Learning path to education in Machine Learning

An unconventional learning path to go from a total beginner in machine learning to a professional A.I engineer

Contents

• About
• How to use this guide
• Motivation
• Curriculum

About

This is a path for those who is interested in learning Machine Learning/Deep Learning/AI course on your own time, for free. The list comprises of videos, articles, courses, lectures, books etc.

How to use this guide (Daily Plan)

The intention of this path is not to conclude everything on this list. Each subject does not require a whole day to be able to understand it fully, and can do multiple of these in a day. Each day I take one subject from the list below, read it cover to cover, take notes, do the exercises and write an implementation in Python. The list is to guide me on the things I can focus on in the coming weeks. I am expected to spend at least 3 hours very day (to cover 10 - 30 hours per week). That is what is important to me, progress daily.

Motivation

Here is an interesting link the Twitter page of an AI community in Nigeria, doing great things. I am inspired by their works.

• AI Saturday Lagos

Curriculum

PREREQUISITES

• 3 blue 1 brown linear algebra YouTube playlist.
• Khan academy course.
• 3 blue 1 brown calculus YouTube playlist.
• Edx introduction to Probability - The science of uncertainty.
• Edx algorithm design and analysis(Python)

MONTH 1

Week 1

• Learn python for data science by Siraj Raval on YouTube.
• The math of intelligence by Siraj Raval on YouTube.
• Intro to Tensorflow by Siraj Raval
• www.explained.ai
• Google ML Crash Course

Week 2

• Introduction to ML by Udacity.
• aischool.microsoft.com
• www.analyticsvidhya.com

Week 3&4

• ML project Ideas: Awesome project ideas on GitHub by NirantK
• Convolutional neural networks for visual recognition (cs231n.github.io)

• Simple practical course on Tensorflow from Kadenze
• Tensorflow cookbook
• Tensorflow-101 tutorial set
• IBM Code Patterns
• Code patterns from IBM which also includes Data Science & Analytics

• Fast Style Transfer Network
• This will show how you can use neural network to transfer styles from famous paintings to any photo.
• Image segmentation
• Object detection with SSD
• One of the fastest (and also simpler) models for object detection.
• Fast Mask RCNN for object detection and segmentation

MONTH 2

Week 1

• Intro to deep learning on Udacity
• Week 1 - Feedforward Neural Networks and Backpropagation
• Read Part I of the Deep Learning Book found here
• Use this cheat sheet to help understand any math notation, found here
• Watch Build a Neural Net in 4 Minutes
• Read Neural Net in 11 lines
• Type out the neural network code yourself in a text editor, compile, and run it locally (using no ML libraries)
• Watch Backpropagation in 5 minutes

Week 2

• Theories of Deep Learning (stats385.github.io/readings)
• Youtube : Image Processing Group - UPC/BarcelonaTech

Week 3&4

• Deep learning by Fast.ai

MONTH 3

Week 1

• Youtube: Computer Vision Foundation (CVF)
• Convolutional Networks
• Watch the Convolutional Networks Specialization on Coursera, found here.
• Read all 3 lecture notes under Module 2 for Karpathy CNN course found here
• Watch my video on CNNs here and here
• Write out a simple CNN yourself (using no ML libraries)
• Recurrent Networks
• Watch the Sequence Models Specialization on Coursera, found here
• Watch my videos on recurrent networks, here, here, and here
• Read Trask's blogpost on LSTM RNNs found here
• Write out a simple RNN yourself (using no ML libraries)

Week 2

• Tooling

• Watch CS20 (Tensorflow for DL research). Slides are here. Playlist is here
• Watch my intro to tensorflow playlist here
• Read Keras Example code to quickly understand its structure here
• Learn which GPU provider is best for you here
• Write out a simple image classifier using Tensorflow
• Generative Adversarial Network
• Watch the first 7 videos you see here
• Build a GAN using no ML libraries
• Build a GAN using tensorflow
• Read this to understand the math of GANs, but don't worry if you don’t understand it all. This is the bleeding edge here

Week 3

• Deep Reinforcement Learning
• Watch CS 294 here
• Build a Deep Q Network using Tensorflow
• Reinforcement learning

Week 4

• Deep learning project of Siraj reimplementation.

• www.cv-tricks.com

• Very useful thing especially if you want to build a robot or the next Dota AI :)

• Magenta project from Google Brain team : - Project that aims to creating compelling art and music with the help of neural networks. And the results are remarkable.

• Deep Bilateral Learning for Real-Time Image Enhancement

• New awesome algorithm of the photo enhancement from Google

• Self driving-car project
• Want to make your car fully automatic? — that’s a good starting point.

MONTH 4

• Fill the gap on deep learning with “Deep Learning Book” by Ian goodfellow and Blog posts and "Yearning for A.I" by Andrew NG