Overview
This is a hands-on class covering artificial neural networks, ranging from historic beginnings and biological inspiration, to modern deep learning. An understanding of basic linear algebra is needed, as well as programming skills. Python will be used.
Slides
- Introduction
- Biological Inspiration
- Universal Approximation
- Perceptron Learning
- Historical Neural Network Architectures
- Gradient Descent and beyond
- Tensorflow
- Keras
- Machine Learning bare essentials
- Convolutional Neural Networks
- Data Augmentation & Generators
- Representations and Features
- Recurrent Neural Networks