betaFeedback

WorksComputer Science

Generative Forecasting from Coarse Observations: Does Cascading Help?

Abstract

We study probabilistic forecasting of dynamical systems such as fluids and waves from coarse observations. Since the unseen fine scales influence the dynamics, the same coarse initial state can evolve into a whole distribution of futures. Generative forecasters may produce visually plausible samples, but it is hard to tell whether they sample the right conditional distribution, since trajectory data contain a single future for each initial state. We address this problem by generating reference data in a hierarchical way: for each coarse initial state, we generate many different fine-scale components and numerically evolve the full states. Using these conditional ensembles, we show that standard diffusion models and stochastic interpolants can make substantial distributional errors when trained over the full forecast horizon with finite data and model capacity, despite being exact in principle. We then show that cascading shorter forecasts greatly reduces these errors for the turbulent flows and waves we study, but we also find and characterize cases in which it hurts. Since the coarse process is not Markov, cascading discards information about the unresolved fine scales even with exact transitions. To understand this, we propose a solvable analytic model which shows that the optimal cascade depth decreases as the coupling between the unresolved and observed scales gets stronger, and increases as the dependence of the forecast on the observation becomes more oscillatory. Consistent with our experiments, cascading is detrimental when the direct forecast is "easy" to learn, for example Gaussian as in the passive scalar case. We release our datasets as a first benchmark for probabilistic forecasting with ground-truth conditional ensembles.

Introduction

Suppose we observe a turbulent flow on a coarse grid and want to predict what it will look like some time later. The grid resolves the large scales, but the unobserved fine (subgrid) scales also influence the dynamics. In 1969, Lorenz argued that in a flow with many scales, a fine-scale error propagates to ever larger ones. Two states that differ only in smallest details become as different as two random states, in finite time (Lorenz, [19]; Palmer et al., [26]). Weather forecasting is a classic example of this “butterfly effect” (Lorenz, [19]). The same coarse observation X0X_0 is therefore consistent with many futures XTX_T, one for each subgrid realization, and the forecast is the conditional distribution P(XT∣X0)P(X_T \mid X_0) (Chorin et al., [11]; 2002).

A modern approach is to learn P(XT∣X0)P(X_T \mid X_0) with conditional generative models such as diffusion models (Song et al., [33]) and stochastic interpolants (SIs) (Albergo & Vanden-Eijnden, [3]; Albergo et al., [2])1 trained on pairs of initial and final observed states. These models generate samples of future weather and turbulent flows (Price et al., [28]; Li et al., [18]; Chen et al., [9]; Kohl et al., [17]) which resemble real physical states, but do they actually sample from the right conditional distribution (Theis et al., [35])? Testing this is a challenging problem. One reason is that the conditional distribution describes all futures consistent with a given observation but available data usually only contain one trajectory per initial state.

Here we address this challenge. We design an experiment that lets us show that common generative forecasters can make substantial distributional mistakes. We then study a way to reduce these mistakes by better using the available data: forecasting over shorter time intervals and composing the results. Practical forecasters already do this, but mostly as a matter of convenience. Our goal is to understand when and why it helps, and when it can actually hurt.

The experiment works as follows. For five spatio–temporal systems with different dynamical characteristics we first generate coarse states and then complete each coarse state with many independently drawn fine-scale components (§ 2). The full states are identical under coarse graining, but the dynamics (computed by a numerical solver) evolve them into different coarse futures. For each coarse initial state this yields a reference ensemble from the conditional distribution, whose statistical properties can be studied empirically. We test a standard conditional SI and a conditional diffusion model trained directly on initial–final pairs (X0,XT)(X_0, X_T), and find that they can be substantially wrong. Figure 2 shows an example for Euler flow (second and third rows). Their samples have visually correct local statistics, but the global structure is completely wrong.

