Pith. sign in

REVIEW 2 major objections 4 minor 3 cited by

The paper shows that Kalman–Bucy filtering can be reduced to a moving R-dimensional subspace, with error set by the model noise outside the subspace, and that a particle version converges to this reduced filter at rate 1/√P.

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 · deepseek-v4-flash

2026-08-04 16:55 UTC pith:JFZT4GXS

load-bearing objection Solid, serious paper: send it to review; the new material is the process-level DLR derivation and the Dlr-Enkf propagation-of-chaos analysis, with fixable technical gaps rather than a load-bearing flaw. the 2 major comments →

arxiv 2509.11210 v2 pith:JFZT4GXS submitted 2025-09-14 math.NA cs.NA

Dynamical Low-Rank Approximations for Kalman Filtering

Fabio Nobile , Thomas Trigo Trindade This is my paper
classification math.NA cs.NA MSC 65C3065C3560H1093E11
keywords dynamical low-rank approximationKalman-Bucy processensemble Kalman filterdata assimilationRiccati equationpropagation of chaosstochastic differential equationslow-rank filtering
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.

The paper tries to establish that continuous-time Kalman–Bucy filtering of a linear SDE can be performed on a low-dimensional, time-varying subspace without losing the Gaussian structure of the filter. It constructs the DLR-KBP, a Gaussian process whose law is a dynamical low-rank approximation of the Kalman–Bucy process, and shows that its mean and covariance obey a closed system of reduced Kalman–Bucy equations. This reduction is valid when the model noise is small and the filtering distribution concentrates near an R-dimensional subspace; the paper quantifies the error in terms of the norm of the noise orthogonal to that subspace. The same construction yields the DLR-EnKF, an ensemble filter with particles in reduced space, for which the paper proves well-posedness and a propagation-of-chaos bound at rate 1/√P. If correct, this gives a principled way to run much larger ensembles at modest cost in high-dimensional data assimilation problems.

Core claim

Under the mean-separated dynamically orthogonal ansatz X_t = U^0_t + U_t Y_t^T, with Y_t conditionally zero mean and the gauge conditions (U^i_t)^T α^j_t = 0, (U^i_t)^T β^j_t = 0, the Kalman–Bucy process reduces to equations (23)–(25). The observation terms are always premultiplied by P_t, whose range lies in U_t, so the physical modes evolve independently of observations by the gradient flow dU_t = Π⊥_U A U_t dt on the manifold of orthonormal frames; the stochastic modes are linear; and the conditional law of the solution is Gaussian. Consequently, the mean and covariance of the DLR-KBP are exactly characterized by deterministic reduced Kalman–Bucy equations, including the reduced Riccati e

What carries the argument

The carrier of the argument is the mean-separated dynamically orthogonal (DO) ansatz X_t = U^0_t + U_t Y_t^T, with Y_t conditionally zero mean and gauge conditions that eliminate redundant motion between the deterministic basis U_t and the zero-mean stochastic modes Y_t. For the KBP the observation process drops out of the U-equation because P_t has range in U_t; the basis evolves by the gradient flow dU_t = Π⊥_U A U_t dt on the manifold of orthonormal frames, and the Gram matrix M^Y_t solves the reduced Riccati equation (35). The covariance is exactly P_t = U_t M^Y_t U_t^T, so all full-state operations reduce to R×R matrix algebra.

Load-bearing premise

The load-bearing premise is the mean-separated low-rank ansatz X_t = U^0_t + U_t Y_t^T with Y_t conditionally zero mean, together with Assumption A1 that ||Π⊥_U Σ^{1/2}||_F < ε: if the filtering covariance has substantial support outside the moving subspace U_t, or the model noise feeds significantly into that orthogonal complement, the mode equations are no longer the projection of the KBP and the Gronwall error bounds fail.

What would settle it

Run the paper's linear-advection full-observation experiment with fixed rank R=15 and isotropic model noise Σ=σI for σ=10^{-3} and σ=0.5. Proposition 4.4 predicts that the Frobenius error between DLR-KBP and full KBP covariances grows like σ(d−R) (the square of ε = ||Π⊥_U Σ^{1/2}||_F) and becomes negligible as σ→0; if the σ=0.5 error is no larger than the σ=10^{-3} error, the assumed small-noise mechanism is not the one controlling accuracy.

