A1189
Title: Flexible hierarchical Bayesian process-informed neural network models for spatiotemporal data
Authors: Christopher Wikle - University of Missouri (United States) [presenting]
Abstract: Process (physics)-informed neural models have become ubiquitous across many areas of science in recent years due to the value of regularizing neural networks based on an underlying partial differential equation physical constraint. The purpose is to discuss a generalization in which mechanistic information about the process can be incorporated via a Bayesian hierarchical approach. In many scientific applications where there is substantial a priori process knowledge, incorporating this information can improve model performance and efficiency. The notion of including process knowledge in data-driven models for spatiotemporal data is not new (e.g., data assimilation, physical-statistical modeling, etc.), and considering hybrid statistical/neural approaches can provide more realistic modeling of complex processes while quantifying uncertainty. A brief overview of neural and statistical approaches is presented, and a unifying hierarchical modeling structure is presented that can accommodate flexible, mechanistically informed neural or statistical models for spatiotemporal dynamic processes.
