← Back to all news
arXiv Machine Learning By Pavel Prochazka

Forecasting Multiple Observables with SCROLL: Score-Trained Uncertainty for Stochastic Dynamics

Read the original on arXiv Machine Learning →

The paper introduces SCROLL, a method for forecasting multiple observables in stochastic dynamical systems by composing each observable’s likelihood into per‑task free‑routed last‑layer beliefs on a shared backbone. This approach learns unit‑dependent loss scaling directly from data, enabling accurate predictive variance estimation without separate tuning. Experiments on the Ornstein–Uhlenbeck process, stochastic Lorenz‑63, and real air‑quality data show that SCROLL recovers analytic kernels, achieves superior negative log‑likelihood on state and regime tasks, and maintains calibration while reducing hyper‑parameter search costs.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

Related stories

arXiv Computer Vision
Oct 16

DIAL: Dynamic and Informed Adjustment of the Native Stochastic Law for Reliable Weather Ensembles

The paper introduces DIAL, a method that dynamically adjusts the native stochasticity of a frozen weather forecasting model to improve ensemble reliability across variables, scales, and lead times. DIAL learns from past forecasts to predict necessary spread adjustments and applies small noise updates during each forecast step, avoiding retraining or full backpropagation. Experiments on FourCastNet 3 show that DIAL outperforms post‑processing and reward‑based baselines, achieving an 8.27% improvement in multiscale CRPS over the unmodified model.

By Yucheng Wang, Haonan Guo, Tianhao Zhao, Di Wang, Yu Wu, Zengmao Wang, Bo Du
More like this →
arXiv Machine Learning
Oct 2

Proper Scoring Rule-based Diffusion for Probabilistic Weather Forecasting

The paper introduces a new diffusion-based approach for probabilistic weather forecasting that incorporates auxiliary conditional denoising tasks. By predicting future states from both clean and corrupted contexts, the model reduces ambiguity and learns conditional distributions across noise levels using a proper scoring rule. Experiments demonstrate improved one‑step forecasting, especially at longer horizons, and better calibration in high‑dimensional global weather scenarios.

By Joonhyeong Park, Giung Nam, Hyungi Lee, Kyunghyun Cho, Byoungwoo Park, Juho Lee
diffusionfine-tuning
More like this →