Pith. sign in

REVIEW 1 major objections 6 minor 2 cited by

High-order Taylor surrogates for implicit maps become tractable by representing each derivative tensor as a Tucker tensor train fit from cheap random probes.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · grok-4.5

2026-07-13 21:00 UTC pith:ZP6MPAY7

load-bearing objection Solid methods paper that makes high-order local Taylor surrogates for covariance-preconditioned implicit maps tractable via Tucker-TT + symmetric probes; theory and probe experiments are the real strength. the 1 major comments →

arxiv 2603.21141 v2 pith:ZP6MPAY7 submitted 2026-03-22 math.NA cs.NA

Tucker Tensor Train Taylor Series

Nick Alger , Blake Christierson , Peng Chen , Omar Ghattas This is my paper
classification math.NA cs.NA MSC 65N2115A6965K1065F30
keywords Tucker tensor trainTaylor series surrogatederivative probesRiemannian optimizationimplicit mapsPDE-constrained mapsrank continuationcovariance preconditioning
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

High-order Taylor expansions of maps that depend on the solution of a nonlinear system (for example a PDE) have long been regarded as unusable in high dimensions: the derivative tensors are enormous and can be touched only by probing. This paper shows that those tensors can be replaced by compact Tucker tensor trains, yielding a local surrogate called the Tucker Tensor Train Taylor Series (T4S). The surrogate is trained not from many input-output pairs, but from random directionally symmetric probes of the derivatives at a single expansion point; each such probe needs only a handful of linear solves that share the same operator. Under spectral decay of the input covariance the paper proves that moderate ranks already give controllable approximation error, and the numerical algorithms (derivative-informed dimension reduction, Riemannian Gauss-Newton or Cauchy SGD with rank continuation, and fast sweeping Jacobians) recover nearly optimal accuracy from probes alone. The result matters because outer-loop tasks such as inverse problems, optimal design, and uncertainty quantification repeatedly need both the map and its derivatives; a derivative-accurate local surrogate can replace the expensive implicit solve inside those loops.

Core claim

A truncated Taylor series whose derivative tensors are each replaced by a Tucker tensor train (T4S) is a computationally tractable local surrogate for a covariance-preconditioned, implicitly defined map. The trains can be fitted from random directionally symmetric probes that cost far less than function evaluations or asymmetric probes, and spectral decay of the covariance supplies explicit rank-error bounds that guarantee the representation exists with moderate ranks.

What carries the argument

The Tucker Tensor Train Taylor Series (T4S): each Fréchet derivative tensor is written as a Tucker decomposition whose central core is itself a tensor train, fitted on the Riemannian manifold of fixed-rank trains by trust-region Gauss-Newton or Cauchy-step SGD with rank continuation, using fast sweeping routines for the Riemannian Jacobian.

Load-bearing premise

The method works only when the input covariance spectrum decays fast enough (or the derivative tensors themselves are low-rank) so that the required Tucker and tensor-train ranks stay moderate; otherwise storage and fitting cost explode, and the local Taylor expansion is valid only near the chosen expansion point.

What would settle it

On a family of random preconditioned tensors or Poisson problems whose covariance eigenvalues decay only as i^{-1} or slower, measure whether the relative forward error of the fitted T4S continues to drop with increasing rank, or whether the ranks needed already exceed the data budget before the error reaches the T3-SVD baseline.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Outer-loop algorithms that need many evaluations of an implicit map and its derivatives can replace each nonlinear solve by a cheap T4S evaluation once the trains are built at a single point.
  • Training cost scales linearly with derivative order rather than exponentially, because only directionally symmetric probes are required.
  • Rank-continuation with edge-condition balancing and Cauchy step sizes removes most hyper-parameter tuning from the fitting stage.
  • The same representational guarantees apply to any map whose derivatives are preconditioned by a Hilbert-Schmidt operator with decaying spectrum, not only PDE maps.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • A mixture of several T4S expansions centered at different points could extend the local surrogate into a piecewise-global model without changing the core fitting machinery.
  • The same probe-and-fit pipeline could be used to compress high-order derivatives that appear inside Newton or Gauss-Newton outer loops themselves, turning each outer iteration into a low-rank linear algebra step.
  • If the covariance spectrum is only moderately decaying, hybrid bases that combine the leading eigenmodes of C with a few active-subspace directions may keep ranks practical.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

1 major / 6 minor

Summary. The paper constructs local high-order Taylor surrogates for covariance-preconditioned maps that depend implicitly on the solution of a nonlinear state equation (e.g., a PDE). Each derivative tensor D^j f(0) is represented as a Tucker tensor train (T4S), fit from random directionally symmetric forward/reverse probes at a single expansion point after a derivative-informed dimension reduction. The authors supply Riemannian Gauss–Newton (TR-RMGN) and Cauchy-step SGD (MC-SGD) algorithms with edge-condition rank continuation, fast sweeping methods for the Riemannian Jacobian and its adjoint, and representational error bounds (Theorem 8, Corollary 9) that depend on the spectral decay of C and the induced norm of D^k q. Numerical experiments show that the fitting procedures match quasi-optimal T3-SVD accuracy from probes alone up to data-limited ranks on random preconditioned tensors, and that T4S recovers high-order Taylor structure on two Poisson PDE examples.

