UDL-RT, a new microwave radiative transfer model that uses machine learning to simulate how satellite signals interact with the atmosphere and surface.
UDL-RT: A unified differentiable all-sky microwave radiative transfer model with learnable parameterizations · Zeting Li, Wei Han, Jie Peng et al. · arXiv:2610.24627
An AI model that can simulate complex atmospheric chemistry and climate interactions, like how emissions affect air quality and radiation, about 100 times faster than traditional methods.
Atmospheric Chemistry-Climate Modeling with an AI Foundation Model · James Weber, Makoto Kelp · arXiv:2610.22555
A look at whether new AI weather models can accurately forecast extreme events like tropical cyclones, comparing their performance against established physics-based systems.
Can AI Forecast the Extremes? Evaluating Probabilistic Weather Models on Tropical Cyclones · Nathan Chalumeau, Xavier Renard, Remi Meynadier et al. · arXiv:2610.22077
A critical challenge for self-driving cars: enabling them to reliably understand their surroundings in all weather, even when cameras and laser sensors fail. A novel system that uses radar, which excels in adverse conditions, to identify and categorize objects in real-time.
Real-Time Radar Semantic and Panoptic Scene Understanding · Santosh Parit, Julia Hindel, Rohit Mohan et al. · arXiv:2610.26816
This paper uses a novel AI method to determine how accurately we can map Greenland's melting ice by analyzing the distinct sea level changes it produces, known as "sea level fingerprints."
Neural Operators Reveal Scale-Dependent Limits on Recovering Greenland Ice Loss from Sea Level Fingerprints · Xiyuan Bao, Sophie Coulson, Natasha Valencic et al. · arXiv:2610.23748
AirMPA, an advanced AI model that significantly improves global air quality forecasting by uniquely separating the prediction of weather conditions from the prediction of atmospheric pollutants.
AirMPA: A Meteorology-to-Pollution Adapter for Global Air Quality Forecasting · Yiheng Wang, Kai Zheng, Yuetan Lin et al. · arXiv:2610.25797
UniPolar, a new technique for cleaning up images from specialized cameras that capture light's polarization, which is vital for seeing details invisible to regular cameras.
UniPolar: Joint Image and Polarimetric Restoration under Compound Degradations via Physical Consistency Learning · Yang Jin, Guangyu Guo, Linghao Shen et al. · arXiv:2610.23360
A deep learning method to automatically clean up complex subsurface imaging data by removing unwanted 'echoes' that can obscure the real features. This approach to robust, physics-informed signal extraction offers valuable lessons for enhancing data quality across various remote sensing and Earth observation applications.
Physics-based deep learning for non-stationary adaptive multiple suppression · Chao Li, Yangkang Chen · arXiv:2610.19147
A look at how microwave imaging, which uses radio waves to "see" inside the body, can reliably monitor breast tumor changes during chemotherapy. The critical challenge of separating actual tumor responses from much larger signal variations caused by patient movement or normal tissue changes.
Sensitivity and Identifiability of Tumor-Related Changes in Longitudinal Microwave Breast Imaging · Somayeh B. Shafiei, Ian Jeffrey, Vahab Khoshdel et al. · arXiv:2610.26630
A new method to capture and process incredibly detailed images of the Sun's outer atmosphere, the corona, using observations from the recent total solar eclipse. By measuring polarized light and employing advanced techniques to combine many exposures and precisely align the data, it reveals unprecedented fine structures of the Sun's magnetic field.
Linear Polarimetry of the Solar Corona during the Total Solar Eclipse of 8 April 2024: Multi-Band Observations and HDR Imaging · Suprit Singh, Paras Sharma, Bharti Arora et al. · arXiv:2610.26231