Write up one-page description of Kernel Density Networks in LaTeX

[levinwil]

1. Describe Kernel Density Forests in related works (make sure to mention having a separate voting set)
2. Show some simulation of KDF performance and report their Brier Scores/ECE
3. Describe the algorithm to find polytope membership
4. Show some polytope composition networks for simple networks and then for more complicated networks
5. Describe treating the polytopes of networks as leaves and using the KDF algorithm for those polytopes/leaves (including training the network with a transformer set and filling in the polytope/leaf KDE's with a vote set)
6. Describe simplicity of implementation and convenient utilization of GPU
7. Show estimated posteriors of KDN on XOR and compare with Deep Networks (make sure to emphasize Deep Networks becoming more confident as distance from training data increases, failing to conform to the shape of the data)
8. Show some class accuracy and Brier score results on XOR and compare with Deep Nets
9. Show some class accuracy and Brier score results on MNIST and compare with Deep Nets

[levinwil]

This was silly in that it jumped way too many steps ahead. Closing.