Less is More
See more
…nt to only using Machine Learning algorithms that have a small VC dimension with convex error, while ignoring practical solutions such as Dropout regularization, back propagation and non-convex error f…
