Implement Adam as a LearningRule

[lamblin]

Citation: Adam: A Method for Stochastic Optimization. Diederik Kingma, Jimmy Ba. http://arxiv.org/abs/1412.6980
Implementation in Theano by Alec Radford (@Newmu): https://gist.github.com/Newmu/acb738767acb4788bac3

[JesseLivezey]

Is someone already working on this? If not, I might try it out next week.

I'd like to request this please 👍

[JesseLivezey]

Starting working on this in #1425.