

    Ensemble samplers with affine invariance

    Jonathan Goodman and Jonathan Weare. Ensemble samplers with affine invariance. Communications in Applied Mathematics and Computational Science, 5(1):65–80, January 2010

    2010

  22. [29]

    Equispaced Fourier representations for efficient Gaussian process regression from a billion data points, May 2023

    Philip Greengard, Manas Rachh, and Alex Barnett. Equispaced Fourier representations for efficient Gaussian process regression from a billion data points, May 2023. arXiv:2210.10210 [cs, math, stat]

    2023 arXiv

  23. [30]

    Stuart, and Sebastian J

    Martin Hairer, Andrew M. Stuart, and Sebastian J. Vollmer. Spectral gaps for a Metropolis- Hastings algorithm in infinite dimensions.The Annals of Applied Probability, 24(6):2455–2490, December 2014

    2014

  24. [31]

    Higham, and Lloyd N

    Nicholas Hale, Nicholas J. Higham, and Lloyd N. Trefethen. ComputingAα,log(A), and related matrix functions by contour integrals.SIAM Journal on Numerical Analysis, 46(5):2505–2523, 2008

    2008

  25. [32]

    Hancock, Jeremy Fischer, John Michael Lowe, Winona Snapp-Childs, Marlon Pierce, SureshMarru, J.EricCoulter, MatthewVaughn, BrianBeck, NiravMerchant, EdwinSkidmore, and Gwen Jacobs

    David Y. Hancock, Jeremy Fischer, John Michael Lowe, Winona Snapp-Childs, Marlon Pierce, SureshMarru, J.EricCoulter, MatthewVaughn, BrianBeck, NiravMerchant, EdwinSkidmore, and Gwen Jacobs. Jetstream2: Accelerating cloud computing via Jetstream. InPractice and Experience in Ad...

    2021

  26. [33]

    Inference in deep Gaussian processes using stochastic gradient Hamiltonian Monte Carlo

    Marton Havasi, José Miguel Hernández-Lobato, and Juan José Murillo-Fuentes. Inference in deep Gaussian processes using stochastic gradient Hamiltonian Monte Carlo. In S. Bengio, H. Wallach, H. Larochelle, K. Grauman, N. Cesa-Bianchi, and R. Garnett, editors,Advances in Neural ...

    2018

  27. [34]

    Non-stationary Gaussian process regression with Hamiltonian Monte Carlo

    Markus Heinonen, Henrik Mannerström, Juho Rousu, Samuel Kaski, and Harri Lähdesmäki. Non-stationary Gaussian process regression with Hamiltonian Monte Carlo. In Arthur Gretton and Christian C. Robert, editors,Proceedings of the 19th International Conference on Artificial Intel...

    2016

  28. [35]

    A unified view of some numerical methods for fractional diffusion

    Clemens Hofreither. A unified view of some numerical methods for fractional diffusion. Computers & Mathematics with Applications, 80(2):332–350, 2020. Numerical Methods for Scientific Computations and Advanced Applications II

    2020

  29. [36]

    Codeforagradient-basedanddeterminant-freeframeworkforfullybayesian gaussian process regression, November 2024

    P.MichaelKielstra. Codeforagradient-basedanddeterminant-freeframeworkforfullybayesian gaussian process regression, November 2024

    2024

  30. [37]

    Michael Kielstra and Michael Lindsey

    P. Michael Kielstra and Michael Lindsey. Gaussian process regression with log-linear scaling for common non-stationary kernels, 2024. 22

    2024

  31. [38]

    Knoll and D.E

    D.A. Knoll and D.E. Keyes. Jacobian-free newton-krylov methods: a survey of approaches and applications.Journal of Computational Physics, 193(2):357–397, 2004

    2004

  32. [39]

    Approximate inference for fully Bayesian Gaussian process regression

    Vidhi Lalchand and Carl Edward Rasmussen. Approximate inference for fully Bayesian Gaussian process regression. In Cheng Zhang, Francisco Ruiz, Thang Bui, Adji Bousso Dieng, and Dawen Liang, editors,Proceedings of The 2nd Symposium on Advances in Approximate Bayesian Inference...

