Xaurora: Generative Weather Forecasting with Denoising Stochastic Interpolants from a Foundation Model Prior
cs.LG, physics.ao-ph
Submitted: 2026-10-05
Updated: 2026-10-05
Code: https://github.com/google-research/weatherbenchX
Project page: https://microsoft.github.io/aurora/finetuning.html#exploding-gradients
Terminology
Sources
- WIND: Weather Inverse Diffusion for Zero-Shot Atmospheric Modeling
- Building Normalizing Flows with Stochastic Interpolants
- Stochastic Interpolants: A Unifying Framework for Flows and Diffusions
- Skillful joint probabilistic weather forecasting from marginals
- Flow map matching with stochastic interpolants: A mathematical framework for consistency models
- U-Cast: A Surprisingly Simple and Efficient Frontier Probabilistic AI Weather Forecaster
- FlowDAS: A Stochastic Interpolant-based Framework for Data Assimilation
- Probabilistic Forecasting with Stochastic Interpolants and F\"ollmer Processes
- Gradient-Free Generation for Hard-Constrained Systems
- ArchesWeather & ArchesWeatherGen: a deterministic and generative model for efficient ML weather forecasting
- An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
- Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning
- Improving neural networks by preventing co-adaptation of feature detectors
- Denoising Diffusion Probabilistic Models
- LoRA: Low-Rank Adaptation of Large Language Models
- Deep Networks with Stochastic Depth
- CoCoGen: Physically-Consistent and Conditioned Score-based Generative Models for Forward and Inverse Problems
- Scaling Laws for Neural Language Models
- Elucidating the Design Space of Diffusion-Based Generative Models
- Neural General Circulation Models for Weather and Climate
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks