pith. sign in

arxiv: 2605.18745 · v2 · pith:YGQREDCWnew · submitted 2026-05-18 · 📊 stat.ML · cs.LG· cs.NA· math.NA· math.PR· q-fin.MF· stat.CO

SURGE: Approximation and Training Free Particle Filter for Diffusion Surrogate

Lifu Wei , Yinuo Ren , Naichen Shi , Yiping Lu This is my paper

Pith reviewed 2026-06-30 18:18 UTC · model grok-4.3

classification 📊 stat.ML cs.LGcs.NAmath.NAmath.PRq-fin.MFstat.CO
keywords data assimilationparticle filterdiffusion modelssequential Monte Carloscore-based generative modelsposterior samplingpath measure
0
0 comments X

The pith

Sequential Monte Carlo reweighting over diffusion trajectories corrects guided sampling to the true posterior.

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

The paper introduces a particle filtering approach that uses a diffusion model to simulate dynamical system evolution and then incorporates noisy observations. Guidance from the observation likelihood steers the diffusion process, but this alone does not sample the correct posterior. The authors therefore run Sequential Monte Carlo reweighting and resampling along the entire diffusion path, treated as a path measure, to adjust particle weights and eliminate the mismatch. A reader would care because the resulting procedure fuses data with the diffusion prior in an unbiased way and requires no extra training or approximation steps.

Core claim

Treating the diffusion generation process as a path measure and applying Sequential Monte Carlo reweighting and resampling after observation-guided steering produces an unbiased particle filter that converges to the true posterior distribution.

What carries the argument

Sequential Monte Carlo reweighting and resampling performed on the diffusion trajectory viewed as a path measure

If this is right

  • Guided diffusion sampling is corrected to sample exactly from the observation-conditioned posterior.
  • Observational data is fused with diffusion simulations without introducing additional bias or requiring model retraining.
  • The method supports continuous, sequential correction of predicted states as new noisy observations arrive.
  • The approach remains training-free and approximation-free once a diffusion prior is available.

Where Pith is reading between the lines

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

  • The same path-measure correction idea could be tested on other generative models whose sampling trajectories admit a well-defined measure.
  • The number of particles required for reliable posterior approximation in high-dimensional state spaces remains an open practical question.
  • Because the method is unbiased, it could serve as a reference sampler when evaluating faster but approximate data-assimilation techniques.

Load-bearing premise

Reweighting and resampling particles along the diffusion path is sufficient to remove any bias introduced by the observation guidance and to guarantee convergence to the true posterior.

What would settle it

In a low-dimensional linear-Gaussian system where the exact posterior is known in closed form, generate many independent runs of the method and check whether the empirical particle distribution converges to the exact posterior as the number of particles grows.

Figures

Figures reproduced from arXiv: 2605.18745 by Lifu Wei, Naichen Shi, Yinuo Ren, Yiping Lu.

