Introduction #

Introduction to Machine Learning and Python - required to complete the course.

Pre-read (Required) #

• https://www.kaggle.com/learn/python (Lessons 1-6)
• https://www.kaggle.com/learn/intro-to-deep-learning (Lessons 1-3)
• https://www.kaggle.com/learn/computer-vision (Lessons 1-4)

Literature (Optional) #

• Christopher M. Bishop. (2007). Pattern Recognition and Machine Learning
• Sebastian Raschka (2019). Python Machine Learning: Machine Learning and Deep Learning with Python
• John M. Zelle, Guido van Rossum. (2016). Python Programming: An Introduction to Computer Science

Demo project #

This is a handwritten digit recognizer created during the class. Draw a digit and click on “predict”.

Class content #

• Introduction to the Google Colab environment, characteristics of working in an interactive environment
• Introduction to Python. The basics required to complete the course
• Introduction to Machine Learning (mean square error and gradient descent)
• Construction of a simple neural network using the TensorFlow library
• Introduction of the concepts of accuracy, confusion matrix and optimizer
• Implementation of a neural network (multilayer perceptron) on the MNIST dataset (preparing a dataset, dividing data into a test and training set, intro to loss functions, graph analysis)
• Implementation of the convolutional network for the same dataset. Interactive overview of convolutional layers, max pooling, activation and fully connected (dense)