Computational Methods for Geological Engineers
In this course we will integrate physical simulations with mathematical concepts and computational simulations. See the presentations below
Some additional resources
- Notes and presentations in this page
- Matlab programming
- Scientific computing book
Computational Methods for Machine Learning
In this course we will discuss the use of modern computational techniques and their application to machine learning.
Notes are online (but are subject to change) at
Material that is covered
1. Semisupervised learning and graphical models Algorithms that are based on the graph Laplacian, eigenvalue problems, sparse matrix techniques.
2. Linear and single layer neural networks Numerical optimization, descent methods, Gradient and stochastic gradient descent, Gauss Newton and Newton's method, inexact Newton and Hessian free methods.
3. Deep Neural Networks and their connections to ODE's. Numerical solution of ODE's, back propagation and the adjoint
4. CNN's and numerical methods for PDE's
5. Applications: Classification, segmentation, stylization
The course will involve extensive coding using PyTorch.
- Homework 50%
- Final project presentation 20%
- Final project 30%
- Students that will make the effort and program in Julia will get 5%
Some files to upload
Shot course - Computational methods for AI
2. Unsupervised and semisupervised learning
3. Deep Neural Networks
4. Convolutional Neural Networks