Pith. sign in

REVIEW 3 major objections 7 minor 79 references

An Eulerian approach to regularized JKO scheme with low-rank tensor decompositions for Bayesian inversion

T0 review · 3 major / 7 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read This paper claims that entropy-regularized JKO proximal steps, discretized on a fixed grid and compressed as tensor trains, produce a tractable Eulerian surrogate for high-dimensional Bayesian posteriors whose sampling quality, at equal…

desk verdict A genuine algorithmic advance for structured Bayesian posteriors, but the high-dimensional tractability claim is only tested on targets that are low-rank by construction. read the letter →

arxiv 2411.12430 v1 pith:YNBM5TXW submitted 2024-11-19 math.NA cs.NAmath.OC

classification math.NAcs.NAmath.OC MSC 49Q2265K1062F15
keywords WassersteingradientflowJKOschemeentropicregularizationtensortrainlow-rankapproximationBayesianinverseproblemssamplingfixed-pointiteration
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

Bayesian inverse problems are usually sampled with Markov-chain methods that burn in, correlate, and can reuse almost none of the expensive posterior evaluations. This paper proposes instead to treat the posterior as the stationary point of a Wasserstein gradient flow and to march toward it with entropy-regularized Jordan-Kinderlehrer-Otto (JKO) proximal steps, discretized once on a fixed grid. Because the grid is fixed, posterior evaluations can be cached; because the densities and potentials are stored in the tensor-train (TT) format, the exponential grid growth is compressed away when the target has low-rank structure. The paper shows that, with the same number of unique posterior evaluations as a Metropolis-Hastings chain, the TT model produces samples whose Sinkhorn distance to a reference is comparable or smaller across synthetic and PDE-constrained inverse problems (Tables 2 and 3). If this holds, expensive forward-model calls stop being wasted on burn-in and correlated draws, and a reusable model of the posterior becomes available for marginals and importance sampling.

What carries the argument

The mechanism is the entropy-regularized JKO proximal step in its dynamic form, converted by the Hopf-Cole change of variables $\Phi = 2\beta_k \log \eta$, $\rho = \eta\hat\eta$ into two linear heat equations $\partial_t \hat\eta = \beta_k \Delta \hat\eta$ and $\partial_t \eta = -\beta_k \Delta \eta$ coupled at the terminal time by $\tilde\eta = (\rho_\infty/\hat\eta)^{1/(1+2\beta_k)}$. The paper treats this system as a fixed-point equation $\eta = G(\eta)$, solves each heat application exactly in TT format via Kronecker matrix exponentials, and reconstructs the nonlinear terminal coupling with TT-cross, so the only calls to the unnormalized posterior happen during cross approximation on the fixed grid. This object makes the method gradient-free, cacheable, and, under low-rank assumptions, subexponential in the dimension.

What would settle it

Construct a 30-dimensional target with provably high TT-rank, e.g. a random mixture of many well-separated Gaussian components, and measure the maximal rank and the number of unique posterior evaluations TT-cross needs to reach a fixed Sinkhorn distance to reference samples; if the rank grows with the number of components and the unique-evaluation count approaches or exceeds the Metropolis-Hastings chain length needed for the same accuracy, the central tractability claim fails.

Watch

Extended reading notes

Core claim

The central claim is that a fully Eulerian, gradient-free approximation of a high-dimensional posterior is tractable when each regularized JKO proximal step is reformulated as a fixed point of two heat equations. With the Hopf-Cole transform, the optimality conditions for the proximal step become a forward heat equation for one entropic potential and a backward heat equation for another, coupled through the terminal condition by the explicit formula $\tilde\eta_m = (\rho_\infty / \hat\eta_m)^{1/(1+2\beta_k)}$. The paper solves these equations on a tensor grid with finite differences in tensor-train format, applies TT-cross to build the coupled terms from few pointwise posterior evaluations, and accelerates the fixed-point iteration with Anderson acceleration. The resulting TT density yields a deterministic transport ODE, optionally with a final stochastic correction, for sampling, and the same model gives normalized marginals and an importance distribution without extra posterior calls.

Load-bearing premise

The load-bearing premise is that the posterior and the intermediate entropic potentials are close to low-rank tensor trains, so TT-cross can reconstruct them from a few pointwise evaluations; if that fails, storage and evaluation cost blow up and the Eulerian grid becomes exponential again.

Editorial extensions

If this is right

  • A single JKO step with large $T$ and small $\beta$ can already give a good posterior model (KL $\sim 10^{-6}$ in the Gaussian verification), so multi-step flow may often be unnecessary.
  • Fixed-grid caching makes posterior evaluations reusable across fixed-point iterations and outer steps, multiplying the effective budget by a factor between 1.5 and 7.3 in the test problems.
  • From the fitted TT model one obtains the normalized density, marginals, credible intervals, and samples, all without additional posterior calls.
  • The model can be re-purposed as an importance distribution for a quantity of interest, reducing estimator variance without new posterior calls, as demonstrated for the parabolic inverse problem.
  • The method is gradient-free, so it applies when the posterior is only available through a black-box unnormalized density, such as measurements tied to a PDE or ODE forward solver.

