OlmoEarth：面向多模态地球观测的稳定潜空间图像建模
OlmoEarth: Stable Latent Image Modeling for Multimodal Earth Observation
面向多模态时空遥感数据设计的自监督掩码策略与损失，可作为把掩码图像建模从静态 RGB 扩展到多时相多传感器输入的参考；其嵌入与全微调两套评测协议也值得在遥感下游任务中复用。
原文依据 a multimodal, spatio-temporal foundation model that employs a novel self-supervised learning formulation, masking strategy, and loss
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论地球观测数据带来一个独特挑战：它像图像一样具有空间性，像视频或文本一样具有序列性，并且高度多模态。我们提出 Helios：一个多模态时空基础模型，采用专为地球观测领域设计的新颖自监督学习形式、掩码策略与损失函数。在多项研究基准和来自外部合作方的真实世界任务上，Helios 相比其他 12 个基础模型取得了最优性能。在嵌入评估中，Helios 在 24 个任务中的 15 个上取得最佳表现；在完整微调设置下，它在 29 个任务中的 19 个上表现最佳。我们将 Helios 部署为一个端到端平台的骨干网络，该平台覆盖地球观测模型的数据采集、标注、训练与推理。Helios 平台把前沿基础模型与强大的数据管理工具交到致力于解决全球重大问题的非营利组织与非政府组织手中。Helios 的源代码、训练数据与预训练权重已在 REDACTED 处公开。
英文原文摘要 1049 字符
Earth observation data presents a unique challenge: it is spatial like images, sequential like video or text, and highly multimodal. We present Helios: a multimodal, spatio-temporal foundation model that employs a novel self-supervised learning formulation, masking strategy, and loss all designed for the Earth observation domain. Helios achieves state-of-the-art performance compared to 12 other foundation models across a variety of research benchmarks and real-world tasks from external partners. When evaluating embeddings Helios achieves the best performance on 15 out of 24 tasks, and with full fine-tuning it is the best on 19 of 29 tasks. We deploy Helios as the backbone of an end-to-end platform for data collection, labeling, training, and inference of Earth observation models. The Helios platform puts frontier foundation models and powerful data management tools into the hands of non-profits and NGOs working to solve the world's biggest problems. Helios source code, training data, and pre-trained weights are available at REDACTED.