Scientific Computing III: Data-Driven and Machine Learning Methods
This course introduces graduate students to contemporary numerical methods and techniques for data generation and analysis. The course program includes numerical approximation theory, neural-network-based methods for solving PDEs and inverse problems, neural operators for parametric PDEs, methods for dimensional reduction, including diffusion maps and autoencoders, generative models, and graph data analysis (if time allows).
Prerequisites & course information
Recommended: AMSC660 or AMSC661 . Cross-listed with: AMSC665 . Credit Only Granted for: AMSC808N, CMSC828V, AMSC665 or CMSC665 . Formerly: AMSC808N and CMSC828V. Additional Information: Students are assumed to have background in Real Analysis (MATH 410), probability theory (STAT 410), and Partial Differential Equations ( MATH462 ). If these courses (or their equivalent) have not been taken, ask for the instructor's permission.
Sources
- Course catalogue ↗2026-2027