TerraFlow提升遥感多模态时序学习，助力灾害风险预测
TerraFlow: Multimodal, Multitemporal Representation Learning for Earth Observation
- 基于时序目标实现跨时空模态的序列感知学习
- 在GEO-Bench-2上F1最高提升50%，Brier分数降低24%
- 对真实长序列数据鲁棒，适合灾害预测等实际应用
我们提出TerraFlow，一种面向地球观测的新型多模态、多时相表示学习方法。TerraFlow基于时间训练目标，实现空间、时间与模态间的序列感知学习，同时对真实地球观测数据中常见的变长输入保持鲁棒性。实验表明，TerraFlow在GEO-Bench-2基准的所有时序任务上均优于现有顶尖地球观测基础模型。此外，TerraFlow首次展示了深度学习在自然灾害风险图预测上的初步能力——这一任务常导致其他先进模型失效。TerraFlow在F1分数上最高提升50%，Brier分数降低24%。
原文摘要 · Abstract (English)
We propose TerraFlow, a novel approach to multimodal, multitemporal learning for Earth observation. TerraFlow builds on temporal training objectives that enable sequence-aware learning across space, time, and modality, while remaining robust to the variable-length inputs commonly encountered in real-world Earth observation data. Our experiments demonstrate superiority of TerraFlow over state-of-the-art foundation models for Earth observation across all temporal tasks of the GEO-Bench-2 benchmark. We additionally demonstrate that TerraFlow is able to make initial steps towards deep-learning based risk map prediction for natural disasters -- a task on which other state-of-the-art foundation models frequently collapse. TerraFlow outperforms state-of-the-art foundation models by up to 50% in F1 score and 24% in Brier score.
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。