Completed
Diverse future of computational science and programming languages
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Scientific Machine Learning: Opportunities and Challenges - Keynote
Automatically move to the next video in the Classroom when playback concludes
- 1 Welcome and information about JuliaCon 2020
- 2 Introduction and acknowledgments
- 3 Outline of the talk
- 4 What is Scientific Machine Learning (SciML)?
- 5 What are the opportunities and challenges of SciML?
- 6 BIG DATA alone is not enough
- 7 Using physics base models means that we are doing Computational Science
- 8 Problem 1: Complex multiscale multiphysics phenomena
- 9 Problem 2: High dimensional parameters
- 10 Problem 3: Data are sparse, intrusive and expensive to acquire
- 11 Problem 4: Rare events
- 12 Problem 5: Uncertainty qualification
- 13 SciML and Computational Science, summary
- 14 Example: flow inside a rocket engine combustor
- 15 Example: equations of flow inside a rocket engine combustor
- 16 Physics-based model are powerful but computationally expensive
- 17 Overview of model reduction methods
- 18 Similarities and differences between model reduction and ML
- 19 Can we have the best of two worlds (model reduction and ML)?
- 20 Overview of Lift & Learn approach
- 21 Example: Lift & Learn approach to 2D flow in a rocket engine
- 22 Example: Training of the model
- 23 Example: Comparing results for pressure and temperature
- 24 Outlook of SciML
- 25 Diverse future of computational science and programming languages
- 26 Q&A: Advice for people bringing ML approach to scientific problems
- 27 Q&A: Pitfalls of the interplay between domains knowledge and ML
- 28 Q&A: Does the present approach improve the fidelity of solution of highly non-linear systems?
