Machine Learning Experiments

Experiments not typical for conceptual insights

---

Gradient Descent

• surrogate gradients : Gradient descent using surrogate gradients applied to create a matrix whose determinant is a given number

The Basic GAN

• A simple experiment to find Nash Equilibrium for a problem of generating a Matrix whose determinant is what we intend to be.

Monte Carlo Tree Search

• A template code to use MCTS to drive decision making process towards a specific objective. That is to make best selection of hyperparameters which maximizes any choosen performance function.