Hey, I'm Gene!
Fort Wayne, Indiana
For example, vanilla behavioral cloning on MakeWaterfall leads to an agent that strikes close to waterfalls however doesnt create waterfalls of its personal, presumably because the place waterfall motion is such a tiny fraction of the actions in the demonstrations. For instance, with behavioral cloning (BC), we might perform hyperparameter tuning to reduce the BC loss. Alice is successfully tuning her algorithm to the take a look at, in a manner that wouldnt generalize to sensible tasks, and so the 20% increase is illusory. The problem with Alices approach is that she wouldnt be able to make use of this technique in an actual-world activity, because in that case she cant merely check how a lot reward the agent gets - there isnt a reward operate to check! Simply accessible experts. Domain experts can often be consulted when an AI agent is constructed for real-world deployment. For instance, the net-VISA system used for international seismic monitoring was built with related domain knowledge offered by geophysicists.