Figure 2. Direct forecasters can sample the wrong conditional distribution. One held-out Euler observation (top left). Each row shows three samples of the vorticity at time TT, followed by the conditional mean and the pointwise standard deviation over 10,000 samples. Reference: the simulator, run from 10,000 fine-scale completions of the observation. Direct SI and EDM: generative forecasters trained on pairs (X0,XT)(X_0, X_T). Cascaded SI: the cascaded forecaster of Section 3 at the validation-selected depth S⋆=10S^\star= 10, discussed in Section 5.

Diffusion models and SIs generate samples via stochastic dynamics in an auxiliary “diffusion” time. This dynamics is in general very different from the physical dynamics of the system over the forecasting period. A forecaster trained only on pairs (X0,XT)(X_0, X_T) has to learn in one fell swoop everything the physical dynamics does between 00 and TT. But the physical dynamics is a composition of many short steps. As intermediate observations at times 0<t<T0 < t < T are often available ([31, 30]), it is natural to ask whether they can be leveraged to improve conditional forecasting. Instead of learning P(XT∣X0)P(X_T \mid X_0) directly, one can learn the conditional transitions P(Xti+1∣Xti)P(X_{t_{i+1}} \mid X_{t_i}) over SS shorter intervals and compose them into a cascade of depth SS. In a cascade of SIs, the generative path is then forced to agree with the physical evolution at some intermediate times. In the Euler example, such a cascaded SI approach is much better than direct and it effectively reproduces the conditional distribution (last row of Figure 2).

In practice, the depth of a cascade is set by a fixed time step: GenCast rolls out a twelve-hour transition over fifteen days ([28]), and conditional diffusion models and SIs for fluid simulation use the time step of the training data ([9, 17, 23]). Cascading can, however, also discard information even with exact transitions. Only the full state evolves by composing short steps. After coarse graining, the unresolved fine scales carry information from one interval to the next, so that the coarse process is not Markov. To our knowledge, there is no analysis of how cascading affects the learned distribution. The questions we ask here are: is it better to forecast directly or with a cascade, what properties of the physical system determine the answer, and how deep should a cascade be?

We propose a solvable model to answer these questions (§ 3). The optimal cascade depth results from the interplay between easier learning and memory loss, and depends on the specific system under study. If the conditional distribution is easy to learn over the full horizon, cascading can be detrimental. This is the case for a scalar field passively advected by a steady flow, whose conditional distribution stays Gaussian at every horizon, with a mean affine in the observation and a covariance that does not depend on it (§ 5).

Our main contributions are the following.

  • An experimental design that enables proper evaluation of the conditional distribution. New datasets for five physical systems, with different simulated futures for each observed initial state, allow us to compare learned and true conditional distributions on every observation (§ 2 and § 4). To our knowledge, these are the first datasets designed for this notoriously challenging problem; we release them as an openly accessible benchmark. We find that diffusion models and SIs trained directly over the forecast horizon sample substantially wrong conditional distributions in four of the five systems, despite producing visually plausible samples.

  • A theoretical explanation. For smooth dynamics, we show that, through nonlinear interactions, the conditional mean leaves a fixed approximation class as the horizon grows, and that, under partial observation, a fine cascade of exact transitions loses the conditional spread of the forecast. We construct a stylized model of a rotating system driven by a hidden Ornstein–Uhlenbeck process where we compute the error of a simple SI exactly and prove that it is minimized at a finite depth S∗≥2S^* \ge2; with the rotation switched off, every cascade is worse than the exact direct forecast (§ 3).

  • An experimental test. At a fixed number of supervised transitions, cascading reduces the error of the direct forecasters by factors between 2.3 and 5.7, with the smallest error at an intermediate depth, for nonlinear flows and linear waves alike. For the passive scalar, every added stage increases the error, consistent with the analysis (§ 5).

