This is a collection of resources that either go beyond the basic introductory ML algorithms or go in-depth with theory behind those algorithms. Majority of the courses are recordings of actual classes taught at top universities, rather than typically watered down MOOCs. Since they are taught to advanced undegrad and grad students, the required background for these is strong knowledge of multivariable calculus, linear algebra, probability theory and statistics.
If you don't know where to start, try CS229 - Machine Learning - this is the Andrew Ng's course that the infamous coursera MOOC was based on, but goes much deeper into theory.
- CS294 - Deep Reinforcement Learning
- CS231n - Convolutional Neural Networks for Visual Recognition
- CS224d - Deep Learning for Natural Language Processing
- CS229 - Machine Learning
- Neural Networks for Machine Learning
- Oxford's Machine Learning
- NYU's Deep Learning
- Cambridge Information Theory, Pattern Recognition, and Neural Networks
- MIT Statistical Learning Theory and Applications
- CMU Statistical Machine Learning
- CMU Convex Optimization