[PAOCQ] Andre Souza (MIT)

“Score-Based Generative Modeling for Geophysical Inference and Response Prediction”

This talk introduces a distribution-first approach to geophysical prediction using score-based generative models. I will demonstrate uncertainty-aware reconstruction of the ocean interior from surface observations, as well as forced-response prediction in simple dynamical systems using the Generalized Fluctuation–Dissipation Theorem. Both capabilities arise from learning the score function, which represents the geometry of the underlying probability distribution. The method avoids perturbed-ensemble simulations and natively quantifies uncertainty. These results point toward fast and scalable statistical dynamical inference for geophysical systems.

PAOC Colloquium —

Interdisciplinary seminar series that brings together the whole PAOC (Program in Atmospheres, Oceans, and Climate) community. Seminar topics include all research concerning the physics, chemistry, and biology of the atmospheres, oceans and climate, as well as talks about societal impacts of climatic processes.

Contact: paoc-colloquium-comm@mit.edu