Why we should condition denoising diffusion generative models on windows of past observations

Abstract

Data assimilation (DA) is, traditionally, a cycling process that relies on time-dependent priors to propagate information from past observations to future cycles. Using denoising diffusion generative modeling for DA is challenging because standard approaches use a fixed training data set, which in turn leads to a static prior that ignores information from past observations. Because past observations are ignored, DA systems with static priors lead to larger posterior errors than cycling DA systems. Incorporating time-dependent priors into generative models, however, requires expensive and frequent retraining. Motivated by linear systems theory - where the dependence of a prediction of a Kalman filter on past observations decays exponentially - we condition diffusion models on short windows of past observations. Specifically, we describe training procedures for two frameworks: a diffusion DA system predicting the current state given a set of past observations, and a diffusion ``direct observation prediction'' (DOP) system, predicting future observations given a set of past observations. Using a canonical linear system, we show that both systems can achieve the minimal posterior error characteristic of a fully-cycled DA/DOP system, without re-training, provided the time windows are long enough. The linear setup ensures analytical tractability, avoids confounding neural network training errors, and confirms that conditioning on windows of past observations is required for efficient and accurate diffusion-based DA or DOP.

Connections between the Föllmer process and the denoising diffusion probabilistic model
Yuta Koike
arXiv2 Sep 2026PreprintarXiv licenceFull textProbability StatisticsarXiv
Exact Global MCMC with Denoising Diffusion
Mitch Hill
arXiv1 Sep 2026PreprintarXiv licenceFull textStatisticsarXiv
Denoising as Projection: Constrained Optimization with Gradient-Guided Diffusion
Runyu Zhang, Jiawei Zhang, Gioele Zardini +2
arXiv11 Sep 2026PreprintarXiv licenceFull textOptimisation & controlarXiv
Spatio-temporal Latent Denoising Diffusion Probabilistic Models for Reduced-order Modeling of Parametrized Dynamical Systems
Michiel Nikken, Nicolò Botteghi, Federico Califano +1
arXiv15 Sep 2026PreprintCC BY 4.0Full textNumerical analysis & scientific computingarXiv
RDDMPI: Residual Denoising Diffusion Model for Probabilistic Multivariate Time Series Imputation
Ramiro Valdes Jara, David Chapman, Adam Meyers
arXiv10 Sep 2026PreprintarXiv licenceFull textStatisticsarXiv

Cite

BibTeX RIS

BibTeX

@misc{morzfeld2026should,
  title = {{Why we should condition denoising diffusion generative models on windows of past observations}},
  author = {Matthias Morzfeld and Daniel Hodyss},
  howpublished = {arXiv},
  year = {2026},
  month = sep,
  doi = {10.48550/arxiv.2609.10924},
  eprint = {2609.10924},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.10924},
  note = {CC BY 4.0},
}

RIS

TY  - UNPB
TI  - Why we should condition denoising diffusion generative models on windows of past observations
AU  - Matthias Morzfeld
AU  - Daniel Hodyss
PY  - 2026
DA  - 2026/09/10
PB  - arXiv
DO  - 10.48550/arxiv.2609.10924
UR  - https://arxiv.org/abs/2609.10924
AB  - Data assimilation (DA) is, traditionally, a cycling process that relies on time-dependent priors to propagate information from past observations to future cycles. Using denoising diffusion generative modeling for DA is challenging because standard approaches use a fixed training data set, which in turn leads to a static prior that ignores information from past observations. Because past observations are ignored, DA systems with static priors lead to larger posterior errors than cycling DA systems. Incorporating time-dependent priors into generative models, however, requires expensive and frequent retraining. Motivated by linear systems theory - where the dependence of a prediction of a Kalman filter on past observations decays exponentially - we condition diffusion models on short windows of past observations. Specifically, we describe training procedures for two frameworks: a diffusion DA system predicting the current state given a set of past observations, and a diffusion ``direct observation prediction'' (DOP) system, predicting future observations given a set of past observations. Using a canonical linear system, we show that both systems can achieve the minimal posterior error characteristic of a fully-cycled DA/DOP system, without re-training, provided the time windows are long enough. The linear setup ensures analytical tractability, avoids confounding neural network training errors, and confirms that conditioning on windows of past observations is required for efficient and accurate diffusion-based DA or DOP.
ER  -

Abstract

Related

Cite