Temporal Generative Flow Networks (Temporal GFNs) offer a novel approach to probabilistic time series forecasting by adapting GFN principles for continuous data.

We tackle prediction of partially observable systems with a multiscale inference scheme for diffusion models, improving accuracy and stability on solar dynamics without increasing computational cost.

A sequence-level EDM-derived diffusion model using progressively increasing noise levels is designed for probabilistic forecasting of complex dynamics like weather.