Meta-Guide to Mathematical foundations of ML

This is a guide of guides of what's available over the internet. Machine Learning might not be as hard as we once thought..
This list curates somewhat less popular resources on the internet, with a focus on the math background. There are so many other videos, MOCCs that are good and accessible. Chris Olah's blog is still my favourite so far

Conceptual Understanding

What we want to acheive:
Learning = Representation + Optimzation
https://medium.com/@devnag/machine-learning-representation-evaluation-optimization-fc7b26b38fdb
Manifold and Topology
http://colah.github.io/posts/2014-03-NN-Manifolds-Topology/
KL Divergence
https://www.reddit.com/r/MachineLearning/comments/6kohsi/geometric_interpretation_of_kl_divergence/
Quora
https://www.quora.com/What-are-the-best-ways-to-pick-up-Deep-Learning-skills-as-an-engineer

Information Theory

Visualised
http://colah.github.io/posts/2015-09-Visual-Information/
Probability and Info T. from Deep Learning Book
http://www.deeplearningbook.org/contents/prob.html
Reddit ftw
https://www.reddit.com/r/MachineLearning/comments/6aebk2/p_kullbackleibler_divergence_explained

NN

Calculus on Backpropagation
http://colah.github.io/posts/2015-08-Backprop/
Calculus on CNN (credits to Anthony)
http://www.jefkine.com/general/2016/09/05/backpropagation-in-convolutional-neural-networks/
Complicated RNN/LSTM
https://distill.pub/2016/augmented-rnns/

Research Direction

From OpenAI
https://openai.com/requests-for-research/

If you have more to recommend, please let me know!

cover photo from OpenAi