Predicting partially observable dynamical systems via diffusion models with a multiscale inference scheme

Conditional diffusion models provide a natural framework for probabilistic predic tion of dynamical systems and have been successfully applied to fluid dynamics and weather prediction. However, in many settings, the available information at a given time represents only a small fraction of what is needed to predict future states, either due to measurement uncertainty or because only a small fraction of the state can be observed. This is true for example in solar physics, where we can observe the Sun’s surface and atmosphere, but its evolution is driven by internal processes for which we lack direct measurements. In this paper, we tackle the probabilistic prediction of partially observable, long-memory dynamical systems, with applications to solar dynamics and the evolution of active regions. We show that standard inference schemes, such as autoregressive rollouts, fail to capture long-range dependencies in the data, largely because they do not integrate past information effectively. To overcome this, we propose a multiscale inference scheme for diffusion models, tailored to physical processes. Our method generates trajectories that are temporally fine-grained near the present and coarser as we move farther away, which enables capturing long-range temporal dependencies without increasing computational cost. When integrated into a diffusion model, we show that our inference scheme significantly reduces the bias of the predicted distributions and improves rollout stability.

Keywords

Subject (DDC)

520 - Astronomie, Kartografie

Project

Event

39th Conference on Neural Information Processing Systems (NeurIPS 2025)

Exhibition start date

Exhibition end date

Conference start date

02.12.2025

Conference end date

07.12.2025

Date of the last check

ISBN

979-8-331-33827-5

ISSN

Language

English

Created during FHNW affiliation

Yes

Strategic action fields FHNW

Publication status

Published

Review

peer-reviewed

Open access category

Closed

License

Citation

Morel, R., Ramunno, F., Shen, J., Bietti, A., Cho, K., Cranmer, M., Golkar, S., Gugnin, O., Krawezik, G., Marwah, T., McCabe, M., Meyer, L., Mukhopadhyay, P., Ohana, R., Parker, L., Qu, H., Rozet, F., Leka, K. D., Lanusse, F., et al. (2025). Predicting partially observable dynamical systems via diffusion models with a multiscale inference scheme. Advances in Neural Information Processing Systems 38. 39th Conference on Neural Information Processing Systems (NeurIPS 2025), 74387–74405. https://doi.org/10.52202/085713-2240

Full item page