Figure 1
Figure 1. Figure 1: Many modern systems come with complemen [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Conceptual description of SURGE for Data As [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 2
Figure 2. Figure 2: Conceptual description of SURGE for Data Assimila￾tion. We present SURGE, an approximation-free data assimilation framework. SURGE presents diffusion-based process as a path distribution and perform reweighting and resampling on multi stochastic trajectories during inference to achieve approximation￾free output. To construct an approximation-free data assimilation frame￾work, we draw inspirations from part… view at source ↗
Figure 3
Figure 3. Figure 3: Performance comparison between baseline meth [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Performance comparison between baseline meth [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Qualitative comparison of vorticity field recon [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Qualitative and quantitative comparison of [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: More trajectory wise comparison between baselines and SURGE on Lorenz system. The blue SURGE’s trajectory [PITH_FULL_IMAGE:figures/full_fig_p021_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: More trajectory wise comparison between baselines and SURGE on Lorenz system. The blue SURGE’s trajectory [PITH_FULL_IMAGE:figures/full_fig_p022_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Failure case analysis of the Lorenz system under partial observation. When the trajectory is initialized near the null-isocline (x ≈ 0), the drift model exhibits significant oscillatory instability (Left). In this scenario, although the SURGE guidance attempts to correct the bias, the poor quality of the underlying proposal distribution leads to excessive concentration of particle weights and trajectory ov… view at source ↗
Figure 9
Figure 9. Figure 9: Failure case analysis of the Lorenz system under partial observation. When the trajectory is initialized near [PITH_FULL_IMAGE:figures/full_fig_p023_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Illustration of SURGE behavior under unstable and erroneous predictions from the diffusion surrogate, and [PITH_FULL_IMAGE:figures/full_fig_p023_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: More trajectory wise comparison between baselines and SURGE on Navier-stokes flow in terms of Energy Spectrum Relative Error and RMSE, demonstrating that SURGE consistently outperforms all baselines [PITH_FULL_IMAGE:figures/full_fig_p023_11.png] view at source ↗
Figure 11
Figure 11. Figure 11: More trajectory wise comparison between baselines and SURGE on Navier-stokes flow in terms of Energy [PITH_FULL_IMAGE:figures/full_fig_p024_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: More trajectory wise comparison between baselines and SURGE on Navier-stokes flow in terms of Energy [PITH_FULL_IMAGE:figures/full_fig_p024_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Impact of Ensemble Averaging. Individual particles (left) exhibit high stochastic variance, whereas the ensemble mean (right) effectively outperform in RMSE. This shows that SURGE relies on all particles for robust estimation. 24 [PITH_FULL_IMAGE:figures/full_fig_p024_13.png] view at source ↗
Figure 13
Figure 13. Figure 13: Impact of Ensemble Averaging. Individual particles (left) exhibit high stochastic variance, whereas the ensemble [PITH_FULL_IMAGE:figures/full_fig_p025_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: More trajectory wise comparison between baselines and SURGE on weather forecasting in terms of VIL [PITH_FULL_IMAGE:figures/full_fig_p025_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: More trajectory wise comparison between baselines and SURGE on weather forecasting in terms of VIL [PITH_FULL_IMAGE:figures/full_fig_p026_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: More trajectory wise comparison between baselines and SURGE on weather forecasting in terms of VIL [PITH_FULL_IMAGE:figures/full_fig_p027_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: More trajectory wise comparison between baselines and SURGE on weather forecasting in terms of VIL [PITH_FULL_IMAGE:figures/full_fig_p028_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: Ablation of the guidance term. Without guidance (right), the trajectory fails to correct drift and degrades to [PITH_FULL_IMAGE:figures/full_fig_p028_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Ablation of the reward term. Without reward (right), the trajectory fails to correct drift and degrades to the [PITH_FULL_IMAGE:figures/full_fig_p029_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: Ablation of SURGE weight computing and resampling. The trajectory is same as FlowDAS predicted. [PITH_FULL_IMAGE:figures/full_fig_p029_20.png] view at source ↗
read the original abstract

Data assimilation (DA) addresses the problem of sequentially estimating the state of a dynamical system from noisy and incomplete observations. In this work, we employ a diffusion model as a world model to simulate and predict the system's dynamics. Recently, score-based diffusion models have learned global diffusion priors that effectively model (stochastic) dynamics, revealing strong potential for data assimilation. In this paper, we investigate how information from noisy observations can be incorporated to enable continuous correction and refinement of the predicted system state when using a diffusion prior. Motivated by particle filtering methods, we represent the posterior distribution using a set of particles. After receiving noisy observations, the diffusion model is guided using the observation likelihood to steer the generation process toward observation-consistent states. Nevertheless, such guidance does not guarantee sampling from the true posterior. We therefore employ a Sequential Monte Carlo approach over the diffusion trajectory, viewed as a path measure, to reweight and resample particles, thereby correcting the generation process and ensuring convergence toward the desired posterior distribution. This leads to an unbiased particle filtering method that rigorously fuses observational data with diffusion model simulations.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 0 minor

Summary. The paper proposes SURGE, a training- and approximation-free particle filter for data assimilation that employs a diffusion model as a world model. After guiding the diffusion generation process via the observation likelihood to steer toward observation-consistent states, the method applies Sequential Monte Carlo reweighting and resampling over the diffusion trajectory (treated as a path measure) to correct the guided samples and produce particles from the true posterior.

