Fetching the paper…

Align-RUDDER: Learning From Few Demonstrations by Reward Redistribution · Around