Learning Path for Applied ML/DL

gistfile1.md

LINEAR ALGEBRA

• Chapter 2 of Deep Learning book (by Ian Goodfellow, short path) Link

• Lecture series on Linear Algebra by three blue one brown (Highly recommended) Link + singular value decomposition (shortpath).

• And if you have sufficient time, then highly recommended to take “Introduction to Linear Algebra” by Gilbert Strang (on YouTube, long path) Link.

PROBABILITY

• Statistics and probability by Khan academy (short and highly recommended path)Link,

• STATS 110 by Joe Blitzstein, Harvard (long path)Link

Information Theory

• Best way to learn is to do Google search like this ( “what is intuition behind X” ) ( Try to focus on link between KL Divergence and Entropy)

Basic Machine Learning

• Andrew Ng’s coursera

• Chapter 5 (ML basics) Deep Learning book Link

• CS 109 by Harvard Link

• Data science group (IITR) blog posts on basic ML technique.Link

If you don't have any time constraints then follow step 3 otherwise step 1,2,4.

After completing this do some kaggle problems and get familiar yourself with basic ML implementation. You can start with this awesome by Sebastian Raschka [Link] (https://github.com/rasbt/python-machine-learning-book)

BASIC ML to DL:

• It is very important to understand why basic ML techniques failed on high dimension inputs and about Representation Learning. This is must for your journey in Deep Learning.

• Section 5.11 challenges motivating Deep Learning and chapter 1 from deep learning book

• Representation Learning : A review and new perspectives by Yoshua Bengio (You can leave Probabilistic models and Auto Encoder as of now)

COMPUTER VISION

• Chapter 6, 7, 8, 9 in Deep Learning book Link

• CS231n Convolutional Neural Networks for Visual Recognition (2016 version + generative adversarial networks + deep reinforcement learning from 2017 version)Link

Natural language processing:

• Chapter 1, 2 of Kyunghyun Cho’s lecture notes at NYU (highly recommended - first read this)Link,

• CS224n Deep Learning for NLP Link

Reinforcement Learning :

• It is recommended to first get the idea of convolution neural networks. Later, start with

• Chapter 1 - Sutton's book on Reinforcement Learning link,

• “Introduction to Reinforcement Learning” by David silver (person behind alphago) Link,

• CS294 - Deep Reinforcement Learning (little bit Advanced course) Link.

If you find this is as sub-optimal policy for your agent on your deep learning quest or this as insufficient data to predict your path and want to find optimal policy then please follow FAQ’s on Artificial Intelligence and Deep Learning group (AIDL) on Facebook.