We are back to reinforcement learning and we'll start discussing sampling one distribution using another distribution - https://drive.google.com/file/d/1Y0dFZZhl8hqXS1u9nvU__fJGtrjA4yY6/view?usp=sharing

Nahum composed a MDP process that should help clarify our re-reinforcement learning discussion.  See https://www.dropbox.com/sh/xrnv132ep0n9en2/AADX8gDC22FONqVfvRd1eRG2a?dl=0 for a html and notebook versions.   Enjoy.   

Will revisit convex optimization and see an interesting generalization of the deviate.   
https://drive.google.com/file/d/1HGDj_BEi5XBGHENeQ7ac77uNzlDNDmq0/view?usp=sharing

We'll pose with reinforcement learning and learn about the relation between KNN and regression -

https://drive.google.com/file/d/1K3axkCXQbc_k0x7tcu6MjJ0JYnI1mYLz/view?usp=sharing

Interesting automated ML blog entry -
https://towardsdatascience.com/tpot-automated-machine-learning-in-python-4c063b3e5de9