Reading between the lines

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

  • The same fixed-point machinery could be applied to any energy whose first variation is computable in TT format, turning the method into a general-purpose constrained sampling tool for fairness, safety, or interpretability constraints rather than a KL-specific one.
  • Because the Eulerian model is a discretized function, one could build a joint density $\rho(x,y)$ over parameters and measurements and condition on new data by tensor contractions, letting a single fitting amortize many measurement campaigns.
  • The observed superlinear convergence of Anderson acceleration suggests that the relaxation parameter, currently chosen heuristically, might be tuned automatically, removing one user-dependent robustness knob.
  • If the low-rank hypothesis is confirmed across a wider class of posteriors, the resource metric for comparing samplers may shift from chain length to unique posterior evaluations, which better reflects expensive forward-model calls.
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

3 major / 7 minor

Summary. The paper proposes an Eulerian, entropy-regularized JKO scheme for approximating and sampling from Bayesian posteriors that are available only through unnormalized pointwise evaluations. The proximal step is reformulated via a Hopf-Cole transform as a pair of coupled heat equations, which is then solved as a fixed-point problem in the Tensor Train format using TT-cross approximation, matrix-exponential heat solves, and Anderson acceleration. The authors demonstrate the method on a Gaussian verification problem, several synthetic targets (Gaussian mixture, double-moon, nonconvex potential), and two PDE-based Bayesian inverse problems, comparing sample quality against Metropolis-Hastings MCMC with equal numbers of posterior evaluations. They also show that the fitted TT model can be reused to construct an importance distribution for a quantity of interest. The central claim is that low-rank TT compression makes the Eulerian approach tractable in dimensions up to d=30 and that the resulting samples are comparable or better than MCMC with the same resource budget.

Significance. If the central claims are fully supported, this would be a valuable contribution: it offers a gradient-free, deterministic alternative to MCMC with reusable posterior evaluations, access to marginals and normalized densities, and a natural way to reuse the fitted model for related tasks such as importance sampling. The derivation in Section 3 is clear and the algorithmic pipeline is well motivated. The authors are also transparent about the main hypothesis, and the release of code at the stated repository is a strength. However, the significance is currently limited by the fact that all numerical demonstrations use targets that are low-rank by construction, and the convergence properties of the fixed-point iteration are only empirical. The headline claim about overcoming the curse of dimensionality therefore remains conditional on an unverified structural assumption.

major comments (3)
  1. [1.3, 4.3, Tables 2–3] The central tractability claim—that the TT format overcomes the curse of dimensionality—is load-bearing but is only tested on targets that are low-rank by construction. In Section 1.3 the low-rank assumption is explicitly a hypothesis, and Theorem 4.1 (Eq. (4.3.1)) provides only a conditional error bound depending on the existence of a rank-r approximation and on conditioning constants κ. The Gaussian mixture is a sum of five product Gaussians (rank ≤ 5), the double-moon and nonconvex potentials have very low algebraic separability, and both PDE posteriors are reported with rmax = 1 (Table 3). No experiment studies a target whose TT rank grows with dimension or grid resolution, and the ranks (or truncation errors) of the intermediate potentials η_m and \hatη_m during the fixed-point iteration are never reported. To support the headline claim, the paper should add at least one experiment with a genuinely non-product target (e.g., a correlated Gaussian with a dense covariance, or a rotated/elliptical distribution) and should report the evolution of TT ranks and approximation errors of η_m and \hatη_m, together with memory and runtime scaling in d and N.
  2. [3.2, 3.3, 4.4] The fixed-point iteration is the engine of each proximal step, but its convergence is only empirical. The operator G in (3.2.5)–(3.2.8) is derived from the optimality conditions; no theorem states that G is a contraction or that the Picard/Anderson iterates converge. The only theoretical result, inequality (3.3.1) in Section 3.3, concerns the outer proximal sequence under a subgradient bound Ξ_k, and the paper explicitly hypothesizes that Ξ_k can be controlled by (T_k, β_k). The convergence evidence is a single Gaussian verification (Figure 3a). Because divergence of the fixed-point iteration would invalidate every reported experiment, the paper should either prove convergence under explicit conditions on T, β, and the target, or provide convergence diagnostics for all test cases, including the effect of the truncation rank rmax and the Anderson relaxation parameter q.
  3. [5.1, Tables 2–3] The headline comparison against Metropolis-Hastings is made in terms of the number of unique posterior evaluations, but the TT method's total posterior calls (including cache hits) are substantially larger and its internal cost (TT-cross, SVD, Anderson updates, ODE/SDE integration) is not counted. If the abstract's 'same amount of resources' means only posterior evaluations, this should be stated explicitly; otherwise wall-clock time or a comparable cost model should be reported. As written, the claim is stronger than the measured quantity.
minor comments (7)
  1. [3.1, Eq. (3.1.8)] The displayed initial condition for \hatη is garbled; it should read \hatη(0,x)=ρ_k(x)/η(0,x), as used in (3.2.6). Please fix the typo.
  2. [2.1, Corollary 2.2.1] The notation W_2^2 and W_2 is used interchangeably for the squared and unsquared 2-Wasserstein distance; please standardize the notation.
  3. [4.3, Theorem 4.1] The theorem states a bound 'for sufficiently small ε>0', but ε is the best-approximation error; the quantifier should be clarified, and the dependence on κ should be discussed since κ is never estimated.
  4. [5.2, Table 1] The convergence criterion and the definition of 'not converged' (the '-' entries) are not given; please specify the maximum number of iterations and the stopping tolerance.
  5. [4.5] The choice of ϵ_SDE is heuristic (0.01–0.001); a brief sensitivity study or a rule for selecting it would strengthen the sampling section.
  6. [6.1] The importance-sampling experiment reports only the spread of estimates; please provide a quantitative comparison of variance or mean squared error and state how the bias is controlled.
  7. [1.4, 5.3] The related methods in [32] and [34] are discussed but never compared numerically; adding such a comparison would clarify the practical advantage of the proposed approach.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the regularized JKO fixed-point cycle and TT discretization are self-contained, and the low-rank condition is an acknowledged tractability hypothesis rather than a fitted prediction.