Significance. If the unbiasedness claim holds, the work would provide a rigorous, training-free mechanism for fusing noisy observations with diffusion-based dynamical simulations, extending SMC to diffusion path measures in a manner that avoids the bias of guidance alone. This could strengthen connections between score-based generative models and classical filtering methods for sequential state estimation.

major comments (1)
  1. [Abstract] Abstract: the central claim that SMC reweighting and resampling over the diffusion trajectory 'corrects the generation process and ensuring convergence toward the desired posterior distribution' and yields an 'unbiased particle filtering method' is asserted without any derivation of the importance weights, explicit Radon-Nikodym derivative between the target posterior path measure and the guided proposal, or error analysis for score estimation, time discretization, or surrogate approximation. This is load-bearing for the unbiasedness guarantee.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their constructive feedback. We address the concern regarding the abstract's assertion of unbiasedness below, providing the strongest honest defense of the manuscript while acknowledging where revisions are warranted.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that SMC reweighting and resampling over the diffusion trajectory 'corrects the generation process and ensuring convergence toward the desired posterior distribution' and yields an 'unbiased particle filtering method' is asserted without any derivation of the importance weights, explicit Radon-Nikodym derivative between the target posterior path measure and the guided proposal, or error analysis for score estimation, time discretization, or surrogate approximation. This is load-bearing for the unbiasedness guarantee.

    Authors: We agree that the abstract, by design, asserts the unbiasedness result without including the full derivation. The main manuscript derives the importance weights via the Radon-Nikodym derivative between the target posterior path measure and the guided proposal (Section 3), establishes that the SMC procedure over diffusion trajectories yields unbiased samples from the filtering distribution in the continuous-time limit, and discusses discretization and score-estimation errors with supporting bounds. The abstract's phrasing is therefore a high-level summary of these results rather than a standalone claim. To improve clarity, we will revise the abstract to qualify the unbiasedness statement and explicitly reference the theoretical derivation in the main text. revision: yes

Circularity Check

0 steps flagged

No circularity: derivation relies on standard external SMC and diffusion path measures

full rationale

The paper presents guidance via observation likelihood followed by SMC reweighting/resampling on the diffusion trajectory as a path measure. This is framed as applying known Sequential Monte Carlo techniques to correct guided diffusion sampling, without any self-definitional reduction, fitted parameter renamed as prediction, or load-bearing self-citation chain. The unbiasedness claim rests on the standard Radon-Nikodym property of importance weights between path measures, which is an external mathematical fact not derived inside the paper. No equations or steps reduce the target posterior convergence to quantities defined by the paper's own fits or prior self-citations. The derivation is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

Abstract supplies insufficient detail to enumerate concrete free parameters or invented entities; the core assumptions are the effectiveness of diffusion priors for dynamics and the corrective power of the SMC step.

axioms (2)
  • domain assumption Score-based diffusion models have learned global diffusion priors that effectively model stochastic dynamics.
    Stated directly in the abstract as motivation for using the diffusion model as world model.
  • ad hoc to paper Sequential Monte Carlo over the diffusion trajectory corrects guided sampling and ensures convergence to the true posterior.
    This is the load-bearing step asserted to deliver the unbiased filter.

pith-pipeline@v0.9.1-grok · 5744 in / 1295 out tokens · 28125 ms · 2026-06-30T18:18:52.918781+00:00 · methodology

Review history (2 revisions) →

discussion (0)

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

Reference graph

Works this paper leans on

31 extracted references · 27 canonical work pages · 10 internal anchors

  1. [1]

    Albergo, M. S. and Vanden-Eijnden, E. Building normal- izing flows with stochastic interpolants. arXiv preprint arXiv:2209.15571,

    work page internal anchor Pith review Pith/arXiv arXiv

  2. [2]

