Cheers to Dopamine Driven Random Walks

Whats the best and most enjoyable way to learn ?

Learning Representations by back-propagating errors

The backpropagation paper