full rationale

No significant circularity found. The core proximal-step construction is derived from the first-order optimality system (3.1.1)-(3.1.3) via the Hopf-Cole transform (3.1.4)-(3.1.9), and the fixed-point update \tilde\eta_m=(\rho_\infty/\hat\eta_m)^{1/(1+2\beta_k)} in Eq. (3.2.9) is the explicit first variation of the KL functional, not a parameter fitted to the target output. The approximate posterior \rho_{TT}=\eta_M\hat\eta_M is the output of the fixed-point solve, and the reported comparisons in Tables 2 and 3 use external reference samples (exact Gaussian-mixture draws or long-run emcee chains), so the sample-quality claims are not forced by construction. TT-cross is used as a black-box reconstruction (Section 4.3) with an error bound (Theorem 4.1) cited from external work [54]; the low-rank assumption is explicitly labelled a hypothesis in Section 1.3 ('We hypothesize that if the posterior has some additional structure...'), which is a tractability assumption rather than a circular reduction. The paper also honestly flags the incompleteness of the convergence analysis in Section 3.3 ('We have to admit that the standard analysis presented above is rather incomplete...' and 'We hypothesize that the subgradient norm \Xi_k can be controlled...'), again as a limitation of the theoretical support, not as a self-referential derivation. Self-citations ([7], [62], [65]) appear only as contextual references for comparison methodology, fixed-point acceleration practice, and Langevin post-processing; none of these carries the central derivation, which relies on external results [9,10,25,47,48,49,51,54]. Therefore the honest finding is no significant circularity, with score 0.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The central derivation is self-contained, but the numerical method rests on unproven practical assumptions: low-rank compressibility and empirical fixed-point convergence. The paper introduces no new physical entities or fitted physical constants; the free parameters are algorithmic hyperparameters.

free parameters (7)
  • T (outer JKO step) = Not reported for main experiments; verification scans 0.1 to 1e5
    Time horizon of each proximal step; chosen by hand and strongly affects accuracy.
  • beta (entropic regularization) = Not reported for main experiments; verification best at 1e-4
    Entropic regularization strength; controls bias and fixed-point convergence.
  • rmax (TT rank truncation) = 1 to 5 in Tables 2-3
    Truncation rank of the TT approximation; determines storage and accuracy.
  • N_n (grid nodes per dimension) = 30 in verification; not reported for main experiments
    Number of nodes per dimension in the Eulerian grid.
  • N_cache (cache size) = Not reported
    Number of stored posterior evaluations; affects the caching speedup.
  • q (Anderson relaxation) = 0.8 to 0.9
    Relaxation parameter for Anderson acceleration; must be tuned to avoid divergence.
  • eps_SDE (SDE fraction) = 0.001 to 0.01
    Heuristic fraction of the time interval integrated with SDE dynamics to avoid ODE artifacts.
assumptions (5)
  • domain assumption The posterior and intermediate entropic potentials have small tensor-train rank.
    Hypothesis stated in Section 1.3; the entire compression strategy depends on it without verification beyond the test problems.
  • domain assumption The fixed-point operator G defined by (3.2.5)-(3.2.9) converges under Picard or Anderson iteration for the chosen parameters.
    No proof is provided; Section 4.4 reports only empirical convergence in preliminary experiments.
  • domain assumption The finite-difference discretization on the truncated cube accurately approximates heat evolution on R^d.
    Continuous derivation uses the heat kernel on R^d (3.2.1)-(3.2.4), while numerics use a matrix exponential of a discrete Laplacian on a bounded grid (Section 4.2); the approximation error is not analyzed.
  • standard math Standard gradient-flow and optimal-transport theory (Ambrosio-Gigli-Savare) applies to the chosen functionals.
    Used in Section 2 for existence, uniqueness, and EVI inequalities; the authors themselves note the theory is incomplete for multimodal targets.
  • standard math The TT-cross approximation error bound (Theorem 4.1) holds with acceptable constants for the tensors in question.
    The method relies on TT-cross to reconstruct tensors from few evaluations; the condition number kappa and approximation error epsilon are not verified in practice.

