Sign Up

Antonis Anastasopoulos

In this talk, I will first cover our group's efforts to measure the geospatial knowledge already encoded in large language models trained on large amounts of text, discussing different ways to extract and quantify it. In the second part, I will discuss our work towards building geospatial-native foundation models inspired by the incredible success of language models, such as trajectory models. We will discuss applications of such models on tasks such as trajectory anomaly detection, image localization, and urban mobility simulation.

This talk is co-sponsored by AI-ACCESS.