Artificial neural networks have come to dominate machine learning over the last decade, enabling substantial advances in areas such as computer vision, robotics, speech recognition and natural language processing.  In this tutorial we provide an introduction to Artificial Neural Networks, covering fundamental concepts such as gradient descent, back-propagation, multi-layer perceptrons, auto-encoders, convolutional networks, recurrent networks, and Transformers. Concepts will be developed in a roughly historical order, with reference to historical context and enabling technologies. We will review network architectures, as well as techniques for training and performance evaluation, including methods for supervised, semi-supervised and self-supervised learning.

This tutorial is based on course material taught to 4th and 5th year engineering students at Grenoble Institut Polytechnique. While this is an introductory course, students will be assumed to have a basic understanding of linear algebra, basic calculus, probability and statistics.