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

If this is right

  • DLR-KBP is the exact equation of motion for a Gaussian process whose law is the rank-R low-rank filter; for rank-R Gaussian initial data it remains Gaussian and rank-R for all times.
  • If Assumption A1 holds (model noise orthogonal to the subspace has Frobenius norm < ε), the mean-square error of the mean and squared Frobenius error of the covariance relative to KBP are O(initial error + ε²) with explicit Gronwall constants.
  • The reduced deterministic equations for U^0, U, and M^Y are not an ad hoc covariance truncation: they are the exact moment equations of the DLR-KBP process, giving a rigorous derivation of the low-rank Riccati flow.
  • For fully observed dissipative linear systems with P > 4R − 1 particles, DLR-EnKF is well-posed, and propagation of chaos gives uniform-in-time convergence of any particle to the DLR-KBP at rate 1/√P.
  • Because DLR-EnKF evolves only R-dimensional stochastic modes, it permits P ≫ d particles at roughly the cost of a d-particle standard EnKF, reducing Monte Carlo error and spurious correlations.

Where Pith is reading between the lines

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

  • Editorial extension: the fact that the subspace U_t evolves without using observations suggests the method works best when the signal operator A itself has a slowly rotating dominant invariant subspace; when A varies rapidly or the eigenvalue gap is small, the low-rank subspace may lag the filter and need rank adaptation.
  • Editorial extension: the propagation-of-chaos condition depends on the reduced rank R rather than the state dimension d, so the practical promise is in very high-dimensional settings where a small R suffices; a cost–accuracy comparison of DLR-EnKF versus standard EnKF at fixed RMSE would make this concrete.
  • Editorial extension: the same DO construction could be carried over to nonlinear signal dynamics with a particle-in-reduced-space empirical measure, provided hyper-reduction is used to avoid evaluating full-state nonlinear terms; the paper identifies this as future work.

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

2 major / 4 minor

Summary. The paper develops a dynamical low-rank approximation of the Kalman-Bucy process (DLR-KBP) by applying the dynamically orthogonal/DO formalism to the Kalman-Bucy SDE. It derives evolution equations for the mean, the physical subspace, and the stochastic modes, shows that the DLR-KBP is Gaussian and that its covariance obeys a reduced Riccati equation, and gives Gronwall-type moment error bounds under a small-noise assumption. It then proposes a particle approximation, the DLR-ENKF, and proves well-posedness and a propagation-of-chaos result under full-observation and dissipative-dynamics assumptions. Numerical experiments on a linear advection model and a 2D advection-diffusion model illustrate the accuracy and reduced-cost claims. The authors state explicitly that the reduced Kalman-Bucy equations are not new, and that the method is intended for small or vanishing model noise.

Significance. If the proof gaps identified below are repaired, the paper would be a solid contribution to reduced-order filtering. Its main value lies in deriving the known reduced Kalman-Bucy equations from a stochastic-process ansatz rather than an ad hoc covariance truncation, and in providing a particle formulation with rigorous large-particle asymptotics. The paper is honest about its assumptions and about the limited applicability of the method to small model noise. The numerical experiments support the claimed P^{-1/2} decay and the computational advantage of the DLR-ENKF. The treatment of the DLR-KBP as a Gaussian process and the reduced Riccati equation are clean and well motivated. The main weaknesses are technical: several stated hypotheses in the moment-bound lemmas are not sufficient for the proofs as written, and the error-bound proof in Proposition 4.4 contains an unstated boundedness assumption and an incorrect norm identity.

