About
Independent research at the frontiers of geospatial science.
Mosab Hawarey
Director of Geospatial Research, an independent virtual research organization focused on theoretical foundations of in-context learning, geospatial artificial intelligence, quantum geodesy, and geospatial intelligence. My work establishes rigorous mathematical frameworks characterizing when and why transformer architectures succeed on structured scientific problems.
Current research centers on the ICL-Easy/ICL-Hard dichotomy through the SSC/AC-SSC (Sufficient Statistic Complexity/Attention-Computable SSC) framework. I develop complete dichotomy theorems across geospatial domains—climate prediction, remote sensing, multi-modal foundation models, navigation systems, and intelligence analysis—with formal proofs, tight sample complexity bounds, and systematic quality assessment methodologies.
Research Philosophy
I believe that rigorous theory provides the foundation for impactful practice. By establishing precise characterizations—necessary and sufficient conditions, matching upper and lower bounds, complete dichotomy theorems—we gain principled understanding that guides application and reveals fundamental limits.
My approach emphasizes modularity, reproducibility, and systematic quality assessment. Complex research questions are decomposed into self-contained modules with clear dependencies, developed through phased execution with continuous quality evaluation. This methodology produces work that meets the highest standards of academic scholarship.
I have a purpose: to elevate the knowledge of humanity, no matter what.
Geospatial Research
Geospatial Research is an independent virtual research organization dedicated to advancing the theoretical foundations of in-context learning, geospatial artificial intelligence, quantum geodesy, and geospatial intelligence. We operate at the intersection of theoretical machine learning and geospatial science—developing rigorous frameworks that characterize computational efficiency of transformer architectures across climate models, remote sensing, navigation systems, intelligence analysis, and quantum-enhanced geodesy.
Our work advances the entanglements of geospatial artificial intelligence, quantum geodesy, and geospatial intelligence through mathematical precision, complete dichotomy theorems, and explicit acknowledgment of assumptions and limitations. We publish in peer-reviewed venues and make our research accessible to the broader scientific community.