Title: Spatially-explicit machine learning for geographic applications

Abstract: In this talk I will provide an overview on the topic of embedding spatial context and dependencies into neural network models to improve their performance when working with geographic data. I will first talk about how spatial dependencies can manifest in different real-world data and how this has been approached traditionally in different applied domains like ecology or epidemiology. I will then explore existing parametric and non-parametric embeddings capturing spatial and spatio-temporal dynamics and will discuss how these may be leveraged for predictive and generative neural network models. Moving from methods to applications, I will highlight how geospatial machine learning can be deployed to tackle problems in urban science and Earth observation. Lastly, I will outline the intersection between the geospatial domain and large (foundation) models, including the challenges and opportunities of these models.

About this Series: The Atmospheres, Ocean and Climate Sack Lunch Seminar Series is an informal seminar series within PAOC that focuses on more specialized topics than the PAOC Colloquium. Seminar topics include all research concerning the science of atmospheres, ocean and climate. The seminars usually take place on Wednesdays from 12-1pm. The presentations are either given by an invited speaker or by a member of PAOC and can focus on new research or discussion of a paper of particular interest. Contact: sacklunch-committee@mit.edu for more information and Zoom password