major comments (2)
  1. [Lemma 5.2, Lemma 5.3, Theorem 5.5] The moment bounds are stated for P > 4R−1, but the proof of Lemma 5.2 via Lemma 5.6(a) gives the restriction n < 1 + (P−1)/(2R), so for a given n one needs P > 2R(n−1)+1. Lemma 5.3's proof uses the assumption 4R+1 < P to make a cubic coefficient non-positive, yet its statement and Theorem 5.5 assume only P > 4R−1. Consequently Theorem 5.5 is not proven under its stated hypothesis. This is load-bearing for the well-posedness of the DLR-ENKF and for the propagation-of-chaos results that rely on these moment bounds. The fix is to restate Lemma 5.2 with the n-dependent condition, adjust Lemma 5.3 to P > 4R+1, and update Theorem 5.5 and Proposition 5.9 hypotheses accordingly.
  2. [§4.1.3, Proposition 4.4] The proof invokes 'owing to the boundedness of P_KBP and P_DLR' to justify the Gronwall step, but no such uniform boundedness is assumed or proven. For fixed T the Riccati solutions are continuous on [0,T] and hence the relevant coefficients are integrable even if they grow exponentially, so the argument can be repaired by replacing uniform boundedness with integrability on [0,T]. Separately, the displayed identity for ||ΔΣ_t||_F² is incorrect: ΔΣ = Σ − Π_U Σ Π_U also contains the term Π_⊥ Σ Π_⊥ and cross terms, not only the two mixed terms shown. The bound ||ΔΣ_t||_F² ≤ 2λmax(Σ)ε² does not follow as written; a triangle-inequality estimate gives a bound of order ε² with a different constant. These issues do not invalidate the qualitative statement of the proposition but need correction.
minor comments (4)
  1. [Abstract / Section 1] The phrase 'when the filtering distribution concentrates around a low dimensional subspace' is motivational; the actual sufficient condition is Assumption A1 on the noise component orthogonal to U_t. Clarify this in the abstract or introduction to avoid overstating the regime of validity.
  2. [§5.3, Proposition 5.9] The initial-condition estimate for n ∈ [1,2) is written as E[G^n]^{1/n} ≤ c_n^2 / √P; the displayed inequality should be c_n / P^{1/2} (after absorbing constants). The conclusion is correct, but the notation is sloppy.
  3. [§6.2] The BUG-like integrator for the DLR-ENKF is described algorithmically and checked numerically, but no stability or consistency analysis is given for it. Since the paper makes theoretical claims elsewhere, a brief remark that the integrator analysis is outside the scope would help set expectations.
  4. [Appendix B, proof of Lemma 5.2] The proof says 'the result follows' after identifying β/τ2 = (P−1)/(4R), but it does not state the resulting condition on n. Please make the n-dependence explicit, as this affects the statements of the lemmas.

Circularity Check

0 steps flagged

No significant circularity: the DLR-KBP and DLR-ENKF equations are derived in-paper from an explicit ansatz and constraints; self-citations are background or standard lemmas, and the overlap with [48,51,52] is acknowledged.

full rationale

The central derivation chain is self-contained. Section 3 starts from the ansatz (11) and the orthogonality constraints (13), and derives the mean-separated DO equations (14)-(16) by Itô calculus and coefficient matching (Lemmas 3.1, 3.2). Substituting the Kalman-Bucy coefficients yields the DLR-KBP equations (23)-(25); the covariance equation (29) and reduced Riccati equation (35) are direct consequences, not fitted quantities. Proposition 4.4 is a conditional Gronwall error bound whose assumption A1 explicitly quantifies the small-noise condition, so the bound is not a self-definitional restatement. The paper expressly states that the reduced Kalman-Bucy equations are not new: 'these reduced equations are not new – having been introduced in [48, 51, 52] as ad hoc model-order reduction schemes' (Section 7). The DLR-ENKF well-posedness and propagation-of-chaos results (Theorem 5.5, Proposition 5.11) are proved in the text, using standard external ingredients ([36, Theorem II.3.5], [15, Lemma 7.2]) and a standard Gaussian sample-covariance bound ([13, Lemma 10]); the self-citations to [26, 27, 37, 38] are background for the DLR formalism, not load-bearing substitutes for proof. No parameter is fitted to a target output; numerical experiments validate the derived rates. Minor technical caveats (unstated boundedness of P_KBP and P_DLR in the proof of Proposition 4.4; moment-range precision in Lemma 5.2) are correctness concerns, not circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 6 axioms · 0 invented entities

No fitted parameters; all constants are either from prior literature or generic Gronwall constants. The low-rank ansatz and assumptions A1-A3 are the key domain assumptions.

