Goal: This module introduce students to the basic concepts of neural networks taking a do it by yourself approach. More specifically the students will learn how a neural network is a structured as a collection of perceptron forming multiple layers. They will also learn how these systems can be trained using backpropagation and gradient descent techniques. Finally, they will have the opportunity to implement their own neural network in python without the need of particular libraries. Moreover, the will put into practice these techniques by implementing their first patter recognition
Assignment: Students will be evaluated as a combination of their final examination that will cover the theory and the result of the group assignment. The assignment consist in the development of a pattern recognition using a neural network library that they will implement by themselves. Groups of maximum four students will be formed during the first week of the module
Resources: Material that will be used for the classes includes:
- Michael Nielsen, Neural Networks and Deep Learning http://neuralnetworksanddeeplearning.com/
- THE MNIST DATABASE of handwritten digits http://yann.lecun.com/exdb/mnist/
- Additional material is available in the form of articles at the following: https://francescolelli.info/
This is a short overview of the work that will be done each week: