Third person imitation learning project for CS101 at Caltech
The goal is to learn from demonstrated states from an "expert" policy. In practice, we inted to capture states from successfully learned policies from reinforcement learning algorithms, then learn new networks using those states as input and compare the learning curves.