    Albergo, M. S. and Vanden-Eijnden, E. Nets: A non-equilibrium transport sampler. arXiv preprint arXiv:2410.02711,

    work page arXiv

  3. [3]

    The Ensemble Schr{\"o}dinger Bridge filter for Nonlinear Data Assimilation

    Bao, F. and Sun, H. The ensemble schr ¨odinger bridge filter for nonlinear data assimilation. arXiv preprint arXiv:2512.18928,

    work page internal anchor Pith review Pith/arXiv arXiv

  4. [4]

    A score-based nonlinear filter for data assimilation

    Bao, F., Zhang, Z., and Zhang, G. A score-based nonlinear filter for data assimilation. arXiv preprint arXiv:2306.09282,

    work page arXiv

  5. [5]

    and Han, J

    Bruna, J. and Han, J. Posterior sampling with denoising ora- cles via tilted transport. arXiv preprint arXiv:2407.00745,

    work page arXiv

  6. [6]

    Cardoso, G., Idrissi, Y . J. E., Corff, S. L., and Moulines, E. Monte carlo guided diffusion for bayesian linear inverse problems. arXiv preprint arXiv:2308.07983,

    work page arXiv

  7. [7]

    R., Ying, L., and Izzo, Z

    Chen, H., Ren, Y ., Min, M. R., Ying, L., and Izzo, Z. Solving inverse problems via diffusion-based priors: An approximation-free ensemble sampling approach. arXiv preprint arXiv:2506.03979, 2025a. Chen, S., Jia, Y ., Qu, Q., Sun, H., and Fessler, J. A. Flow- das: A stochastic interpolant-based framework for data assimilation. arXiv preprint arXiv:2501.16...

    work page arXiv

  8. [8]

    Split gibbs discrete diffusion posterior sampling

    Chu, W., Wu, Z., Chen, Y ., Song, Y ., and Yue, Y . Split gibbs discrete diffusion posterior sampling. arXiv preprint arXiv:2503.01161,

    work page arXiv

  9. [9]

    9 SURGE Filtering Domingo-Enrich, C., Drozdzal, M., Karrer, B., and Chen, R. T. Adjoint matching: Fine-tuning flow and diffusion generative models with memoryless stochastic optimal control. arXiv preprint arXiv:2409.08861,

    work page arXiv

  10. [10]

    Discrete feynman-kac correctors

    Hasan, M., Ohanesian, V ., Gazizov, A., Bengio, Y ., Aspuru- Guzik, A., Bondesan, R., Skreta, M., and Neklyudov, K. Discrete feynman-kac correctors. arXiv preprint arXiv:2601.10403,

    work page arXiv

  11. [11]

    Adjoint sampling: Highly scalable diffusion samplers via adjoint matching.arXiv preprint arXiv:2504.11713, 2025

    Havens, A., Miller, B. K., Yan, B., Domingo-Enrich, C., Sriram, A., Wood, B., Levine, D., Hu, B., Amos, B., Karrer, B., et al. Adjoint sampling: Highly scalable diffusion samplers via adjoint matching. arXiv preprint arXiv:2504.11713,

    work page arXiv

  12. [12]

    RNE: plug-and-play diffusion inference-time control and energy-based training

    He, J., Hern ´andez-Lobato, J. M., Du, Y ., and Vargas, F. Rne: a plug-and-play framework for diffusion density estimation and inference-time control. arXiv preprint arXiv:2506.05668, 2025a. He, J., Jeha, P., Potaptchik, P., Zhang, L., Hern´andez-Lobato, J. M., Du, Y ., Syed, S., and Vargas, F. Crepe: Con- trolling diffusion with replica exchange. arXiv p...

    work page internal anchor Pith review Pith/arXiv arXiv

  13. [13]

    Flow Matching for Generative Modeling

    Lipman, Y ., Chen, R. T., Ben-Hamu, H., Nickel, M., and Le, M. Flow matching for generative modeling. arXiv preprint arXiv:2210.02747,

    work page internal anchor Pith review Pith/arXiv arXiv

  14. [14]