Conditional forecasting from coarse observations

We consider a physical system whose full state YtY_t evolves according to Yt=Gt(Y0)Y_t = G_t(Y_0). A forecaster sees only the coarse state Xt=H(Yt)X_t = \mathcal{H}(Y_t). At the initial time, we write Y0=(X0,Z0)Y_0 = (X_0, Z_0), where Z0Z_0 contains the fine-scale components. Since the observation X0=x0X_0 = x_0 generally does not determine Z0Z_0, different full initial states consistent with the same observation can lead to different observed futures. The forecasting task is therefore to learn the conditional distribution

P(XT∣X0=x0)(1)P(X_T \mid X_0 = x_0) \tag*{(1)}

for every observation x0x_0. In our systems, YtY_t consists of one or two fields on a two-dimensional periodic domain, simulated on grids ranging from 1282128^2 to 102421024^2 points. The observation map H\mathcal{H} keeps only their Fourier modes with 0<∣k∣≤80 < |k| \le8 (196 real coordinates per field), while the unresolved initial modes extend to higher wavenumbers that depend on the system (Appendix § 11.§ 1).

Why trajectory data are not enough. A generative forecaster learns an approximation P^(XT∣X0=x0)\widehat{P}(X_T \mid X_0 = x_0) from which it can draw samples. To measure its accuracy, we would like to compare the learned and the true conditional distributions for each observation, using some metric DD on probability distributions, and then average over observations,

E(P^)=Ex0∼P(X0)[D(P(XT∣X0=x0),P^(XT∣X0=x0))].(2)\mathcal{E}(\widehat{P}) = \mathbb{E}_{x_0 \sim P(X_0)}\left[D\left(P(X_T \mid X_0 = x_0), \widehat{P}(X_T \mid X_0 = x_0)\right)\right]. \tag*{(2)}

Estimating E\mathcal{E} requires many samples from P(XT∣X0=x0)P(X_T \mid X_0 = x_0) for each of many observations x0x_0. But data usually contain only one future per observed initial state, with all observed states being different. Proper scoring rules can rank forecasters on such data [13], but they cannot determine how far they are from the true conditional. Averaging final spectra or other statistics over initial conditions only tests the marginal distribution P(XT)P(X_T). Comparing the joint distributions of (X0,XT)(X_0, X_T) could in principle work because they determine the conditionals. The issue is that a forecaster that generates the “correct” forecast for a slightly perturbed observation would have a nearly correct joint distribution, but the forecasts could be completely wrong if they are sensitive to the initial state (which is often the interesting regime).

Hierarchical data generation. When the system can be simulated, this problem can be addressed by sampling the initial data hierarchically. For each conditioning ensemble ee, we first sample the coarse scales x0(e)∼P(X0)x_0^{(e)} \sim P(X_0), and then many different fine scales for this coarse state,

z0(e,i)∼P(Z0∣X0=x0(e)),Y0(e,i)=(x0(e),z0(e,i)),i=1,…,n.(3)z_0^{(e,i)} \sim P(Z_0 \mid X_0 = x_0^{(e)}), \qquad Y_0^{(e,i)} = \left(x_0^{(e)}, z_0^{(e,i)}\right), \qquad i = 1,\ldots,n. \tag*{(3)}

A numerical solver sees the full initial state, coarse and fine, and evolves it into (generally) diverging future states, because the fine scales influence the evolution of the coarse ones. But the forecaster only sees the coarse scales, which are the same for every member of the ensemble. The observed endpoints XT(e,i)=H(GT(Y0(e,i)))X_T^{(e,i)} = \mathcal{H}(G_T(Y_0^{(e,i)})) are thus independent samples from P(XT∣X0=x0(e))P(X_T \mid X_0 = x_0^{(e)}), and they form a reference ensemble against which any forecast from x0(e)x_0^{(e)} can be compared (Figure 1). Each conditioning ensemble belongs as a whole to the training, validation, or test split, so every test observation is new to the forecaster.

Figure 1. Hierarchical data generation. Many unobserved fine initial states z0(i)z_0^{(i)} are combined with the same observed coarse state x0x_0. The solver evolves each full initial state Y0(i)Y_0^{(i)}, and observing the result at time TT gives a reference ensemble xT(1),xT(2),…x_T^{(1)}, x_T^{(2)}, \ldots for P(XT∣X0=x0)P(X_T \mid X_0 = x_0).

