Scientific Machine Learning
Details
"Computational scientific discovery is at an interesting juncture. While we have mechanistic models of lots of different scientific phenomena, and reams of data being generated from experiments - our computational capabilities are unable to keep up. Our problems are too large for realistic simulation. Our problems are multi-scale and too stiff. Our problems require tedious work like calculating gradients and getting code to run on GPUs and supercomputers.
While traditional deep learning methodologies have had difficulties with scientific issues like stiffness, interpretability, and enforcing physical constraints, this blend with numerical analysis and differential equations has evolved into a field of research with new methods, architectures, and algorithms that overcome these problems while adding the data-driven automatic learning features of modern deep learning.
The next step forward is a combination of science and machine learning, which combines mechanistic models with data-based reasoning, presented as a unified set of abstractions and a high-performance implementation." [1]
**********************************************************************************
Bogdan Burlacu is a senior researcher at the University of Applied Sciences Upper Austria where he develops algorithms and methodologies for interpretable physical modeling using symbolic regression. He has a bachelor's degree in systems and computer engineering from the "Gheorghe Asachi" Technical University of Iasi, Romania, and a Ph.D. in Software Engineering from the Johannes Kepler University of Linz, Austria. His main research interests are machine learning, symbolic regression, physical-based modeling, and high-performance computing.
Scientific Machine Learning