    2020

  33. [40]

    Teckentrup, and Simon Urbainczyk

    Jonas Latz, Aretha L. Teckentrup, and Simon Urbainczyk. Deep Gaussian process priors for Bayesian image reconstruction, 2024

    2024

  34. [41]

    Albergo, and Michael Lindsey

    Peter Lunts, Michael S. Albergo, and Michael Lindsey. Non-Hertz-Millis scaling of the antiferromagnetic quantum critical metal via scalable hybrid Monte Carlo.Nature Communications, 14(1):2547, May 2023

    2023

  35. [42]

    Neal.MCMC Using Hamiltonian Dynamics

    Radford M. Neal.MCMC Using Hamiltonian Dynamics. Chapman and Hall/CRC, 2011

    2011

  36. [43]

    Noack, Harinarayan Krishnan, Mark D

    Marcus M. Noack, Harinarayan Krishnan, Mark D. Risser, and Kristofer G. Reyes. Exact Gaussian processes for massive datasets via non-stationary sparsity-discovering kernels. Scientific Reports, 13(1):3155, March 2023

    2023

  37. [44]

    Noack, Hengrui Luo, and Mark D

    Marcus M. Noack, Hengrui Luo, and Mark D. Risser. A unifying perspective on non-stationary kernels for deeper Gaussian processes.APL Machine Learning, 2(1):010902, 02 2024

    2024

  38. [45]

    Geoff Pleiss, Martin Jankowiak, David Eriksson, Anil Damle, and Jacob R. Gardner. Fast matrix square roots with applications to Gaussian processes and Bayesian optimization. In Proceedings of the 34th International Conference on Neural Information Processing Systems, NIPS ’20,...

    2020

  39. [46]

    Carl Edward Rasmussen and Christopher K. I. Williams.Gaussian processes for machine learning. Adaptive computation and machine learning. MIT Press, Cambridge, Mass, 2006

    2006

  40. [47]

    Sparse Gaussian processes revisited: Bayesian approaches to inducing-variable approximations

    Simone Rossi, Markus Heinonen, Edwin Bonilla, Zheyang Shen, and Maurizio Filippone. Sparse Gaussian processes revisited: Bayesian approaches to inducing-variable approximations. In Arindam Banerjee and Kenji Fukumizu, editors,Proceedings of The 24th International Conference on...

    2021

  41. [48]

    J. M. Sanz-Serna. Runge-Kutta schemes for Hamiltonian systems.BIT, 28(4):877–883, December 1988

    1988

  42. [49]

    R. T. Scalettar, D. J. Scalapino, R. L. Sugar, and D. Toussaint. Hybrid molecular- dynamics algorithm for the numerical simulation of many-electron systems.Physical Review B, 36(16):8632–8641, December 1987

    1987

  43. [50]

    OCO-2 Level 2 bias-corrected XCO2 and other select fields from the full-physics retrieval aggregated as daily files, retrospective processing v11.2r, 2017

    Jet Propulsion Laboratory Science Computing Facility. OCO-2 Level 2 bias-corrected XCO2 and other select fields from the full-physics retrieval aggregated as daily files, retrospective processing v11.2r, 2017

    2017

  44. [51]

    Stan modeling language users guide and reference manual, 2.35, 2024

    Stan Development Team. Stan modeling language users guide and reference manual, 2.35, 2024. 23

    2024

  45. [52]

    Alex Toth and C. T. Kelley. Convergence analysis for anderson acceleration.SIAM Journal on Numerical Analysis, 53(2):805–819, January 2015

    2015

  46. [53]

    Oliphant, Matt Haberland, Tyler Reddy, David Cournapeau, Evgeni Burovski, Pearu Peterson, Warren Weckesser, Jonathan Bright, Stéfan J

    Pauli Virtanen, Ralf Gommers, Travis E. Oliphant, Matt Haberland, Tyler Reddy, David Cournapeau, Evgeni Burovski, Pearu Peterson, Warren Weckesser, Jonathan Bright, Stéfan J. van der Walt, Matthew Brett, Joshua Wilson, K. Jarrod Millman, Nikolay Mayorov, Andrew R. J. Nelson, E...

    2020

  47. [2023]

    Association for Computing Machinery

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.