how reviews work

0 comments
Cite this review

Pith. "Pith review of An Eulerian approach to regularized JKO scheme with low-rank tensor decompositions for Bayesian inversion." pith.science (2026). https://pith.science/paper/YNBM5TXW

@misc{pith2026241112430,
  author       = {Pith},
  title        = {Pith review of: An Eulerian approach to regularized JKO scheme with low-rank tensor decompositions for Bayesian inversion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YNBM5TXW}},
  note         = {Machine review of arXiv:2411.12430}
}
read the original abstract

The possibility of using the Eulerian discretization for the problem of modelling high-dimensional distributions and sampling, is studied. The problem is posed as a minimization problem over the space of probability measures with respect to the Wasserstein distance and solved with entropy-regularized JKO scheme. Each proximal step can be formulated as a fixed-point equation and solved with accelerated methods, such as Anderson's. The usage of low-rank Tensor Train format allows to overcome the \emph{curse of dimensionality}, i.e. the exponential growth of degrees of freedom with dimension, inherent to Eulerian approaches. The resulting method requires only pointwise computations of the unnormalized posterior and is, in particular, gradient-free. Fixed Eulerian grid allows to employ a caching strategy, significally reducing the expensive evaluations of the posterior. When the Eulerian model of the target distribution is fitted, the passage back to the Lagrangian perspective can also be made, allowing to approximately sample from it. We test our method both for synthetic target distributions and particular Bayesian inverse problems and report comparable or better performance than the baseline Metropolis-Hastings MCMC with same amount of resources. Finally, the fitted model can be modified to facilitate the solution of certain associated problems, which we demonstrate by fitting an importance distribution for a particular quantity of interest.

Figures

Figures reproduced from arXiv: 2411.12430 by the authors.

Figure 1
Figure 1. Schematic of proposed fixed-point algorithm. [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. Contour lines of the test distribution and the samples generated with the deterministic method (left) and the [PITH_FULL_IMAGE:figures/full_fig_p014_2.png] view at source ↗
Figure 3
Figure 3. Reconstruction for a Gaussian target distribution. [PITH_FULL_IMAGE:figures/full_fig_p017_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Contour lines of two-dimensional marginals of the test distributions shown with the reference sample, the [PITH_FULL_IMAGE:figures/full_fig_p019_4.png]
Figure 5
Figure 5. Figure 5: Posterior distribution and reconstruction of the initial solution for the hyperbolic problem. [PITH_FULL_IMAGE:figures/full_fig_p022_5.png]
Figure 6
Figure 6. Figure 6: Posterior distribution and reconstruction of the initial solution for the parabolic problem. [PITH_FULL_IMAGE:figures/full_fig_p022_6.png]
Figure 7
Figure 7. Figure 7: Distribution of the estimate of Eρ∞F via the posterior mean (blue) and the importance estimator (6.1.3) along with the true mean (vertical blue line). can see that the estimates without importance sampling (blue) are quite spread out. The estimates via the importance d…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

79 extracted references · 64 canonical work pages

  1. [34]

    Tensor train based sampling algorithms for approximating regularized Wasserstein proximal operators

    Fuqun Han, Stanley Osher, and Wuchen Li. Tensor train based sampling algorithms for approximating regularized Wasserstein proximal operators. arXiv preprint arXiv:2401.13125, 2024

  2. [32]

    Solution of the Fokker–Planck equation by cross approximation method in the tensor train format

    Andrei Chertkov and Ivan Oseledets. Solution of the Fokker–Planck equation by cross approximation method in the tensor train format. Frontiers in Artificial Intelligence, 4:668215, 2021

  3. [1]

    Planck 2018 results-V

    Nabila Aghanim, Yashar Akrami, Mark Ashdown, J Aumont, Carlo Baccigalupi, M Ballardini, Anthony J Banday, RB Barreiro, N Bartolo, S Basak, et al. Planck 2018 results-V. CMB power spectra and likelihoods. Astronomy & Astrophysics, 641:A5, 2020

  4. [2]

    Population-based deep image prior for dynamic PET denoising: A data-driven approach to improve parametric quantification

    Qiong Liu, Yu-Jung Tsai, Jean-Dominique Gallezot, Xueqi Guo, Ming-Kai Chen, Darko Pucar, Colin Young, Vladimir Panin, Michael Casey, Tianshun Miao, Huidong Xie, Xiongchao Chen, Bo Zhou, Richard Carson, and Chi Liu. Population-based deep image prior for dynamic PET denoising: A data-driven approach to improve parametric quantification. Medical Image Analys...

