Talk details

In schedule:
Reinforce Stage
October 6, 15:40 - 16:25 CET
Foundation Models in the Geospatial Domain: Challenges, Opportunities and Applications
AI Conference
machine learning
geospatial data science
foundation models
large language models
computer vision
location analytics
Level: General

A large share of all globally available data can in some shape or form be geo-referenced; mapped onto the sphere of planet Earth. This can be leveraged to train large "foundation" models that fuse different layers of spatial information and learn meaningful representations of every location on the planet. Such models will be transformative for private and public sector alike. From visionary moonshots like planetary digital twins, to high-impact commercial applications in the insurance sector, Geospatial Foundation Models promise to be a general-purpose toolbox—similar to Large Language Models (LLMs). This talk will explain the essential concepts needed to understand Geospatial Foundation Models and their potential applications, outline challenges for building these models and compare them to existing foundation models such as LLMs and large vision models (e.g. Segment Anything).

Reinforce 2023 - Konstantin Klemmer
Konstantin Klemmer
Postdoctoral Researcher at Microsoft
Reinforce 2023 - Microsoft

I am a postdoctoral researcher at Microsoft Research New England. My research focuses on the representation of geospatial phenomena in machine learning methods. I am also interested in the deployment of these methods in urban environments and to tackle climate change within both private and public sector organizations. In currently serve as a member of the board for Climate Change AI, a global non...