Code from the paper Active Robot Imitation Learning with Autoencoders and Imagined Rollouts, by Norman Di Palo and Edward Johns. Published at the NeurIPS 2019 Workshop on Robot Learning. robot-learning.ml Code was written by Norman Di Palo, for questions contact normandipalo - gmail - com.

To run the Active Learning algorithm, run

python main.py

The various modules contain different models for policy networks, uncertainty networks, dyanmics networks.

Work in progress!