Significance. High-order Taylor surrogates for high-dimensional implicit maps have long been regarded as intractable because the derivative tensors are enormous and accessible only through probes. The combination of directionally symmetric probing (O(mk) shared-operator linearized solves), Tucker-tensor-train compression, and derivative-informed sketching makes such surrogates practical under spectral decay of C. The representational theory (peeling argument, symmetry-to-Tucker reduction, hyperbolic-cross eigenvalue sums) is carefully developed, the algorithms are specified at the level of gauged tangent vectors and sweeping contractions, and the random-tensor experiments provide an independent T3-SVD baseline. If the claims hold under the stated hypotheses, the work supplies a concrete, derivative-accurate alternative to global operator learning for outer-loop problems that only need local accuracy near a design or prior mean.

major comments (1)
  1. The central claim is supported under the paper’s own hypotheses (spectral decay of C or additional low-rank structure in D^j q; local Taylor validity). No load-bearing internal inconsistency was found in the peeling argument (Proposition 2, Lemmas 3–5), the infinite-dimensional reduction (Lemma 7, Theorem 8), or the probe-cost analysis (Table 1, §3.5). The experiments match T3-SVD from probes alone (Figs. 11–14) and recover high-order Taylor structure on the Poisson examples (Figs. 16–22). I therefore raise no major technical objections that would require a rewrite of the core contribution.
minor comments (6)
  1. §1.1 and the abstract correctly flag locality and spectral-decay requirements; a short forward pointer in the abstract to the precise hypotheses of Theorem 8 / Corollary 9 would help readers who stop at the abstract.
  2. Figure 2 and the surrounding discussion of graphical tensor notation are clear, but a one-sentence reminder that the output mode is the last index would reduce momentary confusion when reading the T3 definition (Definition 4).
  3. In §4.4.1 the “useless rank removal” three-phase sweep is described only in prose; a short algorithmic box or pseudocode would make the procedure easier to re-implement.
  4. The MC-SGD stopping criterion (§4.3.2) uses fixed constants C_τ=1, C_t=3 and |B|=⌊n_s/10⌋. A brief sensitivity remark (or a single additional panel) would strengthen the claim of “little hyperparameter tuning.”
  5. Typographical: “dimen-sion” hyphenation artifact appears in a figure caption in §7.2.1; “co-vector” in Fig. 21 is fine but could be “covector” for consistency with the rest of the text.
  6. Related-work placement: the connection to [65] and [15,16] is noted, but a sentence contrasting T4S training data (symmetric probes at one point) with those works’ moment/correlation constructions would clarify novelty for readers coming from the stochastic-PDE literature.

Circularity Check

0 steps flagged

No significant circularity: representational bounds derive from spectral decay of C and operator norms via peeling/Tucker arguments; experiments benchmark against independent T3-SVD and true Taylor series.

full rationale

The central representational claim (Theorem 8 / Corollary 9) is obtained by a self-contained peeling argument (Proposition 2 + Lemmas 3–5) that constructs TT cores from eigenvalue decay of Kronecker products of C, followed by a symmetry-to-Tucker reduction (Lemma 5) and infinite-dimensional reduction (Lemma 7). These steps use only the induced norm of D^k q(θ0) and the spectrum of C; they do not invoke the fitting procedure or any target accuracy that is later “predicted.” Random-tensor experiments compare probe-based TR-RMGN/MC-SGD fits against an independent dense T3-SVD baseline (quasi-optimal in Frobenius norm). PDE experiments compare T4S output to the true (unreduced) Taylor series and to dimension-reduced Taylor series, not to a quantity defined by the fit itself. Self-citations to the authors’ prior TT-probing work [7] supply algorithmic background for derivative probes and are not used as uniqueness theorems or load-bearing premises that force the present claims. Rank-continuation and Riemannian optimization are standard manifold techniques applied to a least-squares loss on probes; no parameter is fitted to a subset of data and then re-labeled a prediction of a closely related quantity. The paper’s own stated limitations (spectral decay of C, locality of Taylor expansion) are hypotheses of the theorems, not hidden circular assumptions. Consequently the derivation chain is independent of its inputs and the circularity score is minimal.

Axiom & Free-Parameter Ledger

5 free parameters · 6 axioms · 3 invented entities

The central claim rests on standard multilinear algebra and Fréchet calculus, domain assumptions about uniquely solvable smooth state equations with computable partials, spectral decay of the covariance (or low-rank structure of derivatives), and several algorithmic hyperparameters for ranks and optimization. The T4S factorization and the specific Riemannian fitting pipeline are the main invented constructs; they are algorithmic rather than physical entities and are tested against T3-SVD and true Taylor series.

