See here for the complete weekly lesson plan.

Assignment 1:

Implement a neural network by following this tutorial. Note that the author’s use of notation is slightly different, and that he has dropped the bias parameter. Discussion points (write your answers in a reply to this thread):

- Use the chain rule to check whether the author’s derivation of gradients are correct ( I suspect there is an error).
- Note that the X and y values are binary (either 0 or 1). Can we estimate y using linear regression? Why and why not?
- We will estimate y using a three layer neural network (1 input layer, 1 hidden layer and 1 output layer). Which loss function we should use?
- Try adding an additional hidden layer - will this improve model performance?

Assignment 2:

Classify the MNIST dataset using a neural network tutorial.

This post is a wiki; add resources as you see fit.