At some point, every machine learner should implement a network that learns XOR. The folder LearnXOR is my implementation from the early days.

To see a live one-shot learning demo with a simple laptop and webcam, take a look at Demo.

The table data structure in the Lua language is quite versatile. In Decision-Trees, I use Lua tables alone to construct decision trees and random forests from arbitrary numerical and categorial datasets, such as MNIST or iris.txt.