Inverse KKT - Learning Cost Functions of Manipulation Tasks from Demonstrations.

Peter Englert, Marc Toussaint

Browse the full ISRR paper archive.