    Adjoint Schrödinger Bridge Sampler.arXiv preprint arXiv:2506.22565, 2025

    Liu, G.-H., Choi, J., Chen, Y ., Miller, B. K., and Chen, R. T. Adjoint schr\” odinger bridge sampler. arXiv preprint arXiv:2506.22565,

    work page arXiv

  15. [15]

    Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow

    Liu, X., Gong, C., and Liu, Q. Flow straight and fast: Learning to generate and transfer data with rectified flow. arXiv preprint arXiv:2209.03003,

    work page internal anchor Pith review Pith/arXiv arXiv

  16. [16]

    Inference-Time Scaling for Diffusion Models beyond Scaling Denoising Steps

    doi: 10.1175/1520-0469(1963)020⟨0130:DNF⟩2.0.CO;2. URL https://journals.ametsoc.org/view/ journals/atsc/20/2/1520-0469_1963_ 020_0130_dnf_2_0_co_2.xml. Ma, N., Tong, S., Jia, H., Hu, H., Su, Y .-C., Zhang, M., Yang, X., Li, Y ., Jaakkola, T., Jia, X., et al. Inference-time scaling for diffusion models beyond scaling denoising steps. arXiv preprint arXiv:2...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1175/1520-0469(1963)020 1963

  17. [17]

    GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models

    Nichol, A., Dhariwal, P., Ramesh, A., Shyam, P., Mishkin, P., McGrew, B., Sutskever, I., and Chen, M. Glide: Towards photorealistic image generation and editing with text-guided diffusion models. arXiv preprint arXiv:2112.10741,

    work page internal anchor Pith review Pith/arXiv arXiv

  18. [18]

    Large Language Diffusion Models

    10 SURGE Filtering Nie, S., Zhu, F., You, Z., Zhang, X., Ou, J., Hu, J., Zhou, J., Lin, Y ., Wen, J.-R., and Li, C. Large language diffusion models. arXiv preprint arXiv:2502.09992,

    work page internal anchor Pith review Pith/arXiv arXiv

  19. [19]

    M., and Han, J

    Ren, Y ., Gao, W., Ying, L., Rotskoff, G. M., and Han, J. Driftlite: Lightweight drift control for inference-time scal- ing of diffusion models.arXiv preprint arXiv:2509.21655, 2025a. Ren, Y ., Rotskoff, G. M., and Ying, L. A unified approach to analysis and design of denoising markov models. arXiv preprint arXiv:2504.01938, 2025b. Robinson, M., Evans, J....

    work page arXiv

  20. [20]

    arXiv preprint arXiv:2511.22688 , year=

    Sabour, A., Albergo, M. S., Domingo-Enrich, C., Boffi, N. M., Fidler, S., Kreis, K., and Vanden-Eijnden, E. Test- time scaling of diffusions with flow maps. arXiv preprint arXiv:2511.22688,

    work page arXiv

  21. [21]

    Singhal, Z

    Singhal, R., Horvitz, Z., Teehan, R., Ren, M., Yu, Z., McK- eown, K., and Ranganath, R. A general framework for inference-time scaling and steering of diffusion models. arXiv preprint arXiv:2501.06848,

    work page arXiv

  22. [22]

    Feynman- kac correctors in diffusion: Annealing, guidance, and product of experts.arXiv preprint arXiv:2503.02819, 2025

    Skreta, M., Akhound-Sadegh, T., Ohanesian, V ., Bondesan, R., Aspuru-Guzik, A., Doucet, A., Brekelmans, R., Tong, A., and Neklyudov, K. Feynman-kac correctors in diffu- sion: Annealing, guidance, and product of experts. arXiv preprint arXiv:2503.02819,

    work page arXiv

  23. [23]

    Score-Based Generative Modeling through Stochastic Differential Equations

    Song, Y ., Sohl-Dickstein, J., Kingma, D. P., Kumar, A., Er- mon, S., and Poole, B. Score-based generative modeling through stochastic differential equations. arXiv preprint arXiv:2011.13456,

    work page internal anchor Pith review Pith/arXiv arXiv 2011

  24. [24]