  5. [3]

    Roberts and Richard L

    Gareth O. Roberts and Richard L. Tweedie. Exponential convergence of Langevin distributions and their discrete approximations. Bernoulli, 2(4):341 – 363, 1996

  6. [4]

    A single series from the Gibbs sampler provides a false sense of security

    Andrew Gelman and Donald B Rubin. A single series from the Gibbs sampler provides a false sense of security. Bayesian statistics, 4(1):625–631, 1992

  7. [5]

    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, 2010

  8. [6]

    Affine invariant interacting Langevin dynamics for Bayesian inference

    Alfredo Garbuno-Inigo, Nikolas Nüsken, and Sebastian Reich. Affine invariant interacting Langevin dynamics for Bayesian inference. SIAM Journal on Applied Dynamical Systems, 19(3):1633–1658, 2020

Show all 79 references
  1. [7]

    Less interaction with forward models in Langevin dynamics

    Martin Eigel, Robert Gruhlke, and David Sommer. Less interaction with forward models in Langevin dynamics. arXiv preprint arXiv:2212.11528, 2022

  2. [8]

    Computation of exact gradients in distributed dynamic systems

    Yuri Evtushenko. Computation of exact gradients in distributed dynamic systems. Optimization Methods and Software, 9(1-3):45–75, 1998

  3. [9]

    The variational formulation of the Fokker–Planck equation

    Richard Jordan, David Kinderlehrer, and Felix Otto. The variational formulation of the Fokker–Planck equation. SIAM journal on mathematical analysis, 29(1):1–17, 1998

  4. [10]

    Gradient flows: in metric spaces and in the space of probability measures

    Luigi Ambrosio, Nicola Gigli, and Giuseppe Savaré. Gradient flows: in metric spaces and in the space of probability measures. Springer Science & Business Media, 2005. 24 A PREPRINT - NOVEMBER 20, 2024

  5. [11]

    The geometry of dissipative evolution equations: the porous medium equation

    Felix Otto. The geometry of dissipative evolution equations: the porous medium equation. 2001

  6. [12]

    The Wasserstein gradient flow of the fisher information and the quantum drift-diffusion equation

    Ugo Gianazza, Giuseppe Savaré, and Giuseppe Toscani. The Wasserstein gradient flow of the fisher information and the quantum drift-diffusion equation. Archive for rational mechanics and analysis, 194(1):133–220, 2009

  7. [13]

    A gradient flow approach to the Keller-Segel systems (progress in variational problems : Variational problems interacting with probability theories)

    Adrien Blanchet. A gradient flow approach to the Keller-Segel systems (progress in variational problems : Variational problems interacting with probability theories). 2013

  8. [14]

    Lagrangian discretization of crowd motion and linear diffusion

    Hugo Leclerc, Quentin M’erigot, Filippo Santambrogio, and Federico Stra. Lagrangian discretization of crowd motion and linear diffusion. SIAM J. Numer. Anal., 58:2093–2118, 2019

  9. [15]

    An augmented lagrangian approach to Wasserstein gradient flows and applications

    Jean-David Benamou, Guillaume Carlier, and Maxime Laborde. An augmented lagrangian approach to Wasserstein gradient flows and applications. ESAIM: Proceedings and surveys, 54:1–17, 2016

  10. [16]

    Primal dual methods for Wasserstein gradient flows

    Jose A Carrillo, Katy Craig, Li Wang, and Chaozhen Wei. Primal dual methods for Wasserstein gradient flows. Foundations of Computational Mathematics, pages 1–55, 2022

  11. [17]

    Convergence of entropic schemes for optimal transport and gradient flows

    Guillaume Carlier, Vincent Duval, Gabriel Peyré, and Bernhard Schmitzer. Convergence of entropic schemes for optimal transport and gradient flows. SIAM Journal on Mathematical Analysis, 49(2):1385–1418, 2017

  12. [18]

    Entropic approximation of Wasserstein gradient flows

    Gabriel Peyré. Entropic approximation of Wasserstein gradient flows. SIAM Journal on Imaging Sciences , 8(4):2323–2351, 2015

  13. [19]

    Carrillo, and Jingwei Hu

    Rafael Bailo, José A. Carrillo, and Jingwei Hu. Fully Discrete Positivity-Preserving and Energy-Dissipating Schemes for Aggregation-Diffusion Equations with a Gradient-Flow Structure. Communications in Mathematical Sciences, 18(5):1259–1303, 2020

  14. [20]

    A variational finite volume scheme for Wasserstein gradient flows

    Clément Cances, Thomas O Gallouët, and Gabriele Todeschi. A variational finite volume scheme for Wasserstein gradient flows. Numerische Mathematik, 146:437–480, 2020

  15. [21]

    Stein variational gradient descent: A general purpose Bayesian inference algorithm

    Qiang Liu and Dilin Wang. Stein variational gradient descent: A general purpose Bayesian inference algorithm. Advances in neural information processing systems, 29, 2016

  16. [22]

    Stein variational gradient descent as gradient flow

    Qiang Liu. Stein variational gradient descent as gradient flow. Advances in neural information processing systems, 30, 2017

  17. [23]

