Decision-Oriented Uncertainty Quantification for Risk Control in Earth System Spatiotemporal Foundation Models
Read the original on arXiv Machine Learning →The Flow has not summarised this story yet — read it at arXiv Machine Learning.
The Flow has not summarised this story yet — read it at arXiv Machine Learning.
arXiv:2607. 07758v1 Announce Type: new Abstract: Foundation models (FMs) have transformed machine learning from isolated task-specific model development toward general-purpose models pretrained on broad data and adapted to multiple downstream tasks.
arXiv:2509. 25017v2 Announce Type: replace Abstract: Wildfires are among the most severe natural hazards, posing a significant threat to both humans and natural ecosystems.
The paper presents a modular, scalable framework for estimating the probability of failure (PoF) of power‑line assets using geospatial machine learning. It integrates diverse environmental predictors—topography, vegetation indices, lightning climatology, proximity features, and operational records—to model vegetation‑ and lightning‑related failure modes. The architecture is designed to be computationally efficient, easily extensible to new data sources, and suitable for large‑scale utility deployment, enabling asset‑level risk stratification for inspection and resilience planning.
arXiv:2606. 27277v1 Announce Type: new Abstract: Earth Observation (EO) forecasting aims to predict future Earth surface dynamics from satellite observations under changing meteorological conditions.
arXiv:2608.24665v1 Announce Type: new Abstract: Power outage prediction models are increasingly used in assessments of climate-driven infrastructure risk, yet current evaluation practices obscure whe...
arXiv:2608. 00012v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) are increasingly used to interpret Earth observation data, yet their capability to support real-world disaster emergency response remains insufficiently evaluated.