A Physics-Based Reduced Order Model Capturing the Topology of Dynamical Manifolds
Explore a transformational approach to dynamics inspired by Mallat's Scattering Transformation in this 1-hour 23-minute talk by Michael E. Glinsky, CEO of BNZ Energy Inc. Delve into the Heisenberg Scattering Transformation, a method based on a canonical approach that projects dynamics onto the Renormalization Basis. Learn how this technique constrains dynamics to a low-dimensional complex linear subspace and quantifies topology using a Multi-Layer Perceptron. Discover the application of this approach in creating a fast surrogate model for 2D pulsed power liner implosions (MagLIF). Gain insights from Glinsky's 30-year career in data science, deep learning, and Bayesian data analysis, and understand how this innovative method can be applied to make better business decisions in various industries.