free parameters (5)
  • Taylor order k = typically 3–5 in experiments
    User-chosen truncation order; paper recommends modest k=3 or 4 from experience; controls both accuracy and cost.
  • Tucker and TT ranks (n, r) via rank continuation = τ=10, n_chunk=1 (defaults)
    Adapted by edge condition numbers with τ and n_chunk; selected by validation error and data-to-manifold-dimension ratio.
  • Dimension-reduction tolerance ε and stagnation p = ε ∈ {0.25,0.05,0.01}, p=5
    Controls reduced dimensions N,M in Algorithms 1–2; directly limits achievable Taylor error (Fig. 18).
  • Training sample count n_s and MC-SGD batch/smoothing (C_τ, C_t, |B|) = |B|=⌊n_s/10⌋, C_τ=1, C_t=3
    Determines data-limited ranks and stopping; chosen by user; experiments vary n_s from 100 to 6400.
  • Expansion point θ_0 and operator C (or local Gaussian approx.) = mean/cov of N or local logistic linearization
    For non-Gaussian parameters (Example 2) θ_0 and C are taken from a local normal approximation; quality of the local model is a modeling choice.
axioms (6)
  • domain assumption State equation R(θ,u)=0 is uniquely solvable near θ_0; Q and R are smooth; directional partials of R and Q are computable.
    Required for Fréchet derivatives of q and for the forward/reverse probing lattices in §3.
  • domain assumption C is Hilbert–Schmidt, self-adjoint, positive semidefinite; eigenvalues of Kronecker products of C control TT ranks (Theorem 8).
    Load-bearing for representational guarantees and for the power-law Corollary 9.
  • standard math Standard multilinear algebra: induced norms, matricizations/unfoldings, Kronecker identities (Lemma 1), TT/Tucker geometry.
    Used throughout §§2,5,6 and Appendices A–B.
  • standard math Fixed-rank nondegenerate T3 tensors form an embedded manifold; gauged variations and doubled-rank retractions via T3-SVD are valid.
    Extends known TT-manifold results; Appendix A.4; enables TR-RMGN and MC-SGD.
  • standard math Directionally symmetric probes determine the full multilinear derivative by polarization/symmetry.
    Justifies training only on symmetric probes (§3.5).
  • domain assumption Local Taylor expansion is an adequate surrogate near the expansion point for the intended outer-loop use.
    Stated limitation in §1.1; not proved globally.
invented entities (3)
  • T4S model (Tucker tensor train Taylor series) independent evidence
    purpose: Local surrogate storing each derivative tensor as a T3 factorization inside a truncated Taylor series.
    Core proposed object (Definition 1); algorithmic construct with independent checks via T3-SVD and true Taylor comparisons.
  • Derivative-informed shared input/output sketching (Algorithms 1–2) independent evidence
    purpose: Coarse dimension reduction from symmetric forward/reverse probes before T3 fitting.
    Paper-specific procedure; validated by reduced-Taylor error KDEs.
  • TR-RMGN and MC-SGD with edge-condition rank continuation for T3 independent evidence
    purpose: Fit fixed-rank T3 from probes with little hyperparameter tuning and adaptive ranks.
    New algorithmic combination for this problem; compared to T3-SVD baseline.

pith-pipeline@v1.1.0-grok45 · 60383 in / 3842 out tokens · 44719 ms · 2026-07-13T21:00:31.198933+00:00 · methodology

0 comments
read the original abstract

Learning derivative-accurate surrogates for implicit simulators is a key challenge in scientific machine learning. High-order Taylor surrogates have long been considered intractable in high dimensions, because the derivative tensors are enormous and accessible only through probes. We make such surrogates tractable with the Tucker tensor train Taylor series (T4S), a local surrogate that represents each derivative tensor of a truncated Taylor expansion as a Tucker tensor train. T4S targets a different learning problem than global operator learning: rather than training from input-output pairs at many parameter values, it is trained from random directionally symmetric derivative probes at a single expansion point. Computing $m$ probes of the $k$th derivative requires only $O(mk)$ linearized solves sharing one operator, cheaper than the $O(m)$ nonlinear solves for function evaluations or $O(m\,2^k)$ linearized solves for asymmetric probes. We develop derivative-informed dimension reduction, Riemannian Gauss-Newton and Cauchy SGD fitting algorithms with rank continuation, requiring little hyperparameter tuning, and fast sweeping routines for the Riemannian Jacobian. We prove representational guarantees under spectral decay of the input covariance. Experiments show that our methods match quasi-optimal T3-SVD accuracy on random tensors from probes alone, up to data-limited ranks, and recover high-order Taylor structure in Poisson PDE examples.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

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

  1. Universal Approximation of Nonlinear Operators and Their Derivatives

    cs.LG 2026-05 unverdicted novelty 8.0

    Proves the first universal approximation theorems for k-times differentiable nonlinear operators between Banach spaces and their derivatives uniformly on compact sets in weighted Sobolev norms via encoder-decoder oper...

  2. Universal Approximation of Nonlinear Operators and Their Derivatives

    cs.LG 2026-05 unverdicted novelty 7.0

    Proves first UATs for k-times differentiable nonlinear operators and their derivatives via OL architectures uniformly on compact sets in weighted Bastiani-Sobolev spaces on general Banach spaces.