Direct forecasters can sample the wrong distribution. Using our reference ensembles, we can check whether a learned forecaster samples the right conditional distribution. This leads to our first major finding. Figure 2 shows results for a two-dimensional Euler flow, where the forecaster observes a coarse-grained initial vorticity. We trained two strong generative forecasters on pairs (X0,XT)(X_0, X_T) from the training ensembles: a conditional SI [9], which we call Direct SI, and a conditional elucidated diffusion model (EDM) [16]. We then drew 10,000 samples from each model for a held-out observation. We find that both models generate plausible individual samples. However, a closer look reveals that the large-scale structure of their forecasts is wrong. Indeed, the reference mean has a distinct large-scale geometry, and the reference standard deviation is strongly heterogeneous, while the means of the two forecasters are “washed out”, and their standard deviations are nearly uniform. The same happens for waves in heterogeneous media (Appendix § 11.§ 5).

Direct and cascaded forecasts

As in most forecasting datasets, our datasets contain observations at intermediate times 0<t<T0 < t < T. A direct forecaster does not use them: it learns P(XT∣X0)P(X_T \mid X_0) from the pairs (X0,XT)(X_0, X_T) alone. A cascaded forecaster divides the forecast horizon into SS intervals, with ti=iT/St_i = iT/S, and learns the conditional distributions Ki(x,⋅)=P(Xti+1∈⋅∣Xti=x)K_i(x,\cdot) = P(X_{t_{i+1}} \in\cdot\mid X_{t_i} = x) of the transitions between consecutive observed states, for i=0,…,S−1i = 0,\ldots,S-1. Let K^i\widehat{K}_i denote the learned approximation of KiK_i. A cascaded forecast from x0x_0 applies these approximations in turn: starting from X^0=x0\widehat{X}_0 = x_0, it samples

X^ti+1∼K^i(X^ti,⋅),i=0,…,S−1.(4)\widehat{X}_{t_{i+1}} \sim\widehat{K}_i(\widehat{X}_{t_i},\cdot), \qquad i = 0,\ldots,S-1. \tag*{(4)}

Each transition is thus conditioned only on the state generated by the previous one, and the forecast X^T\widehat{X}_T is distributed according to the composition K^0K^1⋯K^S−1\widehat{K}_0\widehat{K}_1\cdots\widehat{K}_{S-1} of the learned transitions. We implement both forecasters with stochastic interpolants (Section 4). The direct one is the Direct SI of Figure 2, and the cascaded one is Cascaded SI. The two coincide for S=1S = 1.

Consider first an ideal setting, in which the model class of the forecaster contains the true conditional distributions, training reaches the population optimum, and sampling is exact. In this setting, Direct SI samples P(XT∣X0)P(X_T \mid X_0) exactly, even though the state is only partially observed, since (2.1) is a well-defined conditional distribution. Cascaded SI need not be exact even in this ideal setting, because the composition K0K1⋯KS−1K_0K_1 \cdots K_{S-1} of the exact transitions generally differs from P(XT∣X0)P(X_T \mid X_0). In practice, however, the transitions are learned from finite data with a model of finite capacity, which is why shorter transitions can help. We analyze the two effects in turn and then show how they trade off.

Why can shorter transitions be easier to learn? Over a long horizon, the future observation can depend on the initial one in a nonlinear or oscillatory way, which a model of finite capacity may not be able to represent. Over a short horizon this dependence may be simpler. We make this precise for the conditional mean mt(x)=E[Xt∣X0=x]m_t(x) = \mathbb{E}[X_t \mid X_0 = x], since an error in the conditional mean implies an error in the distribution (Proposition C.7). We model limited capacity by a fixed class F\mathcal{F} of functions of the observation, such as the polynomials of a given degree. For smooth dynamics, the distance from mtm_t to F\mathcal{F} grows like cntn/n!c_n t^n/n! for short horizons, where tnt^n is the first power in the expansion of mtm_t whose coefficient lies outside F\mathcal{F} (Theorem C.1). For quadratic dynamics, such as Galerkin truncations of the Euler equations, and quadratic F\mathcal{F}, this coefficient is cubic in the observation and appears at order t2t^2 (Theorem C.2). Repeated nonlinear interactions then create a dependence that a quadratic map cannot represent.