axioms (6)
  • domain assumption The filtering process satisfies the low-rank ansatz X_t = U^0_t + U_t Y_t^T for all t (eq. (11)), with U, U^0 adapted to the observation filtration and Y zero-conditional-mean.
    All DLR equations (14)-(16) are derived from this ansatz; if it fails the mode equations are not exact.
  • domain assumption Smallness of the unmodelled noise: ||Π⊥_U Σ^{1/2}||_F < ε (Assumption A1).
    Used in Prop 4.4 to bound Dlr-Kbp mean and covariance error; without it the low-rank approximation can be poor or unstable.
  • domain assumption Full observations: H = I_d and Γ = ρ^{-1} I_d (Assumption A2).
    Used for well-posedness of Dlr-Enkf and propagation of chaos (Thm 5.5, Prop 5.9, 5.11).
  • domain assumption Dissipativity: λmax(A + A^T) < 0 (Assumption A3).
    Provides uniform moment bounds on the sample covariance, needed for well-posedness and propagation of chaos.
  • standard math Lemma 5.6 from [15] (Del Moral and Tugaut) giving uniform moment estimates for processes with drift and quadratic-variation bounds.
    Quoted as a black box for the propagation of chaos proof.
  • standard math Global Riccati well-posedness from [8, Lemma 2] and the Gaussian property of inhomogeneous Ornstein-Uhlenbeck processes.
    Used in Lemma 4.6 and Prop 4.2.

pith-pipeline@v1.3.0-alltime-deepseek · 42341 in / 18074 out tokens · 201234 ms · 2026-08-04T16:55:06.698794+00:00 · methodology

0 comments
read the original abstract

We propose a dynamical low rank approximation of the Kalman-Bucy process (DLR-KBP), which evolves the filtering distribution of a partially continuously observed linear SDE on a small time-varying subspace at reduced computational cost. This reduction is valid in presence of small noise and when the filtering distribution concentrates around a low dimensional subspace. We further extend this approach to a DLR-ENKF process, where particles are evolved in a low dimensional time-varying subspace at reduced cost. This allows for a significantly larger ensemble size compared to standard EnKF at equivalent cost, thereby lowering the Monte Carlo error and improving filter accuracy. Theoretical properties of the DLR-KBP and DLR-ENKF are investigated, including a propagation of chaos property. Numerical experiments demonstrate the effectiveness of the technique.

Figures

Figures reproduced from arXiv: 2509.11210 by Fabio Nobile, Thomas Trigo Trindade.

Figure 1
Figure 1. Figure 1: Mean and covariance errors between Fom-Kbp and Dlr-Kbp. The Dlr￾Kbp rank is R = 15. BAP = Best Approximation Possible = ∥P Fom t − TR(P Fom t )∥F . Note that this definition of the RMSE also takes into account the “uncertainty” of the error via the covariance. The former metrics are shown in the two plots of fig. 2. Both the errors in the mean and covariance of Dlr-Kbp with respect to their Kbp coun￾terpar… view at source ↗
Figure 2
Figure 2. Figure 2: Dlr-Kbp approximation properties of Kbp for varying R. (a) (Dlr)-Kbp RMSEs (b) Integrated RMSEs v. computing cost [s] [PITH_FULL_IMAGE:figures/full_fig_p031_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Signal tracking abilities of Dlr-Kbp for varying R via RMSE metrics. 6.2 Air pollution model This section investigates the numerical performance of Dlr-Enkf with respect to Fom￾Enkf, specifically the relation between the number of particles, RMSE and rank. The model is given by the Finite Element discretisation of a 2D stochastic advection-diffusion model over the time-interval [0, 1] and either full or pa… view at source ↗
Figure 4
Figure 4. Figure 4: Dlr-Enkf approximation error of Dlr-Kbp with P particles and Wft is a finite-dimensional Q-Wiener process that will be specified below. The initial condition has distribution u0(x) = exp(−(x1 − 0.5)2 − (x2 − 0.5)2 ) + R Xtrue i=1 1 i 2 sin(iπx1) cos(iπx2)ξi , ξi iid∼ N (0, 1) where Rtrue = 12. This SPDE admits a solution in the sense of [33, Theorem 4.2.4], i.e., a continuous H-valued process whose equival… view at source ↗
Figure 5
Figure 5. Figure 5: Each indicator function χℓ is associated to one of the red squares. where if ⋆ = full, Vet is a Q-Brownian motion on the finite element basis (like Wft), and if ⋆ = part, then the observation process is a finite-dimensional SDE and Vet = Vet a 25-dimensional Brownian motion. Note that for either observation operator h : H → K, its transpose h ⊤ : K → H verifies ⟨hx, y⟩K = ⟨x, h⊤y⟩H for x ∈ H and y ∈ K. For… view at source ↗
Figure 6
Figure 6. Figure 6: Errors [PITH_FULL_IMAGE:figures/full_fig_p035_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Sample mean RMSE ± standard deviation (over 10 runs) of Enkf(P = 10), Enkf(P = 425) and Dlr-Enkf(P = 425) with R = 10 35 [PITH_FULL_IMAGE:figures/full_fig_p035_7.png] view at source ↗

discussion (0)