Reference graph

Works this paper leans on

119 extracted references · 9 linked inside Pith · cited by 1 Pith paper

  1. [1]

    Trust-region methods on Riemannian manifolds.Foundations of Computational Mathematics, 7(3):303–330, 2007

    P-A Absil, Christopher G Baker, and Kyle A Gallivan. Trust-region methods on Riemannian manifolds.Foundations of Computational Mathematics, 7(3):303–330, 2007

    2007

  2. [2]

    Princeton University Press, 2008

    P-A Absil, Robert Mahony, and Rodolphe Sepulchre.Optimization algorithms on matrix manifolds. Princeton University Press, 2008

    2008

  3. [3]

    SIAM, 2022

    Ben Adcock, Simone Brugiapaglia, and Clayton G Webster.Sparse polynomial approximation of high-dimensional functions, volume 25. SIAM, 2022. 93

    2022

  4. [4]

    Randomized algorithms for rounding in the tensor-train format.SIAM Journal on Scientific Computing, 45(1):A74–A95, 2023

    Hussam Al Daas, Grey Ballard, Paul Cazeaux, Eric Hallman, Ag- nieszka Miedlar, Mirjeta Pasha, Tim W Reid, and Arvind K Saibaba. Randomized algorithms for rounding in the tensor-train format.SIAM Journal on Scientific Computing, 45(1):A74–A95, 2023

    2023

  5. [5]

    Optimal de- sign of large-scale nonlinear Bayesian inverse problems under model uncertainty.Inverse Problems, 40(9):095001, 2024

    Alen Alexanderian, Ruanui Nicholson, and Noemi Petra. Optimal de- sign of large-scale nonlinear Bayesian inverse problems under model uncertainty.Inverse Problems, 40(9):095001, 2024

    2024

  6. [6]

    Alen Alexanderian, Noemi Petra, Georg Stadler, and Omar Ghat- tas. Mean-variance risk-averse optimal control of systems governed by PDEs with random parameter fields using quadratic approxima- tions.SIAM/ASA Journal on Uncertainty Quantification, 5(1):1166– 1192, 2017

    2017

  7. [7]

    Tensor Train Construction From Tensor Actions, With Application to Compression of Large High Order Derivative Tensors.SIAM J

    Nick Alger, Peng Chen, and Omar Ghattas. Tensor Train Construction From Tensor Actions, With Application to Compression of Large High Order Derivative Tensors.SIAM J. Sci. Comput., 42(5):A3516–A3539, January 2020

    2020

  8. [8]

    Methods of applied mathematics

    Todd Arbogast and Jerry L Bona. Methods of applied mathematics. University of Texas at Austin, 2008

    2008

  9. [9]

    A stochastic collocation method for elliptic partial differential equations with random input data.SIAM Journal on Numerical Analysis, 45(3):1005–1034, 2007

    Ivo Babuˇ ska, Fabio Nobile, and Ra´ ul Tempone. A stochastic collocation method for elliptic partial differential equations with random input data.SIAM Journal on Numerical Analysis, 45(3):1005–1034, 2007

    2007

  10. [10]

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

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

    2023

  11. [11]

    Springer, 1989

    H Thomas Banks and Karl Kunisch.Estimation techniques for dis- tributed parameter systems. Springer, 1989

    1989

  12. [12]

    Gradient-based data and parameter dimension reduction for Bayesian models: an infor- mation theoretic perspective.arXiv preprint arXiv:2207.08670, 2022

    Ricardo Baptista, Youssef Marzouk, and Olivier Zahm. Gradient-based data and parameter dimension reduction for Bayesian models: an infor- mation theoretic perspective.arXiv preprint arXiv:2207.08670, 2022

    Pith/arXiv arXiv 2022

  13. [13]

    Hessian-based model reduction for large-scale systems with initial-condition inputs.International Journal for Numerical Methods in Engineering, 73(6):844–868, 2008

    Omar Bashir, K Willcox, O Ghattas, B van Bloemen Waanders, and J Hill. Hessian-based model reduction for large-scale systems with initial-condition inputs.International Journal for Numerical Methods in Engineering, 73(6):844–868, 2008. 94

    2008

  14. [14]

    Model reduction and neural networks for parametric PDEs.The SMAI Journal of computational mathematics, 7:121–157, 2021

    Kaushik Bhattacharya, Bamdad Hosseini, Nikola B Kovachki, and An- drew M Stuart. Model reduction and neural networks for parametric PDEs.The SMAI Journal of computational mathematics, 7:121–157, 2021

    2021

  15. [15]

    Bonizzoni, F

    F. Bonizzoni, F. Nobile, and D. Kressner. Tensor train approximation of moment equations for elliptic equations with lognormal coefficient. Computer Methods in Applied Mechanics and Engineering, 308:349– 376, 2016

    2016

  16. [16]

    Analysis and approximation of moment equations for PDEs with stochastic data

    Francesca Bonizzoni. Analysis and approximation of moment equations for PDEs with stochastic data. 2013

    2013

  17. [17]

    Regularity and sparse approxi- mation of the recursive first moment equations for the lognormal darcy problem.Computers & Mathematics with Applications, 80(12):2925– 2947, 2020

    Francesca Bonizzoni and Fabio Nobile. Regularity and sparse approxi- mation of the recursive first moment equations for the lognormal darcy problem.Computers & Mathematics with Applications, 80(12):2925– 2947, 2020

    2020

  18. [18]

    Cambridge University Press, 2023

    Nicolas Boumal.An introduction to optimization on smooth manifolds. Cambridge University Press, 2023

    2023

  19. [19]

    RTRMC: A Riemannian trust-region method for low-rank matrix completion.Advances in neu- ral information processing systems, 24, 2011

