EarthAlign: Learning Language-Aligned Representations of Earth System Data
cs.CV, cs.AI, cs.LG
Submitted: 2026-10-10
Updated: 2026-10-10
Code: https://github.com/mbzuai-oryx/EarthAlign
Sources
- Qwen3-VL Technical Report
- UniMed-CLIP: Towards a Unified Image-Text Pretraining Paradigm for Diverse Medical Imaging Modalities
- AtmoRep: A stochastic model of atmosphere dynamics using large scale representation learning
- WeatherQA: Can Multimodal Language Models Reason about Severe Weather?
- ClimaX: A foundation model for weather and climate
- FourCastNet: A Global Data-driven High-resolution Weather Model using Adaptive Fourier Neural Operators
- Prithvi WxC: Foundation Model for Weather and Climate
- MedGemma Technical Report
- InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models