模型可自适应调整预报时间步，兼顾短时精度与长期稳定性。
Timestep-Conditioned Transformers for Global Weather Forecasting
- 通过多时间步推理实现动态时间步配置
- 在13400万参数下达到近顶尖中短期概率预报性能
- 适合需要灵活预报时间的气象决策场景
现有机器学习气象预报模型依赖预设且固定的自回归时间步。时间步选择存在根本权衡：短时间步（如1至6小时）能精细捕捉日变化大气动力学，但会增加给定预报周期内的误差累积；长时间步（如24小时）虽降低误差累积，却限制了短时预报的可用性。本文提出GEM-3，一种概率性全球气象模型，通过显式多时间步推理解决此权衡问题。仅需一组训练权重，模型可在推理时配置时间步，平衡不同预报周期下的可预测性与实用性。此外，混合时间步训练显著提升滚动推演稳定性，优于专用时间步模型。GEM-3为轻量级邻域注意力变压器，在等距圆柱投影网格上约1.34亿参数，相较前代GEM-2有若干架构改进。结果表明，该系统具备接近顶尖的中短期概率预报能力、稳定的超长周期推演、高效训练与推理，以及对决策有用的关键诊断信息。
原文摘要 · Abstract (English)
Existing machine-learning weather forecasting models rely on predetermined and fixed autoregressive timesteps. The choice of model timestep involves a fundamental trade-off: shorter timesteps (e.g. 1 to 6 hours) finely resolve atmospheric dynamics within the diurnal cycle but increase error accumulation for a given forecast horizon, while longer timesteps (e.g. 24 hours) reduce error accumulation but limit the usability of short-range forecasts where sub-daily predictability is high. In this work, we present GEM-3, a probabilistic global weather model that addresses this trade-off through explicit multi-timestep inference. With a single set of trained weights, the model timestep can be configured at inference time to balance predictability and usability across a broad forecast horizon. Additionally, we find that mixed-timestep training consistently improves rollout stability relative to timestep-specialist models. Under the hood, GEM-3 is a lightweight neighborhood-attention transformer with ~134M parameters on an equirectangular grid with a number of architectural advancements beyond its predecessor GEM-2. The result is a practical forecasting system that couples near-SOTA medium-range probabilistic skill, stable extended-range rollouts, efficient training and inference, and decision-relevant diagnostics.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。