Model-Constrained Deep Learning Approaches for Scientific Machine Learning

Explore model-constrained deep learning approaches in this comprehensive lecture by Dr. Tan Bui, an associate professor at the University of Texas at Austin. Delve into the emerging field of scientific machine learning, which combines traditional scientific computing with modern machine learning techniques. Learn about the advantages and challenges of both approaches, and discover how scientific machine learning aims to develop explainable, data-driven models that require less data than traditional methods. Examine two specific efforts in model-aware machine learning: the ROM-ML approach and the Autoencoder-based Inversion (AI) approach. Gain insights into theoretical results for linear PDE-constrained inverse problems and observe numerical results for various nonlinear PDE-constrained inverse problems, demonstrating the validity of these innovative approaches. This 1-hour 18-minute talk provides a deep dive into the cutting-edge research at the intersection of computational science, engineering, and mathematics.

Model-constrained deep learning approaches