Gradient descent in the real world, a.k.a. lab mini golf

lab mini golf

A great end-of-summer lab outing to Grand Old Golf. We saw examples of setting the learning rate too high, one-shot learning, optimization through quantum tunneling, and even deep networks!