It can however be represented by a composition of short maps from the same class. When the full state is observed, composing SS fitted polynomial maps of a fixed degree over intervals of length δ=T/S\delta= T/S converges to the true map over [0,T][0,T] as δ→0\delta\to0, whereas a single fitted polynomial keeps a positive error (Theorem C.5).

References

  1. [1]Abernathey, R., rochanotes, Ross, A., Jansen, M., Li, Z., Poulin, F. J., Constantinou, N. C., Sinha, A., Balwada, D., SalahKouhen, Jones, S., Rocha, C. B., Wolfe, C. L. P., Meng, C., van Kemenade, H., Bourbeau, J., Penn, J., Busecke, J., Bueti, M., & Tobias. (2022). pyqg/pyqg: v0.7.2. Zenodo. https://zenodo.org/records/6563667zenodo.org/records/6563667
  2. [2]Albergo, M., Boffi, N. M., & Vanden-Eijnden, E. (2025). Stochastic interpolants: A unifying framework for flows and diffusions. Journal of Machine Learning Research, 26(209), 1–80.
  3. [3]Michael S. Albergo and Eric Vanden-Eijnden. Building normalizing flows with stochastic interpolants. In International Conference on Learning Representations (ICLR), 2023.arxiv.org/abs/2209.15571
  4. [4]Lorenzo Baldassari, Ali Siahkoohi, Josselin Garnier, Knut Solna, and Maarten V de Hoop. Conditional score-based diffusion models for bayesian inference in infinite dimensions. Advances in Neural Information Processing Systems, 36:24262–24290, 2023.arxiv.org/abs/2305.19147
  5. [5]Georgios Batzolis, Jan Stanczuk, Carola-Bibiane Schönlieb, and Christian Etmann. Conditional image generation with score-based diffusion models. arXiv preprint arXiv:2111.13606, 2021.arXiv
  6. [6]Kaifeng Bi, Lingxi Xie, Hengheng Zhang, Xin Chen, Xiaotao Gu, and Qi Tian. Accurate medium-range global weather forecasting with 3D neural networks. Nature, 619(7970):533–538, 2023. doi: 10.1038/s41586-023-06185-3.DOI
  7. [7]Gerard Brunick and Steven Shreve. Mimicking an Itô process by a solution of a stochastic differential equation. The Annals of Applied Probability, 23(4), August 2013. ISSN 1050-5164. doi: 10.1214/12-aap881. URL http://dx.doi.org/10.1214/12-AAP881.
  8. [8]Keaton J. Burns, Geoffrey M. Vasil, Jeffrey S. Oishi, Daniel Lecoanet, and Benjamin P. Brown. Dedalus: A flexible framework for numerical simulations with spectral methods. Phys. Rev. Res., 2:023068, Apr 2020. doi: 10.1103/PhysRevResearch.2.023068. URL https://link.aps.org/doi/10.1103/PhysRevResearch.2.023068.DOI
  9. [9]Yifan Chen, Mark Goldstein, Mengjian Hua, Michael Samuel Albergo, Nicholas Matthew Boffi, and Eric Vanden-Eijnden. Probabilistic forecasting with stochastic interpolants and fӧllmer processes. In Ruslan Salakhutdinov, Zico Kolter, Katherine Heller, Adrian Weller, Nuria Oliver, Jonathan Scarlett, and Felix Berkenkamp (eds.), Proceedings of the 41st International Conference on Machine Learning, volume 235 of Proceedings of Machine Learning Research, pp. 6728–6756. PMLR, 21–27 Jul 2024. URL https://proceedings.mlr.press/v235/chen24n.html.proceedings.mlr.press/v235/chen24n.html
  10. [10]Guillaume Chevillon. Direct multi-step estimation and forecasting. Journal of Economic Surveys, 21(4):746–785, 2007. doi: 10.1111/j.1467-6419.2007.00518.x.DOI
  11. [11]Alexandre J Chorin, Ole H Hald, and Raz Kupferman. Optimal prediction and the mori–zwanzig representation of irreversible processes. Proceedings of the National Academy of Sciences, 97(7):2968–2973, 2000.pubmed.ncbi.nlm.nih.gov/10737778