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

Forward citations

Cited by 3 Pith papers

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

  1. Dynamical Low-Rank Filters for Data Assimilation

    math.NA 2026-07 conditional novelty 6.0

    Dynamical low-rank filters minimize joint mean-covariance error for SDE data assimilation and extend to Kalman-Bucy, ensemble, and particle forms for nonlinear problems.

  2. Long-time Behaviour of DLRA for SDEs

    math.PR 2026-06 unverdicted novelty 5.0

    Proves existence of invariant measures for strong DO systems approximating SDEs via moment estimates and fixed-point arguments under standard dissipativity assumptions.

  3. Dynamical Low-Rank Smoothing

    math.NA 2026-07 conditional novelty 4.0

    DLRA-JMCO filtering is extended via RTS recursion to a reduced-order smoother, with a Kalman–Bucy form for affine drift, cutting cost and storage while keeping an adaptive low-rank basis.

Reference graph

Works this paper leans on

52 extracted references · 1 linked inside Pith · cited by 3 Pith papers

  1. [1]

    On-the-fly reduced-order mod- elling of the filter density function with time-dependent subspaces.Combustion 36 Theory and Modelling, pages 1–19, 2025

    Aidyn Aitzhan, Peyman Givi, and Hessam Babaee. On-the-fly reduced-order mod- elling of the filter density function with time-dependent subspaces.Combustion 36 Theory and Modelling, pages 1–19, 2025

    2025

  2. [2]

    Multi-level data assimilation for ocean forecasting using the shallow-water equations.Journal of Computational Physics, 524:113722, 3 2025

    Florian Beiser, H ˚ avard Heitlo Holm, Kjetil Olsen Lye, and Jo Eidsvik. Multi-level data assimilation for ocean forecasting using the shallow-water equations.Journal of Computational Physics, 524:113722, 3 2025

    2025

  3. [3]

    A mollified ensemble Kalman filter.Quar- terly Journal of the Royal Meteorological Society, 136:1636–1643, 7 2010

    Kay Bergemann and Sebastian Reich. A mollified ensemble Kalman filter.Quar- terly Journal of the Royal Meteorological Society, 136:1636–1643, 7 2010

    2010

  4. [4]

    Bishop and Pierre Del Moral

    Adrian N. Bishop and Pierre Del Moral. On the mathematical theory of ensemble (linear-Gaussian) Kalman–Bucy filtering.Mathematics of Control, Signals, and Systems, 35:835–903, 12 2023

    2023

  5. [5]

    A Strongly Convergent Numerical Scheme from Ensemble Kalman Inversion.SIAM Journal on Numerical Analysis, 56:2537–2562, 1 2018

    Dirk Bl¨ omker, Claudia Schillings, and Philipp Wacker. A Strongly Convergent Numerical Scheme from Ensemble Kalman Inversion.SIAM Journal on Numerical Analysis, 56:2537–2562, 1 2018

    2018

  6. [6]

    Benign Landscape of the Brockett Cost Function.url: www.racetothebottom.xyz/posts/brockett-symmetric/, visited on 2023-12-04

    Nicolas Boumal. Benign Landscape of the Brockett Cost Function.url: www.racetothebottom.xyz/posts/brockett-symmetric/, visited on 2023-12-04

    2023

  7. [7]

    Cambridge University Press, 2023

    Nicolas Boumal.Introduction to Optimization on Smooth Manifolds. Cambridge University Press, 2023

    2023

  8. [8]

    R.S. Bucy. Global theory of the Riccati equation.Journal of Computer and System Sciences, 1:349–361, 12 1967

    1967

  9. [9]

    An Ensemble Kalman Filter for Numerical Weather Prediction Based on Variational Data As- similation: VarEnKF.Monthly Weather Review, 145:617–635, 2 2017

    Mark Buehner, Ron McTaggart-Cowan, and Sylvain Heilliette. An Ensemble Kalman Filter for Numerical Weather Prediction Based on Variational Data As- similation: VarEnKF.Monthly Weather Review, 145:617–635, 2 2017

