ScientificComputingCWI / SemesterProgramme-SciML

Material for workshop and autumn school on scientific machine learning 2023

Workshop, December 2023

Below you can find the presentations during the workshop on scientific machine learning, 6-8 December 2023, CWI Amsterdam.

Wednesday 6 December, Scientific computing with machine learning

Machine-learning-enhanced numerical methods, physics-informed machine learning

• Jan Hesthaven (EPFL): slides
• Xiaodong Cheng (Wageningen): slides
• Benjamin Sanderse (CWI): slides
• Stefania Fresca (Milan)
• Richard Dwight (Delft): slides
• Dongwei Ye (Twente): slides

Thursday 7 December, Scientific computing for machine learning

Mathematical and computational theories for machine learning

• Sid Mishra (EPFL): slides
• Aron Jansen (eScience center): slides
• Silke Glas (Twente): slides
• Andrea Walther (Berlin): slides
• Matthias Möller (Delft): slides

Friday 8 December, Scientific machine learning in applications

Hybrid modeling and simulation for large-scale systems

• Dirk Hartmann (Siemens): slides
• Quercus Hernández Laín (Zaragoza): slides

---

Autumn school, October 2023

Course material for the autumn school in scientific machine learning, 9-13 October 2023, CWI Amsterdam.

Day 1: Closure models and turbulence

• Andrea Beck: slides
• Romit Maulik: slides
• Syver Agdestein, Benjamin Sanderse:
  • slides
  • code

Day 2: Reduced-order models

• Andrea Manzoni & Mengwu Guo, see Mengwu's website
• Karen Veroy: slides

Day 3: SINDy

• Steve Brunton: slides
• Urban Fasel:
  • Matlab code and slides
  • Python code PySINDy

Day 4: Operator and equivariant learning

• Paris Perdikaris: slides
• Erik Bekkers: slides and code

Day 5: Differentiable simulators

• Chris Rackauckas slides
  • Demonstration that standard uncorrected AD can produced incorrect results on ODE codes code
  • Demonstration code for how PDEBench's Burgers implementation is >100x slower than a proper classical inverse problem simulator and the optimization process code
• Hod Lipson: slides
• Benjamin Sanderse: wrap-up