Resources for those that want to skip ahead

[jeremy (Jeremy Howard (Admin))]

I know that some folks in the class have some prior experience with deep learning, and/or are putting in 40+ hours a week into this course - I want to ensure that you get as much out of the course as possible, so I figured I’d start a thread with resources and ideas for those that want to skip ahead faster and/or deeper than the lessons.

Here’s a great resource to start off the discussion with - a well-curated list of deep learning papers in a good order for self-study: https://github.com/songrotek/Deep-Learning-Papers-Reading-Roadmap . Also, those interested in understanding the math at a deep level, the Deep Learning Book is a good resource.

[jeremy (Jeremy Howard (Admin))]

Some beginner-intermediate level resources:

• Neural Networks and Deep Learning - an excellent book with some really nice interactive pieces. Much less math intensive that the book I mentioned in the previous post, although still more math-focused than code-focused
• Convolutional Neural Networks for Visual Recognition (Stanford) - truly exceptional presentation of everything you need to know about CNNs. I couldn’t recommend it more highly
• Hinton’s Coursera course - a bit outdated, but still excellent. Unfortunately the assignments are in Matlab/Octave

Advanced resources:

• University of Oxford Deep Learning - presents advanced concepts in a way that I, frankly, found unnecessarily complex and math-heavy. But it’s the only course (so far) that includes cutting edge recent advances in deep learning