Set up leaderplot plots with links to actual runs

[dustinvtran]

@dustinvtran @dusenberrymw has a setup for plotting several metrics, compared across methods, and each method of which is averaged over multiple seeds. The plotting data is obtained directly from the experiments.

This improves our existing leaderboard (a table) by 1. automatically averaging over multiple seeds instead of manually; and 2. programmatically going from experiment runs -> results for visualization purposes.

It would be great to set this up for all datasets. Later, we should also look into how to enable this publically once Tensorboard.dev supports making available all the different experiment runs we'd like.

https://github.com/google/edward2/issues/279