    Nicolas Boumal and Pierre-antoine Absil. RTRMC: A Riemannian trust-region method for low-rank matrix completion.Advances in neu- ral information processing systems, 24, 2011

    2011

  20. [20]

    A Riemannian trust region method for the canonical tensor rank approximation problem.SIAM Journal on Optimization, 28(3):2435–2465, 2018

    Paul Breiding and Nick Vannieuwenhoven. A Riemannian trust region method for the canonical tensor rank approximation problem.SIAM Journal on Optimization, 28(3):2435–2465, 2018

    2018

  21. [21]

    Greedy inference with structure-exploiting lazy maps.Advances in Neural Information Processing Systems, 33:8330– 8342, 2020

    Michael Brennan, Daniele Bigoni, Olivier Zahm, Alessio Spantini, and Youssef Marzouk. Greedy inference with structure-exploiting lazy maps.Advances in Neural Information Processing Systems, 33:8330– 8342, 2020

    2020

  22. [22]

    Hand-waving and in- terpretive dance: an introductory course on tensor networks.Journal of physics A: Mathematical and theoretical, 50(22):223001, 2017

    Jacob C Bridgeman and Christopher T Chubb. Hand-waving and in- terpretive dance: an introductory course on tensor networks.Journal of physics A: Mathematical and theoretical, 50(22):223001, 2017

    2017

  23. [23]

    Tan Bui-Thanh, Omar Ghattas, James Martin, and Georg Stadler. A computational framework for infinite-dimensional Bayesian inverse problems Part I: The linearized case, with application to global seismic 95 inversion.SIAM Journal on Scientific Computing, 35(6):A2494–A2523, 2013

    2013

  24. [24]

    Tensor com- pletion via tensor train based low-rank quotient geometry under a pre- conditioned metric.arXiv preprint arXiv:2209.04786, 2022

    Jian-Feng Cai, Wen Huang, Haifeng Wang, and Ke Wei. Tensor com- pletion via tensor train based low-rank quotient geometry under a pre- conditioned metric.arXiv preprint arXiv:2209.04786, 2022

    Pith/arXiv arXiv 2022

  25. [25]

    Derivative-informed neural operator acceleration of geometric MCMC for infinite-dimensional bayesian inverse problems.Journal of Machine Learning Research, 26(78):1–68, 2025

    Lianghao Cao, Thomas O’Leary-Roseberry, and Omar Ghattas. Derivative-informed neural operator acceleration of geometric MCMC for infinite-dimensional bayesian inverse problems.Journal of Machine Learning Research, 26(78):1–68, 2025

    2025

  26. [26]

    Residual-based error correction for neural operator accelerated infinite-dimensional Bayesian inverse problems

    Lianghao Cao, Thomas O’Leary-Roseberry, Prashant K Jha, J Tinsley Oden, and Omar Ghattas. Residual-based error correction for neural operator accelerated infinite-dimensional Bayesian inverse problems. Journal of Computational Physics, 486:112104, 2023

    2023

  27. [27]

    Jean C´ ea. Conception optimale ou identification de formes, cal- cul rapide de la d´ eriv´ ee directionnelle de la fonction coˆ ut.ESAIM: Mod´ elisation math´ ematique et analyse num´ erique, 20(3):371–402, 1986

    1986

  28. [28]

    Hessian-based sampling for high- dimensional model reduction.International Journal for Uncertainty Quantification, 9(2), 2019

    Peng Chen and Omar Ghattas. Hessian-based sampling for high- dimensional model reduction.International Journal for Uncertainty Quantification, 9(2), 2019

    2019

  29. [29]

    Peng Chen and Alfio Quarteroni. A new algorithm for high-dimensional uncertainty quantification based on dimension-adaptive sparse grid ap- proximation and reduced basis methods.Journal of Computational Physics, 298:176–193, 2015

    2015

  30. [30]

    Cou- pled input-output dimension reduction: Application to goal-oriented bayesian experimental design and global sensitivity analysis.arXiv preprint arXiv:2406.13425, 2024

    Qiao Chen, Elise Arnaud, Ricardo Baptista, and Olivier Zahm. Cou- pled input-output dimension reduction: Application to goal-oriented bayesian experimental design and global sensitivity analysis.arXiv preprint arXiv:2406.13425, 2024

    Pith/arXiv arXiv 2024

  31. [31]

    First order k-th moment finite element analysis of nonlinear operator equations with stochastic data

    Alexey Chernov and Christoph Schwab. First order k-th moment finite element analysis of nonlinear operator equations with stochastic data. Mathematics of Computation, 82(284):1859–1888, 2013

    2013

  32. [32]

    Andrzej Cichocki, Namgil Lee, Ivan Oseledets, Anh-Huy Phan, Qibin Zhao, Danilo P Mandic, et al. Tensor networks for dimensionality 96 reduction and large-scale optimization: Part 1 low-rank tensor decom- positions.Foundations and Trends®in Machine Learning, 9(4-5):249– 429, 2016

    2016

  33. [33]

    Approximation of high-dimensional parametric PDEs.Acta Numerica, 24:1–159, 2015

    Albert Cohen and Ronald DeVore. Approximation of high-dimensional parametric PDEs.Acta Numerica, 24:1–159, 2015

