Created for Machine Learning class at WPI.
Implemented a neural network from scratch that trains on the mnist data set.
Created a hyperparamter optimization algorithm that optimizes each hyperparamter individually and combines then together.
Blue surface is the training waits from an inverse transform of 2 value PCA.
Red dots are the approximated neural network weights array.
The Z dimension is the cross entropy loss.
