I taught a short course of Deep Learning in 2016 Fall and presented 2 lectures of Deep Learning in 2017 Spring. So I want to share some useful materials.

Slides

Below is PPT that I used in lectures:

  1. Introduction to Deep Learning(17/1/11) / Neural Networks Basics(17/6/25)
  2. Supervised Learning and Optimization Tricks(17/1/11) / Neural Networks Learning(17/6/25)
  3. Varients of Neuron
  4. Varients of Connection
  5. General View

Images and references are all contained in folder images and ref.

Because of short of time, I didn’t make PPTs of ‘Unsupervised Learning and Regularization’ and other stuff in plan. But I did cover some of these contents in the last lecture. Maybe one day I will upload such missing PPTs.

Textbook

As for which book should be textbook, I recommand Deep Learning, written by I. Goodfellow, Y. Bengio and A. Courville. Its contents are mostly driven by intuition and contain only necessary mathematics. It is also a nice practical guide although its guidances are scattered through the book.

Here is an annotated PDF version of this textbook, which contains about 1200 comments, including notes, highlights of errors and etc.

Since Deep Learning is a very active research field, the larger the gap between the day you read this textbook and it published became, the more inaccurate future trend pointed out by this textbook is. As a supplement, you should read a most recent review or survey of Deep Learning.

Appendix

There are many resources of Deep Learning, including courses, vedios, books and etc. However, most of them either are only focusing on basic concepts without mentions of current research trend, or contained too much contents so that readers cannot have a clear understanding of this field and are out of date rapidly.

Because of situations above I am managing to write a online book Deep Learning Short Course in moderate size so as to provide a clear view but still not lost complexity. Since I have a lot of work to do, finishing it might be a long process.