Overhaul to MapObjective

[dannys4]

Currently MapObjective is a nice idea, but has some flaws

• KLObjective is not really measuring the KL, it's off by a constant. Maybe we just disregard this and call it a quirk.
• In practice, we need regularization. It would be really cool to have another abstract class like RegularizedObjective::MapObjective, where we create an abstract class RegularizerBase with e.g. child class L2Regularizer(lambda). Then, we could do something like RegularizedObjective obj = KLObjective(train, test, dim) + L2Regularizer(0.05).

[dannys4]

Additionally, I think we should extricate the data samples from the Objective. I believe that the best code would be

MapObjective obj = KLObjective() + MapSamples(train, test, dim) + L2Regularizer(0.05)

so then we could easily do a map from density objective via

MapObjective obj = KLObjective() + LogDensity(density_fcn) + L2Regularizer(0.05)

Please let me know any thoughts