    2017

  10. [10]

    A rank-adaptive robust integrator for dynamical low-rank approximation.BIT Numerical Mathematics, 62:1149–1174, 12 2022

    Gianluca Ceruti, Jonas Kusch, and Christian Lubich. A rank-adaptive robust integrator for dynamical low-rank approximation.BIT Numerical Mathematics, 62:1149–1174, 12 2022

    2022

  11. [11]

    An unconventional robust integrator for dynamical low-rank approximation.BIT Numerical Mathematics, 62:23–44, 3 2022

    Gianluca Ceruti and Christian Lubich. An unconventional robust integrator for dynamical low-rank approximation.BIT Numerical Mathematics, 62:23–44, 3 2022

    2022

  12. [12]

    Sorensen

    Saifon Chaturantabut and Danny C. Sorensen. Nonlinear Model Reduction via Discrete Empirical Interpolation.SIAM Journal on Scientific Computing, 32:2737– 2764, 1 2010

    2010

  13. [13]

    Alexey Chernov, H ˚ akon Hoel, Kody J. H. Law, Fabio Nobile, and Raul Tempone. Multilevel ensemble Kalman filtering for spatio-temporal processes.Numerische Mathematik, 147:71–125, 1 2021

    2021

  14. [14]

    Jana de Wiljes, Sebastian Reich, and Wilhelm Stannat. Long-Time Stability and Accuracy of the Ensemble Kalman–Bucy Filter for Fully Observed Processes and Small Measurement Noise.SIAM Journal on Applied Dynamical Systems, 17:1152– 1181, 1 2018

    2018

  15. [15]

    Del Moral and J

    P. Del Moral and J. Tugaut. On the stability and the uniform propagation of chaos properties of Ensemble Kalman–Bucy filters.The Annals of Applied Probability, 28, 4 2018

    2018

  16. [16]

    Geir Evensen. Sequential data assimilation with a nonlinear quasi-geostrophic model using Monte Carlo methods to forecast error statistics.Journal of Geophys- ical Research: Oceans, 99:10143–10162, 5 1994

    1994

  17. [17]

    A mathematical perspective on transformers.arXiv preprint arXiv:2312.10794, 2023

    Borjan Geshkovski, Cyril Letrouit, Yury Polyanskiy, and Philippe Rigollet. A mathematical perspective on transformers.arXiv preprint arXiv:2312.10794, 2023. 37

    Pith/arXiv arXiv 2023

  18. [18]

    Givens and Rae Michael Shortt

    Clark R. Givens and Rae Michael Shortt. A class of Wasserstein metrics for prob- ability distributions.Michigan Mathematical Journal, 31, 1 1984

    1984

  19. [19]

    A hybrid particle-ensemble Kalman filter for problems with medium nonlinearity.PLOS ONE, 16:e0248266, 3 2021

    Ian Grooms and Gregor Robinson. A hybrid particle-ensemble Kalman filter for problems with medium nonlinearity.PLOS ONE, 16:e0248266, 3 2021

    2021

  20. [20]

    On the Relaxation Dynamics of Lohe Oscillators on Some Riemannian Manifolds.Journal of Statistical Physics, 172:1427–1478, 9 2018

    Seung-Yeal Ha, Dongnam Ko, and Seung-Yeon Ryoo. On the Relaxation Dynamics of Lohe Oscillators on Some Riemannian Manifolds.Journal of Statistical Physics, 172:1427–1478, 9 2018

    2018

  21. [21]

    John Harlim and Andrew J. Majda. Catastrophic filter divergence in filtering nonlinear dissipative systems.Communications in Mathematical Sciences, 8(1):27 – 43, 2010

    2010

  22. [22]

    P. L. Houtekamer and Herschel L. Mitchell. Data Assimilation Using an Ensemble Kalman Filter Technique.Monthly Weather Review, 126:796–811, 3 1998

    1998

  23. [23]

    Ensemble Kalman methods for inverse problems.Inverse Problems, 29:045001, 4 2013

    Marco A Iglesias, Kody J H Law, and Andrew M Stuart. Ensemble Kalman methods for inverse problems.Inverse Problems, 29:045001, 4 2013

    2013

  24. [24]

    Springer International Pub- lishing, 2021

    Olav Kallenberg.Foundations of Modern Probability. Springer International Pub- lishing, 2021

