时空知识融合：一种面向大气时间序列预测的轻量级方法
On the Integration of Spatial-Temporal Knowledge: A Lightweight Approach to Atmospheric Time Series Forecasting
- State Key Laboratory of AI Safety, Institute of Computing Technology, Chinese Academy of Sciences（人工智能安全国家重点实验室，计算技术研究所，中国科学院）
- University of Chinese Academy of Sciences（中国科学院大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
本文提出轻量级模型 STELLA，仅用时空位置嵌入和MLP替代Transformer，以1万参数和1小时训练在5个数据集上超越先进方法，证明时空知识整合优于复杂架构。
AI中文摘要：
Transformer 因其捕捉全局时空相关性的能力，在大气时间序列预测（ATSF）中受到关注。然而，其复杂的架构导致参数数量过多和训练时间延长，限制了其在大规模预测中的可扩展性。在本文中，我们从大气动力学的理论视角重新审视 ATSF，并揭示了一个关键见解：即使没有注意力机制，时空位置嵌入（STPE）也能固有地建模时空相关性。其有效性源于地理坐标和时间特征的整合，这些特征与大气动力学内在相关。基于此，我们提出了 STELLA，一种用于 ATSF 的时空知识嵌入轻量级模型，仅使用 STPE 和 MLP 架构替代 Transformer 层。凭借 1 万参数和一小时训练，STELLA 在五个数据集上相较于其他先进方法取得了优越性能。本文强调了时空知识整合相对于复杂架构的有效性，为 ATSF 提供了新颖见解。代码可在 https://github.com/GestaltCogTeam/STELLA 获取。
英文摘要：
Transformers have gained attention in atmospheric time series forecasting (ATSF) for their ability to capture global spatial-temporal correlations. However, their complex architectures lead to excessive parameter counts and extended training times, limiting their scalability to large-scale forecasting. In this paper, we revisit ATSF from a theoretical perspective of atmospheric dynamics and uncover a key insight: spatial-temporal position embedding (STPE) can inherently model spatial-temporal correlations even without attention mechanisms. Its effectiveness arises from the integration of geographical coordinates and temporal features, which are intrinsically linked to atmospheric dynamics. Based on this, we propose STELLA, a Spatial-Temporal knowledge Embedded Lightweight modeL for ASTF, utilizing only STPE and an MLP architecture in place of Transformer layers. With 10k parameters and one hour of training, STELLA achieves superior performance on five datasets compared to other advanced methods. The paper emphasizes the effectiveness of spatial-temporal knowledge integration over complex architectures, providing novel insights for ATSF. The code is available at https://github.com/GestaltCogTeam/STELLA.
