Overview
Syllabus
Intro
Awesome reduced order model team and collaborators
Physical simulations play an important role in modern scienc
How does conditional generative adversarial network perform?
Pro and cons of black-box approach
How can we get an interpretability?
DMD accelerates 3D printing process simulation
Time-windowing Wavelet DMD improves accuracy
Are there other data-driven interpretable methods?
Parameterized latent space dynamics identification (LaSDI)
Performance of LaSDI to radial advection problem
gLaSDI: physics-informed greedy latent space dynamics identificat
How about physics-constrained model?
Projection-based linear subspace reduced order model
Space-time ROM achieves the maximal compression
Component-wise ROM accelerates lattice-structure design optir
PROM accelerates wind turbine blade design optimization
Database local ROMs accelerate multi-start airplane wing optin
Category of data-driven methods via level of intrusiveness
Taught by
DataLearning@ICL