    2021

  25. [25]

    Shreve.Brownian motion and stochastic calculus

    Ioannis Karatzas and Steven E. Shreve.Brownian motion and stochastic calculus. Springer-Verlag, 1991

    1991

  26. [26]

    Stability properties of a projector-splitting scheme for dynamical low rank approximation of random parabolic equations.Numerische Mathematik, 149:973–1024, 12 2021

    Yoshihito Kazashi, Fabio Nobile, and Eva Vidliˇ ckov´ a. Stability properties of a projector-splitting scheme for dynamical low rank approximation of random parabolic equations.Numerische Mathematik, 149:973–1024, 12 2021

    2021

  27. [27]

    Dynamical low-rank ap- proximation for stochastic differential equations.Mathematics of Computation, 94(353):1335–1375, Aug 2025

    Yoshihito Kazashi, Fabio Nobile, and Fabio Zoccolan. Dynamical low-rank ap- proximation for stochastic differential equations.Mathematics of Computation, 94(353):1335–1375, Aug 2025

    2025

  28. [28]

    Well-posedness and accuracy of the ensemble Kalman filter in discrete and continuous time.Nonlinearity, 27:2579– 2603, 10 2014

    D T B Kelly, K J H Law, and A M Stuart. Well-posedness and accuracy of the ensemble Kalman filter in discrete and continuous time.Nonlinearity, 27:2579– 2603, 10 2014

    2014

  29. [29]

    Majda, and Xin T

    David Kelly, Andrew J. Majda, and Xin T. Tong. Concrete ensemble Kalman filters with rigorous catastrophic filter divergence.Proceedings of the National Academy of Sciences, 112:10589–10594, 8 2015

    2015

  30. [30]

    Dynamical Low-Rank Approximation.SIAM Journal on Matrix Analysis and Applications, 29:434–454, 1 2007

    Othmar Koch and Christian Lubich. Dynamical Low-Rank Approximation.SIAM Journal on Matrix Analysis and Applications, 29:434–454, 1 2007

    2007

  31. [31]

    On numerical properties of the ensemble Kalman filter for data assimilation.Computer Methods in Applied Mechanics and Engineering, 197:3574–3583, 8 2008

    Jia Li and Dongbin Xiu. On numerical properties of the ensemble Kalman filter for data assimilation.Computer Methods in Applied Mechanics and Engineering, 197:3574–3583, 8 2008

    2008

  32. [32]

    Shiryaev.Statistics of random processes

    Robert Liptser and Albert N. Shiryaev.Statistics of random processes. Springer, 2001

    2001

  33. [33]

    Springer International Publishing AG, 2015

    Wei Liu and Michael R¨ ockner.Stochastic Partial Differential Equations - An In- troduction. Springer International Publishing AG, 2015

    2015

  34. [34]

    Lord, Catherine E

    Gabriel J. Lord, Catherine E. Powell, and Tony Shardlow.Introduction to Com- putational Stochastic PDEs. Cambridge University Press, 2014

    2014

  35. [35]

    Oseledets

    Christian Lubich and Ivan V. Oseledets. A projector-splitting integrator for dy- namical low-rank approximation.BIT Numerical Mathematics, 54:171–188, 3 2014. 38

    2014

  36. [36]

    Woodhead Pub- lishing, 2008

    Xuerong Mao.Stochastic Differential Equations and Applications. Woodhead Pub- lishing, 2008

    2008

  37. [37]

    Musharbash, F

    E. Musharbash, F. Nobile, and T. Zhou. Error Analysis of the Dynamically Or- thogonal Approximation of Time Dependent Random PDEs.SIAM Journal on Scientific Computing, 37:A776–A810, 1 2015

    2015

  38. [38]

    Dual Dynamically Orthogonal approxima- tion of incompressible Navier Stokes equations with random boundary conditions

    Eleonora Musharbash and Fabio Nobile. Dual Dynamically Orthogonal approxima- tion of incompressible Navier Stokes equations with random boundary conditions. Journal of Computational Physics, 354:135–162, 2 2018

    2018

  39. [39]

    Nikolas N¨ usken, Sebastian Reich, and Paul J. Rozdeba. State and Parameter Estimation from Observed Signal Increments.Entropy, 21:505, 5 2019

    2019

  40. [40]