    Stein variational gradient descent without gradient

    Jun Han and Qiang Liu. Stein variational gradient descent without gradient. In International Conference on Machine Learning, pages 1900–1908. PMLR, 2018

  18. [24]

    Sampling in unit time with kernel Fisher-Rao flow

    Aimee Maurais and Youssef Marzouk. Sampling in unit time with kernel Fisher-Rao flow. arXiv preprint arXiv:2401.03892, 2024

  19. [25]

    TT-cross approximation for multidimensional arrays

    Ivan Oseledets and Eugene Tyrtyshnikov. TT-cross approximation for multidimensional arrays. Linear Algebra and its Applications, 432(1):70–88, 2010

  20. [26]

    The alternating linear scheme for tensor optimiza- tion in the tensor train format

    Sebastian Holtz, Thorsten Rohwedder, and Reinhold Schneider. The alternating linear scheme for tensor optimiza- tion in the tensor train format. SIAM Journal on Scientific Computing, 34(2):A683–A713, 2012

  21. [27]

    Tensor-train density estimation

    Georgii S Novikov, Maxim E Panov, and Ivan V Oseledets. Tensor-train density estimation. InUncertainty in artificial intelligence, pages 1321–1331. PMLR, 2021

  22. [28]

    High-dimensional density estimation with tensorizing flow

    Yinuo Ren, Hongli Zhao, Yuehaw Khoo, and Lexing Ying. High-dimensional density estimation with tensorizing flow. Research in the Mathematical Sciences, 10(3):30, 2023

  23. [29]

    Deep composition of tensor-trains using squared inverse Rosenblatt transports

    Tiangang Cui and Sergey Dolgov. Deep composition of tensor-trains using squared inverse Rosenblatt transports. Foundations of Computational Mathematics, 22(6):1863–1922, 2022

  24. [30]

    Low-rank tensor methods for partial differential equations

    Markus Bachmayr. Low-rank tensor methods for partial differential equations. Acta Numerica, 32:1–121, 2023

  25. [31]

    Solving high-dimensional parabolic PDEs using the tensor train format

    Lorenz Richter, Leon Sallandt, and Nikolas Nüsken. Solving high-dimensional parabolic PDEs using the tensor train format. In International Conference on Machine Learning, pages 8998–9009. PMLR, 2021

  26. [33]

    Inverse problems: a bayesian perspective

    Andrew M Stuart. Inverse problems: a bayesian perspective. Acta numerica, 19:451–559, 2010

  27. [35]

    Gradient structures and geodesic convexity for reaction–diffusion sys- tems

    Matthias Liero and Alexander Mielke. Gradient structures and geodesic convexity for reaction–diffusion sys- tems. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , 371(2005):20120346, 2013

  28. [36]

    Nonconvex gradient flow in the Wasserstein metric and applications to constrained nonlocal interactions

    Katy Craig. Nonconvex gradient flow in the Wasserstein metric and applications to constrained nonlocal interactions. Proceedings of the London Mathematical Society, 114(1):60–102, 2017. 25 A PREPRINT - NOVEMBER 20, 2024

  29. [37]

    On parameter estimation with the Wasserstein distance

    Espen Bernton, Pierre E Jacob, Mathieu Gerber, and Christian P Robert. On parameter estimation with the Wasserstein distance. Information and Inference: A Journal of the IMA, 8(4):657–676, 10 2019

  30. [38]

    A survey on optimal transport for machine learning: Theory and applications

    Luis Caicedo Torres, Luiz Manella Pereira, and M Hadi Amini. A survey on optimal transport for machine learning: Theory and applications. arXiv preprint arXiv:2106.01963, 2021

  31. [39]

    A survey of optimal transport for computer graphics and computer vision

    Nicolas Bonneel and Julie Digne. A survey of optimal transport for computer graphics and computer vision. In Computer Graphics Forum, volume 42, pages 439–460. Wiley Online Library, 2023

  32. [40]

    Stability of flows associated to gradient vector fields and convergence of iterated transport maps

    Luigi Ambrosio, Stefano Lisini, and Giuseppe Savaré. Stability of flows associated to gradient vector fields and convergence of iterated transport maps. manuscripta mathematica, 121(1):1–50, 2006

  33. [41]

    A computational fluid mechanics solution to the Monge-Kantorovich mass transfer problem

    Jean-David Benamou and Yann Brenier. A computational fluid mechanics solution to the Monge-Kantorovich mass transfer problem. Numerische Mathematik, 84(3):375–393, 2000

  34. [42]

    Quadratically regularized optimal transport

    Dirk A Lorenz, Paul Manns, and Christian Meyer. Quadratically regularized optimal transport. Applied Mathe- matics & Optimization, 83(3):1919–1949, 2021

  35. [43]

    Interpolating between optimal transport and KL regularized optimal transport using Rényi divergences

    Jonas Bresch and Viktor Stein. Interpolating between optimal transport and KL regularized optimal transport using Rényi divergences. arXiv preprint arXiv:2404.18834, 2024

  36. [44]