    Improving and generalizing flow-based generative models with minibatch optimal transport

    Tong, A., Fatras, K., Malkin, N., Huguet, G., Zhang, Y ., Rector-Brooks, J., Wolf, G., and Bengio, Y . Improving and generalizing flow-based generative models with mini- batch optimal transport. arXiv preprint arXiv:2302.00482,

    work page internal anchor Pith review Pith/arXiv arXiv

  25. [25]

    Inference-time alignment in diffusion models with reward-guided generation: Tutorial and review.arXiv preprint arXiv:2501.09685, 2025

    Uehara, M., Zhao, Y ., Wang, C., Li, X., Regev, A., Levine, S., and Biancalani, T. Inference-time alignment in diffu- sion models with reward-guided generation: Tutorial and review. arXiv preprint arXiv:2501.09685,

    work page arXiv

  26. [26]

    Training-free adaptation of diffusion models via doob’s h-transform

    11 SURGE Filtering Zhu, Q., Ye, Z., Liu, H., Wang, Z., and Chen, M. Training-free adaptation of diffusion models via doob’s h-transform. arXiv preprint arXiv:2602.16198,

    work page arXiv

  27. [27]

    On the power of (approximate) reward models for inference-time scaling.arXiv preprint arXiv:2602.01381,

    Zhu, Y . and Lu, Y . On the power of (approximate) re- ward models for inference-time scaling. arXiv preprint arXiv:2602.01381,

    work page arXiv

  28. [28]

    Then for any integrable test function ϕ, E " 1 N NX i=1 ϕ( ˜X(i)) {(X(j), ˜w(j))}N j=1 # = NX j=1 ˜w(j) ϕ(X(j))

    Let { ˜X(i)}N i=1 be obtained by multinomial resampling fromP i ˜w(i)δX (i) and assigning equal weights 1/N. Then for any integrable test function ϕ, E " 1 N NX i=1 ϕ( ˜X(i)) {(X(j), ˜w(j))}N j=1 # = NX j=1 ˜w(j) ϕ(X(j)). 14 SURGE Filtering Proof. Conditioned on the current weighted particles, the resampling indices are i.i.d. with P(A(i) = j) = ˜w(j) and...

    1963

  29. [29]

    SURGE consistently improves both SDA and FlowDAS backbones

    Full results on the Lorenz 1963 experiment. SURGE consistently improves both SDA and FlowDAS backbones. METHOD RMSE ↓ W1 ↓ BPF (N=20) 0.0625 0 .0448 DM 0.0766 0 .0549 ENKF 0.0624 0 .0448 SDA 0.0589 0 .0426 + SURGE 0.0555 0 .0396 FLOWDAS 0.0545 0 .0388 FLOWDAS AVG 0.0923 0 .0698 + SURGE 0.0502 0 .0363 Table

    1963

  30. [30]

    Additional Results Baselines

    SO ( 5% → 100%) METHOD KES-RE ↓ RMSE ↓ KES-RE ↓ RMSE ↓ BPF (N=20) 0.490 1 .143 0 .486 1 .133 DM 0.657 1 .310 0 .663 1 .320 ENKF 0.551 0 .847 0 .676 0 .800 SDA 0.473 0 .987 0 .231 0 .590 + SURGE 0.417 0 .966 0.207 0 .564 FLOWDAS 0.401 1 .018 0 .543 0 .872 FLOWDAS AVG 0.329 0 .898 0 .315 0 .723 + SURGE 0.317 0 .851 0.278 0 .673 B.4. Additional Results Basel...

    1993

  31. [31]

    Ensemble Kalman Filter (EnKF) maintains a finite ensemble and applies a Kalman-style update under a Gaussian approximation of the forecast distribution(Evensen, 2003)

    Diffusion Model (DM) refers to a plain diffusion sampler that generates trajectories from the learned prior without observation guidance, included to isolate the contribution of guidance. Ensemble Kalman Filter (EnKF) maintains a finite ensemble and applies a Kalman-style update under a Gaussian approximation of the forecast distribution(Evensen, 2003). S...

    2003