Learning Disentangled Representations:

from Perception to Control

NIPS 2017 Workshop