Deep Learning in an Hour, Day, Season, or Decade

Whether we like it or not, deep learning is eating the world. Several weeks ago, I decided to learn how it works. It’s the Wild West out here. Everyone recommends different resources and makes different assumptions about how much time you have.

I looked at a bunch and removed some guesswork on both fronts: These resources are simply the best, better than all the rest, and maybe more importantly, they give you reliable options whether you want your journey to be 60 minutes or 6 years.

I’ve included some details on prerequisites at the end. Regardless of background, everyone should check out the first link. 3Blue1Brown is a legend.

1. Hour

• Neural networks playlist, 3Blue1Brown

• Deep Learning with PyTorch: A 60 Minute Blitz, Soumith Chintala (PyTorch documentation)

2. Day

• Learn PyTorch for deep learning in a day, Daniel Bourke

• I tried to learn as much deep learning math as I could in 24 hours, Phosphorus (Less Wrong blog)

3. Season

• Books

  • Deep Learning, Ian Goodfellow, Yoshua Bengio, and Aaron Courville

  • Deep Learning with Python, François Chollet

  • Neural Networks and Deep Learning, Michael Nielsen

  • Machine Learning with PyTorch and Scikit-Learn, Sebastian Raschka

    
    

• Courses
  • course.fast.ai, Jeremy Howard
  • deeplearning.ai, Andrew Ng & others
  • Elements of AI, University of Helsinki
  • Dive Into Deep Learning, Aston Zhang, Zachary C. Lipton, Mu Li, and Alexander J. Smola
  • Neural Networks: Zero to Hero, Andrej Karpathy
  • Learn Machine Learning in 3 Months, Siraj Raval

4. Decade

• Go to grad school

• Roadmaps

  • Machine Learning Roadmap, Daniel Bourke

  • Complete Roadmap to be a Deep Learning Engineer, Let the Data Confess

    
    

• Resource lists

  • Awesome Deep Learning, ChristosChristofidis

  • Crème de la crème of AI courses, SkalskiP

  • Deep Learning.md, brylevkirill

    
    

• Miscellaneous
  • Papers With Code, Meta AI Research
  • Two Minute Papers (Youtube Channel), Károly Zsolnai-Fehér
  • Sentdex (Youtube Channel), Harrison Kinsley

A Note on Prerequisites

It’s a free country. Learn what you want along the way. All the resources above are fairly clear about what’s needed.

That said, if you want to embark on a serious course of study, here are a few central building blocks they all share that you’ll need to be comfortable with.

• Manipulating vectors and matrices
• Finding derivatives
• Gradient descent
• Basic Python programming

It’ll also help to know:

• How to learn

• How to prompt tools like ChatGPT effectively, for broad explanations.

• How other kinds of advanced machine learning besides deep learning work, if you want to become a domain expert one day.

Happy learning 💡