Description

Optimization for Machine Learning
Theoretical Exercises
Solve Exercises 38 and 39 from the lecture notes.
Practical Exercises
EPFL
Martin Jaggi & Nicolas Flammarion github.com/epfml/OptML course

The theory of non-convex optimization is unfortunately not very illuminative. However, their practical performance is usually unmatched by convex methods. In this exercise, we will use the PyTorch framework to train a small neural network on some simple datasets.
Problem 1 (PyTorch Refresher):
If you run notebooks from your own computer, install PyTorch following the instructions on
pytorch.org
We recommend using the following online tutorial:
pytorch.org/tutorials/beginner/pytorch with examples.html
You can optionally also look at the following exercise from the Machine Learning course:
https://github.com/epfml/ML course/blob/master/labs/ex12/
Problem 2 (Simple Neural Network):
Follow the notebook provided here:
colab.research.google.com/github/epfml/OptML course/blob/master/labs/ex07/template/Lab 7.ipynb