  12. [12]Chorin, A. J., Hald, O. H., & Kupferman, R. (2002). Optimal prediction with memory. Physica D: Nonlinear Phenomena, 166(3-4), 239–257.
  13. [13]Tilmann Gneiting and Adrian E Raftery. Strictly proper scoring rules, prediction, and estimation. Journal of the American Statistical Association, 102(477):359–378, 2007. doi: 10.1198/016214506000001437. URL https://doi.org/10.1198/016214506000001437.
  14. [14]Isaac M. Held, Raymond T. Pierrehumbert, Stephen T. Garner, and Kyle L. Swanson. Surface quasi-geostrophic dynamics. Journal of Fluid Mechanics, 282:1–20, 1995. doi: 10.1017/S0022112095000012.DOI
  15. [15]Heusel, M., Ramsauer, H., Unterthiner, T., Nessler, B., & Hochreiter, S. (2017). Gans trained by a two time-scale update rule converge to a local nash equilibrium. Proceedings of the 31st International Conference on Neural Information Processing Systems, NIPS’17, 6629–6640.
  16. [16]Tero Karras, Miika Aittala, Timo Aila, and Samuli Laine. Elucidating the design space of diffusion-based generative models. In S. Koyejo, S. Mohamed, A. Agarwal, D. Belgrave, K. Cho, and A. Oh (eds.), Advances in Neural Information Processing Systems, volume 35, pp. 26565–26577. Curran Associates, Inc., 2022. doi: 10.52202/068431-1926. URL https://proceedings.neurips.cc/paper_files/paper/2022/file/a98846e9d9cc01cfb87eb694d946ce6b-Paper-Conference.pdf.DOI
  17. [17]Georg Kohl, Li-Wei Chen, and Nils Thuerey. Benchmarking autoregressive conditional diffusion models for turbulent flow simulation. Neural Networks, 199:108641, 2026. doi: 10.1016/j.neunet.2026.108641.DOI
  18. [18]Lizao Li, Rob Carver, Ignacio Lopez-Gomez, Fei Sha, and John Anderson. Seeds: Emulation of weather forecast ensembles with diffusion models. arXiv preprint arXiv:2306.14066, 2023.arxiv.org/abs/2306.14066
  19. [19]Edward N. Lorenz. Deterministic Nonperiodic Flow. Journal of the Atmospheric Sciences, 20(2):130–148, March 1963. doi: 10.1175/1520-0469(1963)020⟨0130:DNF⟩2.0.CO;2.DOI
  20. [21]Ma, C., Wang, J., & E, W. (2019). Model reduction with memory and the machine learning of dynamical systems. Communications in Computational Physics, 25(4), 947–962. https://doi.org/10.4208/cicp.OA-2018-0269
  21. [22]Massimiliano Marcellino, James H. Stock, and Mark W. Watson. A comparison of direct and iterated multistep AR methods for forecasting macroeconomic time series. Journal of Econometrics, 135 (1–2):499–526, 2006.DOI
  22. [23]Nikolaj T. Mücke and Benjamin Sanderse. Physics-aware generative models for turbulent fluid flows through energy-consistent stochastic interpolants. Computers & Fluids, 314:107098, 2026. ISSN 0045-7930. doi: https://doi.org/10.1016/j.compfluid.2026.107098. URL https://www.sciencedirect.com/science/article/pii/S0045793026001404.
  23. [24]Charlie Nash, Jacob Menick, Sander Dieleman, and Peter Battaglia. Generating images with sparse representations. In Marina Meila and Tong Zhang (eds.), Proceedings of the 38th International Conference on Machine Learning, volume 139 of Proceedings of Machine Learning Research, pp. 7958–7964. PMLR, 18–24 Jul 2021. URL https://proceedings.mlr.press/v139/nash21a.html.proceedings.mlr.press/v139/nash21a.html
