These notes are things that I am currently learning. Many of the notes are handwritten, and in the future I plan to type up some of them up. Note that in many places I will sacrifice clarity for rigor :)

Reinforcement Learning

Derivation of Bellman’s Equation

Banach Fixed Point Theorem, Iterative Policy Evaluation

Depth First Learning 1: Policy Gradients

RL: Missing Theory

Machine Learning

Deriving the loss function for softmax regression

Some backprop notes

Real and Convex Analysis

Three hard theorems

Convex sets

Distance functions

Convex separation

Convex functions

Normal cones

Subgradients

Optimal control

Optimal control problems

Robotics and Dynamics

Swing-up and Balance Control for the Simple Pendulum

Derivation of EOM for the Double Pendulum

Derivation of EOM for the Inverted Pendulum (Cartpole)

Derivation of EOM for the Dual Inverted Pendulum

Derivation of EOM for the Wheeled Inverted Pendulum

Talks

Is (Deep) Reinforcement Learning Barking Up The Wrong Tree

Intrinsically motivated reinforcement learning