arXiv：cs.LG（机器学习，全量分类）· Joonhyeong Park, Giung Nam, Hyungi Lee, Kyunghyun Cho, Byoungwoo Park, Juho Lee·· 5 小时前AI 评分33
基于 Proper Scoring Rule 的扩散模型用于概率天气预报
Proper Scoring Rule-based Diffusion for Probabilistic Weather Forecasting
AI 导读
研究者提出一种基于 proper scoring rule 的扩散方法，通过辅助条件去噪任务让随机预测器从上下文及其损坏版本中预测同一未来状态，从而更有效地学习预测分布。该方法在推理时仍可在完全损坏端点单次前向生成每个集合成员，标准 CRPS 训练被恢复为其仅端点特例。受控实验显示辅助任务在多种架构上改善单步预报，预报时效越长增益越大，并扩展至全球高维天气预报的从头训练与微调。
正文
Abstract:Recent probabilistic weather forecasters train stochastic predictors with the continuous ranked probability score (CRPS) to generate each ensemble member in a single forward pass. These models learn the predictive distribution from the forecast context alone, which becomes difficult at longer forecast horizons where uncertainty is high. To learn the predictive distribution more effectively, we introduce auxiliary conditional denoising tasks that predict the same future state from the context and its corrupted version, which provides partial future information that can reduce prediction ambiguity. Building on distributional diffusion models, we learn the conditional distributions of these tasks with a single stochastic predictor by minimizing a proper scoring rule across noise levels. At inference, the predictor can still generate each ensemble member in a single forward pass at the fully corrupted endpoint. Standard CRPS training is recovered as the endpoint-only special case of our formulation, so our framework extends existing CRPS-based forecasters with only additional conditioning inputs. Controlled experiments show that the auxiliary tasks improve one-step forecasting across architectures, with larger gains at longer forecast horizons. The gains extend to high-dimensional global weather forecasting under both training from scratch and fine-tuning, along with improved calibration and potential benefits for generalization under distribution shift.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2609.38632 [cs.LG] |
| (or arXiv:2609.38632v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2609.38632 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Joonhyeong Park [view email]
[v1]
Tue, 29 Sep 2026 22:47:10 UTC (7,128 KB)
来源：arXiv：cs.LG（机器学习，全量分类） · arxiv.org