    Entropic optimal transport: Convergence of potentials

    Marcel Nutz and Johannes Wiesel. Entropic optimal transport: Convergence of potentials. Probability Theory and Related Fields, 184(1):401–424, 2022

  37. [45]

    Computational optimal transport: Complexity by accelerated gradient descent is better than by Sinkhorn’s algorithm

    Pavel Dvurechensky, Alexander Gasnikov, and Alexey Kroshnin. Computational optimal transport: Complexity by accelerated gradient descent is better than by Sinkhorn’s algorithm. In International conference on machine learning, pages 1367–1376. PMLR, 2018

  38. [46]

    Sinkhorn distances: Lightspeed computation of optimal transport

    Marco Cuturi. Sinkhorn distances: Lightspeed computation of optimal transport. Advances in neural information processing systems, 26, 2013

  39. [47]

    Fisher information regularization schemes for Wasserstein gradient flows

    Wuchen Li, Jianfeng Lu, and Li Wang. Fisher information regularization schemes for Wasserstein gradient flows. Journal of Computational Physics, 416:109449, 2020

  40. [48]

    A kernel formula for regularized Wasserstein proximal operators

    Wuchen Li, Siting Liu, and Stanley Osher. A kernel formula for regularized Wasserstein proximal operators. arXiv preprint arXiv:2301.10301, 2023

  41. [49]

    On the relation between optimal transport and schrödinger bridges: A stochastic control viewpoint

    Yongxin Chen, Tryphon T Georgiou, and Michele Pavon. On the relation between optimal transport and schrödinger bridges: A stochastic control viewpoint. Journal of Optimization Theory and Applications, 169:671– 691, 2016

  42. [50]

    Entropic and displacement interpolation: a computational approach using the hilbert metric

    Yongxin Chen, Tryphon Georgiou, and Michele Pavon. Entropic and displacement interpolation: a computational approach using the hilbert metric. SIAM Journal on Applied Mathematics, 76(6):2375–2396, 2016

  43. [51]

    Convergence of flow-based generative models via proximal gradient descent in Wasserstein space

    Xiuyuan Cheng, Jianfeng Lu, Yixin Tan, and Yao Xie. Convergence of flow-based generative models via proximal gradient descent in Wasserstein space. arXiv preprint arXiv:2310.17582, 2023

  44. [52]

    Kronecker products and matrix calculus with applications

    Alexander Graham. Kronecker products and matrix calculus with applications. Courier Dover Publications, 2018

  45. [53]

    Al-Mohy and Nicholas J

    Awad H. Al-Mohy and Nicholas J. Higham. A new scaling and squaring algorithm for the matrix exponential. SIAM Journal on Matrix Analysis and Applications, 31(3):970–989, 2010

  46. [54]

    Error analysis of tensor-train cross approximation

    Zhen Qin, Alexander Lidiak, Zhexuan Gong, Gongguo Tang, Michael B Wakin, and Zhihui Zhu. Error analysis of tensor-train cross approximation. In S. Koyejo, S. Mohamed, A. Agarwal, D. Belgrave, K. Cho, and A. Oh, editors, Advances in Neural Information Processing Systems, volume...

  47. [55]

    Parallel cross interpolation for high-precision calculation of high- dimensional integrals

    Sergey Dolgov and Dmitry Savostyanov. Parallel cross interpolation for high-precision calculation of high- dimensional integrals. Computer Physics Communications, 246:106869, 2020

  48. [56]

    Fast adaptive interpolation of multi-dimensional arrays in tensor train format

    Dmitry Savostyanov and Ivan Oseledets. Fast adaptive interpolation of multi-dimensional arrays in tensor train format. In The 2011 International Workshop on Multidimensional (nD) Systems, pages 1–8. IEEE, 2011

  49. [57]

    Fast solvers for unsteady thermal fluid structure interaction

    Philipp Birken, Tobias Gleim, Detlef Kuhl, and Andreas Meister. Fast solvers for unsteady thermal fluid structure interaction. International Journal for Numerical Methods in Fluids, 79(1):16–29, 2015

  50. [58]

    Exact optimal accelerated complexity for fixed-point iterations

    Jisun Park and Ernest K Ryu. Exact optimal accelerated complexity for fixed-point iterations. In International Conference on Machine Learning, pages 17420–17457. PMLR, 2022

  51. [59]

    Convergence analysis for Anderson acceleration

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

  52. [60]

    Two classes of multisecant methods for nonlinear acceleration

    Haw-ren Fang and Yousef Saad. Two classes of multisecant methods for nonlinear acceleration. Numerical linear algebra with applications, 16(3):197–221, 2009. 26 A PREPRINT - NOVEMBER 20, 2024

  53. [61]

    Anderson acceleration for fixed-point iterations

    Homer F Walker and Peng Ni. Anderson acceleration for fixed-point iterations. SIAM Journal on Numerical Analysis, 49(4):1715–1735, 2011

  54. [62]

    Application of accelerated fixed-point algorithms to hydrodynamic well-fracture coupling

