Yao Zhang - Scientific Machine Learning, SciMLScientific Machine Learning (SciML) is a rapidly emerging discipline within the data science community, aimed at addressing domain-specific data challenges and deriving insights from scientific datasets through innovative methodologies. By integrating tools from both machine learning and scientific computing, SciML develops scalable, domain-aware, robust, reliable, and interpretable approaches for data analysis and learning. These advancements are poised to fuel the next wave of data-driven scientific breakthroughs, particularly in the physical and engineering sciences. Like scientific computing, SciML is inherently multidisciplinary, drawing upon expertise in applied and computational mathematics, computer science, and the physical sciences to tackle complex problems. Machine Learning
Deep Learning
Reinforcement Learning
Artificial Intelligence in the Sciences and Engineering\(\quad\) It covers deep learning, partial differential equations in science, physics-informed neural networks, operator learning, diffusion models, differentiable physics, hybrid workflows, neural differential equations, JAX, symbolic regression, and machine learning in chemistry and biology. Nonconvex Optimization for Deep Learning
Graphical Models and Causal Inference
Numerical Analyis Meets Maching Learning
\(\quad \) A highly recommended reading is the Handbook of Numerical Analysis, Volume 25: Numerical Analysis Meets Machine Learning. Its content is as follows:
|