    2015

  34. [34]

    SIAM, 2015

    Paul G Constantine.Active subspaces: Emerging ideas for dimension reduction in parameter studies. SIAM, 2015

    2015

  35. [35]

    Dimension- independent likelihood-informed MCMC.Journal of Computational Physics, 304:109–137, 2016

    Tiangang Cui, Kody JH Law, and Youssef M Marzouk. Dimension- independent likelihood-informed MCMC.Journal of Computational Physics, 304:109–137, 2016

    2016

  36. [36]

    Likelihood-informed dimension reduction for nonlin- ear inverse problems.Inverse Problems, 30(11):114015, 2014

    Tiangang Cui, James Martin, Youssef M Marzouk, Antti Solonen, and Alessio Spantini. Likelihood-informed dimension reduction for nonlin- ear inverse problems.Inverse Problems, 30(11):114015, 2014

    2014

  37. [37]

    Mitigating the influence of the boundary on pde-based covariance operators.arXiv preprint arXiv:1610.05280, 2016

    Yair Daon and Georg Stadler. Mitigating the influence of the boundary on pde-based covariance operators.arXiv preprint arXiv:1610.05280, 2016

    Pith/arXiv arXiv 2016

  38. [38]

    A multi- linear singular value decomposition.SIAM journal on Matrix Analysis and Applications, 21(4):1253–1278, 2000

    Lieven De Lathauwer, Bart De Moor, and Joos Vandewalle. A multi- linear singular value decomposition.SIAM journal on Matrix Analysis and Applications, 21(4):1253–1278, 2000

    2000

  39. [39]

    Rank-adaptive ten- sor methods for high-dimensional nonlinear PDEs.Journal of Scientific Computing, 88(2):36, 2021

    Alec Dektor, Abram Rodgers, and Daniele Venturi. Rank-adaptive ten- sor methods for high-dimensional nonlinear PDEs.Journal of Scientific Computing, 88(2):36, 2021

    2021

  40. [40]

    Model-based geostatistics.Journal of the Royal Statistical Society Series C: Applied Statistics, 47(3), 1998

    Peter J Diggle, Jonathan A Tawn, and Rana A Moyeed. Model-based geostatistics.Journal of the Royal Statistical Society Series C: Applied Statistics, 47(3), 1998

    1998

  41. [41]

    Learning op- timal aerodynamic designs through multi-fidelity reduced-dimensional neural networks

    Xiaosong Du, Joaquim R Martins, Thomas O’Leary-Roseberry, Anir- ban Chaudhuri, Omar Ghattas, and Karen E Willcox. Learning op- timal aerodynamic designs through multi-fidelity reduced-dimensional neural networks. InAIAA SCITECH 2023 Forum, page 0334, 2023

    2023

  42. [42]

    Springer, 2016

    Dinh D˜ ung, Vladimir N Temlyakov, and Tino Ullrich.Hyperbolic cross approximation. Springer, 2016. 97

    2016

  43. [43]

    Finitely correlated states on quantum spin chains.Communications in mathe- matical physics, 144(3):443–490, 1992

    Mark Fannes, Bruno Nachtergaele, and Reinhard F Werner. Finitely correlated states on quantum spin chains.Communications in mathe- matical physics, 144(3):443–490, 1992

    1992

  44. [44]

    H Pearl Flath, Lucas C Wilcox, Volkan Ak¸ celik, Judith Hill, Bart van Bloemen Waanders, and Omar Ghattas. Fast algorithms for Bayesian uncertainty quantification in large-scale linear inverse problems based on low-rank partial Hessian approximations.SIAM Journal on Scien- tific Computing, 33(1):407–432, 2011

    2011

  45. [45]

    Dissertation, 2017

    Patrick Gelß.The Tensor-Train Format and Its Applications: Model- ing and Analysis of Chemical Reaction Networks, Catalytic Processes, Fluid Flows, and Brownian Dynamics. Dissertation, 2017

    2017

  46. [46]

    Shape derivative-informed neural oper- ators with application to risk-averse shape optimization.arXiv preprint arXiv:2603.03211, 2026

    Xindi Gong, Dingcheng Luo, Thomas O’Leary-Roseberry, Ruanui Nicholson, and Omar Ghattas. Shape derivative-informed neural oper- ators with application to risk-averse shape optimization.arXiv preprint arXiv:2603.03211, 2026

    arXiv 2026

  47. [47]

    Gradient-based optimiza- tion for regression in the functional tensor-train format.Journal of Computational Physics, 374:1219–1238, 2018

    Alex A Gorodetsky and John D Jakeman. Gradient-based optimiza- tion for regression in the functional tensor-train format.Journal of Computational Physics, 374:1219–1238, 2018

    2018

  48. [48]

    Reverse-mode differentiation in arbitrary tensor network format: with application to supervised learning.Journal of Machine Learning Research, 23(143):1– 29, 2022

    Alex A Gorodetsky, Cosmin Safta, and John D Jakeman. Reverse-mode differentiation in arbitrary tensor network format: with application to supervised learning.Journal of Machine Learning Research, 23(143):1– 29, 2022

    2022

  49. [49]

    Variants of alternating least squares tensor completion in the tensor train format