  24. [25]Tung Nguyen, Rohan Shah, Hritik Bansal, Troy Arcomano, Romit Maulik, Veerabhadra Kotamarthi, Ian Foster, Sandeep Madireddy, and Aditya Grover. Scaling transformer neural networks for skillful and reliable medium-range weather forecasting. In Advances in Neural Information Processing Systems, volume 37, pp. 68740–68771, 2024.arxiv.org/abs/2312.03876
  25. [26]Tim N Palmer, Andreas Döring, and Gregory Seregin. The real butterfly effect. Nonlinearity, 27(9):R123–R141, 2014.DOI
  26. [27]Gaurav Parmar, Richard Zhang, and Jun-Yan Zhu. On aliased resizing and surprising subtleties in gan evaluation, 2022. URL https://arxiv.org/abs/2104.11222.
  27. [28]Ilan Price, Alvaro Sanchez-Gonzalez, Ferran Alet, Tom R. Andersson, Andrew El-Kadi, Dominic Masters, Timo Ewalds, Jacklynn Stott, Shakir Mohamed, Peter Battaglia, Remi Lam, and Matthew Willson. Probabilistic weather forecasting with machine learning. Nature, 637(8044):84–90, 2025. ISSN 1476-4687. doi: 10.1038/s41586-024-08252-9. URL https://doi.org/10.1038/s41586-024-08252-9.
  28. [29]Tobias Rohner and Siddhartha Mishra. Efficient computation of large-scale statistical solutions to incompressible fluid flows. In Proceedings of the Platform for Advanced Scientific Computing Conference, PASC ’24, New York, NY, USA, 2024. Association for Computing Machinery. ISBN 9798400706394. doi: 10.1145/3659914.3659922. URL https://doi.org/10.1145/3659914.3659922.
  29. [30]François Rozet and Gilles Louppe. Score-based data assimilation, 2023. URL https://arxiv.org/abs/2306.10574.
  30. [31]Salva Rühling Cachay, Bo Zhao, Hailey Joren, and Rose Yu. DYffusion: A dynamics-informed diffusion model for spatiotemporal forecasting. In Advances in Neural Information Processing Systems, volume 36, 2023. URL https://openreview.net/forum?id=WRGldGm5Hz.openreview.net/forum?id=WRGldGm5Hz
  31. [32]Aliaksandra Shysheya, Cristiana Diaconu, Federico Bergamin, Paris Perdikaris, Jose Miguel Hernández-Lobato, Richard E. Turner, and Emile Mathieu. On conditional diffusion models for PDE simulations. In Advances in Neural Information Processing Systems, volume 37, 2024. URL https://openreview.net/forum?id=nQl8EjyMzh.openreview.net/forum?id=nQl8EjyMzh
  32. [33]Yang Song, Jascha Sohl-Dickstein, Diederik P. Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations, 2021. URL https://arxiv.org/abs/2011.13456.
  33. [34]Gábor J. Székely and Maria L. Rizzo. Energy statistics: A class of statistics based on distances. Journal of Statistical Planning and Inference, 143(8):1249–1272, 2013. ISSN 0378-3758. doi: https://doi.org/10.1016/j.jspi.2013.03.018. URL https://www.sciencedirect.com/science/article/pii/S0378375813000633.
  34. [35]L. Theis, A. van den Oord, and M. Bethge. A note on the evaluation of generative models. In International Conference on Learning Representations, Apr 2016. URL http://arxiv.org/abs/1511.01844.
  35. [36]David Zwicker. py-pde: A Python package for solving partial differential equations. Journal of Open Source Software, 5(48):2158, 2020. doi: 10.21105/joss.02158. URL https://doi.org/10.21105/joss.02158.

Paper details

vlnc:W_fkrk6x83
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 7

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Computer Science

Not recorded for this paper

View statement
Looking for related papers…
Contents