GRI Guest: Hamed Alemohammad

Hamed Alemohammad, director of the Clark University Center for Geospatial Analytics,will present:

Advancing Foundation Models for Geospatial Applications

GeoAI is transforming how we consume geospatial data, what insights we drive from them, and ultimately how we interact with the world around us. This transformation is being expedited by the rapid growth in developing AI foundation models for Earth observations (EO) including multi-modal models that combine inputs from various sensors in Earth orbit. Following more than a decade of research and development in building and deploying supervised and semi-supervised techniques for EO, geospatial foundation models are now changing the paradigm to learn geospatial patterns from raw (aka unlabeled) observations.

In this presentation, I will review the current state of GeoAI by providing an overview of the applications at the intersection of AI and EO. Next, I will showcase how geospatial foundation models are changing the paradigm for GeoAI applications. I will discuss the state-of-the-art geospatial foundation models and strategies for their design and validation. I will close with sharing some thoughts on future direction of this field, and priorities for advancing geospatial foundation models.