Model Scale Versus Domain Knowledge in Long-Term Forecasting of Chaos
45:04Operator Learning in Theory and Practice
49:23Science Education for AI: Beyond AI in Science Education
40:19From Polar Ice to Rising Seas: Advancing Climate Science with Machine Learning
43:02From Scheduled to Predictive Maintenance in Hydropower using AI
58:04Data-Driven Approaches to Mitigate Chalk Influxes
39:43Physical Insights from Discriminative and Generative Models for Frictional Surface Design
40:32Model Scale Versus Domain Knowledge in Long-Term Forecasting of Chaos
William Gilpin explores the interplay between model complexity and foundational domain understanding for robust long-term prediction in chaotic natural systems.
William Gilpin (UT Austin) - Model scale versus domain knowledge in long-term forecasting of chaos
Welcome to a conference on discovery and prediction in complex geo systems. AI/ML shows promise for accelerating Bayesian inversions, is able to capture and identify controlling features of high fidelity models, and appears to improve complex system analysis, including extending predictive horizons, for reasons that are not fully understood. This conference focuses on efforts to advance the use of AI/ML to understand physical processes in the geosciences and adjacent fields.