The Thing that we Tried Didn't Work very Well: Deictic Representation in Reinforcement Learning.

Sarah Finney, Natalia Gardiol, Leslie Pack Kaelbling, Tim Oates

Browse the full UAI paper archive.