NaN.

CourseDeep LearningLast updated 13 hours ago

Learn Deep Learning: From Beginner to Expert

This course consists of ~20 tutorials to learn deep learning. You can think of this course as a "Free Online Nano Book". You'll learn core deep learning concepts and algorithms, including neural nets, convolutional nets, word embeddings, and even advanced concepts like generative adversarial networks. Tutorials have been written in a way to make learning productive and easy to the extent possible.

This is an updated, more cohesive version of our earlier deep learning course.

Why deep learning?: Deep Learning has caused the revival of artificial intelligence. It has become the dominant method for speech recognition (this allows us to talk to Amazon Alexa and Siri), computer vision (this allows us to search for "my pictures on the beach" on Google Photos), language translation and even game artificial intelligence (think AlphaGo and DeepMind). If you'd like to learn how these systems work and maybe make your own, deep learning is for you.

Structure: In the first section, you'll learn what machine learning is and the main concepts within machine learning. This background in machine learning is required in-order to be able to start learning about neural networks and deep learning, which is a sub-field of machine learning. In the second section, you'll learn what deep learning is, what neural networks look like, and how they are trained. In the following sections, we'll cover some popular neural networks architectures. This includes (a) convolutional neural networks, which have revolutionized computer vision, (b) long short-term memory networks, which have revolutionized natural language processing, and (c) generative adversarial networks, which are the first practical and successful algorithm for generating images.

Expected time to completion: Depends on how frequently you come back, but non-stop reading time is about 4-5 hours. Obviously, it is recommended that after every couple of tutorials you take time out and ask yourself what you learnt, or better yet, try to explain it to someone else.

How to use this course: Each section of this course depends on the previous one, apart from sections 3 and 4 (CNNs and LSTMs), which do not depend on each other and can be read in any order. Section 5 (GANs) depends on sections 1, 2 and 3, but not on section 4.

Pre-requisites: The course assumes some mathematical maturity, mostly basic fluency in algebra and calculus 1 (differentiation and chain rule).

Pro-tip: If you already know machine learning, you can skip tutorials by marking them completed.

An Introduction to Machine Learning

In this section, we'll learn about what machine learning is, and in particular, how it differs from general programming. We'll learn a specific machine learning model, linear regression, and see how the model can be learnt using gradient descent. Gradient descent is also the how neural networks are trained, and in many ways, our linear regression model can be looked at as a very simple neural network.

We'll also learn about different types of problems in machine learning (supervised vs unsupervised), and talk about core concepts such as overfitting and regularization. This section ends with a beautiful visual recap of machine learning.

An Introduction to Deep Learning and Neural Networks

In this section, we'll learn about how deep learning differs from machine learning, i.e. what deep means and why it is important. We'll learn specifically what neural networks look like and how they are trained using back-propagation. We'll also introduce the concept of computational graphs which is how neural networks are implemented in popular deep learning libraries such as TensorFlow, Torch, etc. We'll end the section with learning regularization techniques specific to deep learning, such as early-stopping and dropout.

What is deep learning? How does it relate to machine learning?
5 mins
A First Look at Neural Networks
9 mins
Computational Graphs and Backpropagation
11 mins
Regularization methods in Deep Learning
5 mins
Dropout (neural network regularization)
7 mins

Convolutional Neural Networks and Computer Vision

We'll start by introducing the image classification problem in computer vision and a small note on how images are stored in a computer. Thereafter, we'll describe the architecture of a Convolutional Neural Network (CNN) in substantial detail, and see why the convolutional layer is very well suited for computer vision applications.

Computer Vision tasks: Image classification, localization, etc
3 mins
How do computers see an image?
4 mins
Convolutional Neural Network
18 mins

Long Short-Term Memory Networks and Natural Language Processing

We'll start by introducing Recurrent Neural Networks (RNN), which are neural networks with loops in them. We'll see how RNNs can be used for inputting and outputting sequences and how they maintain an internal state. Then we'll introduce Long Short-Term Memory (LSTM) networks and see how they overcome some significant shortcomings of general RNNs. We'll see how this model can be used for performing state-of-the-art language translation and along the way introduce the concept of word embeddings.

Generative Adversarial Networks and Image Generation

In this section (single tutorial), we'll learn about Generative Adversarial Networks (GANs). GANs are one of the few machine learning techniques which has given good performance for generative tasks, or more broadly unsupervised learning. In particular, they have given splendid performance for a variety of image generation related tasks. We'll introduce the image generation problem, describe the GAN setup and look at the results of applying GANs to image generation.