This groundbreaking talk dives into the intriguing intersections between geospatial data science and the high-dimensional embeddings generated by large language models (LLMs) like ChatGPT. Just as geospatial experts map the physical world, data scientists use embeddings to chart abstract spaces, creating a unique opportunity for geospatial professionals to leverage their expertise in the emerging field of explainable AI.
Through visually engaging examples, the talk will address questions such as:
- How is embedding similar to surveying, and what does “distance” signify in abstract space?
- How do dimensionality reduction techniques and clustering enable cartographic representations of conceptual spaces?
- Can foundation models serve as a “GPS” for meaning and context?
By drawing parallels between geospatial and AI methodologies, attendees will gain new perspectives on their potential role in this burgeoning, interdisciplinary domain.