    Vitalii Aksenov, Maxim Chertov, and Konstantin Sinkov. Application of accelerated fixed-point algorithms to hydrodynamic well-fracture coupling. Computers and Geotechnics, 129:103783, 2021

  55. [63]

    Henderson, and Ravi Varadhan

    Bohao Tang, Nicholas C. Henderson, and Ravi Varadhan. Accelerating fixed-point algorithms in statistics and data science: A state-of-art review. Journal of Data Science, 21(1):1–26, 2022

  56. [64]

    Score-based generative modeling through stochastic differential equations

    Yang Song, Jascha Sohl-Dickstein, Diederik P Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations. arXiv preprint arXiv:2011.13456, 2020

  57. [65]

    Generative modelling with tensor train approximations of Hamilton–Jacobi–Bellman equations

    David Sommer, Robert Gruhlke, Max Kirstein, Martin Eigel, and Claudia Schillings. Generative modelling with tensor train approximations of Hamilton–Jacobi–Bellman equations. arXiv preprint arXiv:2402.15285, 2024

  58. [66]

    Black box approximation in the tensor train format initialized by ANOV A decomposition.SIAM Journal on Scientific Computing, 45(4):A2101–A2118, 2023

    Andrei Chertkov, Gleb Ryzhakov, and Ivan Oseledets. Black box approximation in the tensor train format initialized by ANOV A decomposition.SIAM Journal on Scientific Computing, 45(4):A2101–A2118, 2023

  59. [67]

    Faster Wasserstein distance estimation with the Sinkhorn divergence

    Lenaic Chizat, Pierre Roussillon, Flavien Léger, François-Xavier Vialard, and Gabriel Peyré. Faster Wasserstein distance estimation with the Sinkhorn divergence. Advances in Neural Information Processing Systems, 33:2257– 2269, 2020

  60. [68]

    Stochastic optimization for large-scale optimal transport

    Aude Genevay, Marco Cuturi, Gabriel Peyré, and Francis Bach. Stochastic optimization for large-scale optimal transport. Advances in neural information processing systems, 29, 2016

  61. [69]

    Interpolating between optimal transport and MMD using Sinkhorn divergences

    Jean Feydy, Thibault Séjourné, François-Xavier Vialard, Shun-ichi Amari, Alain Trouve, and Gabriel Peyré. Interpolating between optimal transport and MMD using Sinkhorn divergences. In The 22nd International Conference on Artificial Intelligence and Statistics, pages 2681–2690, 2019

  62. [70]

    Strong equivalence between metrics of Wasserstein type

    Erhan Bayraktar and Gaoyue Guo. Strong equivalence between metrics of Wasserstein type. 2021

  63. [71]

    Alaya, Aurélie Boisbunon, Stanislas Chambon, Laetitia Chapel, Adrien Corenflos, Kilian Fatras, Nemo Fournier, Léo Gautheron, Nathalie T.H

    Rémi Flamary, Nicolas Courty, Alexandre Gramfort, Mokhtar Z. Alaya, Aurélie Boisbunon, Stanislas Chambon, Laetitia Chapel, Adrien Corenflos, Kilian Fatras, Nemo Fournier, Léo Gautheron, Nathalie T.H. Gayraud, Hicham Janati, Alain Rakotomamonjy, Ievgen Redko, Antoine Rolet, Ant...

  64. [72]

    emcee: the MCMC hammer

    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, 2013

  65. [73]

    Numerical methods for Bayesian inverse problems

    Robert Scheichl and Jakob Zech. Numerical methods for Bayesian inverse problems. 2021

  66. [74]

    Inverse determination of boundary conditions and sources in steady heat conduction with heat generation

    TJ Martin and GS Dulikravich. Inverse determination of boundary conditions and sources in steady heat conduction with heat generation. 1996

  67. [75]

    Inverse determination of temperatures and heat fluxes on inaccessible surfaces

    TJ Martin and GS Dulikravich. Inverse determination of temperatures and heat fluxes on inaccessible surfaces. WIT Transactions on Modelling and Simulation, 8, 2024

  68. [76]

    Space marching difference schemes in the nonlinear inverse heat conduction problem

    AS Carasso. Space marching difference schemes in the nonlinear inverse heat conduction problem. Inverse Problems, 8(1):25, 1992

  69. [77]

    Constructive representation of functions in low-rank tensor formats.Constructive Approximation, 37:1–18, 2013

    Ivan V Oseledets. Constructive representation of functions in low-rank tensor formats.Constructive Approximation, 37:1–18, 2013

  70. [78]

    Conditional wasserstein distances with applications in bayesian ot flow matching

    Jannis Chemseddine, Paul Hagemann, Christian Wald, and Gabriele Steidl. Conditional wasserstein distances with applications in bayesian ot flow matching. arXiv preprint arXiv:2403.18705, 2024

  71. [79]

    Sampling with trusthworthy constraints: A variational gradient framework

    Xingchao Liu, Xin Tong, and Qiang Liu. Sampling with trusthworthy constraints: A variational gradient framework. Advances in Neural Information Processing Systems, 34:23557–23568, 2021. 27

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.