Research Seminar on Geospatial AI

1 unit

This seminar will study and discuss recent research on geospatial AI and spatial intelligence: machine learning methods for problems with a meaningful spatial component. We will cover algorithmic themes such as representation learning, neural fields, world models and digital twins, generalization and domain adaptation, and multimodal fusion, applied to data including satellite and aerial imagery, ground-based and mobile sensors, and GPS trajectories, with applications spanning environmental and public health monitoring to field robotics and autonomous driving. Papers will be drawn from top computer vision and machine learning venues (e.g., CVPR, ICCV/ECCV, NeurIPS, ICML, ICLR, AAAI).