    Lars Grasedyck, Melanie Kluge, and Sebastian Kramer. Variants of alternating least squares tensor completion in the tensor train format. SIAM Journal on Scientific Computing, 37(5):A2424–A2450, 2015

    2015

  50. [50]

    Stable ALS approximation in the TT-format for rank-adaptive tensor completion.Numerische Mathematik, 143(4):855–904, 2019

    Lars Grasedyck and Sebastian Kr¨ amer. Stable ALS approximation in the TT-format for rank-adaptive tensor completion.Numerische Mathematik, 143(4):855–904, 2019

    2019

  51. [51]

    SIAM, 2002

    Max D Gunzburger.Perspectives in flow control and optimization. SIAM, 2002. 98

    2002

  52. [52]

    Geometry of matrix product states: Metric, parallel transport, and curvature.Journal of Mathematical Physics, 55(2), 2014

    Jutho Haegeman, Micha¨ el Mari¨ en, Tobias J Osborne, and Frank Ver- straete. Geometry of matrix product states: Metric, parallel transport, and curvature.Journal of Mathematical Physics, 55(2), 2014

    2014

  53. [53]

    Most tensor problems are NP-hard.Journal of the ACM (JACM), 60(6):1–39, 2013

    Christopher J Hillar and Lek-Heng Lim. Most tensor problems are NP-hard.Journal of the ACM (JACM), 60(6):1–39, 2013

    2013

  54. [54]

    The alternating linear scheme for tensor optimization in the tensor train for- mat.SIAM Journal on Scientific Computing, 34(2):A683–A713, 2012

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

    2012

  55. [55]

    On manifolds of tensors of fixed tt-rank.Numerische Mathematik, 120(4):701–731, 2012

    Sebastian Holtz, Thorsten Rohwedder, and Reinhold Schneider. On manifolds of tensors of fixed tt-rank.Numerische Mathematik, 120(4):701–731, 2012

    2012

  56. [56]

    Tobin Isaac, Noemi Petra, Georg Stadler, and Omar Ghattas. Scal- able and efficient algorithms for the propagation of uncertainty from data through inference to prediction for large-scale problems, with ap- plication to flow of the Antarctic ice sheet.Journal of Computational Physics, 296, September 2015

    2015

  57. [57]

    Tobin Isaac, Noemi Petra, Georg Stadler, and Omar Ghattas. Scal- able and efficient algorithms for the propagation of uncertainty from data through inference to prediction for large-scale problems, with ap- plication to flow of the Antarctic ice sheet.Journal of Computational Physics, 296:348–368, 2015

    2015

  58. [58]

    Scalable symmetric Tucker tensor decomposition.SIAM Journal on Matrix Analysis and Applications, 45(4):1746–1781, 2024

    Ruhui Jin, Joe Kileel, Tamara G Kolda, and Rachel Ward. Scalable symmetric Tucker tensor decomposition.SIAM Journal on Matrix Analysis and Applications, 45(4):1746–1781, 2024

    2024

  59. [59]

    Springer- Verlag New York, 2005

    Jari Kaipio and Erkki Somersalo.Statistical and Computational Inverse Problems, volume 160 ofApplied Mathematical Sciences. Springer- Verlag New York, 2005

    2005

  60. [60]

    Inexact trust-region algorithms on Riemannian manifolds.Advances in neural information processing systems, 31, 2018

    Hiroyuki Kasai and Bamdev Mishra. Inexact trust-region algorithms on Riemannian manifolds.Advances in neural information processing systems, 31, 2018

    2018

  61. [61]

    Walter de Gruyter GmbH & Co KG, 2018

    Boris N Khoromskij.Tensor numerical methods in scientific computing, volume 19. Walter de Gruyter GmbH & Co KG, 2018. 99

    2018

  62. [62]

    Effi- cient time-stepping scheme for dynamics on TT-manifolds

    Boris N Khoromskij, Ivan V Oseledets, and Reinhold Schneider. Effi- cient time-stepping scheme for dynamics on TT-manifolds. 2012

    2012

  63. [63]

    Adam: A method for stochastic optimization.arXiv preprint arXiv:1412.6980, 2014

    Diederik P Kingma and Jimmy Lei Ba. Adam: A method for stochastic optimization.arXiv preprint arXiv:1412.6980, 2014

    Pith/arXiv arXiv 2014

  64. [64]

    Kolda and Brett W

    Tamara G. Kolda and Brett W. Bader. Tensor Decompositions and Applications.SIAM Rev., 51(3):455–500, August 2009

    2009

  65. [65]

    Low-rank tensor approximation for high-order correlation functions of Gaussian random fields.SIAM/ASA Journal on Uncertainty Quantifi- cation, 3(1):393–416, 2015

    Daniel Kressner, Rajesh Kumar, Fabio Nobile, and Christine Tobler. Low-rank tensor approximation for high-order correlation functions of Gaussian random fields.SIAM/ASA Journal on Uncertainty Quantifi- cation, 3(1):393–416, 2015

    2015

  66. [66]

    Low- rank tensor completion by Riemannian optimization.BIT Numerical Mathematics, 54(2):447–468, 2014