    Springer Interna- tional Publishing, 2017

    Alfio Quarteroni.Numerical Models for Differential Problems. Springer Interna- tional Publishing, 2017

    2017

  41. [41]

    Springer Berlin Heidelberg, 1994

    Alfio Quarteroni and Alberto Valli.Numerical Approximation of Partial Differen- tial Equations. Springer Berlin Heidelberg, 1994

    1994

  42. [42]

    Pavel Sakov and Peter R. Oke. A deterministic formulation of the ensemble Kalman filter: an alternative to ensemble square root filters.Tellus A: Dynamic Meteorology and Oceanography, 60:361, 1 2008

    2008

  43. [43]

    Sapsis and Pierre F.J

    Themistoklis P. Sapsis and Pierre F.J. Lermusiaux. Dynamically orthogonal field equations for continuous stochastic dynamical systems.Physica D: Nonlinear Phe- nomena, 238:2347–2360, 12 2009

    2009

  44. [44]

    The rank- reduced Kalman filter: Approximate dynamical-low-rank filtering in high dimen- sions.Advances in Neural Information Processing Systems, 36:61364–61376, 2023

    Jonathan Schmidt, Philipp Hennig, J¨ org Nick, and Filip Tronarp. The rank- reduced Kalman filter: Approximate dynamical-low-rank filtering in high dimen- sions.Advances in Neural Information Processing Systems, 36:61364–61376, 2023

    2023

  45. [45]

    Thomas Sondergaard and Pierre F. J. Lermusiaux. Data Assimilation with Gaus- sian Mixture Models Using the Dynamically Orthogonal Field Equations. Part I: Theory and Scheme.Monthly Weather Review, 141:1737–1760, 6 2013

    2013

  46. [46]

    Thomas Sondergaard and Pierre F. J. Lermusiaux. Data Assimilation with Gaus- sian Mixture Models Using the Dynamically Orthogonal Field Equations. Part II: Applications.Monthly Weather Review, 141:1761–1785, 6 2013

    2013

  47. [47]

    High- dimensional ensemble Kalman filter with localization, inflation, and iterative up- dates.Quarterly Journal of the Royal Meteorological Society, 150:4870–4884, 10 2024

    Hao-Xuan Sun, Shouxia Wang, Xiaogu Zheng, and Song Xi Chen. High- dimensional ensemble Kalman filter with localization, inflation, and iterative up- dates.Quarterly Journal of the Royal Meteorological Society, 150:4870–4884, 10 2024

    2024

  48. [48]

    Low-rank approximated Kalman–Bucy filters us- ing Oja’s principal component flow for linear time-invariant systems.IEEE Control Systems Letters, 2024

    Daiki Tsuzuki and Kentaro Ohki. Low-rank approximated Kalman–Bucy filters us- ing Oja’s principal component flow for linear time-invariant systems.IEEE Control Systems Letters, 2024

    2024

  49. [49]

    Low-rank approximated Kalman filter using Oja’s principal component flow for discrete-time linear systems

    Daiki Tsuzuki and Kentaro Ohki. Low-rank approximated Kalman filter using Oja’s principal component flow for discrete-time linear systems. In2024 SICE Festival with Annual Conference (SICE FES), pages 242–247. IEEE, 2024

    2024

  50. [50]

    Helmke, and J.B

    None Wei-Yong Yan, U. Helmke, and J.B. Moore. Global analysis of Oja’s flow for neural networks.IEEE Transactions on Neural Networks, 5:674–683, 1994

    1994

  51. [51]

    Comparison of estimation error between two different low-rank Kalman-Bucy filters

    Shuto Yamada and Kentaro Ohki. Comparison of estimation error between two different low-rank Kalman-Bucy filters. In2021 60th Annual Conference of the Society of Instrument and Control Engineers of Japan (SICE), pages 462–467. IEEE, 2021. 39

    2021

  52. [52]

    Z t 0 tr( cM bYs )2d[ ˜Ns, ˜Ns]sds 1/2 # . To bound the last term, notice that 8E

    Shuto Yamada and Kentaro Ohki. On a new low-rank Kalman-Bucy filter and its convergence property. InProceedings of the ISCIE International Symposium on Stochastic Systems Theory and its Applications, volume 2021, pages 16–20. The ISCIE Symposium on Stochastic Systems Theory and Its Applications, 2021. A Alternative DO condition In this section, we conside...

    2021