    Daniel Kressner, Michael Steinlechner, and Bart Vandereycken. Low- rank tensor completion by Riemannian optimization.BIT Numerical Mathematics, 54(2):447–468, 2014

    2014

  67. [67]

    Sparse adaptive tensor Galerkin approximations of stochastic PDE-constrained control prob- lems.SIAM/ASA Journal on Uncertainty Quantification, 4(1):1034– 1059, 2016

    Angela Kunoth and Christoph Schwab. Sparse adaptive tensor Galerkin approximations of stochastic PDE-constrained control prob- lems.SIAM/ASA Journal on Uncertainty Quantification, 4(1):1034– 1059, 2016

    2016

  68. [68]

    Operator learning with PCA-Net: upper and lower complexity bounds.Journal of Machine Learning Research, 24(318):1– 67, 2023

    Samuel Lanthaler. Operator learning with PCA-Net: upper and lower complexity bounds.Journal of Machine Learning Research, 24(318):1– 67, 2023

    2023

  69. [69]

    Fourier neural operator for parametric partial differential equations

    Zongyi Li, Nikola Kovachki, Kamya Azizzadenesheli, Burigede Liu, Kaushik Bhattacharya, Andrew Stuart, and Anima Anandkumar. Fourier neural operator for parametric partial differential equations. International Conference on Learning Representations, 2021

    2021

  70. [70]

    Parameter and state model reduction for large-scale statistical inverse problems.SIAM Journal on Scientific Computing, 32(5):2523–2542, 2010

    Chad Lieberman, Karen Willcox, and Omar Ghattas. Parameter and state model reduction for large-scale statistical inverse problems.SIAM Journal on Scientific Computing, 32(5):2523–2542, 2010

    2010

  71. [71]

    Finn Lindgren, H˚ avard Rue, and Johan Lindstr¨ om. An explicit link be- tween Gaussian fields and Gaussian Markov random fields: the stochas- tic partial differential equation approach.Journal of the Royal Statis- tical Society Series B: Statistical Methodology, 73(4):423–498, 2011. 100

    2011

  72. [72]

    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, and George Em Zhang, Zhongqiang nd Karniadakis. Learning nonlinear operators via deeponet based on the universal approximation theorem of operators.Nature machine intelligence, 3(3):218–229, 2021

    2021

  73. [73]

    Gaus- sian mixture taylor approximations of risk measures constrained by PDEs with Gaussian random field inputs

    Dingcheng Luo, Joshua Chen, Peng Chen, and Omar Ghattas. Gaus- sian mixture taylor approximations of risk measures constrained by PDEs with Gaussian random field inputs. August 2024

    2024

  74. [74]

    Dimension reduction for derivative-informed operator learning: An analysis of approximation errors.arXiv preprint arXiv:2504.08730, 2025

    Dingcheng Luo, Thomas O’Leary-Roseberry, Peng Chen, and Omar Ghattas. Dimension reduction for derivative-informed operator learning: An analysis of approximation errors.arXiv preprint arXiv:2504.08730, 2025

    Pith/arXiv arXiv 2025

  75. [75]

    Efficient PDE-constrained optimization under high- dimensional uncertainty using derivative-informed neural operators

    Dingcheng Luo, Thomas O’Leary-Roseberry, Peng Chen, and Omar Ghattas. Efficient PDE-constrained optimization under high- dimensional uncertainty using derivative-informed neural operators. SIAM Journal on Scientific Computing, 47(4):C899–C931, 2025

    2025

  76. [76]

    Low-rank tensor estimation via rie- mannian gauss-newton: Statistical optimality and second-order con- vergence.Journal of Machine Learning Research, 24(381):1–48, 2023

    Yuetian Luo and Anru R Zhang. Low-rank tensor estimation via rie- mannian gauss-newton: Statistical optimality and second-order con- vergence.Journal of Machine Learning Research, 24(381):1–48, 2023

    2023

  77. [77]

    Automated calculation of higher order partial differential equation con- strained derivative information.SIAM Journal on Scientific Comput- ing, 41(5):C417–C445, 2019

    James R Maddison, Daniel N Goldberg, and Benjamin D Goddard. Automated calculation of higher order partial differential equation con- strained derivative information.SIAM Journal on Scientific Comput- ing, 41(5):C417–C445, 2019

    2019

  78. [78]

    Dimensionality reduction of parameter-dependent problems through proper orthogo- nal decomposition.Annals of Mathematical Sciences and Applications, 1(2):341–377, 2016

    Andrea Manzoni, Federico Negri, and Alfio Quarteroni. Dimensionality reduction of parameter-dependent problems through proper orthogo- nal decomposition.Annals of Mathematical Sciences and Applications, 1(2):341–377, 2016

    2016

  79. [79]

    Sampling via measure transport: An introduction

    Youssef Marzouk, Tarek Moselhy, Matthew Parno, and Alessio Span- tini. Sampling via measure transport: An introduction. InHandbook of uncertainty quantification, pages 1–41. Springer, 2016

    2016

  80. [80]

    Founda- tional research gaps and future directions for digital twins.Washington, DC: The National Academies Press, 2024

    National Academies of Sciences, Engineering, and Medicine. Founda- tional research gaps and future directions for digital twins.Washington, DC: The National Academies Press, 2024. 101

    2024

Showing first 80 references.